I’

MICHIGAN STATE UNIV

1. ll n lljllllllflljlll

 

 

ll

3193015

 

 

 

 

 

 

 

 

 

 

 

.1
31:, iii

 

 

 

 

 

 

 

This is to certify that the

dissertation entitled

Novel Methodologies to Study Cysteine Status
in Proteins by Chemical Cleavage and Mass—
Mapping by MALDI-TOF Mass Spectrometry

presented by

Jiang Wu

has been accepted towards fulfillment
of the requirements for

Ph. D. degreein Chemistry

We“

 

V Major professor \
Date May 16, 1997
MSUicnnAfI-Irmnliim‘ ‘ '1, m” ‘ I ' ‘ 0-12771

 

 

) A
'v—v '_v

‘
fivf‘vv

 

4 4'.

d“

0' * «8 04‘444—~

 

LIBRARY
Michigan State
University

 

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
‘ TO AVOID FINES return on or before date due. \.-.-

DAME DATE DUE DATE DUE '\

DEC 2 81998i

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU Is An Affirmative Action/Equal Opportunity Institution
clclmmpma-DJ

 

NOVEL METHODOLOGIES TO STUDY CYSTEINE STATUS IN
PROTEINS BY CHEMICAL CLEAVAGE AND MASS—MAPPING BY
MALDI-TOF MASS SPECTROMETRY

By

Jiang Wu

‘Ii
”‘1

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Chemistry

1997 I

 

ABSTRACT
NOVEL METHODOLOGIES TO STUDY CYSTEINE STATUS IN PROTEINS BY
CHEMICAL CLEAVAGE AND MAS S-MAPPING BY MALDI-TOF MASS
SPECTROMETRY
By

J iang Wu

Current methodology for characterizing free sulfhydryls and disulfide bonds in
proteins involves the modification of cysteine residues, enzymatic digestion of protein
chain for cleavage between half cystinyl residues, I-IPLC fractionation, and sequence
determination of peptide fragments. This strategy is tedious, cumbersome, and can
promote disulfide exchange.

A simple methodology has been developed to localize free cysteine groups in
peptides and proteins. This new approach employs a specific reaction between free
sulfliydryls and 2-nitro-5-thiocyanobenzoic acid (NTCB) or 1-cyano-4-dimethy1amino-
pyridinium tetrafluoroborate (CDAP) to specifically cyanylate cysteine thiols. The N-
terminal peptide bond of the cyanylated cysteinyl residue can then be cleaved under
alkaline conditions to form an amino-terminal peptide and a series of 2-iminothiazolidine-
4-carboxylyl peptides which can be mapped to the sequence by MALDI-MS. The
cleavage conditions have been systematically investigated using a number of peptides
containing different amino acids adjacent to the N-terminal side of cysteine residues.
While the cleavage reaction was traditionally performed in pH 8.5-9.5 buffer at 37-65°C

for 12-80 hours, optimal results have been obtained in 1M ammonium solution in which

 

 

 

the cleavage is complete within an hour at ambient temperature. This improvement also
minimizes side reactions resulting from prolonged exposure to the alkaline medium.

A novel strategy is described for assignment of disulfide pairings in proteins. A
denatured protein is subjected to limited reduction in acidic solution to produce a mixture
of partially reduced protein isomers; the nascent sulfhydryls are immediately cyanylated by
CDAP under the same buffered conditions. The cyanylated protein isomers, separated by
and collected from reversed-phase HPLC, are subjected to the cleavage reaction to form
truncated peptides, which, after further reduction of the remaining disulfide bonds, can be
mass-mapped by MALDI-MS. This simple, fast, and sensitive strategy avoids disulfide
scrambling and is applicable to disulfide characterization in proteins containing adjacent
cysteines. Several proteins with various disulfide structures have been studied to
demonstrate the feasibility of the methodology.

Two techniques are introduced for the study of disulfide structures of protein
folding intermediates. The cyanylation of thiol groups by the CDAP under acidic
conditions has been applied to trapping folding intermediates of recombinant human
epidermal grth factor (hEGF). Disulfide structures of seven CDAP-trapped
intermediates have been identified by the approach described above. Both native and non-
native intermediates were found in the folding process. The analysis of the disulfide

intermediates provides new insight into the folding pathway of hEGF.

 

 

 

 

 

ACI

Iwould like to express my 51
Watson for his guidance. encourage
study at MSL’. Iwould also like to II
and support, and for sening as a sec

Eugene LeGoii for sening on my

I

COl

tram to acknowledge Drs l
knowledge in protein chemistry an
members, past and present. thank yo

mani'lhanks to other members of the

Bti’, Mel. Without their technical sup

And last, but not the least, I
daUShier, Mengyni, for their love, sup;

80 very gratefiil to my parents moth:

their '
calmg, encouragement, and und

State University.

ACKNOWLEDGIVIENTS

Iwould like to express my sincere gratitude to my research advisor, Dr. J. Throck
Watson, for his guidance, encouragement, and financial support throughout my graduate
study at MSU. I would also like to thank Dr. Doug Gage for his continual encouragement
and support, and for serving as a second reader of my dissertation; Drs. John Allison and
Eugene LeGoff for serving on my committee.

I want to acknowledge Drs. Eugene Zaluzec, and Zhi H. Huang for sharing their
knowledge in protein chemistry and organic chemistry. To all the Watson group
members, past and present, thank you for your friendship and cooperation. I also owe

many thanks to other members of the Mass Spectrometry Facility: Melinda, Paochi, Gary,
Bev, Mel. Without their technical support, my graduate study would have been difficult.
And last, but not the least, I would like to thank my wife, Ping Gong, and my
daughter, Mengyu, for their love, support, and patience that make my life enjoyable. I am
also very grateful to my parents, mother-in—law, brothers and sister, and sister-in-laws, for

their caring, encouragement, and understanding during my graduate study at Michigan

State University.

 

 

 

 

ll Chemj

TAE

List of Tables. . .

List ofFigureS , . -

Chapter 1. Introduction

1. introduction _. .

ll. Characterization of Cysteine Sun
A Quantitative Determination of
B. Localization of Sulihydryl Grc
C. Disulfide Bond Assignment

Ill. innit-Assisted laser Desorption
(MALDI-MS)

n The Matrix ....... _

B. Sample Preparation.
C Instrumentation.....
D. Characterization Features of M
E. New Techniques in MALDl-TC

F. Application of MALDl-TOF M
W. References ..........

Cha t
Per 2. A Strategy to Locate Cystei

Specific Chemical Cleavage

byMALDi-TOF MS ..
1. Introduction...

 

TABLE OF CONTENTS

List of Tables ......................................................................................... viii
List of Figures .......................................................................................... x
Chapter I. Introduction .............................................................................. l
I. Introduction ........................................................................................ 1
II. Characterization of Cysteine Status in Proteins ............................................... 3
A. Quantitative Determination of SH and 8-8 Groups in Proteins ........................ 5

B. Localization of Sulfhydryl Groups in Proteins ........................................... 16

C. Disulfide Bond Assignment ................................................................ 21

III. Matrix-Assisted Laser Desorption/Ionization Mass Spectrometry

(MALDI-MS) ................................................................................... 33

A. The Matrix ................................................................................... 35

B. Sample Preparation ......................................................................... 36

C. Instrumentation ............................................................................. 38

D. Characterization Features of MALDI-TOF MS ........................................ 40

E. New Techniques in MALDI-TOF MS ................................................... 41

F. Application of MALDI-TOF MS ......................................................... 43

IV. References ........................................................................................ 46

Chapter 2. A Strategy to Locate Cysteine Residues in Proteins by
Specific Chemical Cleavage Followed by Mass Mapping

by MALDI-TOF MS ................................................................. 57

I. Introduction ..................................................................................... 57
II. Chemical Cleavage at Cysteine Residues .................................................... 58
III. Experimental Section ........................................................................... 65
IV. Results and Discussion ........................................................................ 70

 

 

A. Cyanylation and Cleavag
B. Effects of pH and Tempe
C. Identification of Free SH
D. Choice ofReducrng Ager
E. Stability of Disulfide Bon
F. Analysis by MALDI-h-lS

G. Reactions of Proteins imn

l’. Conclusions ._,

l’l References. . . .

Chapteri. Further Study on the Lee
Residues by Chemical C it

Optimization and Applica
i. introduction ..... .

H. Materials and Methods .

ill. Results and Discussion.

A. Cyanylation of SH groups l
B. Optimization of Cleavage F
C. Effects of Amino Acid Stru

B-elimination Reactions ._

D. Application of O
W. Conclusions... . . . .. .

V. Future Work ........
VI. References .........

ptimized N

Chiliter 4.

I u
.....
....

.....
-----

A. Cyanylation and Cleavage of SH Groups by NTCB ............................... 70
B. Effects of pH and Temperature on Cleavage Kinetics... . ........................ 75
C. Identification of Free SH and Total Cysteines in Proteins ........................ 76
D. Choice of Reducing Agent ............................................................ 89
E. Stability of Disulfide Bonds during NTCB Reaction .............................. 90
F. Analysis by MALDI-MS .............................................................. 92
G. Reactions of Proteins immobilized on a Zetabind Membrane .................... 94
V. Conclusions ....................................................................................... 99
VI. References ...................................................................................... 100

Chapter 3. Further Study on the Localization of Protein Cysteine
Residues by Chemical Cleavage and MALDI—MS:

Optimization and Application ...... . .................................... , .......... 105

I. Introduction ..................................................................................... 105
II. Materials and Methods ................................. . ...................................... 107
III. Results and Discussion ........................................................................ 110
A. Cyanylation of SH groups by CDAP ............................................... 110

B. Optimization of Cleavage Reactions ................................................ 113

C. Effects of Amino Acid Structures on Cleavage and

tit-elimination Reactions .............................................................. 125

D. Application of Optimized Methodology ............................................ 127

IV. Conclusions ..................................................................................... 134
V. Future Work .................................................................................... 135
VI. References ................ . ..................................................................... 139

Chapter 4. A Novel Strategy for Assignment of Disulfide Bond

Pairings in Proteins ................................................................... 141

I. Introduction .................................................................................... 141
11, Our Novel Strategy for Disuiftde Assignment ............................................. 144
III. Experimental Section ......................................................................... 148

vi

 

 

IV. Results and Discussion
A. Partial Reduction of Prot
B. Cyanylation of Nascent S
C. HPLC Separation of Part
D. Cleavage of Peptide C hai
E. Interpretation of MALDI
F. Application to Proteins C

Adjacent Cysteine Residu

V. Pioomole Scale Reaction

VT. Characteristics of the MCLDOdOlOl

Vii. Conclusions, .. ..

ill]. References. . . . . . .

Chapter 5. Disrlfide Mapping of Folc
Epidermal Growth Factor

Cleavage, and Mass Spect
I Introduction ..... .

ii. Experimental Section. ,.

ill. Results and Discussion. . .

A Refolding of hEGF under C

B. Trapping of Folding Intenn

C. Separation of Folding Inten

D. Disulfide Mapping of Well I
- Aspects for Further I
- Conclusions ................

Vi. References

 

IV. Results and Discussion ........................................................................ 151
A. Partial Reduction of Proteins ....................................................... 15.1
B. Cyanylation of Nascent Sulthydryls ................................................ 155
C. HPLC Separation of Partially Reduced and Cyanylated Protein Isomers. . . ..156
D. Cleavage of Peptide Chains .......................................................... 158
E. Interpretation of MALDI Data ...................................................... 159
F. Application to Proteins Containing Closely Spaced and
Adjacent Cysteine Residues ......................................................... 176
V. Picomole Scale Reaction ...................................................................... 194
VI. Characteristics of the Methodology ................................ . ........................ 196
VII. Conclusions ................................................................................... 198
VIII. References .................................................................................... 199

Chapter 5. Disulfide Mapping of Folding Intermediates of Human
Epidermal Growth Factor by CDAP Trapping, Chemical

Cleavage, and Mass Spectrometry ................................................. 201
I. Introduction ..................................................................................... 201
11. Experimental Section ......................................................................... 21 0
III. Results and Discussion ........................................................................ 214
A. Refolding of hEGF under Controlled Conditions ................................ 214

B. Trapping of Folding Intermediates ................................................ 216

C, Separation of Folding Intermediates by HPLC ................................... 217
D. Disulfide Mapping of Well Populated Intermediates ............................. 222
IV, Aspects for Further Improvement ........................................................... 240
V. Conclusions .................................................................................... 242
VI. References ...................................................................................... 244

vii

 

 

 

 

TableZl Calculated "”2 “11”“ °f
peptides after the NTCB
Table2.2 Mass assignment of pepti
Tablcll Calculated m/z values of l
peptides afier the cleavat
latest Relative yields (%) of B-t
conditions... . .. _
Tablet] Calculated and observed n
resulting from the cleavag
at designated cysteine pai;
Table42 Calculated and observed m
resulting from the cleavag
at deSlgnated cysteine pair
tablet] Calculated and observed ml
resulting from the cleavage
inhibitor chains at designatt
able 4-4 Calculated and observed ml;

resulting from the cleavage

at sites of designated Cystei
Table

- Calculated and observed m/z

LIST OF TABLES
Table 2.1 Calculated m/z values of fragments for sulfhydryl—containing
peptides after the NTCB reaction .................................................. 71
Table 2.2 Mass assignment of peptide fragments after different treatments ............... 78
Table 3.1 Calculated m/z values of fragments for sulfhydryl-containing
peptides after the cleavage reaction .............................................. 112
Table 3.2 Relative yields (%) of B-elimination under different cleavage
conditions ........................................ . ................................... 1 15
Table 4.1 Calculated and observed m/z values for possible fragments
resulting from the cleavage reaction of ribonuclease A chains
at designated cysteine pairs ....................................................... 161
Table 4.2

Calculated and observed m/z values for possible fragments
resulting from the cleavage reaction of or—lactalbumin chains
at designated cysteine pairs ................. . ..................................... 168
Table 4.3 Calculated and observed m/z values for possible fragments
resulting from the cleavage reaction of soybean trypsin
inhibitor chains at designated cysteine pairs ................................... 173
Table 4.4 Calculated and observed m/z values for possible fragments
resulting from the cleavage of recombinant hEGF chains
at sites of designated cysteine pairs ............................................. 178

Table 4.5 Calculated and observed m/z values for possible fragments

viii

 

 

 

 

resulting from the cleav

sites of designated cyst:

Table 4.6 Calculated and observed

Table 5.1

Table 5.2

fragments resulting from
chains at designated cvst
Expecred mfz values for 1
the cleavage of 15 singlv
Fragments of cleavage pr.

cyanylated l-disulfideinte

 

resulting from the cleavage reaction of LRslGF-I chains at
sites of designated cysteine pairs .................................................. 187
Table 4.6 Calculated and observed m/z values [M-H]' for possible
fragments resulting from the cleavage reaction of insulin
chains at designated cysteine pairs ................................................ 191
Table 5.1 Expected m/z values for possible fragments resulting from
the cleavage of 15 singly reduced/cyanylated hEGF isomers ................. 224
Table 5.2 Fragments of cleavage products corresponding to 15

cyanylated l-disulfideintermediates of hEGF .................................. 237

ix

 

 

 

 

 

Figure 1.1 Intra- and Inter-chain d
figure 1.2 The modification of cys

S-carboxymethvlcvsteil
Figure 1.3 Chemical reaction betwr
Figure 1.4 The reaction of Ellman‘:

acid), with cysteinyl res

Figure 15 Modification of cysteine

fobenzoate (37331

Figure 1.6 Derivatization of the cyst

Figure] 7 S-[Z-((iodoacety1)amin0)

(IAEDANS)

Tigue 1.8 Illustration of sulfhydryV<

lg11te1.9 Classical strategy for disu

F191161.10 Current protocol for disul

1'
litre 1.11. Scheme of matrix-assistei

 

Figure 1.1
Figure 1.2

Figure 1.3
Figure 1.4

Figure 1.5

Figure 1.6
Figure 1.7

Figure 1.8
Figure 1.9
Figure 1.10

Figure 1.11.

Figure 2.1

Figure 2.2

Figure 2.3
Figure 2.4

 

LIST OF FIGURES

Intra- and Inter-chain disulfide bonds in proteins .................................. 4
The modification of cysteine with iodoacetic acid to form
S-carboxyrnethylcysteine ............................................................. 6
Chemical reaction between p-mercun'benzoate and sulfliydryl group ........... 8
The reaction of Ellman’s reagent, 5,5’-dithio-bis-(2-nitrobenzoic
acid), with cysteinyl residues in proteins ........................................... 9
Modification of cysteine sulfhydryl by 2-nitro-5-thiosul-
fobenzoate (NT SB) ................................................................. 15
Derivatization of the cysteinyl group by SBD-F or ABD-F .................... 18

5-[2-((iodoacetyl)amino)-ethyl]naphthalene—1-sulfonic acid

(IAEDAN S) ........................................................................... 19
Illustration of sulthydryl/disulfide exchange in proteins ......................... 22
Classical strategy for disulfide bond assignment of a protein ................... 23
Current protocol for disulfide bond assignment by mass spectrometry ....... 28
Scheme of matrix-assisted laser desorption/ionization time-

of-flight mass spectrometer (MALDI-TOF MS) ................................ 39
Reaction between cysteine residue and 2-nitro-5-thiocyanobenzoic

acid (N TCB), (A) Cyanylation and (B) Cleavage reaction ..................... 59
A mechanism for base catalyzed cleavage reaction .............................. 61
Chemical modification to identify free sulthydryls and total cysteines ........ 64
Conceptional scheme for microscale analysis of proteins

immobilized on an inert membrane ............................................... 69

 

 

figure 25 MALDI mass spectra c
after reaction with a 5-
peptide and (B) cleava
to doubly charged spec
calculated and observe

Figurezt’; MALDI Wm Ofcleav

TCVEWLRRYLKN, (E
(‘31 WIRQEAVDCLKr
with 5-fold molar excess
identities of the peaks ,
Figure 2.7 strum feature; of me
Farm MALDl mass spectrum (
reaction With 500 Pmol ‘
“131er was applied to :1
Peaks corresponding to 1

for the identities of mark

and (B) TCEP/NTCB trea

of other marked peaks anc

observed m/z values .....
Flam I i
.10 Imass spectra of or

(B) TCEP/NTCB treatment

Cyanylated/uncleaved peptit

Products. See Tab1e2.2 for

 

 

 

 

Figure 2.5 MALDI mass spectra of peptide RYVVLPRPVCFBKGMNYTVR
after reaction with a 5-fold molar excess of NTCB, (A) cyanylated
peptide and (B) cleavage products. ++ indicates peaks corresponding
to doubly charged species. See Table 2.1 for the correlation of
calculated and observed m/z values ............................................... 72
Figure 2.6 MALDI spectra of cleavage products of peptides (A)
TCVEWLRRYLKN, (B) DRVYIHPCHLLYYS, and
(C) MHRQEAVDCLKKFNARRKLKGA, after reaction
with 5-fold molar excess of NTCB. See Table 2-1 for the

 

identities of the peaks ............................................................. 73
Figure 2.7 Structural features of model proteins ........................................... 77

Figure 2.8 MALDI mass spectrum of 100 pmol of spinach ferredoxin after

 

reaction with 500 pmol of NTCB. Approximately 1.5 pmol of
analyte was applied to the probe tip. Asterisks (*) indicates
peaks corresponding to B-elimination products. See Table 2.2
for the identities of marked peaks and the correlation of
calculated and observed m/z values ............................................ 80
Figure 2.9 MALDI mass spectra of papain after (A) NTCB/TCEP treatment
and (B) TCEP/NTCB treatment. See Table 2.2 for the identities
of other marked peaks and the correlation of calculated and
observed m/z values ............................... . ................................ 82
Figure 2.10 MALDI mass spectra of ovalbumin after (A) NTCB/TCEP and
(B) TCEP/NTCB treatment. * indicates peaks corresponding to
cyanylated/uncleaved peptides and/or their B-elimination

products. See Table 2.2 for the identities of marked peaks and

 

 

 

the correlation of calcu

Pigurele MALDI mass spectra C
(B) NTCB/Dialysis/l C l

Table 2.2 for the correl

figure 2.12 MALDI mass spectrum
reaction with NTCB 1

calculated and observed

Figure 2.13 Stability study of somau
molar excesses of MC]

excess of NTCB reagen‘
somatostatin decompose

ngre2l4 MALDI mass spectrum ‘
NTCBflCEP treatment

Figural-15 MALDI mass spectrum c

TCEP/NTCB treatment l

Em“ (A) Cyanylation of sulfhy

bond cleavage catalde
Figure 3.2 HPLC chromatograms of
in(A) pH 9.0, 37°C, 18h,
marked with I, II, B, C/U,

B-elimination product cya

 

 

 

Figure 2.11

Figure 2.12

Figure 2.13

Figure 2.14

Figure 2.15

Figure 3.1

Figure 3.2

Figure 3.3

 

the correlation of calculated and observed m/z values ........................... 84
MALDI mass spectra of B-lactoglobulin after (A) NTCB/TCEP,
(B) NTCB/Dialysis/TCEP, and (C) TCEP/NTCB treatment. See

Table 2.2 for the correlation of calculated and observed m/z values .......... 87
MALDI mass spectrum of or- and B-hemoglobin mixtures alter

reaction with NTCB. See table 2.2 for the correlation of

calculated and observed m/z values ................................................ 88
Stability study of somatostatin in the presence of different
molar excesses of NTCB reagent. After exposed to 50-fold

excess of NTCB reagent in pH 9.0 buffer at 37°C for 16 hours,
somatostatin decomposes completely .............................................. 91
MALDI mass spectrum of 10 pmol of papain after in situ
NTCB/TCEP treatment on a Zetabind membrane ............................... 96
MALDI mass spectrum of 15 pmol of ovalbumin after in situ
TCEP/NTCB treatment on a Zetabind membrane ...................... . ........ 97
(A) Cyanylation of sulfhydryl group by CDAP and (B) peptide

bond cleavage catalyzed by ammonia ............................................ 116
HPLC chromatograms of cleavage products of SLRRSSCFGGR
in (A) pH 9.0, 37°C, 18h, and (B) 1 NINE-14011, rt, 1h. The peaks
marked with I, II, B, C/U, and D are fragment 1, fragment 11,

B-elimination product, cyanylated/uncleaved species, and peptide

dimer, respectively ................................................................. \\9

HPLC chromatograms of cleavage products of

DRVYIHPCHLLYYS in (A) pH 9.0, 3'7"C, 1811, and (B)

1 M NH40H, rt, 1h, respectively ............... . ................................ \ZQ

xii

 

 

Figure 3.4 HPLC chromatograms
EKPLQNFFLCFR in (

l M NlLOll rt, In re

Figure 3.5 HPLC chromatograms
MHRQEM’DCLKKFT
18h, and (B) l M MM
53111816 MALDI spectra of the c
alter cleavage in (A) pH

SOlUIiOrL rt llt respecti
represent fragment 1. fra

uncleaved peptide, and i
5311163 7 MALDI spectra of the c‘

DRVYIHPCHLLYYS, 2
18h and (B) l M NlL0l
Symbol ++ represents a <
Figure3.8 MALDI spectra of the cy
after cleavage in (A) pH

solution, rt, lh, respectiv

Flgure3 9
' MALDI Spectra of the cy:

 

 

 

Figure 3.4 HPLC chromatograms of cleavage products of
EKPLQNFTLCFR in (A) pH 9.0, 37°C, 18h, and (B)
lMNH40H, rt, 1h, respectively....................................... ..........121
Figure 3.5 HPLC chromatograms of cleavage products of peptide
MHRQEAVDCLKKFNARRKLKGA in (A) pH 9.0, 37°C,
18h, and (B) 1 M NH40H, rt, 1h, respectively ................................ 122
Figure 3.6 MALDI spectra of the cyanylated peptide SLRRSSCFGGR,
after cleavage in (A) pH 9.0, 37°C, 18h and (B) 1 M NH40H
solution, rt, 1h, respectively. Symbols 1, 11, D, C/U, and (3
represent fragment 1, fragment 11, peptide dimer, cyanylated/
uncleaved peptide, and its [ii-elimination product, respectively... ... .......l23
Figure 3.7 MALDI spectra of the cyanylated peptide
DRVYIHPCHLLYYS, after cleavage in (A) pH 9.0, 37"C,
18h and (B) l M NH40H solution, rt, 111, respectively.
Symbol ++ represents a doubly charged species ............................. 123
Figure 3.8 MALDI spectra of the cyanylated peptide EKPLQNFTLCFR,
after cleavage in (A) pH 9.0, 37°C, 18h and (B) l MNH40H
solution, rt, lh, respectively. [P represents the intact peptide ........... 12.4
Figure 3.9 MALDI spectra of the cyanylated
MHRQEAVDCLKKFN ARRKLKGA, after cleavage in (A) p11 9 .0,
37°C, 18h and (B) l M NH40H solution, rt, 1h, respectively ............ 12.4
Figure 3.10 MALDI spectra of spinach ferredoxin (W. 10483) after
cyanylation by the CDAP and subsequent cleavage in (A)
pH 9.0 bufi‘er for 18 hours at 37°C and (B) 1 MNH4011

solution for one hour at room temperature, respective1y ................ 12%

 

Figure 3.11 MALDI mass SW” OfO‘
cyanylation by CDAP and
buffer at 37C for 18 bout
c and d represent cyanylal
B-elimination products. rt

ligire312 MALDI mass spectra of r:
after cyanylation by C DA]
pH 9.0 at 37C for 18 hos
respectively. The peaks "
impurity in the sample T
related to two unidentifie

Fl11116.13 HPLC chromatograms of
phosphokinase afier react
for 18 hours, and (B) 1 hr
The main HPLC peaks cc
at cysteine residues are rr

Figure314 Chemical modification of
and thiocyanoPYIidine, re

Flgure 4.

1 Descriptive overview of c

assignment of disulfide b1

11 '
gure42 Chemical overview of on
ii i
gu 64.3 Conceptional scheme to i
Flgure44
- C separation of dena

partially reduced/ cyanyl‘

protein were separated b

 

Figure 3.11 MALDI mass spectra of ovalbumin (MW. 42699) after
cyanylation by CDAP and subsequent cleavage in (A) pH 9.0
buffer at 37°C for 18 hours and (B) 1M NH40H, rt, 1h , respectively.
c and d represent cyanylated/uncleaved peptides and its
1.3-elimination products, respectively ............................................ 129
Figure 3.12 MALDI mass spectra of rabbit muscle creatine phosphokinase
after cyanylation by CDAP and subsequent cleavage in (A)
pH 9.0 at 37°C for 18 hours and (B) 1 M NH40H, rt, 1h,
respectively. The peaks "*" are due to "carry-over" from an
impurity in the sample. The question marks "7" in (A) are

related to two unidentified species ............................................... 131

 

Figure 3.13 HPLC chromatograms of cleavage products of creatine

phosphokinase after reaction in (A) pH 9.0 buffer at 37°C

for 18 hours, and (B) 1 M NH40H solution, rt, 1h respectively.

The main HPLC peaks corresponding to the specific cleavage

at cysteine residues are marked ................................................. 133
Figure 3.14 Chemical modification of SH groups by 2,2'-dipyridyldisulfide

and thiocyanopyridine, respectively ............................................ 137

Figure 4.1 Descriptive overview of our proposed methodology for

assignment of disulfide bond pairings in proteins ............................ 145
Figure 4.2 Chemical overview of our methodology ...................................... 146
Figure 4.3 Conceptional scheme to illustrate "Partial Reduction" ..................... 153

Figure 4.4 HPLC separation of denatured ribonuclease A and its
partially reduced/ cyanylated isomers. Ten-nmol of the

protein were separated by reversed-phase HPLC on a

xiv

 

 

Figure 4.5

Figure4.6

FlSure 4.7

Fiturets

Vydac C18 column at a fit
gradient 204076 B in 90 ll
watetand3=0 l%TFA
singly reduced/Cyanylated
MALDI-TOP analysis

Disulfide structure of ribor
The MALDI mass spectra
the cleavage of the four 5
A isomers, corresponding
respectively The symbo
species and protonated B
Thepeaks 1-124 in ( A )
Behrnination only occuri
do not provide any speci
calculated and observed
The MALDI mass spectn
the cleavage of a lelUlt
H and " represent the d
Products, respectively ..
HPLC separation of dena
reduoed/ cyanylated isom
Vydac C4 column at a fit
gradient 40.50%

Bin 90

”dB=0.1% TFA in C1

cyanylated species, as de'

 

Figure 4.5

Figure 4.6

Figure 4.7

Figure 4.8

Vydac C18 column at a flow rate of 1.5 ml/min with a linear

gradient 20-40% B in 90 minutes, where A = 0.1% TFA in

water and B = 0.1% TFA in CH3CN. Peaks 14 represent

singly reduced/cyanylated species, as determined by

MALDI-TOF analysis .............................................................. 157
Disulfide structure of ribonuclease A ............................................ 160
The MALDI mass spectra of peptide mixtures resulting from

the cleavage of the four singly reduced/cyanylated ribonuclease

. A isomers, corresponding to the HPLC peaks 1-4 in Figure 4.4,

respectively. The symbols ++ and * represent the doubly-charged
species and protonated B-elimination products, respectively.
The peaks 1-124 in ( A) and (B) represent intact proteins in which
B-elimination only occurred at one cysteinyl residue. These products
do not provide any specific information. See also Table 4.1 for
calculated and observed m/z values ............................................. 162
The MALDI mass spectrum of peptide mixtures resulting from
the cleavage of a mixture of HPLC peaks 1 and 2. The symbols
++ and * represent the doubly-charged species and B-elimination
products, respectively ............................................................ 165
HPLC separation of denatured oe-lactalbumin and its partially
reduced/ cyanylated isomers. Separation was carried out on a
Vydac C4 column at a flow rate of 1.5 ml/min with a linear
gradient 40-50% B in 90 minutes, where A = 0.1% TFA in water
and B = 0.1% TF A in CH3CN. Peaks 1-4 represent singly reduced/

cyanylated species, as determined by MALDI-TOP analysis ............... 167

 

 

 

Figure” The MALDI mass spectra
cleavage of the four 51an
isomers, corresponding tr
respectively The symbol
species and B-elimination
4.2 for the calculated and

Figure 4.10 HPLC separation of dena
its partially reduced" cya
carried out on a Vydac (
“1111 a linear gradient 15

TFA in water and B = 8‘

 

0.1% TFA. Peaks 1 and
species, as determined b
Figure 4.11 The MALDI mass spectr
the cleavage of the two
prsin inhibitor (STl) is
peaks 1 and 2 in Figure
peaks are due to impunt
Figure 4.12 The MALDI mass spectr
the cleavage of unsepari
soybean trypsin inhibitor
the mixture was dialyset

excess reagents. See tal

m/z values ......

Figure4 '
.13 may structure and dis

 

 

 

Figure 4.9 The MALDI mass spectra of peptide mixtures resulting from the

cleavage of the four singly reduced/cyanylated or-lactalbumin
isomers, corresponding to the HPLC peaks 1-4 in Figure 4.8.,
respectively. The symbols ++ and "‘ represent the doubly-charged
species and B-elimination products, respectively. See also Table

4.2 for the calculated and observed m/z values ................................. 169

Figure 4.10 HPLC separation of denatured soybean trypsin inhibitor and

 

'its partially reduced/ cyanylated isomers. Separation was
carried out on a Vydac C4 column at a flow rate of 1.2 ml/min
with a linear gradient 15-50% B in 30 minutes, where A = 0.1%
TFA in water and B = 80% 1-propanol/20% H20 containing
0.1% TFA. Peaks 1 and 2 represent singly reduced/cyanylated

species, as determined by MALDI-TOP analysis ............................. 172

Figure 4.11 The MALDI mass spectra of peptide mixtures resulting from

the cleavage of the two singly reduced/cyanylated soybean
trypsin inhibitor (STI) isomers, corresponding to the HPLC
peaks 1 and 2 in Figure 4.10, respectively. The unidentified

peaks are due to impurities in the sample ...................................... 174

Figure 4.12 The MALDI mass spectrum of peptide mixtures resulting fi'om

the cleavage of unseparated, partially reduced/cyanylated
soybean trypsin inhibitor (STI) isomers. Before the cleavage,
the mixture was dialysed against 1% HAc solution to remove

excess reagents. See table 4-3 for possible fragments and their

m/z values .......................................................................... 175

Figure 4.13 Primary structure and disulfide bond linkage of recombinant

 

 

 

human epidermal grOMh
Figure 4.14 HPLC separation of denat
cyanylated isomers Sepa
column at a flow rate of 1
B in 15 min and 35-55%
in water and B = 0 1% Tl
reduced cyanylated specie
iigue415 The MALDI mass spectra
cleavage of the three sing
HPLC peaks 1-3 in Fig

and u represent the dOU

 

products, and cyanylated
also Table 4.4 for the cal

Figure416 The MALDI mass spectra

the cleavage of the three
isomers and a completeli
pending to the HPLC pe

The symbols ++ , ‘, and

B-elimination products a

respectively. See also T

m/z values .....

”BM 1
- 7 C separation of dena

reduced/ cyanylated ison

Vydac C18 column at a 8

gradient 30-50% B in 45

 

 

 

 

human epidermal grth factor (hEGF) (MW. 6347.2) ...................... 177
Figure 4.14 HPLC separation of denatured hEGF and its partially reduced/
cyanylated isomers. Separation was carried out on a Vydac C18
column at a flow rate of 1.0 ml/min with a linear gradient 15-3 5%
B in 15 min, and 35-55% B from 15—50 min, where A = 0.1% TFA
in water and B = 0.1% TF A in CH3CN. Peaks 1-3 represent singly
reduced/cyanylated species, as determined by MALDI-TOP analysis ....... 179
Figure 4.15 The MALDI mass spectra of peptide mixtures resulting from the
cleavage of the three singly reduced/cyanylated hEGF isomers,
HPLC peaks 1-3 in Fig. 4.14, respectively. The symbols ++ , *,
and ** represent the doubly-charged species, B-elimination
products, and cyanylated/uncleaved products, respectively. See
also Table 4.4 for the calculated and observed m/z values ................... 181
Figure 4.16 The MALDI mass spectra of peptide mixtures resulting from
the cleavage of the three doubly reduced/cyanylated hEGF
isomers and a completely reduced/cyanylated hEGF, corres-
ponding to the HPLC peaks 4—7 in Figure 4.14, respectively.
The symbols -H- , *, and ** represent doubly-charged species,
B-elimination products, and cyanylated/uncleaved products,
respectively. See also Table 4.4 for the calculated and observed
m/z values .......................................................................... 183
Figure 4.17 HPLC separation of denatured LR3IGF-I and its partially
reduced/ cyanylated isomers. Separation was carried out on a
Vydac C18 column at a flow rate of 1.0 ml/min with a linear

gradient 30-50% B in 45 minutes, where A = 0.1% TFA in

xvii

 

 

 

 

 

 

water and B = 0 1% TFA
singl.v reduwdicyanylaled
10F anaii'siS

iigUre 4.18 The MALDI mass 513%"2
the cleavage of the two 5
isomers. conesponding l
4.17, respeciiVel." The 5

text. See also Table 4 5
Figure 4.19 HPLC separation ofdena
cyanylated isomers Sep
column at a flow rate of
B in 40 minutes. where .1

TFA in CH3CN Peaks
species, as determined b'
A and B are insulin A-cl
FigureiZO The negative ion MALDI
resulting from the cleave
cyanylated insulin isome
1-3 in Figure 4.19, respe
the B-elimination produc
reSPCCthely. See also T

observ

Figure421 Separ ~ ed mil values. H
ation of nbonucleasi
Species by microbore C 1:

conditions are the same a

 

 

water and B = 0.1% TF A in CH3CN. Peaks 1 and 2 represent
singly reduced/cyanylated species, as determined by MALDI-
TOF analysis ........................................................................ 185
Figure 4.18 The MALDI mass spectra of peptide mixtures resulting from
the cleavage of the two singly reduced/cyanylated LR3IGF-I
isomers, corresponding to the HPLC peaks 1 and 2 in Figure
4.17, respectively. The symbol * represent the B-elimination the
text. See also Table 4.5 for the calculated and observed m/z values ....... 188
Figure 4.19 HPLC separation of denatured insulin and its partially reduced/
cyanylated isomers. Separation was carried out on a Vydac C18
column at a flow rate of 1.0 m1/min with a linear gradient 20-50%
B in 40 minutes, where A = 0.1% TFA in water and B = 0.1%
TFA in CH3CN. Peaks 1-3 represent singly reduced/cyanylated
species, as determined by MALDI-TOP analysis. The peaks marked
A and B are insulin A-chain and B-chain, respectively ....................... 191
Figure 4.20 The negative ion MALDI mass spectra of peptide mixtures
resulting from the cleavage of the three singly reduced/
cyanylated insulin isomers, corresponding to the HPLC peaks
1-3 in Figure 4.19, respectively. The symbols * and ** represent
the B-elimination products and cyanylated/uncleaved peptides,
respectively. See also Table 4.6 for the calculated and
observed m/z values .......................................................... 193
Figure 4.21 separation of ribonuclease A and its partially reduced/cyanylated
species by microbore C18 HPLC column (1 x100mm). HPLC

conditions are the same as those in Figure 4.4, except the flow

xviii

 

 

 

rate is 005 ml/min
Figure 51 Conventional approach fc
intermediates
Figrre 5.2 New approach for disulfrr
Figure53 Primary structure and dis
factor (hEGF)Figure 5 4
folding intermediates of l
Fl8111e55 Comparison of HPLC prr
and CDAP-trapped foldi
iigure5.6 HPLC separation of 111-!

11gure57 The MALDI mass spectr

 

the cleavage of the three
of non-native hEGF 111-;
peaks 1-3 in Figure 56 1
represent the doubly-cha
products, respectively
“8116.8 Comparison of HPLC prr
singly redUCed/cyanylate
represented by peaks 1 a
Film-9 HPLC separation of “H

Fl
81610 The MALDI mass spectr;

of the 3 singly reduced/c

A~C corresponding to 11

respectively. . . .
rims ........ . .
. l
1 C separation of TM

 

Figure 5.1

Figure 5.2

Figure 5.3

Figure 5.5

Figure 5.6

Figure 5.7

Figure 5.8

Figure 5.9

 

rate is 0.05 ml/min .............................................................
Conventional approach for disulfide mapping of protein folding

intermediates ..................................................................

New approach for disulfide mapping of protein folding intermediates... . . ..

Primary structure and disulfide linkage of human epidermal growth
factor (hEGF)Figure 5.4. HPLC separation of CDAP-trapped
folding intermediates of hEGF at different time courses ..................

Comparison of HPLC profiles of iodoacetate-trapped, acid-trapped,
and CDAP-trapped folding intermediates of hEGF ......................

HPLC separation of III-A and its partially reduced/cyanylated species

The MALDI mass spectra of peptide mixtures resulting from

the cleavage of the three singly reduced/cyanylated species

of non-native hEGF III-A, A-C corresponding to the HPLC

peaks 1-3 in Figure 5-6, respectively. The symbols 4+ and *

represent the doubly-charged species and B-elimination

..... 195

..... 203

208

..... 218

..... 220

..... 225

products, respectively ............................................................. 226

Comparison of HPLC profiles of cleavage products of two
singly reduced/cyanylated HI-A isomers. (A) and (B) are

represented by peaks 1 and 2 in Figure 5.6, respectively ................

HPLC separation of III-B and its partially reduced/cyanylated species. . . ..

Figure 5.10 The MALDI mass spectra of peptide mixtures from the cleavage

..... 228

.230

of the 3 singly reduced/cyanylated species of non-native hEGF III-B,

A-C corresponding to the HPLC peaks 1-3 in Figure 5.9,

respectively ...................................................................

..... 231

Figure 511 HPLC separation of II-A and its partially reduced/cyanylated species ...... 233

xix

 

 

 

—

Figure 5.12 The MALDI mass spectra
of (A) the cyanylated hE
reduced/cyanylated speci
peaks 1 and 2 in Figure

ligureSl3 Disulfide mapping ofl

Figure 5.14 Disulfide structures of

ofhEGF............ .

 

 

Figure 5.12 The MALDI mass spectra of peptide mixtures from the cleavage

of (A) the cyanylated hEGF H-A, (B) and (C) the partially

reduced/cyanylated species of II-A, corresponding to the HPLC

peaks 1 and 2 in Figure 5.11, respectively ..................................... 235
Figure 5.13 Disulfide mapping of l-disulfide folding intermediates of hEGF ........... 238
Figure 5.14 Disulfide structures of seven well populated folding intermediates

of hEGF ........................................................................... 241

 

 

 

1 Introduction

Since its discovery in 1912,
broad and diverse disciplines. Howcv
of biomolecules stems in large
techniques in the early 19803 (1, 2).
sciences have led to a dramatic '
inforination of biologitu active
undiscipline-biological mass spec
Structural isms of biomedically impor
ins accelerated dramatically, due to
ionization techniques: matrix—mister
electrospiay ionization (EST) (4). Con:
routine the analysis of high mass com]:
mural products, and drug metabolites
MS has been revolutionary in spurrin
and biotechnology, and most recently ir

Protein analytical techniques a

Siructrrre as well as detecting and defir

thatthe complete structure of a proteir

 

 

 

CHAPTER 1

INTRODUCTION

1. Introduction

Since its discovery in 1912, mass spectrometry (MS) has provided key insight for
broad and diverse disciplines. However, the involvement of mass spectrometry in studies
of biomolecules stems in large part from the discovery of desorption ionization
techniques in the early 1980s (1, 2). At the same time, rapid advances in biological
sciences have led to a dramatic increase in the demand for chemical and structural
information of biologically active materials. The combination led to the creation of a
new discipline-biological mass spectrometry—which addresses the challenging unsolved
structural issues of biomedically important molecules. In the past decade, this expansion
has accelerated dramatically, due to the discovery of two new mass spectrometry
ionization techniques: matrix-assisted laser desorption/ionization (MALDI) (3) and
electrospray ionization (ESI) (4). Commercial availability of these instruments has made
routine the analysis of high mass compounds including proteins, peptides, carbohydrates,
natural products, and drug metabolites with picomole to femtomole sensitivity. Today,
MS has been revolutionary in spurring research in protein biochemistry, glycobiology,
and biotechnology, and most recently in DNA sequencing (5).

Protein analytical techniques are essential for determining a protein’s primary
structure as well as detecting and defining posttranslational modifications of proteins so
that the complete structure of a protein can be obtained (6, 7). Nucleotide sequences of

l

 

 

 

   
   
   
   
 

chnnically modified in viva after
Defining postnanslational modificati
are destined for use as pharmaceuti
same chemical structure as their
Umceptable structtm could
pharmacological or immunological

Defining the locations and
isone of the mom important contribu
tomake in the field of protein bi
evident when it is uwd in conjunctii
protein sequence and amino acid an
which identification is hm entirely
20 commonly occurring amino acids
molecular mass, an intrinsic physical
of choice for characterization of postr
welldeveloped, accurate, and semi
available.

Disulfide bond formation at
Postttanslational modifications frequr

Mass spectrometry, combined with

extensively applied to the recognitii

 

 

 

genes embody the information required to deduce the primary structure (amino acid
sequence) of proteins but they do nOt reveal whether the side chains of amino acids are
chemically modified in vivo after translation of genes or the extent of any modifications.
Defining posttranslational modifications is essential in the case of expressed proteins that
are destined for use as pharmaceuticals. This is required to ensure that they have the
same chemical structure as their natural counterparts or an acceptable alternate structure.
Unacceptable structures could be detrimental to recipients due to unwanted
pharmacological or immunological activities.

Defining the locations and structures of posttranslational modifications in proteins
is one of the most important contributions that mass spectrometry has made and continues
to make in the field of protein biochemistry (8). The unique potential of MS is most
evident when it is used in conjunction with conventional methodologies like automated
protein sequence and amino acid analysis. In contrast to the latter two approaches in
which identification is based entirely on chromatographic retention relative to one of the
20 commonly occurring amino acids or a derivative thereof, MS relies on measurement of
molecular mass, an intrinsic physical pr0perty. Mass spectrometry is clearly the method
of choice for characterization of posttranslationally modified proteins, since in most cases
well-developed, accurate, and sensitive chemical or biochemical approaches are not
available.

Disulfide bond formation after gene expression is one of the most important
POSttranslational modifications frequently encountered in protein characterization (9, 10).
Mass spectrometry, combined with other chemical/biochemical techniques, has been

extensively applied to the recognition of both free sulthydryls and disulfide bonds in

 

 

 

 

 

proteins. In this disserta' tion. novel

disuliide bonds, and disulfide struc
graterdetail inthechapter 2. 4. an
introduce basic aspects of MALDl

statisinproteins.

ll. Characterization of Cysteine S
Among the 20 amino acids

Iheredwed form of cysteine contai
cystine, is formed by linking two sul
wntnhutes to protein biological fun
active site for enzyme catalysis such
metals (ll). Sulthydryl groups c
physiological and biochemical procc:
division, oxidative phosphorylation at
hand comprises the major covalent 1
Chain in nature (Figure 1.1). Intra-ch
inribonuclease A, serve to confer
Chain Additionally, by limiting or
the correct orientation of the amino
antibodies and other biologically

functional in maintatmn' ' g the quate

only linkage between subunits or pro

 

proteins. In this dissertation, novel methodologies for recognizing free sulfliydryl groups,
disulfide bonds, and disulfide structures of folding intermediates will be discussed in
greater detail in the chapter 2, 4, and 5, respectively. The objective of this chapter is to

introduce basic aspects of MALDI and current methodologies for characterizing cysteine

status in proteins.

 

H. Characterization of Cysteine Status in proteins

Among the 20 amino acids that compose proteins, cysteine has unique properties.
The reduced form of cysteine contains a sulihydryl (thiol) group while its oxidized form,
cystine, is formed by linking two sulfltydryls together to form a disulfide bond. Cysteine
contributes to protein biological functions by using its free sulthydryl (-SH) group as the

active site for enzyme catalysis such as in cysteine proteases, and as the chelating site for

 

metals (11). Sulfhydryl groups can also play an important role in a variety of
physiological and biochemical processes such as muscular contraction, nerve activity, cell
division, oxidative phosphorylation and photosynthesis (11). Disulfide bond, on the other
hand, comprises the major covalent cross-linkage in proteins and may be intra- or inter-
chain in nature (Figure 1.1). Intra-chain disulfide bOnds, such as the four disulfide bonds

in ribonuclease A, serve to confer conformational stability on the folded polypeptide

 

chain. Additionally, by limiting or directing the folding these bonds may contribute to
the correct orientation of the amino acid residues that form the active sites of enzymes,
antibodies and other biologically active proteins. Inter-chain disulfide bonds are
functional in maintaining the quaternary structure of multi-chain proteins, serving as the

only linkage between subunits or providing covalent stability to structures otherwise

 

 

 

 

 

 

 

Inter

Figure 1.1. Inna-and

mannined by non-covalent forces
proteins play a unique role in its act
themechanism of activity in glutathi
that cysteine sulfliydryls and cysti
hmctions, the recognition of cyst:
characterization (9, ii).

For a new protein with
cysteine content is very beneficial t

the amino acid sequence of the pr

CDNA sequence, difi'erentiating be

 

 

 

 

 

m—
U)..—

Intra—chain disulfide bonds

 

l
s
l
s
J

Inter-chain disulfide bonds

Figure 1.1. Intra- and Inter-chain disulfide bonds in proteins.

maintained by non-covalent forces (e. g., insulin, IgG). Disulfide bonds in particular
proteins play a unique role in its activity. For example, disulfide bonds are involved in
the mechanism of activity in glutathione reductase and thioredoxin. Due to diverse roles
that cysteine sulfhydryls and cystine disulfide bonds play in protein structures and
functions, the recognition of cysteine status is a very important aspect of protein
characterization (9, 11).

For a new protein with unknown sequence, the quantitative determination of
cysteine content is very beneficial to further characterization of the protein. Even though
the amino acid sequence of the protein can be easily deduced from the corresponding

cDNA sequence, differentiating between free cysteine residues and those involved in

 

 

 

  
  
   
 
  
  
 

disulfide bonding (cystines) is also
the cysteine residues after translati
pimarystructure. Characterization
bonds in proteins has become a cruci
hearse molecules having inco
biological activity than that of th
cysteine status in protein chemistry
cystine; (2) localization of free sul

linkage.

A Quantitative Determination of 1
The cysteine sulfhydryl grou
acids by its high reactivity and by it
they participate, such as alkylatio
exchange, charge-transfer complexe
(ll). Sulihydryls can dissociate into
Sulfltydryl groups is attributable to
ions, which exist at reasonable con

(pK, 8—105). Thiolate-dependent r

at moderate speed in the pH range 6-

 

 

 

 

disulfide bonding (cystines) is also very essential because the state and connectivity of
the cysteine residues after translation can not be predicted readily from the protein’s
primary structure. Characterization of the location of sulfltydryl groups and disulfide
bonds in proteins has become a crucial part of the analysis of recombinant DNA products,
because molecules having incorrect disulfide bond linkages may have much lower
biological activity than that of the desired product. The problems associated with .
cysteine status in protein chemistry include: (1) quantitative evaluation of cysteine and
cystine; (2) localization of free sulflrydryl groups; and (3) assignment of disulfide bond

linkage.

A. Quantitative Determination of SH and S-S groups in proteins

The cysteine sulfhydryl group is distinguished from other side chains of amino
acids by its high reactivity and by the exceptionally diverse chemical reactions in which
they participate, such as alkylation, acylation, oxidation, sulthydryl-disulfide bond
exchange, charge-transfer complexes, reactions with organic heavy metal compounds
(11). Sulthydryls can dissociate into a proton and a thiolate anion. The high reactivity of
sulfhydryl groups is attributable to the high nucelophilicity of the correSponding thiolate
ions, which exist at reasonable concentrations at neutral to weakly alkaline pH-values
(pK, 8-10.5). Thiolate-dependent reactions thus proceed readily at pH 8 and above, and

at moderate speed in the pH range 6-8.

RSH —-> RS" + H’“

 

 

A variety of methods,

disulfidc bonds participate, have
disulfidc bond content in proteins.

chaptetsUl-B).

L Methods Based on Alkyhtion

0
II + I
-NH*~CH -C —

Figure 1.2. The modifica
S-carboxymethylcysteine

Alkylation (S-carboxymcthyl
isthe most frequently used reaction
be accomplished with compounds
icdoacctamide, bromoacetate, etc,
and specificity to SH (Figure 1.2
electrochemical method (14) or ab

bi "(I-labeled iodoacetate permits

radioactivity of the product (16.

 

 

 

A variety of methods, based on numerous reactions in which sulfliydryls and
disulfide bonds participate, have been studied for the determination of sulfhydryl and
disulfide bond content in proteins. Reviews on this subject can be found in several book

chapters (1 1-13).

1. Methods Based on Alkylation

CIIOOH
SH

1 CH2

CH2 0 l

I II + ICH2COOH -——> S + H1
—NH—-CH —C -— «. 1

cc c
—NH —CH -—C —

Figure 1.2. The modification of cysteine with iodoacetic acid to form
S-carboxymethylcysteine.

Alkylation (S-carboxymethyltion) of SH groups to form S-carboxymethylcysteine
is the most frequently used reaction in protein characterization (11:13). Alkylation can
be accomplished with compounds containing an alkyl halide group such as iodoacetate,
iodoacetamide, bromoacetate, etc., among which, iodoacetamide shows highest reactivity
and specificity to SH (Figure 1.2). The iodide liberated can be determined by an
electrochemical method (14) or absorbance at 226 nm (15). As an alternative, alkylation
by “C-labeled iodoacetate permits the number of thiol groups to be determined from the

radioactivity of the product (16, 17). To avoid side reactions connected with the

 

 

 

 

 

fonnafion of iodine by photooxid:
performedindiedark.

Based on the same principlt
developed in order to improve the :
alkylation by iodoacetate is still t1

nilfliydryls because of its simplicity

2. Methods Based on Memptidat

Organomercurial compounds
Sllgmups (Figure 1.3). In con‘
crumble. Among various organor
mmpormds have widely been e
mectnibenzoate proposed by Boyer
sensitivity, selectivity, and precisi:

measurement of the increase in absor

  
  
  
  
  
  
  
   

binding of p-mercuribenzcate (p~M
only the number of SH groups in a
of loss of enzymatic activity and
rate and the extent of the reaction 0
PH 4.6 than at pH 7-8 because of th
result, the reactivity of the reagent i
reagents, such as l-(4-chloromercu

and p-(hydroxymercuri)-benzoic

formation of iodine by photooxidation of iodide ions, the carboxymethylation was
performed in the dark.

Based on the same principle, a number of other alkylating reagents have been
developed in order to irnprove the sensitivity and specificity of the reaction (18). But
alkylation by iodoacetate is still the most popular derivatization reaction to modify

sulthydryls because of its simplicity and specificity.

2. Methods Based on Mercaptidation

Organomercurial compounds are high affinitive and specific reagents for protein
SH groups (Figure 1.3). In contrast to alkylation, the mercaptidation reaction is
reversible. Among various organomercurial reagents, monofunctional organomercurial
compounds have widely been exploited. Spectrophotometric titration with p-
mercuribenzoate proposed by Boyer stands in first place because of its simplicity, high
sensitivity, selectivity, and precision (19). This classical method is based on the
measurement of the increase in absorbance in the 25 0-255 nm region that occurs upon the
binding of p-mercuribenzoate (p-MB) to SH groups. It provides a means to measure not
only the number of SH groups in a protein, but also the relationship between the degree
of loss of enzymatic activity and the number of blocked SH groups. Interestingly, the
rate and the extent of the reaction of SH groups of many proteins with p—MB is higher at
pH 4.6 than at pH 7-8 because of the high affinity of p—MB for hydroxide ions (20). As a
result, the reactivity of the reagent increases as the pH of the solution is decreased. Other
reagents, such as 1-(4-chloromercuriphenylazo)—napphthol-2 (CMPN; mercury-orange),

and p-(hydroxymercuri)—benzoic acid (21), were also proposed to improve the

 

 

 

 

 

 

 

 

Spectrophotometric detection of SI

carboxyhemoglobin (23).
SIH Ct
CH2 0
I II +
—NH—CH—-c-— @

Figure 1.3. Chemical reaction be

3. Methods Based on Sulthydrylll

RISSR, + R:

R,SSR +

Among oxidants of SH gro
reactions with sulfliydryls are ab
sclflrydryl/disulfide exchange. As

of two steps of nucleophilic substi

intermediate stage.

 

 

 

Spectrophotometric detection of SH groups in lysine monoxygenase (22) and in

carboxyhemoglobin (23).

SH COO'

CH2 0 Hg
l H + "—‘—> I + H+
—NH-—CH —'C -— S

Figure 1.3. Chemical reaction between p-mercuribenzoate and sulfliydryl group.

3. Methods Based on SulfhydryllDisulfide Exchange

RISSR1 + RSH ——-> RISH + RlSSR

RISSR + RSH ——> RISH + RSSR

Among oxidants of SH groups, disulfides occupy a special position since their
reactions with sulfhydryls are absolutely specific. This reaction is referred to as
sulfhydryl/disulfide exchange. As can be seen fiom the equations, this reaction consists
of two steps of nucleophilic substitution with the formation of a mixed disulfide at the

intermediate stage.

 

 

 

 

 

Among the various disulfi

nitrobemic acid), DTNB, has
thiolate anion with excess E11
stoichiometric formation of TNB l
disolfide (RS-TNB) (Figure 1.4). 1
his bufl’er, pH 8, at a concentratit
ccntaininglmMEDTAinafivet
groups (EDTA is included not only
metal ions can affect the developme
=13,600 Worn"), very fast and st:

ofSHgmupsinboth native and (let

 

Figure 1.4. The reaction of Eli
with cysteinyl residues in prot

 

 

 

 

Among the various disulfides proposed, Ellman’s reagent, 5,5’-dithio—bis-(2-
nitrobenzoic acid), DTNB, has received the widest application (24). The reaction of
thiolate anion with excess Ellman’s reagent at pH 7-8 is favored toward the.
stoichiometric formation of TNB thiolate (5-thio-2-nitrobenzoate, TNB') and a mixed
disulfide (RS-TNB) (Figure 1.4). Ellman’s reagent usually is dissolved in phosphate or
tris bufi‘er, pH 8, at a concentration of 10 mM. It is added to a solution of protein
containing lmM EDTA in a five to ten-fold excess with respect to the number of SH
groups (EDTA is included not only to protect SH groups from oxidation but also because
metal ions can affect the development of the color). The method is highly sensitive (a412nm '
= 13,600 M'1cm"), very fast and strictly specific, and may be used for the determination

of SH groups in both native and denatured proteins (25-27).

N02 N02

s. 9 SI.
e + 9
COOH

1|
—NH-—CH --C -
‘et '
| COOH —NH—CH —C —

NO2

—————>

+
—m——m—
_m_m _

Figure 1.4. The reaction of Ellman’s reagent, 5,5’-dithio-bis-(2-nitrobenzoic acid),
with cysteinyl residues in proteins.

