PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 c!CIHC/DateDue.p65—p.15

 

FURTHER DEVELOPMENT, APPLICATIONS, AND DATA PROCESSING
ALGORITHMS FOR DISULFIDE MASS MAPPING OF PROTEINS BASED ON
PARTIAL REDUCTION, CYANYLATION, AND CN-INDUCED CLEAVAGE
By

J ianfeng Qi

A DISSERTATION
Submitted to
Michigan State University
in partial fulﬁllment of the requirements

for the degree of

DOCTOR OF PHILOSOPHY
Department of Chemistry

2001

UMI Number: 3036736

®

UMI

 

UMI Microform 3036736

Copyright 2002 by ProQuest Information and Learning Company.

All rights reserved. This microform edition is protected against
unauthorized copying under Title 17. United States Code.

 

ProQuest Information and Learning Company
300 North Zeeb Road
PO. Box 1346
Ann Arbor, MI 48106-1346

ABSTRACT
FURTHER DEVELOPMENT, APPLICATIONS, AND DATA PROCESSING
ALGORITHMS FOR DISULFIDE MASS MAPPING OF PROTEINS BASED ON
PARTIAL REDUCTION, CYANYLATION, AND CN-INDUCED CLEAVAGE
By

J ianfeng Qi

The location of the free cysteines in a protein can be determined by
cyanylation/CN—induced cleavage chemistry with subsequent analysis by mass
spectrometry. The free cysteines can be cyanylated and the protein backbone can be
cleaved in aqueous ammonia at the N-terrninal side of the cyanylated cysteines. The
mass spectral data of the resulting mixture of cleavage fragments can be mass mapped to
the sequence of the protein by comparing the experimental mass of the cleavage
fragments with expected mass. The location of the free cysteines can then be deduced.

Partial reduction of a multi-disulﬁde protein with tris-carboxyethyl-phosphine
(TCEP) hydrochloride under kinetically limiting conditions can produce a mixture of
singly reduced isoforrns of the protein. The location of the two nascent free cysteines in
a singly reduced isoforrn, resulting ﬁom reduction of a disulﬁde bond, can be determined
by mass mapping the cleavage fragments; thus, the connectivity of the disulﬁde bonds
can be deduced. Various aspects of this methodology were further investigated.
Utilization of doubly reduced isoforrns offers ﬂexibility in the chemistry of partial
reduction and provides complementary data to the previous partial reduction protocol,

which utilizes only singly reduced isoforrns. An additional analysis step by mass

spectrometry aﬂer the CN-induced cleavage reaction of partially reduced/cyanylated
protein isoforms, but before complete reduction of the residual disulﬁde bonds, may
provide important diagnostic information. When CN-induced cleavage reactions are
conducted in aqueous ammonia, the resulting CN-induced fragments may have an amide
or a free acid C-terminus; thus, the mass of each form must be computed for use in data
interpretation. In cases of isomass CN-induced fragments, sequence analysis of such
fragments by low-energy collision induced dissociation (CID) -tandem mass
spectrometry (MS/MS) can be used to distinguish the isomeric species.

The cysteine-rich and apparently highly knotted anti-microbial peptide sillucin,
which is resistant to the proteolytic mapping approach to disulﬁde mapping, is amenable
to partial reduction/cyanylation/CN-induced cleavage mass mapping methodology. One
of four possible singly reduced isoforms and one of six possible doubly reduced isoforrns
of sillucin generated sufﬁcient information from mass mapping of the cyanylation/CN-
induced cleavage products to assign the disulﬁde connectivity (Cys2-Cys7, Cyle-
Cy824, Cysl3-Cy830, and Cysl4-Cys21) and rule out the other 104 isomeric possibilities.

The cysteine status of the chloroplast glyceraldehyde-3-P dehydrogenase subunits
was determined by cyanylation, CN-induced cleavage, and mass mapping.

Algorithms for data processing for the disulﬁde mass mapping based on
cyanylation and CN-induced cleavages were proposed. The concept of using negative
signature masses (masses that can be used to reject certain disulﬁde linkages if present)
facilitates the disulﬁde assignment process and, in some cases, allows identiﬁcation of

the disulﬁde structure even when an incomplete data set is obtained.

To my wife, Tong Pei, for her love, support, and patience that makes my life enjoyable

iv

ACKNOWLEDGEMENTS

I would like to express my sincere gratitude to my mentor, Dr. J. Throck Watson,
for his guidance, encouragement, and support throughout my graduate study at Michigan
State University. I would also like to thank Dr. John Allison for his continual
encouragement, support and serving as a second reader of my dissertation; Dr. Eric Tomg
and Mr. Dehua Hang for their guidance and advice on development of the computer
algorithms; Dr. Marcos Dantus for serving on my committee.

I especially thank my wife, Tong Pei, for her love, support, and patience that
make my life enjoyable. I am also very grateful to my mom, parents-in-law, brother, and
brother-in-law, for their caring, encouragement, and understanding during my graduate
study at Michigan State University.

To all Watson group members, Mr. Wei Wu, Mr. Yingda Xu, Dr. Ying Yang, Ms.

Susan Li, and Mr. Jose-Luis Gallegos-Perez, for your ﬁ'iendship and collaboration.

TABLE OF CONTENTS
LIST OF TABLES .................................................................................... xi
LIST OF FIGURES ................................................................................. xiii
ABBREVIATIONS ................................................................................ xxii
CHAPTER 1
INTRODUCTION: DISULFIDE MAPPING OF PROTEINS
BY MASS SPECTROMETRY ..................................................................... 1
I. Introduction .................................................................................... 1
H. Disulﬁde Mass Mapping Based on Proteolysis ........................................... 4
1. Overview of the Method ................................................................. 4
2. Constraints/Caveats ....................................................................... 7
3. An Experimental Protocol ............................................................... 8
4. A Representative Application ............................................................ 9
5. Advanced Topics: Using MALDI-PSD or ESI-MS/MS to Obtain
Additional Information on Disulﬁde-Linked Peptides .............................. 18
6. Conclusions ................................................................................ 22

III. Determination of Disulﬁde Bonds by Partial Reduction, Chemical Tagging,

Edman Sequencing Analysis and Mass Spectrometry ................................. 23
1. Overview of the Method ................................................................. 23
2. Constraints/Caveats ...................................................................... 25
3. An Experimental Protocol ............................................................... 26
4. A Representative Application .......................................................... 27
5. Conclusions ............................................................................... 29

vi

IV. Disulﬁde Mass Mapping Based on Partial Reduction,

Cyanylation, and CN-induced Cleavage ................................................. 29
1. Overview of Method .................................................................... 29
2. Constraints, Caveats, and Other Aspects ............................................. 33
3. Experimental Protocols ................................................................. 35
A. A Protocol for Determining the Location of Free Cysteines by
Alkylation/Cyanylation /CN-Induced Cleavage ................................. 35
B. A Protocol for Determining the Location of Free Cysteines by
Cyanylation/CN-Induced Cleavage ................................................ 35
C. A Protocol for Determining the Location of Disulﬁde Bonds by Partial
Reduction/ Cyanylation/CN-Induced Cleavage .................................. 36
4. A Representative Application: Disulﬁde Mass Mapping
Using Singly Reduced/Cyanylated Isoforms ........................................ 37
5. Advanced topic: Combining Partial reduction/Cyanylation/CN-induced
Cleavage with Proteolytic Mass Mapping ............................................ 41
6. Conclusions ............................................................................... 44
V. References .................................................................................... 45
CHAPTER 2

DEVELOPMENT OF MORE FLEXIBLE PROTOCOLS FOR THE PARTIAL
REDUCTION AND SUBSEQUENT CYANYLATION/CN-INDUCED

CLEAVAGE/MASS MAPPING METHODOLOGY ........................................... 50
I. Introduction .................................................................................... 50
11. Materials and Methods ...................................................................... 53

HI. Expand Partial Reduction Strategy
to Include Doubly Reduced Protein Isoforms ........................................... 57

IV. Addition of an Analysis Step by Mass Spectrometry after CN-induced Cleavage,
but Before Complete Reduction of the Residual Disulﬁde Bonds ................... 66

V. Further Investigation of the Status of C-terminus (Amide or Free Acid) of
CN-induced Cleavage Fragments .......................................................... 71

Vii

VI. Sequence Analysis of hninothiazolidine-Blocked Peptides by CID-MS/MS ....78

VII. Conclusions ................................................................................ 81
VIII. References ................................................................................ 82
CHAPTER 3

DETERMINATION OF THE DISULFIDE STRUCTURE OF SILLUCIN,
A HIGHLY KNOTTED, CYSTEINE-RICH PEPTIDE, BY

CYANYLATION/CN-INDUCED CLEAVAGE MASS MAPPING ....................... 84
I. Introduction ................................................................................... 84
11. Materials and Methods ..................................................................... 90
III. Results ....................................................................................... 93
1. Accurate Mass Measurement of Intact Sillucin .................................... 93
2. Partially Reduced/Cyanylated Species of Sillucin ................................. 94
3. A Single Reduced/Cyanylated Species .............................................. 96
4. A Doubly Reduced/Cyanylated Species ........................................... 106
IV. Discussions .............................................................................. 1 11
1. Partial Reduction ...................................................................... 111
2. Incomplete Cyanylation, Incomplete CN-induced cleavage,
and Disulﬁde Scrambling ............................................................ 113
V. Conclusion ................................................................................ 122
VI. References ................................................................................ 122
CHAPTER 4
DETERMHNIATION OF THE CYSTEINE STATUS OF THE CHLOROPLAST
GLYCERALDEHYDE-B-P DEHYDROGENASE SUBUNITS ......................... 125
I. Introduction ................................................................................. 125
11. Materials and Methods .................................................................. 131
HI. Results .................................................................................... 134

viii

1. Mass Mapping Based on Cyanylation and CN-induced Cleavage . . . . . 134

2. Subunit Composition ................................................................ 149

IV. Discussions .............................................................................. 149

V. References ................................................................................. 151
CHAPTER 5

COMPUTER ALGORITHMS FOR DISULFIDE MASS MAPPING OF
PROTEINES BASED ON PARTIAL REDUCTION AND CYANYLATION/

CN-INDUCED CLEAVAGE ................................................................... 154
I. Introduction ................................................................................. 154
H. A Naive Algorithm ....................................................................... 156
III. A Basic Algorithm ....................................................................... 158

1. Consideration of Singly Reduced Isoforms ........................................ 162

2. Consideration of Doubly Reduced Isoforms ....................................... 167

3. Consideration of Triply Reduced Isoforms ......................................... 170

IV. Negative Signature Masses and an Improved Algorithm ............................ 171
1. Negative Signature Masses ........................................................... 171

2. An Algorithm Based on Negative Signature Masses .............................. 173
V. Dealing with Imperfect Data .............................................................. 180
VI. Implementation and Testing of the Algorithms ....................................... 181
1. Implementation of the Basic Algorithm ............................................. 181
A. C++ Program Files ................................................................ 181

B. Preliminary Testing using Data from a

Model Protein: ribonuclease-A .................................................. 182
C. The Content of Input and Output Text Files ..................................... 183

ix

2. Implementation of the Algorithm Based on Negative Signature Masses .. . 1 89

A. C++ Program Files ................................................................ 189

B. Preliminary Testing using Data from a
Model Protein: ribonuclease-A ................................................... 189
C. The Content of Input and Output Text Files .................................... 190
VH. Conclusions ............................................................................... 197
VIII. References ................................................................................ 198

LIST OF TABLES

Table 1.1. Calculated m/z values of protonated CN-induced cleavage ﬁ'agments
from all possible singly reduced/cyanylated isofonns of an unknown
protein having the same sequence as RnaseA. The boldface entries
correspond to native RnaseA, one of the 105 possible isomeric
arrangements of the four cystines in a protein with this sequence. ........... 39

Table 1.2. m/z values of CN-cleavage products of two singly reduced/cyanylated
isoforrns GM2AP(110-142) after reduction of the residual disulﬁde
bond (adopted from reference 17). ............................................... 44

Table 2.1. m/z values for possible fragments resulting ﬁom the cleavage reaction
of one of the doubly reduced/cyanylated isoforrn of ribonuclease-A
represented by I-IPLC peak B in Figure 2.3. ..................................... 64

Table 2.2. Some itz-fragments studied by ESI—CID-MS/MS
on an ion trap. ....................................................................... 79