 

 

 

 

Grassetti and Murray (28,

dithiodipyridineweregood substi
stronger reactivity to SH groups '
widely md as the binding bed 01

proteins(3l).

4. Methods Based on Eledrochem

One of the important pmpert
each other though oxidation/reductit
Undermild conditions (e.g., in the
bonds Finther oxidation results in

reducing agent, disulfide bonds can

Those specific oxidation/red
thebasis of amperometric titration c
the first portion of AgNO, binds t
small current to increase. As soo

appear in solution and are redu

Current which is proportional to the

 

10

Grassetti and Murray (28, 29) demonstrated 2, 2’-dithiodipyridine or 4, 4’-
dithiodipyridine were good substituents for Ellman’s reagent. These compounds showed
stronger reactivity to SH groups in a wide pH range (pH 1-8) (30) and are therefore

widely used as the binding bed of affinity chromatography for sulfhydryl-containing

proteins (3 1).

4. Methods Based on Electrochemical Titration
One of the important properties of sulfhydryl/disulfide is that they can convert to

each other though oxidation/reduction reactions. Cysteine is very sensitive to oxidation.
Under mild conditions (e.g., in the air), SH groups undergo oxidation to form disulfide

bonds. Further oxidation results in the formation of cysteic acid. Under the presence of

reducing agent, disulfide bonds can be reduced to form sulfhydryls.

[O] RSSR + 2H+ + 2e'
[R]

 

 

2RSH

Those Specific oxidation/reduction properties of a sulfllydryl/disulfide pair form

the basis of amperometric titration of sulfliydryls with silver nitrate (32). In this method,
the first portion of AgNO3 binds to the SH groups of the protein and do not cause the
small current to increase. As soon as all the SH groups are blocked, free metal ions
appear in solution and are reduced on the platinum electrode, resulting in a diffusion

current which is proportional to the concentration of the metal ions released to solution.

 

 

 

 

The amperometn'c titration

sulfhydryl groups. With prior re
enmded to quantitation of disulfid
(33) niece-shiny measured the con
unified pig kidney Na‘. K'-ATPa:
ammonium nitrate buffer (pH 7.6).
Recently Sun et a1 (34)
imultaneously both sulfliydryls an
detectorsystem actually has five ele
steel amtiliary electrode, and three I
used to distinguish between sulfhy
chromatographic analysis, as illustrt
mdpartially reduwd bovine insulin.
Although the advantage r
spectrophotometry lies in the possi?
solutions, in enzymological studies,
image as the Spectrophotometric

sensitivity, the necessity of special a

5. Methods based on Oxidation

Shipton et a1 (35) found th

selectively and quantitatively oxid

 

 

11

The amperometric titration method was first proposed for the measurement of
sulfhydryl groups. With prior reduction of disulfide bonds, this method was also
extended to quantitation of disulfide bonds in proteins. For example, Gevondyan et al
(33) successfully measured the content of free SH groups and disulfide bonds in the
purified pig kidney Na“, K+-ATPase by amperometric titration with silver nitrate in

ammonium nitrate buffer (pH 7.6).

 

Recently Sun et al (34) developed an electrochemical detector to detect
simultaneously both sulfliydryls and disulfides in peptides. The three-electrode EC
detector system actually has five electrodes, one Ag/AgCl reference electrode, a stainless
steel auxiliary electrode, and three Hg/Au amalgam working electrodes, and thus can be
used to distinguish between sulfliydryl- and disulfide—containing peptides in a single

chromatographic analysis, as illustrated with proteolytic digests of bovine OLA-crystallin

 

and partially reduced bovine insulin.

Although the advantage of electrochemical titration in comparision with
spectrophotometry lies in the possibility of carrying out analysis in cloudy or colored
solutions, in'enzymological studies, the electrochemical method has not received as wide
a usage as the spectrophotometric method. This is evidently because of its lower

sensitivity, the necessity of special apparatus.

5. Methods based on Oxidation
Shipton et a! (35) found that benzofuroxan (benzo-Z-cxa-l, 3-diazole N-oxide)

selectively and quantitatively oxidizes the SH groups of proteins, the reagent being

 

 

 

 

 

reduced to o—benzoquinone dioxime.
absorbance at wavelengths far remo
Alternatively, proteins are treated

amiable residues of cysteine into cy

 

thenbereliably determined by amin
perfonnic acid leads to destruction
quantitative conversion of methioni

is therefore limited

h Methods Based on Enzyme Ass

RSH + Papain-S-SCP

Singh et a1 (37) developed
group determination using an inset
llle sulthydryl/disulfide exchange n
SCH, results in the stoichiometri
motivated papain catalyzes the hy
nnplified spectrophotometn'c signal
is about 100-fold more sensitive

sulfhydryls, e. g., glutathione, cys

determined by this approach. Howe

 

f_———

12

reduced to o-benzoquinone dioxime. The reduction is accompanied by a large increase in
absorbance at wavelengths far removed from the region where proteins absorb strongly.
Alternatively, proteins are treated with performic acid (3 6). This is to convert the
unstable residues of cysteine into cysteic acid, which withstands acid hydrolysis and may
then be reliably determined by amino acid analysis. However, treatment of a protein with
performic acid leads to destruction of tryptophan, partial destruction of tyrosine, and
quantitative conversion of methionine into methionine sulfone. The practical application

is therefore limited.

6. Methods Based on Enzyme Assay
RSH + Papain-S-SCH3 -——> RS-SCH3 + Papain-SH

Singh et al (37) developed a sensitive Spectrophotometric assay for sulfliydryl
group determination using an inactive disulfide derivative of papain (papain-S-SCH3).
The sulfliydryl/disulfide exchange reaction of a protein cysteinyl residue with papain-S-
SCH3 results in the stoichiometric formation of active papain (papain-SH). The
reactivated papain catalyzes the hydrolysis of a chromogenic substrate, resulting in an
amplified Spectrophotometric signal proportional to the initial amount of SH. The assay
is about 100-fold more sensitive than that using Ellman’s reagent. A variety of
sulfliydryls, e. g., glutathione, cysteamine, penicillamine, etc. have been successfully

determined by this approach. However, the feasibility of the method for the measurement

 

 

 

 

of thiols in large proteins remain 1

proteins, the accessibility of protein

7. Methods Based on Mass Spec
Mass spectrometry is
determining protein cysteine/cysti

limitations of sensitivity, multiple
many classical methods (38-40).
Feng er al (38) demonstrat
PSI-MS for rapidly counting cyste
proteins. In this analysis, a peptid
iodmceticacidintheabsence ofa
ndfliydryl group(s). Parallel analys
datafiom which the number of fr:
molecular weight of the alkylated d
usefor iodoacetic acid, a mass shifi
protein. A complementary experimi
protein to a disulfide reducing r
differentiation of the total numbe
Recently Sun et a1 applied this pr

littoglobulin B (39). However, the

Shift from alkylation by iodoacetic

 

 

 

13

of thiols in large proteins remain to be tested. Because of the steric hindrance of large

proteins, the accessibility of protein sulfliydryls to papain might be a problem.

7. Methods Based on Mass Spectrometry

Mass spectrometry is becoming an alternative means of quantitatively
determining protein cysteine/cystine content in a way that promises to overcome
limitations of sensitivity, multiple derivatization, and protein contamination that burden
many classical methods (3 8-40).

Feng et al (3 8) demonstrated a method that combines reduction/alkylation and
ESI-MS for rapidly counting cysteines, free sulfhydryl groups, and disulfide bonds in
proteins. In this analysis, a peptide or denatured protein is first allowed to react with
iodoacetic acid in the absence of a disulfide reducing agent to selectively label any free
sulfhydryl group(s). Parallel analyses using underivatized protein provide mass spectral
data from which the number of free sulfhydryls can be determined by the shift in the
molecular weight of the alkylated derivative relative to that of the native protein. In the
case for iodoacetic acid, a mass shift of 59 u is observed for each derivatized cysteine of a
protein. A complementary experiment involving the same reaction, after exposure of the
protein to a disulfide reducing reagent, provides mass spectral data that allows for
differentiation of the total number of cysteine(s) and/or cystine(s) originally present.
Recently Sun et al applied this procedure to ribonuclease A, lysozyme, insulin, and [3-
lactoglobulin B (3 9). However, the above method is limited to small proteins as the mass

shift from alkylation by iodoacetic acid is small and the limited mass resolution available

 

 

 

 

and potentially prevent small in

drawback, Zaluzec er a1 (40)
iydroxymercuribenzoate (pHMB)
derivatizationthemass shifl of32l
masshitt minimizes the error whic
topoor resolution of the instrumen
protein samples and provide an e

cysteine content even in a protein in

& Determination of S-S Content

Conventional methodologie
reductive cleavage of the S-S bonds
the methods mentioned above. Th
deduced

A sensitive and quantitative
thiosulfobenzoate (NTSB) has dra‘
actually composed of two sequentir

dimlfide bond with sodium sulfite:

RSSR' + so;

 

 

 

14

could potentially prevent small mass shift from being detected. To circumvent the
drawback, Zaluzec et al (40) used monofunctional organomercurial reagent (p-
hydroxymercuribenzoate (pHMB) for selective derivatization of thiol groups. After
derivatization, the mass shift of 321 u is obtained for each derivatized cysteine. The large
mass shift minimizes the error which might occur during mass spectrometric analysis due
to poor resolution of the instrument. Both MALDI and E81 can detect low picomole of
protein samples and provide an excellent method for the quantitative assessment of

cysteine content even in a protein mixture.

8. Determination of S-S Content

Conventional methodologies for S—S counting in proteins are based on the
reductive cleavage of the S-S bonds to form cysteinyl residues that can be determined by
the methods mentioned above. The content of disulfide bonds can then indirectly be
deduced.

A sensitive and quantitative method involving the use of the reagent 2-nitro-5-
thiosulfobenzoate (NTSB) has drawn great attention (41, 42). The NTSB assay is
actually composed of two sequential reactions. The first reaction is the cleavage of a

disulfide bond with sodium sulfite:

RSSR' + 8032' ——> RSSO3' + R'S'

 

 

 

 

producedinreactiononN'lSBto

du'cbcnzoate (NTB), the latter can
the reaction of NTSB with

of both dissolved oxygen and the

inconvenience associated with the

wellastheneedtoremovethered

SH sso,
I

c 0
iHi n +
—NH—CH-—c—— (01

Figiue 1.5. Modification of cysteir

Himse er a! (43) descri
wlyacrylamide gel electrophoresi
quantitative determination of irn
electrophoresis is one of the most j
biological systems, since it requii

Visualized by selective methods of

Withers of cleaved disulfide bonds

 

 

 

 

 

 

15

The second reaction (Figure 1.5) involves nucleophilic attack of the thiolate
produced in reaction on NTSB to yield 1 mol each of a thiosulfonate and 2-nitro-5-
thiobenzoate (NTB), the latter can be measured by the absorbance at 412 nm.

The reaction of NTSB with thiols and disulfides can be carried out in the presence
of both dissolved oxygen and the reducing agent which eliminates the inaccuracy and
inconvenience associated with the necessity to work under an oxygen-free atmosphere as

well as the need to remove the reducing agent.

 

SH S.‘
|
CH2 0 s ‘
~NH—éH—t': + T ('3 +
No2 —NH—CH —-C — No2

Figure 1.5. Modification of cysteine sulfhydryl by 2-nitro-5-thiosu1fobenzoate (N TSB).

Hirose et al (43) described another novel method that makes use of
polyacrylamide gel electrophoresis (PAGE) following two-step alkylation for the
quantitative determination of intramolecular disulfide bonds in proteins. Gel
electrophoresis is one of the most powerful techniques for protein analysis in complex
biological systems, since it requires very small amounts of proteins which can be
visualized by selective methods of gel staining. In this method, proteins with different

numbers of cleaved disulfide bonds are alkylated with iodoacetic acid or iodoacetamide

 

 

 

 

 

  
  
  
  
 
 
  
  
 

asthefirst step. The remaining di
and the newly generated free sulfli
(iodoacetamide, iodoacetic acid. 0
acid-urea PAGE, different intermed
hands, depending on difierences
alkylation used in combination

analysis of disulfide bonds in protei

B. localization of Sulfltydryl G

localization of cysteine res
bonds (see section C) in protei
manipulations in protein chemistry
been tested and uwd both in the re
available SH groups, and there is nt

Yd.

1- Classical Approach

The general strategy for
conventional methods involves seve
trollps, usually, by an irreversible r

intent or minimize sulfliydryl/dis

 

enzymes or chemical reagents betw

lSseparated by chromatography. F

 

 

 

 

 

 

16

as the first step. The remaining disulfide bonds were reduced by excess dithiothreitol,
and the newly generated free sulfliydryls were alkylated with the reagent not yet used
(iodoacetamide, iodoacetic acid, or vinylpyridine) as the second step. By subsequent
acid-urea PAGE, different intermediates formed in two steps can be resolved into distinct
bands, depending on differences in their net charge and conformation. Two-step
alkylation used in combination with autoradiography was especially useful for the

analysis of disulfide bonds in proteins synthesized in complex biological systems.

B. Localization of Sulfhydryl Groups

Localization of cysteine residues and their possible pairings in forming disulfide
bonds (see section C) in proteins involves some of the oldest and best studied
manipulations in protein chemistry. A large number of reagents and procedures have
been tested and used both in the reduction of disulfide bonds and in the reaction of the
available SH groups, and there is no indication that the search for better methods is over

yet.

1. Classical Approach

The general strategy for locating free sulfhydryl groups in proteins by
conventional methods involves several steps. The first step is to modify free sulfhydryl
groups, usually, by an irreversible reaction such as alkylation, under conditions that can
prevent or minimize sulfhydryl/disulfide exchange. Second, the protein is cleaved by
enzymes or chemical reagents between cysteine residues. Third, the mixture of the digest

is separated by chromatography. Finally, the derivatized sulfhydryl-containing peptides

—

are recognized, mapped to sequem
andrelated to specific segments of

A key to the success of thi
reaction and reagent for sulfhydryl
medium and label sulfhydryl gror
conditions (ideally under weak a
exchange which is minimized at 1
strong UV or fluorescent absorp
absorption of proteins, or easily an
of derivatimd peptides afier HPLC
bedistinguishable from other min
the identification of amino acids
derivatives.

While iodoacetate and pyriv
hzsbeenfargreater interestinthe u
larger molecule which can serve a
2,1,3—benzoxadiazole (47), 7-fluor

(aminosulfonyl)-7-fluoro~2, 1 ,3 —barn

 

assay of sulfliydryls by reason of
The adducts exhibit fluorescence a
shill corresponding to environment
under borate buffer, pH 9.5, on

oxidation of the sulthydryls. For 1

 

 

 

17

are recognized, mapped to sequence by the Edman technique and/or mass spectrometry,
and related to specific segments of the protein.

A key to the success of this approach is to choose an appropriate derivatization
reaction and reagent for sulfliydryl groups. The reagent should be soluble in the reaction
medium and label sulfhydryl groups selectively, rapidly, and irreversibly under mild
conditions (ideally under weak acidic conditions to avoid sulthydryl/disulfide bond
exchange which is minimized at pH 2-6.5). Furthermore, the reagent should possess
strong UV or fluorescent absorption which does not overlap with the maximum
absorption of proteins, or easily attach a radioactive element to facilitate the recognition
of derivatized peptides after HPLC separation. Finally, the derivative of cysteine should
be distinguishable from other amino acids by the Edman degradation technique in which
the identification of amino acids exclusively relies on the retention time of PTH-
derivatives.

While iodoacetate and pyridylethylation are still extensively used (44-46), there
has been far greater interest in the use of this chemistry as a mechanism for introducing a
larger molecule which can serve as a sulfhydryl probe. Two derivatives of 7-fluoro-
2,1,3-benzoxadiazole (47), 7—fluoro-2,1,3-benzoxadiazole 4-sulfonate (SBD-F) and 4-
(aminosulfonyl)-7-fluoro-2,1,3-banzoxadiazole (ABD-F), have proven usefiil for the
assay of sulfhydryls by reason of their formation of fluorescent products (Figure 1.6).
The adducts exhibit fluorescence at a long wavelength (~515 nm) and show a spectral ,
shift corresponding to environmental hydrophobicity (48, 49). The reaction is performed
under borate buffer, pH 9.5, containing 1 mM EDTA to prevent metal-catalyzed

oxidation of the sulflrydryls. For the derivatization of sulfltydryls in large proteins, the

 

 

—

proteinshavetobedenann'edtom

51).

l ‘i
N
col .
/ N/
l |
sown,+ 502NH2

SBD-F ABD-F

Figure 1.6. Derivatization

Chin and Wold (52) emplc
reduce disulfide bonds and ABD-l
sulfhydryls and disulfide bonds i
apparently do not react with each c
general method for the complete

Uting the similar procedure, Kirk:

bonds and one free sulfltydryl group

Another reagent, 5-[2-((i
(IAEDANS) (Figure 1.7) has been 1
Recently, Stnrroek et a1 (58) used ‘

testicular angiotensin-eonverting er

from enzymatic digests by HPLC,

 

 

 

 

18

proteins have to be denatured to ensure that every SH can be attacked by the reagents (50,

 

51).
F. N F. N ‘ r N a
/ \ / \ / \ /
‘N/O ‘N/O sod::rate> ‘N /O 6 NE)
SIO3'NH4+ SIOzNHz $O3'Na+ SIOZNHQ
SBD-F ABD-F . 3313-311 ABD-SR

Figure 1.6. Derivatization of the cysteinyl group by SBD-F or ABD-F.

Chin and Wold (52) employed the combinatiOn of tributylphosphine (Bu3P) to
reduce disulfide bonds and ABD-F. to block free sulfliydryls to characterize both free
sulfhydryls and disulfide bonds in a. number of proteins. } Since the two reagents
apparently do not react with each other, their combination offers a convenient and quite
general method for the complete characterization of free and cross-linked cysteines.
Using the similar procedure, Kirley (53, 54) determined the location of three disulfide
bonds and one free sulfhydryl group in the B-subunit of (Na+,K*)-ATPase.

Another reagent, S-[2-((iodoacetyl)amino)-ethyl]naphthalene-1 -sulfonic acid
(IAEDANS) (Figure 1.7) has been proven powerful for sulfhydryl derivatization (55-57).
Recently, Sturrock et al (58) used this reagent to label a free cysteine residue in human
testicular angiotensin-converting enzyme (tACE). After isolating the fluorescent peptide

from enzymatic digests by HPLC, the sequence of the fluorescent peptides was mapped

 

 

 

by MALDI-PSD. The sequence

LACE.

r
rnzcc-

Figure 1.7. 5-[2-((iodoacetyl)arr

2. Affinity Chromatography
The only functional group

enable covalent bond which can
The sulflrydryl-containing protei
interaction (59). After enzymati
peptides do not bind to the colum
containing fragments are then or
structural characterization by Ednu
' interactions include (60): a) bind
binding to affinity media with tea
exchange, and (iii) binding to af

methods were originally used to pt

 

 

 

 

 

19

by MALDI-PSD. The sequence data established that Cys496 is a free thiol form in

tACE.

u
IHZCC—NH-CHZCHZ-NH

©©

SO3H

Figure 1.7. 5-[2-((iodoacetyl)amino)-ethyl]naphthalene-l~sulfonic acid (IAEDANS).

2. Affmity Chromatography

The only functional group for which conditions are available for the formation of
a stable covalent bond which can be Split under mild conditions is the sulfhydryl group.
The sulfhydryl-containing protein is attached to an' affinity column by covalent
interaction (59). After enzymatic digestion on column, the nonsulfhydryl-cbntaining
peptides do not bind to the column and can be washed away. The attached sulfliydryl-
containing fragments are then removed from the column and - subjected to further
structural characterization by Edman degradation and/or mass spectrometry. The affinity
interactions include (60): (i) binding to affinity media that contain heavy metals, (ii)
binding to affinity media with reactive disulfide that undergo facile sulfhydryl/disulfide
exchange, and (iii) binding to affinity media that contain chelated zinc. While those

methods were originally used to purify sulflrydryl-containing proteins, this approach was

 

 

  
  
  
 
 

recently used to locate cysteine
snlfliydryl-containing peptides (61
groups in high mass proteins bwa
But conditions have to be establ'

grontpseanbeattachedtothecol

3. Cleavage at Cysteine Residues

The third method utilizes a l
ofanSH group into an SCN gn
ream ofa protein with Ellman
showed that 2-nitro-5-thiocyanobe
sulfliydryls, which subsequently cle
residues under mildly alkaline conc
of 2-iminothiazolidine-4—carboxyly
residues, the cleavage reaction rest
alignment of which indicates the
potentially a very useful method, 1‘
because all but the N-terminal p
cuboxylyl group, which is not am
convenient method to remove the
Because cysteines are relatively sc
‘ Produces large fragments which 4

Wllacrylamide gel electrophoresi:

 

 

 

 

 

20

recently used to locate cysteine residues in proteins by mass mapping of bound
sulfhydryl-containing peptides (61). This approach is powerful for locating sulfhydryl
groups in high mass proteins because the HPLC separation of the digests is unnecessary.
But conditions have to be established and optimized under which all the sulfhydryl

groups can be attached to the column.

3. Cleavage at Cysteine Residues

The third method utilizes a cleavage reaction at cysteine residues. The conversion
of an SH group into an SCN group was first achieved in two stages by successive
treatment of a protein with Ellman’s reagent and then with cyanide (62, 63). Stark (64)
showed that 2-nitro—5-thiocyanobenzoic acid (NTCB) specifically cyanylates cysteine
sulfhydryls, which subsequently cleave at the N-terminal side of the cyanylated cysteinyl
- residues under mildly alkaline conditions to form an amino-terminal peptide and a series
of 2-iminothiazolidine-4-carboxylyl (ITC) peptides. If a protein contains 11 cysteine
residues, the cleavage reaction results in the formation of n+1 peptide fragments, mass
alignment of which indicates the number and location of cysteine residues. While
potentially a very useful method, it has seldom been used for sequence determination
because all but the N-terminal peptide become blocked by the iminothiazolidine-
carboxylyl group, which is not amenable to Edman degradation. Thus far, there is no
convenient method to remove the ITC, group from the cleavage products (65, 66).
Because cysteines are relatively scarce in proteins, cleavage at these residues usually
produces large fragments which can be mass mapped by sodium dodecyl sulfate-

polyacrylamide gel electrophoresis (SDS-PAGE) (63, 64, 67). However, peptide

 

 

fi

assignments using this approach an
5%) (64, 67). Papayannopoulos
spectrometry to sequence the NTC
isolated from Sarcophaga buIlala.
mass assignment and can be applier
tlie NTCB approach has primarily
proteins although selective identifi
bonds was also reported (68-73).
The cleavage reaction men
mas spectrometry are the theme 0

theclwvage reaction will be discus

C Disulfide Bond Assignment
Although the amino acid 5
corresponding cDNA, modification;
predicted accurately. Disulfide l
POSttranslaticnal modifications, is e:
As a result, most disulfide-containi
bonding is correct. Improved anal

0f S-S are thus important to m

However, although there is 800d m

   

 

21

assignments using this approach are often complicated by its poor mass accuracy ( error >
5%) (64, 67). PapayannOpoulos and Biemann (68) first used CID tandem mass
spectrometry to sequence the NTCB cleavage reaction products of a protease inhibitor
isolated from Sarcophaga bullata. Mass spectrometry provides much more accuracy for
mass assignment and can be applied to the sequencing of ITC blocked peptides. To date,
the NTCB approach has primarily been employed to locate total cysteines in various
proteins although selective identificatiOn of free cysteine in the presence of disulfide
bonds was also reported (68-73).

The cleavage reaction mentioned above and mass mapping of the products by
mass spectrometry are the theme of this dissertation. The chemistry and mechanism of

the cleavage reaction will be discussed in detail in chapter 2.

C. Disulfide Bond Assignment

Although the amino acid sequences of proteins are readily deduced from the
corresponding cDNA, modifications occurring to the protein after translation can not be
predicted accurately. Disulfide bonding, one of the most frequently encountered
posttranslational modifications, is essential for stabilization of protein’s tertiary structure.
As a result, most disulfide-containing enzymes have maximum activity only if disulfide
bonding is correct. Improved analytical methods for the rapid and definitive assignment
of SS are thus important to many areas of biochemical research and technology.

However, although there is good methods for quantifying the number of disulfide bonds

 

i

inproteins (41, 42), the unambiguo
taskwhich is sometimes only resoli
Two frequently used ten
exchange and disulfide bond sen
sulflli’dfltl/disulfide exchange (sue
both sulfliydryl/disulfide exchange
mildly alkaline media when stru
d'uulfide bond surnames. Figun

protein

 

Figure 1.8. Illustration 0

Sanger and his coworkers 1
insulin, recognized that an interc
Protein was subjected to partial
formation of mixed disulfides. S
disulfide exchange, disulfide bond

was minimal over the range of pH

   

 

22

in proteins (41, 42), the unambiguous determination of disulfide bonds remains a difficult
task which is sometimes only resolved by a main commitment of time and sample.

Two frequently used terms in cysteine chemistry are sulfliydryl/disulfide
exchange and disulfide bond scrambling (or disulfide bond interchange). Although
sulfhydryl/disulfide exchange (such as Ellman’s reagent) is the most specific reaction,
both sulfhydryl/disulfide exchange and disulfide bond scrambling in proteins occur in
mildly alkaline media when structures permit and result in mismatched nonnative
disulfide bond structures. Figure 1.8 shows the sulfhydryl/disulfide exchange in a

protein.

   

Figure 1.8. Illustration of sulfliydryl/disulfide exchange in proteins.

Sanger and his coworkers (74-76), in their pioneering work on the structure of
insulin, recognized that an interchange reaction took place with disulfide bonds when the -
protein was subjected to partial hydrolysis in cold concentrated HCl, leading to the
formation of mixed disulfides. Subsequent studies demonstrated that , like sulfhydryl/
disulfide exchange, disulfide bond scrambling also occurs in slightly alkaline buffers but

was minimal over the range of pH 2 to pH 6.5 (77, 78).

 

 

 

 

R'S-SR' + R"S

In alkaline and neutral met
from the hydrolytic cleavage of dis

[menu of alkylating agents. such

1. Classical Strategy

Multiple step en

Fractn

Modific

 

 

 

R'S-SR' + R"S-SR" ——-—-> 2R'S—SR"
In alkaline and neutral media, the reaction is catalyzed by thiols which can arise
from the hydrolytic cleavage of disulfides. Disulfide bond interchange is inhibited by the I

presence of alkylating agents, such as iodoacetate (79).

1. Classical Strategy

 

Multiple step enzymatic/chemical cleavage of protein

l

Fractionate peptides by HPLC

l

Reduce disulfide bonds

l

Modification of nascent SH groups

l

Separation of the resulting peptides

l

Sequencing by Edman degradation

Figure 1.9. Classical strategy for disulfide bond assignment of a protein.

 

 

 

As illustrated in Figure

remgnin'ng disulfide bond strucu
similar to the assignment of fre<
enzymes or chemical reagents bet
avoid or minimize disulfide scram
reversed-phase HPLC. Third, the
peptides are determined using Ed
identified peptides are related to sp
the methodology focused on the
specially various mass spectrom

claracteristic aspects of the strategj

2. Cleavage of Cystine Peptides
The choice of enzymes for
specificity of the enzyme to pro
identification process; the condi
capability of cleaving between ev
no more than one disulfide bond.
factors. The identification of di
Specific cleavage reagent is used (
with a variety of enzymes, such as

or less specifically, chymotrypsin.

highly specific for Arg and Lys resi

 

 

 

 

24

As illustrated in Figure 1.9, a well established approach (65, 80, 81) for
recognizing disulfide bond structure of proteins involves several steps which are very
similar to the assignment of free sulfliydryl groups. First, a protein is cleaved by
enzymes or chemical reagents between half-cystinyl residues under conditions that can
avoid or minimize disulfide scrambling. Second, the mixture of digests is separated by
reversed-phase HPLC. Third, the amino acid sequence or molecular masses of these
peptides are determined using Edman degradation or mass spectrometry. Finally, the

identified peptides are related to specific segments of the protein. Recent development of
the methodology focused on the identification of the disulfide-containing peptides,
especially various mass spectrometric techniques are more and more involved. The

characteristic aspects of the strategy will be discussed below.

2. Cleavage of Cystine Peptides
The choice of enzymes for cleavage is largely dictated by three requirements: the
specificity of the enzyme to produce well-defined fragments that can simplify the
identification process; the conditions to minimize disulfide interchange; and the
capability of cleaving between every half—cystinyl residue to obtain peptides containing
no more than one disulfide bond. Usually compromise has to be made among these three
factors. The identification of disulfide-containing peptides is facilitated greatly if a
specific cleavage reagent is used (10, 39, 82, 83). Proteins may be cleaved selectively
with a variety of enzymes, such as trypsin, staphyococcus aureus V8, cyanogen bromide,
or less specifically, chymotrypsin. Trypsin has been used most frequently because it is

highly specific for Arg and Lys residues. Unfortunately, trypsin has maximum activity at

 

 

 

 

pll 8.3, and is not active in acid.

enzymolysis. However, in spite of
published papers on the disulfide
cleavage reagent, among which
Chyrnotrypsin is less specific
conditions. staphyococcw aureus
4.0 and 7.8, cleaving on the GI
Cyanogen bromide is also used fn
Me. This reaction is very attractir
of side reactions, and because th
volatile. In addition, disulfide br
proteins in their native state are of
isnsed to open up or unfold the prr
isnot a particularly abundant cons
large peptides. Subsequent digesti
peptides (82).

Unfortunately, if proteins c
must be cleaved by non-specific r
lhemrolysin, and partial acid hydro
sllt‘tcific proteases it is difficult
SCtlmcttts of the protein simply by

case,multiple-step Edman degradat

0fthe disulfide-bridged peptides.

 

 

 

25

pH 8.3, and is not active in acid. As a result, disulfide scrambling may occur during
enzymolysis. However, in spite of the jeopardy, our survey indicated that among over 60
published papers on the disulfide assignment in various proteins, 70% used trypsin as a
cleavage reagent, among which 80% of the digestion was performed at pH>8.
Chymotrypsin is less specific than trypsin, and likewise active only under alkaline
conditions. staphyococcus aureus V8 is useful because it has maximum activity at pH
4.0 and 7.8, cleaving on the C-terminal side of Glu or Glu and Asp, respectively.
Cyanogen bromide is also used frequently to cleave proteins on the C-terminal side of
Met. This reaction is very attractive as it is highly specific for Met and is generally free
of side reactions, and because the reagents (cyanogen bromide and formic acid) are
volatile. In addition, disulfide bonds are stable during treatment with CNBr. Since
proteins in their native state are often resistant to enzymatic attack, cleavage with CNBr
is used to open up or unfold the protein, rendering it susceptible to enzymolysis. As Met
is not a particularly abundant constituent of proteins, cleavage with CNBr usually gives
large peptides. Subsequent digestion by other enzyme(s) is required to produce smaller
peptides (82).

Unfortunately, if proteins can not be cleaved by specific cleavage reagents, they
must be cleaved by non-specific reagents. Pepsin (maximum activity around pH 3.0),
thermolysin, and partial acid hydrolysis are useful for this purpose. However, using non-
specific proteases it is difficult to relate the disulfide-bridged peptides to specific
segments of the protein simply by determining their molecular masses by MS. In such a
case, multiple-step Edman degradation and/or MS/MS is a useful tool for characterization

of the disulfide-bridged peptides.

 

 

 

3. Purification of Cystine-Contai

Following digestion of a 1
original disulfide bonds remain inta
toidentify, by N-terminal amino :
dinilfide bonded peptide. Reverse
acetonitrile in 0.1% trifluomacetie :
rapid and employing conditions
nsuallydetected byUV at215 nm‘
tinm high-mass proteins is still 8
HPLC columns are used, because
hydrolysis.

An alternative method I
identification of disulfide-bonded
based on the chemical modification
Orpaper electrophoresis is carried
performic acid vapour. After th
repeated in the perpendicular di
migrate along a line at 45° to ea

formed cysteic acid peptides migra

then eluted and analyzed by amin

method is diagonal electrophoresis

 

 

 

 

 

3. Purification of Cystine-Containing Peptides

Following digestion of a protein to give a mixture of peptides in which the

original disulfide bonds remain intact, the next step is to pry each of these peptides and
to identify, by N-terminal amino acid and if necessary, partial sequence analysis, each
disulfide bonded peptide. Reversed-phase HPLC under acidic conditions ( a gradient of
acetonitrile in 0.1% trifluoroacetic acid, for example), is ideal for this purpose, being both
rapid and employing conditions favoring stability of disulfide bonds. Peptides are
usually detected by UV at 215 nm with acceptable sensitivity. The purification of digests
from high-mass proteins is still a challenging task even though microbore or capillary
HPLC columns are used, because so many peptide fragments might be produced after
hydrolysis.

An alternative method to reversed-phase HPLC for the separation and
identification of disulfide-bonded peptides is diagonal electrophoresis. This method is
based on the chemical modification of cysteinyl residues to cysteic acids. The thin-layer
or paper electrophoresis is carried out in the horizontal direction, and then exposed to the
performic acid vapour. After the performic acid is removed, the electrophoresis is
repeated in the perpendicular direction. All peptides unaffected by performic acid
migrate along a line at 45° to each direction of the electrophoresis, while the newly
formed cysteic acid peptides migrate off the diagnal. Peptides containing cysteic acid are
then eluted and analyzed by amino acid composition or sequences. An example of this

method is diagonal electrophoresis map of a peptic digests of or—chymotrypsin.

 

 

 

4. Identification of Disnlfide Fe
The conventional approac
after HPLC separation is very t
made by a comparison between c
the disulfide bonds. Peptide peaks
SS; for each such peak, two new
S,itislikelythat there will be at 1
However, this strategy depends
before and after reduction, and i
rodwed species are alkylated and
Performic acid oxidation (36) co
modification has the disadvantagr
residues will also have their elutio
halfcystine residues may be. identi
The advent of modem mass :
disulfide bonds in proteins more
bombardment (FAB), matrix-as
electrospray ionization (ESI) have
mass spectrometry, as shown in Fir
their unique masses and by comps
the signal corresponding to the

corresponding to the respective thi

 

 

 

 

 

 

4. Identification of Disulfide Peptides

The conventional approach for the identification of disulfide-containing peptides
after HPLC separation is very tedious and time consuming. The identification can be
made by a comparison between chromatograms of peptides before and after reduction of
the disulfide bonds. Peptide peaks that disappear upon reduction are presumed to contain
S—S; for each such peak, two new peaks are formed. If a single peptide has an internal S-
S, it is likely that there will be at least a small shift in the elution position upon reduction.
However, this strategy depends too much on the variation of hydrophobicity of peptides
before and after reduction, and is therefore not always reliable. More effectively, the
reduced species are alkylated and re-purified by HPLC followed by Edman sequencing.
Performic acid oxidation (36) could also be used to identify peptides with S-S, but this
modification has the disadvantage that peptides containing methionine and trytophan
residues will also have their elution positions altered. Alternatively, peptides containing
half-cystine residues may be. identified colorimetrically (52-54). .

The advent of modern mass spectrometric techniques has made the task of locating
disulfide bonds in proteins more tractable (10, 83). In the last decade, fast atom
bombardment (FAB), matrix-assisted laser desorption/ionization (MALDI), and
electrospray ionization (ESI) have widely been used for disulfide bond assignment. In
mass spectrometry, as shown in Figure 1.10, disulfide linked peptides are identified from
their unique masses and by comparison with the spectrum of reduced samples in which
the signal corresponding to the S-S linked peptide(s) is replaced by two signals

corresponding to the respective thiol peptide components, if inter-bridged, or shifted by

 

 

 

 

Figure 1.10. Current protocol

tvt0 mass units (dithiol) if infra-l
for identifying peptides because it

that becomes important when tl

hldrclysis of large proteins wlr

 

 

 

 

 

 

28

 

 

SH
S *3 Disulfide bonded
E protein
B s ~s E
Enzymic/chemical
digestions
ir
Mixture of peptides
SH '
[+Y + M + \’S—S7/ +npeptides
l /
\ ‘ \ /
\ l \ /
r \ ,
l r
\ l I
, Ml / identification by MS
l ///
/ 4/
/ /\

 

reduction and MS analysis

Figure 1.10. Current protocol for disulfide bond assignment by mass spectrometry.

two mass units (dithiol) if intra-bridged. Mass spectrometry is a particularly good tool
for identifying peptides because it does not require rigorously purified peptides, a feature
that becomes important when the amount of protein is small or when investigating

hydrolysis of large proteins where purification of peptides to homogeneity may be

 

 

 

 

 

difficult Furthermore, modem

MALDI-PSD, FAB/CIDIMS/MS,
provides confirmation of a specific
Takao er al (84) first
disulfide-containing peptides from
be determined Almost at the
strategy for the recognition of di
became an accepted method an
inhibitor of amylase (86) and h
group srrbwquently demonstrated
ribonuclease A and lysozyme
methodologies based on mass 5
containing peptides (39, 83).
Peptide—mapping by E81 (1
increasingly in the characterization
uscdhereisexactly same as in
sensitivity (low picomole to high
HPLC has made the purification
MALDI-MS provides the capacity
Wins and peptides. It is particu
mapping without the previous isola

interchain disulfide-containing pep

excellent tool for the direct diagno

 

 

 

 

 

 

29

difficult. Furthermore, modern mass spectrometric techniques, such as ESI/MS/MS,
MALDI-PSD, FAB/CID/MS/MS, permit the sequence determination of peptides which
provides confirmation of a specific peptide segment in a protein.

Takao et al (84) first recognized the potential of FAB-MS for identifying
disulfide-containing peptides from which the location of disulfide bonds in proteins may
be determined. Almost at the same time, Morris and Pucci (85) reported the same
strategy for the recognition of disulfide bond structure in insulin. Their protocol soon
became an accepted method and successfully applied to proteins such as bacterial
inhibitor of amylase (86) and human tissue inhibitor of metalloproteinases (87). Smith’s
group subsequently demonstrated the feasibility of the FAB/MS for S-S assignment in
ribonuclease A and lysozyme, ’ and further developed several supplementary
methodologies based on mass spectrometry to facilitate the identification of disulfide-
containing peptides (3 9, 83).

Peptide-mapping by ESI (88, 89) or MALDI-MS (90, 91) are being used
increasingly in the characterization and identification of disulfide bridges. The strategy
used here is exactly same as in the F AB-MS. However, they provide much more
sensitivity (low picomole to high femtomole). Besides, the easy interface of ESI to
HPLC has made the purification of the disulfide-containing peptides an easy task.
MALDI-MS provides the capacity of rapidly analyzing small quantities of mixtures of
proteins and peptides. It is particularly well suited to the determination of peptide mass
mapping Without the previous isolation or fractionation. The in situ disulfide cleavage of
interchain disulfide-containing peptide during the analysis by MALDI-MS provides an

excellent tool for the direct diagnosis of disulfide bonded peptides (90, 91). The poor

 

 

 

resolution of both E31 and MAL

identification of intrachain disu]
reduction of the S-S bond.

Bean and Carr (92) used
electronvolt) collisionOactivated di
mmes up to 2,000 Da containin
cleavage at the disulfide bridge wi
fragment, giving a characteristic
32u These peaks are remarkabl

products of a disulfide bond cleav

1 Other Methodologies for SS
Despite their many successe
locating disulfide bond linkages. I
inde alkaline solutions. The er
because the requirement to find 3
always of priority and most of the ;
Another problem is that the cleav
the locations of S-S can be deduce
separated by the amino acids r
reactions. This is especially the

example, trypsin is useful only if

Even when the half-cystinyl resid

 

 

 

 

3O

resolution of both ESI and MALDI instruments, however, could be a problem for the
identification of intrachain disulfide bond which has only two mass units shift after
reduction of the S-S bond.

Bean and Carr (92) used tandem mass spectrometry with high—energy (kilo-
electronvolt) collision-activated dissociation (CAD) to examine a number of peptides of
masses up to 2,000 Da containing a single inter-chain disulfide bridge. They observed
cleavage at the disulfide bridge with retention of zero, one, or two sulfurs on the charged
fragment, giving a characteristic triplet of peaks in the mass spectrum with separation of
32 u. These peaks are remarkably intense, so they can provide ready identification of

products of a disulfide bond cleavage.

5. Other Methodologies for S-S Assignment-

Despite their many successes, two problems remain in the traditional methods for
locating disulfide bond linkages. One problem is the disulfide scrambling which occurs
in mild alkaline solutions. The endeavor to minimize the scrambling are often fruitless
because the requirement to find a specific enzyme to produce recognizable fragments is
always of priority and most of the specific enzymes only work under alkaline conditions.
Another problem is that the cleavage reaction does not always give peptides from which
the locations of S-S can be deduced. This may be because half-cystinyl residues are not
separated by the amino acids required by specific chemical or enzymatic cleavage
reactions. This is especially the case for a protein containing adjacent cysteines. For
example, trypsin is useful only if the half-cystinyl residues are separated by Lys or Arg.

Even when the half-cystinyl residues are separated by amino acids suitable for cleavage

 

   
  
  
  
  
  
   
 

by specific enzymes, the reaction
highly folded, as it is when disulfi
for the reduced and carboxymethy
94).

if a protein fails to produ
method, partial acid hydrolysis,
nixtme of disulfide-containing
hdnoed (95). Partial acid hydro
ternperatrne, etc.) under which 0
hydrolysis is attractive as disulfid
Steric effects play less of a role i
under acid conditions, and beca
enzymes. However, due to the lo
evenamedium-sized protein (e. g.
of peptide fragments. Many overlz
the cleavage pattern is not predicta
identify peptide fragments, com]
assisted data interpretation have to

Another problem occurs whe
any enzyme. lhat is, it is unlike
cleave between the two cysteine re:

one disulfide bond. The resulting

Slmcture and there are two possible

 

 

 

31

by specific enzymes, the reaction may not proceed at an appreciable rate if the protein is
highly folded, as it is when disulfide bonds are intact. Hence, cleavage that occur readily
for the reduced and carboxymethylated protein may not occur in the native protein (93,
94).

If a protein fails to produce desirable fragments after digestion, an alternative
method, partial acid hydrolysis, was proposed to replace enzymatic digestion to produce a
mixture of disulfide-containing peptides from which the disulfide connections may be
deduced (95). Partial acid hydrolysis is performed under controlled conditions (time,
temperature, etc.) under which only limited peptide fragments are formed. Partial acid
hydrolysis is attractive as disulfide bonds are particularly stable in dilute or weak acids.
Steric effects play less of a role in acid hydrolysis because most proteins are denatured
under acid conditions, and because the catalyst, H3O+, is much smaller than proteolytic
enzymes. However, due to the low specificity of partial acid hydrolysis, the digests of
even a medium-sized protein (e. g., hen egg-white lysozyme) are a very complex mixture
of peptide fragments. Many overlapped disulfide-containing peptides were produced and
the cleavage pattern is not predictable. Therefore, even if mass spectrometry was used to
identify peptide fragments, complicated chromatographic separation and computer-
assisted data interpretation have to be applied to the definitive identification.

Another problem occurs when a protein has a -Cys-Cys- structure that is resistant to
any enzyme. That is, it is unlikely, if impossible, to find a‘ cleavage reagent that can
cleave between the two cysteine residues and produce peptide fragments that only contain
one disulfide bond. The resulting peptide always contains two disulfide bonds in its

structure and there are two possible isomers (suppose they are inter-chain disulfides).

 

 

 

Different approaches have

 

used the combination of organi
containing peptide and high-perf
assign the disulfide structure of
large investment of time and ma
Strategy takes advantage of multi
(99) demonstrated that during seqt
nophenylthiohydantoin (PTH) can
premably remains attached to tht
the second Cys, a PTH upper
dehydroalanine, which can cast}
sequencer. Zhang and Liang (100)
Httwentoxin-I that contains two adj

Gray (103-105) developed a
cRhine-rich small proteins. This
disulfide bonds and sequence anan

subjected to limited degree of re

 

 

 

32

-——C __
XS

0/ c\
\ I

\ /

Different approaches have been reported to solve the problem. Stults et al (96)
used the combination of organic synthesis of the possible isomers of a disulfide-
containing peptide and high-performance tandem mass spectrometry in their attempt to
assign the disulfide structure of recombinant relaxin. This method obviously requires '
large investment of time and material, and is therefore not widely accepted. Another
strategy takes advantage of multiple step Edman degradation (97-102). Nokihara et al
(99) demonstrated that during sequence analysis of a peptide containing a disulfide bond,
no phenylthiohydantoin (PTH) can be seen after cleavage of the first Cys residue, since it
presumably remains attached to the second Cys by the disulfide bond. After cleavage of
the second Cys, a PTH appears as a cysteine and a dithiothreitol adduct of
dehydroalanine, which can easily be distinguished from other PTH-AAs on the
sequencer. Zhang and Liang (100) used this strategy for studyingthe disulfide linkage of
Huwentoxin-I that contains two adjacent cysteine residues.

Gray (103-105) developed a new approach for disulfide mapping of tightly folded,
cysteine-rich small proteins. This approach utilized a combination of partial reduction of
disulfide bonds and sequence analysis of alkylated peptides. In this strategy, a protein is

subjected to limited degree of reduction under acidic conditions to form a series of

 

 

 

partially opened disulfide bonds

sulflrydryls, followed by sequence
bond(s) that had been reduced.
and has been successfully applied
However, one limit on the applica
fragments after digestion of the
demonstrated by lshibashi e! a1(l
In all, disulfide linkage assi
Two troublesome problems, the d
Cyscontaining protein and the p
present There is a considerabl
techniques for disulfide bond assi
areas in studying protein structur
answers to the problems in deter

protein chemists.

111. Matrix-Assisted Laser Decor

Since the 1960s mass spectrt
liltide and protein analysis, but (
c(Impounds real progress in the tie
ionization methods; plasma desor
bombardment mass spectrometry (

ttthrriques, matrix-assisted laser i

 

 

 

 

 

 

33

partially opened disulfide bonds that are separated by HPLC. Alkylation of free
sulfliydryls, followed by sequencer analysis, provided explicit assignment of the disulfide
bond(s) that had been reduced. This approach can minimize disulfide bond scrambling
and has been successfully applied to a number of proteins containing adjacent cysteines.
However, one limit on the application of the method concerns peptide size. Analysis of
fragments after digestion of the protein can hopefully overcome the drawback, as
demonstrated by Ishibashi et al (106).
In all, disulfide linkage assignment in proteins is still challenging protein chemists.
Two troublesome problems, the detemiination of disulfide bond arrangement in a -Cys-
Cys-containing protein and the prevention of disulfide scrambling, remain bothersome at
present. There is a considerable interest in the protein society in developing novel
techniques for disulfide bond assignment. Mass spectrometry is one of the most active
areas in studying protein structure. It is hoped that mass spectrometry can provide
answers to the problems in determination of disulfide linkage that remain bothering

protein chemists.

III. Matrix-Assisted Laser Desorption/Ionization Mass Spectrometry (MALDI-MS)

Since the 19603 mass spectrometry has been considered a promising technique for
peptide and protein analysis, but owing to the size, polarity and low volatility of these
compounds real progress in the field was only achieved after discovery of the desorption
ionization methods; plasma desorption mass spectrometry (PDMS) (l), and fast atom
bombardment mass spectrometry (FAB MS) (2). The introduction in 1988 of two new

techniques, matrix-assisted laser desorption/ionization (MALDI) (3) and electrospray

 

 

 

 

ionization (ESI) (4), has extended

scieneeiarbeyond the scope of P
very difiaent ideas and princip
difficulties with MS methods. i.
biomolecules, and shown consi
accurate determination of molec
combination with biochemical tec
lire first attempts to g
desorpnonfronimtion date back to
Which can be desorbed and ion
molecules are analyzed, more cncr
clergy flux leads to pyrolytic (
breakthrough came in the late
experimented with the use of a m
desorptionfronization of large, non
obtained when such samples were
large molar excess of an energy-ab:
not strongly absorb at the wavelet
this technique was conducted with
an The ejected ions were analyze
demonstrated that the ions generat
67,000 daltons could be formed

Despite lack of an established io

 

 

 

 

34

ionization (ESI) (4), has extended the applicability of mass spectrometry in the biological
science far beyond the scope of PDMS and FAB-MS. Although MALDI and E81 utilized
very different ideas and principles, both new techniques have overcome the main
difficulties with MS methods, i. e., the desorption and ionization of large and labile
biomolecules, and shown considerable promise in characterizing biomolecules by
accurate determination of molecular mass up to a few hundred thousand daltons. In
combination with biochemical techniques, numerous applications have been achieved (5).

The first attempts to generate ions of organic molecules by direct laser
desorption/ionization date back to early 1970s (107). However, the size of the analytes
which can be desorbed and ionized was limited to ~1,000 daltons. When larger
molecules are analyzed, more energy is required to desorb them. However, such a high
energy flux leads to pyrolytic or photochemical decomposition of analytes. The
breakthrough came in the late 19803 when Hillenkamp and Karas successfully
experimented with the use of a matrix, nicotinic acid (3, 108). It was discovered that
desorption/ionization of large, nonvolatile molecules such as proteins could be similarly
obtained when such samples were irradiated by a laser after being codeposited with a
large molar excess of an energy-absorbing “matrix” material, even though the analyte did
not strongly absorb at the wavelength of the laser radiation. The first experiment with
this technique was conducted with frequency-quadrupled Nd-YAG laser operated at 266
nm. The ejected ions were analyzed by a time-of-flight (TOF) mass spectrometer. It was
demonstrated that the ions generated from proteins with molecular mass range 10,000-
67,000 daltons could be formed and detected from picomolar amounts of analyte.

Despite lack of an established ionization mechanism, the dependence on the matrix

 

 

 

material in the desorption/ionized

laser desorptionftonization (MAL
powerful technique for accurate
groaerthan 200,000 Da. With the
(PSD) and delayed extraction (DE
formicmral elucidation of biopol
The majority of the work
immanents being available in th
(Chapter 2) was conducted on a
Biosystem Voyager Elite, which
Spectrafor structural analysis; thi:
extraction hardware and sofiwan

however, are based on the same pri

”he Matrix

Matrix is the key to the suc
discovery of nicotinic acid as t
meounds have been investigated
applicable. Among them, sinap
tinnamic acid, and a-cyano—4-hyt
(111-113). It was considered imp

maximum of the matrix compound,

 

 

 

 

35

material in the desorption/ionization process gave rise to the terminology matrix-assisted
laser desorption/ionization (MALDI). Today MALDI has developed into an extremely
powerful technique for accurate and sensitive analysis of molecular masses extending to
greater than 200,000 Da. With the introduction of two new techniques, post-source decay
(PSD) and'delayed extraction (DE), MALDI-TOP MS is also becoming an effective tool
for structural elucidation of biopolymers.

The majority of the work that will be presented was performed on two MALDI
instruments being available in the MSU/NIH mass spectrometry facility. Early work
(Chapter 2) was conducted on a Vestec 2000. Others were gained on a PerSeptive
Biosystem Voyager Elite, which can be Operated in the reflectron mode to acquire PSD
spectra for structural analysis; this instrument has recently been equipped with delayed
extraction hardware and software for improving resolution. The two instruments,

however, are based on the same principle.

A. The Matrix

Matrix is the key to the successful analysis of high-mass biomolecules- Since the
discovery of nicotinic acid as the first matrix, several hundred different organic
compounds have been investigated (109, 110). However, only a few of these are widely
applicable. Among them, sinapinic acid (SA), 2,5-dihydroxybenzoic acid (DHB),
cinnamic acid, and or-cyano-4-hydroxycinnamic acid (a-CHCA) are extremely useful
(111-113). It was considered important that the laser wavelength match the absorption

maximum of the matrix compound. All the matrices above have strong UV absorption in

—

the320-350 nm range and can be
tenancy-doubled Nd-YAG lase
include miscibility with the anal}
required for the dissolution of the
chemical composition that promot
protons to the analyte, nonreactivi
retire low heat of sublimation and

The physioochemical even
and their ionization in MALDI he
believed to serve to minimize sam1
incident laser energy, resulting in 1
gas phase. One model for the me
of matrix are induced to undergo l
submquent expansion of these n
isolated protein molecules into the
protein undergoes ionization throu;

Protesses that remain to be explain

B. Sample Preparation
Proper sample preparation i
(“’0- Matrix solutions are typi

mixtures at a concentration of :

 

 

36

the 320-350 nm range and can be used with a much cheaper nitrogen laser (337 nm) or
frequency-doubled Nd-YAG lasers (355 nm). Other important matrix characteristics
include miscibility with the analyte in the solid phase, solubility in the same solvents
required for the dissolution of the analyte, vacuum compatibility (low vapor pressure), a
chemical composition that promotes the ionization of matrix substituents that can donate
protons to the analyte, nonreactivity with the analyte, and other physical prOperties such
as the low heat of sublimation and a capacity to crystallize readily (114).

The physicochemical events leading to the transfer of proteins to the gas phase
and their ionization in MALDI have not yet been fully elucidated (115). The matrix is
believed to serve to minimize sample fragmentation from the laser beam by absorbing the
incident laser energy, resulting in the sample and matrix molecules being ejected into the
gas phase. One model for the mechanism (116, 117) assumes that the uppermost layers
of matrix are induced to undergo a phase transition fiom the solid to the gas phase. The
subsequent expansion of these matrix molecules into the vacuum drags the matrix-
isolated protein molecules into the gas phase. During the transfer to the gas phase, the
protein undergoes ionization through proton transfer reactions with the matrix by reaction

processes that remain to be explained.

B. Sample Preparation
Proper sample preparation is also critical for successful analysis by MALDI-MS
(118). Matrix solutions are typically prepared in water/ethanol or water/acetonitrile

mixtures at a concentration of 5-10 pig/pl (~40 mM), depending on the solubility

 

 

 

 

   
   

properties of the matrix. The
solvent that is miscible with the m
triflmroacetic acid (IPA) is frequ
matrix and analyte solutions are
~5,000:1 to 10,000:l, and final co
reportively. An aliquot of ~lul
(V12000), and allowed to dry by
warm air, or under vacuum. lnstet
he sample plate holder used in Va
pumit a much higher sample ti
codeposits from solution with thi
chamber by a probe and a vacuum
hand” in the Voyager Elite MALL
d1uation(l to 10 ns) pulses ofan L'

To date, the chemrsu’y' of
the choice of matrix and method 0
difierent matrices and solvents(11

122) afl‘ects the resolution and se

analyses under different conditions

 

 

37

properties of the matrix. The analyte is prepared at a concentration of ~0.1 pig/p11 in a
solvent that is miscible with the matrix solution (for peptides and proteins, aqueous 0.1%
trifluoroacetic acid (TFA) is frequently used to maintain the acidity of the solution). The
matrix and analyte solutions are mixed to give a final matrixzanalyte molar ratio of
~5,000:1 to 10,00021, and final concentration of ~40 nmol of matrix/ ~1pmol of analyte,
respectively. An aliquot of ~1ptl of the mixture is applied to MALDI-MS probe tip
(V T2000), and allowed to dry by either ambient evaporation, heating with a stream of
warm air, or under vacuum. Instead of the stainless steel probe tips used in the VT2000,
the sample plate holder used in Voyager Elite system can accommodate 100 samples and
permit a much higher sample throughput. During the drying process, the matrix
codeposits from solution with the analytes. The sample is inserted into the vacuum
chamber by a probe and a vacuum lock in the VT2000, and by an “automatic mechanical
han ” in the Voyager Elite MALDI-TOF mass spectrometer, and irradiated with short
duration (1 to 10 ns) pulses of an UV laser beam.

To date, the chemistry of “matrix assistance” remains incompletely understood;
the choice of matrix and method of application is still empirical. Sample preparation by
different matrices and solvents(119), matrix additive (120), and evaporation rate (121,
122) affects the resolution and sensitivity of MALDI. Optimal results require parallel

analyses under different conditions.

 

 

 

 

(L instrumentation

lnMALDl (Figure 1.”)
laser which is directed and fee
controlled by an attenuator and is
interaction of photons with the m
of the cocrystallized sample/man
spread of ions generated by Mr
moderating voltage or a reflectror
resolution Experimentally, a statir
sample by application of a two-star
nith regard to a closely spaced at
the some kinetic energy (assuming
towards long (12 m) field-free tin

MALDI is typically used in
because TOF is simple, cheap, an
MALDI. It has virtually no uppe
Which can produce very high m/z
its capacity to generate the entire
information, in contrast to a s

traverse the flight tube (T OF) is

relationship: TOF = L/l) = L(m/Zz

velocity, in is the mass of the ion

 

 

 

 

38

C. Instrumentation

In MALDI (Figure 1.11) the cocrystal of sample/matrix is irradiated by a pulsed
laser which is directed and focused by a prism and Optical lens. The irradiance is
controlled by an attenuator and is increased gradually until the threshold is reached. The
interaction of photons with the matrix and protein results in the desorption and ionization
of the cocrystallized sample/matrix from a metal surface. The initial kinetic energy
spread of ions generated by MALDI is large, so either a linear TOF with a high
accelerating voltage or a reflectron with an ion mirror (or both) is used to improve mass
resolution. Experimentally, a static electric field is imposed upon ions generated from the
sample by application of a two-stage high voltage (typically i 25 kV) to the sample probe
with regard to a closely spaced accelerating electrode. The ions are thus accelerated to
the same kinetic energy (assuming the initial kinetic energy is zero) by the electric field
toward a long (1-2 m) field-free time-of-flight (TOF) analyzer.

MALDI is typically used in conjunction with a time-of-flight mass analyzer (123),
because TOF is simple, cheap, and well-suited to the pulsed nature of laser desorption in
MALDI. It has virtually no upper mass range and is therefore compatible with MALDI,
which can produce very high m/z ions. Another advantage of the TOF mass analyzer. is
its capacity to generate the entire spectrum from every single laser shot without losing
information, in contrast to a scanning mass analyzer. The time required for ions to
traverse the flight tube (TOF) is dependent on their masses and is described by the
relationship: TOF = Up = L(m/22 eV)”, where L is the length of flight tube, to is the ion

Velocity, m is the mass of the ion, and V is the acceleration potential. Thus, low-mass

 

 

 

no. 77—].—

 

Pmbe Ti

Figure 1.11. Scheme of matrix-a:
Spectrometer (MALDI-TOF MS)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

39
Beam Splitter
#:2212533} Laser
/ / \ I
I :3' Transient
/// - Recorder
// Trigger
//
I] _
Sample // H I I J—
Probe \ 6'? G)
$

éf—‘érrr 6% ‘9 l

n \¥ _1 T‘\ FlightTube U

' \ T \ Detector

I Ion \

I Source |

\
\
\
\
Probe Tip

 

 

Figure 1.11. Scheme of matrix-assisted laser desorption/ionization time-of—flight mass
spectrometer (MALDI-TOF MS)

 

 

 

ions have a shorter flight time

spatially discrete individual ion
mlz ratio. A detector positioned
as each ion packet strikes it. A
yields a TOF spectrum. The di
pulse and common to all ions, an
proportional to (m/z)“" and can b
time, tot) into a m/z ratios axis
highly efficient because all ions 0
manned; they simply arrive at
recorded for a particular ion refle
ion source (such as the time/lo
distributions). The poor resolving
with the mass of the ionized ma
isotopic masses are utilized in the
detect certain protein modificati
masses. In spite of this, it is
Pmteins with molecular masses

accuracy for proteins above 40 k

1). Characteristic Features of

One of the important fea

MALDI is a typical soft ionizatio

 

 

40

ions have a shorter flight time than heavier ions. They are separated into a series of
spatially discrete individual ion packets, each traveling with a velocity characteristic of its
m/z ratio. A detector positioned at the end of the field-free flight-tube produces a signal
as each ion packet strikes it. A recording of the detector signal as a function of time
yields a TOF spectnun. The difference between the start time, triggered by the laser
pulse and common to all ions, and the arrival time of an individual ion at the detector is
proportional to (m/z)W2 and can be used to convert the x-axis of the spectrttm (ion arrival
time, tof) into a m/z ratios axis ( a conventional mass spectrum). The tof analyzer is
highly efficient because all ions of different m/z ratio arising from a single laser shot are
measured; they simply arrive at the ion detector at different times. However, The tof

recorded for a particular ion reflects many different initial conditions experienced in the

 

ion source (such as the time/location of ion formation and initial kinetic energy
distributions). The poor resolving power is reflected by peak broadening which increases
with the mass of the ionized macromolecules. Therefore, average masses rather than
isotopic masses are utilized in the MALDI-TOF measurement. This limits the capacity to
detect certain protein modifications and protein sequence variations, especially at high
masses. In spite of this, it is possible to achieve a mass accuracy of 0.1-0.01% for

proteins with molecular masses between 1 and 40 kDa, and with somewhat poorer

accuracy for proteins above 40 kDa ( 123).

D. Characteristic Features of MALDI-TOF MS
One of the important features of MALDI spectra is its simplicity (116, 118).

MALDI is a typical soft ionization technique that yields little fragmentation. Unlike ESI

 

 

 

spectra which are dominated by

in MALDI is the singly proto
also observed to varying degrees
or analyte/alkali metal interactio

Compmed to other io '

common in biological samples
unseparated, heterogeneous corn
mixtures or synthetic reaction mi:

The most striking feature
has been umd to measure protei
500,000 Da (124). MALDI is als
Spectra with as little as 1 femt
rcported (112). The analysis at I

obtained.

E. Novel Techniques in MAL
MALDI has been a su

and peptides, and other biologi

because it is a soft ionization te

introduction of post-source deca

technique (125, 126). is based 0

collision-induced decay taking P

 

 

41

spectra which are dominated by multiply charged species, the predominant analyte signal
in MALDI is the singly protonated species, though the doubly protonated molecule is
also observed to varying degrees. Other adducts arising from analyte/matrix interactions
or analyte/alkali metal interactions are of relatively low abundance.

Compared to other ionization techniques such as ESI or FAB, MALDI-MS is
relatively tolerant to contaminants, such as buffers, salts, and denaturants which are
common in biological samples (116). It has therefore the capacity to analyze
unseparated, heterogeneous complex mixtures, such as enzymatic or chemical digest
mixtures or synthetic reaction mixtures.

The most striking feature of MALDI-MS is its very large practical mass range. It
has been used to measure proteins and glycoproteins with molecular weight as high as
500,000 Da (124). MALDI is also a very sensitive technique for protein characterization.
Spectra with as little as l femtomole of protein applied on the probe tip have been
reported (112). The analysis at low picomole to high femtomole range can be routinely

obtained.

E. Novel Techniques in MALDI-TOF MS

MALDI has been a superior method for molecular mass determination of proteins
and peptides, and other biological materials, but lacks the capacity for structural analysis,
because it is a soft ionization technique. This deficiency has been overcome since the
introduction of post-source decay (PSD) analysis of MALDI-generated ions. The PSD
technique (125, 126). is based on mass analysis of product ions from unimolecular or

collision-induced decay taking place in the field-free region between the ion source and

 

 

 

 

 

 

thereflectron. Incontrast to a li

energy analyzer to differentiate i
velocity in the linear mode. lons

accelerated precursors are detect

 

precursor in the linear mode of
because the product ions have 1
reflectron mode of the TOF ins
PSD, like MS/MS techniques, car
sized peptides (@500 Da). Sinc
PSD has quickly evolved into a p
atthe low picomole scale.

Another limitation of
distribution of ions in the MALD
nearly independent of mass and
the analyte. In addition, when
presumably lost by collision wi
energy dispersion To attain hi
ion extraction. In contrast to
generated by the laser beam n
potential, in delayed extraction,
ionization and ion extraction ev

isfield-free during the delay. F

speller. Application of the ap

 

 

42

the reflectron. In contrast to a linear instrument, reflection instruments can be used as an
energy analyzer to differentiate ions that are otherwise detected as species with the same
velocity in the linear mode. Ions formed as a result of metastable decomposition of fully
accelerated precursors are detected at the same arrival time (same apparent mass) as their
precursor in the linear mode of TOF because they have the same velocity. However,
because the product ions have lower kinetic energy, they can often be resolved in the
reflectron mode of the TOF instrument by lowering potential of reflector. Therefore,
PSD, like MS/MS techniques, can provide full or partial sequence information of medium
sized peptides (<2500 Da). Since its commercialization about three years ago, MALDI-
PSD has quickly evolved into a powerful technique for sequence determination of peptide
at the low picomole scale. r

Another limitation of traditional MALDI-TOF MS is the broad energy
distribution of ions in the MALDI source. The initial velocity of desorbed analyte ions is
nearly independent of mass and the initial kinetic energy is proportional to the mass of
the analyte. In addition, when desorption occurs in a strong electric field, energy is
presumably lost by collision with the neutral plume, resulting in further mass-dependent
energy dispersion. To attain high resolution, a delayed extraction technique is used for
ion extraction. In contrast to conventional MALDI instruments in which the ions
generated by the laser beam near the surface of the sample probe are extracted by a dc
potential, in delayed extraction, a short time delay (<3 00 ns) is inserted between the laser
ionization and ion extraction event. The region between the repeller and extraction grid
is field-free during the delay. Following the delay, a pulsed potential is applied to the

repeller. Application of the appropriate pulse voltage provides the energy correction

 

 

 

 

 

 

necessary to simultaneously det
initial energy. The initially less

pulseisappliedandtraverseal

 

ions. An energy/spatial correcti
reach the detector plane sirnul
accuracy, and the quality of
reducing chemical noise, and m
lhe potential of delayed extrac

peptides (1271130) as well as for

F. Application of MALDI-TO]
Over the past few years, 1
yet available for macromolecula
achievements in protein, nucleotir
MALDI has been exte
structures of proteins derived fro
provides routine and reliable m
typically have a high degree of
extraneous compounds has bee
malaria parasite. This remarkab
now within the reach of mass 3

MALDI is an excellent m

modifications. Covalently mod'

 

 

 

 

43

necessary to simultaneously detect all ions of the same mass/charge regardless of their
initial energy. The initially less energetic ions are closer to the repeller at the time the
pulse is applied and traverse a longer segment of the electric field than more energetic
ions. Arr energy/spatial correction is thus provided such that all ions of the same m/z
reach the detector plane simultaneously. Delayed extraction improves resolution, mass
accuracy, and the quality of MALDI mass spectra by suppressing matrix background,
reducing chemical noise, and minimizing the effect of laser intensity on performance.
The potential of delayed extraction MALDI has been demonstrated for proteins and

peptides (127:130) as well as for oligonucleotides (131).

F. Application of MALDI-TOF MS

Over the past few years, MALDI has become among the most powerful methods
yet available for macromolecular characterization of living systems. A wide range of
achievements in protein, nucleotide, and glycobiology have been attained in a short time.

MALDI has been extensively used for the determination of primary covalent
structures of proteins derived from both natural and recombinant sources (132). MALDI
provides routine and reliable means to analyzing tryptic digests and glycoproteins which
typically have a high degree of heterogeneity (133). The high tolerance of MALDI to
extraneous compounds has been used to study the degradation of hemoglobin by a
malaria parasite. This remarkable achievement shows direct analysis of cell contents is
now within the reach of mass spectrometry (134).

MALDI is an excellent method of choice for dealing with protein posttranslational

modifications. Covalently modified protein N- and C- termini (135), disulfide bonds

 

 

 

(136, 137), phowhorylation (13

DNA interactions (14]) have all

A dramatic demonstratio
samples is shown with “protein
analysis of a mixture of pep
degradation (142). Each of the
canbeidentified fi'om the mass d

Another highly promisin:
identification by MALDI of Coc
electrophoresis (143-145). This
protein spot identification as the
protein quantities in 2D gel S]
proteins, poor protein recoveries r

In additional to the use r
analysis of combinatorial librarie
etal (146) used MALDI-MS for
peptides isolated from support-
monoclonal antibody was scree
peptides isolated were shown to

The capacity for mass

lagged behind that of oligopeptid

last few years (147, 148), especi

 

 

44

(136, 137), phosphorylation (138), glycosylation (139), lipidation (140), and protein-
DNA interactions (141) have all been studied by MALDI.

A dramatic demonstration of the ability of MALDI to analyze heterogeneous
samples is shown with “protein ladder sequencing” which involves the simultaneous
analysis of a mixture of peptide/proteins that have undergone a stepwise Edman
degradation (142). Each of the fiagments differs from the next by one amino acid and
can be identified from the mass difference between successive peaks.

Another highly promising technique in the last three years is the unambiguous
identification by MALDI of Coomassie-stained protein spots from two dimensional gel
electrophoresis (143-145). This technique is superior to Edman microsequencing for
protein spot identification as the latter often suffers from problems such as insufficient
protein quantities in 2D gel spots, widespread occurrence of N-terminally blocked
proteins, poor protein recoveries from gels, and other unknown factors.

In additional to the use of MALDI as a sequencing tool, its ability toward the
analysis of combinatorial libraries has also been explored. As an early example, Keough
et al (146) used MALDI-MS for the rapid sequence determination of biologically active
peptides isolated from support-bound combinatorial peptide libraries. An anti-gp120
monoclonal antibody was screened against a hexapeptide library and six of the eight
peptides isolated were shown to possess the exact recognition sequence for the antibody.

i The capacity for mass spectrometric analysis of oligonucleotides has generally
lagged behind that of oligopeptides. However, this position has changed markedly in the

last few years (147, 148), especially with the development of new matrices (149, 150).

 

 

 

 

 

 

investigation of a 50-mer is now
reports of the analysis of much lar
It is believed that MAL
biological studies. As is often
exciting breakthroughs may. in th

this suiting.

 

45

Investigation of a SO—mer is now somewhat routine (151, 152) and there are individual
reports of the analysis of much larger RNAs (15 3).

It is believed that MALDI-MS will play more and more important roles in
biological studies. As is often true in a new and rapidly developing field, the most
exciting breakthroughs may, in the end, occur in areas not even anticipated at the time of

this writing.

 

 

 

 

IV, References

l. Thorgersen, D- F , skowrons
Commurr., 60, 616-620(1974§

2. Barber, M, Bordoli, R S. S
(1981).

3 Karas,ir.,andrrrrrenkamp. F

4. Meng C. K, Mann, M». at
Spectrometry and Allied TOP‘

5 Burlingame, A. L, Boyd, R
(1996).

6. Carr, S. A, Roberts, G. D, .
Materials, C. N. McEwen a
(1990), Ch.3, p87.

7- Yarn S. C. B., Grinnell, B W

8- Carr, S. A, and Biemann, K .

9. Creighton, T. E., in Mecham
(1992).

10.Hirayama, K., and Akashi, S
Edited by Matsuo, T, Caprir
Sons Ltd, (1994).

11. Torchinsky, Y. M,

Word. (1981).

in Surfu.

 

 

 

46

IV. References

1.

10.

11

Thorgersen, D. F ., Skowronski, R. P., and Macfarlane, R. D., Biochem. Biophys. Res.
Commun, 60, 616-620(1974).
Barber, M., Bordoli, R. S., Sedgwick, R. D., and Tyler, A. N., Chem. Commun., 325

(1981).

. Karas, M., and Hillenkamp, F., Anal. Chem, 60, 2299-2301 (1988).

Meng, C. K., Mann, M., and Penn, J. B., Proc. 36th ASMS Conference on Mass
Spectrometry and Allied Topics, San Francisco, CA, (1988), p771.

Burlingame, A. L., Boyd, R. K., and Gaskell, S. J., Anal. Chem, 68, 599R—651R
(1996)

Carr, S. A., Roberts, G. D., and Hemling, M. E., inMass Spectrometry of Biological
Materials, C. N. McEwen and B. S. Larsen (Editors), Marcel Dekker, New York,
(1990), Ch.3, p87.

Yan, S. C. B., Grinnell, B. W., and Wold, F., Trends Biochem. Sci, 14, 264 (1989).

. Carr, S. A., and Biemann, K., Methods Enzymol., 106, 29-58(1984).

Creighton, T. E., in Mechanism of Protein Folding, Edited by Pain, R. H., IRL Press,
(1992)

Hirayama, K., and Akashi, S., in Biological Mass Spectrometry: Present and Future,
Edited by Matsuo, T., Caprioli, R. M., Gross, M. L., and Seyama, Y., John Wiley &

Sons Ltd, (1994).

. Torchinsky, Y. M., in Surfur in Protein; (Metzler, D., Translation Ed), Pergamon,

Oxford, (1981).

 

 

 

 

12. Jocelyn, P. C ., Biochemistry 0_
13. Jocelyn, P. C, in Methods in
W.B.,anf1th. O. W. Inc. (1'
14. Stark G. R.._ and Stein. \‘1'. H.
15. Swoboda G.. and Hasselbac

( 1973).

16. Oliveira, B., and Lamm..M. E
110211111113. Cutmingham B.
M., Biochemism'. 7. 1973-191
18. Lundblad, R. L.. in Technique
19.80yer. P. D.. JACS, 76. 4331
30.Kochkina_ V.‘M.. and Torchn
(1975).
21- Sakai, H., Anal. Biochem.. 26
22-Ya1nuchi, T., Yamamoto, S..
23. Gibson, Q. H., J. Biol. Chem.
24' 511111311, G. L., Arch. Biochem

25. Riddles, P. W., Blakeley R
(1983).

2 Q
6.G11bert, H. F., Adv. Enzymol

27. H
UXtable, R. J ., Biochemistry

. 2 -
g-Grassetn, D. R. and Mun
9 ay’

47

12. Jocelyn, P. C., Biochemistry of the SH Group, Academic Press, London, (1974).

13. Jocelyn, P. C., in Methods in Enzymol., Vol. 143, Academic Press, edited by Jakoby,
W. B., Griffith, O. W. Inc., (1987).

14. Stark, G. R., and Stein, W. H., J. Biol. Chem, 239, 3755 (1964).

15. Swoboda, G., and Hasselbach, W., Hoppe-Seyler ’s Z. Physio]. Chem, 354, 1611
(1973).

16. Oliveira, B., and Lamm,,M. E., Biochemistry, 10, 26-31 (1971).

17. Gall, W. E., Cunningham, B. A., Waxdal, M. J ., Konigsberg, W. H., and Edelman, G.
M., Biochemistry, 7, 1973-1982 (1968).

18. Lundblad, R. L., in Techniques in Protein Modification, CRC Press, (1995).

19. Boyer, P. D., JACS, 76, 4331 (1954).

20. Kochkina, V. M., and Torchinsky Yu, M., Biochem. Biophys. Res. Communs. 63, 392
(1975).

21. Sakai, H., Anal. Biochem., 26, 387 (1968).

22. Yamuchi, T., Yamamoto, S., and Hayaishi, 0., J. Biol. Chem, 250, 7127(1975).

23. Gibson, Q. H., J. Biol. Chem, 248, 1281 (1973).

24. Ellman, G. L., Arch. Biochem. Biophys., 82, 70 (1959).

25. Riddles, P. W., Blakeley, R. L., and Zemer, B., in Methods in Enzymol., 91, 49-60
(1983).

26. Gilbert, H. F., Adv. Enzymol., 63, 69-172 (1990).

27. Huxtable, R. J ., Biochemistry of Sulfitr, Plenum, New York, 1986.

' 28. Grassetti, D. R., and Murray, J. F., Arch. Biochem. Biophys., 119, 41 (1967).

 

 

 

 

 

29. Grassetti, 1). R. and Murray--
30. Stuchbury. T., Shipton. M.. N

l, A. L, and Suschitzky. H.. E
lllanson, 1.; and Ryden. L.. Prc

and Applications. VC H Press.
32. Gruen, L. C.. and Harrap. B. E
33.Gevondyan, .\'.. M., GeVond

FEBS Letters. 255(2). 265-26
34.Sun, Y.. Smith. D. L.. and Sh
35. Shipton. M. and Brooklehurs
36. Sun. Y.. and Smith. D. I... An
375318113. Blattler. W. A.. anr
ll-Feng, F. Bell. A.. Dumas.

Conference on Mass Spectror
3931111, Y., Bauer, M, 1),, Keo

Biology, Vol: 61: Protein ar

Chapman,J R, Humana Prer
0.ZalllZec,E.J,Gagc,D A a
366(1994)
1Thannhauser, T W. Konish
(1984).
42.
Thallnhauser, T W McWh
., <
322(1985).

 

48

29. Grassetti, D. R., and Murray, J. F. J. chromat'ogr. 41, 121 (1969).

30. Stuchbury, T., Shipton, M., Norris, R., Malthouse, J. P. G., Brocklehurst, K., Herbert,
J. A. L., and Suschitzky, H., Biochem. J., 151, 417 (1975).

31. Janson, J .; and Ryden, L., Protein Purification: Principles, High Resolution Methods,
and Applications, VCH Press, (1993), p252.

32. Gruen, L. C., and Harrap, B. SJ, Anal. Biochem., 42, 377 (1971).

33. Gevondyan, N., M., Gevondyan, V. S., Gavrilyeva, E.,E., and Modyanov, N., N.,
FEBS Letters, 255(2), 265-268 (1989).

34. Sun, Y., Smith, D. L., and Shoup, R. E., Anal. Biochem., 197, 69-76 (1991).

35. Shipton, M., and Brocklehurst, K., Biochem. J., 167, 799 (1977).

36. Sun, Y., and Smith, D. L., Anal. Biochem., 172, 130-138 (1988).

37. Singh, R., Blattler, W. A., and Collinson, A. R.,, Anal. Biochem., 213, 49-56(1993).

38. Feng, F., Bell, A., Dumas, F., and Konishi, Y., Proceedings of the 38th Annual
Conference on Mass Spectrometry and Allied Topics, Tucson, AZ, (1990), p273.

39. Sun, Y., Bauer, M. D., Keough, T. W., and Lacey, M. P., Methods in Molecular
Biology, Vol: 61: Protein and Peptide Analysis by Mass Spectrometry, edited by:
Chapman, J. R., Humana Press Inc., Totowa, NJ, (1996).

40. Zaluzec, E. J ., Gage, D. A., and Watson, J. T., J. Amer. Soc. Mass Spectrom, 5, 359-

366(1994)

41.'Thannhauser, T. W., Konishi, Y., and Scheraga, H. A., Anal. Biochem., 138, 181

(1984)

42. Thannhauser T. W., McWherter, C. A., and Scheraga, H. A., Anal. Biochem., 149,

322(1985)

 

 

 

 

 

43.11irose, M., Takahashi. N" 06
44. Makinen, A. L.. and Nowak. '
45. Wang, Z. X., Preiss. B.. and 1
46. Male, A. S., and Jones. B. L.
47. Toyo’oka T., and Imai. K.. A
48.103'o’oka T., and Imai. K.. A
49. Sueyoshi. T.. Miyata T.. I'
(10166), 97. 1811 (1985).
50. loyo’oka T.. and Imai. K. .4
51Kobayashi, 1.. Rude. 1.. Kar
Biochem. (Tokyo). 98. 1017 (?
53.Chin, C. C. Q.. and Wold. F..
53. Kirley, T. L.,J. Biol. C hem.
54- Kirley. T. L, Anal. Biochem.
55- Pardo, J. P., and Slayman. C.
56' Miyanishj, T., and Borejdo, l

57. Bishop, J. E., Squier, T. C..

(1988).
58.8mmock,E.D.,Yu X C “

95609566 (1996).

59. J
anson, 1.; and Ryden, L. Pr

and Applications,VCH Pres
s

 

49

43. Hirose, M., Takahashi, N., Oe, H., and Doi, E., Anal. Biochem., 168, 193-201 (1988).

44. Makinen, A. L., and Nowak, T., J. Biol. Chem, 264, 12148 (1989).

45. Wang, Z. X., Preiss, B., and Tsou, C. L., Biochemistry, 27, 5095 (1988).

46. Mak, A. S., and Jones, B. L., Anal. Biochem., 84, 432 (1978).

47. Toyo’oka, T., and Imai, K., Anal. Chem, 56, 2461-2464 (1984).

48. Toyo’oka, T., and Imai, K., Analyst, 109, 1003 (1984).

49. Sueyoshi, T., Miyata, T., Iwanaga, S., Toyo’oka, T., and Imai, K., J Biochem.
(Tokyo), 97, 1811 (1985).

50. Toyo’oka, T., and Imai, K., Anal. Chem, 57, 1931 (1985).

51. Kobayashi, T., Kudo, 1., Karasawa, K., Mizushima, H., Inoue, K., and Nojima, S., J.
Biochem. (Tokyo), 98, 1017 (1985).

52. Chin, C. C. Q., and Wold, F ., Anal. Biochem., 214, 128-134 (1993).

53. Kirley, T. L., J. Biol. Chem, 264, 7185-7192 (1989).

54. Kirley, T. L., Anal. Biochem., 180, 231-236 (1989).

55. Pardo, J. P., and Slayman, C. W., J. Biol. Chem, 264, 9373 (1989).

56. Miyanishi, T., and Borejdo, J ., Biochemistry, 28, 1287 (1989).

57. Bishop, J. E., Squier, T. C., Bigelow, D. J., and Inesi, G., Biochemistry, 27, 5233
(1988)

58. Sturrock, E. D., Yu, X. C., Wu, Z., Biemann, K., and Riordan, J. F ., Biochemistry, 35,

9560-9566 (1996).

59. Janson, J .; and Ryden, L., Protein Purification: Principles, High Resolution Methods,

and Applications, VCH Press, (1993), p252.

 

 

 

 

 

60. Kalef, E, Walfrsh, P. G.‘. and
61Zaluzec, E. J., Gage, D. A.
Conference on Mass Spectror
62. Catsimpoolas, -\.. and Wood.
63. Degani, Y., and Patchornik. r‘
64. Stark, G. R, .lletltotis m 1511:;

65_Biemann, K, in .t-lethods m

(1993), p119.

66. Schafier, M. 11.. and Stark, (

(1976).

67. Darbre, A., Practical Protett

p149.

68. Papayarmopoulos, 1A., and I

69. Mahboub, 8., Richard, C, Dr
(1986).

70. '
Altamirano, M. M., Plumbrid

(1992).

71.
Tang, C, Yuksel, K. 11. Jac

283. 12.19 (1990).

72. N
efsky, B, and Bretscher, A

73. ’
Mmard, P, Desmadril, M

B'
tochem., 185, 419-423 (198

50

60. Kalef, E.; Walfish, P. G.; and Gitler, 0, Anal. Biochem., 212,325-334( 1993)

61. Zaluzec, E. J .; Gage, D. A., and Watson, J. T., in Proceeding of the 42nd ASMS
Conference on Mass Spectrometry and allied Topics, Chicago, (1994), p20.

62. Catsirnpoolas, N., and Wood, J. L., J. Biol. Chem, 241, 1790 (1966).

63. Degani, Y., and Patchornik, A, Biochemistry 13, 1 (1974).

64. Stark, G. R., Methods in Enzymology, vol. 47, 129-132 (1977).

65. Biemann, K., in Methods in Protein Sequence Analysis, Plenum Press, New York,
(1993), p119.

66. Schafl'er, M. H., and Stark, G. R., Biochem. Biophys. Res. Commun, 71, 1040-1047
(1976)

67. Darbre, A., Practical Protein Chemistry--A Handbook, John Wiley & Sons, (1985),
p149.

68. Papayannopoulos, LA, and Biemann, K., Protein Sci. 1, 278-288 (1992).

69. Mahboub, S., Richard, C., Delacourte, A., and Han, K., Anal. Biochem., 154, 171-182
(1986)

70. Altamirano, M. M., Plumbridge, J. A., and Calcagno, M. L., Biochem., 31, 1153-1158
(1992)

71. Tang, C., Yuksel, K. U., Jacobson, T. M., and Gracy, R. W., Arch. Biochem. Phys,
283, 12-19 (1990).

72. Nefsky, B., and Bretscher, A., Proc. Natl. Acad Sci. USA, 86, 3549-3553 (1989).

73. Minard, P., Desmadril, M., Ballery, N., Perahia, D., and MOuawad, L., Eur. J.

Biochem., 185, 419-423 (1989).

 

 

 

 

 

 

74, Sanger, F ., Nature flondonl. l

75. Ryle, A. P., and Sanger. F.. B

76. Ryle, A. P., Sanger, P. Smith

77. Lumper, L., and Zahn. 11.. Ad

78. Spackman. D. 11.. Stein. W. l

79. Marti, T., Rosselet. S. J .. Tit.
(1987).

80. Stasnick P. E.. Hermodson.
(1984).

31. Haeliner-Gormley. L. Paren‘
83 (1985).

82' Yazdanpafast R.. Andrews. 1

25072513 (1987).

31. Smith, D. L., and Zhou, Z.. A

84.18180, T., Yoshida, M., Hon

Spectrom, 11, 549-556 (198.

8 .
5. Moms, H. R., and Pucci, P.

86I '
Akash1,S.; and Hirayama, K
87. Williamson, R. A Marsto
C, n,

88. '
KHShnamurthy, T.; and Haur

, , 0 e

51

74. Sanger, F ., Nature (Iondon), 171, 1025-1026 (1953).

75. Ryle, A. P., and Sanger, F., Biochem. J., 60, 535-540 (1955).

76. Ryle, A. P., Sanger, F., Smith, L. F., and Kitai, R., Biochem. J., 60, 541-556 (1955).
77. Lumper, L., and Zahn, H., Adv. Enzymol., 27, 199 (1965).

78. Spackman, D. H., Stein, W. H., and Moore, S., J. Biol. Chem, 235, 648-659 (1960).

79. Marti, T., Rosselet, S. J., Titani, K., and Walsh, K. A., Biochemistry, 26, 8099-8109

 

(1987).

80. Staswick, P. E., Hermodson, M. A., and Nielsen, N. C., J. Biol. Chem, 259, 13431
(1984).

81. Haeffner-Gormley, L., Parente, L., and Wetlaufer, D. B., Int. J. Pept. Protein Res, 26,
83 (1985).

82. Yazdanparast, R., Andrews, P. C., Smith, D. L., and Dixon, J. E., J. Biol. Chem, 262,

 

2507-2513 (1987).
83. Smith, D. L., and Zhou, Z., Methods in Enzymol., Vol. 193, 374-389 (1990).
84. Takao, T., Yoshida, M., Hong, Y. M., Aimoto, S., and Shimonishi, Y., Biomed. Mass.
Spectrom., 11, 549-556 (1984).
85. Morris, H. R., and Pucci, P., Biochem. Biophys. Res. Commun. 126, 1122 (1985).
86. Akashi, S., and Hirayama, K., Biomed. Environ. Mass Spectrom., 15,541-546 (1988).
87. Williamson, R. A., Marston, F. A.; and Angal, S., Biochem. J., 268, 267-274 (1990).
38. Krishnamurthy, T.; and Hauer, C. C., Biol. Mass Spectrom. 23, 719-726 (1994).

89. Bauer, M.; Sun, Y.; Degenhardt, C.; and Kozikowski, B., J Protein Chem. 12, 759-

764(1993)

 

 

 

90. Patterson. 8.; 811d Katta. V" A

9(Crimmins. D. L., saylor. M"
36111995).

92.Bean,M. P: and Carr. 5. A...

93.Chung, D.. Sairom. M. R.. z
(1975).

94.lhannhauser. T. W. Konishi
115-119(1987).

95..Zhou, Z.. and Smith. D. L. J.

96861115.]. 1.. Bourell. J. 11.. C
111-". Griffin. P. R.. Rinder
Spectrom., 19, 655-664 (1991

97.1011, J. W., Elzinga M., and

93. Serensen. H. H.. Thomsen. J
11713720 (1990).

99. Nokihara, K, Morita, N., Ya
(1992).

1011. Zhang, 1)., and Liang, 8. J.

10 '
1.Knshnamurthy, T., Hauer

Biol. Mass Spectrom., 23 71

102.
Tang, Y., and Selsted, M. E

 

52

90. Patterson, S.; and Katta, V., Anal. Chem. 66, 3727-3732 (1994).

91. Crimmins, D. L., Saylor, M., Rush, J., and Thoma, R. S., Anal. Biochem., 226, 355-
361(1995)

92. Bean, M. F .; and Carr, S. A., Anal. Biochem., 201, 216-226 (1992).

93. Chung, D., Sairom, M. R., and Li, C. H., Int. J. Peptide Protein Res, 7, 487-493
(1975)

94. Thannhauser, T. W., Konishi, Y., and Scheraga, H. A., Methods Enzymol., Vol: 143,
115-119 (1987).

95. Zhou, Z., and Smith, D. L., J. Protein Chem, 9, 523-532 (1990).

96. Stults, J. T., Bourell, J. H., Canova-Davis, E., Ling, V. T., Laramee, G. R., Winslow,
J. W., Griffin, P. R., Rinderknecht, E., and Vandlen, R. L., Biomed. Environ. Mass
Spectrom., 19, 655-664 (1990).

97. Fox, J. W., Elzinga, M., and Tu, A. T., Biochemistry, 18, 678-684 (1979).

98. Serensen, H. H., Thomsen, J ., and Bayne, S., Biomed. And Environ. Mass Spectrom.,
19, 713-720 (1990).

99. Nokihara, K., Morita, N., Yamaguchi, M., and Watanabe, M., Anal. Lett, 25, 513-533
(1992)

100. Zhang, D., and Liang, S., J. Protein Chem, 12, 735-740 (1993).

101. Krishnamurthy, T., Hauer, C. R., Prabhakaran, M., Freedy, J. G., and Hayashi, K.,

Biol. Mass Spectrom., 23, 719-726 (1994).

102. Tang, Y., and Selsted, M. E., J. Biol. Chem, 268, 6649-6653 (1993).

 

 

 

 

103.Gray, W. R.. Luque. F. A.. (
Reyes, A., Alford. 1.. Mclnl
Biochemistry, 23. 2796 4 1984
104. Gray, W. R., Protein Sci. 2.
105. Gray W. R., Protein Sci, 2.
106.15h1'bashi, J., Kataoka, 11..
Mizoguchi, A., Ishizaki. 11.. a
107. Posthumus. M. A., Kistema
Chem, 50. 985-991 (1978 ).
108.Karas. M., Bachmnn. D.. l
Processes. 78. 53-68 (1987‘).

109. Ehring, H.. Karas. M.. an
(1992).

110. Fitzgerald, M. C., Parr, G. 1

lll.Beavis, R. C., and Chait, B.
112. Strupat, K, Karas, M., ant
111, 89-102 (1991).

113.Beavis, R. C., Chaudhary
(1992).

114.
Juhasz, P., Costello, C. E..
409 (1993).

115. Chait,B.T.; and Kent 8 E

53

103. Gray, W. R., Luque, F. A., Galyean, R., Atherton, E., Sheppard, R. 0, Stone, B. L.,
Reyes, A., Alford, J ., McIntosh, M., Olivera, B. M., Cruz, L. J ., and Rivier, J .,
Biochemistry, 23, 2796 (1984).

104. Gray, W. R., Protein Sci., 2, 1732-1748 (1993).

105. Gray W. R., Protein Sci., 2, 1749-1755 (1993).

106.1shibashi, J., Kataoka, H., Isogai, A., Kawakami, A., Saegusa, H., Yagi, Y.,
Mizoguchi, A., Ishizaki, H., and Suzuki, A., Biochemistry, 33, 5912-5919 (1994).

107. Posthumus, M. A., Kistemaker, P. G., Meuzelaar, H. L. C., and Brauw, M. C., Anal.
Chem, 50, 985—991 (1978).

108. Karas, M., Bachmann, D., Bahr, U., and Hillenkamp, F., Int. J. Mass Spectrom. Ion
Processes, 78, 53-68 (1987).

109. Ehring, H., Karas, M., and Hillenkamp, F., Org. Mass Spectrom., 27, 472-480
(1992).

110. Fitzgerald, M. C., Parr, G. R., and Smith, L. M., Anal. Chem, 65, 3204-3211 (1993).

111. Beavis, R. C., and Chait, B. T., Rapid Commun. Mass Spectrom., 3, 432-435 (1989).

112. Strupat, K., Karas, M., and Hillenkamp, F., Int. J. Mass Spectrom. Ion Processes,
111, 89-102 (1991).

113. Beavis,.R. C., Chaudhary, T., and Chait, B.T., Org. Mass Spectrom., 27, 156-158
(1992).

114. Juhasz, P., Costello, (3. 13., and Biemann, K., J. Am. Soc. Mass Spectrom, 4. 399-
409 (1993).

115. Chait, B. T.; and Kent. S. B. H., Science, 257, 1885-1894 (1992).

 

 

 

116. Siuedak, G.. Proc. Natl. Acad.
117. Busch, K. L., J. Mass Spectror
118.Zaluzec. E. J.; Gage. D. As. a
6109-123 (1995).
119.Cohen. S. L., and Chait. B. T.
120. Preston, L. M.. Murray. K. 1-
55011993).
121.Vonn, 0.. and Roepstorff. P.
133.1-"mm. 0.. and Mann. M.. .1. .

1218433115. R. C.. and Chait. B.

 

124.11111enkamp. F.. Karas. M..
1198A(1991).

125. Kaufmann. R.; Kirsch, D.;
355-385 (1994).

126. Kallfmann, R.; Spengler, B
7, 902-910 (1993).

127. Brown, R. S., and Lennon

128.c61by,s.M..Kin8 T B

868 (1994).

 

54

116. Siuzdak, G., Proc. Natl. Acad. Sci. USA, 91, 11290-11297 (1994).

117. Busch, K. L., J. Mass Spectrom., 30, 233-240 (1995).

118. Zaluzec, E. J .; Gage, D. A.; and Watson, J. T., Protein Expression and Purification,
6, 109-123 (1995).

119. Cohen, S. L., and Chait, B. T., Anal. Chem, 68, 31-37 (1996).

120. Preston, L. M., Murray, K. K., and Russell, D. H., Biol. Mass Spectrom., 22, 544-
550(1993) '

121. Vorm, 0., and Roepstorfi, P., and Mann, M., Anal. Chem, 66, 3281-3287 (1994).

122. Vorm, 0., and Mann, M., J. Am. Soc. Mass Spectrom., 5, 955-958 (1994).

123. Beavis, R. C., and Chait, B. T., Proc. Natl. Acad. Sci. USA, 87, 6873-6877 (1990).

124. Hillenkamp, F., Karas, M., Beavis, R. C., and Chait, B. T., Anal. Chem, 63, 1193-
1198A (1991).

125. Kaufmann, R.; Kirsch, D.; and Spengler, B., Int. J. Mass Spectrom. Ion Proc. 131,
355-385 (1994). .

126. Kaufmann, R.; Spengler, B.; and Lutzenkirchen., Rapid Commun. Mass Spectrom.,
7, 902-910 (1993).

127. Brown, R. S., and Lennon, J. J., Anal. Chem, 67, 1998-2003 (1995).

128. Colby, S. M., King, T. B., and Reilly, J. P., Rapid Commun. Mass Spectrom., 8, 865-

868(1994)

129. Vestal, M. L., Juhasz, P., and Martin, S. A., Rapid Commun. Mass Spectrom., 9,

1044-1050(1995)

130. Whittal, R. M., and Li, L., Anal. Chem, 67, 1950-1954 (1995).

 

 

 

 

 

 

15111111581..Rosk9~ M- T" 3”
A.,Anal. Chem, 68. 941'946 1
132Carr. S. A., in Mass Spectre
Carr. S. A., Eds; Humana PICS
133.Billeci,1. M.. and Stults. J. '1
114. Goldberg, D. E., Slater. A.
Henderson G. B.. J. Exp. Mea
135. Specht B.. Oudenampsen-K
and Spener, 1.. J. Biotechnol.
136.Robertson J. G.. Adams. (
V'illafranca. J. J., Biocherrristr
137. Chang, 1. Y., Schindler. P.
(1995).
138.1180, P-C., Leykam, J .. Anc
219, 9-20 (1994).
139‘ Burlingame, A. L., Curr. Q
140.Wilcox, M. D., Schey, K

Biophys. Res. Commun. 212

14
1. Zhao, Y., and Chait, B. T.

142.Chait,B.T. Wang R B
, , ,, e2

143. Patterson, S. D

144. '
Erdjument,~Bromage H '

, and Aeber

Protein Sci., 3, 2435-2446 (

55

131. Juhasz, P., Roskey, M. T., Smirnov, I. P., Haff, L. A., Vestal, M. L., and Martin, S.
A., Anal. Chem, 68, 941-946 ( 1996).

132. Carr, S. A., in Mass Spectrometry in the biological sciences, Burlingame, A. L.,
Carr, S. A., Eds; Humana Press: Totowa, NJ, (1996)

133. Billeci, T. M., and Stults, J. T., Anal. Chem, 65, 1709-1716 (1993).

134. Goldberg, D. E., Slater, A. F. G., Beavis, B. T., Chait, B. T., Cerami, A., and
Henderson, G. B., J. Exp. Med, 173, 961 (1991).

135. Specht, B., Oudenampsen-Kruger, E., Ingendoh, A., Hillenkamp, F., Leaius, A. G.,
and Spener, F., J. Biotechnol., 33, 259-269 (1994).

136. Robertson, J. G., Adams, G. W., Medzihradszky, K. F., Burlingame, A. L., and
Villafranca, J. J ., Biochemistry, 33, 11563-11575 (1994).

137. Chang, J. Y., Schindler, P., and Chatrenet, B., J Biol. Chem, 270, 11992-11997
(1995).

138. Liao, P-C., Leykam, J ., Andrew, P. C., Gage, D. A., and Allison, J ., Anal. Biochem.,
219, 9-20 (1994).

139. Burlingame, A. L., Curr. Opin. Biotechnol., 7, 4-10 (1996).

140. Wilcox, M. D., Schey, K. L., Busman, M., and Hil'debrandt, J. 1)., Biochem.
Biophys. Res. Commun., 212, 367-374 (1995).

141. Zhao, Y., and Chait, B. T., Anal. Chem, 66, 3723-3726 (1994).

142. Chait, B. T., Wang, R., Beavis, R. C., and Kent, S. B., Science, 262, 89-92 (1993).

143. Patterson, S. D., and Aebersold, R., Electrophoresis, 16, 1791-1814 (1995).

144. Erdjument,-Bromage, H., Lui, M., Sabatini, D. M., Snyder, D. H., and Tempst, P.,

Protein Sci., 3, 2435-2446 ( 1994).

 

 

145.Clauser, K R., Hall. S. C.. Sr
Epstein. L. B., and Burlingan
(1995).

146Youngquist. R. S.. Fuentes
Commun. Mass Spectrom.. 8. '

l47.1\’ordhoff, E.. Kirpeltar 1..
(1996).

148.11mbach. P. A.. Cram P. F.
102(1995).

149.1itzgera1d. M. C .. Parr. C
(1993).

150. Wu, K 1.. Shaler. T. A.. anc

151.Nordhofi°. E.. Kirpekar. F.
Kristiansen, K, Roepstorff. 1

152.Nordhoff, 13., Karas, M.,
Lem) A., Muth, J ., Meier.
(1995).

153.141.6614, F., Nordhoff, E.

Karas, M., and Hillenkamp

56

145. Clauser, K. R., Hall, S. C., Smith, D. M., Webb, J. W., Andrews, L. E., Tran, H. M.,
Epstein, L. B., and Burlingame, A. L., Proc. Natl. Acad Sci. USA. 92, 5072-5076
(1995).

146.Youngquist, R. S., Fuentes, G. R., and Lacey, M. P., and Keough, T., Rapid
Commun. Mass Spectrom., 8, 77-81 (1994).

147. Nordhoff, E., Kirpekar, F ., and Roepstorff, P., Mass Spectrom. Rev., 15, 67-138
(1996).

148. Limbach, P. A., Crain, P. F., and McCloskey, J. A., Curr. Opin. Biotechnol., 6, 96-
102 (1995).

149. Fitzgerald, M. C., Parr, G. R., and Smith, L. M., Anal. Chem, 65, 3204-3211
(1993)

150. Wu, K. J ., Shaler, T. A., and Becker, C. H., Anal. Chem, 66, 1637-1645 (1994).

151.Nordhoff, E., Kirpekar, F., Karas, M., Cramer, R., Hahner, S., Hillenkamp, F.,
Kristiansen, K., Roepstorff, P., Lezius, A., Nucleic Acids Res., 22, 2460-2465 (1994).

152. Nordhoff, E., Karas, M., Cramer, R., Hahner, S., Hillenkamp, F., Kirpekar, F .,
Lezius, A., Muth, J., Meier, C., and Engles, J. W., J Mass Spectrom., 30, 99-112
(1995)

153. Kirpekar, F ., Nordhoff, E., Kristiansen, K., Roepstorff, P., Lezius, A., Hahner, S.,

Karas, M., and Hillenkamp, F., Nucleic Acids Res., 22, 3866-3870 (1994).

 

 

 

ASTRATEGY TO LOCA'
BY SPECIFIC CHEMICA'
ASSISTED LASER DESOl

MA!

1 Introduction

Cysteine sulfhydryl gror
using its free sulfhydryl (SH) 8
cysteine proteases, as the chelati
reshufiling enzymes. Pinpointi:
becomes an important strategy ft

1418e majority of new protein se

cloned cDNA (or, in the case

flagments), differentiating betwc

(disulfide bonds) is still cumbers

conventional techniques, cannot

Classical approaches for
modification of free sulfhydryls

d - .

envatlzed protein chemically r
b .

9 LC fractionation of the p1

11. 2
). Recently, peptide mappi

CHAPTER 2
A STRATEGY TO LOCATE CYSTEINE RESIDUES'IN PROTEINS
BY SPECIFIC CHEMICAL CLEAVAGE FOLLOWED BY MATRIX-
ASSISTED LASER DESORPTION/IONIZATION TIME-OF—F LIGHT

MASS SPECTROMETRY

I. Introduction

Cysteine sulfliydryl group (thiol) contributes to protein biological functions by
using its free sulfhydryl (-SH) group in the active site for enzyme. catalysis such as in
cysteine proteases, as the, chelating site for metal ions, and as the active site of disulfide
reshufiling enzymes. Pinpointing the number and location of cysteine residues thus
becomes an important strategy for determining protein structure. While it is true that the
large majority of new protein sequences are now deduced from nucleotide sequences of
cloned cDNA (or, in the case of prokaryotic proteins, from cloned genomic DNA
fragments), differentiating between free cysteine residues (sulfhydryl groups) and cystines
(disulfide bonds) is still cumbersome and tedious, because the cDNA technique, like other
conventional techniques, cannot identify posttranslational modifications.

Classical approaches for localizing protein free sulfhydryl groups usually involve
modification of free sulfhydryls before and after reduction of a protein, degradation of the
derivatized protein chemically or enzymatically into smaller peptide fragments, followed
by HPLC fractionation of the peptides and Edman degradation of the derivatized peptides

(1, 2). Recently, peptide mapping by mass spectrometry of proteolytic digests before and

57

 

 

afterHPLC fractionation also has
the sulfhydryl derivatization und
disulfide bond exchange is not ar
becleaved at both sides of cys
cysteine. Multiple-step enzymat
case which is tedious and requil
dcrivatizing reagents have to be
and to avoid the retention time (
a1113411112111 sequencer. Althoug
1110111310813th of sulfhydryl ;

feasibility remains to be tested.

11. Chemical Cleavage at CY“

Cleavage at cysteine re:
cyanolysis of disulfide bonds
However, because of the severa
and the elimination of thiocyan
not acceptable to protein chemis

Jacobson et al (7) s
SPeciflcally cyanylates cysteine
Side of the cyanylated cystein
WHO-terminal peptide and a :

(Figur
62.1). f a protein conta:

 

58

after HPLC fractionation also has been reported (3-5). However, this procedure requires
the sulflaydryl derivatization under a carefully controlled conditions so that sulfliydryl-
disulfide bond exchange is not an issue. Secondly, this approach requires that a protein
be cleaved at both sides of cysteine residues to give peptides that contain only one
cysteine. Multiple-step enzymatic or chemical digestions usually are performed in this
case which is tedious and requires a sample size on the order of nanomoles. Thirdly,
derivatizing reagents have to be chosen to facilitate HPLC separation with UV detection
and to avoid the retention time overlap of derivatized cysteine and other amino acids on
an Edman sequencer. Although other approaches such as methods based on affinity
chromatography of sulfhydryl groups have also been proposed (see chapter 1), their

feasibility remains to be tested.

11. Chemical Cleavage at Cysteine Residues

Cleavage at cysteine residues of peptide chains under alkaline conditions after
cyanolysis of disulfide bonds was first observed by Catsimpoolas and Wood (6).
However, because of the several side reactions, such as the reversibility of the cyanolysis
and the elimination of thiocyanylate, the yields were low and the cleavage reaction was
not acceptable to protein chemists.

Jacobson et al (7) showed that 2-nitro-5-thiocyanobenzoic acid (NTCB)
specifically cyanylates cysteine thiols. Subsequent cleavage occurs on the N-terminal
side of the cyanylated cysteinyl residue under mildly alkaline conditions to form an
amino-terminal peptide and a series of 2-iminothiazolidine-4-carboxylyl (ITC) peptides

(Figure 2.1). If a protein contains 11 cysteine residues, the cleavage reaction results in the

 

 

11,.\"—' ”C —
(A1 C59“:
(3118-91
NE

11

11,18- ~ 4

(B) Cleavag
pH 9, 37°C

//0
H3N+ ‘ ’ ‘T C\

11H
H3N+~ 4130-1411 -C -

Figure2l. Reaction betwee

, (A) Cyanylation ar.
cleavage reaction under alkal

59
sH
0 CH 0
11 1 2 11 //0
H3N+—- -—C —NH—CH —C —NH-—- ——c\
_ o-

HOOC
. \
(A) Cyanylation NTCB
pH8-9; rt, 15' ’ 02N@-SCN

V

 

 

NEC -S
O CH O
11 I 2 11 //0
H3N+ — - ——C —NH—-CH —C —NH —- —C
\ _
\ t 0
0H0
(B) Cleavage _
pH 9, 37°C, >16hr 0H
V
O HN S '
// O
H3N+——-—C\ + Y I II //0
OH HN C -NH - - —- \
. O'
('sz /O
H3N+—_ - --CO—NH -—C —_CO --NH-- - —C< (side reaction: B-elimination)

0‘

Figure 2.1. Reaction between cysteine residue and 2-nitro-5-thiocyanobenzoic acid
(NTCB), (A) Cyanylation and (B) Cleavage reaction. B-Elimination competes with
cleavage reaction under alkaline conditions.

 

fi

formation of n+1 peptide fragme
and location of cysteine residues
reaction can come to completion
proteins tested, Degani and Pat
alkaline conditions, competes wi
peptides showed marked variant
depending on the structural prop
ems; of NTCB and a low tom
the side reaction of displacemen
thiol groups (8).

Other reagents for cyan)
(9) and by Bnocklehurst et al (1
111GB, but their utility needs to

Jacobson et al (7) pro
mechanism implicates that hyt
influence of rising pH on the rat
ofOH' is shown in Figure 2.2. '

.out the nucleophilic attack (th

 

reaction probably proceeds only
This generates a much more I

cl’CIization and cleavage, withor

 

 

60

formation of n+1 peptide fragments, mass analysis of the fragments indicates the number
and location of cysteine residues. Although the original paper claimed that the cleavage
reaction can come to completion with little side reactions for most of the peptides and
proteins tested, Degani and Patchornik (8) found B-elimination, occurring also under
alkaline conditions, competes with the cleavage as an adverse reaction. The cyanylated
peptides showed marked variations in their tendency to undergo B-elimination reaction
depending on the structural properties of the respective peptides. Experimentally, a large
excess of NTCB and a low total concentration of thiol groups must be applied to avoid
the side reaction of displacement of CN- from S-cyanocysteine residues by the unreacted
thiol groups (8).

Other reagents for cyanylation of SH groups have been prepared by Wakselman
(9) and by Brocklehurst et al (10, 11). These reagents may offer some advantages over
NTCB, but their utility needs to be further examined.

Jacobson et al (7) proposed a mechanism for the cleavage reaction. This
mechanism implicates that hydroxide ion catalyzes the cleavage, as shown by the
influence of rising pH on the rate of cleavage reaction. A possible explanation of the role
of OH‘ is shown in Figure 2.2. The amide nitrogen is apparently too weak a base to carry
.out the nucleophilic attack (the pKa of a protonated amide is -4.0 or less), and the
reaction probably proceeds only after attack of OH‘ on the carbonyl carbon of the amide.
This generates a much more basic nitrogen, which then can participate in concerted

cyclization and cleavage, without formation of an acyliminothiazolidine intermediate.

 

 

C
l'IJ‘Iq+ - ' -
H
‘
\
H3N*—- —
x}
+ ._ . .—
H,N c\

Figure 2.2. A me

 

61

NEC -S

 

O CH Q
q I 2 II //
H3N+ —- — —NH—CH —C —NH—- —c\
0'
OH-
v
HEB
N\\ /S\ O
G
O ,, /
/NH——‘—CO—NH— —C\
H3N+— “(I3 \A 0'
OH

 

HN s
H3N+ — - ———C/;) + Y (ll) //0

\OH HN C —NH—- ——C\
0'

Figure 2.2. A mechanism for base catalyzed cleavage reaction.

 

 

fi

One featme of the reactir
related to the easy displacement
nitrothiophenolate. The cyanyla
are unreactive to NTCB. The
sulflrydryls can thus be achieve
reaction without prior reduction

While potentially a very
used for sequence determination
by the iminothiamlidineearboxy
to remove the ITC group from
cleavage by NTCB has been to
ntunber of side reactions that cc
the reversibility of the cyanolysi

Because cysteines are 1
usually produces large fragmen‘
polyacrylamide gel electrophore
using this approach are often co

Recently, Papayannopr

 

,\ Spectrometry to sequence the l
isolated from Sarcophaga bul.

could be used to sequence Pet

 

62

One feature of the reaction is that the reactivity of NTCB toward a thiol group is
related to the easy displacement by the surfur nucleophile of the good leaving group p-
nitrothiophenolate. The cyanylation is selective to free sulfhydryl groups, disulfide bonds
are unreactive to NTCB. The selective cyanylation and subsequent cleavage of free
sulfhydryls can thus be achieved in the presence of disulfide bonds by carrying out the
reaction without prior reduction of the protein (7).

While potentially a very useful method, this cleavage reaction has seldom been
used for sequence determination because all butthe N-terminal peptide becomes blocked
by the irninothiazolidinecarboxylyl (ITC) group. Thus far, there is no convenient method
to remove the ITC group from the cleavage products (12). Another reason that Cys
cleavage by NTCB has been used infrequently is that the reaction conditions cause a
number of side reactions that compete with the desired cleavage reaction. These include
the reversibility of the cyanolysis reaction and the B-elimination of thiocyanylate (8).

Because cysteines are relatively scarce in proteins, cleavage at these residues
usually produces large fragments which can be mass mapped by sodium dodecyl sulfate-
polyacrylamide gel electrophoresis (SDS-PAGE) (13-32). However, peptide assignments
using this approach are often complicated by its poor mass accuracy ( error > 5%).

Recently, Papayannopoulos and Biemann (34) used CID tandem mass
spectrometry to sequence the NTCB cleavage reaction products of a protease inhibitor
isolated from Sarcophaga bullata. Their work demonstrated that mass spectrometry

could be used to sequence peptides from the NTCB cleavage reaction in. spite of the

 

 

 

fi

blocked N-terminus. However.
and hence cysteine-rich proteins,
We developed a simple a
location of both cystine and fro
chemical cleavage reaction desc
peptides by MALDI-TOF MS (1
ordenatrn'ed protein is first allo

chain at free sulfliydryl sites.
disulfide bonds by tris(2-carbc
reagent for this purpose since, 1:
thiol groups that may react wit
thermltiug peptide fragments
deduce the number and locatic
the same cleavage reaction aft

on the number and location of
of the protein. The selectiy

sensitivity, speed, and mass a

approach. We also demonstrr

to remove excess reagents an

spectra. Experimental condi

sulflrydryl cyanylation, and

Zetabind membranes.

 

 

63

blocked N-terminus. However, tandem MS is practically limited to low-mass peptides,
and hence cysteine-rich proteins, because of the effective mass limit of CID (< 3000 Da).
We developed a simple and sensitive methodology to recognize the number and
location of both cystine and free sulfhydryls in peptides and proteins using the specific
chemical cleavage reaction described above, followed by mass mapping of the resulting
peptides by MALDI-TOF MS (35). In these analyses, as shown in Figure 2.3, a peptide
or denatured protein is first allowed to react with NTCB to selectively cleave the peptide
chain at free sulfhydryl sites. The reaction mixture is then subjected to reduction of
disulfide bonds. by tris(2-carboxyethyl)phosphine hydrochloride (TCEP) (36), a useful
reagent for this purpose since, unlike other common reducing reagents, it does not contain
thiol groups that may react with NTCB under the reaction conditions. Mass mapping of
the resulting peptide fragments by MALDI-MS provides information from which one can
deduce the number and location of the free sulfhydryls. Parallel experiments involving
the same cleavage reaction after reduction of disulfide bonds provide mass spectral data
on the number and location of total cysteines, which can be used to confirm the sequence
of the protein. The selectivity of the NT CB cleavage reaction combined with the
sensitivity, speed, and mass accuracy of MALDI-TOF MS are attractive features of this
approach. We also demonstrated the reactions performed on a Zetabind membrane used
to remove excess reagents and other salts and buffers in order to improve the MALDI
spectra. Experimental conditions are described for protein disulfide bond reduction,
sulflrydryl cyanylation, and cleavage reactions performed either in solution or on

Zetabind membranes.