Table 3.1. Calculated m/z values for MI-I+ (based on average masses of amino
acid residues) of possible fragments resulting from CN-induced
cleavage of singly reduced/cyanylated isoforms of all disulﬁde
structural isomers of sillucin. ..................................................... 97

Table 3.2. Calculated m/z values (monoisotopic) for b and y ions of CN-induced

fragments itz-13-29-NH2 and itz-14-3O-NH2 from sillucin. ................ 104
Table 3.3. Calculated m/z values (monoisotopic) for b and y ions of CN-induced

ﬁ'agments itz-13-29-NH2 and itz-14-30-OH from sillucin. .................. 105
Table 3.4. Identities of the MALDI-MS peaks in Figure 3.5. ............................ 115
Table 3.5. Identities of the MALDI-MS peaks in Figure 3.9. ............................. 116
Table 3.6. Identities of the MALDI-MS peaks in Figure 3.11. ............................ 117

Table 4.1. The experimental and calculated m/z values for the protonated
CN-induced cleavage fragments of the cyanylated
GAPD subunit A. .................................................................... 139

Table 4.2. The experimental and calculated m/z values for the protonated

CN-induced cleavage fragments of the cyanylated GAPD
subunit B-oxidized .................................................................. 141

xi

Table 4.3. The experimental and calculated m/z values for the protonated
CN-induced cleavage fragments of the cyanylated GAPD
subunit B-reduced. .................................................................. 143

Table 4.4. The experimental and calculated m/z values for the fragments from B-
elimination and incomplete CN-induced cleavage of subunit A of
GAPD. ................................................................................ 146

Table 4.5. The experimental and calculated m/z values for the fragments from B-
elimination and incomplete CN-induced cleavage of subunit B of
GAPD. ................................................................................ 147

Table 5.1. Some of the same CN-induced fragments from two singly reduced
isoforrns of two hypothetical protein isomers with different disulﬁde
structures. The hypothetical protein has 90 amino acid residues,
among which 8 cysteines located at positions 10, 20, 30, 40, 50, 60,

70, and 80. ........................................................................... 159
Table 5.2. The number of possible disulﬁde bonds and disulﬁde structures for a

protein having n disulﬁde bonds. ................................................. 160
Table 5.3. The CN-induced fragments from all the possible singly reduced/

cyanylated isofonns of ribonuclease-A. ......................................... 164
Table 5.4. Filtering the List B for ribonuclease-A using three negative masses ........ 1 76

Table 5.5. The shortened disulﬁde bond List C for ribonuclease—A after ﬁltering
using negative masses. ............................................................. 177

Table 5.6. Forming the disulﬁde structure list D from the shortened disulﬁde list
C (Table 5.5) for ribonuclease—A. ................................................ 179

xii

Figure 1.1.

Figure 1.2.

Figure 1.3.

Figure 1.4.

Figure 1.5.

Figure 1.6.

LIST OF FIGURES

Different congeners of a hypothetical protein with 4 cysteines: 1

completely reduced form, 6 one-disulﬁde forms, and 3 two-disulﬁde

forms. The number of congeners increases geometrically with the

number of cysteines. ................................................................... 2

Conceptual representation of mass changes observed upon reduction
of peptides containing intermolecular vs. intramolecular disulﬁde
linkages. “m/z” desiganates mass to charge ratio. ................................ 3

Conceptual illustration of the different array of fragments that can be

used to distinguish the three possible isomeric arrangements of two

disulﬁde bonds in a hypothetical protein containing four cysteines by

the general proteolytic approach for disulﬁde mass mapping. Note

that fragments C, C’ , and C” have the same mass in each case

because they contain no cysteine, whereas the mass of the fragments
containing cysteine vary substantially for the different isomeric

disulﬁde arrangements. ................................................................ 6

Overall sequence of tACE showing tryptic cleavage sites at K491 and

K511; these deﬁne the origin of tryptic fragment Tyr492-Ly5511 at the

lower left that shifts from 2145.4 Da to 2453.7 Da upon alkylation of

the free sulﬂiydryl on Cys496. The mass of the protonated alkylated-

peptide (2453.7 Da) is very close to the observed peak at m/z 2453.4

in Figure 1.6. .......................................................................... 10

HPLC chromatograms of the endoproteinase Lys-C digestion mixture

of AEDANS-labeled tACE as recorded by two different detectors. The

upper panel was monitored at 340 nm for the fluorescence of the

AEDAN S ﬂuorophore to selectively detect the peptide with the
AEDANS-modiﬁed cysteine; the lower panel was monitored at 214

nm to detect all the digestion peptides (adopted from reference 39) ........... 12

The MALDI mass spectrum of the HPLC fraction represented by the

single peak in the top panel of Figure 1.5. The peak at m/z at 2453.4
represents the protonated proteolytic fragment containing the modiﬁed
cysteine as illustrated in Figure 1.4 and described in the text. The other

two peaks represent other proteolytic peptides that coelute

with the alkylated one (adopted from 39). ........................................ 13

xiii

Figure 1.7. Conceptual representations of disulﬁde arrangement among the six
cysteines involved in the three cysteine-containing fragments from
CNBr digestion of tACE. Category (i) shows the simplest
arrangement, which derives from three isolated disulﬁdes in tACE;
experimental results indicate that this category represents the disulﬁde
arrangement in native tACE. Categories (ii) and (iii) consist of six and
eight isomeric arrangements, respectively, some of which involve
overlapping disulﬁde bonds (indicated, in part for two possibilities, by
dashed lines). .......................................................................... 15

Figure 1.8. MALDI mass spectrum of the putative peptide of a CNBr fragment
(see Figure 1.9) of the tACE protein. Peaks at m/z 856.6 and m/z
4388.0 correspond to the individual peptides connected by the
disulﬁde as cleaved during the MALDI process
(adopted ﬁ'om reference 39). ....................................................... 16

Figure 1.9. Conceptual illustration of digestion of tACE by CNBr and trypsin to
produce a cysteine-containing fragment, the analysis of which by
MALDI-MS identiﬁes the disulﬁde Cys538-Cy8550. ........................... 17

Figure 1.10. MALDI-PSD-MS mass spectrum of a disulﬁde-linked proteolytic
fragment ﬁ'om dual digestion of the TNFbp protein with trypsin and
thermolysin. The numbers by the structure give the mass of the
indicated fragment ions; the peaks marked by an asterisk at m/z 825.6,
940, and 1027 conﬁrm the disulﬁde linkage pattern shown (adopted
from reference 28). ................................................................... 19

Figure 1.11. ESI—CID-MS/MS mass spectrum of a triply charged ion [M+3H+]3+
at m/z 1406.1 corresponding to a disulﬁde—linked tryptic fragment (in
which a disulﬁde bond links two peptides, S130-K156 and Y171-
K181) of bovine beta-1,4-galactosyltransferase (adopted
ﬁom reference 34). .................................................................... 21

Figure 1.12. A diagrammatic presentation of sequencing results for the partial
reduction/alkylated isoforrns of AVR9 represented by peaks in the
HPLC chromatogram. Peak 2 represents a singly reduced isoform
from partial reduction and indicates that AVR9 contains a disulﬁde
bond between Cys2-Cysl6. Peak 3 represents a doubly reduced
isoform containing a rediual disulﬁde bond between Cys6-Cys1 9
before complete reduction; peak 4 represents a doubly reduced isoform
containing a residual disulﬁde Cys12-Cy326 before complete
reduction. The peaks 5 and 6 represent the completely reduced/NEM-
modiﬁed species (as determined by mass spectrometry); both species
have the same mass and apparently result from epimeric forms of the
prochiral alkylating reagent (adopted from reference 44). ..................... 28

xiv

Figure 1.13. (A) Cyanylation of sulﬂiydryl group by CDAP and (B) peptide bond
cleavage by ammonia. (itz = iminothiazolidine-carboxyl; B—elimination
= loss of HSCN from a cyanylated cysteine residue.)
(adopted from reference 47) ......................................................... 30

Figure 1.14. Chemical overview of the partial reduction/cyanylation/CN-induced
cleavage methodology using singly reduced isoforrns of a hypothetical
2-disulﬁde protein (adopted from reference 16). ................................. 32

Figure 1.15. HPLC separation of denatured ribonuclease A and its partially
reduced/cyanylated isoforrns. The separation was done by reversed-
phase HPLC on a Vydac C18 column (5pm, 4.6mmx250mm) at a ﬂow
rate of 1.5 mI/min with a linear gradient 20-40% B in 90 minutes,
where A = 0.1% TFA in water and B = 0.1% TFA in 90% CH3CN.
The peak marked by IP is the intact protein (ribonuclease A). Peaks 1-
4 represent singly-reduced/cyanylated ribonuclease-A isoforms, as
determined by MALDI—TOF MS (adopted ﬁ'om reference 16). ............... 38

Figure 1.16. The MALDI mass spectra of four peptide mixtures resulting from the
CN-induced cleavage of the singly reduced/cyanylated ribonuclease A
isoforrns. Spectra (a)- ((1) correspond to the HPLC peaks 1-4 in Figure
1., respectively. The symbols # and * represent the doubly charged
fragments and protonated B-elimination products, respectively. The
double mark #* in (c) shows a doubly charged
B-elimination product (adopted from reference 16). ............................. 40

Figure 1.17. C18-RP-HPLC separation of partially reduced and cyanylated
isoforrns of GM2AP(1 10-142) (adopted from reference 17). .................. 43

Figure 2.1. Overview of chemical reactions involved in the disulﬁde mass
mapping of two doubly reduced/cyanylated isoforrns (BI) and (C1) of
a hypothetical protein (A). (C2): the cleaved fragments from (C1)
before complete reduction; (B3) and (C3): the corresponding cleaved
fragments from (B1) and (Cl) after complete reduction. ....................... 58

Figure 2.2. Disulﬁde mass mapping using a doubly reduced (B1) and a singly
reduced/cyanylated (D1) isoform of a hypothetical protein (A)
containing four disulﬁde bonds. (B3) and (D3): the corresponding
cleaved fragments from (B1) and (D1) after complete reduction. ............ 59

XV

Figure 2.3. HPLC chromatogram of partially reduced/cyanylated isoforrns of
ribonuclease-A: IP represents the intact ribonuclease-A; Peaks 1-4
represent the singly reduced/cyanylated isoforrns (as also shown in
Figure 1.15 in Chapter 1); Peaks A-F represent the doubly
reduced/cyanylated isoforms; Peaks g-j represent the triply
reduced/cyanylated isoforrns, as determined
by MALDI-MS analysis. ............................................................ 61

Figure 2.4. The MALDI mass spectrum of a peptide mixture resulting from the
cleavage of a doubly reduced/cyanylated ribonuclease-A isoform
corresponding to HPLC peak B in Figure 2.3. .................................... 63

Figure 2.5. The MALDI mass spectrum of the peptide mixture resulting from CN-
induced cleavage (but before complete reduction) of the doubly
reduced/cyanylated ribonuclease-A isoform represented by HPLC
peak B in Figure 2.3. Itz-40—64 represents the itz-blocked fragment of
ribonuclease-A from residue 40 to 64. Details about prompt
fragmentation (or in-source fragmentation) during analysis by
MALDI—MS can be found in references 4-8. ..................................... 68

Figure 2.6. Eliminating a possible disulﬁde structure (B) by mass mapping
analysis of the CN—cleavage products before the complete reduction
step. Disulﬁde structure (A) shows the correct disulﬁde linkages in the
hypothetical protein; structures (B) and (C) are two of 15 possible
structures of arranging six cysteines as three disulﬁde bonds
in such a protein. ..................................................................... 69

Figure 2.7. The MALDI mass spectrum of the peptide mixture resulting from CN-
induced cleavage (but before complete reduction) of the doubly
reduced/cyanylated ribonuclease-A isoform represented by HPLC
peak A in Figure 2.3. Itz-26-64 represents the itz-blocked fragment
of ribonuclease-A from residue 26 to 64. ......................................... 70

Figure 2.8. A proposed mechanism (9) involving nucleophilic attack by 0H for
the CN-induced cleavage reaction at a cyanylated cysteine residue. .......... 72

Figure 2.9. Production of CN-induced cleavage ﬁ'agments with an amide C-
terminus via a mechanism involving nucleophilic attack by NH3

at a cyanylated cysteine residue. ................................................... 73

Figure 2.10. Isotope peak intensity distribution in expanded ESI mass spectrum of
itz-13-29-red-NH2 (itz-CCGDSGYWCRQCGIKYT—NHZ) produced

from sillucin. “M” represents the molecule and “MHZ“ represents the
doubly protonated molecule. ........................................................ 75

xvi

Figure 2.11. Isotope peak intensity distribution in expanded MALDI mass

spectrum of itz-12-B@1 3-23-ox-OH (itz-C(B@C)CGDSGYWCRQ-
OH) produced from sillucin. ........................................................ 76

Figure 2.12. Isotope peak intensity distributions for two forms of the CN-induced

fragment, 1-12-red (ACLPNSCVSKGC; 1-12-red-NH2 and 1-12-red-

OH), produced from sillucin, were detected during ESI infusion on an

ion trap. “0.3x” in the middle panel indicates that the peak intensity is

30% of the intensity for the corresponding peaks in the bottom panel,
respectively. ........................................................................... 77

Figure 2.13. ESI-CID-MS/MS of itz-CEGNPYVPVHFDASV (residue 110-124 of

Figure 3.1.