 

 

 

/

HS2t

\
coo

Cyanylation
& Cleavage

”K? )H,‘
39

ZOK—i /
40 \jii/COO-

Reduction
by TCEP

19 NH +
10 3
2 KSH —/ 3.
0
¥_SH _f_/
30

MALDI
Analysis

CDAP/FCEP approach to
locate free sulfhydrst)

Figure 2.3. Chemical modi

 

 

r\

HS 20

C00‘

Cyanylation
& Cleavage

19

NH;
K180 _/

Reduction
by TCEP

 

V

19 10 NH;

20 \SH/ 39
30

MALDI
Analysis

 

V
CDAP/TCEP approach to

locate free‘ sulfhydryl(s)

  

Reduction
by TCEP

Cyanylation
& Cleavage

 

‘l NH3+

20 V29
30 39
V

MALDI
Analysis

 

v
TCEP/CDAP approach to
identify total cysteines

Figure 2.3. Chemical modification to identify free sulflrydryls and total cysteines.

 

 

 

ll]. Experimental Section

MALDI-TOF MS

MALDI mass spectra w
Vestec Corp, Houston, TX) t
nitrogen laser (337 nm, 3-ns pul
26W. Data were acquired in ‘
recorder with Z-ns resolution.
or internal calibration using star
(1 pmol of each standard) obt
experiments were performed to
C0,, leaukee, WI) as the nu
(vlv) solution of acetonitrile/ar
peptide or protein samples ant
allowed to air dry before beir
immobilized on an inert Zetal
mounted membrane containing
matrix/protein solution was a

acetonitrile/O. 1% TFA solutio

introduction into the mass spec

 

 

 

 

 

 

III. Experimental Section

MALDI-TOF MS

MALDI mass spectra were obtained on a Vestec LaserTec Research (VT 2000,
Vestec Corp, Houston, TX) time—of-flight (TOF) mass spectrometer equipped with a
nitrogen laser (337 nm, 3-ns pulse). The accelerating voltage in the ion source was set to
26 kV. Data were acquired in the linear mode of operation, using a SOD-MHz transient
recorder with 2-ns resolution. Time-to-mass conversion was achieved by either external
or internal calibration using standards of bradykinin (m/z 1061.2) and insulin (m/z 5734.5)
(1 pmol of each standard) obtained from Sigma Chemical Co. (St. Louis, MO). All
experiments were performed using or-cyano-4-hydroxycinnamic acid (Aldrich Chemical
Co., Milwaukee, WI) as the matrix. Saturated matrix solutions were prepared in a 50%
(v/v) solution of acetonitrile/aqueous 0.1% TFA and mixed in equal volumes with the
peptide or protein samples and applied to a stainless-steel probe tip. The mixture was
allowed to air dry before being introduced to the mass spectrometer. For a protein
immobilized on an inert Zetabind membrane, matrix solution was applied to a probe-
mounted membrane containing the adsorbed protein or reaction products (37). The
matrix/protein solution was allowed to dry under the saturated vapor of the 50%

acetonitrile/O. 1% TF A solution and cocrystallized on the membrane surface prior to

introduction into the mass spectrometer.

 

 

 

 

 

Chemicals
Tris(2-carboxyethyl)phos
Rockford, IL. Guanidine hy
Biochemicals (Indianapolis,
MHRQEAVDCLKKFNARRKI
peptides and proteins, 2—nitro-5-
pruchased from Sigma and use
NTCB was prepared in eitl
grnnidine-HCl solution and th
aqueous solution was prepared
base, and stored under N2 in -1

suitable concentration before us

Analysis of Free sulfhydryls l

Peptide or protein samp
M tris-HCl buffer (pH 8.0)
Cyanylation of the free sulth:
excess of NTCB (over free sul:
then raised to 9.0 by adding

protein was accomplished afie

bonds was performed at 37°C

 

 

Chemicals

Tris(2-carboxyethy1)phosphine hydrochloride. (TCEP) was purchased from Pierce,
Rockford, IL. Guanidine hydrochloride was a product of Boehringer-Mannheim
Biochemicals (Indianapolis, IN). Peptides DRVYIHPCHLLYYS and
MHRQEAVDCLKKFNARRKLKGA were purchased from Bachem, California. Other
peptides and proteins, 2-nitro-5 -thiocyanobenzoic acid (NTCB), and Trizma buffer, were
purchased from Sigma and used without further purification. The 0.01 M solution of
NTCB was prepared in either 0.1 M tris-HCl buffer or the buffer made in 6 M
guanidine-HCI solution and the pH of the solution was adjusted to pH 8.0. TCEP
aqueous solution was prepared as 0.1 M stock solution and adjusted to pH 8.0 with tris
base, and stored under N2 in -20°C freezer. The stock solution was further diluted to a

suitable concentration before use.

Analysis of Free sulfhydryls for Proteins in Solution (NTCB/TCEP procedure)
Peptide or protein samples (1 nmol to 100 pmol) were solubilized in 2~5 pl of 0.1
M tris-HCl buffer (pH 8.0) containing 6 N guanidine-HCI as denaturing agent.
Cyanylation of the free sulfliydryl groups was achieved by adding a 5-10 fold molar
excess of NTCB (over free sulfhydryls) and reacting at 37°C for 30 minutes. The pH was
then raised to 9.0 by adding 3.0 M tris base. Cleavage of the cyanylated peptide or
protein was accomplished after incubation at 37°C for 16 hours. Reduction of disulfide

bonds was performed at 37°C for 30 minutes by providing a 5-10 fold molar excess of

 

 

 

 

 

 

 

TCEP reducing reagent (over

  
   
  
  
  
  
 

volmne of a 50% (v/v) acetonitn'

Confirmation of Total Cystein

The above protein sampl
TCEP (over disulfide bonds) at
protein. Then a 5-10 fold mol
modify the total cysteines (nati
with3.0Mtrisbaseand the m'
dihrted 50—fold with a 50% (v/v

MS.

Microscale Analysis of Proteir
Figure 2.4 illustrates a
treatment and processing of
membrane (OAS-um pore siz
Products Inc, Denver, CO) war
A sample containing 10
tris-HCl buffered (pH 8.0) 6
molar excess of NTCB soluti

The membrane-attached prob

water that served to prevent

 

 

 

67

TCEP reducing reagent (over disulfide bonds). Samples were diluted with a 50-fold

volume of a 50% (v/v) acetonitrile/O. 1% TF A solution prior to analysis by MALDI-MS.

Confirmation of Total Cysteines (T CEP/NT CB procedure)

The above protein samples were allowed to react with a 5-10 fold molar excess of
TCEP (over disulfide bonds) at 37°C for 30 minutes to reduce the disulfide bonds in the
protein. Then a 5-10 fold molar excess of NTCB (over total cysteines) was applied to
modify the total cysteines (native and nascent sulfliydryls). The pH was adjusted to 9.0
with 3.0 M tris base and the mixture was incubated at 37°C for 16 hours. The sample was
diluted 50-fold with a 50% (v/v) acetonitrile/O. 1% TFA solution for analysis by MALDI-

MS.

Microscale Analysis of Proteins Immobilized on an Inert Membrane

Figure 2.4 illustrates a schematic flowchart for various protocols of chemical
treatment and processing of the sample prior to analysis by MALDI. A Zetabind
membrane (OAS-um pore size and SO-um thickness, purchased fiom Life Science
Products Inc., Denver, CO) was fixed to a probe tip as described previously (3 7).

A sample containing 10 to 15 pmol of protein (0.5111 volume, dissolved in 0.05 M
tris-HCl buffered (pH 8.0) 6 N guanidine-HCI) was applied to the membrane. A SO-fold
molar excess of NTCB solution was then added to the protein on the membrane surface.
The membrane-attached probe tip was put in a closed vessel containing a few drops of

water that served to prevent drying of the reaction mixture droplet on the membrane

 

 

 

surface. The vessel was incuba

h'l'CB. The reaction mixture
immersion in 0.01% NH.Ol-l f
other salts.

Afier air drying, 1.0 ul
HCI solution was applied to th
for 16 hours to promote cleav
SO-fold molar excess (~l ul,
applied to the membrane to
37"C for another 30 minutes.
io0.1% TFA aqueous solution

A 1.5-pl aliquot of 8
acid in a 50% acetonitrile/O. l ‘M
to resolubilize the peptides.
period) in a vessel saturated
solution so that the peptide:
effectively resolubilized fror
subsequently cocrystallized ‘
analyzed directly by MALDI.

Parallel experiments

membrane to confirm the to

 

68

surface. The vessel was incubated at 37°C for 30 minutes to promote cyanylation with
NTCB. The reaction mixture was allowed to dry on the membrane and then washed by
immersion in 0.01% NH4OH for 30 seconds to remove the excess NTCB reagents and
other salts.

After air drying, 1.0 1.11 of 0.05 M sodium borate-buffered (pH 9.0) 6 N guanidine-
HCl solution was applied to the membrane. The probe tip was incubated again at 37°C
for 16 hours to promote cleavage in a closed vessel containing a few drops of water. A
SO-fold molar excess (~1 pl, in this case) of TCEP (over. disulfide bonds) was then
applied to the membrane to reduce the disulfide bonds. The probe tip was incubated at
37°C for another 30 minutes. After air drying, the membrane was washed by immersion
in 0.1% TFA aqueous solution for 30 seconds to remove the reagents and buffer salts.

A 1.5-pl aliquot of a saturated matrix solution of or-cyano-4-hydroxycinnamic
acid in a 50% acetonitrile/O.1% TFA aqueous solution was applied to the dried membrane
to resolubilize the peptides. The solution was allowed to dry slowly (over one hour
period) in a vessel saturated by the vapor of the 50% acetonitrile/0.1% TFA aqueous
solution so that the peptides and proteins immobilized on the membrane could be
effectively resolubilized from the membrane surface by the matrix solution and
subsequently cocrystallized with the matrix. The resulting peptide mixtures were
analyzed directly by MALDI-MS (37). .

Parallel experiments with the TCEP/NTCB procedure also were performed on the

membrane to confirm the total number of cysteines in the protein.

 

 

 

 

 

593,36 ha mUHZ
Alllrlllll Alllllllll
9923.0 6 m m: 521% dogged“? 3v
Ame...

   

0G NH - E08

 

\

:5 Pa
uunwomue woumohu 530k“

 

 

 

69

 

 

 

.oSfiQEoE tog am no Hogan—088m 9:8.on .«o $3128 2885? How 080:8 3859850 4mm 059m

 

Ill'lllll

 

 

 

 

 

 

 

 

 

 

fleece an
a: 520 A3)
has...
>
58%. 6.?
azma amok
actuated Gov
mung...
065m
t 0er
/ owmaéflo A0
amen

 

 

 

0mm e $9 $2:
Zommu sow /

 

Ellie]

 

:22
Mesa

nouaNEnEOme 6W

 

 

 

 

$ng

5553 he
a: :83 Amv

A

\

£8.on
cassava

Emmet» @320

 

 

 

   

flfimbooo
gwagopoa
gem mu 3on
Comm
:3
mOHZ

 

 

 

 

 

coverage 2v

    

330% £805

 

 

 

lV. Results and Discussion

A. Cyanylation and Cleavage
Preliminary experimen
by MALDI-MS were conduct
amino acid sequences and cy
expected masses of their expec
Typical MALDI mass
cleavage products, obtained b
under the described experim
cyanylation, a cyano group re
the original peptide (Figure 2.
the observed value exactly m
at 37°C for 16 hours, the cyan:
2.53), two of them are cleave
respectively. The peak at m/z
with the cleavage reaction. '
the mass shift from the cyany
Likewise, Figure 2.6 :
remaining three peptides. T
peptide mixtures. Most of t]

such as the cyanylated/melt

 

70

IV. Results and Discussion

A. Cyanylation and Cleavage of Peptides by NTCB

Preliminary experiments involving NTCB cleavage of analytes prior to analysis
by MALDI-MS were conducted with sulfhydryl-containing peptides having different
amino acid sequences and cysteine locations. Table 2.1 lists the peptides and the
expected masses of their expected products after cyanylation and cleavage.

Typical MALDI mass spectra of the cyanylated peptide (MW 2427.9) and its
cleavage products, obtained by reacting 100 pmol of peptide with 500 pmol of NTCB
under the described experimental conditions, are shown in Figure 2.5. After the
cyanylation, a cyano group replaces hydrogen, resulting in a mass increase of 25 Da from
the original peptide (Figure 2.5A). The expected cyanylated peptide has a m/z of 2453.9;
the observed value exactly matches the expected data. After incubation in pH 9.0 buffer
at 37°C for 16 hours, the cyanylated peptide is replaced by three new components (Figure
2.5B), two of them are cleavage products with the expected m/z of 1099.5 and 1373.5,
respectively. The peak at m/z 23194.3 corresponds to the B-elimination product competing
with the cleavage reaction. The B-elimination product can be easily identified because
the mass shift from the cyanylated peptide is -59 Da.

Likewise, Figure 2.6 shows the MALDI spectra of the cleavage products from the
remaining three peptides. These spectra reveal general features of MALDI spectra of
Peptide mixtures. Most of the major peaks can be assigned to expected specific species,

such as the cyanylated/uncleaved peptide, the B-elimination products, and the specific

 

 

 

cocoa“: 95.2 a

 

 

71

 

 

mdmg . memo Tbomm Yemom v.83“ <©M§<§q0§<m0§

 

 

was? mam? méamm QMQN mevm . EOMmmo>mmA§m
memo mooa wee: v.32 v.83 mtwqdmogmm
meow fl IVS @an M @603 mgmm fl Zdemmgmxwoh
a is car age ..fiww .fim
. . . . 00553.0. 28 888
Lmifi

 

20388 m0 H Z 05 Saw 829.qu wgewcooebafifism 8m $533.“ mo 83? N? 333030 .mm 035.

 

 

 

 

 

_______._._fi._.—

5 0 1000

 

 

Figure 2.5. MALDI mass :
reaction with a 5-fold mola
products. ++ indicates pea
for the correlation of calcul

 

72

 

2453.9 A

 

4A A A AAAA. A_ A .A A AA A —‘ AA —A - ‘ ‘ A A L I
m '— - u '

 

1100.2
1373.9

 

 

 

 

 

 

 

 

Wimgr J hag

500 who 15 o 2(360 2500 3000
m/z '

 

 

Figure 2.5. MALDI mass spectra of peptide RYVVLPRPVCFEKGMNYTVR after
reaction with a 5-fold molar excess of NTCB, (A) cyanylated peptide and (B) cleavage
products. -H- indicates peaks corresponding to doubly charged species. See Table 2.1
for the correlatiOn of calculated and observed m/z values.

 

 

 

 

73

 

1507.1 A

 

 

t
t
15491)
160715

 

 

 

1746.2
B
0‘.
O\
2.: 3
++ I a
AM J.,] A
1659.7
C .

2568.0
627.3

9865
r
i 2

 

 

++
M:_WJ:1A - +Lt.' A A 4A..) L

.._

1000 1500 2000 2500 3000
m/z

 

 

Figure 2.6. MALDI spectra of cleavage products of peptides (A) TCVEWLRRYLKN,
(B) DRVYIHPCHLLYYS, and (C) MHRQEAVDCLKKFNARRKLKGA, after reaction
with 5-fold molar excess of NTCB. See Table 2-1 for the identities of the peaks.

 

 

 

 

 

cleavage produce Wit“ 3 ms:
used. Mass mapping by MAL
the peaks. However, the rt35130I
greatly, depending on the natu
cleavage products. m/z 986.5
amount however, the respon
different. Another observatior
elimination of cyanylated pept
of B-elirnination products (it:
cyanylated cysteine residue (
amino acid residues. such as
Products, perhaps though 3'
underwent little B-eliminatior
DRVYIHPQHLLYYS. Exp
changing reaction conditions
NTCB/analyte were not succ
have the same pH dependenc
elimination products can als.
reactions affect the yield of
information on the sulflrydry

Shut
tures 0n the cleavage rea

 

 

74

cleavage products, with a mass accuracy of better than 0.05% if internal standards are
used. Mass mapping by MALDI, therefore, unambiguously recognizes the identities of
the peaks. However, the responses of peptides on MALDI, i. e., the peak intensities, vary
greatly, depending on the nature of the peptides. For example, in Figure 2.6C, the two
cleavage products, m/z 986.5 and m/z 1659.7 respectively, should be of equal molar
amount, however, the responses of the two fragments on MALDI are considerably
different. Another observation is that, in addition to the expected cleavage products, [3-
elimination of cyanylated peptides always competes as a main side reaction. The yields
of B-elimination products depend on the structures of amino acids adjacent to the
cyanylated cysteine residue (19). Our experiments indicate that certain neighboring
amino acid residues, such as Pro, His, or Phe, may increase the yield of B-elimination
products, perhaps though steric hindrance. Therefore, while TQVEWLRRYLKN
underwent little B-elimination, the [El-elimination product predominated for the peptide
DRVYIHPQHLLYYS. Experiments carried out to minimize this side reaction by
changing reaction conditions such as pH, temperature, reaction time, and molar ratio of
NTCB/analyte were not successful, because both cleavage and B-elimination reactions
have the sarrie pH dependency. Fortunately, both cyanylated/uncleaved peptides and [3-
elimination products can also be detected by MALDI-MS. Therefore, although these
reactions affect the yield of the NTCB cleavage reaction, they provide complementary
information on the sulfhydryl location and cleavage pattern. The effects of amino acid

structures on the cleavage reaction will further be discussed in detail in Chapter 3.

 

 

 

 

 

Price (14) reported tha'
during the cyanylation of sulfh)
between protein sulfhydryl gror
also observed this side reactior
experimental conditions. No

mixed disulfides has been obse

B. The Effects of pH and Tel
Although the cleavage

systematic study and optimize
reaction. The experimental co
protocol proposed by J acobsor
9.5 buffer for 16-80 hours alt
to promote the cleavage (38).

by conducting the cleavage re
and 9.5 buffer solutions, 1
cValidation and cleavage car
using tIiS-HCl or sodium bor;
be cOmpleted in a 5-fold m
minutes at 37°C, the completr

Promote both cleavage and B

 

75

Price (14) reported that another concurrent reaction may occur less frequently
during the cyanylation of sulfliydryl groups and lead to the formation of a mixed disulfide
between protein sulfhydryl group and NTCB reagent. Denslow and Nguyen (3 8) recently
also observed this side reaction. But such a side reaction seems to be minor under our
experimental conditions. No trace of the products corresponding to the formation of

mixed disulfides has been observed on the analysis by MALDI-MS.

B. The Effects of pH and Temperature on the Kinetics of the Cleavage Reaction
Although the cleavage reaction by NTCB was proposed two decades ago (7), little
systematic study and optimization have been carried out regarding to the kinetics of the
reaction. The experimental conditions used today are exactly the same as that used in the
protocol proposed by Jacobson et al (7). Typically, the cleavage is performed in pH 8.0 ~
9.5 buffer for 16-80 hours although higher temperature was also occasionally employed
to promote the cleavage (3 8). We have studied the effects of pH on the cleavage reaction
by conducting the cleavage reaction on the above model peptides in pH 7.5, 8.0, 8.5, 9.0,
and 9.5 buffer solutions, respectively. Time-course studies indicated that both
cyanylation and cleavage can be carried out under mildly alkaline conditions (pH 8-9)
using tris-HCl or sodium borate as buffers. While cyanylation of sulfliydryls can easily
be completed in a 5-fold molar excess of NTCB solution even at pH 7.5 within 30
minutes at 37°C, the complete cleavage of the cyanylated peptide chains takes as long as
16‘ hours even at pH 9.0. Experimental results showed that higher pH can greatly

Promote both cleavage and B-elimination, while the relative yields of the two competitive

 

 

 

 

 

 

 

reactions do not vary Significa
resulted in incomplete cleavag
resulting from long hours Of in'
It should be pointed out that thi
from peptide to peptide, depen
tocysteines. Efficient cleavag
The cleavage reactior
temperature (e. g. 50°C) has
were usually conducted in me
reaction system by dialysis or
analysis on microscale (pico
although higher temperature .
after long hours of incubatior
of solvents. In addition, or
re“milling reagents (e. g.,

it
mperatures have to be caret

CI . o
Identification of Free Sr

Proteins may contain
f
011118. The NTCB/TCEP an

to '
determine the free sulth)

76

reactions do not vary significantly over a wide pH range. The lower pH (< 7.5) often
resulted in incomplete cleavage (cyanylated/uncleaved product) and other side reactions
resulting from long hours of incubation, as will be discussed in greater detail in chapter 3.
It should be pointed out that the kinetics of both cleavage and B-elimination vary greatly
from peptide to peptide, depending considerably on the structure of amino acids adjacent
to cysteines. Efficient cleavage requires careful monitoring of the reaction.

The cleavage reaction was previously performed at 37°C, although higher
temperature (e. g., 50°C) has also been used. However, previous cleavage experiments
were usually conducted in macroscale (mg) after removing the excess reagents from the
reaction system by dialysis or other purification procedures (7, 13-20). Our results from
analysis on microscale (picomole of samples in microliters of solution) indicated that
although higher temperature can accelerate the cleavage, it is not recommended because
after long hours of incubation, the reaction vial can easily get dry due to the evaporation
of solvents. In addition, our cleavage reaction is performed without removal of the
remaining reagents (e. g., NTCB), the side reactions caused by NTCB at higher

temperatures have to be carefully considered.

C. Identification of Free Sulfliydryls and Total Cysteines in Proteins
Proteins may contain free sulfluydryls, disulfide bonds, or a combination of both
forms. The NTCB/TCEP and TCEP/NTCB procedures described in Figure 2.3 were used

to detennine the free sulfl’rydryls and total cysteines, respectively. Figure 2.7 lists the

 

 

 

 

1-—-'-T—“"'T
SH Sl
Sp'
8
1 l 25 m:
22 1
SH
66 i
1 k
l 10
s
[3

macaw

SH SHQ

Figure

77

 

 

 

 

 

 

 

AcNH

 

 

 

 

18 39 44 47 77 97
l | I I T
SH SH SH SH SH
Spinach Ferredoxin (MW. 10483)
S S s .19
l 25 56 I 95 I 212
22 | I 63 I 153 200
SH S S
Papain (MW. 23426)
S ‘ H
66 I i i 160 162
I 106 119 121 1
S S
B-Lactoglobulin A (MW. 18368)
11 3O 68 i i 344 367 382 385
| | 73 120 I l
SH SH (3 03 SH SH
Ovalbumin (MW. 42699)
1 “i4 141
SH
oc-Hemoglobin A (MW. 15053)
93 112 1 4 6

la la.

B-Hemoglobin A (MW. 15954)

Figure 2.7. Structural features of model proteins.

 

 

Table 2.2. Mass assignrr

4—q

NTCl

#

Fragment

%

————

Protein

Ferredoxin‘ I - I 7
18-38
3943
44.-46
47-76
77-97
18-43
44-76

Papain 1-24

25-212

 

Ovalbumin 1_ 10

11-29
30-366
367-381
382-385
I-29
I-29
367-385
367-335

B'LactOg.

l- 12
lObulin A 0

121-162

Omen‘(Iglobint‘
‘Hmoglobrn

 

78

Table 2.2. Mass assignment of peptide fragments after different treatments

 

NTCB/TCEP Treatment TCEP/NTCB Treatment

 

 

 

Pr oteln Fragment [M+H]+CI - [M+I-1]+obi Fragment [M+H]+calc [M+I‘I]+ob§_
Ferredoxina 1-17 1839.1 1333.8
18-38 2352.2 2347.7
3943 517.5 ndb
44-46 320.3 320.0
47-76 3337.6 3337.6
77-97 2332.5 2332.0
18-43 2792.00 2792.8
44-76 3580.8c 3581.5
Papain 1-24 2621.5 2622.1 1-21 2372.7 2371.9
225-212 20848 nd 2224 290.3 288.4
25-55 3547.0 3546.7
56-62 880.9 880.9
63-94 3794.2 3792.4
95-152 6309.2 6308.9
153-199 5039.7 5038.9
200-212 1503.7 1503.5
Ovalbumin 1-10 1011.1 nd 1-10 1011.1 nd
11-29 2378.7 2378.5 11-29 2378.7 2377.4
30-366 37684.7 nd 30-72 4762.5 4761.4
367-381 1715.1 1715.2 73-119 5427.1 5426.1
382-385 429.5 429.4 120-366 27581.1 nd
1-29 3312.8c 3312-3 367-381 1715.1 1714.3
129 337134 3371-5 382-385 429.5 nd
367-385 20676.: 2067-7 129 3312.8° 3311.2
367-385 2126 68 2126-6 1-29 3371.8d 3370.3
. 367-335 2067.6c 2068.1
B-Lactog— 1420 4905.7 4901.1 1-65 7247.4 7238.2
lobulin A 121-162 13506.7 13510 66-105 4651.6 4653.5
106-118 1461.6 1462.1
119-120 274.3 nd
121-159 4551.3 4548.1
160-162 396.5 nd
or-Hemoglobina 1-103 1 1073-6 1 1065
B-Hemoglobin 104-141 4097-3 4093-0
192 9919.0 9915.0
93-1 1 1 2207.6 2207.0
112-146 3828.4 3831.6

_

a Only NTCB treatment was applied to spinach ferredoxin and hemoglobin samples since the proteins
only contain free sulfliydryl groups.

b Not detected by MALDI-MS.

° B-elimination products.

d Cyanylated/uncleaved peptides.

 

 

 

structural features of the prote

observed masses of the fragmer

I. Spinach Ferredoxin

Spinach ferredoxin cont
biological systems. some of thi
uith Feand/or M0 (41). The
applied to the protein solution
complexes. The MALDI mass
of ferredoxin reacted with 2.5
the MALDI probe tip). sh!
conesponding to 39-43. The

good agreement, with a relatii

the TOP instrumnt, the peak

with the peak at m/Z 2332.5

Ill/Z 2347.7 deviated from thi

1
he peak at m/z 1820.1 is at

183 ‘
9-1) dunng the cleavage 1

m/
22792.8 and 3581.5 cm

44-76' ‘
1n Wthh B-elimination

79

structural features of the proteins studied by the methodologies. The calculated and

observed masses of the fragments from these proteins are listed in Table 2.2.

1. Spinach Ferredoxin

Spinach ferredoxin contains 5 free sulfhydryls and no disulfide bonds (3 9, 40). In
biological systems, some of the free sulfl1ydryl groups in ferredoxin can form complexes
with Fe and/or Mo (41). Therefore, prior to treatment with NTCB, 1 mM EDTA was
applied to the protein solution to liberate the free sulfhydryls from possible S-Fe or S-Mo
complexes. The MALDI mass spectrum in Figure 2.8, recorded fiom a 100-pmol sample
of ferredoxin reacted with 2.5 nmol of NTCB (~1.5 pmol of the analyte was applied to
the MALDI probe tip), shows all the protonated peptide fragments except that
correSponding to 39-43. The calculated and observed [M + H]+ values (Table 2.2) are in

good agreement, with a relative mass deviation of < 0.1%. But, due to poor resolution of

the TOF instrument, the peak at m/Z 2352.5 (corresponding to fragment 18-3 8) overlaps
with the peak at m/z 2332.5 (corresponding to fragment 77-97). The observed peak at
m/Z 2347.7 deviated from the calculated value by 48 Da, indicating much higher error.
The peak at m/Z 1820.1 is attributable to the dehydration product of fragment 1-17 (m/z

1839.1) during the cleavage reaction, as observed by Papov and Biemann (42). Peaks at

m/z 2792.8 and 3581.5 correspond to the incomplete cleavage of fragments 18-43 and

44-76 in which 1.3-elimination occurs at Cys39 and Cys47, respectively.

 

 

Intensity
\
320.0

Relative:
\

 

 

\1 019 Q

///18ZOJ

 

rooo 2r

Figure 2.8, MALDI mass S
5001111101 of NTCB
Asterisks ( i
the identiti

a . Apprt
)lndicates peak
38 of marked pea

80

 

2332.0
2347.7

1

\320.0

 

Relative Intensity

 

,//1820J
\1838 8

E *2792.8
3337.6
; *3581.5

 

 

 

J “ - LL

1000 2000 3000 4000 5000 6000
m/z

 

 

 

Figure 2.8. MALDI mass spectrum of 100 pmol of spinach ferredoxin after reaction with
500 pmol of NTCB. Approximately 1.5 pmol of analyte was applied to the probe tip.
Asterisks (*) indicates peaks corresponding to B-elimination products. See Table 2.2 for
the identities of marked peaks and the correlation of calculated and.observed m/z values.

 

 

2. Papain

Papain, a member Of a
study because it has one 0““
15). The NTCB-(TCEP 1310'
sulfhydryl (Figure 2.9A). 86‘
0525 should give two PCP‘i‘
bond (CysZZ-Cys66). The rec
and produces peptide fiagmen
m’z 2621.5 and 20848.0. res
was observed corresponding
conesponding t0 peptide ch
probably due to suppression 1

Another interesting 01
reaction mixture from papair
frtrgment 1-24 at m/z 2622
iminothiazolidinyl residue a1
are Still linked by an inter-c1
at C1825. Thus, observatic
diSUIfide bond might have be
011 B-lactoglobulin A (MW 1
at Cylel, B-lactoglobulin

iI - ° '
ner charm disulfide bond (1

 

2. Papain

Papain, a member of a family of over 40 thiol proteases (43), was chosen for this
study because it has one cysteine (Cys25) at the active site and three disulfide bonds (44,
45). The NTCB/TCEP procedure was utilized to recognize the locus of the free
sulfhydryl (Figure 2.9A). Because papain is a single chain polypeptide (44), cleavage at
Cys25 should give two peptide chains which remain linked by an inter-chain disulfide
bond (Cys22-Cys66). The reduction of disulfide bonds by TCEP cleaves the two chains
and produces peptide fragments 1-24 and 25-212 having calculated values for [M + H]+ at
m/z 2621.5 and 20848.0, respectively. Experimentally, an intense peak at m/z 2622.1
was observed corresponding to the peptide chain 1-24. The expected high-mass peak-
corresponding to peptide chain 25-212 was not seen in the MALDI mass spectrum,
probably due to suppression by the low-mass peptide fragment present in the sample.

Another interesting observation is that the MALDI analysis of the NTCB cleavage
reaction mixture from papain prior to disulfide bond reduction also showed the peptide
fragment 1-24 at m/z 2622.1. In principle, only an intact protein derivatized by an
iminothiazolidinyl residue at Cys25 should be detected because the two peptide chains
are still linked by an inter-chain disulfide bond (Cy522-Cys66) after the NTCB cleavage
at Cys25. Thus, observation of the peak at m/z 2622.1 indicates that the inter-chain
disulfide bond might have been cleaved during the MALDI process. Similar experiments
on B-lactoglobulin A (MW 18368) (46) gave a similar result. After cleavage with NTCB
at Cys121, B-lactoglobulin A should produce two peptide chains that are linked by an

inter-chain disulfide bond (Cys66-Cysl60). However, the MALDI mass spectrum of B-

 

 

 

 

 

 

l-24

 

l<clati\_/_L_: Intensitv

 

 

 

 

 

Relative Intensity

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

82
1-24
A
.E
.5
0)
.>
E.
0)
ad
W 25212
5000 10000 15000 20000 25000
m/z '
B
121
.9
a
8
.8
_°>’ 63-94
:3.
32
3 9‘.
M H
N w).
0\
1000 2000 3000 4000 5000 6000 7000 8000
m/z

Figure 2.9. MALDI mass spectra of papain after (A) NTCB/TCEP treatment and (B)
TCEP/NTCB treatment. See Table 2.2 for the identities of other marked peaks and ,the
correlation of calculated and observed m/z values.

 

 

 

 

 

lactoglobulin A after reaction
resulting from the cleavage 0f
the recent report by P3116150"
undergo "prompt fragmentatio
111).

Total cysteine assessm
procedure. Most of the expect
seven cysteine residues were
peaks unable to be assigned tc
inthe commercial sample. T:
their observed masses. In th
Elimination were found, whic

POlypeptide chain at all the c

3. Ovalbumin

Ovalbumin (48‘50)
studied by the proposed PTO

N
TCB/TCEP procedure (Fi

gave two more peaks (at m/

an
dcleavage of the disulfid

a .

greement Wrth the previou

Furrh
ermore, two clusters

83

lactoglobulin A after reaction with NTCB showed peaks corresponding to fragments
resulting from the cleavage of the inter-chain disulfide bond. Our observations support
the recent report by Patterson et al that the inter-chain disulfide bond in peptides can
undergo "prompt fragmentation" or “in-source fragmentation” in the MALDI ion source
(47).

Total cysteine assessment of papain was performed according to the TCEP/NTCB
procedure. Most of the expected fragments corresponding to the cleavage at each of the
seven cysteine residues were found in the MALDI mass spectrum (Figure 2.9B). The
peaks unable to be assigned to any of the fragments were most likely due to an impurity
in the commercial sample. Table 2.2 also lists the calculated masses of the fragments and
their observed masses. In this case, only a few fragments of incomplete cleavage or B-
elimination were found, which indicates that the NTCB reagent can effectively cleave the

polypeptide chain at all the cysteine residues in papain.

3. Ovalbumin

Ovalbumin (4850), containing four free cysteines and one disulfide bond, was
studied by the pr0posed procedures. Compared to the mass spectrum obtained from the
NTCB/TCEP procedure (Figure 2.10A), that obtained with the TCEP/NTCB treatment
gave two more peaks (at m/z 4761.4 and 5426.1) that were due to reduction, cyanylation,
and cleavage of the disulfide bond Cys73-Cy5120 (Figure 2.108). This conclusion is in
agreement with the previous assignment of the disulfide bond linkage in ovalbumin (50).

Furthermore, two clusters of peaks were observed in the MALDI mass spectra which

 

 

 

 

’_____.____———————

367-381 \

367-385*

/’367'385*

_85

 

 

 

367-381 N

 

3675485"

 

 

 

 

 

 

 

 

1000 2000

Figure 2.10. MALDI mas
TUSP/NTCB treatment. '
Peptrdes and/or their B-eli

and the correlation 1

 

84

 

 

 

 

 

 

 

 

 

 

 

 

 

367-38l\ *
.3 A
(i
\O
m
* O\
In 01
oo .1.
<13 ...
[\
g /
as \ *
I 9(-
83 33
UM
flea-.4344 4 -4 -.._ T.,...”
367-381V * .,
2’8 B
m
if.
‘9 cost
32
* (\l
\/ {it
1000' 2000' 3000' 4000 5000 6000
m/z

Figure 2.10. MALDI mass spectra of ovalbumin after (A) NTCB/TCEP and (B)
TCEP/NTCB treatment. * indicates peaks corresponding to cyanylated/uncleaved
peptides and/or their B-elimination products. See Table 2.2 for the identities of marked
Peaks and the correlation of calculated and observed m/z values.

 

 

 

 

 

correspond to the incompltite
peak at m/z 3371.5 is due 10
Cysll (expected Ml 33718)
at Cysll (expected m/z 3312.
peptide and its Belimination '
and 2067.7 are likely due

uncleaved Cys382 and its

ovalbumin contains two ph
observed [M+H]' ion of the

phosphate group is still att
indicating that phosphoryla
exIltrirnental/ reaction condi
”I1 be used to detect the 1
segment, because it is prime
Shift. This example demon:

be used to confirm a protein

4. B-Lactoglobulin A

B-Lactoglobulin A

(46). After NTCB/TCEP t1

12
1 are expected, with a 1

E I
Xpemnentally, the fragmt

 

85

correspond to the incomplete cleavage and/or B-elimination of sulfliydryl groups. The
peak at m/z 3371.5 is due to the peptide chain 1-29 containing a cyanylated/uncleaved
Cysll (expected m/z 3371.8), while m/z 3311.1 corresponds to the B-elimination product
at Cysll (expected m/z 3312.8). The mass difference between the cyanylated/uncleaved
peptide and its B-elimination product is 59 Da. For the same reason, peaks at m/z 2126.6
and 2067.7 are likely due to the peptide chain 367-3 85, containing a cyanylated]
uncleaved Cys3 82 and its [3-elirnination product, respectively. It is known that
ovalbumin contains two phosphorylated residues at serines 68 and 344 (50). The
observed [M+H]+ ion of the fragment 30-72 (expected at m/z 4761.5 Da) shows that the
phosphate group is still attached to Ser68 after cyanylation and cleavage reactions,
indicating that phosphorylated proteins should be amenable to analysis under these
experimental/ reaction conditions. On the other hand, the mass mapping of the fragments
can be used to detect the possible posttranslational modification in a specific protein
segment, because it is principally possible to recognize the modification from the mass
shift. This example demonstrated that total cysteine determination by this approach can

be used to confirm a protein’s structure.

4. B-Lactoglobulin A

B-Lactoglobulin A contains five cysteines, of which Cys12l is a free cysteine
(46). After NTCB/TCEP treatment, two fragments corresponding to the cleavage at Cys
121 are expected, with a m/z of 13506.7 (1-120) and 4905.7 (121-162), respectively.

Experimentally, the fragment 1-120 presents as a small peak (Figure 2-11A), probably

 

due to the suppreSSion from c
two expeaed fragments. 1'12
the other two fragments 121‘]
cyanylation and cleavage at C
160 is insufficiently stable if
thiolate ions that subsequen
Given the fact that B-lactoglr
observed nonspecific cleav:
exchange and subsequent c
mixture against 20% acetic
molecular weight salts and 0
detected by MALDI (Fig. 2-
The peak at m/z 18390 is 111
containing fragment 1-120
dOUbly Charged ion. The IV
Shows fragments 1-65, 66-1
a[111160-162, which fall it

it
agments are essentially

MALDI responses of indi‘

S. .
lrnrlar MALDI spectra of

the disulfide bonds and sul:

86

due to the suppression from other components (peptides and salts). In additional to the
two expected fragments, 1-120 and 121-162, MALDI can also identify trace amounts of
the other two fragments 121-159 (m/z 4551.2) and 1-65 (m/z 7247.0), attributable to both
cyanylation and cleavage at Cys66 and Cysl60, implying that the disulfide bond pair 66-
160 is insufficiently stable in the presence of NTCB "reagent and is hydrolyzed to form
thiolate ions that subsequently participate in the cyanylation and cleavage reactions.
Given the fact that B-lactoglobulin A is particularly liable to SH/S-S exchange (51), the
observed nonspecific cleavage at Cys66 and Cys160 could also result from such
exchange and subsequent cyanylation and cleavage. After dialysis of the cleavage
mixture against 20% acetic acid for 12 hours (MW cut-off: 3000 Da) to remove low
molecular weight salts and other contaminants, the fragment 1-120 can be unambiguously
detected by MALDI (Fig. 2-11B), confirming that Cys121 is a cyanylation/cleavage site.
The peak at m/z 18390 is likely due to the reoxidization during the dialysis of sulfliydryl-
containing fragment 1-120 and 121-162, while the peak at m/z 9191 is obviously its
doubly charged ion. The MALDI spectrum after TCEP/NTCB treatment (Figure 2-11C)
shows fragments 1-65, 66-105, 106-118, 121-159, and two smallest fragments, 119-120
and 160-162, which fall in the region reserved for matrix peaks. Therefore, all the
fragments are essentially detected, despite of the greatly variable differences in the
MALDI responses of individual fragments. Denslow and Nguyen (3 8) have reported
similar MALDI spectra of cleavage products of B-lactoglobulin obtained by reduction of

the disulfide bonds and subsequent cyanylation and cleavage of the sulfhydryl groups.

 

 

 

 

    

2000 4000

6000

Figure 2.11. MALDI n
NTCB/Dialysis/TCEP,

correlation of calculatec

87

 

 

 

 

 

121-162
A
a
;i
.._1
xi 1-65 ++ 1-120 1-162
2.. ._- ‘TJA‘“““T
1-162
L120
121-162 1-65 -++
121'15\9 66-105
/
C

 

 

 

 

 

2000 4000 6000 8000 10000 12000 14000 16000 18000 20000
m/z

Figure 2.11. MALDI mass spectra of b-lactoglobulin after (A) NTCB/TCEP, (B)
NTCB/Dialysis/TCEP, and (C) TCEP/NTCB treatment. See Table 2.2 for the
correlation of calculated and observed m/z values.

4. Hemoglobin

Human hemoglobin A
one free cysteine at 104, whi
112. Since hemoglobin A car
commercially available. The
mixture of 01- and B-hemoglc
the expected fragments can ‘
observation is that each of tl
chain, is accompanied by a p
The identity of these peaks

Optimization of the deterrnin

0104-141

 

 

1393—1 11
-146

Relative Intensity
81 1 2

fl:

 

 

 

4. Hemoglobin

Human hemoglobin A contains 0t- and [3- chains (52). oe-Hemoglobin A contains
one fi'ee cysteine at 104, while B-hemoglobin A contains two free cysteines at 93 and

112. Since hemoglobin A can be easily oxidized in the air. The native hemoglobin is not

 

commercially available. The hemoglobin we acquired was a service sample containing a

mixture of or- and B-hemoglobin A variants. After NTCB cyanylation and cleavage, all

 

the expected fiagments can be unambiguously identified (Figure 2.12). An interesting
observation is that each of the fragments, 1-92 and 93-111 of B chain, and 1-103 of A

chain, is accompanied by a peak which shows a 102 Da mass increase fiom that fragment. '

 

The identity of these peaks are unclear. Because of the limited sample available, the

optimization of the determination was not pursued further.

 

 

 

 

insulin
oelO4-14l

.E

.9 .4 .0

£1 :3 E

a) «t, 0',

'81 a .2;

T)

94 ' 1-92

 

 

 

 

2000 4000 ‘ 6000 8000 10000 12000 14000 16000
m/z

Figure 2.12. MALDI mass spectrum of oe- and B-hemoglobin mixtures after reaction
with NTCB. See table 2.2 for the correlation of calculated and observed m/z values.

 

 

E. Choice of Reducing Rea:
Several reducing rea
reduction of protein disu
dithiothreitol (DTT) (54)~ ar
However, thiol-containing re
be avoided because excess
interferes with analyses
tarboxyethyl)ph0sphine (TI
disulfide bonds (36, 55, 56)
used in the presence of an 0:
isstable not only in acidic
reducing activity than or
reaction solution is therefon
Sufficient for complete red

larger amounts of TCEP d
MALDI—MS.

F. o 0
Stability of Disulfide ll

Stoichiometric expe

for complete cyanylation (

Performed under the same

 

E. Choice of Reducing Reagent

Several reducing reagents have been deve10ped for the efficient and selective
reduction of protein disulfide bonds. Among these, B-mercaptoethanol (53),
dithiothreitol (DTT) (54), and sodium borohydride (NaBH4) (55) are traditionally used.
However, thiol-containing reducing reagents such as B-mercaptoethanol and DTT must
be avoided because excess reagents will react with NTCB. Sodium borohydride
interferes with analyses by MALDI-MS (55). Tributylphosphine and tris(2-
carboxyethyl)phosphine (TCEP) have proven to be selective and reactive towards
disulfide bonds (36, 55, 56). However, tributylphosphine is water-insoluble and must be
used in the presence of an organic solvent. TCEP is a water-soluble reducing reagent and
is stable not only in acidic solution, but in basic solution as well. TCEP shows higher ‘
reducing activity than DTT over a wide pH range (57). The pH adjustment of the
reaction solution is therefore unnecessary. Usually, a 5-fold molar excess of TCEP is
sufficient for complete reduction of disulfide bonds in denaturing conditions although
larger amounts of TCEP do not appear to have any adverse effect on the analysis by

MALDI-MS.

F. Stability of Disulfide Bonds during the NTCB Reaction
Stoichiometric experiments indicated that the NTCB/thiol ratio of 5:1 is sufficient
for complete cyanylation of thiol groups. Because both cyanylation and cleavage were

Performed under the same condition without removal of the NTCB reagent, the excess of

 

 

 

 

 

 

 

 

 

NTCB reagent can always

along period of incubation,
The mechanism of th
acting as a nucleophile to a
cyanylated product, which 0
amino peptide bond (7). I
groups. Selective cyanylati
presence of cystines beca
specificity is not without li
containing peptides indica
incubation time result in cle
the stability study of disul:
excess of NTCB reagent u
disulfide bond. The same 11
excess of NTCB reagent i:
hours. MALDI spectra, as
generally stable in the pre:
incubation at 37°C, cor
cyanylation/cleavage of the
NTCB exceeds SO-fold und
correspond to the cleavage

disulfide bonds undergo l

 

 

90

NTCB reagent can always have a chance to attack the unreacted sulfhydryl groups during

a long period of incubation, which ensures the cyanylation goes completion.

The mechanism of the NTCB reaction involves the dissociated thiolate anion (-S')

acting as a nucleophile to attack the cyanate moiety under alkaline conditions to form a
cyanylated product, which can then undergo cleavage catalyzed by hydroxide ions at the
amino peptide bond (7). In principle, the NTCB reagent is specific for free sulfhydryl
groups. Selective cyanylation/cleavage at free sulfhydryl groups can be performed in the
presence of cystines because the disulfide bonds do not react with NTCB. However, this
specificity is not without limits under the reaction conditions. Experiments on disulfide-
containing peptides indicate that a large excesses of NTCB reagent and excessive
incubation time result in cleavage at disulfide bond residues. This has been confirmed by
the stability study of disulfide-containing peptide(s) in the presence of different molar
excess of NTCB reagent under the same buffer conditions. Somatostatin contains one
disulfide bond. The same molar amount of somatostatin was mixed with 2, 5, 10, 50-fold
excess of NTCB reagent in pH 9.0 buffer, respectively, and incubated at 37°C for 16
hours. MALDI spectra, as shown in Figure 2.13, show that while the disulfide bond is
generally stable in the presence of a 10-fold molar excess of NTCB after 16 hours of
incubation at 37°C, complete splitting of the disulfide bond and subsequent
Cyanylation/cleavage of the nascent cysteine residues take place when the molar excess of
NTCB exceeds 50-fold under the same conditions. The masses of the resulting fragments
correspond to the cleavage at one or both cysteine sites. One possible explanation is that

disulfide bonds undergo hydrolysis under mildly alkaline conditions (58, 59). The

 

 

 

 

inta

4km

degradation pr

W.,

fi—fi—‘fl"
1200 1300 1400

 

Figure 2.13. Stability stt
of NTCB reagent. After
at 37°C for 16 hours, 801

 

91

 

intact somatostatin
\ 2 fold

 

 

WW

5 fold

 

 

10 fold

/l

degradation products \ ‘ 50 fold

 

 

N

1200 1300 1400 1500 1600 1700 1800 1900 2000
m/z

 

 

 

 

 

Figure 2.13. Stability study of somatostatin in the presence of different molar excesses
of NTCB reagent. After exposed to 50-fold excess of NTCB reagent in pH 9.0 buffer
at 37°C for 16 hours, somatostatin decomposes completely.

 

 

 

 

resulting thiolate anions atta

cleave the peptide chain
NTCBII‘CEP procedure is
NTCB must be applied to e
purified by chromatography
subsequent cleavage reacti
total cysteines (T CEP/NTC
prior to the NTCB reaction.
Another commonly
bond exchange that usually
9.0, 161101118 incubation) di
not affect the assignment
srtlfltydryls become blocke
hand, the described methot

bond pairing patterns.

G. Analysis by MALDI-ll

Because of the wid
obtained fi'om NTCB cleaV
are common. Previous 1

employed sodium dodecl

Unfortunately, its accuracl

 

 

92

resulting thiolate anions attack the NTCB reagent to form cyanylated products that further
cleave the peptide chain at the original disulfide bond sites. Therefore, if the
NTCB/TCEP procedure is used for sulfhydryl—rich proteins where a high molar excess of
NTCB must be applied to ensure complete cyanylation, the cyanylated proteins are better
purified by chromatography or dialysis to remove the excess NTCB reagent prior to the
subsequent cleavage reaction. This limitation is not a problem for the determination of
total cysteines (TCEP/NTCB procedure) since all the disulfide bonds must be reduced
prior to the NTCB reaction.

Another commonly encountered problem in cysteine characterization is disulfide
bond exchange that usually occurs at pH > 8.0. Under our experimental conditions ( pH
9.0, 16 hours incubation) disulfide bond exchange cannot be avoided. However, it should
not affect the assignment of 'free sulflrydryl groups and/or total cysteines since free
sulfhydryls become blocked immediately after cyanylation with NTCB. On the other
hand, the described method cannot directly be employed for the assignment of disulfide

bond pairing patterns.

G. Analysis by MALDI-MS

Because of the widely variable occurrence of cysteine in proteins, the fragments
obtained from NTCB cleavage vary greatly in size. Fragments with a mass over 5000 Da
are common. Previous reports on the mass mapping of NTCB cleavage products
employed sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE).

Unfortunately, its accuracy of molecular weight determination is usually limited to :t 5-

 

 

 

10%. In many cases, pep

masses of two or more pep
Incomplete cleavage of prot
and the SDS-PAGE techni
products corresponding t
cyanylated/uncleaved pepti
MALDI-TOF MS
determine molecular weig
0.01% and thereby provid
comparison of calculated
assignment is better than i
for calibration. This accura
MALDI spectra art
though peaks for a sod
Occasionally, matrix addur
that the MALDI signal intt
because the reaction produ
peptides may not give a
discrimination effects are r

of salts, or due to differe

peptides are more suppress

 

 

93

10%. In many cases, peptide assignments are inconclusive or are ambiguous if the
masses of two or more peptides obtained from the NTCB digestion are similar (19, 21).
Incomplete cleavage of protein chains is often observed with the NTCB reaction (60-62),
and the SDS-PAGE technique has difficulty distinguishing among incomplete cleavage
products corresponding to the peptide with an unreacted sulfhydryl group, a
cyanylated/uncleaved peptide, or a B-elimination product.

MALDI-TOF MS has been a valuable complement to SDS-PAGE. It can
determine molecular weights as high as 300,000 Da with the mass accuracy of i01-
0.01% and thereby provide an unambiguous answer to questions mentioned above. A
comparison of calculated values and our observed data indicates that the error of mass
assignment is better than i 0.05% for most of the peptides if internal standards are used
for calibration. This accuracy is sufficient for the mass mapping of cleavage products.

MALDI Spectra are dominated by peaks for singly protonated species [M+H]+,
though peaks for a sodium adduct [M+Na]+ often appear at a lower intensity.
Occasionally, matrix adduct peaks are observed. A limitation of our present method is
that the MALDI signal intensities vary considerably from peptide to peptide in a mixture
because the reaction products are directly analyzed without purification by HPLC. Some
peptides may not give a detectable signal during analysis of a mixture. These
discrimination effects are either due to MALDI contaminants such as high concentrations
of salts, or due to differential responses of the peptides, or both. Usually high-mass

Peptides are more suppressed and yield weaker signals.

 

 

 

 

 

 

 

 

 

 

 

H. Reactions of Proteins 1

One of the great i
techniques, such as FAB a
However, the high concei
interferes with a successful
form a thick cake or a resi
Samples containing 00an
of the sample to minimize
analyte often can be obser
where only trace quantities

Recently, various tr
supports for direct analysis
and proteins on an inert tr
soluble MALDI contamina
while the analyte is retaine
matrix solution. Moreove
proteins are possible on the

The in situ disulfi‘
reaction for protein samph
Zetabind membrane and a

of which have been emplc

64). Experiments with ti

 

 

94

H. Reactions of Proteins Immobilized on A Zetabind Membrane

One of the great advantages of MALDI-MS over other mass spectrometric
techniques, such as FAB and E81 mass spectrometry, is its tolerance of contaminants.
However, the high concentration of denaturing agents used in our procedure still
interferes with a successful analysis by MALDI-MS. The protein-matrix rrrixtures often
form a thick cake or a residue that does not completely cocrystallize on the probe tip.
Samples containing contaminants frequently give no or very weak signals. Upon dilution
of the sample to minimize the effect of the contaminants, a MALDI signal from the
analyte often can be observed. Diluting the sample is, however, self-defeating in cases
where only trace quantities of the analyte are available.

Recently, various transfer membranes (3 7, 63, 64) have been studied as sample
supports for direct analysis of proteins by MALDI-MS. The immobilization of peptides
and proteins on an inert transfer membrane can be used to facilitate removal of water-
soluble MALDI contaminants, such as high concentrations of denaturing agents and salts,
while the analyte is retained when the membrane is immersed in water prior to adding the
matrix solution. Moreover, in situ chemical and enzymatic modifications of adsorbed
proteins are possible on the membrane.

The in situ disulfide reduction by TCEP, cyanylation by NTCB, and cleavage
reaction for protein samples as small as 10 pmol have been performed on a nylon-based
Zetabind membrane and a porous polyethylene membrane (purchased from Fisher), both
of which have been employed for direct MALDI-MS analysis of high-mass proteins (37,

64). Experiments with the NTCB reaction on the two membranes showed comparable

 

 

 

 

 

 

results. However, the pol

operationally difficult to ha

The MALDI mass
reactions with 10 pmol of }
membrane, are shown in P
NTCB reagents was appl
cyanylation of sulfhydryl:
membrane by washing, W]
the high concentration of
subsequent analysis by I
disulfide bond sites, in the
the excess NTCB reagent
performed in solution and
spectra for papain. The c
complete for ovalbumin
observed in the MALDI rr
Cysll. However, the res
poorer for the analytes imr
adequately controlled.

One problem with
maintaining an aqueous s<

This problem can be mini

containing a few drops c

 

 

95

results. However, the polyethylene membrane is a porous thin film on which it was
operationally difficult to handle the chemical reactions.

The MALDI mass spectra of the reaction products, obtained respectively from
reactions with 10 pmol of papain and 15 pmol of ovalbumin immobilized on a Zetabind
membrane, are shown in Figures 2.14 and 2.15. A SO—fold molar excess of TCEP and
NTCB reagents was applied to ensure complete reduction of disulfide bonds and
cyanylation of sulflrydryls. Because all reagents can readily be removed from the
membrane by washing, while the protein and/or peptide reaction products are retained,
the high concentration of TCEP and NTCB do not present any large problem during
subsequent analysis by MALDI-MS. Moreover, the possible cleavage reaction at '
disulfide bond sites, in the presence of a high molar excess of NTCB, was minimized as
the excess NTCB reagent was washed out after the cyanylation reaction. The reactions
performed in solution and on the surface of the Zetabind membrane gave ahnost the same
spectra for papain. The cleavage reaction performed on the membrane was even more
complete for ovalbumin relative to that performed in the solution; no peaks were
observed in the MALDI mass spectrum that corresponded to the incomplete cleavages at
Cysl 1. However, the resolution and accuracy of mass assignment can be substantially
poorer for the analytes immobilized on the membrane if the power level of the laser is not
adequately controlled.

One problem with performing microscale chemical reactions on a membrane is in
maintaining an aqueous solution of the reactants on the membrane for an extended period.
This problem can be minimized by performing reactions in a closed vessel (see Fig. 2-3)

containing a few drops of water to saturate the air space with water vapor so that the

 

 

 

1-24

Relative '

 

 

 

5000

Figure 2.14. MALDI n
treatment on a Zetabind

 

96

 

 

 

 

 

 

 

 

1-24
.42
U)
r:
8
.3
0
.>
:3.
d)
94
""1 -:..: J4-‘ramt‘vfiew‘ that: - :- :.-:-M>
5000 10000 15000 ’ 20000 25 000

m/z

Figure 2.14. MALDI mass Spectrum of 10 pmol of papain after in situ NTCB/TCEP
treatment on a Zetabind membrane. -

 

 

367-381

 

Relative Intensity

 

 

Figure 2.15. MALDI m
treatment on a Zetabind

 

 

97

 

 

 

 
 

 

 

 

367-381
\
9:
.12 7‘
{D m
G l\
8
5 m
g 1‘.
:2: e» 3
d) 00
m m
1006 2060 3060 4060 5060 6000

m/z

Figure 2.15. MALDI mass spectrum of 15 pmol of ovalbumin after in situ TCEP/NTCB
treatment on a Zetabind membrane.

 

 

droplet of aqueous reactior
reach completion.

The resolubilizatir
subsequent cocrystallizati
successful analysis by MA
an aqueous solution of tl
membrane and the analy
saturated with the vapor (
drying/cocrystallizing prc
improves the resolution in

After some of the
Nguyen (38) reported a s
after cyanylation on PVl
MALDI-MS. In their pr
ribonuclease, bovine B-la
membrane. After rinsing

by placing the membrane

 

98

droplet of aqueous reaction mixture will be sustained long enough to allow the reaction to
reach completion.

The resolubilization of reaction products from the membrane surface and
subsequent cocrystallization of the analyte/matrix mixture are the key to eventual
successful analysis by MALDI-MS. To assure the complete resolubilization of samples,
an aqueous solution of the matrix in 50% acetonitrile/0.1% TFA was applied to the
membrane and the analyte-matrix mixture was allowed to dry- slowly in a container
saturated with the vapor of the 50% acetonitrile/O.1% TFA aqueous solvent. This long
drying/cocrystallizing process also helps produce fine analyte-matrix cocrystals and
improves the resolution in the MALDI mass spectrum.

After some of the results presented in this chapter were published, Denslow and
Nguyen (3 8) reported a similar approach to cleave blotted proteins at cysteine residues
after cyanylation on PVDF membranes. The resulting fragments were analyzed by
MALDI-MS. In their protocol, the cyanylation of 50 pmol each of bovine pancreatic
ribonuclease, bovine B-lactoglobulin, and bovine insulin was carried out on, a PVDF
- membrane. After rinsing the membrane to remove buffers, cleavage was then carried out
by placing the membrane in alkaline buffer and the cleavage products were extracted into
MALDI solvent by sonication. Although most of the expected fragments were detected,
the efficiency of cleavage is somewhat less than that performed in free solution. Several
partial cleavage fragments were found, some of them correspond to B—elimination or

mixed disulfide bonds (38). These results support our conclusion that the cleavage

 

 

 

 

reaction can be achieved 0

that the MALDI analysis 0

V. Conclusions

We have demonstr
treatment of a protein an
simple and sensitive me
sulfliydryl residues in the
follouing TCEP and NTC
posttranslational modifica
reaction products. We als
remove excess reagents a
NTCB reagent is unique

protocol provides the adv:

high sensitivity.

 

 

 

99

reaction can be achieved on membranes. Moreover, our experiments further demonstrate

that the MALDI analysis of cleavage products can directly be performed on membranes.

V. Conclusions ‘

We have demonstrated here that the combination of NTCB and TCEP chemical
treatment of a protein and subsequent mass-mapping by MALDI-TOF MS provides a
simple and sensitive methodology for determining the number and location of free
sulfhydryl residues in the presence of disulfide bonds. On the other hand, mass-mapping
following TCEP and NTCB procedures can be used to confirm the primary sequence and
posttranslational modification of a protein and to identify the cleavage products and side
reaction products. We also demonstrated that an inert Zetabind membrane can be used to
remove excess reagents and other salts to improve the MALDI spectra. The use of the
NTCB reagent is unique in that it specifically targets the sites being analyzed. This
protocol provides the advantages of fast analysis, easy operation, high mass accuracy, and

high sensitivity.

 

 

 

Vl. References
l. Darbre, A. Practical }
149.
2. canoe. Q-aand W
3, Hirayama, K., and AI‘
(Matsuo, T., CaPFiOHa
Ltd, (1994). p. 299-3j
4. Robertson, J. G.. Ad
Villafranca J. J Bioc
5. Glocker,M. 0., Arbor
Acad. Sci. USA., 91, 5
. Catsirnpoolas,N., and
. Jacobson, G. R., Sch:
248,6583-6591 (1973
- Degani, Y., and Patch
. Wakselman, M., Gui
Chem. Commun., 1, 2
10. Brocklehurst, K., Ma.
A., Mushiri, M. S., an
11. Brocklehurst, K., Ma]

12' SChaffer, M. H. and
(1976).

 

 

 

100

VI. References

1.

10.

Darbre, A. Practical Protein Chemistry: A Handbook, John wiley & Sons, (1985) p.
149.

Chin, C. C. Q., and Wold, F., Anal. Biochem., 214, 128-134 (1993).

Hirayama, K., and Akashi, S., in Biological Mass Spectrometry: Present and Future
(Matsuo, T., Caprioli, R. M., Gross, M. L., and Seyama, Y, Eds), John Wiley & Sons
Ltd., (1994). p. 299-312.

Robertson, J. G., Adams, G. W., Medzihradszky, K. F ., Burlingame, A. L., and
Villafranca, J. J ., Biochemistry, 33, 11563-11575 (1994).

Glocker, M. 0., Arbogast, B., Milley, R., Cowgill, C., and Deinzer, M. L., Proc. Natl.
Acad. Sci. USA., 91, 5868-5872 (1994).

Catsimpoolas, N., and Wood, J. L., J. Biol. Chem, 241, 1790-1796 (1966).

Jacobson, G. R., Schaffer, M. H., Stark, G. R., and Vanaman, T. C., J. Biol. Chem.
248, 6583-6591 (1973).

Degani, Y., and Patchornik, A., Biochemistry, 13, 1-11 (1974).

Wakselman, M., Guibe-Jampel, E., Raoult, A., and Busse, W. D., J. Chem. Soc.
Chem. Commun., 1, 21(1976).

Brocklehurst, K., Malthouse, J. P. G., Baines, B. S., Blenkinsop, R. D., Churcher, J.

A., Mushiri, M. S., and Ormerod, F ., Biochem. Soc. Transactions, 6, 261 (1978).

11. Brocklehurst, K., Malthouse, J. P. 6., Biochem, J., 175, 761 (1978).

12. Schaffer, M. H., and Stark, G. R., Biochem. Biophys. Res. Comm, 71(4), 1040-1047

(1976).

 

 

13. Casey, R., and Lang. 4
14. Price N. C., Biochem.
15. Schaffer, M. H., and
(1976).
16111, H. S., and Gracy.
17. Thomas, M. L., Janat
USA, 79, 1054-1058 (
18. Roberts, D. D., and G
19. Matsudaira, P., lakes,
82, 6788-6792 (1985)
20. Mahboub, S., Richard
(1986).
21. Bahler, M., Benfenati
108,1841-1849(198‘.
22. Minard, P., Desmad
Biochem., 185, 419-4
23' NefSky, 3., and Brets
24.13shdat, Y., Chapot, l\
25' Katayama. E, J. Bioc

26. Sutherland, C., and V

27. Tang, 0, Yuksel, K.

283(1), 12.19 (1990).

101

13. Casey, R., and Lang, A., Biochem. J., 145, 251-261 (1975).

14. Price N. C., Biochem. J., 159, 177-180 (1976).

15. Schaffer, M. H., and Stark, G. R. Biochem. Biophys. Res. Comm. 71, 1040-1047
(1976) i

16. Lu, H. S., and Gracy, R. W., Arch. Biochem. Biophys., 212(2), 347-359 (1981).

17. Thomas, M. L., Janatova, J., Gray, W. R., and Tack, B. F ., Proc. Natl. Acad. Sci.
USA, 79, 1054-1058 (1982).

18. Roberts, D. D., and Goldstein, I. J ., J. Biol. Chem, 259, 909-914 (1984).

19. Matsudaira, P., Jakes, R., Cameron, L., and Atherton, B, Proc. Natl. Acad. Sci. USA.
82, 6788-6792 (1985).

20. Mahboub, S., Richard, C., Delacourte, A., and Han, K., Anal. Biochem., 154, 171-182
(1986)

21. Bahler, M., Benfenati, F., Valtorta, F., Czemik, A. J ., and Greengard, P., J. Cell Biol,
108, 1841-1849 (1989).

22. Minard, P., Desmadril, M., Ballery, N., Perahia, D., and Mouawad, L., Eur. J.
Biochem., 185, 419-423 (1989).

23. Nefsky, B., and Bretscher, A., Proc. Natl. Acad. Sci. USA, 86, 3549-3553 (1989).

24. Eshdat, Y., Chapot, M. P., and Strosberg, A. D., FEBS Letters, 1-2, 166-170 (1989).

25. Katayama, E., J. Biochem., 106, 988-993 (1989).

26. Sutherland, C., and Walsh, M. P., J. Biol. Chem, 264, 578-583(1989).

27. Tang, C., Yuksel, K. U., Jacobson, T. M., and Gracy, R. W., Arch. Biochem. Phys,

283(1), 12-19 (1990).

 

 

 

 

28. May, J. M., Buchs, A.

29. Wines, B. D., and Ear
(1991)

30. 80111, S., and Friedman

31.Altarnirano, M. M., P
(1992).

32. Goold, R., and Baines

33. Stark, G. R. in Metht
(1977), Academic Pre

34. Papayannopoulos, 1.,A

35. Wu, J., Gage, D. A.,;

36.er3, J. A., Butler,
(1991).

37.2e11uzec, E. J., Gag.
Spectrom., 5, 230-23‘

38. Denslow, N. D., and
11241.

39. Matsubara, H” and S

40. Matsubara, H., and S

 

102

28. May, J. M., Buchs, A. and Carter-Su, C., Biochem., 29, 10393-10398 (1990).

29. Wines, B. D., and Easterbrook-Smith, S. B., Molecular Immunology, 28(8), 855-863
(1991)

30. Som, S., and Friedman, S., J. Biol. Chem, 266, 2937-2945 (1991).

31. Altamirano, M. M., Plumbridge, J. A., and Calcagno, M. L., Biochem., 31, 1153-1158
(1992)

32. Goold, R., and Baines, A. J ., Eur. .J. Biochem., 224, 229-240 (1994).

33. Stark, G. R. in Methods Enzymol. (Jakoby, W. B., and Wilchek, M., Eds), vol. 47,
(1977), Academic Press, San Diego, p. 129.

34. Papayannopoulos, LA, and Biemann, K., Protein Sci., 1, 278-288 (1992).

35. Wu, J ., Gage, D. A., and Watson, J. T., Anal. Biochem., 235, l61-174(1996).

36. Burns, J. A., Butler, J. c., and Whitesides, G. M., J. Org. Chem, 56, 2648-2650
(1991)

37. Zaluzec, E. J ., Gage, D. A., Allison, J ., and Watson, J. T., J. Am. Soc. Mass
Spectrom., 5, 230-237 (1994).

38. Denslow, N. D., and Nguyen, H. P., in Techniques in Protein Chemistry V11, (1996),
p.241.

39. Matsubara, H., and Sasaki, R.M., Proc. Natl. Acad. Sci. USA., 57, 439-445 (1967).

40. Matsubara, H., and Sasaki, R.M., J. Biol. Chem. 243, 1732-1757 (1968).

41. Chakrabarti, P., Biochem., 28, 6081-6085(1989).

42. Papov, V. V., and Biemann, K. Proceedings of the 42nd ASMS Conference on Mass

Spectrometry and Allied Topics, Chicago, IL, May29-June3, (1994), pp.688.

 

 

 

 

43.Thibault, P., PleasanC<
Mass Spectrom. Ion PI
44. Amon, R., in Methods
Press, San Diego: "01'
45. Cohen, L. W., Coghlat
46. Mckenzie, H. A. R3
(1972).
41. Patterson, S., and Kati
48. Tompson, E.0.P., and
49.1akatera, K., and Wat
50. Nisbet, A.D., Saundry
J. Biochem., 115, 335
51. Phelan, P., and Malthu
52. Voet, D., and Voet, J.
53. Kawamura, M., and 1
54'K09183berg, K., in M
Academic Press, San
55.Za1uzec, E. J., Gage,
(1994).
56- FiSCher, W. H., Rivit

225-228 (1993).

 

103

43. Thibault, P., Pleasance, S., Laycock, M.V., MacKay, R.M., and Boyd, R.K., Int. J.
Mass Spectrom. Ion Proc., 111, 317-353 (1991).

44. Arnon, R., in Methods Enzymol. (Perlmann, G. E., and Lorand, L., Eds), Academic
Press, San Diego, vol. 19, (1970), pp. 226-244.

45. Cohen, L. W., Coghlan, V. W., and Dihel, L. C., Gene, 48, 219-227(1986).

46. Mckenzie, H. A., Ralston, G. B., and Shaw, D. c., Biochem., 11(24), 4539-4547
(1972).

47. Patterson, S., and Katta, V., Anal. Chem, 66, 3727-3732(1994).

48. Tompson, E.O.P., and Fisher, W.K., Aust. J. Biol. Sci., 31, 433-442 (1978).

49. Takatera, K., and Watanabe, T., Anal. Chem, 65, 3644-3646(1993). '

50. Nisbet, A.D., Saundry, R.H., Moir, A.J.G., Fothergill, LA, and Fothergill, J .E., Eur.
J. Biochem., 115, 335-345 (1981).

51. Phelan, P., and Malthouse, J. P. G., Biochem. J., 302, 511-516 (1994).

52. Voet, D., and Voet, J. G., Biochemistry, John Wilet & Sons, (1990), p. 218-219.

53. Kawamura, M., and Nagano, K., Biochim. Biophys. Acta, 774, 188 ( 1984).

54. Konigsberg, K., in Methods in Enzymol., (Hirs, C. H. W., and Tirnasheff, S. N., Eds),
Academic Press, San Diego,vol. 25, pp. 185.

55. Zaluzec, E. J., Gage, D. A., Watson, J. T., J. Am. Soc. Mass Spectrom. 5, 359-366
(1994)

56. Fischer, W. H., Rivier, J .E., and Craig, A.G., Rapid Commun. in Mass Spectrom. 7,

225-228 (1993).

57. Han, J. C., and Han, G. Y., Anal. Biochem., 220, 5-10 (1994).

 

 

 

 

58. Allen, G., in Lahore
(Bordon, R. 11., and Kn

59.Torchinsky, Y. M., in
Oxford, (1981).

60. May, J. M., Buchs. A..

61.Tang, C. Y., Yuksel,
Biophys., 283, 1.2-19 (l

62. Goold R, and Baines,

63.Vestling, MM, and Fe

64. Blacldedge, J .A., and A

 

104

58. Allen, G., in Laboratory Techniques in Biochemistry and Molecular Biology,
(Burdon, R. H., and Knippenberg, P. H., Eds), Elsevier, Vol.9, (1989), pp. 318.

59. Torchinsky, Y. M., in Sulfur in Protein; (Metzler, D., Translation Ed.), Pergamon,
Oxford, (1981).

60. May, J. M., Buchs, A., and Carter-Su, C., Biochemistry, 29, 10393-10398 (1990).

61. Tang, C. Y., Yuksel, U., Jacobson, T. M., and Gracy, R. W., Arch. Biochem.
Biophys., 283, 12-19 (1990).

62. Goold, R., and Baines, A. J ., Eur. J. Biochem., 224, 229-240 (1994).

63. Vestling, M.M., and Fenselau, 0., Anal. Chem, 66, 471—477 (1994).

64. Blackledge, J .A., and Alexander, A.J., Anal. Chem, 67, 843-848 (1995).

 

 

FURTHER STUD
CYSTEINE RESIDU

MS: 01".

I. Introduction

It has been known 1
the N-terminus of cysteine
amino-tenninal peptide ant
(1,2). The cleavage reactit
37°C for a prolonged pf
Polypeptide chain at the cy
bonds are turreactive to thi
frequency of the occurrenc
insize and can usually be

electrophoresis (SDS-PAG

CYanylated cysteine can a

thiocyanate and dehydroal

mildly alkaline conditions,

a
moderate degree of cleav

near
ly to completion for a

P .
atchonnk (9) found that d

 

CHAPTER 3
FURTHER STUDY ON THE LOCALIZATION OF PROTEIN
CYSTEINE RESIDUES BY CHEMICAL CLEAVAGE AND MALDI-

MS: OPTIMIZATION AND APPLICATION

I. Introduction

It has been known for some time that peptide chains can selectively be cleaved at
the N-terminus of cysteine residues after cyanylation of sulflaydryl groups to form an
amino-terminal peptide and a series of 2-iminothiazolidine-4-carboxylyl (ITC) peptides
(1, 2). The cleavage reaction can be carried out in mildly alkaline conditions (pH8-10) at
37°C for a prolonged period of incubation (12-80h). Selective cleavage of the
polypeptide chain at the cyanylated cysteine residues may be achieved because disulfide
bonds are unreactive to the cyanylation and cleavage. Because of the widely variable
frequency of the occurrence of cysteine in proteins, the resulting fragments vary greatly
in size and can usually be mass-mapped by sodium dodecyl sulfate-polyacrylamide gel
electrophoresis (SDS-PAGE) (3-7) or more recently by mass spectrometry (2, 8).
Cyanylated cysteine can also undergo a base-catalyzed B-elimination reaction to form
thiocyanate and dehydroalanine. Since both cleavage and [J-elimination occur under
mildly alkaline conditions, the two reactions are competitive, sometimes resulting in only
a moderate degree of cleavage of a peptide bond. Although the cleavage seemed to go
nearly to completion for a variety of proteins in Jacobson’s experiments, Degani and

Patchornik (9) found that dehydroalanyl residues were formed in variable amounts from

105

 

 

the [l-elirnination of cyany
pelimination over cleav:
peptides. Higher pH was
Nakagawa er a1 (10) rCC€
hormone, hPTH [1-84]. if
biologically active or-amic'
to produce with an 80% j
recombinant techniques.

been extensively used wit

In chapter 2, we d
inproteins using this spe
TOF MS. This methodo
SimPle, fast, and sensitive
Proteins. One of the inhe
reaction takes too long to
exPosure to alkaline 3011
results.

In this chapter, v
function 0f PH, solvents.
amino acids adjacent to
show that a higher pH .

ne .
vertheless, 1n contras'

 

106

the B-elimination of cyanylated cysteine residues at pH 8-10. Increasing pH favored the
B-elimination over cleavage, depending on the respective structural properties of
peptides. Higher pH was therefore avoided in order to suppress the B-elimination. But
Nakagawa et al (10) recently reported the cleavage of cyanylated human parathyroid
hormone, hPTH [1-84], in a high concentration of ammonia in an attempt to produce a
biologically active oc-amidated peptide. By the specific cleavage reaction they were able

to produce with an 80% yield the or-amidated peptide which could not be produced by

recombinant techniques. In all, the protocol (1, 9) developed by Jacobson et al (1) has

been extensively used with little modification over the past two decades.

In chapter 2, we described a new methodology to localize free sulfhydryl groups
in proteins using this specific cleavage reaction followed by mass-mapping by MALDI-
. TOF MS. This methodology is advantageous over conventional approaches in that it is
simple, fast, and sensitive, and is especially useful for locating sulfhydryl groups in large
proteins. One of the inherent problems with this procedure, however, is that the cleavage
reaction takes too long to accomplish. Various side reactions may rise after long hours of
exposure to alkaline solution, which brings about some uncertainty into the analytical
results.

In this chapter, we present our systematic studies on the cleavage reaction as a
function of pH, solvents, and amino acid structures for polypeptides containing different
amino acids adjacent to N-terminus of cyanylated cysteines. Our experimental results
show that a higher pH can greatly promote both cleavage and B-elimination reactions,

nevertheless, in contrast to the previous reports, for most of the peptides studied, the

 

 

extent of B-elimination d
Ammonia, a stronger nucl
of peptide chains and min
most of the peptides stur
itithin an hour at room ter
effects of peptide structur
The optimized conditions

in a variety of proteins wi
H. Experimental Sectio

Mass Spectrometry
MALDI mass spe
Spectrometer (PerSeptivf
VSL-337ND nitrogen la
the ion source was set t1
Time-to-mass conversio
Standards of bradykinin
skeletal myoglobin (m/z
All experiments were

Chemical Co, Milwauk

In a 50% (v/v) solution

107

extent of B-elimination does not increase significantly even at a pH as high as 12.
Ammonia, a stronger nucleophile than hydroxyl anions, greatly accelerates the cleavage
of peptide chains and minimizes side reactions related to the prolonged incubation. For
most of the peptides studied, the cleavage in l M NH4OH solution can be complete
within an hour at room temperature. Based on the results from over a dozen peptides, the
effects of peptide structure on the rate and yield of the cleavage reaction were evaluated.
The optimized conditions have been applied to the recognition of free sulfhydryl groups

in a variety of proteins with both known and unknown cysteine structures.

II. Experimental Section

Mass Spectrometry

MALDI mass spectra were obtained on a Voyager Elite time-of-flight (TOF) mass
spectrometer (PerSeptive Biosystems Inc., Framingham, MA) equipped with a model
VSL-337ND nitrogen laser (Laser Science, Newton, MA). The accelerating voltage in
the ion source was set to 20 kV. Data were acquired in the linear mode of operation.
Time-to-mass conversion was achieved by external and/or internal calibration using
standards of bradykinin (m/z 1061.2), bovine pancreatic insulin (m/z 5734.5), and horse
skeletal myoglobin (m/z 16,952) obtained from Sigma Chemical Co. (St. Louis, MO).
All experiments were performed using oc-cyano-4-hydroxycinnamic acid (Aldrich
Chemical Co., Milwaukee, WI) as the matrix. Saturated matrix solutions were prepared

in a 50% (v/v) solution of acetonitrile/aqueous 0.1% TFA, and mixed in equal volumes

 

 

 

 

with peptide or protein 1

mixture was allowed to air

Chemicals

Guanidine hydroc
(Indianapolis, IN). l-C
peptides TCVEWLRRYI
ovalbumin, and rabbit 111
used without further puri
DRVYIHPCHLLYYS, .4
American Peptide Com;
California Inc. The 0.10
HCl was freshly prepare

prepared in 0.1 M citrate

Cyanylation of SH Gro

To the peptide 2
cyano-4-dilnethylamino-

sulfhydryl content. Cya

1
00m temperature for 1l

reversed-phase HPLC l

manually, the masses of

 

108

with peptide or protein samples, and applied to a stainless-steel sample plate. The

mixture was allowed to air dry before being introduced into the mass spectrometer.

Chemicals

Guanidine hydrochloride was a product of Boehringer-Mannheim Biochemicals
(Indianapolis, IN). 1-Cyano-4-dimethylamino-pyridinium tetrafluoroborate (CDAP),
peptides TCVEWLRRYLKN and RYVVLPRPVCFEKGMNYTVR, spinach ferredoxin,
ovalbumin, and rabbit muscle creatine phosphokinase were purchased from Sigma and
used without further purification. Acetonitrile and TFA were of HPLC grade. Peptides
DRVYIHPCHLLYYS, ALLETYCATPAKSE, and SLRRSSCFGGR were products of
American Peptide Company. The rest of the peptides were purchased from Bachem
California Inc. The 0.10 M CDAP solution in pH 3.0, 0.1 M citrate buffer-4M guanidine-
HCl was freshly prepared prior to use. The 1 mM protein and peptide solutions were

prepared in 0.1 M citrate buffer, pH 3.0, containing 4 M guanidine-HCI.

Cyanylation of SH Groups

To the peptide and protein solutions was added a 10-fold molar excess of 1-
cyano-4-dimethylamino-pyridinium tetrafluoroborate (CDAP) solution over the
sulfllydryl content. Cyanylation of sulfhydryl groups was accomplished by incubation at
room temperature for 10-15 min. The modified proteins or peptides were purified by
reversed-phase HPLC under gradient elution. The HPLC fractions were collected

manually, the masses of which were determined by MALDI-MS. Those corresponding to

 

 

 

 

the cyanylated proteins at

further use.

Cleavage of the Cyanl’la

The Cyanylated p6
4M Mme-HG solut:
equal volume of 4. M gut
0.25 M tris-HCl buffer, t
The pH 8.0 and pH 9.0 \
the cleavage reaction, w'
ambient temperature for
were taken and diluted tc
MS. The rest of the so
HPLC fractions were co
of the cleavage and B-eli
under the peaks identifle

The cleavage c
PhOSPhOkinase at a cone
tris-HCI buffer, pH 9.0,
NHJOH solution-4 M g
Aliquots of 1 pl of th

CH3CN/0.1%TFA for '

109

the cyanylated proteins and peptides were dried in a speed vac and kept in a freezer for

fiirther use.

Cleavage of the Cyanylated Peptides and Proteins

The cyanylated peptide samples were reconstructed to a concentration of 1 mM in
4 M guanidine-HCI solution. A portion of the above solutions (~5 pl) was mixed with an
equal volume ’of 4 M guanidine-HCI solution in pH 8.0, 0.25 M tris-HCl buffer, pH 9.0,
0.25 M tris-HCI buffer, 0.02 M NaOH (pH 12), and 2 M NH4OH solution, respectively.
The pH 8.0 and pH 9.0 vials were incubated in 37°C water bath for 18 hours to promote
the cleavage reaction, while the vials with NaOH and NH4OH solutions were placed at
ambient temperature for 1 hour. After the reaction, l-ul aliquots of the above solutions
were taken and diluted to 100 ill with 1:1 CH3CN/0. 1%TF A for analysis by MALDI-TOF
MS. The rest of the solutions was acidified and injected into an HPLC column. The
HPLC fractions were collected and identified by MALDI-TOF MS. The relative yields
of the cleavage and B-elimination products were calculated by integrating the HPLC areas
under the peaks identified by MALDI.

The cleavage of cyanylated proteins, ferredoxin, ovalbumin, and creatine
phosphokinase at a concentration of 0.5 mM, was carried out by incubation both in 0.1 M
tris-HCl buffer, pH 9.0, containing 4 M guanidine—HCI at 37°C for 18 hours and in 1 M
NH4OH solution-4 M guanidine-HCI at ambient temperature for one hour, respectively.
Aliquots of 1 pl of the above solutions were taken and diluted to 100 pl with 1:1

CH3CN/0.1%TFA for analysis by MALDI-TOF MS. The rest of the solution was

 

 

 

acidified and then injectet

identified by MALDI-TO

HPLC Separation

The purification
cleavage products were
using Waters model 6001
11m. The column used th
size, 300-A pore, 4.6x2.
solution and CH3CN co:
from 5-50% B in 50 r
conditions for individua

major HPLC fractions .

determined by MALDI-l

HI. Results and Discus

A‘ Cyanylation of SH

The cJ'anylation

tluocyanobenzoic acid (
was believed to be spe

fo '
tmatlon of mixed disr

 

 

110

acidified and then injected to the HPLC column. The HPLC fractions were collected and

identified by MALDI-TOF MS. '

HPLC Separation

The purification of the cyanylated proteins and peptides and the separation of
cleavage products were achieved by reversed-phase HPLC with linear gradient elution
using Waters model 6000 pumps controlled by a PC computer. UV detection was at 215
nm. The column used throughout the study was a Vydac C18 (#218TP54, 10 um particle
size, 300-A pore, 4.6x250 mm). Mobile phases A and B contain 0.1% TFA aqueous
solution and CH3CN containing 0.1% TFA, respectively. Typically, a gradient elution
from 5-50% B in 50 minutes was applied in preliminary experiments. The HPLC
conditions for individual proteins and peptides were then modified and optimized. The
major HPLC fractions were collected manually, and the masses of the fractions were

determined by MALDI-MS.

III. Results and Discussion

A. Cyanylation of SH Groups by CDAP

The cyanylation of sulfliydryl groups was traditionally accomplished by 2-nitro—5-
thiocyanobenzoic acid (NTCB) under mildly alkaline conditions (pH 8-10). The reagent
was believed to be specific to sulfliydryl groups, although side reactions, such as the

formation of mixed disulfide bonds between the NTCB and protein SH groups, were also

 

 

 

reported (9. 11, 12)‘ Sine
under alkaline conditions
be difficult even though t
1-cyano-4-dimethylamin‘
cyanylation of SH group
reagent has not been user

The CDAP is ad\
groups ill the presence c
minimize sulfhydryl/(iii
peptides are stable und
subsequent cleavage rear
showed that complete cy
molar excess of the rea
temperature. Our study
the published results. I
excess of CDAP (~50-f
POUIS) could result in t1
Product can be separ.
implicating the notable
M988 analysis of the 311
corresponding Cyanylatt
b

y CDAP needs to be

reac '
tron. After extensir

 

111

reported (9, 11, 12). Since both the cyanylation and the consequent cleavage are achieved
under alkaline conditions, the separate study of cyanylation and cleavage reactions might
be difficult even though the reaction conditions are carefully controlled. Another reagent,
1-cyano-4-dimethylamino-pyridinium tetrafluoroborate (CDAP), was proposed for the
cyanylation of SH groups under slightly acidic conditions (pH 3-7) (14). However, this
reagent has not been used as extensively. as the NTCB.

The CDAP is advantageous over the NTCB for specific cyanylation of protein SH
groups in the presence of disulfide bonds, because the acidic conditions can effectively
minimize sulfhydryl/disulfide exchange. Furthermore, the cyanylated proteins or
peptides are stable under acidic conditions, the kinetics of the cyanylation and the
subsequent cleavage reaction can be studied independently. Previous papers (10, 14, 15)
showed that complete cyanylation by the CDAP can be carried out using three to five fold
molar excess of the reagent over free SH groups under very mild conditions at room
temperature. Our study with sulfhydryl-containing peptides listed in Table 3.1 confirmed
the published results. Further experiments with the model peptides showed that a large
excess of CDAP (~50-fold over peptide SH groups) and excessive incubation time (> 2
hours) could result in the formation of an unidentified side reaction product. The side
product can be separated from the corresponding cyanylated peptide on HPLC,
implicating the notable differences between the hydrophobicity of the two products.
Mass analysis of the side product by MALDI shows a mass increase of 43 Da from the
corresponding cyanylated peptide. Therefore, like cyanylation by NTCB, the cyanylation
by CDAP needs to be performed under controlled conditions to minimize the side

reaction. After extensive examination with different peptides, we optimized the reaction

 

 

 

 

 

 

   
      
    

Table 3.1. Calculated
cleavage reaction

amino acid sequenc
PHCKRM
RGPCRAFI
SRNRCNDQ
SLRRSSCFGGR
MSRPACPNDKYE
YK'ITICGKGLSATV
EKPLQNFTLCFR
ALLETYCATPALSE
TCVEWLRRYLKN
DRVYIHPCHLLYYS
KRNPGSNKRFPSNCGI

RYVVLPRPVCFEKGM
MHRQBAVDCLKKFNJ

 

 

112

Table 3.1. Calculated m/z values of fragments for sulfliydryl-containing peptides after the

cleavage reaction

 

 

 

IM+HJ +
amino acid sequence M.W. , , ,
fragment I fragment II B-ellmlnatlon
PHCKRM 771.0 253.3 562.7 738.0
RGPCRAFI 919.3 329.3 636.0 887.3
SRNRCNDQ 991.2 532.6 503.6 958.2
SLRRSSCFGGR 1225.7 705.9 564.8 1192.7
MSRPACPNDKYE 1410.0 561.7 892.3 1377.0
YK’I‘TICGKGLSATV 1441.9 625 .8 861.1 1407.9
EKPLQNFTLCFR 1493.9 1088.4 450.5 1460.9
ALLETYCATPALSE 1496.6 709.7 831.9 1463 .6
TCVEWLRRYLKN 1580.9 120.1 1506.8 . 1547.9
DRVYIHPCHLLYYS 1779.4 900.1 924.3 1746.4
KRNPGSNKRFPSNCGRD 1947.3 1517.8 476.5 1914.3
RYVVLPRPVCFEKGMNYTVR 2427.9 1099.5 1373.5 2394.9
MHRQEAVDCLKKFNARRKLKGA 2600.4 986.1 1659.3 2567.4

 

 

 

 

 

conditions: the cyanylati

CDAP (over sulfhydryl
Under these conditions

reaction was observed.

B. Optimization of Cl
Although the sel
of cyanylated cysteine r
today are almost the
(1). Degani and Patcho
reactions and the effects
elimination are both bas
while B-elimination seer
conditions (pH 8-9) W
incubation time differs .
hours, depending on the
also affect the rate of Cl
cleavage, no appreciab
conclusion made by Dr

dipeptides and tripeptidt

model peptides represe

 

 

113

conditions: the cyanylation was carried out by treatment with a 10-fold molar excess of
CDAP (over sulfhydryl groups) in pH 3.0 buffer for 15 min at ambient temperature.
Under these conditions the cyanylation reaction went to completion and very little side

reaction was observed.

B. Optimization of Cleavage Reactions

Although the selective chemical cleavage of peptide chains at the N-peptide bonds
of cyanylated cysteine residues was proposed long ago, the experimental conditions used
today are almost the same as those described in the protocol developed by Jacobson et al
(1). Degani and Patchornik (9) studied the pH [dependence of cleavage and B-elimination
reactions and the effects of amino acid structures on both reactions. The cleavage and [3-
elimination are both base catalyzed and therefore competitive under alkaline conditions,
while B-elimination seems to be favored over cleavage at higher pH. Mildly alkaline
conditions (pH 8-9) were suggested for the cleavage reaction in early study. The
incubation time differs significantly under the above conditions, ranging from 16 to 80
hours, depending on the structures of proteins under study. Structural features of proteins
also affect the rate of cleavage reaction (16). Proteins have to be denatured prior to the
cleavage, no appreciable cleavage was observed for native proteins. However, the
conclusion made by Degani and Patchornik was based on the study on a few simple
dipeptides and tripeptides with special structural factors. The behavior of cysteines in the

model peptides represents to a less extent general structural features of cysteine residues

 

 

 

 

in peptides and proteins.

better understanding and
We have system
filnction of pH and reac
amino acids adjacent to
were first cyanylated an
cyanylated peptides, afi
12.0, and 1 M NILOH
The cleavage mixtures
fractions identified by
side reactions were eval
Results of the k
cleavage can be accom
within one hour at amt
37°C in pH 8-9 buffers.
the greatest concerns e:
favor B-elimination ow
However, our study on

that it is not the case. .

differ significantly fo:

conditions. Some of

 

 

 

114

in peptides and proteins. Further examination of the cleavage reaction is needed for the
better understanding and optimization of the cleavage reaction.

We have systematically studied the rate and yield of the cleavage reaction as a
function of pH and reaction medium for a variety of polypeptides containing different
amino acids adjacent to the cyanylated cysteines. The peptides under study (Table 3.1)
were first cyanylated and the completion of cyanylation was monitored by MALDI. The
cyanylated peptides, after purified by HPLC, were subjected to cleavage at pH 8.0, 9.0,
12.0, and 1 M NH4OH (pH~11) solutions containing 4 M guanidine-HCI, respectively.
The cleavage mixtures were separated by HPLC, the fractions collected, the masses of the
fractions identified by MALDI. The relative yields of the cleavage products and other
side reactions were evaluated by measuring HPLC peak areas.

Results of the kinetics study showed that, for most of the peptides studied, the
cleavage can be accomplished in both ammonia solution and 0.01 M NaOH solution
within one hour at ambient temperature, whereas it requires 18 hours of incubation at
37°C in pH 8-9 buffers. Although elevated pH can accelerate cleavage reactions, one of
the greatest concerns expressed in previous papers was that a higher pH would largely
favor B-elimination over cleavage, causing a lower yield of the cleavage reaction.
However, our study on the relative yields under various pH and reaction media indicated
that it is not the case. As shown in Table 3.2, the relative yields of B-elimination do not
differ significantly for most of the peptides studied under the four experimental

conditions. Some of the peptides show even higher cleavage yields in 1 M NH4OH or

 

 

 

 

       
 

Table 3.2. Relative

PHCKRMd
RGPCRAFI
SLRRSSCFGGR
MSRPACPNDKYE

DRVYIHPCHLLYYS

KRNPGSNKRFPSNC<
RYVVLPRPVCFEKGI
MHRQEAVDCLKKFI‘

‘ in each case,B-elirnin
" side reactions other t1
° peptide undergoes ap‘
d incomplete conversio

 

115

Table 3.2. Relative yields (%) of B-elimination under different cleavage conditions

 

pH 8.0, pH 9.0, pH 12 , 1M NH 40H,

amino acid sequence 37°C,18h 370C,18h rt, 1h rt, 1h

 

PHCKRM“ 9.9a 15.9 25.1 14.8
RGPCRAFI 31.6b 87.5 82.5 70.1
SLRRSSCFGGR 2.8 2.9 3.9 4.6
MSRPACPNDKYE 6.1 5.8 1 1.3 3.9
YK'IVI‘ICGKGLSATV 16.8 18.2 22.2 20.3
EKPLQNFTLCFR 15.2c 29.9 49.8 20.9
ALLETYCATPALSE ~100 ~100 ~100 ~100
TCVEWLRRYLKN 20.9 13.8 13.1 18.3
DRVYIHPCHLLYYS 56.2 90.7 98.5 93.8
KRNPGSNKRFPSNCGRD 5.0 5.0 5.0 5.0

RYVVLPRPVCFEKGMNYTVR 34.3 41.2 52.4 51.5
MHRQEAVDCLKKFNARRKLKGA 12.8 1 1.1 18.4 13.2

 

3‘ in each case,B-elimination is a main side reaction

b side reactions other than B-elirnination and dimerization are observed
° peptide undergoes appreciable dimerization

d incomplete conversion

 

 

 

 

 

 

11,1

11.14;.

Figure 3.1.
(B) Peptide

 

116

SH

 

 

l
O CH 0
ll 2 n //0
H3N+— - —C -NH—CH —C —NH— - —C
\ -
O
(A) Cyanylation CD AP M N-<——\N+—CN
pH3-7; rt, 15' ’ 62 9 _
BF4
V
NEC —S
O \CH O
n l 2 I //0
H3N+ — —C —NH-CH —C —NH — —c
\ -
O
:NH3
(B) Cleavage
1 M NH4OH,
~1h, rt.
v
O HN S
// O
H3N+ *' —C\ + Y 11 //0
NH2 HN c—NH—- —fc\

0'

Figure 3.1. (A) Cyanylation of sulfliydryl group by CDAP and
(B) peptide bond cleavage catalyzed by ammonia

 

 

 

NaOH solutions. The?"
reaction under stronger a]
The mechanism c
2) indicates that the hydi
rising pH on the rate of
nucleophilic attack of (
mechanism, a stronger
condition that the size
nucleophilic group to t
requirement. Ammonia
the hydroxyl anion, the .
carbonyl carbon would
as shown in Figure 3.1.
hydroxyl anion catalyze
Structure replaces carbo:
Although comp;
and 1 M NH40H solutit
1 M NH4OH solution p
0'01 M OH‘ is available

agent will greatly accel

1s a better nucleophile t

anion in terms of the c

 

117

NaOH solutions. These results demonstrate the promise to carry out the cleavage
reaction under stronger alkaline conditions without sacrificing the cleavage yields.

The mechanism of cleavage reaction proposed by Jacobson et al (1) (see chapter
2) indicates that the hydroxide ion catalyzes the cleavage, as shown by the influence of
rising pH on the rate of cleavage reaction. The reaction probably proceeds only after
nucleOphilic attack of OH‘ on the carbonyl carbon of the amide. According to the
mechanism, a stronger nucleophile should promote the cleavage reaction, given the
condition that the size of the nucleophilic group is small so that the attack of the
nucleophilic group to the carbonyl atom is sterically feasible. Ammonia meets the
requirement. Ammonia has a stronger nucleophilic property than hydroxyl anion. Like
the hydroxyl anion, the size of ammonia is small. Nucleophilic attack of ammonia to the
carbonyl carbon would facilitate the cleavage giving an or-amidated N—terminal peptide,
as shown in Figure 3.1. Therefore, the N-terminal fragments obtained by ammonia and
hydroxyl anion catalyzed cleavage have one mass difference in mass spectrometry (amide
structure replaces carboxyl structure).

Although comparable results were obtained for peptide cleavage in 0.01 M NaOH
and 1 M NH4OH solutions, the latter is preferred as it provides several advantages. First,
1 M NH4OH solution provides much more nucleophilic species (:NH3 ~1 M), while only
0.01 M OH’ is available in pH 12 NaOH solution. The high concentration of nucleophilic
agent will greatly accelerate the concerted cleavage reaction. Second, although ammonia
is a better nucleophile than hydroxyl anion, it has a smaller proton affinity than hydroxyl

anion in terms of the capability of catalyzing B-elimination reaction. Therefore, the B-

 

 

 

 

 

 

 

elimination is anticipated
from our experimental 1
NaOH solution are high
solution containing 4 M
chain is more stable ar
Finally, the excess v01:
cleavage.

The representatix
i111 M NH40H solutior
gave comparable result
elimination reactions a1
typical MALDI spectra
NH40H solution. Ther
amain competitive sidc
observed. One such mi
flee SH group again,
rePorted by others (9, 1
Which probably arises

grOUps. However, th

neEligible.

 

118

elimination is anticipated to be minor in 1 M NH4OH. This tendency has been supported
from our experimental results (Table 3.2) that the yields of B-elimination in 0.01 M
NaOH solution are higher than that in 1 M NH4OH solution. Third, the 1 M NH4OH
solution containing 4 M guanidine-HCI has a moderate pH (~11), at which the peptide
chain is more stable and alkali-induced damage of peptide chains can be prevented.
Finally, the excess volatile ammonia can easily be removed from the mixture after
cleavage.

The representative chromatograms of peptides, after cleavage in pH 9.0 buffer and
in 1 M NH4OH solution, are presented in Figure 3.2 to Figure 3.5. The two conditions
gave comparable results in terms of relative yields of competitive cleavage and B-
elimination reactions and other possible side reactions. Figure 3.6 to Figure 3.9 show
typical MALDI spectra of these peptides, after cleavage in pH 9.0 buffer and in 1 M
NH4OH solution. The results from both HPLC and MALDI confirm that B-elimination is
a main competitive side reaction with cleavage reaction. Other side reactions were also
observed. One such minor reaction was the conversion of cyanylated peptides to form a
free SH group again, indicating the reversibility of the cyanylation reaction as also
reported by others (9, 10). Another minor side reaction was the dimerization of peptides,
which probably arises from nucleophilic attack of nascent thioate anions on thiocyano
groups. However, these two side reactions were always minor and in most cases

negligible.

 

 

 

Figure 3.2. HPLCt
in (A) pH 9.0, 37°C

1,11, [1, CM, and D
uncleaved species,

119

 

 

 

 

 

 

 

 

 

 

 

 

 

11
A
I
L L B
d 1 ,. __
11
B
1
grimy”
5 1'0 1'5 2? 25
min

Figure 3.2. HPLC chromatograms of cleavage products of SLRRSSCFGGR
in (A) pH 9.0, 37°C, 18h, and (B) 1 M NH4OH, rt, 1h. The peaks marked with

I, II, B, C/U, and D are fragment 1, fragment II, B-elimination product, cyanylated/
uncleaved species, and peptide dimer, respectively.

 

 

 

 

120

 

U

 

 

 

 

 

 

 

 

10 15 20 25 30

Figure 3.3. HPLC chromatograms of cleavage products of DRVYIHPCHLLYYS

in (A) pH 9.0, 37°C, 18h, and (B) 1 M NH4OH, It, 1h, respectively.

A
II
M 1 w
i3
B
I H CH]
WW
j i j I 35

 

 

 

 

 

 

 

 

 

 

 

Figure 3-4- HPLt
1n(A111H 9.0, 37

121

 

II

 

 

 

 

 

 

(in;

CM

 

 

 

 

 

 

 

 

Figure 3.4. HPLC chromatOgrams of cleavage products of EKPLQNFTLCFR

I

20
min

25 ' 30

in (A) pH 9.0, 37°C, 18h, and (B) 1 M NH4OH, rt, 1h, respectively.

 

 

 

 

 

“We 35 HPLt
MHRQEAVDCL
40H, rt, 1h, r-

 

122

 

II

C/U 1

 

 

 

 

 

 

II

 

C/U

 

 

 

 

WA

 

 

I I I I 1 I I

6 8 10 12 14 16 18 20

Figure 3.5., HPLC chromatograms of cleavage products of peptide
MHRQEAVDCLKKFNARRKLKGA in (A) pH 9.0, 37°C, 18h, and (B) 1 M
NH4OH, rt, 1h, respectively.-

 

 

 

 

 

b

500 1000

Figure 3.6. MALDI
M) pH 9.0, 37°C,
1,11, D, CM, and B I'C
uncleaved peptide, ar
N

 

++III

111
+“\/

 

1000
Figure 3.7.
cleavage in (A) pH
Symbol 4+ represer

123

 

 

 

 

 

 

 

 

 

 

11
A
B C/U
2.111221 h_/.
11
B
[3 9U
“Jinn 41 m
500 , 1000 1500 2600 2500 3000

m/z
Figure 3.6. MALDI spectra of the cyanylated peptide SLRRSSCFGGR, after cleavage

in (A) pH 9.0, 37°C, 18h and (B) 1 M NH4OH solution, rt, 1h, respectively. Symbols
1, H, D, C/U, and [3 represent fragment 1, fiagment II, peptide dimer, cyanylated/
uncleaved peptide, and its B-elimination product, respectively.

 

 

 

 

 

 

 

 

 

 

[3
A
11
++
HIL/ when»
' B
B
I 11
_ “in mnl L
500 1000 1500 m2 2000 2500 3000

Figure 3.7. MALDI spectra of the cyanylated peptide DRVYIHPCHLLYYS,'after
cleavage in (A) pH 9.0, 37°C, 18h and (B) 1 M NH4OH solution, rt, 1h, respectively.

SYmbol +1- represents a doubly charged species.

 

 

V,
.L__

l

 

 

g

500 1001

Figure 3.8. MALDI sp
in (A) pH 9.0, 37°C, 1:
IP represents the intac
\

++I

++
M
1000

 

Figure 3.9,
after cleavage in (A)

124

 

 

 

 

 

 

 

 

 

 

 

 

11 At

B

IP

2’
.. .._. r .1 IL
11 o

B
115:
I

ethnrn ,m_ r- W I ,
560 _ 1060 1500 . 2600 2500 3000

m/z

Figure 3.8. MALDI spectra of the cyanylated peptide EKPLQNFTLCFR, after cleavage
in (A) pH 9.0, 37°C, 18h and (B) 1 M NH4OH solution, rt, 1h, respectively. Symbol

IP represents the intact peptide.

 

II

 

 

 

+.+_-L. _ --JiLr M

500 10b0 (goo 2660 2560 3000
m/z

Figure 3.9. MALDI spectra of the cyanylated MHRQEAVDCLKKFNARRKLKGA,

after cleavage in (A) pH 9.0, 37°C, 18h and (B) 1 M NH4OH solution, rt, 1h, respectively.

 

 

 

 

 

 

 

 

 

 

C. Effects of Amino 1"
Degani and Patc
rates of B-elimination
dinitrophenylcysteine
Negatively charged earl
modified cysteine resid'
take place, retard the 1
esterified amino acid a
rate, presumably becau:
the functional groups
encountered if the cyar
C-terminal side. In mt
the cyanylated cystein
Slmctural diversity on
acids. This conclusic
adjacent to the cyany
undergo B-elimination
r’v‘Sidues may have the
The structural
We“ Significantly t
denatured peptides ex

with rigid or bulky s

 

 

125

C. Effects of Amino Acid Structures on Cleavage and B-Elimination Reactions

Degani and Patchornik (9) showed that the effects of amino acid structures on the
rates of B-elimination closely resemble those found in the B-elimination of S-2,4-
dinitrophenylcysteine compounds whose B-elimination is also base-catalyzed.
Negatively charged carboxylate groups located in the vicinity of the C-oc hydrogen of the
modified cysteine residue, where the attack of the catalyzing hydroxyl ion is supposed to
take place, retard the rate of the elimination. On the other hand, the presence of an
esterified amino acid adjacent to the cysteine residue greatly enhances the elimination
rate, presumably because of the electron-withdrawing effect of the ester group. However,
the frmctional groups mentioned in Degani and Patchornik’s (9) examples are rarely
encountered if the cyanylated cysteine residue in peptides does not locate in either N— or
C-terminal side. In most cases, the group located in the vicinity of the C-oc hydrogen of
the cyanylated cysteine is an amide back bone of peptide chain, which imposes little
structural diversity on the C-oc hydrogen, regardless of structural features of the amino
acids. This conclusion was supported by the observation that the same amino acid
adjacent to the cyanylated cysteine presents considerable variation in the tendency to
undergo B~elimination, whereas different amino acids adjacent to cyanylated cysteine
residues may have the same relative yield (Table 3.2).

The structural features of the amino acid on the N-terminal side of cyanylated
cysteine significantly affects both rate and yield of the cleavage reaction. For most of the
denatured peptides examined, the cleavage products are predominant. But amino acids

with rigid or bulky side chains, such as Pro and Tyr, are more resistant to cleavage,

 

 

 

 

giving B-elimina‘ion as
Son and Friedman,S r
cySteine residues after C
the reaction, in which tl
atom to provide the req
SCN group must be ori
nitrogen atom of the c
steric hindrance 0f the
residue, resulting in ‘
neighboring amino ac
reaction, which was de
itith the identical amii
other hand, is less dept
cysteine. However,
tendency and the rela
variation in pH has
Products, whereas di
tendency to form cle
dominates the relative
It should be n.
22 amino acid residue

role in the cleavage e

 

126

giving B-elimination as a main product (Table 3.2). This observation is consistent with
Som and Friedman’s report (17) that the -Pro-Cys- was protected from cleavage at
cysteine residues after cyanylation. This is understandable in terms of the mechanism of
the reaction, in which the hydroxyl anion (or ammonia) must attack the scissile carbonyl
atom to provide the required anionic tetrahedral intermediate, and the carbon atom of the
SCN group must be oriented appropriately for nucleophilic attack by the activated amide
nitrogen atom of the cyanylated cysteine. Both steps are significantly retarded by the
steric hindrance of the side chain of the amino acid adjacent to the cyanylated cysteine
residue, resulting in the decrease of the cleavage reaction rate. In additional to
neighboring amino acids, the peptide’s solution structure also affects the cleavage
reaction, which was demonstrated by the slightly different yields obtained from peptides
with the identical amino acid on the N-tel’minal side of cysteine. B-Elimination, on the
other hand, is less dependent on the structure of the amino acid adjacent to the cyanylated
cysteine. However, both cleavage and B-elimination reactions show the same pH
tendency and the relative rate can be partially compensated. This explains why the
variation in pH has little effect on the relative ratio of cleavage and B-elirnination
products, whereas different neighboring amino acids give a large variation in the
tendency to form cleavage products. It is the structure of the peptide, not pH, that
dominates the relative rate and yield of cleavage and B-elimination reactions.

It should be noted that the data presented here are deduced from peptides with 6-
22 amino acid residues. For large proteins, solution structures may play a more important

role in the cleavage, even if the protein is denatured. Many exceptions may be observed.

 

 

 

 

 

1), Application of Opt:
The Optimized
peptide chains at cya
methodology was exam
in proteins with diff
phosphokinase were ev
Figure 3.10 sho
NH,0H solution and i1
“CXperimental Section’
47, and 77, respective
Experimentally, we or
(m/z 320.3 and 517.5
cleavage products und
0f incubation at 37°C 1
an overlapped peptide

not observed if the cl

1820.1 in pH 9.0 bufi

1.17 (tn/z 1839.1) du

dehydration did not 01

One of the gt:

a
nd subsequent clear

 

D. Application of Optimized Methodology

The optimized conditions described above greatly facilitated the cleavage of
peptide chains at cyanylated cysteine residues. The feasibility of the improved
methodology was examined by experiments carried out to localize free sulflrydryl groups
in proteins with different cysteine status. Ferredoxin, ovalbumin, and creatine
phosphokinase were evaluated under different cleavage conditions to compare the results.

Figure 3.10 shows the MALDI spectra of spinach ferredoxin after cleavage in 1. M
NH4OH solution and in pH 9.0 buffer under the experimental conditions described in the
“experimental Section”. Ferredoxin contains five free sulfhydryls at positions 18, 39, 44,
47, and 77, respectively (18). Six fragments are anticipated after complete cleavage.
Experimentally, we could detect by MALDI four out of six, two low-mass fragments
(m/z 320.3 and 517.5, respectively) were missing. The mass spectra obtained from
cleavage products under both conditions show a similar pattern. However, after 18 hours
of incubation at 37°C in pH 9.0 buffer, a peak at m/z 3584.0 was observed, attributable to
an overlapped peptide 44-76 with B-elimination at Cys47. This overlapped peptide was
not observed if the cleavage was performed in 1 M NH,OH solution. The peak at m/z
1820.1 in pH 9.0 buffer is probably attributable to the dehydration products of fragment
1-17 (m/z 1839.1) during prolonged exposure to the alkaline conditions (2), while this
dehydration did not occur in l M NH4OH solution.

One of the great advantages of the proposed methodology is that the cyanylation

and subsequent cleavage are specific to free cysteine residues, selective cleavage of

 

 

 

 

128

 

 

 

 

 

18 39 44 47 77
1-'-QC|3P---SCR---SCIJSSCA---TC'A---97

 

 

 

 

 

 

SH SH SH SH SH
Fragment Expected m/z
1-17 1839.1
18-38 2352.2
39—43 517.5
44-46 320.3
47-76 3337.6
77-97 2332.5
2332” 18-43 2792.0
44-76 3580.8
2347.7 3337.6
/ A
g :2
53 E?
</ 35815
i ( 2792.8 A '
”Lyn :r A f hi, M.,.
1838.6
2332.5
2352.6 B
3337.3
500 1000 1500 2600 2500 3000 3500 4000 4500 5000
m/z

Figure 3.10. MALDI spectra of spinach ferredoxin (MW. 10483) after cyanylation
by the CDAP and subsequent cleavage in (A) pH 9.0 buffer for 18 hours at 37°C and
(B) 1 M NH4OH solution for one hour at room temperature, respectively.

 

 

 

 

 

 

1715.2

 

/2067.7
2126.6

1716.0

\71’274

V)
O\
\D
O
N
\/

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 3.11. MA
CDAP and subset

40H solution
cyanYiaied/uncle:

 

 

129

 

 

 

 

11 3O 73 120 367 382
1—--Fcl:F---Y P—-—Q Gw—QCV--—F(E1---RCITV-'-385
SH H SH SH
Fragment Expected m/z
1-10 1011.1
11-29 2378.7
30-366 37684.7
367-381 1715.1
382-385 - 429.5
1715.2 1-29 3312.8°; 3371.8d
367-385 2067.6°; 2126.6d
A
(50. V).
{323 :3 00 In
N r-t [\
\ / :8 a
I l \/
. Mat—V.—
1716.0
2379.0
B
m
In H
5323 :3 i3
8 $3 a f“
J1Li¢M

 

 

 

 

 

 

1600 2600 3600 4600 5600 6600 7600 8000
m/z

Figure 3.11. MALDI mass spectra of ovalbumin (MW. 42699) after cyanylation by
CDAP and subsequent cleavage in (A) pH 9.0 buffer at 37°C for 18. hours and (B) 1M
NH4OH solution at room temperature for one hour, respectively. c and d represent

cyanylated/uncleaved peptides and its B-elimination products, respectively.

 

 

 

protein chains at N-PCP
cystinyl residues. This
both types of residue
demonstrated by the
ovalbumin, a protein c
cyanylation by CDAP,
HCl buffer for 18 ho
temperature, respectir
complete cleavage at
detected by MALDI
informative because 1
cyanylation and cleav
385, corresponding to
WM and Cys382, rt
cleavage sites. In cor
cleavage occurs at th
almost identical. An
occurs at ‘Tyr-Cys-
fragments 11-29 and
Our methodc
sulihydryl groups in

amino acids. The p1

 

 

130

protein chains at N-peptide bonds of cysteinyl residues can be achieved in the presence of
cystinyl residues. This option is particularly useful in cases of native proteins containing
both types of residues, or in partially reduced proteins (15). This feasibility was
demonstrated by the selective cyanylation and cleavage of the free SH groups in
ovalbumin, a protein containing four free sulfliydryls and one disulfide bond ( 19). After
cyanylation by CDAP, the modified protein was subjected to cleavage both in pH 9.0 tris-
HCl buffer for 18 hours at 37°C and in 1 M NH4OH solution for one hour at room
temperature, respectively. Among five fragments expected corresponding to the
complete cleavage at the four free SH residues, only two, 367-381 and 11-29, were
detected by MALDI (Figure 3.11). These two fragments were, however, very
informative because they indicated that Cys10, 30, 367, and 382 had undergone the
cyanylation and cleavage. In addition, two sets of overlapped fragments, 1-29 and 367-
385 , corresponding to the cyanylated/uncleaved species and the B-elimination products at
Cys10 and Cys382, respectively, can be detected, which implies that Cys30 and 367 are
cleavage sites. In contrary, Cys 73 and Cys 120 are linked by a disulfide bond, and no
cleavage occurs at these residues. The results obtained by the cleavage at pH 9.0 are
almost identical. An exception to our previous discussion, however, is that the cleavage
occurs at -Tyr-Cys- (Cys30) structure in this example, resulting in the formation of
fragments 11-29 and 1-29, whereas no such cleavage was found for the model peptide.
Our methodology was further applied to recognizing the location of free
sulfliydryl groups in rabbit muscle creatine phosphokinase, a protein containing 380

amino acids. The primary structure of this protein was established from cDNA clones

 

 

 

 

3412.4

 

 

++

++

3412.3

2000 4000

 

 

Figure 3.12. MA
42977) after CYan
for 18 hours and
reSPeCfiVely Th1
in the sample. T1

131

 

73 145 253 282
1---GCV-°-HCS---FCV--- TCP —--380

SH SH SH SH

Fragment Expected m/z

1-72 8234.2
73-144 8161.0
145-252 12625
253-281 3414.0
282-380 10719
253-380 14055

3412.4

10721

 

 

 

 

8231.9

3412L3 ' 13
815952

 

 

++
2600 4000 6000 8600 10600 12000 14000
m/z

 

 

 

Figure 3.12. MALDI mass spectra of rabbit muscle creatine phosphokinase (MW.
42977) after cyanylation by CDAP and subsequent cleavage in (A) pH 9.0 at 37°C
for 18 hours and (B) 1 M NH4OH solution at room temperature for 1 hour,

respectively. The peaks marked with "*" are due to "carry-over" from an impurity
in the sample. The question marks "7" in (A) are related to two unidentified species.