Figure 3.2.

Figure 3.3.

ribonuclease-A), a doubly charged ion (m/z 830.4), derived from
cyanylation/CN-induced cleavage of a singly reduced ribonuclease-A
isoform, obtained on an ion trap MS. The nomenclature of Biemann
(13) for b and y ions is used throughout, although for convenience,
lower case letters are also used to indicate the peptide bond that is
cleaved to produced key fragments (by and y9) in a manner analogous

to that introduced by Roepstorff et al. (14) using capital letters. .............. 80

Overview of chemical reactions involved in the disulﬁde mass

mapping of a singly reduced/cyanylated species of sillucin.

Recognition of the cleaved fragments by mass spectrometry allows the
linkage of Cysl 3-Cys30 to be deduced. itz represents the 2-
iminothiazolidine-4-carboxyl blocked N-terminus resulting from CN-
induced cleavage by aqueous ammonia on the N-terminal side of the
cyanylated cysteines. * The structure in the parenthesis designates a [3-
elimination product, which results in this case from loss of HSCN at
cyanylated Cys 13 (IS-elimination) and CN-induced cleavage

at cyanylated Cys30. ................................................................. 88

Overview of chemical reactions involved in the disulﬁde mass

mapping of a doubly reduced/cyanylated isoform (C1) of sillucin (A)

as well as a singly reduced/cyanylated isoform (B1). Recognition of

the cleaved fragments (C2) ﬁ’om Cl before complete reduction allows

the connectivity of Cys14-Cys21 to be deduced. Recognition of the

cleaved fragments (B3) from B2 (shown in Figure 3.1) after complete
reduction combined with the determination of Cysl4-Cy521 as

described above allows the connectivity of Cyle-Cys24 to be

deduced; Z could be 2, 7, or 30. ..................................................... 89

The MALDI of spectrum of intact sillucin. ....................................... 94

xvii

Figure 3.4.

Figure 3.5.

Figure 3.6.

Figure 3.7.

Figure 3.8.

Figure 3.9.

HPLC chromatogram of denatured sillucin (IP) and some of its

partially reduced/cyanylated species. Separation was carried out on a

Vydac C18 column at a ﬂow rate of 1.0 mL/min with a linear gradient

of 15% to 40% B in 50 minutes, where A = 0.1% TFA in water and B

= 90% (v/v) acetonitrile/0.1% TFA. Peaks IP, 1, and 2 represent the

intact peptide, 3 singly reduced/cyanylated species, and a doubly
reduced/cyanylated species, respectively; peak 3 represents the

completely reduced/cyanylated species, as determined from analysis

by MALDI-MS. ...................................................................... 95

The MALDI-MS spectrum of the peptide mixture resulting from CN-

induced cleavage and subsequent complete reduction of the singly
reduced/cyanylated sillucin species represented by HPLC peak 1 in

Figure 3.4. “IS” designates an internal standard. ................................ 98

Assignment of the disulﬁde Cysl3-Cys30 in sillucin. .......................... 99

Detection of CN-induced fragment 1-12-OH (two forms: 1-12-ox—OH
and 1-12-red-OH) by MALDI-MS aﬁer it was separated from
itz-12-29-red-NH2 and 1-B@13-29-red-NH2. ................................. 102

ESI CID-MS/MS spectrum of CN-induced ﬁagment itz-13-29-NH2
from sillucin. The precursor is a doubly protonated molecule; [MHz-
HZO]2+ designates a dehydrated and doubly protonated fragment of
the precursor; “B13” indicates the peptide bond that is cleaved in the

peptide backbone between residue G and I (see deﬁnition of the
nomenclature in reference 21) to form an amino terminal ion

represented by “1113” (singly charged) and b132+ (doubly charged);

the deﬁnition of a, b, and y ions can be found in reference 22. .............. 103

The MALDI-MS spectrum of the peptide mixture resulting from CN-

induced cleavage (before complete reduction) of the doubly
reduced/cyanylated sillucin species represented by HPLC peak 2

in Figure 3.4. ........................................................................ 107

Figure 3.10. Assignment of disulﬁde bonds Cysl4-Cy821, Cyle-Cy824,

and Cys2-Cys7. ..................................................................... 108

Figure 3.11. The MALDI-MS spectrum of the peptide mixture resulting from CN-

induced cleavage and subsequent complete reduction of the doubly
reduced/cyanylated sillucin species represented by HPLC peak 2
in Figure 3.4. ........................................................................ 110

xviii

Figure 3.12. Formation of itz-13-SCN@24—29-ox-NH2 and

itz-13-SCN@24-29-red-NH2 because of no
CN-induced cleavage at SCN@24. ............................................... 118

Figure 3.13. Formation of itz-l 3-30-red-OH due to incomplete cyanylation at

Cys 30; Cys3OSH is used to designate the ﬂee cysteine residue. ............ 120

Figure 3.14. Didactic illustration of possible scrambling in a singly reduced

Figure 4.1.

Figure 4.2.

Figure 4.3.

Figure 4.4.

isoform, [sillucin(l3SH-30SH)], illustrated for the case in which the

disulﬁde Cysl 3-Cys30 has been reduced, and a trace amount has

dissociated to form the thiolate ion, [sillucin(13S‘-3OSH)], which

scrambles to [sillucin(lZS’-3OSH)] before cyanylation is

accomplished. ........................................................................ 1 2 1

Sequence alignment of the pea chloroplast glyceraldehyde-3-P
dehydrogenase A (gi [“gi” is the National Center for Biotechnology
Information protein identiﬁcation number] 66025) and B (gilZO663)
subunits (pea a and pea b), the Sulfolobus solfataricus glyceraldehyde-
3-P dehydrogenase (Protein Data Bank (PDB) ﬁle 1b7g), and the
Bacillus stearothermophilus glyceraldehyde-3-P dehydrogenase (PDB
ﬁle 1 gdl). Identical residues in the B subunit (pea b) and in one or
both of the sequences ﬂom the Protein Data Bank are bolded.
Residues that are superimposable in the two crystal structures are
underlined. Cysteines in pea a and pea b are numbered, above the
residue for pea a, below for pea b. Dashes indicate gaps introduced

to optimize the alignment. .......................................................... 126

Conventional methodology for determining the status of the cysteines

in a hypothetical protein of 70 amino acid residues. The straight line
designates the protein backbone; “SH” designates a ﬂee cysteine; “S—

S” designates a disulﬁde bonds; “SR” designates an alkylated cysteine
residue; and “E” together with an arrow indicates where the enzyme

cleaves. .............................................................................. 128

Cyanylation and CN-induced cleavage process to identify ﬂee

cysteines in a hypothetical protein of 70 amino acid residues. The

straight line designates the protein backbone; “SH” designates a ﬂee

cysteine; “S—S” designates a disulﬁde bond; “SCN” designates a

cyanylated cysteine residue; and “itz” designates an
iminothialzolidine-carboxyl blocked amino terminus. ........................ 130

MALDI mass spectrum of the chloroplast glyceraldehyde-3-P

dehydrogenase. 1 ul of a solution of 9 pmol enzyme subunits in
12 mM pH 7-ammonium acetate buffer was used. ............................. 135

xix

Figure 4.5. HPLC separation of the cyanylated A and B subunits of chloroplast
glyceraldehyde-3-P dehydrogenase (GAPD). Cyanylated proteins
(approximately 3.8 nmol total of the A and B subunits) were separated
by reverse phase HPLC on a Vydac C18 column. UV detection was
done at 215 nm. For details see ‘Methods’. Peaks 1 and 2 represent
an 8.3k-Da protein; peaks 4 and 5 represent the cyanylated B and A
subunits of GAPD, respectively, as determined by MALDI-TOF-MS.

Nothing was detected ﬂom HPLC ﬂaction 3 by MALDI-TOF-MS.
The arrows indicate the starting and stopping points for collecting
ﬂactions for peaks 4 and 5, respectively. ........................................ 136

Figure 4.6. The MALDI spectrum of the CN-induced cleavage ﬂagrnents of the
cyanylated A subunit of chloroplast glyceraldehydes-3-P
dehydrogenase represented by HPLC peak 5 in Figure 4.5. Portions 1
and 2 are the low-mass and high-mass portions, respectively. The
mark “B” in parenthesis indicates that the peak represents a CN-
induced cleavage ﬂagrnent of the cyanylated B subunit because HPLC
peaks 4 and 5 (Figure 4.5) were not completely resolved. The mark
“++” designates a peak representing a doubly protonated (also doubly
charged) protein ﬂagrnent. All other peaks without such marks are
singly protonated protein ﬂagments. The mark “7” indicates that
the source of the peak is not identiﬁed. .......................................... 138

Figure 4.7. The MALDI spectrum of the CN-induced cleavage ﬂagrnents of the
cyanylated B subunit of chloroplast glyceraldehyde-3-P
dehydrogenase represented by HPLC peak 4 in Figure 4.5. Portions 1
and 2 are the low-mass and high-mass portions, respectively. The
mark “A” in parenthesis indicates that the peak represents a CN-
induced cleavage ﬂagrnent of the cyanylated A subunit because HPLC
peaks 4 and 5 (Figure 4.5) were not completely resolved. The mark
“++” designates a peak representing a doubly protonated (also doubly
charged) protein ﬂagrnent. All other peaks without such marks
represent singly protonated protein ﬂagrnents. The mark “7” indicates
that the source of the peak is not identiﬁed. The peak at m/z 6962.1
represents both a singly protonated CN-induced cleavage ﬂagrnent
(itz-291-353, marked as “+, B,”; calculated 01/2 for MH+: 6960.8) ﬂom
the cyanylated subunit B-reduced and a doubly protonated CN-induced
cleavage/B-elimination ﬂagrnent (itz-159-B@280-290) [Table 4.2]
formed by B-elimination at cyanylated Cys280 instead of
cleavage (marked as “++, B”; calculated m/z for MHZ”: 6962.6). .......... 140

Figure 4.8. The CN-induced cleavage ﬂagrnents ﬂom the oxidized and reduced
forms of the cyanylated subunit B of GAPD. The straight line
designates the protein backbone; “S—S” designates a disulﬁde bond;
“SCN” designates a cyanylated cysteine residue; and “itz” designates
iminothialzolidine-carboxyl blocked amino terminus. ........................ 144

XX

Figure 4.9.

Figure 5.1.

Figure 5.2.

Figure 5.3.

Figure 5.4.

The formation of some ﬂagments resulted ﬂom B-elimination,

incomplete CN-induced cleavage, and incomplete cyanylation of

subunit A of GAPD. The straight line designates the protein

backbone; “SCN” designates a cyanylated cysteine residue; “SH”

designates a ﬂee cysteine residue; “[3” designates B-elimination (loss

of HSCN ﬂom a cyanylated cysteine residue); and “itz” designates
iminothialzolidine-carboxyl blocked amino terminus. ........................ 148

Information ﬂow for a naive algorithm for disulﬁde mass mapping

based on partial reduction, cyanylation and CN-induced cleavage. The
sequence of the protein to be analyzed and the locations of ﬂee

cysteines are used as input 1; MS spectra of the CN-induced ﬂagments

and m/z measurement accuracy are used as input 2. The output is a

ranked disulﬁde structure list, in which candidate disulﬁde structures

are ranked according to the number of CN-induced ﬂagments matched
(Experimental m/z values and calculated m/z values are compared). ......157

A basic algorithm for disulﬁde mass mapping based on partial
reduction, cyanylation and CN-induced cleavage. The sequence of the
protein to be analyzed and the locations of ﬂee cysteines are used as
input 1; MS spectra of the CN-induced ﬂagments and m/z
measurement accuracy are used as input 2. The output is a ranked
partially reduced isoform list, in which candidate partially reduced
isoforms are ranked according to matching percentage = (the number
of CN-induced ﬂagments matched/the number of such ﬂagments

expected for a partially reduced isoform). ....................................... 161
Using a negative signature mass (m/z 6546.4) ﬂom ribonuclease—A

to reject the existence of 8 disulﬁde bonds. ..................................... 172
An algorithm based on negative signature masses. ............................ 174

xxi

2-D NMR
4-VP
AEDAN S

Ala

Asn

Asp

CDAP

CID

CNBr

Cys

Cys i - Cys j
D

Da

DEAE

ESI

FAB

ABBREVIATIONS

two-dimensional nuclear magnetic resonance

4-vinyl pyridine

5-([2—(acetyl)amino]ethy1amino) naphthalene-l-sulfonic acid group
alanine residue

arginine residue

asparagine residue

aspartic acid residue

B-elimination of HSCN ﬂom cyanylated cysteine residue
cysteine residue

1 -cyano-4-dimethylamino-pyridinium

collision induced dissociation

cyanogen bromide

cysteine residue

the disulﬁde bond between Cys i and Cys j

aspartic residue

dalton

diethylaminoethyl

glutamic acid residue

electrospray

phenylalanine residue

fast atom bombardment

xxii

FTICR

GAPD

Glu

Gln

GMZAP

His

HPLC

IAEDAN S

IGFs

Ile

IS

itz

LC

Leu

ﬂourier transform ion cyclotron resonance
glycine residue
glyceraldehydes-3 -P dehydrogenase

National Center for Biotechnology Information protein identiﬁcation
number

glutamic acid residue

glutamine residue

GM2-activator protein

histidine residue

histidine residue

high performance liquid chromatography

isoleucine residue

5-([2-(iodoacetyl)amino]ethylamino) naphthalene-1 —sulfonic acid
insulin-like grth factors

isoleucine residue

internal standard

iminothiazolidine-4carboxyl terminus, resulting ﬂom the CN-induced
cleavage by aqueous ammonia on the N-terminal side of the cyanylated
cysteine residues

kilo

lysine residue

leucine residue

liquid chromatography

leucine residue

xxiii

LR3 IGF- 1

Lys

Met
MH+
MS
MH22+
MS/MS

m/z

NAD

NADP

OH

OX

PDB

long insulin-like grth factor-1

lysine residue

molecule or mass of a molecule

methionine residue

matrix assisted laser desorption/ionization

methyl

methionine residue

protonated molecule

mass spectrometry, mass spectral or mass spectrometer

doubly protonated molecule

mass spectrometry/mass spectrometry or tandem mass spectrometry
mass to charge ratio

asparagine residue

nicotinamide adenine dinucleotide

nicotinamide adenine dinucleotide phosphate

N-ethylmaleimide

designates an amide C-terminus

designates a ﬂee acid C-terminus

designates that there are residual disulﬁde bonds within a CN-induced
cleavage ﬂagment

proline residue
protease

Protein Data Bank

xxiv

PEG

Phe

PSD

PTH

red

RnaseA

Ser

tACE
TCEP
Thr

TFA
TNFbp
TOF
Tris-HCl

TIP

polyethylene glycol

phenylalanine residue

proline residue

post source decay

phenylthiohydantoin

glutamine residue

arginine residue

resolution

indicates that the residual disulﬁde bonds, which survived the partial
reduction, have been reduced during the complete reduction step; thus, all
the cysteine residues within the ﬂagrnent are ﬂee cysteines
ribonuclease A

reversed phase

serine residue

serine residue

threonine residue

testis angiotensin-converting enzyme
tris(2-carboxyethyl)phosphine

threonine residue

triﬂuoroacetic acid

tumor necrosis factor binding protein

time of ﬂight

tris(hydroxymethyl)aminomethane hydrochloride

tryptophan residue

XXV

Tyr

tyrosine residue
ultra violet

valine residue
valine residue
tryptophan residue

tyrosine residue

xxvi

CHAPTER 1
INTRODUCTION:

DISULFIDE MAPPING OF PROTEINS BY MASS SPECTROMETRY

I. Introduction

The molecular weight and amino acid sequence are available at an early stage in
characterizing a protein, while details of posttranslational modiﬁcation usually require
more time and effort to elucidate. Determination of the cysteine status in a protein is an
important part of its structural characterization. Concerning cysteine status, there are
usually two simple questions: Are there any disulﬁde bonds? If there are, what is the
connectivity of the cysteines in each disulﬁde bond? A sense of the complexity of
possible disulﬁde bonding structures that must be considered for a given number of
cysteines can be gleaned ﬂom Figure 1.1.

The question of whether there are disulﬁde bonds is relatively easy to answer.
The difference in mass between a protein containing two free cysteines (“Free cysteines”
are cysteines that have ﬂee sulﬂrydryls and are not involved in disulﬁde bond formation.)
and that containing two disulfide cysteines (“Disulﬁde cysteines” are cysteines that are
involved in disulﬁde bonds.) forming a disulﬁde bond is two Daltons. However,
conﬁdence in the experimental mass measurement will depend on the molecular weight
of the protein as well as the resolution and mass accuracy of the mass spectrometer. For
example, for a protein in excess of 10 kDa and a mass spectrometer with a resolving
power of 10,000 and a mass accuracy no better than 1 in 104, the analyst may wish to

chemically modify the protein in a way that will exaggerate the mass difference between

the oxidized and reduced state. Chemical reactions with sulﬂiydryls ranging ﬂom
alkylation with iodoacetamide to 4-hydroxymercuri-benzoic acid (1) would shift the mass

difference between a cystine (disulﬁde) and two ﬂee or modiﬁed cysteines ﬂom 2 Da to

116.1 Da up to 643.4 Da, respectively.

Possible Redox States in A Protein Containing Four Cysteines

10 20 3'0 40
SH SH SH SH

 

 

 

 

 

 

10 20 30 4o 10 20 30 4o 10 20 30 40
I I § I I I I I I g I I
s — H SH SH S—s SH SH H s—s
10f 21;” 30 i1 1'0 18H I” 4'0 7‘” 210 i” 4'0
3 g «r 20 a) s 10 s 30 3
1'0 I20 310 40 1'0 ‘1 Is 40 1'0 1" 3'0 ‘1‘
s-—-s s-——-§ 8 2° 3° S s———J§L———s 4°

 

Figure 1.1. Different congeners of a hypothetical protein with 4 cysteines: 1 completely
reduced form, 6 one-disulﬁde forms, and 3 two-disulﬁde forms. The number of
congeners increases geometrically with the number of cysteines. The straight line
designates the protein backbone; “S—S” designates a disulﬁde bond; “SH” designates a

ﬂee cysteine.

Disulﬁde bonds come in two forms, intramolecular and intermolecular disulﬁde
bonds. Recognition of intermolecular disulﬁde bonding can be accomplished by mass
analysis of the protein before and after treatment with an excess of a reducing agent. An

increase in the mass of the protein by only a few daltons (as each cystine acquires two

hydrogens upon reduction) will indicate intramolecular disulﬁdes exclusively. On the
other hand, a dramatic shift (by hundreds to thousands of daltons) to lower mass for the
reduced species ﬂom the original protein indicates that at least one intermolecular

disulﬁde bond is present (see Figure 1.2, with further explanation in the next section).

 

 

 

SH
r +Y + W + \eS-S—i/
. s—s ,
\ l \ /
\ l \ /
\ l \, / Initial analysis of peptide
‘ mixture by mass spectrometry (MS)
I l
- M /
/ 4‘ 0”
/ .
/ /\ Subsequent reduction of
/ . .
peptrde mixture
| l I I and re-analysis by MS
' ’ M+2
m/z

Figure 1.2. Conceptual representation of mass changes observed upon reduction of
peptides containing intermolecular vs. intramolecular disulﬁde bonds. “m/z” desiganates
mass to charge ratio.

Here, we will examine different analytical strategies and describe some
representative experimental protocols for determining the connectivity of various
disulﬁde bonds. The emphasis in this chapter will be on controlled modiﬁcation and
degradation of the original protein with intervening analyses by mass spectrometry to
map the degradation products according to mass shift ﬂom that of the primary structure.
Three major strategies for controlled degradation of the protein will be highlighted: mass
mapping based on proteolysis (2-6), partial reduction coupled with chemical tagging and

mass mapping/Edman Sequencing (6-14), and partial reduction coupled with CN-induced

cleavage involving cyanylation followed by mass mapping (15-20). All mapping
strategies are based on a knowledge of the amino acid sequence and cleavage of the
protein backbone at speciﬁc sites; the involvement of particular disulﬁde cysteines in a

given disulﬁde bond is deduced ﬂom mass analysis of the cleaved ﬂagments.

II. Disulﬁde Mass Mapping Based on Proteolysis

1. Overview of the Method

Given the sequence of arrrino acid residues in a protein and an enzyme of known
speciﬁcity, one can anticipate the ﬂagments (and thus, compute their masses) into which
the protein would be cleaved upon proteolytic digestion. The analytical process of
correlating experimentally detected fragments with those predicted based on a knowledge
of the sequence is called ‘mapping’. Experimentally, one subjects the protein to
proteolytic hydrolysis, analyzes the digestion mixture, and compares the experimental
results with the expected results. The larger the proportion (at least 25% coverage) of the
expected ﬂagments detected, the greater the conﬁdence in correct identiﬁcation of the
protein. In early applications of this mapping strategy, the components of the digestion
mixture were separated analytically by HPLC with identiﬁcation by Edman sequencing,
but more recently, the digestion mixture has been analyzed directly by mass spectrometry
(21), in which case the methodology is called ‘mass mapping’.

Ideally, the protease (P) will cleave the protein at least once between the cysteines
as indicated by ‘P’ in the analytical scheme involving a hypothetical protein, as illustrated

Figure 1.3. Proteolytic cleavage gives a different array of ﬂagments depending on the

connectivity of the disulﬁdes. A tedious aspect of proteolytic disulﬁde mapping is the
isolation and identiﬁcation of the cysteinyl and cystinyl peptide ﬂagments ﬂom a usually
large group of non-sulfur containing peptides. A cysteinyl ﬂagrnent usually contains a
ﬂee cysteine residue; whereas a cystinyl ﬂagrnent usually contains an intra- or inter-
disulﬁde bond (Figure 1.2). Cysteinyl peptides that are ‘missing’ in the results of
analysis of the digestion mixture (compared to the list of anticipated proteolytic peptides)
indicate which cysteines are involved in the disulﬁdes (cystines). Of prime importance is
the isolation and identiﬁcation of the cystinyl proteolytic peptides, as these indicate
which two cysteines are involved in a particular disulﬁde bond. The cystinyl peptides
will be represented by unexpected HPLC peaks in a chromatogram of the digestion
mixture of the protein or by mass Spectral peaks for ﬂagments heavier than anticipated
from consideration of individual proteolytic ﬂagments of the completely reduced protein.
Recognition of cysteinyl peptides can be facilitated by the clever stable-isotope labeling

strategies recently reported by Chen et al. (22) and Adamczyk (23).

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

J—i-T—j ;——I l g—L 1
mi czicai ml mi 02M or our czicai or
P P P P P P P P P P P P
Proteolytic
Digestion *
f ‘C‘ ‘1 ‘3 *‘i‘ i‘
c3 —CT 02 1; C4 C2
Mass (A) (B) (C) (A') (B') /(C') (A") (53") (0")
Spectrometry \ \,// \\ l// \\ ///’/
RI 4’V V R. g S I RI 4’ V\
(C) (A) (B) M 7 (0') (A') (B) m (0") (3") m m/z

Figure 1.3. Conceptual illustration of the different array of ﬂagments that can be used to
distinguish the three possible isomeric arrangements of two disulﬁde bonds in a
hypothetical protein containing four disulﬁde cysteines by the classic proteolytic
approach for disulﬁde mass mapping. Note that ﬂagments C, C’ , and C” have the same
mass in each case because they contain no cysteine, whereas the mass of the ﬂagments
containing cysteines vary substantially for the different isomeric disulﬁde arrangements.
To the extent possible, it is prudent to perform conﬁrmatory experiments on the
primary mapping results. For example, in Figure 1.2, even though the peak of the highest
m/z value in the ﬁrst mass spectrum is consistent with the mass of the disulﬁde—linked
peptides, subjecting the putative structure to chemical reduction and detecting peaks for
the individual peptides, as illustrated in the bottom panel of Figure 1.2, substantially
increases conﬁdence in the ﬁnal results. Similarly, chemical reduction of the structure
containing an intramolecular disulﬁde in the top panel of Figure 1.2 Should shift the mass
of the putative species by 2 Da, as illustrated in the lower panel. Alternatively, reduction
of disulﬁdes during analysis by matrix assisted laser desorption/ionization (MALDI)

mass spectrometry (MS) (also called ‘prompt’ ﬂagrnentation; references 24-28) and

sequencing analysis by MALDI-post source decay (PSD)-MS (reference 17, 27-30) or

electrospray (ESI) tandem mass spectrometry (MS/MS) (references 31-35) can provide
conﬁrmation for preliminary results based solely on mass; examples of these procedures
are described below in section 4: “Representative Application” and section 5: “Advanced

Topic”.