 

 

 

 

 

 

 

and confirmed by 0tht
cysteines located at P0
isareactive site. Buti
the cysteines or if all
available creatine phO
modified protein was
other coexisting protei
9.0 buffer and l M N1
products under both c<
each of the four cystei
to “carry-over” from
creatine phosphokina:
cleavage products fro
fragment, m/z 14062
Cyszss, but B-elimin
are in reduced cystei
creatine PhOSphokina:

It should be 1
Shows two unidentifi
peaks are likely due t
long exposure (18 ho

was reported by Nap.

 

 

132

and confirmed by other physical methods (20). Creatine phosphokinase contains four
cysteines located at positions 73, 145, 253, and 282, respectively, among which Cy5282
is a reactive site. But it is unclear yet if there is a disulfide bond linkage between two of
the cysteines or if all the four cysteines are in the reduced form. The commercially
available creatine phosphokinase was subjected to cyanylation by the CDAP and the
modified protein was purified by reversed-phase HPLC to remove excess reagents and
other coexisting proteins. The purified creatine phosphokinase was then cleaved in pH
9.0 buffer and 1 M NH4OH solution, respectively. MALDI mass spectra of the cleavage
products under both conditions show all five fragments, corresponding to the cleavage at
each of the four cysteine residues (Figure 3.12). The peaks, marked with asterisk are due
to “carry-over” from cleavage products of a protein impurity poorly resolved from
creatine phosphokinase, as demonstrated by the analysis of MALDI spectrum of the
cleavage products from the corresponding protein. In addition, an overlapped peptide
fragment, m/z 14062, was observed in pH 9.0 buffer, corresponding to cleavage at
Cys253, but 1.3-elimination at Cy8282. It is concluded that all the four cysteine residues
are in reduced cysteine form and there is no disulfide-linked cystine in rabbit muscle
creatine phosphokinase.

It should be noted that analysis of the cleavage products from the pH 9 buffer
shows two unidentified peaks (with question mark) on the MALDI spectrum. These
peaks are likely due to protein chain decomposition or other side reactions because of the
long exposure (18 hours) to alkaline conditions during incubation. A similar observation

was reported by Nakagawa et a1 (21) who noticed that prolonged exposure of cyanylated

 

 

 

 

 

 

 

 

 

 

133

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

O
00
m
cx't
91 $3:
N
vb
:‘2
A
d.
m E .—
‘T 6'» $3. 8
F4.“ I 0°
3% ‘1‘
N m
In
CL
m.
3.
E“.
B
I; j I r r r I l r
10 15 20 25 30 35 40 45 50 55
min

Figure 3.13. HPLC chromatograms of cleavage products of creatine phosphokinase
in (A) pH 9.0, 37°C, 18h, and (B) 1 M NH4OH, rt, 1h. The peaks corresponding

to the specific cleavage at cysteine residues are marked.

 

 

 

 

 

peptide under mildly 2
succinimide formation
incubation in a modera

The MALDI s
analytes. For instance
digests, whereas a ver}
results obtained from
followed by mass-ma
Obtained under both Ct
fragments correspond
Ful‘thcrrnore, the B-el
result confirms the or

form; there is no disul

IV. Conclusions

In this chapte
structural factors on
elimination. The C
sulfhydryl groups be
oxidation of SH grou
might occur during c

two reactions are ct

 

—f—————7

134

peptide under mildly alkaline conditions resulted in the oxidative modification of Met,
succinimide formation of Asp, and other side reactions. It seems that the short time of
incubation in a moderate concentration of ammonia minimizes such side reactions.

The MALDI signal is not readily related to the relative concentration of the
analytes. For instance, the less intense peak might represent a main component in the
digests, whereas a very intense peak might represent a minor component. To confirm the
results obtained from MALDI, reversed-phase HPLC fractionation of the fragments
followed by mass-mapping of the fragments by MALDI was carried out for digests
obtained under both conditions. All the main HPLC peaks can be assigned to individual
fragments corresponding to specific cleavage at the cysteine residues (Figure 3.13).
Furthermore, the B-elimination product, 353-380, presents as a minor product. This
result confirms the conclusion that the cysteine in creatine phosphokinase is in reduced

form; there is no disulfide bond between cysteine residues.

IV. Conclusions

In this chapter, we have studied the effects of pH conditions and amino acid
structural factors on the kinetics and relative yields of the cleavage reaction and [3-
elimination. The CDAP was found superior to the NTCB for the cyanylation of
sulfhydryl groups because it reacts with SH groups under acidic conditions, where the
oxidation of SH groups is slow, the protein is kept relatively stable, and the cleavage that
might occur during cyanylation is minimal. In contrast to previous reports, although the

two reactions are competitive, higher pH does not elevate the relative yields of B-

 

 

 

elimination significant
because it can greatly
nucleophile than hydrc
The cleavage in l M
prolonged incubation
terminal side of cyar
reaction. But for most

The optimized
number of proteins w
fast, Simple, sensitive

presence of disulfide l

V. Future Work
The work desI
cysteine residues afte
provide insight intc
modification of prott
cleavage may not b
MALDI, either beca
because of the suppr
MALDI. In the ch:
sample support for d

of peptides and We

 

135

elimination significantly. 1 M ammonia solution is superior to other buffer solutions
because it can greatly accelerate the cleavage reaction because it provides a stronger
nucleophile than hydroxyl anions, and it can easily be removed from the reaction mixture.
The cleavage in 1 M NH4OH solution also minimizes the side reactions related to the
prolonged incubation of proteins. The structures of amino acids adjacent to the N-
terminal side of cyanylated cysteine residues are critical to the rate of the cleavage
reaction. But for most of the peptides studied, the cleavage products are dominant.

The optimized procedure has been used to locate free sulflrydryl groups in a
number of proteins with both known and unknown cysteine structures. This approach is
fast, simple, sensitive, and can be applied to the localization of cysteine residues in the

presence of disulfide bonds.

V. Future Work

The work described here and in chapter 2 demonstrates that chemical cleavage at
cysteine residues after cyanylation and subsequent mass-mapping by MALDI-TOF MS
provide insight into the location of free sulfltydryl groups and posttranslational
modification of protein structures. However, some of the segments from the specific
cleavage may not be detectable by the direct analysis of the digestion mixture by
MALDI, either because of , the interference of salts, buffers, and denature agents, or
because of the suppression of the responses from coexisting peptides during analysis by
MALDI. In the chapter 2, we proposed to use an immobilized inert membrane as a
sample support for direct analysis of proteins. We demonstrated that the immobilization

of peptides and proteins on an inert Zetabind membrane can be used to facilitate removal

 

 

 

 

of water-soluble MALi
and salts, while the an:
adding the matrix solu
possible on the memi
MALDI analysis is su
As a matter of fact, thi
possible reason is the
between sample holde
The electric field the i
positions.
‘ To overcomet
be tested. Instead of a
membrane material
MALDI probe tips as
mixtures. This thin
exPected to improve t
30 far, two 1
eXamined’in this lab
cyanylate SH groups
quantitative Cyanylat

cyanylation is slow (1

reagent, thiocyanopy

NTCB. The TCP, 1i

 

 

136

of water—soluble MALDI contaminants, such as high concentrations of denaturing agents
and salts, while the analyte is retained when the membrane is immersed in water prior to
adding the matrix solution. Moreover, in Situ chemical cleavage of adsorbed proteins is
possible on the membrane. However, the resolution, accuracy, and reproducibility of
MALDI analysis is substantially poorer for the analytes immobilized on the membrane.
As a matter of fact, this becomes a common drawback of using various membranes. One
possible reason is that the membrane is too thick, it behaves as an insulating layer
between sample holder (high voltage) and the protein/matrix sample on the membrane.
The electric field the ions experience is not identical for those desorbed from the different
positions.

To overcome the limitation of the current membranes, a new sample support can
be tested. Instead of a thick membrane layer, a polymer film can be formed by dissolving
a membrane material, such as nitrocellulose (22), in an organic solvent, coat it onto
MALDI probe tips as a thin fihn, and allow it to air dry prior to deposition of peptide
mixtures. This thin film would act as an inert membrane mentioned above, but is
expected to improve the MALDI signal.

So far, two reagents for specific cyanylation of sulflrydryl groups have been
examined, in this laboratory. The CDAP is superior to the NTCB in its capacity to
cyanylate SH groups under acidic conditions. However, CDAP is unstable at pH>4.5,
quantitative cyanylation of SH groups thus may not be achieved in cases where
cyanylation is slow due to structural factors of proteins. Another unexplored cyanylation
reagent, thiocyanopyridine (TCP) (23), may provide advantages over the CDAP and

NTCB. The TCP, like NTCB, can cyanylate sulfhydryl groups though the exchange of

 

 

 

was-

(2,2'-dipj

Figure 3.14. c]
thiocyanopyrid

 

137

@—S—S—NO_ ———>Rss—-@ + HS—@

(2,2'-dipyridyldisulfide) (thiol-form)

H/
(thione-form)

l
RSH + NCS—@ ————> RSCN + HS—@

(thiocyanopyridine) (thiol-form)

Figure 3.14. Chemical modification of SH groups by 2 ,2 d—ipyridyldisulfide and
thiocyanOpyridine, respectively.

 

 

 

 

the cyano group- BU‘
reactive to SH group
expectation is deduo
dipyridyldisulfide (2-1
with thiol groups in a
resonance structure of
compensated for by
protonation of its ring
shares the same struct
(Figure 3.14). The
potentially powerful
this reagent is retarde
In all, the use
reaction conditions :

Siutpler and more pla

 

138

the cyano group. But we expect the TCP, due to its structural features, will be more
reactive to SH groups and can be used in lower pH than the NTCB reaction. This
expectation is deduced from the comparison of structural similarity between 2,2’-
dipyridyldisulfide (2-PDS) and TCP (Figure 3.16). The former can quantitatively react
with thiol groups in a wide pH range (pH 1-9), as the releasing group is stabilized by
resonance structure of the thione. The low nucleophilicity of thiol groups at acidic pH is
compensated for by an increased electrophilicity of the 2-PDS as a result of the
protonation of its ring nitrogen (pKa ~3). The releasing group of TCP after cyanylation
shares the same structure as the product of 2-DPS, both having the stable thione structure
(Figure 3.14). The structural similarity between 2-PDS and TCP makes TCP a
potentially powerful reagent for cyanylation. Unfortunately, the further exploration of
this reagent is retarded at this moment, because the TCP is not commercially available.

In all, the use of a new cyanylation reagent, together with the optimized cleavage
reaction conditions and new membrane techniques, is promising to provide an even

simpler and more plausible application of our cyanylation methodology.

 

 

 

 

VI. References
1. Jacobson, G. R., E
248, 6583-6591 ( lS

2. Wu, J., Gage, D. A

U)

. Nefsky, B., and BI
4. Thomas, M. L., J
USA, 79, 1054-10
S. Roberts, D. D.,an
6. Matsudaira, F., 13
82, 6788-6792(19
7. Mahboub, S., Ric
182(1986).
8. Papayannopoulos
9- Degani, Y., andl
10. Nakagawa, S., T;
Nishimura, 0., F‘
11.DeI1slow, N, D,
Academic Press,
12. Price, N. C., BiOt
13.Liao, 1H,, and r
14. Wakselrnan, M.,

15. Wu, J., and Wat:

 

VI. References

1.

10.

11.

12.

13.

14.

15.

Jacobson, G. R., Schaffer, M. H., Stark, G. R., and Vanaman, T. C., J. Biol. Chem.
248, 6583-6591(1973).

Wu, J., Gage, D. A., and Watson, J. T., Anal. Biochem., 235, 161-174(1996).

. Nefsky, B., and Bretscher, A., Proc. Natl. Acad. Sci. USA, 86, 3549-3553(1989).

Thomas, M. L., Janatova, J ., Gray, W. R., and Tack, B. F ., Proc. Natl. Acad. Sci.
USA, 79, 1054-1058(1982).

Roberts, D. D., and Goldstein, I. J ., J. Biol. Chem, 259, 909-914(1984).

Matsudaira, P., Jakes, R., Cameron, L., and Atherton, B, Proc. Natl. Acad. Sci. USA.
82, 6788-6792(1985).

Mahboub, S., Richard, C., Delacourte, A., and Han, K., Anal. Biochem., 154, 171-
182(1986).

Papayannopoulos, LA, and Biemann, K., Protein Sci., 1, 278-288(1992).

Degani, Y., and Patchornik, A., Biochemistry, 13, 1-11(1974).

Nakagawa, S., Tamakashi, Y., Hamana, T., Kawase, M., Taketomi, S., Ishibashi, Y.,
Nishimura, 0., Fukuda, T., J Am. Chem. Soc, 116, 5513-5514(1994).

Denslow, N. D., and Nguyen, H. P., (in Techniques in Protein Chemistry VII,
Academic Press, Inc., (1996), p.241-248.

Price, N. C., Biochem. J., 159, 177-180(1976).

Liao, T-H., and Wadano, A., J Biol. Chem, 254(19), 9602-9607(1979).

Wakselman, M., and Guibe-Jampel, 13., J. C. 3. Chem. C0mm.,21-22(1976).

Wu, J ., and Watson, J. T., Protein Sci., 6, 391-398(1997)

 

 

 

 

16. Stark, G, R., in Me
17. Som, S., and Fried:
18. Matsubara, H., and
19. Nisbet, A.D., Saur
J. Biochem., 115, I
20. Putney, S., Herli
Belagaje, R., Bier
259(23), 14317-14
21.Nakagawa, S., Ta
0., Fukuda, T., Bi
22. Preston, L. M., h
550(1993).
23. Brocklehurst, K.,

As Mushiri, M, S

 

140

16. Stark, G, R., in Methods Enzymol., 47, 129-132(1977).

17. Som, S., and Friedman, S., J. Biol. Chem, 266(5), 2937-2945( 1991).

18. Matsubara, H., and Sasaki, R.M., J. Biol. Chem, 243, 1732-1757(1968).

19. Nisbet, A.D., Saundry, R.H., Moir, A.J.G., Fothergill, L.A., and Fothergill, J.E., Eur.
J. Biochem., 115, 335-345(1981).

20. Putney, S., Herlihy, W., Royal, N., Pang, H., Aposhian, H. V., Pickering, L.,
Belagaje, R., Biemann, K., Page, D., Kuby, S., and Schimmel, P., J. Biol. Chem,
259(23), 14317-14320(1984)

21. Nakagawa, S., Tamakashi, Y., Ishibashi, Y., Kawase, M., Taketomi, S., Nishimura,
0., Fukuda, T., Biochem. Biophys. Res. Commun., 200, 1735-1741(1994).

22. Preston, L. M., Murray, K. K., and Russell, D. H., Biol. Mass Spectrom., 22, 544-
550(1993).

23. Brocklehurst, K., Malthouse, J. P. G., Baines, B. S., Blenkinsop, R. D., Churcher, J.

A., Mushiri, M. S., and Ormerod, F., Biochem. Soc. Trans, Vol. 6, 261-263(1978).

 

 

 

I. Introduction

However, of the two,
chain. Cysteine is c
the codons UGC an
dehydrogenation of
proteins. Inside the
glutathione. Next to
observation suggests d
disulfide bonds in a
generally lost complett
Many biologit
disulfide bonds. Beca
it is necessary to knov
any unique contributi

With the advent of r

 

 

CHAPTER 4
A NOVEL METHODOLOGY FOR ASSIGNMENT OF DISULFIDE

BOND PAIRINGS IN PROTEINS

I. Introduction

Proteins are synthesized in cells by a stepwise process in which amino acids are
added, one by one, from the N-temrini of the chains. Most of the proteins are found to
contain SH and/or S-S groups, which are associated with cysteine and cystine residues.
However, of the two, only cysteine is incorporated directly into the initial polypeptide
chain. Cysteine is carried to the ribosome by its corresponding tRNA, which recognizes
the codons UGC and UGU. Protein S-S groups arise by subsequent pairing and
dehydrogenation of SH groups. Disulfide bonds occur frequently in extracellular
proteins. Inside the cell the sulflrydryl group is maintained in a reduced state by
glutathione. Next to tryptophan, cyst(e)ine is the most conserved amino acid. This
observation suggests that cyst(e)ine is very important for structure and fimction. If all of
disulfide bonds in a protein are reduced, both the tertiary structure and function are
generally lost completely.

Many biologically and therapeutically important proteins (peptides) contain
disulfide bonds. Because the disulfide bond is an important element of protein structure,
it is necessary to know the locations of disulfide bonds in order to understand more fully
any unique contributions they may make to protein (peptide) structure and function.

With the advent of recombinant DNA techniques for mutagenesis and expression of

141

 

 

 

 

cloned genes, it has b1

disulfide bonds into p
that the products of
proteins, or characteri

an important aspect in

 

methods for quanti

determination of the

cleavage of protein c
such as cyanogen br
bond. The resulting
sequences or molecu
mass spectrometry (
protein leads to the n

Although thi:
it is limited in man:
their amino acid cor
requirement may no
protein is very sma
cleaved between ev
more than one disu

residue by an enzyr

 

 

142

cloned genes, it has become possible to make altered structures, such as insertion of new
disulfide bonds into proteins to increase stability. In these cases, one must either verify
that the products 0f genetic manipulations are identical to those found in the native
proteins, or characterize their differences. The assignment of disulfide bonds is therefore
an important aspect in the structural characterization of proteins. Although there are good
methods for quantifying the; number of disulfide bonds in proteins, the unambiguous
determination of the location or pairing of disulfide bonds continues to challenge protein
chemists. -

Current methodology (1, 2) for assignment of disulfide bonds in proteins involves
cleavage of protein chains between half-cystinyl residues with specific cleavage reagents,
such as cyanogen bromide and trypsin, to obtain peptides that contain only one disulfide
bond. The resulting mixture of peptides is separated, and the amino acid compositions,
sequences or molecular masses of the peptides are determined by Edman degradation or
mass spectrometry or both. Assignment of these peptides to specific segments of the
protein leads to the recognition of disulfide crosslinkages.

Although this approach is well established and has been used with much success,
it is limited in many respects. First, disulfide-containing peptides can be identified by
their amino acid composition or sequence only if they are purified to homogeneity. This
requirement may not be achieved in the case of large proteins, or when the quantity of
protein is very small. Second, the above methodology requires the protein chain be
cleaved between every half-cystinyl residues so that the resulting peptides contain no
more than one disulfide pair. If the protein is not cleaved between every half-cystinyl

residue by an enzyme or chemical reagent, the peptide obtained ‘by the cleavage must be

 

 

 

 

cleaved again using

enzymatic or chemic
proteins are often res
difficult or irnpossibl
spaced or adjacent cy
are frequently used to
disulfide scrambling
approach may not be
The problem of dis
proteins, irrespective
methodologies such
been proposed to av
obtained by these nor
Gray (5, 6) h:
highly bridged sma
experiments, peptide
the partially reduced
positions of alkylate
related to the disulfi
obviously tedious, i

protein using this ap;

 

 

 

 

cleaved again using another enzyme or chemical reagent. However, even multiple-step
enzymatic or chemical digestions rarely yield a full set of diagnostic fragments because
proteins are often resistant to some specific cleavage reagents. It will be even more
difficult or impossible to find an enzyme which is likely to cleave between two closely
spaced or adjacent cysteines (1). It is likewise important to note that the conditions that
are frequently used to cleave a protein chain are also the conditions that frequently lead to
disulfide scrambling (3). As a result, the location of disulfide bonds determined by this
approach may not be the same as the location of disulfide bonds in the native protein.
The problem of disulfide exchange plagues all attempts to locate disulfide bonds in
proteins, irrespective of the general method currently used. Although supplementary
methodologies such as non-specific fragmentation by partial acid hydrolysis (4) have
been preposed to avoid disulfide bond scrambling, it is difficult to deal with the data
obtained by these non-specific techniques.

Gray (5, 6) has described an approach for analyzing disulfide linkage patterns in
highly bridged small peptides with close or adjacent cysteine residues. In his
experiments, peptides were partially reduced under controlled conditions, the isomers of
the partially reduced protein separated by HPLC, the nascent free thiols alkylated, and the
positions of alkylated cysteines recognized from the results of sequence analysis and
related to the disulfide bond pair that had been reduced and cyanylated. However, it is
obviously tedious, if not impractical, to sequence an alkylated high-mass peptide or

protein using this approach.

 

 

 

 

[1. Our Novel Strat
In chapter 2,
free cysteine groups
chemical reaction be
selectively cyanylate
cysteinyl residue can
peptide and a series
mapped to the seque
react with NTCB
groups. In chapte
conditions, and stud‘
elimination. The 0
makes this methodo.
interpretation of the c
In this chapr
assignment of disul:
cleavage of partially
resulting peptides by
for a simple protein r
reduced by tris(2-c:
produce a mixture .

Conditions can be 0;

 

 

II. Our Novel Strategy for Disulfide Assignment

In chapter 2, we developed a simple methodology for recognizing the location of
free cysteine groups in peptides and proteins (7). This approach employs a Specific
chemical reaction between sulflrydryls and 2-nitro-5-thiocyanobenzoic acid (NTCB) to
selectively cyanylate cysteine thiols (8). The N—terminal peptide bond of the modified
cysteinyl residue can then be cleaved under alkaline conditions to form an amino-terminal
peptide and a series of 2-iminothiazolidine-4-carboxylyl peptides which can be mass
mapped to the sequence of the original molecule by MALDI-MS. _ Disulfide bonds do not
react with NTCB and therefore do not interfere with the determination of sulfhydryl
groups. In chapter 3, we further improved and optimized the cleavage reaction
conditions, and studied the effects of pH and amino acid structure on cleavage and [i-
elimination. The optimized cleavage conditions greatly facilitated the cleavage and
makes this methodology much more attractive in terms of speed of the analysis and
interpretation of the data.

In this chapter, we report a novel strategy (outlined in Figure 4.1) for the
assignment of disulfide bond pairings in proteins using the above specific chemical
cleavage of partially reduced and cyanylated protein isomers with mass mapping of the
resulting peptides by MALDI-TOF MS (9). In this methodology, as shown in Figure 4.2
for a simple protein containing only two disulfide bonds, the denatured protein is partially
reduced by tris(2-carboxyethyl)phosphine (TCEP) in a buffer solution at pH 3.0 to
produce a mixture of residual intact protein and isomers of partially reduced species.

Conditions can be optimized so that the predominant products are isoforms in which only

 

 

 

Figure 4.1. I
of disulfide 1:

 

 

 

 

 

 

 

145

 

Denatured Protein

 

 

 

partial reduction TCEP, pH 3

 

Partially Reduced Protein Isomers

 

 

 

cyanylation CDAP, pH 3

 

Cyanylated Protein Isomers

 

 

 

 

 

HPLC separation
Fractionation of Identified by
, ———>
Protern Isomers MALDI-MS

 

 

 

 

 

 

cleavage 1 M NH4OI-I
V
reduction TCEP

 

V
A Mixture of Peptide Fragments

 

 

 

 

 

MALDI-MS

 

 

 

Figure 4.1. Descriptive overview of our proposed methodology for assignment
of disulfide bond pairings in proteins.

 

 

 

 

nee
+ 1
Protein 10
1.... S.“
+ 1
Protein 1°
H‘ClO
Ire:
rrcu
ITC1(
((

Figure 4.2. Chemicr
protein of interest is
each of which corre:
protein containing 11
[TC stands for iminr

 

 

146

 

50

partial reduction; TCEP @ pH 3

 

 

 

SH SH
Intact + 1 I 20 r 40 50 + 1 SI 20 T 40 5 0 Completely
Protein 10 g __39__§ 10 SH 3 0 SH Reduced Species

cyanylation; CDAP @ pH 3

 

 

 

 

 

 

 

 

 

 

 

 

 

 

S N s___
Intact + 1 lIoC 2'0 SICN 4'0 50 + 1 I 2:) 3480 4‘0 50 + Completely Reduced
Protein S 30 S 10 SCN . SCN /Cyanylated Specres
iHPLC fractionation . . .
I er identrficatron
. , b MAI _
lcleavage; 1N NH4OH y DI MS
1 9 1 L.“ 19
1TC10 %Q 29 S_Sr
S—S ITCZO 30 39
ITC30 4'0 50 ITC4O 50
reduction; TCEP
SH
1 9 1 19
§H 10 SH
ITCIO 2." 29 rrC4o—-——so
SH
MS analysis
(Cys10-Cys3 0) (CysZO-Cys40)

Figure 4.2. Chemical overview of our methodology. Partial reduction means that the
protein of interest is reduced under controlled conditions into a mixture of isoforms,
each of which corresponds to reduction of only one of its disulfide bonds; for a
protein containing 11 cystines, n isomers of the singly-reduced protein will result.
ITC stands for iminothiazolidinyl carboxyl residue at the amino terminus.

 

 

 

 

 

 

 

 

asingle disulfide b

cyanylated by l-cyano
the same buffer cond
then separated by re
MALDI-MS to dete
+52 Da correspond
corresponds to doubl
shill from the mass
cleavage in aqueous
disulfide bonds, are
mass-mapped by

to the location of
cleavage. A prim
obtained from our ap
protein isomers are I
needed to define a
confirm an assignm
related to disulfidr
performed in an aci
adjacent or close
llrerefore, two prr

methodology. TI

number of modr

 

147

a single disulfide bond has been reduced. Nascent sulfliydryls are immediately
cyanylated by 1-cyano-4-dimethylamino-pyridinium tetrafluoroborate (CDAP) (10) under
the same buffer conditions. The partially reduced and cyanylated protein isomers are
then separated by reversed-phase HPLC, followed by analysis of HPLC fractions by
MALDI-MS to determine which isomers are singly reduced/cyanylated. Those shifted by
+52 Da correspond to a singly reduced/cyanylated species; a shift of +104 Da
corresponds to doubly reduced/cyanylated species, etc. Those isomers with a 52-Da mass
shift from the mass of the intact protein are dried and subjected to specific chemical
cleavage in aqueous ammonia. The cleaved peptides, which may be linked by residual
disulfide bonds, are then completely reduced to give a mixture of peptides that can be
mass-mapped by MALDI-MS. The masses of the resulting peptide fragments are related
to the location of the paired cysteines that had undergone reduction, cyanylation, and
cleavage. A primary advantage of this approach is its underlying simplicity. The data
obtained from our approach are straightforward and easily interpreted because only a few
protein isomers are produced, and each is relevant. Typically, only n-l intermediates are
needed to define an n-bridge system. Redundant information is always obtained to
confirm an assignment. Secondly, we describe a practical way to circumvent problems
related to disulfide bond scrambling because both reduction and cyanylation are
performed in an acidic medium. Finally, our approach can be used for the assignment of
adjacent or closely spaced cysteines for which conventional methodology fails.
Therefore, two problems associated with current methodologies are solved using our
methodology. The feasibility of the new approach is demonstrated by analyzing a

number of model proteins with various disulfide bond linkages. The reported

 

 

 

 

 

 

 

experimental conditio

cyanylation, HPLC se

III. Experimental S

Mass Spectrometry

MALDI mass

the ion source was
mode of operation.
calibration using 5
5734.5), and horse 5
(St. Louis, MO). [-
acid (Aldrich Chem
were prepared in a ‘
equal volumes with

plate. The mixtur

spectrometer.

 

148

experimental conditions are optimized for partial reduction of the protein, sulflrydryl

cyanylation, HPLC separation, and cleavage.
III. Experimental Section

Mass Spectrometry

MALDI mass spectra were obtained on a Voyager Elite time-of-flight (TOF) mass
Spectrometer (PerSeptive Biosystems Inc., Framingham, MA) equipped with a model
VSL-337ND nitrogen laser (Laser Science, Newton, MA). The accelerating voltage in
the ion source was set to 25 kV. Data were acquired in the. positive or negative linear
mode of operation. Time-to-mass conversion was achieved by external and/or internal
calibration using standards of bradykinin (m/z 1061.2), bovine pancreatic insulin (m/z
5734.5), and horse skeletal myoglobin (m/z 16,952) obtained from Sigma Chemical Co.
(St. Louis, MO). All experiments were performed using oc-cyano-4-hydroxycinnarnic
acid (Aldrich Chemical Co., Milwaukee, WI) as the matrix. Saturated matrix solutions
were prepared in a 50% (v/v) solution of acetonitrile/aqueous 0.1% TF A, and mixed in
equal volumes with peptide or protein samples, and applied to a stainless-steel sample

plate. The mixture was allowed to air dry before being introduced into the mass

Spectrometer.

 

 

 

 

 

 

 

Chemicals

Tris(2—carbox
Chemical Co. (Rock
Mannheim Biocherni
A, bovine milk or-Iac
inhibitor (BPTI), ci

tetrafluoroborate (C

also purchased fro
Acetoninile and TF

pH 3.0 was prepare
with little deteriorati

freshly prepared prio

Partial Reduction 1

Ten-nmol p:
citrate buffer (pH
disulfide bonds war
in the protein (eg.
mol of ribonuclea

temperature for l1

 

 

 

149

Chemicals

Tris(2-carboxyethyl)phosphine (TCEP) hydrochloride was purchased from Pierce
Chemical Co. (Rockford, IL). Guanidine hydrochloride was a product of Boehringer-
Mannheim Biochemicals (Indianapolis, IN). Bovine pancreatic ribonuclease A type HI-
A, bovine milk oc-lactalbumin, soybean trypsin inhibitor (STI), bovine pancreatic trypsin
inhibitor (BPTI), citric acid, sodium citrate, and 1-cyano-4—dimethylamino—pyridinium
tetrafluoroborate (CDAP) were purchased from Sigma and used without further
purification. Recombinant LR3IGF-I (11271), recombinant human epidermal growth
factor (hEGF, containing a Met at the N-terminal as the initiator of biosynthesis) were
also purchased from Sigma and purified by reversed-phase HPLC before use.
Acetonitrile and TFA were of HPLC grade. The TCEP solution in 0.1 M citrate buffer at
pH 3.0 was prepared as 0.10 M stock solution and stored under N2 at -20°C for weeks
with little deterioration. The 0.10 M CDAP solution in 0.1 M citrate buffer at pH 3.0 was

freshly prepared prior to use.

Partial Reduction of Proteins

Ten-nmol protein samples were solubilized and denatured in 10 ul of 0.1 M
citrate buffer (pH 3.0) containing 1 6M guanidine-HCI. Partial reduction of protein
disulfide bonds was carried out by adding an equivalent of TCEP for the cystine content
in the protein (e.g., 40 nmol of TCEP was reacted with 10 nmol of ribonuclease A as 1
mol of ribonuclease A contains 4 moles of cystine), followed by incubation at room

temperature for 10-15 minutes. Depending on the protein understudy, the amount of

 

 

 

 

reducing agent may

reduction products.

computer. UV de
(#218TP54, 10pm

IOum particle size,
proteins were sligh
fractions were coll:
determined by MA
sample size of 10—r
columns; the use 1

sizes.

 

150

reducing agent may be adjusted to ensure the singly reduced isoforms are predominant

reduction products.

Cyanylation of Nascent Sulfhydryls
To the partially reduced protein mixture was added a 20-fold molar excess of
CDAP solution .over the total cysteine content. Cyanylation of the nascent sulfhydryl

groups was accomplished by incubation at room temperature for another 10-1 5 minutes.

HPLC Separation of Partially Reduced and Cyanylated Protein Isomers

Partially reduced and cyanylated species were separated by reversed-phase HPLC
with linear gradient elution using Waters model 6000 pumps controlled by a PC
computer. UV detection was at 215 nm. The columns were either a Vydac C18
(#218TP54, 10pm particle size, 300-A pore, 4.6x250 mm) or a Vydac C4 (#214TP54,
lOum particle size, 300-A pore, 4.6x250 mm). The HPLC conditions for individual
proteins were slightly different (see chromatograms shown below). The major HPLC
fractions were collected manually and the masses of the collected protein isomers were
determined by MALDI-MS. Appropriate fractions were then dried for further use. The
sample size of IO-nanomoles was used for convenient detection from conventional HPLC

columns; the use of microbore columns should allow the use of much smaller sample

sizes.

 

 

 

 

 

 

 

Cleavage of Singly

To the dried
aqueous solution to
of the peptide chain

removed in a vacu

Complete Reductio

Truncated p
reduced by reacting
minimize the possib

acetonitrile/0.1% T

IV. Results and Di

A. Partial Reducti

There are tr
reduction and parti
bonds in a protein
denaturing conditio
selectively split by
three disulfide bon

(ll). Sequential 1

 

 

 

 

151

Cleavage of Singly Reduced and Cyanylated Protein Isomers

To the dried HPLC fraction was added 2ul of 6 M guanidine-HCI in 1M NH4OH
aqueous solution to dissolve the protein residue and then Sul of 1 M NH4OH. Cleavage
of the peptide chain was performed at room temperature for 1 hour. Excess ammonia was

removed in a vacuum system.

Complete Reduction of Remaining Disulfide Bonds

Truncated peptides, still linked by residual disulfide bonds, were completely
reduced by reacting with Zul of 0.1 M TCEP solution at 37°C for 30 minutes at pH 3-5 to
minimize the possibility of reoxidation. Samples were diluted with lOOul of a 50% (v/v)

acetonitrile/O. 1% TFA solution prior to analysis by MALDI-MS.

IV. Results and Discussion

A. Partial Reduction of Proteins

There are two terminologies frequently used in disulfide reduction: sequential
reduction and partial reduction. The former refers to the situation in which disulfide
bonds in a protein are opened one by one by using different reducing agents and/or
denaturing conditions. For example, only the most labile disulfide bond in papain can be
selectively split by the addition of 2-mercaptoethanol in the presence of 8M urea, but all
three disulfide bonds can be reduced in 6M guanidine-HCI by the same reducing agent

(11). Sequential reduction is practically limited in use because it is difficult, if not

 

 

 

 

 

 

 

impossible, to choose

protein can be cleave
latter term, partial re
small portion of eac
desired products are
bridges. As illustra
controlled so that ea
resulting in a mixtur
is understandable
same protein molec
of each disulfide b
which are negligi
Consequently, the 1
minor singly reduce
can easily be ach
stoichiometry of rec
Although e'
reduced to a similar
of various disulfidr
different, dependir
conformational cm

(12-15). Our m1

 

 

152

impossible, to choose appr0priate conditions so that only one of the disulfide bonds in a
protein can be cleaved at a time without the reduction of the other disulfide bonds. The
latter term, partial reduction, refers to limited, controlled conditions under which only a
small portion of each of the disulfide bonds, 6. g., 10% of each, is reduced. That is, the
desired products are those in which reduction has Opened some, but not all disulfide
bridges. As illustrated in Figure 4.3 for ribonuclease A, reduction conditions can be
controlled so that each of the four disulfide bonds is reduced to a minor extent (S. 10%),
resulting in a mixture of four singly reduced isomers as a main reduced form. This result
is understandable as the probabilities of reducing two and three disulfide bonds in the
same protein molecule (under the above conditions where we observed a 10% reduction
of each disulfide bond) are 10%x10% = 1% and 10%x10%x10% = 0.1%, respectively,
which are negligible in comparison to obtaining the singly reduced isoforms.
Consequently, the mixture of partially reduced protein contains major intact proteins,
minor singly reduced isoform, trace doubly reduced isoform, and so on. Practically, this
can easily be achieved by controlling incubation time, reaction temperature, and
stoichiometry of reducing agents.

Although every disulfide bond has the same redox potential and should be
reduced to a similar extent from the standpoint of thermodynamics, the reduction kinetics
of various disulfide bonds in solutions of native or partially denatured protein are quite
different, depending on accessibility of the disulfide bonds to reducing agents, the
conformational energy of the disulfide bonds, and the redox potential of the environment

(12-15). Our methodology requires that individual disulfide bonds be broken at

 

 

 

 

 

 

26
26
SH
(m
58
SH
(11

Figure .

 

 

 

153

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

| 1—1 j 124
26 40 58 65 72 84 95 110
Ribonuclease A
partial TCEP, pH 3,
reduction 15', rt
\l
| J_—" I 124
26 40 58 65 72 84 95 1 10
(major)
["1 84 4O 1—_l 95
| I W
SH SH SH
(minor) (minor)
58 r—j 110 I 65 72 1
I I I T
SH SH SH SH
(minor) (minor)

+ doubly reduced isoforms (trace)
+ triply reduced isoforms (very trace)

OOOOOOOOOOOOOOOOOOOOO

Figure 4.3. Conceptional scheme to illustrate "Partial Reduction".

 

 

 

 

 

   

comparable rates. T

study were denatur
accessibility of reduc
Water-solubl
bonds (5-7, 9). Red
bond scrambling. F
controlled which m
The mixture
reduction states su
conducted with ribo
and LR’IGF-I, sho
cystine content in ii
15min at room tem
20—fold excess of T
From the limited 11
control is not critic

more readily be co

content may not b

protein of unkno

 

 

154

comparable rates. To minimize structural diversity of disulfide bonds, proteins under
study were denatured by dissolution in 6 M guanidine so that differences in the
accessibility of reducing agent to each disulfide bond was minimized.

Water-soluble TCEP has proved to be an excellent reducing agent for disulfide
bonds (5-7, 9). Reduction by TCEP can be carried out at pH 3.0 to suppress disulfide
bond scrambling. Furthermore, at pH 3.0, the reduction of disulfide bonds is kinetically
controlled which makes partial reduction possible (5).

The mixture of partially reduced proteins contained intermediates with different
reduction states such as singly reduced and doubly reduced isoforms. Experiments
conducted with ribonuclease A, insulin, soybean trypsin inhibitor, oc-lactalbumin, hEGF,
and LR3IGF-I, showed that by using approximately an equivalent of TCEP for the total
cystine content in the proteins, about 5~10% of each disulfide bond was reduced within
15 min at room temperature. However, for bovine pancreatic trypsin inhibitor, at least a
20-fold excess of TCEP was required to achieve 10% reduction of each disulfide bond.
From the limited number of proteins we have examined it appears that stoichiometry
control is not critical, although BPTI was an exception. The extent of the reduction can
more readily be controlled by the reaction time and temperature. However, as the cystine

content may not be known a priori, the extent of reduction should be monitored for a

protein of unknown cystine content.

 

 

 

 

 

 

B. Cyanylation of N

Jacobson et
specifically cyanylat
cyanoAdirnethylam‘
same purpose (10).
chemistry. We emp
free and total cystei
selectivity of the
conditions which or
has been reported
conditions (9, 10, 1
in peptide and pro
cyanylation by CD15
cg, pH 3-5 at ror
unnecessary to rem
CDAP. However,
over peptide sulthy
could result in me
control the stoichi
hand, CDAP reac

reaction system w

cyanylation is pos

unknown protein.

 

 

 

155

B. Cyanylation of Nascent Sulfhydryls

Jacobson et al. (8) first showed that 2-nitro-5-thiocyanobenzoic acid (NTCB)
specifically cyanylates cysteine thiols in mildly alkaline solutions. Another reagent, 1-
cyano—4-dimethylamino-pyridiniurn tetrafluoroborate (CDAP) was later proposed for the
same purpose (l0). However, only NTCB has been extensively applied in protein
chemistry. We employed the NTCB reaction to characterize the number and location of
free and total cysteine groups in peptides and proteins (7). In spite of the reactivity and
selectivity of the NTCB reagent, the cyanylation must be performed under alkaline
conditions which may permit some disulfide bond scrambling. CDAP, on the other hand,
has been reported to be a selective and reactive cyanylation reagent under acidic
conditions (9, 10, 16, 17). The utility of the CDAP for cyanylation of sulthydryl groups
in peptide and proteins has been extensively investigated in chapter 3. Complete
cyanylation by CDAP was possible by using a 5-fold molar excess under mild conditions,
e.g., pH 3-5 at room temperature, which is compatible with TCEP. It is therefore
unnecessary to remove excess TCEP, change buffer, or readjust the pH prior to using
CDAP. However, as shown in chapter 3, a large excess of CDAP (~50-fold molar excess
over peptide sulfhydryls) and excessive incubation time (>2 hours at room temperature)
could result in modification of other amino acid side chains. Therefore, it is necessary to
control the stoichiometry of CDAP in order to minimize side reactions. On the other
hand, CDAP reacts instantly with TCEP even at pH 3.0, the excess of TCEP in the
reaction system will be killed upon addition of CDAP, and no firrther reduction during
Cyanylation is possible. This is definitely an advantage for the control of reduction in an

unknown protein. However, it should be noted that a larger amount of CDAP has to be

 

 

 

 

 

applied in case mor

structure, such as B?

C. HPLC Separafio
The mixture
intact protein and a
reversed-phase HPL
bond(s), i. e., partial
4.4 is a typical
ribonuclease A. Be
are four other peaks
greater than that r
reduced/cyanylated
the same, suggestir
disulfide bonds in
other proteins will
The separ
hydrophobicity of
structure, exposes
protein’s hydroph
reduced disulfide

thus, should elute

support this spec

 

 

 

 

 

156

applied in case more TCEP is required for partial reduction of proteins with a tight

structure, such as BPTI.

C. HPLC Separation of Partially Reduced and Cyanylated Protein Isomers

The mixture obtained from the above reactions contains a majority of residual
intact protein and a minority of partially reduced/cyanylated isoforms. The capacity of
reversed-phase HPLC to separate these protein isomers with various residual disulfide
bond(s), i. e., partially reduced isoforms, is demonstrated in following examples. Figure
4.4 is a typical chromatogram of the partially reduced/cyanylated isoforms of
ribonuclease A. Besides the main peak that corresponds to residual ribonuclease A, there
are four other peaks (marked 1-4), each corresponding to a species having a mass ~52 Da
greater than that of the original protein. These peaks are attributable to singly
reduced/cyanylated protein isomers. The intensities of the four peaks are approximately
the same, suggesting that under our reaction conditions, the reducing rates of the four
disulfide bonds in denatured ribonuclease A are comparable. The HPLC separation of
other proteins will be discussed in individual sections below.

The separation mechanism in HPLC is based on the difference in the
hydrophobicity of analytes (5, 18). Opening a given disulfide bond disrupts protein
structure, exposes the protein’s interior hydrophobic amino acids, and increases the
protein’s hydrophobicity to different extents. Therefore, isomers with one or more
reduced disulfide bond(s) should tend to interact more strongly with HPLC columns and,
thus, should elute at later retention time than the original protein. The experimental data

support this speculation. Almost all partially reduced isomers show longer retention

 

 

 

 

 

 

 

20 30

Figure 4.4. HPL‘
reduced/cyanylat
phase HPLC on :
gradient 20-40%
in CH3CN. Pealr
MALDI-TOF an

 

 

157

 

 

 

 

 

 

 

 

 

 

Intact
1 2
3 4
20 3o 40 50 30 i0 8?)

Figure 4.4. HPLC separation of denatured ribonuclease A and its partially
reduced/cyanylated isomers. Ten-nmol of the protein were separated by reversed-
phase HPLC on a Vydac C18 column at a flow rate of 1.5 ml/min with a linear
gradient 20-40% B in 90 minutes, where A = 0.1% TFA in water and B =~0.1% TFA

in CH3CN. Peaks 1-4 represent singly reduced/cyanylated species, as determined by
MALDI-TOF analysis.

 

 

 

 

 

times than their intac

from the HPLC data,
dramatically change
the unusually broad
most likely due to di
suggested by Gray (
and cyanylation, e
column in the water
this case for faster
propanol within a re
It should b
isomers, it is urine
below for ribonucle
two respective prot
unambiguous assigr

described methodol

D. Cleavage of PE

One of the
bond of cyanylatr
usually employed

hours at 37°C) in 2

 

 

 

158

times than their intact original proteins. The differences in hydrophobicity, as reflected
from the HPLC data, indirectly suggests that reduction of individual disulfide bonds can
dramatically change the three-dimensional structure of a protein. Another observation is
the unusually broad HPLC peak for some proteins (e.g., or-lactalbumin, STI), which is
most likely due to different conformations that show slightly different retention times, as
suggested by Gray (5). Some hydrophobic proteins (e.g., STI), eSpecially after reduction
and cyanylation, exhibited a long retention time or even irreversible retention on the
column in the water-acetonitrile mobile phase. l-PrOpanol was substituted for CH3CN in
this case for faster elution. Most of the proteins can be eluted by 40% aqueous l-
propanol within a reasonable time (19).

It should be pointed out that although HPLC is able to separate the protein
isomers, it is unnecessary for all components to be baseline separated. As described
below for ribonuclease A, even with poorly resolved chromatographic peaks representing
two respective protein isomers, interpretation of MALDI data is still possible for the
unambiguous assignment of disulfide bonds. This provides a valuable advantage of the

described methodology for analyzing complicated mixtures.

D. Cleavage of Peptide Chains

One of the key features of our methodology is cleavage at the N—terminal peptide
bond of cyanylated cysteinyl residues. Previously described methods (7, 8, 20-23)
usually employed mildly alkaline conditions (pH~9) and long hours of incubation (>16

hours at 37°C) in attempts to control the B-elimination reaction, a side reaction that results

 

 

 

 

 

 

 

 

 

in a lower yield of

elimination over the
cleavage kinetics as
amino acids adjacent
show that higher pH
for most of the
significantly even at
within an hour at to
1M NH40H (pH~1
our experiments, 1
from the reaction 3
using higher pH i
reversed. Modifica
conditions describe
After cleav:
linked by the remai
the peptide mixtu

subsequent analysi

E. Interpretation
Unlike the

products, MALD

 

 

.159

in a lower yield of the cleavage reaction. An elevated pH was reported to favor B-
elimination over the cleavage reaction (20, 21). We have systematically studied the
cleavage kinetics as a function of pH for cyanylated polypeptides containing different
amino acids adjacent to the N-terminus of cyanylated cysteines. The results in chapter 3
show that higher pH can greatly accelerate the cleavage and B-elimination reactions, but
for most of the peptides studied, the extent of B-elimination does not increase
significantly even at a pH as high as 12, a condition at which cleavage can be complete

within an hour at room temperature. Cleavage can be accomplished in 0.01 M NaOH or

 

l M NH4OH (pH~1 1) solutions containing denaturing agent; both give similar results. In
our experiments, 1 M NH4OH is preferred over 0.01 M NaOH because it can be removed
from the reaction system after completion of the cleavage reaction. One side reaction in
using higher pH is the neutralization of -COOH groups; however, this can be easily
reversed. Modification or cleavage of other side chains was rarely observed under the
conditions described above.

After cleavage at cyanylated cysteinyl residues, the truncated peptide chains, still
linked by the remaining disulfide bonds, can be easily reduced by excess TCEP. Finally,
the peptide mixture is diluted to minimize the adverse effect of guanidine on the

subsequent analysis by MALDI.

E. Interpretation of MALDI Data
Unlike the MALDI spectra of a complicated mixture of enzymatic digestion

products, MALDI data derived from analysis of the cleavage reaction mixture of a

 

 

 

 

 

cyanylated protein ar

cyanylated cysteinyl
reduced/cyanylated
corresponding to the
related to the positio
to deduce the disul
cleavage, provides
otherwise would ha
assignment.
Ribonucleas
four disulfide bon

Cys72. Figure 4.5

,___E

26

Table 4.1 lists the
peptide chains at
cyanylated. Figur
cleavage of isome:

peaks 1-4, respect

 

 

160

cyanylated protein are easy to interpret. Cleavage of the peptide chain takes place only at
cyanylated cysteinyl sites, which in principle yields three fragments for each singly
reduced/cyanylated protein isomer (and sometimes two overlapped fragments
corresponding to the B-elimination at one cysteine site). The mass of each fragment is
related to the position of the two cyanylated cysteinyl residues which in turn can be used
to deduce the disulfide bond. linkage. B-Elimination, an alternative to peptide chain

cleavage, provides mass spectral data corresponding to overlapped peptides (that

 

otherwise would have cleaved) and serves as a confirmation for the disulfide bond pairing
assignment.
Ribonuclease A (Mr = 13,683) (24) contains 124 amino acids that are linked by

four disulfide bonds: Cy326-Cys84, Cys40-Cys95, CysS8-CysllO, and Cys65-

Cys72. Figure 4.5 shows the disulfide structure of ribonuclease A.

 

 

 

 

 

 

 

1 J m l 1,4
26 40 58 65 72 84 95 1 10

 

Figure 4.5. Disulfide structure of ribonuclease A.

Table 4.1 lists the calculated m/z values for possible fragments due to cleavage of the
peptide chains at different sites depending on which disulfide bond was reduced and
cyanylated. Figure 4.6A-D are four MALDI spectra of peptide mixtures resulting from

cleavage of isomers of singly reduced/cyanylated ribonuclease A corresponding to HPLC

peaks 1-4, respectively.

 

 

 

 

 

Table 4.1.

 

CysS 8-1

Cys4C

 

 

161

Table 4.1. Calculated and observed rrr/z values for possible fragments
resulting from the cleavage reaction of ribonuclease A chains at sites of
designated cysteine pairs

 

Reduction

 

 

Fragment Calculated m/z Observed m/z
of Disulfide
125 2702.8 2705.3
26-83 6547.3 6548.5
Cys26-Cys84 84-124 4526.0 4527.4
1-83 9176.2 9176.7
26-124 10995.3 10998.6
1-64 7083.9 7083.8
65-71 789.8 nd '
Cys6S-Cys72 72-124 5906.5 5907.7
1.71 7795.7 7790.0
65-124 . 6618.3 6617.9
1-57 6353.1 6351.1
58-109 5767.4 5766.8
Cys58-Cysl 10 110-124 1659.8 1659.8
1-109 12042.4 12036.7
58-124 7349.1 nd
1-39 4413.9 4414.4
4094 6063.6 6061.2
Cys40-Cys95 95-124 3302.7 3303.7
194 10399.5 10430.5
40-124 9288.3 9293.4

 

 

 

110-124

95'

++

L..—

 

Figure 4.6. Th
of the four sing
HPLC peaks 1
doubly-charge
peaks l~124 in
occurred at on
information. 5

 

 

 

162

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1—25 84-124 A
++ ,L 26'l83 1183* 26-124* 1424
W qt. JN-___
1-64
72124 B
*
++ g] *
.,', :
++ ‘ will A 1-124
Wk - ..x 4 wvf .._ +fi—w-JM
110-124
C
++
58-109 ;57
‘7“ \I i 1409*
In...“ A A.
95124
/1-39
D
4094 *
fl.
++ j A 3‘.
WM M A
5000 10000 15000

m/z

Figure 4.6. The MALDI mass spectra of peptide mixtures resulting from the cleavage
of the four singly reduced/cyanylated ribonuclease A isomers, corresponding to the
HPLC peaks 1-4 in Figure 4.4, respectively. The symbols ++ and * represent the
doubly-charged species and protonated B-elimination products, respectively. The
peaks 1-124 in ( A) and (B) represent intact proteins in which B-elimination only
occurred at one cysteinyl residue. These products do not provide any specific
information. See also Table 4.1 for calculated and observed m/z values.

 

 

 

 

 

 

The mass spectrum

HPLC peak 1 in Ft

fragments 1-25, 26
4526.0) with a rela
peptide chain cleav
m/z 9176.7 corres
cleavage at Cy584
10998.6 is another
at Cy584. Overall,
deduced.

With simi
CysSS-CysllO, al
The MALDI spr
represented by Hi
7083.8, correspor
fragment 65-71 i:
these two fragme
elimination prod
confirmation of t
65~124, while thi

The HPI

carefully collectt

 

 

163

The mass spectrum in Figure 4.6A corresponds to the cleavage products represented by
HPLC peak 1 in Figure 4.4. Three peaks at m/z 2705.3, 6548.5, and 4527.4 are due to
fragments 1-25, 26-83, and 84-124, respectively (expected m/z 2706.8, 6547.3, and
4526.0) with a relative mass deviation of <0.05%. From these data, one can deduce that
peptide chain cleavages occur at Cy826 and Cys84. Additionally, the MALDI peak at
m/z 9176.7 corresponds to an overlapped peptide, 1-83, resulting from peptide chain
cleavage at Cys84, but B-elimination at Cy826. Likewise, the MALDI peak at m/z
10998.6 is another overlapped peptide, 26-124, with cleavage at Cy326, but B-elimination
at Cys84. Overall, a disulfide bond linkage between Cy826—Cys84 can be unambiguously
deduced.

With similar strategy, two other disulfide bond linkages, Cys40-Cys95 and
Cys58-Cys110, also can be recognized from Figure 4.6C and Figure 4.6D, respectively.
The MALDI spectrum in Figure 4.6B, corresponding to the cleavage products
represented by HPLC peak 2 (Figure 4.4), shows two main peaks at m/z 5907.7 and
7083.8, corresponding to fragment 72-124 and 1-64, respectively. Another expected
fragment 65-71 is missing. However, it is still possible to deduce Cys65-Cys72 from
these two fragments because no other combination gives such masses. Two minor [3-
elimination products, m/z 6617.9 and 7790.0, are particularly informative for
confirmation of the assignment in this case. The peak at m/z 6617.9 represents residues
65-124, while that at m/z 7795.7 covers residues 1-71.

The HPLC peaks 1 and 2 in Figure 4.4 were not resolved completely. Even

carefully collected HPLC fractions from one component still contained a small amount of

 

 

 

 

 

 

the other componen

the cleavage produ
corresponding to
assignment of the r
possible because 0
In the earl
optimized, the com
collected as a broa
fraction is given in
spectrum A and B
pairs simultaneous
pairs in a mixture.
and doubly reduce
bond pairs from 1
gives more compl
that the assignmer
more complicate:
responses from t1

isoforrn might be

correct assignme'

 

 

 

164

the other component. This is reflected in the MALDI spectra (Figures. 4.6A and B) of
the cleavage products of these two fractions, each of which contains small fragments
corresponding to the cleavage products of the other fraction. The unambiguous
assignment of the respective disulfide bond pairs in the presence of another isomer is still
possible because only a few fragments are produced.

In the earlier stage of this research, the HPLC conditions were not fully
optimized, the compounds represented by peaks 1 and 2 in Figure 4.4 coeluted and were
collected as a broad peak. The MALDI spectrum of the cleavage products of that broad
fraction is given in Figure 4.7. It is apparent that this spectrum is a perfect overlap of the
spectrum A and B in Figure 4.6. From this spectrum, one can assign two disulfide bond
pairs simultaneously, demonstrating the capability of the methodology to assign disulfide
pairs in a mixture. On the other hand, if the HPLC peak contains a mixture of both singly
and doubly reduced/cyanylated isoforms, Caution should be used in assigning disulfide
bond pairs from the fragments. Because a doubly reduced/cyanylated protein isomer
gives more complicated fragments than a singly reduced/cyanylated protein, it is expected
that the assignment from such a mixture will be challenging. The situation could be even
more complicated, because, due to the discrimination of MALDI responses, the MALDI
responses from the cleavage products of even trace amount of doubly reduced/cyanylated
isoforrn might be more intense than those of the singly reduced/cyanylated isoform. The

correct assignment may be retarded in such a case.

 

 

 

 

 

 

1-25

 

Figure 4.7. Th
of a mixture 01
charged specie

 

 

 

165

 

1-25

72-124
84-124

1-64

 

71*

 

 

 

 

g 26-83

1' .4 1-83* 26—r24* 1'1%4

5600 10000 15000
m/z

 

 

Figure 4.7. The MALDI mass spectrum of peptide mixtures resulting from the cleavage
of a mixture of HPLC peaks 1 and 2. The symbols ++ and * represent the doubly-
charged species and protonated B-elimination products, respectively.

 

 

 

 

 

 

We have obs

chain usually is less
5767.4, and m/z 60
for the middle frag
reduces the yield of
oi-Lactalb
linked as 4 disulfid

and isomers of ‘

isomers have 5'
different disulfide
peaks 1-4 in Figur
corresponding M
calculated m/z va
lactalbumin chain
Although
mass increase 01
protein, MALDI
517.6) which we

not detected, p0:

it impossible to

 

 

166

We have observed that the middle fragment from cleavage of a cyanylated peptide
chain usually is less abundant in the MALDI spectrum (e.g., m/z 6547.3, m/z 789.8, m/z
5767.4, and m/z 6063.6 in Table 4.1). A possible explanation for this observation is that
for the middle fragment, B-elimination is able to occur at either side, which significantly
reduces the yield of the expected cleavage product.

or-Lactalbumin (123 amino acids, Mr = 14,175) also contains 8 cysteines that are
linked as 4 disulfide bonds (25). Figure 4.8 shows the HPLC separation of oc-lactalbumin
and isomers of its partially reduced/cyanylated species. Peaks for four singly
reduced/cyanylated or-lactalbumin isomers (~52-Da mass shift from original molecule,
marked 1-4) are observed in addition to that for residual or-lactalburnin. Again, the four
isomers have similar abundances which suggest comparable reduction rates for the
different disulfide bonds in denatured oc~lactalbumin. The four HPLC fractions (HPLC
peaks 1-4 in Figure 4.8) were subjected to cleavage under the described conditions. The

corresponding MALDI mass spectra are shown in Figure 4.9A-D. Table 4.2 lists the
I calculated m/z values for possible fragments resulting from the cleavage reaction of or-
lactalbumin chains at sites corresponding to different cysteine pairs.