2. Constraints/Caveats

It is desirable, if not a requirement, that at least one proteolytic site be available
between each of the cysteines in the amino acid sequence of the protein. The lack of any
such cleavage site could complicate the analysis because it will not produce separate
proteolytic peptides for each of the cysteines (or cystines). A single proteolytic peptide
containing two or more cysteines (or cystines) will complicate the analysis and/or lead to
ambiguous results. The closer the cysteines are to one another in the sequence, the less
likely it will be to ﬁnd a useful proteolytic site between them. In the extreme case of
adjacent cysteines, the proteolytic approach is unsatisfactory, if not useless (36).

A ﬂequent concern in using the proteolytic approach to disulﬁde mapping is the
problem of disulﬁde scrambling or exchange. Most proteolytic digestions are conducted
in slightly alkaline conditions (e.g., pH 8), in which a signiﬁcant ﬂaction of any ﬂee
sulfhydryl group will be deprotonated to the thiolate form. An existing disulﬁde is
vulnerable to attack by a thiolate group that is in close proximity, resulting in formation
of a different disulﬁde bond; such ‘scrambling’ or interchange of native and non-native
disulﬁdes can give rise to artifactual results. Frequently, the ﬂee sulfhydryls on ﬂee
cysteines are alkylated prior to proteolytic digestion to avoid scrambling; however, some

alkylation reactions themselves are conducted at high pH, e.g., the use of iodoacetamide

(37). During modiﬁcation of the ﬂee cysteines, it was found that the alkylating reagent
with the lower optimal working pH (such as maleimide derivatives) more effectively
suppresses disulﬁde scrambling; the more commonly used iodoacetyl derivative
alkylating reagent, which works at more alkaline conditions, is more liable to generate
scrambled artifacts. It was also found that addition of the alkylating reagent prior to the
denaturant helps prevent scrambling to some extent (34). Disulﬁde scrambling also can
occur during proteolytic digestion. For the same reason, using sample-processing
proteases having a low working pH may be better to minimize the scrambling. However,
proteases that function well at low pH such as pepsin and thennolysin, usually cleave the
substrate with less speciﬁcity, which may bring other problems to mass mapping. For
those more speciﬁc proteases, like trypsin and Glu-C endoproteinase having an alkaline
working pH, use of an immobilized protease cartridge for digestion can effectively
reduce disulﬁde interchange because the higher enzyme—to-substrate ratio reduces the

digestion time ﬂom 2-18 hours to 5 minutes (3 8).

3. An Experimental Protocol

(1) Dissolve the protein in a suitable buffer, add alkylating reagent followed by
denaturant.

(2) Adjust the pH to digestion conditions and incubate the mixture in a thermal bath.
If limited digestion is found, the following approaches might be helpful: a)

Increase the protease concentration. b) Choose a different protease or use a

combination of two or more proteases. c) Pre-treat the protein with cyanogen
bromide, if methionine residues are present in the protein.

(3) Use HPLC to separate the digestion mixture. A comparative HPLC run after
ﬂeeting the digestion mixture with an excess of reducing agent is helpful in
recognizing the disulﬁde-linked peptides.

(4) Use MS (MALDI-MS or ESI-MS) to analyze the HPLC ﬂactions; compare the
experimental data with the mass values calculated for expected digestion ﬂactions
of the protein. Absent ﬂagments plus unexpected ﬂagments (based on
consideration of the fully reduced protein) provide information relating to the
connectivity of disulﬁde cysteines involved in disulﬁde bonds.

(5) Reduce the disulﬁde-linked peptides and re-analyze them by mass spectrometry
or sequencing analysis by MALDI-PSD or ESI-collision induced dissociation

(CID)-MS/MS. These may help conﬁrm the identiﬁcation solely based on mass

mapping.

4. A Representative Application (39)
Testis angiotensin-converting enzyme (tACE) is a protein with 701 amino acid
residues including 7 cysteines, which were asserted by the authors to form three disulﬁde

bonds with one ﬂee cysteine as illustrated in the linear structure in Figure 1.4.

Putative disulﬁde structure of tACE:

r——r I—ﬁ r—i
——c — C c 0 KY C—K — C— c—
1 152 158 352 370 491492 496 511 538 550 701

 

Experimentally observed tryptic fragment :
492-511... incorporating derivatized cys 496

492 496 511 A, ,atio 492 496 511
Tyr— - --Cys- -- -- --Lys i1; Tyr— - --Cys- -- -- --Lys
| R=IAEDANS
SH SR
MH+ = 2145.4 03 MH+ = 2453.7 Da

Figure 1.4. Overall sequence of tACE showing tryptic cleavage sites at K491 and K511;
these deﬁne the origin of the tryptic ﬂagrnent Tyr492-Lys511 at the lower left that shifts
ﬂom 2145.4 Da to 2453.7 Da upon alkylation of the ﬂee sulﬂrydryl on Cys496. The
mass of the protonated alkylated-peptide (2453.7 Da) is very close to the observed peak
at m/z 2453.4 in Figure 1.6.

In general, a protein consisting of an odd number of cysteines will have at least
one ﬂee cysteine (maybe more) in its native form; a protein consisting of an even number
of cysteines might have no ﬂee cysteine. For a protein with one ﬂee cysteine and three
disulﬁde bonds, one must consider 105 different possible disulfide structures. (A
disulﬁde structure is deﬁned as a particular arrangement of all the disulﬁde bonds in a

protein. The formula for calculating the number of possible isomers for this case is as

follows:

M!
2'1 n!(M — 2n)!

 

or(M, n) =

where ot(M, n) represents the number of isomeric disulﬁde structures; 11 is the number of

disulﬁde bonds among M cysteines, and M_>_2n; see reference 40 for details.) All the 105

10

possible disulﬁde structures need to be considered when computing the expected masses
for all possible proteolytic peptide ﬂagments. Given the sequence of a protein and the
speciﬁcity of a protease, calculation of the masses of the expected proteolytic ﬂagrnents
can be automated; several websites (http://prosepctor.ucsf.edu, http://prowl.rockefeller.
edu, etc.) provide this service.

In an effort to simplify the assessment of cysteine status in a protein, the locus of
any ﬂee cysteines is usually determined ﬁrst. In the case of tACE, the protein was
digested by endoproteinase Lys-C after the ﬂee cysteine was modiﬁed by a ﬂuorescent
alkylating reagent, 5-([2-(iodoacetyl)amino]ethylamino) naphthalene-l-sulfonic acid
(IAEDANS) to facilitate detection of the alkylated cysteine-containing peptide. During
analysis of the alkylated digestion mixture by HPLC, because one of the detectors was
adjusted to 340 nm to measure the ﬂuorescence of the 5-([2-(acetyl)amino]ethylamino)
naphthalene-l—sulfonic acid (AEDANS) group, only a single peak was observed in the
chromatogram at the top of Figure 1.5. The ﬂaction corresponding to the single
ﬂuorescence peak was collected and analyzed by MALDI-MS. The corresponding mass
spectrum (Figure 1.6) of the detected ﬂaction (in the top panel of Figure 1.5) shows three
peaks indicating that the HPLC ﬂaction actually consists of at least three coeluting
components. Two of the mass spectral peaks are consistent with the calculated masses of
two of the expected proteolytic peptides (Gln308-Lys317 and Valll9-Lysl37). The
peak at m/z 245 3.4 represents the protonated peptide consisting of residues ﬂom Tyr492
to LysSl l, which after alkylation of Cys496 with the ﬂuorophore (AEDAN S) as
illustrated in Figure 1.4, shift to a mass of 2453.7 Da for (calculated mass of the

protonated species = 2453.7 Da). Sequencing analysis of this peptide by MALDI-PSD-

ll

MS also conﬁrmed the above identiﬁcation (data not shown). These data provide

convincing evidence that Cys496 in the native tACE is a ﬂee cysteine.

 

 

 

 

 

 

 

8

C

e

o

w

2

o l

2

IL 1-

7;

°‘ 0.04 A * r
E 0.03

e

E

3 0.02

C

«I

E

8 0.01

.c

<

O r . r I L J r . AL_1 __.r .A
5 10 15 20 25 30 35
Time (min)

Figure 1.5. HPLC chromatograms of the endoproteinase Lys-C digestion mixture of
AEDANS-labeled tACE as recorded by two different detectors. The upper panel was
monitored at 340 nm for the ﬂuorescence of the AEDAN S ﬂuorophore to selectively
detect the peptide with the AEDANS-modiﬁed cysteine; the lower panel was monitored
at 214 nm to detect all the digestion peptides (adopted ﬂom reference 39).

12

 

 

 

50000

40000 . Gln308-Ly8317
1307.4
3 30000. Val119-Lys137
g 2345.4
8 200001

[Tyr492-Lys51 1]+AEDANS
10000 1 2453.4

w I .L -- ”ELL

500 ‘- 1000 1500 2000 2500 3000 3500
ml:

 

 

 

 

 

Figure 1.6. The MALDI mass spectrum of the HPLC ﬂaction represented by the single peak
in the top panel of Figure 1.5. The peak at m/z 2453.4 represents the protonated proteolytic
ﬂagrnent containing the modiﬁed cysteine as illustrated in Figure 1.4 and described in the
text. The other two peaks represent other proteolytic peptides that coelute with the alkylated
one (adopted ﬂom reference 39).

Having identiﬁed the locus of the ﬂee cysteine in tACE, attention can now be
focused on determining the connectivity of the other six disulﬁde cysteines to form three
cystines (recall that there are 15 isomeric possibilities; see also in Figure 1.7). It is
desirable to degrade the protein by cleaving the peptide backbone at least once between
every two disulﬁde cysteine residues. In this case, alkylated tACE was digested with
cyanogen bromide and, subsequently, further digested with endoproteinase Lys-C.
Cyanogen bromide (CNBr) is a chemical reagent that cuts the peptide backbone
speciﬁcally at the C-terminal side of methionine; because of its small size, CNBr can
readily penetrate to interior cleavage sites that sometimes are less accessible to proteases

of much larger size. There are 19 methionine residues in tACE; a perfect CNBr digestion

with no miscleavages will generate 20 ﬂagments if no disulﬁde bonds link them together.

13

As shown at the top of Figure 1.7, three of the CNBr ﬂagments contain cysteines; two of
these ﬂagments contain two cysteines each and one contains three cysteines, including
Cys496, the alkylated (AEDAN S) cysteine. The disulﬁde cysteines in those three
ﬂagments can form disulﬁde bonds in a variety of ways as shown by three categories in
the lower portion of Figure 1.7. In category (i), each of these of the three pairs of
disulﬁde cysteines forms an intra-molecular disulﬁde bond. In category (ii), two pairs of
the disulﬁde cysteines form two inter-molecular disulﬁdes; there are six possibilities. In
category (iii), all three CNBr fragments are linked by three inter-molecular disulﬁde
bonds; there are eight possible isomeric arrangements. Fortunately for the authors, the
MALDI-MS mass spectral data showed peaks corresponding to the three separate intra-
molecular disulﬁde peptides represented in category (i), indicating that the three intra-

molecular disulﬁde bonds in tACE are isolated or do not overlap one another.

14

Three cysteine-containing fragments from CNBr digestion
of tACE without regard for cysteine status:

AEDANS
—c—c— ——c—c— — —C—C-
152 158 352 370 496 538 550

Possible disulfide structures
(without consideration of cys 496 as it's alkylated):

(i) —c— — (iil—lc\::[— (iii) e/ \o
— _ '— 0

“1:21" ’\:\ /°\
"113'— "1:1—

Figure 1.7. Conceptual representations of the disulﬁde arrangement among the six
cysteines in the three cysteine-containing ﬂagments from CNBr digestion of tACE.
Category (i) shows the simplest arrangement, which derives ﬂom three isolated disulﬁdes
in tACE; experimental results indicate that this category represents the disulﬁde
arrangement in native tACE. Categories (ii) and (iii) consist of six and eight isomeric
arrangements, respectively, some of which involve overlapping disulﬁde bonds
(indicated, in part for two possibilities, by dashed lines).

Mass spectral data (not shown) ﬂom analysis of the reaction mixture ﬂom CNBr
digestion of AEDANS-alkylated tACE include peaks that correspond to the calculated
masses of the three cysteine-containing fragments in Figure 1.7. While these data
strongly suggest three isolated disulﬁdes between the pairs of cysteines within the three
ﬂagments at the top of Figure 1.7., it is prudent to seek conﬁrmatory evidence. The
following text describes an additional experiment to conﬁrm the putative connectivity in
the CNBr digestion ﬂagrnent Ala451-Met566, which contains the alkylated cysteine, as
illustrated in Figure 1.9. Subsequent digestion of this ﬂagrnent (the 2nd structure in

Figure 1.9) by trypsin generates a ﬂagrnent (the 3rd structure in Figure 1.9) consisting of

two peptides linked by a disulﬁde (Cys538-Cys550) and having a calculated mass (for

15

singly protonated species MB”) of 5243.1 Da. Analysis of an HPLC ﬂaction (not
shown) of the trypsin digest of the putative CNBr ﬂagrnent (the 2nd structure in Figure
1.9) by MALDI-MS produced the mass spectrum in Figure 1.8. The peak at m/z 5242.3
in Figure 1.8 corresponds well to the calculated mass of the protonated ﬂagment (5243.1
Da). Furthermore, prompt ﬂagrnentation (also called in-source ﬂagrnentation, see
references 24-28) of the disulﬁde bond in this ﬂagrnent (bottom of Figure 1.9) during
MALDI-MS analysis provided evidence (peaks at m/z 856.6 and m/z 4388.0) for the
individual peptides. Overall, these results provide convincing evidence for the structure

of tACE at the top of Figure 1.9.

 

 

 

 

+
.. arr.
FHIPSSVPYIRYFVSFIIQFQFHEALCIZQAAGHTGPLHK '

CDIYQSK
150001 550

3

C

3 [M+2H]+

4388.0
500013.535 L - 1.. 3..

 

 

1000 2000 3000 4000 5000 6000
m/z

Figure 1.8. The MALDI mass spectrum of the putative peptide of a CNBr ﬂagrnent (see
Figure 1.9) of the tACE protein. Peaks at m/z 856.6 and m/z 4388.0 correspond to the
individual peptides connected by the disulﬁde that was cleaved during the MALDI
process (adopted ﬂom reference 39).

16

Putative Disulﬁde Structure of the Alkylated tACE:

AEDANS
—C — cl ——c—c——MA——L —KF— C—KC—K—M
1 152 158 352 370 450 451496511512 538 549.550 556 5MSS 701

1\/

 

Cyanogen Bromide Cleavage/HPLC
AEDANS

A—c -KF—C—KC-—K—M
451 496511512 538 549550 556 566

KM

Trypsin Cleavage/HPLC
511 538 549
Tyr—- -- --Cys- - - --Lys
H+ W Laser
Cys- - - --Lys
550 556
MW: 5243.1 Da
Prompt Fragmentation during
analysis by MALDI-MS
511 538 549 550 556
H+ Tyr— -- --Cys- - - --Lys HT Cys- -- - --Lys
MH+3 4388.103 MH+Z 856.0 Da
+ or +
S s
H 556 511 n 549
Cys- “ '- --Lys Tyr-- --Cys- - - --Lys
550 538

Neutral Peptided with Dehydrated Cysteines
Figure 1.9. Conceptual illustration of the digestion of tACE by CNBr and trypsin to

produce two peptides linked by a disulﬁde bond, the analysis of which by MALDI-MS
identiﬁes the disulﬁde CysS38-Cy8550.

17

5. Advanced Topics: Using MALDI-PSD-MS and ESI-CID-MS/MS to Obtain
Additional Information on Disulfrde—Linked Peptides

Fragmentation of the protonated-disulﬁde-linked peptides by metastable decay
during analysis by MALDI-PSD-MS (17, 27-30) or by collision-induced dissociation
(CID) during ESI-MS/MS (31-35) sometimes provides sequence information on
disulﬁde-linked proteolytic peptides. This second dimension of mass spectrometric
information can be used for conﬁrmation of the identity of some disulﬁde-linked
peptides.

The ﬁrst example here involves the analysis by MALDI-PSD-MS of a digestion
product consisting of three peptides linked by two disulﬁde bonds as shown by the
structure insert in the mass spectrum in Figure 1.10. The proteolytic ﬂagrnent of interest
here derives ﬂom the thermolysin digestion of a tryptic ﬂagrnent of the recombinant
human tumor necrosis factor binding protein (TNFbp in reference 28), a cysteine-rich
protein with 13 disulﬁdes among 26 cysteines across 162 residues. In this special case,
there are two possible disulﬁde structures; the alternative to the structure Shown in Figure
1.10 would involve switching positions of linking the tetrapeptide and the pentapeptide to
the central peptide. During analysis of the disulﬁde-linked peptides by MALDI-PSD-
MS, it is likely that some of the disulﬁde bonds as well as some of the peptide bonds will
ﬂagrnent. While the peaks at m/z 242.9 and m/z 1350 in the mass spectrum (Figure 1.10)
do not distinguish the two isomeric disulﬁde structures under consideration, the peaks at

m/z 825.6, 940, and 1027 uniquely correspond to the structure shown in Figure 1.10.

18

 

 

2X —1

q-

 

 

 

 

 

 

 

 

 

100. 359 243
: Leu-CysjEeuIPro-Gln
; 1350r-l-
90- S
; 242.9 1318r-l-
'1—1'1027 940
801 1284|—+ [- I— r826
. Val-Ser-CysS er Asn Cys- -Lys-Lys
701 +—|
1 ? 453
3.. 601 Ser-Leu-Glu Cys
'2 1 1560
2 1
.5 50- *
‘3 I
E 40:
30: co, m (D g 940 1350
l {’3 3 '3 v 8 *
l to 00 1027
20: I// g at, 29. 1560 1618
. I» N 32 \ / .
, to ‘- ,1
10.‘ \ \ 1?
J ,
01
200 600 1000 1400 1800
m/z

Figure 1.10. The MALDI-PSD-MS mass spectrum of a disulﬁde-linked proteolytic ﬂagrnent
ﬂom dual digestion of the TNFbp protein with trypsin and thermolysin. The numbers by the
structure give the mass of the indicated ﬂagrnent ions; the peaks marked by an asterisk at m/z
825.6, 940, and 1027 conﬁrm the disulﬁde structure shown (adopted ﬂom reference 28).

19

Cysteines occur in a protein less ﬂequently on average than those residues serving
as proteolytic cleavage sites. If cysteines are somewhat evenly distributed within a
protein (the above example is an exception to this assertion, as the cysteine content of
TNFbp protein is abnormally high), there is often more than one cleavage site between
the cysteines. As a result, a less complicated situation involving a disulﬁde-linked
proteolytic ﬂagrnent consisting of only two peptides linked by one disulﬁde is usually
observed. An example of such a structure is inserted into the ESI-CID-MS/MS mass
spectrum (Figure 1.11) of a disulﬁde-linked tryptic ﬂagrnent of beta-1,4-
galactosyltransferase (34). In describing their mass spectral data, the authors used a
hybridization of the Roepstorff et al. (41) and Biemann (42) nomenclature to designate
ions ﬂom the two peptides linked by a disulﬁde bond. The peaks labeled with an upper
case Y denote the y-ions ﬂom the ﬂagrnentation of the lower peptide; peaks with the
lower case y and b denote the y- and b-series of ions ﬂom fragmentation of the upper

peptide. (Y11b12)2+ denotes a doubly charged b ion of the upper peptide with 12 residues
ﬂom the N—terminus that was linked by a disulﬁde bond to the Yll ions ﬂom the lower
peptide ( with 11 residues). Most of the ﬂagrnentation occurred on the C-terminal side
of the cysteine (residue 5) in the upper peptide, namely, the yn series and Yllbn series of
ions. Only the ions [(Y 11b1,)“ and (Y 9b13)2+ ions] containing the intermolecular

disulﬁde provide information on the disulﬁde connectivity. The mass of the lower

peptide can be deduced by subtracting the mass of the bn series of ions (to the right of
Cys5) of the upper peptide ﬂom the mass of the bnYll ions as observed in the

experiment. Ions such as Y9b13 are not observed often, possibly because they require

ﬂagmentation of two peptide bonds.

20

|
S

II-TLP-M-D-C-IIS-P-H-K

b13I: b:22 b25
S-L-T-A-C-P-E-E-SIP L bﬁi—Ifm. PIE/1w --L 21- E- ngl 1IIPIZIDILK
s
Y14

 

 

 

 

 

 

Y11 Y9 Y4
(1’11”13)2+
100— 1315.4
‘ “11914)2+ 2,,
~ Y b )
80_ 1343.8 ( 11 22
Q; _ Y14 1823.9
a. - 1585.7
2" ‘ ("11912)2+ (Y b )2+
'6 60-1 1265.9 2+ 11 24
E : Y b 2“ 1528.7 1756-0
«I - ( 11 9) ”11925)2+
> 40 1061.1
.3 _ 2+ 1978.8
g - y5 “9913) “5 /
I! ' 571.4 Y8 1183.5 1685-8
20- 945.6 \
- Y4
‘4682 ;
Oj . . 1': J l 1 , . . .
400 800 1200 1600 2000
m/z

Figure 1.11. The ESI-CID-MS/MS mass spectrum of a triply charged ion [M+3H+]3+ at m/z
1406.1 corresponding to a disulﬁde-linked tryptic ﬂagrnent (in which a disulﬁde bond links
two peptides, S130-K156 and Y171-K181) of bovine beta-1,4-galactosyltransferase (adopted
ﬂom reference 34).

It is important to note the following observations about sequencing analysis of

disulﬁde-linked peptides by MALDI-PSD-MS and ESI-CID-MS/MS. First, although

several successful examples by both techniques had been shown (17, 27-35), both

21

techniques failed to provide diagnostic information in other cases (28). In addition, more
diverse disulﬁde-linked peptides need to be studied by both techniques to optimize the
experimental parameters and determine the limitations of the techniques. For example,
while intermolecular disulﬁde bonds in the disulﬁde-linked peptides are usually
ﬂagrnented during analysis by MALDI-PSD-MS, ﬂagrnentation of disulﬁde bonds does
not always occur during analysis by ESI-CID-MSfMS of disulﬁde-linked peptides. No
indication of the ﬂagrnentation of the intermolecular disulﬁde bond, either symmetrical
or non-symmetrical was found in the example described above and several other
examples (34, 35); whereas ﬂagrnentation of such disulﬁde bonds in other disulﬁde-
linked peptides during ESI-CID-MS/MS was observed (27-32, 43). It seems that what
controls the ﬂagrnentation at various sites of the disulﬁde-linked peptides would be the

key to solve this problem.

6. Conclusions

The proteolytic approach to disulﬁde mapping has been used for many years, ﬁrst
with only chromatographic and sequencing analyses, but these have been combined with
analyses by mass spectrometry in recent applications. The proteolytic approach can be
tirne- and labor-intensive primarily because of the effort in recognizing and isolating the
cysteinyl and cystinyl peptides ﬂorn the often-large number of other proteolytic peptides
generated upon digestion of a large protein.

Disulﬁde mapping of proteins containing adjacent cysteines is impractical, if not

impossible, by the proteolytic approach. Proteins containing nearby cysteines in their

22

amino acid sequence may also present intractable problems for the proteolytic approach
because of the lack of a trustworthy proteolytic site between the cysteine residues.
Concerns about disulﬁde scrambling are always an issue in the proteolytic
approach to disulﬁde mapping because of the susceptibility of an existing cystine to
attack by a thiolate ion produced ﬂom signiﬁcant dissociation of any cysteine under the
slightly alkaline conditions required for the enzymatic digestion. Alkylation of ﬂee
cysteines is effective in some case in suppressing artifact formation due to disulﬁde
scrambling; thiolate-disulﬁde exchange was observed during some of the thiolate

alkylations under alkaline pH (9).