Although the compound represented by HPLC peak 1 in Figure 4.8 showed a
mass increase of ~52 Da, corresponding to an isomer of a singly reduced/cyanylated
protein, MALDI-MS (Figure 4.9A) only detected one expected cleavage product (m/z
517 .6) which was due to fragment 120-123. Two other fragments, 15 and 6-119, were
not detected, possibly because of signal suppression. The insufficient information makes

it impossible to deduce a disulfide bond linkage from the MALDI data.

 

 

 

 

 

 

Figure 4.8. in
Cyanylated 130
of 1.5 [111/min
111 Water and I

reduced/cyan:

 

 

 

167

 

Intact

 

 

 

 

 

 

10 2'0 30 40 50 60 70

Figure 4.8. HPLC separation of denatured oc-lactalbumin and its partially reduced!
cyanylated isomers. Separation was carried out on a Vydac C4 column at a flow rate
of 1.5 ml/min with a linear gradient 40-50% B in 90 minutes, where A = 0.1% TFA
in water and B = 0.1% TF A in CH3CN. Peaks 1-4 represent singly
reduced/cyanylated species, as determined by MALDI-TOF analysis.

 

 

 

 

 

Table 4.2. (
resulting frr
designated «

Reductio
of Disulf

Cys6-Cy:

Cys28-C

Cys6l-(

Cys73

 

168

Table 4.2. Calculated and observed m/z values for possible fragments
resulting from the cleavage reaction ofa—lactalbumin chains at sites of

 

 

 

designated cysteine pairs
Reduction Fragment Calculated m/z Observed m/z
of Disulfide
1-5 618.7 nd
6-119 13135.7 nd
cys6-c.y3120 120-123 517.6 517.6
1-118 13676 nd
6-123 13575 nd
1-27 3125.6 3122.5
28-110 9525.6 9517.1
CysZS-Cysl 11 111-123 1620.8 1620.1
1-110 12573 12560
28—123 11068 nd
1-60 6918.9 6914.4
61-76 1800.8 1798.9
Cys6l-Cys77 77-123 5552.5 5557.5
1-76 8641.5 8648.6
61-123 7275.3 7280.6
1-72 8258.1 8264.5
73-90 2098.3 2099.0
Cys73-Cys91 91-123 3915.6 3912.8
1-90 10278 10303
73-123 5936.0 nd

 

 

 

 

 

 

120-123
/

L...__._..

111-123 1-27

 

 

 

 

 

 

 

 

 

r
E --76
(:1

 

111—123

E?

61—76

 

73-90

Kim.

\

 

 

Figure 4.9. The
the four Singly
Peaks L4 in Fi
Charged Specie:
for the calculat

 

169

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

120-123 A
/
..._—A— 4— #— 4 4 Aw——~_l
111-r23 1-27
B
\O
3 77-123 1-60
I \o l l 1-7628-110 _ 1-110
77-123 '
R q, C
E ‘T
S 3 1-60
\/ \61-123
1-76
:r*- ".._“...r- A .._. “‘M
91-123
D
73-90
E, i “I 1 1-72 1-90 '
A: A___.
5600 10000 15000

m/z

Figure 4.9. The MALDI mass spectra of peptide mixtures resulting from the cleavage of
the four singly reduced/cyanylated oc-lactalbumin isomers, corresponding to the HPLC
peaks 1-4 in Figure 4.8., respectively. The symbols ++ and * represent the doubly—
charged species and protonated b-elimination products, respectively. See also Table 4.2
for the calculated and observed m/z values.

 

 

 

 

The HPLC p
peak broadening. H
all the fragments re
singly reduced/cyar
1620.1 and 3122.5
peak at rn/z 9517.
usually shows 10v
elimination. Furti
product l-llO. Tl
Peaks at m/z 179
respectively, resul
76 (m/z 8640.3) 1
1“formation is co
deduced firm the
The Cy36
aIiSmg from anal
3m Figme 4.8).
the fragments 6
Products, 61.12;
is linked to Cy:
fraction 2_ Fig

Products in HP

 

170

The HPLC peaks 2 and 3 in Figure 4.8 were not baseline separated due to severe
peak broadening. However, the MALDI-MS data interpretation permitted recognition of
all the fragments resulting from the cleavage of the mixture of the two isomers of the
singly reduced/cyanylated protein. As seen in Figure 498, two intense peaks at m/z
1620.1 and 3122.5 can be assigned to fragments 111-123 and 1-27, respectively. A third
peak at m/z 9517.1 matches the middle fragment 28110 which, as mentioned above,
usually shows lower abundance due to its opportunity for double participation in B-
elirnination. Furthermore, the peak at m/z 12560 is an indication of the B-elimination
product 1-110. Thus, a disulfide bond, Cys28-Cysl 11, can be deduced. The other three
peaks at m/z 1798.9, 5552.1, and 6909.9 match fragments 61-76, 77-123, and 1-60,
respectively, resulting from cleavage at Cys6l and Cys77. The B—elimination product 1-
76 (m/z 8640.3) further confirms that Cys77 is a cleavage site. The combination of this
information is consistent with the assignment. Thus, two disulfide bond pairs can be
deduced from the mass spectrum of an impure HPLC fraction.

The Cys61-Cys77 pair is further confirmed by the mass spectrum in Fig. 4-9C,
arising from analysis by MALDI-MS of the cleavage products in HPLC fraction 3 ( peak
3 in Figure 4.8). The three main peaks at m/z 1798.9, 5557.5, and 6914.4 correspond to
the fragments 61-76, 77—123, and 1-60, respectively. In addition, two B-elimination
products, 61-123 at m/z 7280.6 and 1-76 at m/z 8648.6, provide confirmation that Cys61
is linked to Cys77. The peak at m/z 1620.0 is likely due to some “carry-over” from
fraction 2. Figure 4.9D was obtained from analysis by MALDI-MS of the cleavage

products in HPLC fraction 4 (Figure 4.8). The three MALDI peaks at m/z 2099.0,

 

 

 

 

 

 

3912.8, and 8264.5
elimination product
Cys73 is linked to C
The other e
811 (Mr. 19977) cc
and Cysl36-Cysl4
MALDI analysis <
indicating the samr
had a very poor i
between STI and (
memw
but several impur
8amilles were sul
4-10, the elevated
difficult. Table
Preliminary MA:
given in Figure
4.11A) correSpc
indicating Cyslj
they are not real
detected for frag

rs 1n3uffici em 1

171

3912.8, and 8264.5 represent fragments 73-90, 91-123, and 1-72, respectively. The B-
elimination product (m/z 10300.2) is due to fragment 1-90. These data indicate that
Cys73 is linked to Cys91.

The other example, soybean trypsin inhibitor (STI), is also demonstrated here.
STI (Mr. 19977) contains 180 amino acids linked by two disulfide bonds, Cys39-Cys86
and Cysl36-Cysl45. ( another variant of STI contains 181 amino acids with Mr. 20095.
MALDI analysis of commercial STI gave a m/z 19988 Da as a predominant peak,
indicating the sample contains 180 amino acids). Being a more hydrOphobic protein, STI

had a very poor HPLC behavior in CH3CN/HZO mobile phase. The strong interaction

 

between STI and C18 or C4 column makes the HPLC separation impractical. The HPLC
analysis of STI was improved greatly by using n-prOpanol as a mobile phase modifier,
but several impurities were still poorly resolved from the STI peak. The purified STI
samples were subjected to partial reduction and cyanylation. As illustrated in Figure
4.10, the elevated base line makes the separation of intact STI and its reduced isoforms
difficult. Table 4.3 lists the m/z values of the expected and observed fragments.
Preliminary MALDI analysis of the cleavage products from HPLC peaks 1 and 2 are
given in Figure 4.11A and B. The MALDI peaks at m/z 4036.3 and 5017.2 (Figure
4.11A) correspond to the fragment 145-180 and an overlapped peptide 136-180,
indicating Cys136 and Cys145 have been cleaved. Other peaks are also observed, but
they are not readily related to the cleavage at any cysteine sites. Since no signals were
detected for fragments 1-136 and 136-144, the information obtained by MALDI analysis

is insufficient to draw any conclusion. Figure 4.11B was obtained by the MALDI

 

 

 

 

 

 

 

 

 

 

 

Figure 4.10. r
Partially reduc

COlumn at a fl
Where A = 0_j
TFA. Peaks

MALDI-T01

 

 

172

 

Intact

 

 

 

20 25 30 35 40

Figure 4.10. HPLC separation of denatured soybean trypsin inhibitor and its
partially reduced! cyanylated isomers. Separation was carried out on a Vydac C4
column at a flow rate of 1.2 mllmin with a linear gradient 15-50% B in 30 minutes,

where A = 0.1% TFA in water and B = 80% l-propanol/20% H20 containing 0.1%

TFA. Peaks 1 and 2 represent singly reduced/cyanylated species, as determined by
MALDI-TOF analysis.

 

 

 

 

 

 

Table 4.3.
resulting 1
at sites of

 

Reducti
0f Disu

 

Cys39-C

Cysl3€

Cys39
Cys13

 

173

Table 4.3. Calculated and observed m/z values for possible fragments
resulting from the cleavage reaction of soybean trypsin inhibitor chains

at sites of designated cysteine pairs

 

Reduction

 

 

Fragment Calculated m/z Observed m/z
ofDisulfide
1-38 4058.4 4059.1
Cys39-Cys86 39-35 5258.1 5259.9
86-180 10747 10752
1-135 14971 nd
136-144 1057.1 nd
Cysl36-Cysl45 145-180 4035.5 4036.3
13(1-144) 15951 nd
8(136-180) 5015.6 5017.2
1-38 4058.4 4057.1
39-85 5258.1 5257.8
95????sz 86-135 5740.4 5738.8
ys ys [3(136-144) 1057.1 nd
8(145—180) 4035.5 4037.9

 

 

 

 

 

 

145—180

 

136-l

1-38

 

 

 

4000

Figure 4.11. T1
of the two sing
correSp0nding 1
96319 are duet

 

 

 

 

174

 

 

 

 

 

 

 

 

 

145-180
A
136-180

1-38

B
39-85
All I / 86-180
4000 6000 8000 - 10000 121100 14b00

m/z

Figure 4.11. The MALDI mass spectra of peptide mixtures resulting from the cleavage
of the two singly reduced/cyanylated soybean trypsin inhibitor (STI) isomers,
corresponding to the HPLC peaks 1 and 2 in Figure 4.10, respectively. The unidentified
peaks are due to impurities in the sample. '

 

 

 

 

 

 

 

3000 3500

Figure 4.12. "I

cleavage of un
isomers. Befo
remove CXCgSS

175

 

 

 

 

 

 

 

 

 

 

1-38
145-180
\1
. O 10
22 or
“"3 L
3000 3500 4000 4500 5000 5500 6000 6500

m/z

Figure 4.12. The MALDI mass spectrum of peptide mixtures resulting from the
cleavage of unseparated, partially reduced/cyanylated soybean trypsin inhibitor (STI)
isomers. Before the cleavage, the mixture was dialysed against 1% HAO solution to
remove excess reagents. See table 4-3 for possible fragments and their m/z values.

 

 

 

 

 

analysis of the clea'
5259.9, 10752 are
unidentified peaks
data, a Cys39-86 10
In the pres
remove excess rea:
poor chromatogra]
separation, instead
two disulfide bc
reduced/cyanylate
_Sp€Ctrum of the t
fragments cones;
at both of the disr
attributable to ;
reduced/Cyatnylat
at each of the fc

Would be difficul

F‘ Application

As We re
bond linkage 13}
Under alkaline c

Closely Spaced

 

 

 

176

analysis of the cleavage products from HPLC peak 2. The fragments at m/z at 4059.1,
5259.9, 10752 are due to the fragments 1-38, 39-85 and 86-180, respectively. Other
unidentified peaks are apparently due to impurities in the HPLC fraction. From these
data, a Cys39-86 linkage is deduced.

In the present procedure, the cyanylated proteins are separated by HPLC to
remove excess reagents and to separate the resulting isomers. For simple proteins with
poor chromatographic performances, such as STI, it could be better to avoid HPLC
separation, instead, use dialysis to get rid of excess reagents. Since STI only contains
two disulfide bonds, the cleavage products from the mixture of the partially
reduced/cyanylated isoforms were expected to be reasonably simple. The MALDI
_ spectrum of the cleavage products from the unseparated isomers (Figure 4.12) shows
fragments corresponding to the cleavage at each of the two disulfide bond sites and also
at both of the disulfide bonds (see Table 4.3). For example, the fragment at m/z 5738.8 is
attributable to fiagment 86~135. Apparently, the mixture contains the doubly
reduced/cyanylated STI, which gives rise to the fragments corresponding to the cleavage
at each of the four cysteine residues. The unambiguous assignment of disulfide pairs

would be difficult in such cases.

F. Application to Proteins Containing Closely Spaced or Adjacent Cysteines

As we reviewed in chapter 1, two problems remain in the recognition of disulfide
bond linkage by conventional methods; one is the disulfide bond scrambling occurring
under alkaline conditions and another is the failure to cleave protein chains between two

closely spaced or adjacent cysteines. Our methodology can circumvent problems with

 

 

 

 

 

disulfide bond excl
conditions. Equall
containing closely s
the proteolytic clea
of cystine and cyan
The feasibi
containing closely
human recombinar
(LR3IGF-l), and ii
Recombin;
resembled its nat
initiator 0f proteir
the disulfide bon.
assignment of dis
by an 15811 residr

containing Only c

thsnsncgp;

Figue 4.13.
epidermal gr

 

 

177

disulfide bond exchange, because the chemical reactions were performed under acidic
conditions. Equally important is that our methodology is also applicable to proteins
containing closely spaced or adjacent cysteines, because our approach does not depend on
the proteolytic cleavage between cysteine residues, but depends on the partial reduction
of cystine and cyanylation of the corresponding nascent cysteine residues.

The feasibility of our methodology to recognize disulfide linkage in proteins
containing closely spaced and adjacent cysteine residues is demonstrated here using
human recombinant epidermal growth factor (hEGF), LONG R3 insulin growth factor-I
(LR3 IGF-I), and insulin as examples.

Recombinant hEGF used in this research was purchased from Sigma. It
resembled its native form (53 amino acids), but with a Met at the N-terminus as an
initiator of protein biosynthesis (26, 27). Figure 4.13 shows the amino acid sequence and
the disulfide bond linkage in recombinant hEGF. Conventional methodologies for the
assignment of disulfide pairs in hEGF is tedious because Cys31 and Cys33 are separated
by an Asn residue, which is usually resistant to proteolytic digestion to give peptides

containing only one disulfide bond.

 

 

* J . 1.. 1
MN SD SEéPLSI-IDGYCILHDGVCMYIEALDKYA WGYIGER QYRDLKWWELR
- 1 6' 1 4 2 0 3 l 3 42 5 3

Figure 4.13. Primary structure and disulfide bond linkage of recombinant human
epidermal growth factor (hEGF) (MW. 6347.2)

 

 

 

 

 

Table 4.4.
resulting 11
designated

 

Reductior
of Disulfi

 

Cys6-Cys

Cysl4-Cy.

Cys33-C}

Cys6-Cy:
Cysl4-Cj

Cys6~Cy
Cys33-C

Cysl4-
Cys33-

Cys6.r
Cysl4
Cys33

 

178

Table 4.4. Calculated and observed m/z values for possible fragments
resulting from the cleavage of recombinant hEGF chains at sites of
designated cysteine pairs

 

Reduction

 

 

of Disulfide Fragment Calculated m/z Observed m/z
-1-5 682.7 nd
6-19 1541.6 1540.1
CYS6‘CYS20 2053 4217.9 4218.3
6-53 5676.5(5735.5) 5734.4
113 1555.6 1554.6
Cys 14-Cys3l 1430 1970.3 1969.2
31-53 2916.3 2917.1
132 3694.1 3692.9
3341 1020.1 1020.2
Cys33-Cys42 4253 1721.0 1722.0
.141 4641.2 4643.1
33-53 2664.1(2723.1) 2666.3(27259)
-1-5 682.7 11d
6-13 915.9 915.9
Cys6-Cy520 14-19 667.7 nd
(”51443531 20-30 1344.6 1345.0
31-53 2916.3 2916.0
14-30 1936.3 1934.7
46 682.7 nd
6-19 1541.6 1541.1
Cys6-Cys20 2032 1560.8 11d
Cy833-Cys42 3341 1020.1 1020.0
4253 1721.0 1721.7
3353 2664. 1(2723. 1) 2664.9(2725.1)
113 1555.6 1555.2
14-30 1970.3 1970.1
Cysl 4-Cy831 3 1-32 260.2 . . nd
Cys33,cys42 3341 1020.1 1020.6
42-53 1721.0 1722.5
3353 2664.1(2723. 1) 2666.9(2726.4)
-1-5 682.7 nd
6-13 915.9 915.8
1419 667.7 nd
Cys6-Cys20 20.30 1344.6 1344.3
Cys 14-Cy331 31-32 260.2 nd
Cys33-Cys42 33.41 1020.1 1020.1
4253 1721.0 1722.0
14-30 1936.3 1935.8
33-53 2664. 1(2723. 1) 2665.9(2727.3)

 

 

 

 

 

 

 

15

Figure 4.14.
Cyanylated i:
rate of 1.0 111
15-50 min, .
reuresent sir

anal1’Sis.

 

179

 

 

 

 

 

 

 

 

 

 

 

Intact
I
2
L7 MW
' 15 20 25 '30 35 40 45
min

Figure 4.14. HPLC separation of denatured hEGF and its partially reduced/
cyanylated isomers. Separation was carried out on a Vydac C18 column at a flow
rate of 1.0 mllmin with a linear gradient 15-35% B in 15 min, and 35-55% B from
15-50 min, where A = 0.1% TFA in water and B = 0.1% TFA in CH3CN. Peaks 1-3
represent singly reduced/cyanylated species, as determined by MALDI-TOF

analysis.

 

 

 

 

After partial
obtained by HPLC
the first peak is int:
doubly reduced/cy
Table 4.4 lists the 1
peptide chains at
reduced and cya
lactalbumin, the r1
experimental cor

characterized. Fi
cleavage of singlj
respectively. Fig
isoforms and the
The mas
represented by p
‘0 fragments 6
cYtlnylated/uncl.
is not detectabl.
chain cleavage
Cleavage produ
.41, and 42_53

4.15B). Additi

 

 

180

After partial reduction and cyanylation of hEGF, eight well-resolved peaks were
obtained by HPLC (Figure 4.14). Mass analysis of the peaks by MALDI indicated that
the first peak is intact protein, while peaks 1-3, 4-6, and 7 are singly reduced/cyanylated,
doubly reduced/cyanylated, and completely reduced/cyanylated isoforms, respectively.
Table 4.4 lists the expected and observed m/z values for fragments due to cleavage of the
peptide chains at different cysteine sites depending on which disulfide bond(s) were
reduced and cyanylated. Unlike the chromatograms of ribonuclease A and or-
lactalbumin, the reduction of two disulfide bonds in hEGF was also observed under the
experimental conditions and, therefore, the corresponding fractions were also
characterized. Figure 4.15A-C are MALDI spectra of peptide mixtures resulting from
cleavage of singly reduced/cyanylated hEGF isomers corresponding to HPLC peaks 1-3,
respectively. Figure 4.16A-D are the MALDI mass spectra of doubly reduced/cyanylated
isoforms and the completely reduced/cyanylated isoform, respectively.

The mass spectrum in Figure 4.15A corresponds to the cleavage products
represented by HPLC peak 1 in Figure 4.14. The peaks at m/z 1540.1 and 4218.3 are due
to fragments 6-19 and 20-53, respectively. The peak at m/z 5734.4 is due to a
cyanylated/uncleaved fragment 6-5 3 with an expected m/z of 5738.5. The fragment -1-5
is not detectable by MALDI in any case. From these data, one can deduce that peptide
chain cleavage occurs at Cys6 and CysZO. Therefore, Cys6 is linked to CysZO. The
cleavage products of HPLC peak 2 shows m/z values attributable to fragments -1-32, 33-
,41, and 42-53, respectively, suggesting Cys33 and Cys42 are cleavage sites (Figure

4.15B). Additionally, two overlapped peptides, 33—53 and -1-41 confirm the conclusion.

 

 

 

 

6-19

42

33-41

++

1’2

4:41

1

 

 

 

 

LE.

1000

Figure 4.15.

“the three si
1‘3 in Fig. 4.
SWiles, B-eli
also Table 4,;

 

 

 

181

 

 

 

 

 

 

 

 

 

 

6-53*
peak 1
-1-53
6-19 ++
20-53
++
42-53
peakZ
* *-
33-41 (13%
('1
m,” .132
++
-1-41
31-53
peak 3
2m
4.3 8
v-< <- 3
:++|| l l ll
(0
L _ r - _ r a. A. A... .
1000 2000 3000 4000 5000 6000 7000

m/z

Figure 4.15. The MALDI mass spectra of peptide mixtures resulting from the cleavage
of the three singly reduced/cyanylated hEGF isomers, corresponding to the HPLC peaks
1-3 in Fig. 4.14, respectively. The symbols ++ , *, and ** represent the doubly-charged
species, B-elimination products, and cyanylated/uncleaved products, respectively. See
also Table 4.4 for the calculated and observed m/z values.

 

 

 

 

Overall, a disulfide
The cleavage prodt
4.15C). In additir
disulfide bond link
detected, correspor
5) confirm that thr
sites. However,l
bring about confu
The inter
reduced/cyanylatt
reduced/cyanylat
Strategy iS help
correSponding to
sites”, that is, th
(Figure 4.16A)
fragments 6-19,
CYS42, reSpecti
Observed. Obv
Strategy, the M
the disulfide ll

MALDI spectr

that CleaVage C

g'> fragment .1

 

182

Overall, a disulfide bond linkage between Cys33-Cys42 can be unambiguously deduced.
The cleavage products from HPLC peak 3 present a little more complicated case (Figure
4.15C). In addition to the desired fragment -1-13, 14-30, and 31-53 implicating the
disulfide bond linkage between Cysl4-Cys31, two fragments, 31-41 and 42-53, were also
detected, corresponding to the undesired cleavage at Cys42. Further experiments (chapter
5) confirm that the cleavage at Cys42 is minor in comparison with the cleavage at other
sites. However, because of the high responses on MALDI, these two fragments could
bring about confusion in data interpretation.

The interpretation of MALDI data from the cleavage products of doubly
reduced/cyanylated isoforms is never as straightforward as that from singly
reduced/cyanylated isomers, because more fragments are observed. A slightly different
strategy is helpful for the assignment. Instead of looking at the MALDI peaks
corresponding to the cleavage sites, the absence of such peaks represents the “uncleaved
sites”, that is, those still linked by a disulfide bond. For example, the MALDI spectrum
(Figure 4.16A) of the cleavage products from HPLC peak 4 shows peaks due to
fragments 6-19, 33-41, and 42-53, indicating the cleavage at Cys6, CysZO, Cys33, and
Cys42, respectively. No fragments due to the cleavage at Cysl4 and Cys31 can be
observed. Obviously, these two cysteines are linked as a disulfide pair. With the same
strategy, the MALDI spectra of the cleavage products from HPLC peaks 5 and 6 reveal
the disulfide linkage Cys33—Cys42 and Cys6-Cy320, respectively. The analysis of the
MALDI spectrum of the cleavage products from the completely cyanylated hEGF shows
that cleavage occurs at each of the cysteine positions, although some of the fragments (e.

g‘a fragment -1-5) might be missing on MALDI.

 

 

 

42-5

33-41

owl

 

 

—r
k
t.

 

 

 

‘3.“ 1
.go

3
éa 5-

o
T

33-41

 

 

 

 

++

20-30
1000

 

Figure 4.16.
’ 0f the threec
cyanylated h
The SMbols

and cyanylat
and Observe.

 

 

lR’Z

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

4253
peak 4
31-53
42-53 peak 5
"" o
«5g 319: 1430*
J-J a .9135: ML - . _
42-53
3341
peak 6
41+ ‘1'13 14-30 ("in
42-53
3341 peak 7
.... at:
+v—4 r M
‘5 M m
Ar)" 20.30 M~ __
1000 2000 3000 4000 5000 6000 7000

m/z

Figure 4.16. The MALDI mass spectra of peptide mixtures resulting from the cleavage
of the three doubly reduced/cyanylated hEGF isomers and a completely reduced/

A cyanylated hEGF,corresponding to the HPLC peaks 4-7 in Fig. 4-14, respectively.
The symbols ++ , *, and ** represent doubly-charged species, B-elimination products,
and cyanylated/uncleaved products, respectively. See also Table 4.4 for the calculated
and observed m/z values.

 

 

 

 

 

Recombina
replacing glutamat
LR3IGF-I shows hi
contains adjacent
scheme, assigned «
below, has never
recombinant s0ur<
as for lGF-l isola
have a major infl

bonded isomers c

Figure 4
reduced/cyanyla
With the tYpica
First, the major
the Singly redur
lemme Smalle:
most of the pi
Second, Only t‘

2in Figure 4.1

 

 

184

Recombinant LR3IGF-I is a variant of human IGF-I that contains arginine
replacing glutamate-3 as well as an amino terminal extension of 13 amino acids (28).
LR3IGF-I shows higher biological activities than its analog, IGF-I. Like IGF-I, LR3IGF-I
contains adjacent cysteine residues at 60 and 61 positions. The disulfide bonding
scheme, assigned on the basis of homology to the insulin (or IGF-I) sequence and shown
below, has never been verified experimentally. In preparing authentic LR3IGF-I from
recombinant sources, it is important to confirm that the disulfide bond linkage is the same
as for IGF-I isolated from natural sources, since a mismatching of disulfide bonds could
have a major influence on any biological activity, as has been observed with disulfide-

bonded isomers of insulin (29).

1 19 31 6061 65 74
l j

r 89“

 

 

 

 

 

Figure 4.17 shows the HPLC separation of LR3IGF-I and isomers of its partially
reduced/cyanylated species. The chromatogram shows two exceptions in comparison
with the typical chromatographic pattern of the partially reduced/cyanylated isoforms.
First, the major peak representing the intact protein shows longer retention .than one of
the singly reduced isoforms. This result indicates that the hydrophobicity of the protein
become smaller than that of the intact protein after opening one disulfide bond, whereas
most of the proteins show larger hydrophobicity after splitting the disulfide bond(s).

Second, only two singly reduced/cyanylated protein isomers were observed (peaks 1 and

2 in Figure 4.17), even though the protein contains three disulfide bonds. Stronger

 

 

 

 

 

 

 

b

5 10

Figure 4.17
cyanylated
rate of 1.0 1
TFA in wa
Tedllced/cy

 

185

 

Intact

 

 

 

 

 

 

 

 

5 1h 1T5 20 35 30 3'5 40 45
min

Figure 4.17. HPLC separation of denatured LR’IGF-I and its partially reduced/
cyanylated isomers. Separation was carried out on a Vydac C1 8'colu1rm at a flow
rate of 1.0 mein with a linear gradient 30-50% B in 45 minutes, where A = 0.1%
TFA in water and B = 0.1% TFA in CH3CN. Peaks 1 and 2 represent singly
reduced/cyanylated species, as determined by MALDI-TOF analysis.

 

 

 

 

 

reducing condition
prolong the reduct
isomers, but the tl
isomer from the
retention time as
MALDI-MS 0f tl
disulfide bond (C
under the reductir
understandable f
stability of IGFJ
in LR3IGF-I) is
might eXplain th
Table 4,;
Cleavage reactir
The two HPLC
under the descr
Figure 4-18 A 2
fragments 65-3
the cleavage re
m/z 2752.9 an
_ Where B‘elimi;

Cleavage sites,

 

186

reducing conditions (increase the ratio of reducing agent, apply higher temperature, and
prolong the reduction time) result in the formation of doubly reduced/cyanylated protein
isomers, but the third disulfide bond still refuses to reduce. At first we assume that the
isomer fiom the reduction and cyanylation of the third disulfide bond had the same
retention time as its intact protein and, therefore, coelutes with the intact one. But
MALDI-MS of the cleavage products indicates that it is not the case. Apparently one
disulfide bond (Cys31-Cys74, as concluded from the next paragraph) is inert to reduction
under the reduction conditions, even if the protein has been denatured. This conclusion is
understandable from the thermodynamic point of view. A recent study on the disulfide

stability of IGF-I showed that the disulfide bond 18-61 (corresponding to disulfide 31-74

 

in LR3IGF-I) is superstable and preferentially formed in the folding process (30), which
might explain that this disulfide pair is structurally stable enough to resist the reduction.
Table 4.5 lists the calculated m/z values for possible fragments resulting from the
cleavage reaction of LR3IGF-I chains at sites corresponding to different cysteine pairs.
The two HPLC fractions (HPLC peaks 1 and 2 in Figure 4.17) were subjected to cleavage
under the described conditions. The corresponding MALDI mass spectra are shown in
Figure 4.18A and B. The peaks at m/z 2188.7 and 6372.9 in Figure 4.18A are due to
fragments 65—83 and 1-59, respectively. The fragment 60-64, a middle fragment during
the cleavage reactions, was missing from the MALDI spectrum. However, the peaks at
m/z 2752.9 and 6936.5, representing two overlapped peptide fragments 60-83 and 1-64,
where B-elimination occurs at Cys65 and Cys60, respectively, confirm that 60 and 65 are

cleavage sites. From these data, one can deduce that Cys60 is linked to Cys65. The

 

 

 

Table 4.5
resulting
designao

 

Reduc
0f Dis

 

Cys60

Cysl'

Cysi

 

 

Table 4.5. Calculated and observed m/z values for possible fiagments

resulting from the cleavage reaction of LR3IGF-I chains at sites of
designated cysteine pairs

 

 

Rafa“ Fragment Calculated m/z Observed m/z
of Dlsulfide
1-59 6371.2 6372.9
60-64 637.7 nd
Cys60-Cys65 65-83 2188.6 2188.7
0(1-64) 6931.9 6936.5
[3(60-83) 2749.3 2752.9
1-18 1978.4 1976.9
19-60 4538.0 4538.6
Cys 1 9-Cys61 61-83 2681.1 2682.5
[3(1-60) 6439.4 6441.9
13(19-83) 7142.1 7139.1
1-30 3223.8
31-73 4964.5
Cys3 1-Cys74 74-83 10092 no reduction
[30-73) 8111.3
501-33) 5897.7

 

 

 

 

 

 

 

1000 2

Figure 4.18.
Of the W0 Si
peaks 1 and
B-eliminatic
also Table 4

 

 

 

 

188

 

 

 

 

 

 

65-83
A
1-59
L A All 60-83* 1-64* 1-83
61-83
’ B
1-18 19-60
\7 922/ *0
“?
Bibi 521A A

 

 

 

 

1000 2000 3000 4000 5000 6600 7600 8000 9000 10000
m/z

Figure 4.18. The MALDI mass spectra of peptide mixtures resulting from the cleavage
of the two singly reduced/cyanylated LR3IGF-I isomers, corresponding to the HPLC
peaks 1 and 2 in Figure 4-17, respectively. The symbol * represent the protonated
B-elimination products. The peaks with question marks are discussed in the text. See
also Table 4.5 for the calculated and observed m/z values.

 

 

 

 

 

 

cleavage products
60, and 61-83, rt
4.18B). Addition
linked to Cys61.
can be easily assi
digestion. As i;
assume adjacent
disulfide structuj
lGF-l and 161?.
Shovvn). 11 is als
to study the refo
it shoulc
Confusing in ha
103 Da from th
the peak with
fragment 19-6(
Caused by the
exchange b ctr
unexplainable,
Cl’S-Gly. Stru.
than disulfide

other hand, 11

 

 

189

cleavage products from HPLC peak 2 show m/z values attributable to fragments 1-18, 19-
60, and 61-83, respectively, suggesting Cysl9 and Cys61 are cleavage sites (Figure
4.18B). Additionally, an overlapped peptide, 1-60, confirm the conclusion that Cysl9 is
linked to Cys61. Overall, the disulfide bond structure containing two adjacent cysteines
can be easily assigned without multiple enzymatic digestions and/or multiple step Edman
digestion. As insulin growth factor contains a number of protein analogs which all
assume adjacent cysteine structures, the data presented here are of great importance to the
disulfide structure study of such a protease. As evidence, preliminary experiments on
IGF-I and IGF-II, both containing adjacent cysteines, show similar results (data not
shown). It is also expected that the approach described here will be sufficiently powerful
to study the refolding intermediates of the IGF analogs (3 0).

It should be pointed out that the question-marked peaks in Figure 4.18B are very
confusing in nature. The peak with a single question mark showed a mass increase of
103 Da from the fragment 1-18, suggesting a cysteine attached to the fragment, whereas
the peak with double question marks showed a mass decrease of 103 Da from the
fragment 19-60, suggesting a cysteine removed from the fragment. The latter might be
caused by the cleavage at Cy559, rather than Cys60, assuming that the disulfide bond
exchange between two adjacent cysteine residues occurs. The former one is
unexplainable, because such an exchange can never occur at Cysl9, which has the -Lys-
Cys-Gly- structure. Obviously structural heterogeneity of recombinant proteins, other
than disulfide exchange, would be critical to wield occurrence of such fragments. On the

other hand, unambiguous assignment of the disulfide structure is evident even if such

 

 

 

 

undesired fragme
native disulfide b

Insulin cc
intrachain (Cys 1
residues (Cys }

cyanylated insulf

Ach

B ch

The chr
disulfide bond
Preferred. 0t
partially reduc
reduced/cyany
disulfide bond

A prol
gives a good
mode, Whilei
exl3ecte,d and

for fragment:

 

 

190

undesired fragments are present, because the fragments from specific cleavage at the
native disulfide bond are sufficient to make a positive conclusion.

Insulin contains two interchain (Cys A7-Cys B7 and Cys A20-Cys B19) and one
intrachain (Cys A6-Cys A1 1) disulfide bridges, and includes a pair of adjacent cysteine
residues (Cys A6, Cys A7). The HPLC chromatogram of the partially reduced]

cyanylated insulin isoforms is shown in Figure 4.19.

l 7 l . 20
A chain GIVEQgCASVCSLYQLBNYCN
‘ I 11 /
B chain FVNQHLCllGSHLVEALYLVgGERGFFYTPKA

The chromatogram shown by Gray (5) indicates that the reduction rates of the
disulfide bonds in native insulin vary widely and that one route of reduction was
preferred. Our chromatogram (Figure 4.19) of denatured insulin and isomers of its
partially reduced/cyanylated form shows a peak for each of the three possible singly
reduced/cyanylated species. Our data indicate that under denaturing conditions, the three
disulfide bonds undergo reduction at comparable rates.

A problem with the analysis of insulin fragments by MALIDI is that its A-chain
gives a good signal in the negative ion mode, but a very weak signal in the positive ion
mode, while the B-chain gives a good response in both modes (31). Table 4.6 shows the

expected and observed m/z values (deprotonated molecules detected in negative mode)

for fragments resulting from the cleavage reaction of insulin chains at sites corresponding

 

 

 

 

15

Figure 4.19
cl’itnylated
rate of 1.0
TFA in wa
reduced/c}

marked A

 

 

 

 

 

 

191

 

Intact

 

 

 

 

 

 

 

 

 

 

 

- 1.1 tum

15 2O 25 30 35 40 45 50

Figure 4.19. HPLC separation of denatured insulin and its partially reduced/
cyanylated isomers. Separation was carried out on a Vydac C18 column at a flow
rate of 1.0 ml/min with a linear gradient 20-50% B in 40 minutes, where A = 0.1%
TFA in water and B = 0.1% TFA in CH3CN. Peaks 13 represent singly
reduced/cyanylated species, as determined by MALDI-TOF analysis. The peaks

marked A and B are insulin A-chain and B-chain, respectively.

 

 

 

 

 

 

 

 

 

 

Table 4.6
fragment
of design

 

Reduc
0f Dis

 

CysA6-l

CysA7-

CysA2'

 

 

192

Table 4.6. Calculated and observed m/z values [M-H]' for possible
fragments resulting from the cleavage reaction of insulirchains at sites
of designated cysteine pairs

 

Reduction

 

 

. Fragment Calculated m/z Observed m/z
of Disulfide
Al-S 543.6 nd
A6-10 505.6 nd
Al 1-21 1373.8 1374.2
CysA6-CysA1 1 A6-21 1803.4 1804.8
A1-21 2340.0 2339.6
B1-30 3399.9 3400.3
Al-6 646.7 nd
A7-21 1734.9 1735.7
C A7-C B7
ys ys B1-6 755.9 nd
B7-3O 2685.1 2686.3
A1-l9 2123.7 2122.6
A20-21 261.3 nd
CysAZO-CysB 19 B 1-18 2043 .4 2041.5
B19—3O 1401.5 1402.3
B1-30 3400.0 (3366.0) 3400.1 (3366.1)

 

 

 

 

 

 

 

 

 

All-21

 

 

 

 

 

Figure 4.20.
the cleavage
HPLC peak
B‘eliminatic
Table 4.6 fC

 

193

 

 

 

 

 

 

 

 

 

 

 

 

131.30
A
_ A1-21
A11 21 A6-21*
\ kw
A7-21
B
13730
A1-19 .. r.
o a
:3 ._'.. C
Q) m
/
11319-30
l | B1-18
1000 w 1500 2000 2500 3000 3560 4000

m/z

Figure 4.20. The negative ion MALDI mass spectra of peptide mixtures resulting from
the cleavage of the three singly reduced/cyanylated insulin isomers, corresponding to the
HPLC peaks 1-3 in Figure 4.19, respectively. The symbols * and ** represent the
B~elimination products and cyanylated/uncleaved peptides, reSpectively. See also
Table 4.6 for the calculated and observed m/z values.

 

 

 

 

to different cystei
mass spectra. Fig
in Figure 4.19) sl
resulting in the fi
Al-S and A610
plagued by the
cleavage. Howe
of the adjacent
4.19 show Simil
corresponding 1
1735.7 is attrib‘
0f the insulin .
SPecific fragme
described here
proteins COHtai
Al-19, B 1-18

linked 10 CySE

V' Picomole:

So far,
by the detect
methOdologie

diSUlflde bOn‘

 

194

to different cysteine pairs. Figures 4.20A-C are the corresponding negative ion MALDI
mass spectra. Figure 4.20A (corresponding to the cleavage products of the HPLC peak 1
in Figure 4.19) shows that chemical cleavage occurs at Cys6 and Cysll of the A chain,
resulting in the fragment A1 1-21 and A6—21. Two low-mass fragments corresponding to
Al-S and A6-10 are not readily detected by MALDI. Furthermore, the mass spectrum is
plagued by the unidentified impurity peaks, probably resulting from the nonspecific
cleavage. However, it is still convincing to conclude that cleavage indeed occurred at one
of the adjacent cysteine sites, CysA6. The cleavage products from HPLC 2 in Figure

4.19 show similar results (Figure 4.20B). In addition to an expected peak at m/z 2686.3,

 

corresponding to fragment B7-30 (calculated m/z 2685.1), another intense peak at m/z
1735.7 is attributed to fragment A7-21 (calculated m/z 1734.9), resulting from cleavage
of the insulin A—chain at Cys A7 which is adjacently linked to Cys A6. These two
specific fragments suggest that Cys A7 be linked to Cys B7. Therefore, the methodology
described here shows the potential for the assignment of disulfide bond pairings in
proteins containing close or adjacent cysteine residues. Figure 4.20C gives fragments
Al-19, B 1-18, B19-30, plus an overlapped peptide B1-30. It is obvious that CysA20 is

linked to CysBl9.

V. Picomole Scale Reactions

So far, the sample size used in our protocol is ~10 nmol. The sensitivity is limited
by the detection limit of our HPLC systems. Although it is better than conventional
methodologies, our methodology described here should have the potential to assign

disulfide bonds at the picomole level of proteins. One improvement in the methodology

—¥—

 

 

 

30

Figure 4.:
Species b:

Safneast

 

 

195

 

 

 

 

awn 41mm“. ____, ,J Lm_m

 

 

30 40 5'0 60 70
minute

Figure 4.21. separation of ribonuclease A and its partially reduced/cyanylated
species by microbore C18 HPLC column (1x100mm). HPLC conditions are the

same as those in Figure 4.4, except the flow rate is 0.05 ml/min.

 

 

 

 

 

is to use a microbr
microscale is a cl
Twice as much th
The CDAP conce
used in protocol ‘
the lower conce
reduced/cyanylat
mm C18 microb
was obtained 0:
preliminary resn

samples.

V1. Character

The an:
facets of simpl
the case of mul
only one of 11
cYsteines from
masses of the
reduced isofo:
itself to 00m]

ehmination p

 

 

196

is to use a microbore column HPLC to separate the protein isoforms. To run chemistry at
microscale is a challenge to our protocol. Appropriate modifications have to be made.
Twice as much the TCEP reducing agent had to be used to promote the partial reduction.
The CDAP concentration for the subsequent cyanylation is five times as much as that
used in protocol to prevent the autohydrolysis of CDAP reagent, which is accelerated in
the lower concentration medium. Figure 4.21 shows the separation of the partially
reduced/cyanylated isoforms of 500 picomoles of ribonuclease A obtained on a 1x100
mm C18 microbore column (thanks to the generosity of Dr. Gage). A better separation
was obtained on the microbore column, while the retention time is shorter. These
preliminary results indicate that our chemistry can be used at microscale level for protein

samples.

VI. Characteristics of the Methodology

The analytical advantage of our proposed methodology accrues from several
facets of simplicity. First, the optimized procedure for partial reduction assures that in
the case of multiple disulfides there will be significant quantities of all isoforms in which
only one of the cystines has been reduced. Second, cleavage of a cyanylated pair of
cysteines from a newly reduced cystine yields only three fragments. Third, the sum of
masses of the three cleavage fragments is equal to the mass of the cyanylated singly
reduced isoform plus the mass of two molecules of water, a feature that readily lends
itself to computerized assessment of the experimentally determined data. Fourth, [3-

elimination products are usually observed which confirms the assignment of cleavage

 

 

 

 

data by peaks thi
sites. This is an
consistency of d
likelihood that a?
is quite low, evr
resolvable by TV
reduced and cyz
results reported
is the risk of dis
Because our e
scrambling are
evidence that :
indeed observe
anticipated fro
Signal
MALDI is stil
does make as:
those Shown i
electrOSpray
problem, Fu
Which may b

batch analySl

 

 

197

data by peaks that represent sequences of residues that overlap one of the two cleavage
sites. This is another feature that lends itself to computerized assessment of the internal
consistency of data interpretation. Fifth, for relatively large proteins (> 10 kDa) the
likelihood that alternative disulfide pairings would lead to the same mass spectral pattern
is quite low, even if some mass Spectral peaks for two alternative pairings may not be
resolvable by MALDI. Sixth, complete HPLC separation of the isomers of the singly
reduced and cyanylated species is unnecessary for identification as demonstrated in our
results reported here. Seventh, one of the severe problems in conventional methodology
is the risk of disulfide bond scrambling which usually occurs in an alkaline medium (1-3).
Because our experiments are performed under acidic conditions, opportunities for
scrambling are kinetically suppressed. Complete analysis of the MALDI data gives no
evidence that scrambling occurred during our procedure. Some unexpected peaks are
indeed observed in our spectra, but their masses do not appear to be related to fragments
anticipated from scrambled disulfide bonds.

Signal suppression for some components of a complex mixture in the analysis by
MALDI is still a problem (7). The suppression of signal from certain peptide fragments
does make assignments risky when they are represented by minor peaks such as some of
those shown in Figure 6. The complementary nature of analytical results by MALDI and
electrospray (ESI) (32) would seem to provide one promising way to deal with this
problem. Furthermore, ESI provides ready access for analysis of mixtures by LC-MS,

which may be an effective alternative to problems of analyte suppression encountered in

batch analysis.

 

 

 

 

VII. Conclusion

We have
chemical modifi
subsequent mas
alternative meth
known primary
minimizing disr
Furthermore, t1
cysteines for w
sensitive. We:
ionization to th

biological sour

 

 

198

VII. Conclusions

We have demonstrated that the combination of partial reduction of a protein,
chemical modification of the sulfliydryl groups, cleavage of the peptide chain, and
subsequent mass-mapping by MALDI-TOF MS provides a simple and effective
alternative methodology for the assignment of disulfide bond pairings in proteins of
known primary structure. Our novel approach offers an important advantage of
minimizing disulfide bond scrambling, a concern in most conventional methodologies.
Furthermore, this approach may be applied to proteins containing close or adjacent
cysteines for which conventional approaches fail. The procedure is simple, fast, and
sensitive. We are currently applying this methodology in combination with electrospray

ionization to the characterization of disulfide bond pairings in proteins from a variety of

biological sources.

 

 

 

 

VIII. Reference
1. Hirayama, K
(Matsuo T, <

p299-312.
2. Smith, D. L.
3. Ryle,A.P.,
4. Zhou,Z.,an
5. Gray, W. R.
6. Gray, W. R.
7. Wu ,J., Gag
8. Jacobson,(
248,6583-1

9- W181. and
10- Wakselma
11.8hapira,E
12. Creighton.
13.Kuwajimg
29, 8240-:

14. Reeve, J,
15-T€l1<eda,1
679-684(
16.Nakaga“

199

VIII. References

1.

 

10.

11

12.

13.

14.

15.

16.

Hirayama, K., and Akashi, S.,. In Biological mass spectrometry: present and future,

(Matsuo T, Caprioli RM, Gross ML, Seyama Y, eds.), New Nork, Wiley, (1994),
p299-312.

Smith, D. L., and Zhou, Z., Methods Enzymol 193, 374-389(1990).

Ryle, A. P., and Sanger, F., Biochem. J, 60, 535-540(1955).

Zhou, Z., and Smith, D. L., J. Protein Chem. 9, 523-532(1990).

Gray, W. R., Protein Sci., 2, 1732-1748(1993).

Gray, W. R., Protein Sci., 2, 1749-1755(1993).

Wu ,J., Gage, D. A., and Watson, J. T., Anal. Biochem., 235, 161—174(1996).
Jacobson, G. R., Schaffer, M.H., Stark, G. R., and Vanaman, T. C., J. Biol. Chem,
248, 6583-6591(1973).

Wu, J ., and Watson, J. T., Protein Sci., 6, 391-398(1997).

Wakselrnan, M., and Guibe-Jampel, 13., J. C. S. Chem. Comm, 21-22(l976).

. Shapira, E., and Arnon, R., J. Biol. Chem, 244(4), 1026-1032(1969).

Creighton, T. E., Prog. Biophys. Mol. Biol, 33, 231-297(1978).

Kuwajima, K., Ikeguchi, M., Sugawara, T., Hiraoka, Y., and Sugai, S., Biochemistry,
29, 8240-8249(1990).

Reeve, J. R., and Pierce, J. G., Int. J. Pept. Protein Res., 18, 79(1981').

Takeda, K., Ogawa, K., Ohara, M., Hamada, S., Moriyama, Y., J. Protein Chem, 14,
679-684(1995).

Nakagawa, S., Tamakashi, Y., Hamana, T., Kawase, M., Taketomi, S., Ishibashi, Y.,

Nishimura, 0., and Fukuda, T., J. Am. Chem. Soc, 116, 5513-5514(1994).

———

 

 

17.Nakagawa, S.
0., and Fukuc
l8. Regnier, F. E
19. Rubinstein, h
20. Degani, Y., a
21. Stark, G. R.,
22. Lu, H. 8., G1
23.Papayann0p
24.Smyth,D.C
25.Brew, K., (
4570-4582(

26. Heck, D. E
27. Chang, M
9207-9216
28Tomas, F.
Ballard, F,
2(1‘-Rinderkne
30~ Huber, S,1
1749-175.
31.Za1uzec,l
5,230-23

32- Siuzdak,

 

200

17. Nakagawa, S., Tamakashi, Y., Ishibashi, Y., Kawase, M., Taketomi, S., Nishimura,

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31

32.

O., and Fukuda, T., Biochem. Biophys. Res. Commun., 200, 1735-1741(1994).
Regnier, F. E.,"and Gooding, K. M., Anal. Biochem., 103, l-25(1980).

Rubinstein, M., Anal. Biochem., 98, 1-7(1979).

Degani, Y., and Patchornik, A., Biochemistry 13, 1-1 1(1974).

Stark, G. R., Methods Enzymol., 47, 129-132(1977).

Lu, H. S., Gracy, R. W., Arch. Biochem. BiOphys., 212, 347-359(1981).
Papayannopoulos, I. A., and Biemann, K., Protein Sci., 1, 278-288(1992).

Smyth, D. G., Stein, W. H., and Moore, S., J. Biol. Chem, 238, 227-234(1963).
Brew, K., Castellino, F. J., Vanaman, T. C., and Hill, R. L., J. Biol. Chem, 245,
45 70-45 82(1970).

Heck, D. E., J. Biol. Chem, 267, 21277(1992).

Chang, J—Y., Schindler, P., Ramseier, U., and Lai, P-H., J Biol. Chem, 270(16),
9207-92l6(1995).

Tomas, F. M., Knowles, S. E., Chandler, C. S., Francis, G. L., Owen, P. C., and
Ballard, F. J., J. Endocrinology, 137, 413-42l(1993).

Rinderknecht, E., Humbel, R. 13., FEBS Lett. 89, 283-286(1978).

Hober, S., Forsberg, G., Palm, G., Hartmanis, M., and Nilsson, B., Biochemistry, 31,

1749-1756(1992).

. Zaluzec, E. J ., Gage, D. A., Allison, J., and Watson, J .T., J. Am. So. Mass Spectrom.,

5, 230-237(1994).

Siuzdak, G., Proc. Natl. Acad. Sci. USA, 91, 11290-1 1297(1994).

 

 

 

 

DISULFII
RECOMB

(hEGF) Bi

1. Introduction
Proteins:
proteins must 2
information for a
acids in the poi
Process involvin
that the major e
Perhaps even art
into their natii
Protein’s prim
scientists have .
is, the folding
attempts to 801'
The stt
(Thestions: (1)

hiOlOgically‘ac

 

 

CHAPTER 5
DISULFIDE MAPPING OF FOLDING INTERMEDIATES OF
RECOMBINAN T HUMAN EPIDERMAL GROWTH FACTOR
(hEGF) BY CDAP TRAPPING, CHEMICAL CLEAVAGE, AND

MASS SPECTROMETRY

I. Introduction
- Proteins are synthesized within the cell on ribosomes. To be biologically active, all
proteins must adopt specific folded three-dimensional structures. Yet the genetic
information for a protein specifies only the primary structure, the linear sequence of amino
acids in the polypeptide backbone. Although folding in the cell is a highly complex
process involving a cascade of helper proteins called the molecular chaperones, it is clear
that the major events in the folding process occur after departure from the ribosome and
perhaps even after release from the chaperones (1, 2). Proteins can spontaneously fold
into their native conformation under physiological conditions. That implies that a
protein’s primary structure dictates its three-dimensional structure. For many years,
scientists have endeavored to understand the process by which protein folding occurs, that
is, the folding pathway. Although a wealth of information has been accumulated in
attempts to solve this problem, the answer is far from complete.
The study of protein folding can be broken down into three different but related
questions: (1) By what kinetic process or pathway does a protein adopt its native and

biologically-active folded conformation? (2) What is the physical basis of the stability of

201

 

 

 

 

 

the folded confor
folding process 2
Protein '
practical point 0
to exploit such
process of in Vil
Many 0
be overcome if
protein-stabiliz
its oxidation/rt
the disulfide l
folding pathw
requires the
refolding accr
been proven 1
6)- Two us:
developed te
amide group;
be alhplied t1
“01 arnenabl
intermediate
bonds, is 01

intehnediatr

202

the folded conformation? (3) Why does the amino acid sequence determine one particular
folding process and resultant three-dimensional structure, instead some other?

Protein folding study is of great significance both from the theoretical and
practical point of View. Recent advances in genetic engineering have provided the means
to exploit such knowledge, thereby adding a new urgency for the need to understand the
process of in vitro protein folding.

Many of the difficulties associated with determining protein folding pathways can
be overcome if folding is coupled to disulfide bond formation. This is the only type of
protein-stabilizing interaction that is susceptible to specific experimental control, due to

its oxidation/reduction nature. Since 80% of proteins in nature contain disulfide bond(s),

 

the disulfide formation study is of great significance for the understanding of protein
folding pathways. Folding is coupled to disulfide formation if the folded conformation
requires the correct disulfide bond linkage. The reduced protein is unfolded, and
refolding accompanies disulfide formation. Characterization of folding intermediates has
been proven to be a very powerful tool in the elucidation of protein-folding pathways (3-
6). Two useful techniques are currently available to tackle this problem. A newly
developed technique of pulsed-label NMR (7, 8) permits trapping and identification of
amide groups that are engaged in the structured elements. This technique can in principle
be applied to all types of proteins. However, intermediates trapped by this method are
not amenable to chromatographic purification because of the transient nature of folding
intermediates (9). An established method, which utilizes the formation of disulfide
bonds, is often used. In this method, as described in Figure 5 .1, the disulfide-bonded

intermediates which form during the folding process are chemically trapped in a time

—;

 

 

 

 

course manner. '
kinetics of the
connected to a 1
containing prote
further purified
governed by the

the simplest apj

small

 

203

course manner. The trapped intermediates are isolated and structurally characterized, the
kinetics of the interconversion are determined, and the different components are
connected to a folding pathway. This technique is limited to the analysis of disulfide
containing proteins, but a unique advantage is that species of trapped intermediates can be
further purified for characterization. Because the formation of disulfide bonds is
governed by the thiol/disulfide concentration of the redox couple used, it has represented

the simplest approach to study the folding of proteins. By varying the concentrations of

small

reduced/unfolded protein

1

fold at pH 8.5

iodoacetate acid quench
at pH 8.5 at pH<4

v

trap intermediates

 

HPLC

separate intermediates
classical approach

determine disulfide structure

Figure 5.1. Conventional approach for disulfide mapping of
protein folding intermediates

 

 

 

 

molecular weigl
intermediates as
predictable man
accumulated ant
Even wi
have been char
have been stuC
(BPTI) (15-1 8)
The fit:
type protease
original mode
intermediates
and those thi
imponant int
modem separ
existed five 8
Species) and
the model 1
intermediate:
Stabilize the
dictating the

mainly 0f 11;

204

molecular weight thiol and disulfide components in solution, the distribution of
intermediates as well as the rate of regeneration of native protein can be controlled in a
predictable manner. This method thus allows construction of disulfide bond isomers
accumulated and trapped along the process of folding.

Even with these advantages, the problem remains nontrivial, and few pathways
have been characterized completely. The folding pathways of several proteins (1 0- 14)
have been studied in this manner. Among them, bovine pancreatic trypsin inhibitor

(BPTI) (15-18), and ribonuclease A (9, 19-22) have been the most extensively studied.

 

The first case of a well-documented disulfide folding pathway is BPTI, a Kunitz-' '
type protease inhibitor which comprises 58 amino acids and three disulfides. In the
original model of BPTI folding (15, 23), eight well-populated 1- and 2-disulfide
intermediates were identified. Five were shown to contain exclusively native disulfides,
and those that adopted non-native disulfides have been suggested to be kinetically

important intermediates. This original BPTI model was recently reexamined using

 

modern separation and analytical methodologies (17, 18 , 24), it was concluded that there

 

existed five species of well-populated intermediates (two l-disulfide and three 2-disulfide
species) and all of them contain only native disulfide bonds. Despite the inconsistency,
the model of BPTI folding maintains three principal conclusions: (a) the folding
intermediates consist of only one- and two-disulfide species; (b) specific interactions that
stabilize the native BPTI play a crucial role in guiding the folding early on and hence
dictating the formation of limited numbers of well-populated intermediates that consist

mainly of native disulfides. The rate-limiting step of BPTI folding is the intramolecular

 

———

 

 

 

rearrangements <
rapidly forms the
However
nontrivial. The
number of cysti:
three or fewer
disulfide protei
internal disulfi
separation of i:
tasks. A clea
pr0posed only
Traditi
reagent that 33
exchange, l
intermediates
ribonuclease
scale as alky
i0doacetate c
scale as the
PIOblem for
0f a"‘Wlatior
Because the

‘0 quench fi

 

 

205

rearrangements of two-disulfide intermediates to generate the immediate precursor that
rapidly forms the third and final native disulfide.

However, although this approach has limited success so far, the problem remains
nontrivial. The number of possible disulfide intermediates increases rapidly with the
number of cystines in the protein, and success has been claimed with proteins having only
three or fewer disulfide bonds. The determination of a folding pathway for a three
disulfide protein is already a difficult undertaking, there being 74 possible different
internal disulfide-bonded intermediates excluding the reduced and native states. The
separation of intermediates and the determination of disulfide structures are formidable
tasks. A clearly defined pathway for formation of a three-disulfide species has been
proposed only in the case of BPTI.