III. Determination of Disulﬂde Bonds by Partial Reduction, Chemical Tagging,

Edman Sequencing Analysis and Mass Spectrometry

1. Overview of the Method

Free cysteines in the native protein are modiﬁed with a particular alkylating
reagent, e.g., with iodoacetamide, to prevent disulﬁde scrambling. As the multiple
disulﬁde bonds are progressively reduced by partial reduction with t1is-(2-carboxyethyl)-
phosphine (TCEP) hydrochloride, the nascent sulﬂrydryls are modiﬁed with different
alkylating reagents to chemically tag the products at different stages of reduction. The
modiﬁed cysteines produce distinguishing responses during separation of the
phenylthiohydantoin (PTH) derivatives by HPLC following Edman degradation. Thus,
this analytical strategy permits recognition of ﬂee cysteines in the original protein as well

as the locus of disulﬁde cysteines that were coupled into speciﬁc disulﬁde bonds.

23

Partial reduction of the protein is accomplished with an excess amount of
reducing agent, but under kinetically limiting conditions (6, 7). In this way, products
isolated ﬂom the reaction mixture at an early stage will consist principally of one or more
of several different singly reduced isoforrns of the protein (e. g., a protein containing four
disulﬁdes can produce as many as four singly reduced isoforrns). At a later stage during
the dynamic process of partial reduction, doubly reduced isoforrns (e.g., a protein
containing four disulﬁdes can produce as many as six doubly reduced isoforrns) as well
as triply reduced isoforrns will appear in the reaction mixture.

This partial reduction/chemical modiﬁcation is especially advantageous in the
analysis of small cysteine-rich proteins, which do not respond well to proteolytic
digestion due to inaccessibility of the protease to sites that are in the closely packed core
region. For proteins consisting of fewer than 50 residues, Edman degradation performed
on the protein after the cysteines have been modiﬁed by different alkylating reagents
provides information on the cysteine status (ﬂee cysteines or disulﬁde cysteines involved
in disulﬁde bonds) (reference 44). Mass spectrometry is helpful in verifying that proper
alkylation has been performed after appropriate stages of partial reduction of the disulﬁde
bonds in the protein (9, 12-14). For proteins consisting of more than 50 residues,
partially reduced/alkylated species of the protein are usually cleaved into smaller sized
ﬂagments by proteolyses to allow different “tags” on various alkylated cysteines residues

to be identiﬁed by Edman sequencing and mass spectrometry (12-14).

24

2. Constraints/Caveats

Because different proteins have different disulﬁde structures, the reduction yields
vary accordingly. Some proteins can be converted into some of each of the possible
singly reduced isoforrns because all the disulﬁdes are accessible to the reducing agent.
For other proteins, however, only certain disulﬁdes are readily accessible to the reducing
agent. In such a case, a stepwise reduction/alkylation approach may be useful in which
the available singly reduced isoforrns of the protein are alkylated before the reducing
conditions are strengthened to force some of the protein to the doubly reduced stage (8).
Different alkylating reagents can be used after different stages of reduction, thereby
chemically ‘tagging’ speciﬁc cysteines as a function of time during the dynamic
reduction process (8, 44). The temporal relationship of certain cysteines as manifested
via chemical 'tagging' can provide a basis for deducing the connectivity of cysteines
involved in a particular disulﬁde bond.

Many widely used alkylating reagents work under slightly alkaline conditions that
bring about the possibility of disulﬁde scrarrrbling. Usually, a reactive reagent like
iodoacetamide is preferred over a less reactive reagent such as iodoacetic acid (7).
Nonetheless, scrambling is sometimes observed if the kinetics of the disulﬁde exchange
(scrambling) reaction is faster than that of the alkylation reaction. In those cases, an
alkylating reagent, such as N-ethylmaleimide (NEM), that can be used under acid
conditions is preferred (44). Other complications may arise during the use of alkylating
reagents, such as the formation of isomeric products that give multiple peaks during
HPLC separation as has been reported for N-ethylrnaleimide, a problem related to the

prochiral nature of the reagent (44).

25

Sometimes there is a problem in separating the partial reduction products by
HPLC, due to their being structurally similar. While coelution of structurally similar
partially reduced species can complicate data interpretation, the involvement of mass

spectrometry can provide useful information regarding the disulﬁde status (9, 12-14).

3. An Experimental Protocol

(1) Dissolve the protein in the appropriate buffer to alkylate any ﬂee cysteines.

(2) Use the reducing agent (TCEP) to partially reduce the protein. The amount of
reagent and the incubation time as well as the temperature should be controlled to
optimize the yield of the partially reduced products. Alkylate the products with a
alkylating reagent.

(3) Subject the mixture to HPLC separation. If a poor yield is observed; repeat step
(2) with stronger reducing conditions.

(4) Use a different alkylating reagent at each progressive stage of reduction/
alkylation.

(5) To each HPLC ﬂaction, add an excess of reducing agent to totally reduce the
residual disulﬁde bonds. Alkylate the nascent ﬂee cysteines (ﬂom complete
reduction of the residual disulﬁdes) with a unique alkylating reagent.

(6) If the protein is small enough, use Edman degradation to sequence the whole
partially reduced/alkylated protein species. If the protein is large (>50 residues), a
combination of proteolytic mapping and Edman degradation is preferred.
Because there is no disulﬁde left at this stage, there is little resistance to the

proteolytic digestion.

26

4. A Representative Application (44)

The AVR9 elicitor protein ﬂom Cladosporium fulvum has three disulﬁde bonds
across the 28-residue structure (44). The AVR9 protein was partially reduced with TCEP
and the nascent ﬂee cysteines were modiﬁed with N-ethylmaleimide (NEM) under acidic
conditions. The reaction mixture was then separated by HPLC; major ﬂactions were
collected and subjected to complete reduction (of residual disulﬁdes)/alkylation with 4-
vinyl pyridine (4-VP). This second alkylating reagent (4-VP) tagged the cysteines in
disulﬁdes that survived the partial reduction procedure, while the ﬁrst reagent (NEM)
tagged the cysteines ﬂom disulﬁdes that had been reduced during partial reduction. In
this way, the nascent cysteines in the singly or doubly reduced species (tagged as NEM)
could be distinguished ﬂom those cysteines (eventually tagged as 4-VP) remaining as
disulﬁdes after the initial step involving partial reduction.

The doubly modiﬁed protein was then subjected to Edman sequencing directly due to its
small size. The retention times for PTH-NEM-modiﬁed cysteine and PTH-4-VP-
modiﬁed cysteine in HPLC are different ﬂom one another as well as different ﬂom that
for PT H-unmodiﬁed cysteine; thus, the sequencing results of the doubly modiﬁed protein
provide a temporal record of its exposure to reduction/alkylation. The chemical tagging
of various cysteines in the sequence allows the connectivity in disulﬁde bonds to be
deduced because certain pairs of cysteines became available for Simultaneous alkylation.
For example, during the second cycle of Edman sequencing of the partially
reduced/alkylated species of AVR9 represented by HPLC peak 2 in Figure 1.12, a peak
having the retention time of PTH-NEM-cysteine was observed. As Edman degradation

progressed to cycle 6, a 4-VP—cysteine was observed, and so forth. As Edman

27

degradation progressed to cycle 6, a 4-VP-cysteine was observed, and so forth. Thus, the
diagram of sequencing results for the species represented by the HPLC peak 2 (in Figure
1.12) indicate that it is a singly reduced species, ﬂom which it can be deduced that AVR9

contains a disulﬁde bond between CysZ and Cysl 6.

2 NEM 2 NEM 2 NEM
6 4VP 6 4VP 6 NEM
12 4VP 12 NEM 12 4VP
16 NEM 16 NEM 16 NEM
19 4VP 1 9 4VP 1 9 NEM
26 4VP 26 NEM 26 4VP

naﬁve
AVR9

J

 

 

L \ll\i II M“

I I I I I I I

10 15 20 25 30 35 40
HPLC elution time (min)

EEP/NEM I

 

 

Figure 1.12. A diagrammatic presentation of sequencing results for the partial
reduction/alkylated species of AVR9 represented by the peaks in the HPLC
chromatogram. Peak 2 represents a singly reduced speceis ﬂom partial reduction and
indicates that AVR9 contains a disulﬁde bond between Cys2-Cysl6. Peak 3 represents a
doubly reduced isoform containing a residual disulﬁde bond between Cys6-Cys19 before
complete reduction; peak 4 represents another doubly reduced isoform containing a
residual disulﬁde Cyle-Cys26 before complete reduction. Peaks 5 and 6 represent the
completely reduced/NEM-modiﬁed species (as determined by mass spectrometry); both
species have the same mass and apparently result ﬂom epimeric forms of the prochiral
alkylating reagent (adapted ﬂom reference 44).

28

From the sequencing results for the species represented by the HPLC peaks 3 and
4 (in Figure 1.12), it can be deduced that these are two doubly reduced isoforrns. Thus,
each of the three ﬂactions represents a different disulﬁde, thereby identifying the three

disulﬁdes as Cys2-Cysl6, Cys6-Cys19, and Cys12-Cys26.

5. Conclusions

The partial reduction/alkylation approach is suitable for disulﬁde assignment in
small cysteine-rich proteins. For larger proteins, this approach will require more steps
than either the proteolytic approach or the partial reduction/cyanylation approach that

will be discussed in the next section.

IV. Disulfide Mass Mapping Based on Partial Reduction, Cyanylation and CN-

induced Cleavage

1. Overview of the Method

This novel approach to disulﬁde mapping is based on chemical cleavage of the
peptide backbone on the N-tenninal side of cyanylated cysteines (Figure 1.13). The ﬁrst
step in the process is cyanylation of ﬂee cysteines by a suitable reagent, e.g., 1-cyano-4-
dimethylamino-pyridium (CDAP) tetraﬂuoroborate (45), as Shown in Figure 1.13. The
cyanylated cysteine is susceptible to nucleophilic attack, e.g., by ammonia, as shown in
the second chemical reaction illustrated in Figure 1.13, producing a truncated peptide and

an irrrinothiazolidine (itz)-carboxyl-blocked peptide (46).

29

SH
H2 0 o

o
H N+ —-_U—NH— H_E_NH_-_ //
3 R0-

(A) Cyanylation +
pH3—7; 1'1, 15, CDAP, MezN ON—CN

 

 

BF4
I!
NEC—s
0 H
I i 2 1 4°
H3N+ —-— —NH— H— -NH—--—c
O—
'NH3
(B) Cleavage
1 M NH40H,
~1h, rt.
I
O HN S
/ O
H3N+—-—C/\ + Y I || //0
NH2 HN C—NH —- -C\
O‘
peptide itz-peptide
0 CH 0
II 2
+ (H3N+—-—C-NH—(li— —NH—'—C< )
O-

B-elimination peptide
Figure 1.13. (A) Cyanylation of sulﬂrydryl group by CDAP and (B) peptide bond

cleavage by ammonia. (itz = iminothiazolidine—carboxyl; B-elimination = loss of HSCN
ﬂom a cyanylated cysteine residue.) (adopted ﬂom reference 47)

30

The cyanylation reaction is speciﬁc for cysteine; it does not modify methionine,
nor does it react with hydroxylated, nor acidic nor basic side chains under analytically
useﬁrl conditions (47). Furthermore, the cyanylating reagent will not react with a cystine
(disulﬁde bond)! How, then, can the cyanylation reaction be used in disulﬁde mapping?
It is ﬁrst necessary to reduce a particular disulﬁde bond, then cyanylate the nascent ﬂee
cysteines to promote cleavage at the sites previously tied up as a cystine. (Free cysteines
in the native protein should be alkylated as described earlier in the previous section, or by
use of the CDAP methodology prior to partially reducing the cystines as described
below.) The analytical challenge comes with proteins containing more than one cystine,
in which case the process of partial reduction (with TCEP) is used to generate a singly
reduced isoform corresponding to each cystine (16).