Traditionally, folding intermediates were trapped by addition of iodoacetate, a
reagent that alkylates free thiols and thereby prevents further oxidation or thiol-disulfide
exchange. However, this method was recently questioned. The rearrangement of
intermediates during trapping with iodoacetate was observed for both BPTI (17) and
ribonuclease A (25) since thiol-disulfide exchange is expected to occur on the same time
scale as alkylation by iodoacetate. For example, alkylation of thiols with 100 mM of
iodoacetate occurs with a half-time of about 1 second at pH 8, which is on the same time
scale as the kinetics of disulfide rearrangement (17). This could be a more severe
problem for partially structured intermediates where steric hindrance would retard the rate
of alkylation of some thiols (26). As an alternative, acid quenching was proposed.
Because the thiolate anion is the reactive species in thiol/disulfide exchange, it is possible

to quench folding extremely rapidly by lowering the pH. The subsequent reversed-phase

 

 

 

 

 

HPLC at pH 2
introducing sigr
acid quenching
quenched inten
occur (17, 27).
before the disu
procedure.
Both 0
(19), who insi
completely, or
2—aminoethyl
because it rea
thiol/disulfidr
exchange rez
bonds, thusc'
Rece'
fOIding inte
phenylarson
the presenc.
allowing tr:
Yet formed
much on th

and) therefi

 

 

206

HPLC at pH 2 is well suited for separation of acid-quenched intermediates without
introducing significant rearrangements during the separation. A practical advantage of
acid quenching is that it is reversible. As a result, it is possible to purify an acid-
quenched intermediate and subsequently to allow further rearrangement or folding to
occur (17, 27). On the other hand, the intermediatestrapped by acid have to be modified
before the disulfide bond structure can be defined. In this respective, it is a cumbersome
procedure.

Both of the above trapping methods were attacked by Rothwarf and Scheraga
(19), who insisted that the criteria for a good blocking agent are that it block quickly,
completely, and without modifying the protein at sites other than thiols. They proposed
2-aminoethyl methanethiosulfonate [(NH2)CZHSSS02CH3] (AEMTS) as a trapping agent
because it reacts specifically and much more quickly with thiol groups and thus prevents
thiol/disulfide bond exchange. However, the trapping by AEMTS is based on the SH/S-S
exchange reaction between an intermediate and the reagent to form mixed disulfide
bonds, thus disulfide scrambling during analytical processing is still questionable.

Recently, Happersberger et al (28) proposed a novel method for trapping protein~
folding intermediates by selective modification of bis-cysteine sulfhydryls with
phenylarsonous acid derivatives (PAA), forming cyclic dithioesters in a wide pH range in
the presence of two suitably spaced sulflaydryl groups of a peptide or protein, thereby
allowing trapping and characterization of protein folding intermediates which have not
yet formed covalent disulfide bonds. But, derivatization by this reagent depends too

much on the structural accessibility of bis-cysteine SH groups which undergo the reaction

and, therefore, is limited in practical use.

 

 

 

 

The isola
based on charge
of disulfide bon
efficiency of it
separation of ir
better altematis

In addi‘
for recognitior
cumbersome.

To 51
intermediates
large demand

In cha
Pairings in pr
demonstrated
applicable to
ChalPter, we r
in folding inr
Figu]
intermediate
free sulfhyd
by reversed

the diSUIfidt

 

 

207

The isolation of intermediates can be achieved by ion exchange chromatography,
based on charge differences of species under separation. The information about the state
of disulfide bonding is thus predictable from the elution order. However, the low column
efficiency of ion chromatography prevents it from becoming a powerful tool for the
separation of individual intermediates. Reversed-phase HPLC has been proven to be a
better alternative for the separation, especially for the separation of small protein isomers.

In addition to the inconsistency of trapping methods, the conventional approach
for recognition of disulfide bond structures of folding intermediates is very tedious and
cumbersome. Problems mentioned in chapter 4 still plague the methodology.

To summarize, disulfide bond structure elucidation of protein folding
intermediates is still a touchy and formidable problem in protein chemistry. There is a
large demand for the development of new techniques.

In chapter 4, we have developed a novel methodology to assign disulfide bond
pairings in proteins by partial reduction/cyanylation/cleavage/mass mapping. We have
demonstrated that the methodology can minimize disulfide bond scrambling and be
applicable to proteins containing closely spaced or adjacent cysteine residues. In this
chapter, we exploit the feasibility of our novel methodology to study disulfide structures
in folding intermediates and its potential advantages over conventional approaches.

Figure 5.2 illustrates our new procedure. After folding is initiated, the folding
intermediates are trapped by CDAP acidic solution (pH 3) which specifically cyanylates
free sulfhydryl groups and thus effectively quenches the folding process. After separation
by reversed-phase HPLC, the folding intermediates are identified by MALDI-MS, and

the disulfide bond structures characterized by our novel approach described in chapter 4.

 

 

 

 

Figur

COIIT
techniques i
mapping tec
the trappin
SUlfllydryl/c‘
quantitative
quench, Wh

delemrine

 

208

reduced/unfolded protein

J

fold at pH 8.5

trapped and quenched
by CDAP acidic solution

 

v

trap intermediates

HPLC

separate intermediates
Our novel approach

determine disulfide structure

Figure 5.2. New approach for disulfide mapping of protein folding intermediates

Compared to the conventional approach, our new methodology introduces two
techniques into the folding area: the disulfide trapping technique and the disulfide
mapping technique. These improvements provide several important advantages. First,
the trapping reaction, occurring in acidic solution, greatly reduces the risk of
sulflrydryl/disulfide exchange that bothers iodoacetate trapping. Secondly, the CDAP
quantitatively cyanylates free sulflrydryl groups to stop further folding. Unlike acid
quench, which is totally reversible and requires further chemical modification in order to

determine the disulfide structure of the folding intermediates, the cyanylated

 

 

 

 

 

intermediates ca
new approach.

advantages over

6 .
usnsncrrsr
L_

Figure

We us
protein contai
an example tc

Prelin
were recently

intermediate:
trap the inte
iOdoacetatet
determined
dlSUlflde Sir

We

CDAP tech

 

209

intermediates can be directly subjected to characterization of disulfide structure by our
new approach. Thirdly, our disulfide bond mapping technique itself provides unique

advantages over conventional techniques, as demonstrated in previous chapters.

6 14 20 31 33 42
NSDSECIIPLSHDGYCLHDGV(EMYIEALDKYACNCIIVVGYIGERJCQYRDLKWWELR

 

 

 

 

Figure 5.3. Disulfide linkage of human epidermal growth factor (hEGF).

We used recombinant human epidermal growth factor (hEGF), a small, compact
protein containing 5 3 amino acid residues linked by three disulfide bonds (Figure 5.3), as
an example to demonstrate the feasibility of our proposed methodology.

Preliminary results on the disulfide structures of folding intermediates of hEGF
were recently reported by Chang et al (29). In this paper, the distribution of disulfide
intermediates is difi‘erent depending on whether iodoacetate or acidic quench was used to
trap the intermediates, suggesting possible thiol/disulfide exchange in the process of
iodoacetate trapping. The disulfide structures of a few well-populated intermediates were
determined by classical methods, indicating that most of them contain non-native
disulfide structure.

We isolated and identified over 18 disulfide folding intermediates trapped by the

CDAP technique. The disulfide structures of seven well-populated were characterized by

 

 

 

 

 

our new appro:
conventional or
is to elucidate t

methodology d

11. Experimer

Mass Spectre
MALI
spectrometer
extraction an
accelerating
voltages wer
acquired in
aChieved by
1061.2), b0.
16,952) obt
Performed 1
W1) as the
acetoIlitrile
Samples, 21;

befom beir

 

210

our new approach. Our methodology is simpler, faster, and more sensitive than the
conventional one. Although the final goal of characterizing disulfide bond intermediates
is to elucidate the folding pathway of the protein, this chapter will be only focused on the

methodology development.
11. Experimental Section

Mass Spectrometry

MALDI mass spectra were obtained on a Voyager Elite time—of-flight (TOF) mass
spectrometer (PerSeptive Biosystems Inc., Framingham, MA) equipped with delayed
extraction and a model VSL-337ND nitrogen laser (Laser Science, Newton, MA). The
accelerating voltage in the ion source was set to 20 kV. Grid voltage and guide wire
voltages were 93.6% and 0.2% of the accelerating voltage, respectively. Data were
acquired in the positive linear DE mode of operation. Time-to-mass conversion was
achieved by external and/or internal calibration using standards of bradykinin (m/z
1061.2), bovine pancreatic insulin (m/z 5734.5), ), and horse skeletal myoglobin (m/z
16,952) obtained from Sigma Chemical Co. (St. Louis, MO). All experiments were
performed using or-cyano-4-hydroxycinnamic acid (Aldrich Chemical Co., Milwaukee,
WI) as the matrix. Saturated matrix solutions were prepared in a 50% (v/v) solution of
acetonitrile/aqueous 0.1% TFA, and mixed in equal volumes with peptide or protein
samples, and applied to a stainless-steel sample plate. The mixture was allowed to air dry

before being introduced into the mass spectrometer.

 

 

 

 

 

Materials
Recomb
Escherichia Cc
purity of the p
active when
hydrochloride
hydrochloride
citric acid, sor
pyridinium te
further purific
0.1 M citrate
at 20°C fOl‘i

buffer at pH j

Folding EXT

The:
absence of r
0-1 M Citrr
reducing ag
hours, The
the reduced;

The

with 0.05

211

Materials

Recombinant human epidermal growth factor (hEGF) was derived from
Escherichia Cali. Cells and was supplied by Protein Institute Inc., Broomall, PA. The
purity of the protein was greater than 97%. The recombinant hEGF is fully biological
active when compared with standard. Tris(2-carboxyethyl)phosphine (TCEP)
hydrochloride was purchased from Pierce Chemical Co. (Rockford, IL). Guanidine
hydrochloride was a product of Boehringer—Mannheim Biochemicals (Indianapolis, IN).
citric acid, sodium citrate, hydrochloric acid, acetic acid, and 1-cyano-4-dimethylamino-
pyridinium tetrafluoroborate (CDAP) were purchased fiom Sigma and used without

further purification. Acetonitrile and TF A were of HPLC grade. The TCEP solution in

 

0.1 M citrate buffer at pH 3.0 was prepared as 0.10 M stock solution and stored under N2
at -20°C for weeks with little deterioration. The 0.10 M CDAP solution in 0.1 M citrate

buffer at pH 3.0 was freshly prepared prior to use.

Folding Experiments

The folding experiments described here were performed in Tris-HCl buffer in the
absence of redox agents to control the folding. hEGF (1.0 mg) was dissolved in 0.2 ml of
0.1 M Citrate buffer, pH 3.0, containing 6 M of guanidine-HCI and 0.1 M of TCEP
reducing agent. The denaturation and reduction of hEGF was carried out at 37°C for 2
hours. The reaction mixture was separated by HPLC and the fraction corresponding to
the reduced hEGF was collected, recovered, dried and stored at -20°C.

The reduced and denatured proteins were refolded by diluting the protein sample

with 0.05 M Tris-HCl buffer, pH 8.5 to a frnal protein concentration of 1 mg/ml. The

—;

 

 

 

 

 

 

protein was sub.
time course ma
solution contair
to 3 if necessz
temperature fo
fractions were
50 Da, 100 L
disulfide (Ill-

respectively.

Disulfrde Ma
Deper
different strar
29). In addit
indicated the
fT0111 native
disulfide b0
Procedure d
Was estimat
exactly the

Sublected n

 

212

protein was subjected to folding in the open air. Folding intermediates were trapped in a
time course manner by mixing aliquots of the sample (0.1 ml) with 10 ul of 0.5 M HCl
solution containing freshly prepared 0.2 M CDAP. The pH of the solution was adjusted
to 3 if necessary. The cyanylation of free sulfhydryl groups was performed at room
temperature for 15 minutes. The trapped intermediates were separated by HPLC. The
fractions were collected manually and identified by MALDI-TOF MS. Those with 0 Da,
50 Da, 100 Da, 150 Da mass increases from the intact protein correspond to the 3-
disulfide (III- or N), 2-disulfide (II-), l-disulfide (I—), and O-disulfide (R) species,

respectively.

Disulfide Mapping of Purified, Well Populated Folding Intermediates

Depending on the number of disulfide bonds formed in the intermediates,
different strategies were applied to the disulfide mapping of folding intermediates (15, 17,
29). In additional to the 2-disulfide and l-disulfide bond intermediates, MALDI analysis
indicated there are two 3-disulfide bond intermediates with quite different retention time
from native hEGF, indicating they have scrambled disulfide bond structures. The
disulfide bond assignment of these two intermediates was performed according to the
procedure described in chapter 4. Again, the approximate amount of each intermediate
was estimated by the relative peak area in the HPLC chromatogram. The procedure is
exactly the same as we described in chapter 4. The quantity of the intermediates

subjected to the treatment can be roughly estimated from the ratio of a given peak to the

 

 

 

 

 

total peak area
TCEP and CD!
For 2-
intermediate w
approach was
mapping of t]
subjected to pr
4, to detennir
protein, the 1
species and
disulfide map
For t
purified and
adding 2 p1 c
the protein r:
the PCptide c
removed in:
disulfide b0
37°C for 30
diluted Wit]

MALD1.M

 

 

213

total peak areas representing all the protein intermediates. The concentration of the
TCEP and CDAP can thus be calculated to ensure optimal results of the treatment.

For 2-disulfide intermediates, a portion of the purified and cyanylated
intermediate was subjected to the same treatment as the 1-disulfide intermediates. This
approach was applied to define the location of the two free sulfhydryl groups by mass
mapping of the resulting peptides. Another portion of the same intermediate was
subjected to partial reduction/cyanylation/cleavage/mass mapping, as described in chapter
4, to determine the disulfide bond structure. In additional to the original 2-disulfide
protein, the mixture of partially reduced/cyanylated isomers contains two l-disulfide
species and one O—disulfide species (completely reduced/cyanylated cysteine). The
disulfide mapping of the products is described below.

For those containing only one disulfide bond (l-disulfide intermediates), the
purified and cyanylated intermediates were directly subjected to chemical cleavage by
adding 2ul of l M NH4OH aqueous solution containing 6 M guanidine-BC] to dissolve
the protein residue. Then an additional 5 ul of 1 M NH4OH were added, and cleavage of
the peptide chain was performed at room temperature for one hour. Excess ammonia was
removed in a vacuum system. Truncated peptides, which may still be linked by a residual
disulfide bond, were completely reduced by reacting with 2141 of 0.1 M TCEP solution at
37°C for 30 minutes at pH 3-5 to minimize the possibility of reoxidation. Samples were

diluted with 100a] of a 50% (v/v) acetonitrile/0.1% TFA solution prior to analysis by

MALDI-MS.

 

 

 

 

 

HPLC Separa‘
Both tb
trapped intern
elution using V
at 215 nm. T
4.6x250 mm).
to facilitate th
acetonitrile/w
linear in 15 r.
ml/min. The
protein isom:

for further us

111. Results

A. Refoldir

The
the Chang’:
However, i
aChieved b
reduction .

Complete

 

214

HPLC Separation

I Both the purification of the reduced and denatured hEGF and the separation of
trapped intermediates were performed by reversed-phase HPLC with linear gradient
elution using Waters model 6000 pumps controlled by a PC computer. UV detection was
at 215 nm. The column was a Vydac C18 (#218TP54, 10pm particle size, BOO-A pore,
4.6x250 mm). The chromatographic conditions resembled those used by Chang et al (29)
to facilitate the comparison. Solvent A was water containing 0. 1% TF A. Solvent B was
acetonitrile/water (9:1, v/v) containing 0.1% TFA. The gradient was 14—34% solvent B
linear in 15 min, 34-56% solvent B linear from 15 min to 50 min. The flow rate was 1
ml/min. The HPLC fractions were collected manually and the masses of the collected

protein isomers were determined by MALDI-MS. Appropriate fractions were then dried

for further use.

III. Results and Discussion

A. Refolding of hEGF under Controlled Conditions

The refolding conditions used in our experiments were the same as described in
the Chang’s paper (29). The hEGF samples were also obtained from the same source.
However, in Chang’s report, the reduction and denaturation of the native protein was
achieved by adding dithiothreitol (DTT) in pH 8.5 buffer. In our experiments, the
reduction and denaturation of the protein was performed in TCEP pH 3 solution.

Complete reduction is possible under this condition and reoxidation of the reduced

 

 

 

 

species is minir
by passing tho
pH 8.5. The f
the same Tris-
with the same
typically low,
recovered. Fr
experiments, I
the desalting
impurities. 1
oxidation for
The
concentratio:
disulfide cm
as the rate (
Various re
(oxidized/re
t0 control 1
alone the f
Obtained fr.
as Standau

 

215

species is minimized. In Chang’s experiments, the reduced/unfolded hEGF was desalted-
by passing though a PD-lO column (Pharmacia) equilibrated in 0.1 M Tris-HCl buffer,
pH 8.5. The folding of the desalted protein was initiated immediately by diluting with
the same Tris-HCl buffer to a final protein concentration of 1 mg/ml. Our experiments
with the same procedure showed that the recovery of the protein from PD-lO column is
typically low, although the procedure is simple. Only ~60% of the protein was
recovered. Furthermore, the recovered protein has to be subjected to irnmediate refolding
experiments, otherwise, the kinetics of the folding will be inaccurate. In our experiments,
the desalting was done on an HPLC column which also purifies the protein from other
impurities. Aliquots of reduced protein could be stored in dryness without detectable
oxidation for months at -20°C, because the mobile phase (pH 2) stabilizes the protein.
The formation of disulfide bonds is governed by the thiol/disulfide bond
concentration. By varying the concentration of small molecular weight thiol and
disulfide components in solution, the distribution of intermediate disulfide forms as well
as the rate of regeneration of native protein can be controlled in a predictable manner.
Various redox agents, GSH/GSSG (reduced/oxidized glutathione), DT’I‘”‘/D'I"I“"’d
(oxidized/reduced dithiothreitol), and Cys/Cys—Cys (cysteine/cystine) were typically used
to control the folding process, while the controlled folding was performed in Tris-HCI.
alone the folding is catalyzed by oxygen in air. The disulfide intermediates of hEGF
obtained from controlled folding were extensively studied by Chang et al (29) and served
as standards for measuring the efficiencies of EGF folding in the presence of various

redox agents. In our experiments, only controlled folding was performed in order to

 

 

 

 

 

 

 

 

 

compare the re

Sll'llCillIC mappt

B. Trapping:
Conver
alkylation of r
alkaline condi
Recently Tor
iodoacetamid
advocated th
thiol/disulfid
Common me
and the pat}
However, t
iodoacetate .
either of the
Weissman;
and further
anion. 0r.
because ty
Conditions,

The
alkaline c.

216

compare the results and to illustrate the feasibility of our methodology for disulfide

structure mapping.

B. Trapping of Folding Intermediates

Conventional methodology to trap (or quench) the folding process utilizes the
alkylation of sulfhydryl groups by iodoacetate (iodoacetic acid or iodoacetamide) under
alkaline conditions, although the fidelity of iodoacetate trapping was recently questioned.
Recently Torella et a1 (30) found that alkylation by iodoacetic acid, other than
iodoacetamide, results in a side reaction at histidine residues. Scheraga’s group (9, 19)

advocated the use of AEMTS as a trapping agent, in Spite of the potential risk of

 

thiol/disulfide exchange, because of the fast reaction kinetics. Acid quenching is another
common method to stop folding. The intermediates quenched by folding can be isolated
and the pathway of the intermediate can be further studied by “stop—go” experiments.
However, the kinetics and the distribution patterns of intermediates trapped by
iodoacetate and acidic solution are frequently different, implying the possible error during
either of the two trapping methods. The plausible kinetic analysis of acid quenching by
Weissman and Kim (17) indicated that acid quenching can effectively slow down folding
and further thiol/disulfide exchange by. reducing the concentration of reactive thiolate
anion. On the other hand, the results from iodoacetate trapping should be cautious
because the thiol/disulfide exchange can hardly be eliminated under the reaction
conditions.

The cyanylation of sulfhydryl groups by NTCB (chapter 2) was performed under

alkaline conditions. However, CDAP was proved to be a useful alternative for the

—;

 

 

 

cyanylation of
reagent indicat
alkylation, the
under which 1
effectively tra
cyanylated pre
from folding
solution with
experimental
as the lower j
would cataly

monitored b

C. Separat

Fold
were separ:
Strllctural it
Ms- A to
represent}
Species. '
reduced/m

The result:

 

 

217

cyanylation of SH groups under acidic conditions. Our experiments with the CDAP
reagent indicated that, although the kinetics of the cyanylation is not as efficient as that of
alkylation, the cyanylation can be finished at pH 3 within 15 min at room temperature,
under which the thiol/disulfide exchange can be minimized. Not only can the CDAP
effectively trap intermediates in low-pH to minimize the exchange reaction, but also the
cyanylated products can be subjected to disulfide mapping after isolation. The pH change
from folding (pH 8.5) to cyanylation (pH 3.0) was easily accomplished by acidifying the
solution with a small volume of HCl solution containing CDAP, as described in the
experimental section. The pH was important for a successful cyanylation by the CDAP,
as the lower pH (<2) would slow down the reaction greatly, whereas the higher pH (>4.5)
would catalyze the hydrolysis of the CDAP reagent. The final pH of the reaction was

monitored by a pH meter.

C. Separation of Folding Intermediates by HPLC

Folding intermediates of hEGF trapped at 30min, 3h, 24h, and 48h, respectively,
were separated by HPLC (Figure 5.4). In order to interpret these chromatograms,
structural information of the fractionated intermediates was analyzed by MALDI-TOF
MS. A total of 18 peaks were identified at different time points. Peaks 1, 3, and 4
represent 3-disulfide species. Peaks 2, 5, 6, 7, 10, and 12 represent 2-disulfide bond
species. Peaks 13, 14, 16, and 17 are l-disulfide bond species. Peak 18 is the
reduced/unfolded hEGF. [Peaks 8, 9, 11, and 15 contain both 2- and 1-disulfide species.

The results revealed that the intermediates consist of eight l-disulfide isomers and nine 2-

 

 

 

 

30min

 

 

 

 

 

 

 

 

 

 

3h

 

24h

 

48h

 

 

 

Flgure :
differer

 

218

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

% 18
/Il-A /R
30mm l‘A l-B I'C l-D
‘ is ‘Z 16 l
J 56 7 8 910“ 12 1415 17
P
3h
r
4
24h
. . 3
I Ml ‘— J\~_ w 4 __
N
\
48h Ill-B
M
zofoo ' 36.00 1 401.09

minutes

Figure 5.4. HPLC separation of CDAP-trapped folding intermediates of hEGF at
different time courses.

 

 

 

 

disulfide isomer
are some overlaj
It was a]
of each disulfie
folding are ess
disulfide specie
of the species
after 3 hours 0
the 2-disulfrde
almost disappe
species were 6
are similar to
Chang’s pape:
3- or l-disulf
This is not sr
than it should
nOn-native 3
MALDI-MS
cthOmatOgrau
As a
(iodoacetate,
Profiles Shc

intermediate

 

 

219

disulfide isomers. Most l-disulfide species eluted after 2-disulfide species, while there
are some overlaps or reversed order. '

It was apparent that along the folding process, equilibrium existed among isomers
of each disulfide species. The HPLC profiles at 10- (data not shown) and 30-min of
folding are essentially the same although the relative ratio of the most populated 2-
disulfide Species and the reduced/unfolded species is different. At 30 minutes, about 40%
of the species presented as the reduced/unfolded isomer, whereas it almost disappeared
after 3 hours of folding. The 3-disulfide species began to appear as a main peak, while
the 2—disulfide species dominated. After 24 hours of folding, the 2-disulfide species
almost disappeared and the native hEGF dominated. After 48 hours, only the 3-disulfide
species were detected, representing complete folding. The HPLC profiles presented here
are similar to some of Chang’s results (29), however, there are some differences. In
Chang’s paper, thirteen 2—disulfide isomers were identified, some of them overlaped with ,.
3- or 1-disulfide species. We were able to detect nine discernible 2-disulfide isomers.
This is not surprising as the sensitivity of our detection system was substantially lower
than it should be. Unlike Chang’s chromatogram in which the two peaks representing the
non-native 3-disulfide isomers contain a considerable amount of 2-disulfide species,
MALDI-MS analysis showed that the corresponding peaks (3 and 4) in our
chromatograms represented pure isomers.

As a comparison, the HPLC profiles obtained by different trapping techniques
(iodoacetate, acid, or CDAP) at different time points are shown in Figure 5.5. The three
profiles show almost the same pattern in terms of the appearance of disulfide

intermediates. It should be noted however that, in Chang’s experiments, the HPLC

 

 

 

Flt- .

wommmuwrmaflflmv vommmbrgoxx
oumwoomowo

 

 

220

 

mom: a essences maze Baggage
Ea .vommwb.Eom 6335-880382 mo 8an Una: we nemfimmaov .m.m oSmE

 

 

 

 

 

 

 

 

 

 

 

 

. . m
Z a
Z Z
d
. = . _
J52}? 133.431 ti \
. seem
: = we
a _
momma?
commebiado eoammbéo< 380882

 

 

 

 

profiles of acid-
counterparts. l
isomers trapper
from the under.
or less. Howe
disulfide inter
disulfide inten
a1 (29) assum
means that all
the hydrophol
farfetched.
cYanylation d
SCN groups:
acid and CD [A
intermediates
intermediates
populated 2.
peak could r
characterizat
our results
imerthedlate

measuremer

 

 

 

221

profiles of acid—trapped intermediates did not fully resemble those of iodoacetate-trapped
counterparts. It is understandable in terms of the difference in the hydrophobicity of
isomers trapped by different manners. The hydrophobicity of alkylated protein differs
from the underivatized isomers. Therefore, the relative retention time could change more
or less. However, it is beyond understanding that the majority of iodoacetate-trapped 2-
disulfide intermediates were eluted within three fractions, whereas acid-trapped 2-
disulfide intermediate(s) were accumulated within one fraction (marked as 11). Chang et
al (29) assumed all the three fractions coeluted in acid-trapped HPLC profile, which
means that all the three isomers had the same hydrophobicity before modification, while
the hydrophobicities were substantially different after alkylation. This eXplanation is
farfetched. Fortunately, the hydrophobicity of hEGF isomers before and after
cyanylation does not change greatly (the polarity of SH and SCN is similar while SH and
SCN groups are both small species), direct comparison of HPLC profiles obtained from
acid and CDAP trappings is reasonable. It is noteworthy that the HPLC chromatogram of
intermediates trapped by the CDAP technique is similar to that representing the
intermediates trapped with acid. In both cases, only one peak corresponding to the most
populated 2-disulfide intermediate(s) was obtained. To exclude the possibility that the
peak could contain more than one 2-disulfide species, as suggested by Chang (29), the
characterization of the disulfide structure of the species was carried out (see below) and
our results indicated unambiguously that the peak contains ONLY one 2-disulfide
intermediate. This result confirms that acid trapping provides a more accurate

measurement of folding process. The two major isomers observed by Chang in the

 

 

 

 

chromatogram
caused by alkj
It was
of each disu
intermediates
fractions for
behaved simf
previous rep
characterized
equilibrated
CXperiments,
(Figure 5.4,
convert the ;
alkaline me.
Tyr, and a r

thiols to cat

D. Disulfir

Tran
one is “Pei
Alternative

a COlOI' 01-

 

 

222

chromatogram representing iodoacetate trapping could be from thiol/disulfide exchange
caused by alkylation under alkaline condition.

It was apparent that along the folding process, equilibrium existed among isomers
of each disulfide species. For instance, the concentration of 1- and 2-disulfide
intermediates ascended and then descended as folding progressed, but the relative ratio of
fractions for each kind of isomers remain constant. Scrambled 3-disulfide species
behaved similarly during the folding. These data supported the conclusion drawn by
previous reports (29) that, like other proteins, the folding pathway of the hEGF was
characterized by a sequential flow of unfolded EGF (R) through three groups of
equilibrated intermediates, namely, 1-, 2-, and 3-disulfide (scrambled) isomers. In our
experiments, about 40% of the folding intermediates remained as scrambled species
(Figure 5.4, 48-h sample). The exposure to air for another 48 hours cannot automatically
convert the scrambled species to the native form (furthermore, long hours of exposure to
alkaline medium results in minor oxidation of some amino acids such as Met, His and
Tyr, and a mass shift from the intact protein would be observed), due to the lack of free

thiols to catalyze the disulfide reshuffling.

D. Disulfide Mapping of Well Populated Intermediates

Traditionally, this can be achieved by a number of strategies. The most common
one is “peptide mapping” by isolating and analyzing every enzyme-fragmented peptide.
Alternatively, it can be done by selective labeling of disulfide bonds after reduction with

a color or fluorescent thiol-specific reagents (17, 31). Both methods need microgram

 

 

 

 

 

 

amounts of inte
and numerous a
As we 6

of cyanylated
sulfhydryl gror
structure deter
defrning disull
chemistry beh
intermediates I
successfully cl
3-Disulfrde Ir
Afterz
Mass analysis
peak showed
confirmecl th;
were two nor
collected ant
Strategy desc
33 Straightfo
isomers We“

CDAP Was t

Table

DOSSlblC CYa

 

223

amounts of intermediates, multiple enzymatic digestions, HPLC separation of peptides,
and numerous attempts of sequence analysis.

As we demonstrated in chapters 2-4, the mass mapping of the cleavage products
of cyanylated proteins provides specific information on the number and location of
sulflaydryl groups and disulfide bond pairsf This strategy is also applied to the disulfide
structure determination of the folding intermediates of hEGF. The procedures for
defining disulfide structure in 1-, 2- and 3-disulfide species differ slightly, but the
chemistry behind the methodology is exactly same. A total of seven well populated
intermediates (marked with III-A, III-B, II-A, I-A, I-B, LC, and I-D in Figure 5.4) were
successfully characterized by our strategy.
3-Disulfide Intermediates

After 48-h folding (Figure 5.4, 48-h), three well populated species were separated.
Mass analysis by MADLI indicated all three species are 3-disulfide isomers. The first
peak showed the identical retention time as the native hEGF. . Disulfide mapping
confirmed that the isomer is the native hEGF (data not shown). The latter two peaks
were two non-native isomers with scrambled disulfide structures. The two peaks were
collected and subjected to the partial reduction/cyanylation/cleavage/mass mapping, a
strategy described in chapter 4. The control of partial reduction is important but is never
as straightforward as the protocols used in chapter 4. The molar amounts of protein
isomers were estimated from the HPLC peak area and the concentration of the TCEP and
CDAP was thus calculated.

Table 5.1 shows the expected nr/z of fragments resulting from cleavage of 15

possible cyanylated disulfide isomers. Figure 5.6 shows the HPLC separation of partially

 

 

 

 

 

Table 5.1. 3
of 15 singly

 

Reductie
of Disuli

6-14

6-21

14

1.

224

Table 5.1. Expected m/z values for possible fragments resulting fiom the cleavage
of 15 singly reduced/cyanylated hEGF isomers

 

 

Reduction Reduction
F t m/ F t m/
of Disulfide ragmen Z 0 f D1S 111 fi d6 ragmen Z

1.5 551.5 1-13 1423.4
6-14 6-13 915.9 14-42 1441 3160.6
1453 4837.6 4253 1721.0
1.5 551.5 1-19 2047.1
6-20 619 1539.6 2031 2030 1344.6
20-53 4214.0 3163 2913.3
1-5 551.5 149 2047.1
6-31 6-30 2840.2 2033 2032 1560.8
31-53 2913.3 33-53 2697.2
15 5515 1-19 2047.1
633 6-32 3056.4 2042 2041 2536.9
3353 2697.2 4253 . 1721.0
1.5 551.5 130 3347.7
642 6-41 4031.5 31-33 31-32 261.3
4253 1721.0 33-53 2697.2
143 1423.4 130 3347.7
1420 14-19 667.7 3142 3141 1236.3
20-53 2913.3 42-53 1721.0
1—13 1423.4 1-32 3563.9
1431 1430 1968.3 3342 33-41 1020.1
31-53 2913.3 4253 1721.0

1.13 1423.4

14-33 1432 2184.5

3353 2697.2

 

 

 

 

10.00

Figure 5.6

225

 

Ill-A

 

 

 

 

 

 

 

 

 

 

 

 

 

W)

10.00 ’— 20.‘oo 7 30.30 T who T 5000
minutes

 

 

 

Figure 5.6. HPLC separation of III—A and its partially reduced/cyanylated species.

 

 

 

 

 

++

++ +1

 

 

33-4l

 

 

 

1000

Figure 5.
0f the th]
COHespO.
and e re]

 

 

226

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

42-53
A
6-41'
++
l [A l 141*
42-53 33-53
B
1
1-13
++ 14—32
H \l 14-53* I
WW. .5...
42-53 31.-53
C
33-41
*
m
V:
S; 1.19 8
W
. f_ .4 f4; W A
1000 2000 3000 4000 5000 6000

m/z

Figure 5.7. The MALDI mass spectra of peptide mixtures resulting from the cleavage
of the three singly reduced/cyanylated species of non-native hEGF III-A, A-C
corresponding to the HPLC peaks 1-3 in Figure 5-6, respectively. The symbols ++
and * represent the doubly-charged species and B-elirnination products, respectively.

 

 

reduced/cyany]
1-3 were singlj
reduced EGF
distribution of
disulfide bond
completely rec
products corre
spectrum (Fig
fragments 6-t
fragment 1-5
Cys6 is conne
From
and 33-53 we
S-TB). An 0
Therefore, C
haglnent 42.
degree of c
Cyanylation,
intensity 0ft
the asstghme
Although fr:

MC peak

 

227

reduced/cyanylated protein isomer III-A. Mass analysis of the peaks showed that peaks
1-3 were singly reduced/cyanylated species, the peak eluted at 42.5 min was a completely
reduced EGF which should be typically present as a small peak. This atypical
distribution of the partially reduced species suggests that the opening of the first two
disulfide bonds facilitate the reduction of the last one, resulting in the accumulation of
completely reduced species. Figure 5.7 shows the MALDI mass mapping of the cleavage
products corresponding to HPLC peaks 1-3 in Figure 5.6, respectively. The MALDI
spectrum (Figure 5.7A) of the cleavage products of HPLC peak 1 in Figure 5.6 shows
fragments 6-41, 42-53, and an overlapped B-elimination product 1-41. The small
fragment 1-5 was missing. It is clear that cleavage occurs at Cys6 and Cys42. Therefore,
Cys6 is connected to Cys42 as a disulfide bond.

From the cleavage products of HPLC peak 2 in Figure 5.7, fragments 1-13, 14-32,
and 33-53 were detected, indicating peptide chain cleavage at Cysl4 and Cys33 (Figure
5.7B). An overlapped B-elimination product, 14-53, corroborates with the conclusion.
Therefore, Cysl4 is linked to Cys42. However, this assignment was questioned by the
fragment 42-53, indicating that Cys42 also underwent cleavage. Apparently a minor
degree of cleavage must have occured at Cys42, either by the disulfide split and
cyanylation, or by some disulfide exchange during cleavage process, or both. The peak
intensity of this fragment is so strong that further experiments were performed to confirm
the assigmnent.

Although fragment 42-53 is indicated in the MALDI spectra of cleavage products from

HPLC peaks 1-3, the absolute concentration of the fragment is quite different in the

 

 

 

 

 

 

 

 

 

 

ro.oo\

Figu]
redue
2 in]

228

 

fragment
42-53 A

 

 

 

 

 

 

 

 

 

 

42-53 . L.

 

 

 

10.00 ' 20Too T 30:00 ' 4600 ’ 50T00
minutes

Figure 5.8. Comparison of HPLC profiles of cleavage products of two singly
reduced/cyanylated III-A isomers. (A) and (B) are represented by peaks 1 and
2 in Figure 5-6, respectively.

 

 

 

three cases. Tl
the same HPLt
that the fractio.
fragment repre
while the fragr
so small corr
fragment inder
fact, the peak
mixture of cle
1% contamin.
the data inter]
The n
the fragment
pair (Cys20-
Again, the M
The 1
Salne strateg;
Species, Pe:
anEllysis by
Wmmfi

elimination -

 

 

229

three cases. The cleavage products from HPLC peaks 1 and 2 were re-fractionated under
the same HPLC conditions (Figure 5.8). The MALDI analysis of the fractions indicated
that the fraction eluting at 17.5 min corresponded to fragment 42—53. It was clear that this
fragment represented one of the major cleavage products in HPLC peak 1 (Figure 5.7A),
while the fragment 42-53 in the mixture of the cleavage products from HPLC peak 2 was
so small compared to other fragment(s) that it could be neglected. However, this
fragment indeed plagued every MALDI spectrum due to its high response. As a matter of
fact, the peak corresponding to fragment 42-53 was still intense on MALDI even if the
mixture of cleavage products from HPLC peak 1 was diluted by 100-fold. That means
1% contamination by fragment 42—53 could generate a false information and jeopardize
the data interpretation.

The mixture of the cleavage products from HPLC peak 3 showed the presence of
the fragments 1-19, 20-30, 31-53 (Figure 5.7C), indicating the linkage of last disulfide
pair (CysZO-Cys31) which is already deducible after the first two pairs were defined.
Again, the MALDI spectrum is also plagued by a trace of the undesired fragment 42-53.

The disulfide structure in hEGF isomer III-B was subjected to analysis by the
same strategy. Figure 5 .9 shows the HPLC separation of the partially reduced/cyanylated
Species. Peaks 1-3 are three singly reduced/cyanylated isomers, as confirmed by mass
analysis by MALDI. The MALDI analysis of cleavage products from these peaks are
Shown in Figure 5.10A-C. Figure 5-10A shows the fragments 6-13, 14-53, and a [3-

elimination product 1-13. It is clear that Cys6 is hooked to Cysl4. A small peak

 

 

 

 

roho‘T

Figure

 

 

230

 

 

 

 

 

 

 

 

 

 

III- B
1
2
10.00 ‘ 20.00 ’ 30T00 T 40.00 T 5000
minutes

Figure 5.9. HPLC separation of III-B and its partially reduced/cyanylated species.

 

 

6-13

1’1‘:

 

 

33-41

31-4'

--A

 

 

 

++ h
1000 —
Figure 5.

3 sing1y 1
to the HI

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6-13
A
14-53
*42-53
52
'I" ++
.. l .__._.22 - 2-
42-53
B
33-41
33-53*
-. L'i’- 1311*
31-41 42’53 31_53
C
1-19
0 «-
C? m
E?» ”9 4
++ I g; 1-30
1000 2000 3000 4000 5000 6000

m/z

Figure 5.10. The MALDI mass Spectra of peptide mixtures from the cleavage of the
3 singly reduced/cyanylated species of non-native hEGF III-B, A-C corresponding
to the HPLC peaks 1-3 in Figure 5 .9, respectively.

 

 

 

 

corresponding
5.108) of the
fragments 1-2
elimination p:
conclusively :
therefore, car
pair, Cy320-(
fragments in
and a B-elimi
the assignme
the MALDI-
31-53 can 1
fractionation

HPLC/RSI e

2‘DtSlllfrde

In or
he disulfid.
55% and >
HPLC isola
11 Shows fl

the 0figinal

 

232

corresponding to the fragment 42-53 can be ignored. The MALDI spectrum (Figure
5.10B) of the cleavage products corresponding to HPLC peak 2 in Figure 5.9 shows
fragments 1-32, 33, 41, 42-53, and two overlapped peptides corresponding to the B-
elimination products 1-41 and 33-53, respectively. Disulfide bond linkage 33-42 can be
conclusively assigned. In this spectrum, the fragment 42-53 presents as a huge peak and,
therefore, cannot be ignored. From the disulfide structure of the first two pairs, the third
pair, CySZO-Cys31, can be deduced, which is confirmed by the analysis by MALDI of the
fragments in Figure 5.10C. In addition to the expected fragments, l-19, 20-30, 31-53,
and a B-elimination product 1-30, two undesired fragments, 31-41 and 42—53, can confuse
the assignment, as these fragments indicate that Cys42 is also a cleavage site. Frankly,
the MALDI is limited in this case for an unambiguous assignment, although the fragment
31-53 can be used as evidence that Cys42 is a minor cleavage site. The HPLC
fractionation of the cleavage products combined with the MALDI identification or

HPLC/ESI experiments should be informative to clarify the uncertainty.

2-Disulfide Intermediates

In our folding experiment, only one 2-S-S intermediate (II-A) was collected and
the disulfide bond structure characterized. This most abundant intermediate counts for
55% and >70% of the total population at 30-min and 3-h folding, respectively. The
HPLC isolated, cyanylated II-A was subjected to partial reduction/cyanylation. Figure 5-
11 shows the chromatogram of the partially reduced/cyanylated products. In addition to

the original 2-disulfide species, two l-disulfide and one O-disulfide species arise from the

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

as?

Figure 5.1

 

 

 

 

233

 

 

 

 

 

 

 

 

 

 

 

 

 

ll-A
R .
1
2
10.00 ' 20.00 ' 30:00 T 45.00 ' 50.00 I
minutes ‘

Figure 5.11. HPLC separation of II-A and its partially reduced/cyanylated species.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

partial reducti
disulfide oper
disulfides ope
It shoe

more than on
of the partial]
pure intermee
one species,
complicated '
Althc
products of '
easier to 100:
in Il-A Will ;
Preducts rel:
Spectra. Th,
Peak HA 3]
Cleavage oc
feSidUes. T
whatsoever
The

2 in Figure

indicatingt‘

 

 

234

 

partial reduction and cyanylation. That is, the HPLC peak 1 (Figure 5.11) has one
disulfide opened, peaks 2 and 3 have two disulfides opened, and peak 4 has all three
disulfides opened, resulting in a familiar partial reduction pattern.

It should be pointed out that the HPLC fraction II-A in Figure 5.4 could contain
more than one unresolved species, as suggested by Chang (29). From the HPLC pattern
of the partially reduced protein isomers, it is clear that the original speak II-A represents a
pure intermediate, other species, if any, are very minor. If the fraction contains more than
one species, the partially reduced products of those species Should be much more
complicated than the present pattern. Obviously it is not the case.

Although the disulfide structure of the II-A can be directly mapped from cleavage

 

products of the species represented by HPLC peaks 2 and 3 in Figure 5.11, it is much
easier to locate first the cyanylated cysteine residues. The localization of the SH groups
in II-A will greatly help the subsequent identification of the MALDI data of the cleavage
products related to the reduced disulfide bonds, as the latter usually give more complex
spectra. The MALDI spectrum (Figure 5.12A) of the cleavage products corresponding to
peak II-A shows fragments 6-19, 20—53, and B-elimination product 6-53, indicating that
cleavage occurred at cysteine residues 6 and 20, which represent two unoxidized SH
residues. The peak 1-53 represents an intact II-A that did not undergo any cleavage
whatsoever under our eXperimental conditions.

The MALDI Spectrum (Figure 5.12B) of the cleavage products from HPLC peak
2 in Figure 5.11 shows fragments 6-19, 33-41, 42-53 and a B-elimination product 33-53,

indicating that another disulfide pair, 33—42, must have been reduced, cyanylated and

 

 

 

 

 

 

235

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

A
6-19
1-53*
y2-53 6-53*
++++ 20-53
3341 42-53
B
0.9
'7 ' 33-53*
++ so
ML. ..._ idem/n .llLMA. .. .3”
31-53
6-13 C
42-53
1000 2000 3000 4000 5000 6000

m/z

Figure 5.12. The MALDI mass spectra of peptide mixtures from the cleavage of (A)
the cyanylated hEGF II-A, (B) and (C) the partially reduced/cyanylated species of
II-A, corresponding to the HPLC peaks 1 and 2 in Figure 5.11, respectively.

 

 

236

cleaved. Therefore, Cys33 is connected to Cys42. Likewise, the MALDI spectrum
(Figure 5.12C) of the cleavage products from HPLC peak 3 in Figure 5.11 shows
fragments 6-13, 20-30, 14-30, and 31-53, indicating cleavage at cysteine residues 6, 14,
20, and 31, respectively. There is a minor degree of undesired cleavage at Cys42,
resulting in the formation of fragments 31-41 and 42-5 3.

In summary, intermediate II-A, the most populated species during the early stage

of folding, contains a native structure, Cysl4-Cys31 and Cys33-Cys42.

l-Disulfide Intermediates

The recognition of disulfide structures in 1-disulfide intermediates ' is
straightforward. The strategy is based on the mass mapping of cleavage products of the
cyanylated 1-disulfide species. Instead of looking at the MALDI peaks corresponding to
the cleavage Sites, the absence of such peaks represents the “uncleaved sites”, that is,
those still linked by a disulfide bond. For a protein containing 6 cysteine residues, there
are fifteen possibly l-disulfide intermediates. The possible fragments corresponding to
cleavage at cyanylated cysteine residues of the 15 isomers are listed in Table 5.2. Since
some of the fragments may not be detected by MALDI, the mass mapping of the products
from the completely reduced/cyanylated hEGF was used as a “figureprint” spectrum of
the cleavage products. It is immediately clear that the fragment 1-5 cannot be detected by
MALDI, while fragment 42-53 is very sensitive to MALDI detection. This profile helps
the interpretation of MALDI data from other intermediates, especially if ambiguity arises.

The MALDI spectra of the cleavage products of the four well populated 1-

disulfide intermediates are shown in Figure 5.13. Figure 5.13A is the MALDI spectrum

 

237

Table 5.2. Fragments of cleavage products corresponding to 15 cyanylated l-disulfide
intermediates of hEGF

 

Disulfide Disulfide

 

Z
Linkage Fragment m/ Linkage Fragment m/z
149 2047.1 1-5 551.5
2030 1344.6 6-19 1539.6
6-14 31-32 262.3 1442 2030 1344.6
3341 1020.1 3132 262.3
4253 1721.0 3363 2697.2
143 1423.4 1.5 551.5
1430 1968.3 6-13 915.9
6-20 31-32 262.3 2031 14-32 2184.5
3341 1020.1 3341 1020.1
4253 1721.0 4253 1721.0
1-13 1423.4 1.5 551.5
1449 667.7 6-13 915.9
6-31 2032 1560.8 2033 14-30 - 1344.6
3341 1020.1 3141 1236.3
42-53 1721.0 4253 1721.0
143 1423.4 14 551.5
14.19 667.7 6-13 915.9
6-33 2030 1344.6 2042 1430 1968.3
3141 1236.3 31-32 262.3
4252 1721.0 3363 2697.3
143 1423.4 16 551.5
1419 667.7 613 915.9
642 2030 1344.6 3133 14-19 667.7
3132 262.3 2041 2536.9
33-53 2697.2 4253 1721.0
1.5 551.5 1.5 551.5
630 2840.2 6-13 915.9
1420 31-32 262.3 3142 14-19 667.7
3341 1020.1 2032 1560.8
4253 1721.0 3353 2697.2
1-5 551.5 16 551.5
6-19 1539.6 613 915.9
1431 2032 1560.8 3342 14-19 667.7
3341 1020.1 2030 1344.6
4253 1721.0 3163 2913.3
16 551.5
619 1539.6
1433 2030 1344.6
3141 1236.3
4253 1721.0

 

 

238

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

42-53
3341
LA
++ 1'13 14-30
\ I l 3353*
3353
LR
_‘ 4253
ms
ER ++ 14-30 31-53
42-53
LC
3341
is a.
«a a
14 - “2.3:..JL- - _+__A- *
3341 42-53
I-D
++
‘I ll 20-30 “'19 33-53*
1000 2000 3000 4000 5000 6000
m/z

Figure 5.13. Disulfide mapping of l-disulfide folding intermediates of hEGF.

 

 

 

 

 

 

239

of cleavage products corresponding to HPLC I-A in-Figure 5.4. Fragments 1-13, 14-30,
33-41 and 42-53 were detected, implying chemical cleavage at a cyanylated cysteine at
position 14, 31, 33, and 42, respectively. Obviously, no cleavage occurs at Cys6 and
Cys20. Otherwise, fragments 1-5, 6-13, 14-19, and 20-30 would arise, giving m/z at
551.5, 915.9, 667.7, and 1344.6, reSpectively. Therefore, Cys6-Cys20 linkage is
deduced. It is notable that this intermediate has a native disulfide structure, which was
not detected in Chang’s experiments (29). The second l-disulfide intermediate (I-B)
Showed MALDI peaks corresponding to fragments 6-13, 14-30, 33-53, and 42-53,
respectively. Cleavage occurs at Cys6, 14, 31, 33, and 42, but not Cy820. Extensive
experiments indicated that the cleavage at Cys42 was very minor in comparison with the
cleavage at other sites. Since fragment 33-53 is very intense, a reasonable conclusion can
still be drawn that Cys42 does not present as a free sulfhydryl. Therefore, Cys20 is
connected to Cys42. The MALDI spectrum for I-C (Figure 5.13C) is Simple to interpret.
Fragments 6-30, 33-41, and 42-53 show that Cys6, 31, 33, 42 are cleavage sites.
Therefore, the disulfide pair 14-20 can be deduced. By the same strategy, the MALDI
spectrum (Figure 5.13D) of I-D shows peaks for the fragments 1-19, 20-30, 33-42, and
42-53, suggesting the Cys6-Cysl4 linkage. The disulfide structures of the last two
intermediates are in agreement with the structures proposed by Chang et al (29). By our
strategy, we were able to recognize the disulfide structure in LA and LB, neither of which
was assigned by Chang et al. Among the four well populated l-disulfide intermediates,

only one has a native disulfide structure.

 

 

 

240

IV. Aspects for Further Improvement

One of the most important features of our methodology is the simplicity. Only a
few fragments were produced, each of them is relevant tolthe assignment of (disulfide
structures. Some undesired fragments may be observed, but they rarely affect the
assignment of disulfide bonds because our methodology only relies on the recognition of
specific cleavage at cysteine residues. However, if the side products correspond to the
cleavage at undesired cysteine site(s), the data interpretation would no longer be
straightforward. This Situation is more complicated by the fact that the MALDI response
is not proportional to concentration or amount of analytes. Even a trace amount of an
impurity may give a high MALDI response, as seen by the fragment 42-53 in the case of
hEGF. Fortunately, HPLC can provide quantitative information on the relative
concentration of analytes, if ambiguity arises. The most straightforward way is to
perform HPLC/ESI analysis of cleavage products so that information on both quantities
and masses of the analytes can be obtained.

Our CDAP trapping provides advantages over both iodoacetate trapping and
acidic quenching. Effective trapping can be achieved in most cases as demonstrated by
the MALDI analysis of the trapped intermediates. However, the CDAP concentration
and the pH must be controlled because of the possible side reactions (see chapter 3) and
the instability of CDAP at higher pH (hydrolysis). As the available SH groups vary in the
folding course and the concentrations of various intermediates differ greatly, the optimal
CDAP concentration is difficult to control. It should be a good investment to study more
stable and specific novel reagents that can cyanylate SH groups in acidic solution.

ThiocyanOpyridine (TCP) (32) (see chapter 3) may act as such a reagent. Another

 

 

 

 

241

 

 

. ”am: we 824608.485 mango.“ Banana :95 :33 mo $525.6. 03:35 .36 8,5me

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

242

reagent, 4-thiocyanatoaniline (TCNA), was introduced in Biemann’s group several years
ago (33). This reagent was claimed to be a usefiil alternative to NTCB. But the

capability of this reagent to modify SH groups in acidic solution remains to be examined.

V. Conclusions

To summarize, seven species of well populated intermediates in the refolding of
hEGF have been isolated and characterized, which included four l-disulfide, One 2-
disulfide, and two 3—disulfide scrambled species (Figure 5.14), of which only two have
native disulfide structure. Among the seven intermediates, I-A and LB are newly
identified; the disulfide structures of the other five intermediates are the same as these
published previously (29). These results demonstrate the feasibility of our methodology.
The significance of these intermediates to the folding pathway of hEGF needs to be
further specified.

Our trapping technique, based on the cyanylation of free thiol groups under acidic
conditions, showed similar results to those of acid trapping in terms of the distribution of
intermediates during the folding process. This technique circumvents problems
associated with the traditional iodoacetate trapping and greatly facilitates the subsequent
characterization of disulfide structures of isolated folding intermediates. Our strategy for
the determination of disulfide structures by chemical cleavage and mass mapping of the
fragments is much simpler, faster, and conclusive, although the undesired cleavage at
Cys42 indeed occurred in some cases.

This methodology opened a new door to the application of our chemistry. The

preliminary experimental results presented in chapters 4 and 5 also show the potential to

 

 

 

T————_————”“_* --

243

characterize disulfide structures of folding intermediates of even more complicated

proteins, such as those containing adjacent cysteine residues (e. g., IGFs).

 

 

 

 

244

VI. References

1. Ellis, R. J., Curr. Opin. Struc. Biol., 4, 117-122(1994).

2. Frydman, R. B., Nirnmesgern, E., Ohtsuka, K., and Hartl, F. U., Nature, Land. 370,
111-117(1994).

3. Kim, P. S., and Baldwin, R. L., Annu. Rev. Biochem., 59, 631(1990).

4. Matthews, C. R., Annu. Rev. Biochem., 62, 653(1993).

5. Ptitsyn, O. B., Curr. Opin. Struc. Biol.,.5, 74( 1995).

6. Li, Y.-J., Rothwarf, D. M., and Scheraga, H. A., Nat. Struct. Biol., 2, 489(1995).

7. Jennings, P. A., and Wright, P. E., Science, 262, 892-896( 1993).

8. Balbach, J ., Forge, V., van Nuland, N. A. J., Winder, S. L., Hore, P. J ., and Dobson,
C. M., Nat. Struct. Biol., 2, 865-870(1995).

9. Xu, X., Rothwarf, D. M., and Scheraga, H. A., Biochemistry, 35, 6406-6417(1996).

10. Frech, C., and Schmid, F. X., J. Mol. Biol., 251, 135-149(1995).

11. Creighton, T. E., and Ewbank, J. J., Biochemistry, 33, 1534-1538(1994).

12. Chang, J .-Y., Biochemistry, 35, 11702-11709(1996).

13. Glocker, M. O., Arbogast, B., Milley, R., Cowgill, C., and Deinzer, M. L., Proc. Natl.
Acad. Sci., USA, 91, 5868-5872(1994).

14. Hober, S., Forsberg, G., Palm, G., Hartmanis, M., and Nilsson, B., Biochemistry, 31,
1749-1756(1992).

15. Creighton, T. E., and Goldenberg, D. P., J. Mol. Biol., 179, 497( 1984).

16. Darby, N. J., Norin, P. E., Talbo, G., and Creighton, T. E., J. Mol. Biol., 249,

463(1995)

 

 

 

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

Chang, J .Y., Biochemistry, 35, 11702-11709(1996).

 

245

Weissman, J. S., and Kim, P. S., Science, 253, 1386(1991).

Weissman, J. S., and Kim, P. S., Proc. Natl. Acad. Sci. USA, 89, 9900(1992).
Rothwarf, D. M., and Scheraga, H. A., Biochemistry, 32, 2671(1993).
Rothwarf, D. M., and Scheraga, H. A., Biochemistry, 32, 2680(1993).
Rothwarf, D. M., and Scheraga, H. A., Biochemistry, 32, 2690(1993).

Rothwarf, D. M., and Scheraga, H. A., Biochemistry, 32, 2698( 1993).

 

Creighton, T. 13., Science, 256, 111-112(1992).
Weissman, J. S., and Kim, P. S., Cell, 71, 841-851(1992).
Rothwarf, D. M., and Scheraga, H. A., J. Am. Chem. Soc., 113, 6293(1991).

Goto, Y., and Hamaguchi, J. Mol. Biol., 146, 321(1981).

 

Happersberger, P., Cowgill, C., Deinzer, M. L., and Glocker, M. 0., Proceedings of
the 44th ASMS Conference on Mass Spectrometry and Allied Topics, Portland,
Oregon, May, 1996, p1050.

Chang, J.-Y., Schindler, P., Ramseier, U., and Lai, P.—H., J. Biol. Chem, 270(16),
9207-9216(1995).

Torella, C., Ruoppolo, M., Marino, G., and Pucci, F., FEBS Letters, 352, 301-
306(1994).

Chang, J .-Y., J. Biol. Chem, 268, 4043-4049(1993).

Brocklehurst, K., Malthouse, J. P. G., Baines, B. S., Blenkinsop, 1R. D., Churcher, J.
A., Mushiri, M. S., and Ormerod, F., Biochem. Soc. Trans, Vol. 6, 261-263(1978).
Zaia, J ., Papayannopoulos, I. A., and Biemann, K., Proceedings of the 39th ASMS

Conference on Mass Spectrometry and Allied Topics, Nashville, TN, May 1991.

'0 HICHIGRN STAT

ruiyijlryiyhijl“

ht...