Partial reduction (6, 7, and 16) is achieved under kinetically limiting reduction
conditions using a phosphine-derived reducing reagent (e.g., tris-carboxyethyl-phosphine,
TCEP). The concept of partial reduction is illustrated in Figure 1.14 with a hypothetical
protein containing two cystines. The cyanylation/CN-induced cleavage chemistry is
applied to each singly reduced isoform. Note that late in Figure 1.14, two or more of the
expected three CN-induced cleavage products of the cyanylated protein might still be
connected by residual disulﬁde bonds; complete reduction at this stage produces three
individual ﬂagments. Also note at the bottom of Figure 1.14 that each cystine is ideally
represented by a unique set of ﬂagments, the composition of which can be anticipated
because the sequence of the original protein is known. For some disulﬁde proteins,
doubly reduced isoforrns, in addition to the singly reduced isoforms, may be utilized to

determine their disulﬁde structure (reference 48, Chapter 2, and Chapter 3). Guidelines

31

for optimizing the partial reduction of disulﬁde proteins was discussed in references 6, 7,

l6 and 48.

 

 

 

 

 

, E —2:—so 40
:IPartial50 reduction; TCEP @ pH 3
SH SH
1 20 r #) F—E 50
10 I 30 S 30 SH
:ICyanyl:tion;:: CDAP @ pH 3
fCN 20 SCN 4190 s—s 40
50
10 l 30 S 30 SCN
SOIHPLC fractionation
|——> MS analysis 1
CN- induced Cleavage
in 1M NH4OH
1— 9110—19
itz 10 2° 29 \T
—i_ itz 20—35— 39

S\
I12 ISO—+50 "Z 40

——> MS analysis 2

 

50

 

Complete Reduction;
TCEP @ pH 3
H
1 9 +_.E__19
_I‘_”_ ‘° .1
"2 3° 40 5° itz 20+3e
H

IMS analysis 3

(CystO—Cys30) (Cyszo-Cys40)

Figure 1.14. Chemical overview of the partial reduction/cyanylation/CN—induced
cleavage methodology using singly reduced isoforrns of a hypothetical 2-disulﬁde
protein. “SCN” designates a cyanylated cysteine residue. (adopted ﬂom reference 16).

32

The individual partially reduced species of the protein can be isolated (either
before, but preferably after cyanylation) usually by reversed phase (rp)-HPLC. In
principle, such an isolation of the partially reduced species is not necessary, but doing so
can simplify data interpretation (Chapter 5). Furthermore, such a separation leads to less
complex mixtures of cleavage products, thereby minimizing the possibility of signal

suppression for some components during analysis by MALDI or ESI mass spectrometry.

2. Constraints, Caveats, and Other Aspects

In addition to, and in competition with, the desired cleavage reaction illustrated at
the bottom of Figure 1.13, the cyanylated cysteine can undergo B-elimination expelling
HSCN producing a dehydroalanine peptide as illustrated. Observing the B-elimination
product provides a ‘parity check’ on mass spectral Signals by representing the sum of all
residues in the expected cleavage ﬂagments for a given cyanylated cysteine, a useﬁrl
accounting feature when two or more cyanylated cysteines are present. The extent of [3-
elimination ranges ﬂom a minor to a major process depending mainly on the type of
amino acid residue on the N-terminal side of the cyanylated cysteine (47); when proline
is on the N-terminal side of a cyanylated cysteine, B-elimination is the dominant process.

Other side reactions, such as incomplete cyanylation and incomplete CN-induced
cleavage, could occur and may complicate data processing (48). How to deal with such
side reactions is discussed in reference 48 and Chapter 3.

In the case where there is an odd number of cysteines or where all the cysteines
are not involved in cystines, the positions of the ﬂee cysteines in the original protein can

be determined by immediate cyanylation and mass analysis of CN-induced cleavage

33

products after complete reduction of the residual disulﬁde bonds, which may still link the
cleavage products together (49). If only two disulﬁde cysteines are involved as a cystine,
the connectivity of the involved disulﬁde cysteines can be deduced by default. If four or
more disulﬁde cysteines are involved as cystines, the partial reduction scheme can be
invoked as described above. Alternatively, the initially ﬂee cysteines can be marked by
alkylation as described earlier (37) (a process that renders them unavailable to the
subsequent cyanylation reaction).

The strategy for determining the connectivity of disulﬁde cysteines involved in an
interrrrolecular cystine is the same as that pursued in analysis of intramolecular cystines.
That is, given the sequence of the polypeptide chains, one would use initial cyanylation
or alkylation (if ﬂee cysteines are present), followed by the use of complete or partial
reduction (depending on the number of cystines) and cyanylation/CN-induced
cleavage/mass mapping.

The classic problem of adjacent cysteines is amenable to the cyanylation/cleavage
approach (18), thereby providing a viable opportunity to solve disulﬁde structures that

have hitherto been reﬂactory to the proteolytic approach.

34

3. Experimental Protocols

A. A Protocol for Determining the Location of Free Cysteines by Alkylation/

Cyanylation/CN-Induced Cleavage

(1) Dissolve the protein in an appropriate buffer and mix it with alkylating reagent to
alkylate (37) any free cysteines. Remove the reagents by using rp-Zip tips or rp-
HPLC.

(2) Completely reduce the disulﬁde bonds in the alkylated protein with TCEP (49).

(3) Cyanylate the nascent ﬂee cysteines in the alkylated protein with CDAP (47).
Because CDAP reacts with TCEP, an excess amount of CDAP is needed.

(4) Purify the product by rp—HPLC; collect and dry the HPLC ﬂactions under
vacuum.

(5) Reconstitute a solution of the product in l M aqueous ammonia to cleave the
peptide bond on the N-terminal side of the cyanylated cysteines (47).

(6) Analyze the mixture by MALDI-MS or liquid chromatography (LC)-MS using

ESI.

B. A Protocol for Determining the Location of Free Cysteines by Cyanylation/CN-
Induced Cleavage
(1) Dissolve the protein in an appropriate buffer and mix it with CDAP to cyanylate
(47) any ﬂee cysteines.
(2) Purify the cyanylated protein by rp-I-IPLC; collect and dry the HPLC ﬂactions

under vacuum.

35

(3) Reconstitute a solution of the cyanylated protein in 1 M aqueous ammonia to
cleave the peptide bond on the N-terminal side of the cyanylated cysteines (47).

(4) Completely reduce any residual disulﬁde bonds in the cleaved protein with TCEP
under strong reducing conditions (49).

(5) Analyze the mixture by MALDI-MS or LC-MS using ESI.

C. A Protocol for Determining the Location of Disulfide Bonds by Partial

Reduction/ Cyanylation/CN-Induced Cleavage

(1) Dissolve the protein in an appropriate buffer and mix it with an appropriate
reagent to alkylate (37) or cyanylate (47) any ﬂee cysteines. Remove the reagents
using rp-Zip-tips or rp-HPLC.

(2) Use the reducing agent (TCEP) to partially reduce the alkylated or cyanylated
protein. The amount of TCEP and the incubation time as well as the temperature
should be adjusted to optimize the yield of the partially reduced products (6, 7, l6
and 48).

(3) Cyanylate the products with CDAP (47). Because CDAP reacts with TCEP, an
excess of CDAP is needed.

(4) Separate the mixture by rp-HPLC. If a poor yield is observed, repeat step (2) with
stronger reducing conditions. Collect and dry the HPLC ﬂactions.

(5) Reconstitute a solution of the cyanylated protein species in 1 M aqueous ammonia
to cleave the peptide bond on the N-terminal side of the cyanylated cysteines (47).

(6) Completely reduce any residual disulﬁde bonds in the cleaved protein mixtures

with TCEP under strong reducing conditions (16, 49).

36

(7) Analyze the mixture by MALDI-MS or LC-MS using ESI.

4. A Representative Application: Disulfide Mass Mapping Using Singly

Reduced/Cyanylated Isoforms (l6)

Bovine ribonuclease A (molecular mass for MH+ = 13,683 D8) (50) contains 124
amino acids including eight disulﬁde cysteines that are linked by four disulﬁde bonds:
Cy326-Cys84, Cys40-Cys95, Cys58-Cysl 10, and Cys65-Cys72.

When ribonuclease A (RnaseA) was subjected to the partial reduction/CN-
cleavage/mass mapping methodology under conditions in reference 16, essentially equal
amounts of the four possible singly reduced/cyanylated isoforrns were produced. Figure
1.15 is the chromatogram of the residual intact RnaseA and its singly reduced/cyanylated
isoforrns, which resulted ﬂom partial reduction and cyanylation reactions of RnaseA.
Only one equivalent of TCEP relative to the total cystine content in RnaseA was used
during partial reduction for 15 minutes at room temperature; analysis by MALDI-MS
showed that each of the four products produced (represented by HPLC peaks 1-4 in
Figure 1.15) had a molecular weight 52 Da higher than the intact RnaseA (2 Da for
reducing a cystine plus 25 Da net shift of each of the two nascent ﬂee cysteines upon
cyanylation; see Figure 1.13.) as expected for singly reduced/cyanylated isoforms of
RnaseA.

Table 1.1 lists the calculated m/z values for possible protonated ﬂagments (MH+)
due to CN-induced cleavage of the peptide chains at different sites depending on which
disulﬁde bond was reduced and cyanylated in forming the singly reduced/cyanylated

isoform of ribonuclease-A.

37

 

J

30200 ' ' 10100 ' ' 50100 ' ' 50100 ‘ ' 70300 ' ' 80200 '
Time(mln)

 

Figure 1.15. The HPLC chromatogram of the denatured ribonuclease-A and its singly
reduced/cyanylated isoforrns. The separation was done by reversed-phase HPLC on a
Vydac C18 column (5pm, 4.6mmx250mm) 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
90% CH3CN. The peak marked by IP is the intact protein (ribonuclease-A). Peaks 1-4
represent singly reduced/cyanylated ribonuclease-A isoforms, as determined by MALDI-
TOF-MS (adopted ﬂom reference 16).

38

Table 1.1. Calculated m/z values of protonated CN-induced cleavage ﬂagrnents ﬂom all
possible singly reduced/cyanylated isoforrns of an unknown protein having the same
sequence as RnaseA. The boldface entries correspond to native RnaseA, one of the 105
possible isomeric arrangements of the four cystines in a protein with this sequence.

 

 

 

No. Cys i — Cys j MH+[1-(i-l)] MH+[itz-i-(j-l)] MH+[itz-j-124]
T— W1 __-__.___-_. 'CyS26iCys40 ., 2704.87-W WW W 1750.05 9322.44

2 Cys26—Cys5 8 2704.87 3689.27 7383.23
3 Cy526-Cys65 2704.87 4420.1 6652.39
4 Cys26-Cys72 2704.87 5165.91 5906.58
5 CysZ6—Cy884 2704.87 6546.42 4526.08
6 Cys26-Cys95 2704.87 7769.74 3302.75
7 CysZ6-Cysl 10 2704.87 9412.68 1659.81
8 Cys40-Cys58 4411.89 1982.25 7383.23
9 Cys40—Cys65 441 1.89 2713.09 6652.39
10 Cys40-Cys72 441 1.89 3458.89 5906.58
1 1 Cys40-Cy584 441 1.89 4839.4 4526.08
12 Cys40-Cys95 441 1.89 6062.73 3302.75
13 Cys40-Cys110 4411.89 7705.66 1659.81
14 CysS8-Cys65 6351.1 1 773.87 6652.39
15 Cys58-Cys72 6351.1 1 1519.68 5906.5 8
l6 Cy858-Cys84 6351.1 1 2900.18 4526.08
17 Cys58-Cys95 6351.11 4123.51 3302.75
18 Cys58-Cysl 10 6351.11 5766.45 1659.81
19 Cys65-Cys72 7081.94 788.84 5906.58
20 Cys65-Cy384 7081.94 2169.35 4526.08
21 Cys65-Cys95 7081 .94 3392.67 3302.75
22 Cys65-Cysl 10 7081 .94 5035.61 1659.81
23 Cys72-Cys84 7828.75 1423.54 4526.08
24 Cys72-Cys95 7828.75 2646.87 3302.75
25 Cys72-Cysl 10 7828.75 4289.8 1659.81
26 Cys84-Cys95 9208.26 1266.36 3302.75
27 Cys84-Cys1 10 9208.26 2909.3 1659.81
28 Cys95-Cysl 10 10431.58 1685.97 1659.81

 

* “Cys i — Cys j” designate the disulﬁde bond that was reduced to form the
corresponding singly reduced/cyanylated isoform of ribonuclease-A.

39

 

2705.3 4527.4

 

 

 

8
6548.5
2265.0 59(Q1 I 7 .3 9‘77 10398.6 13334
WW [Ra—.1
7083.8
D

5907.7

 

 

 

 

 

 

 

 

 

 

 

1659.8
C
60 9.8

576% 6}51.1

3175.0 12036.7

it 0* a
"1.; L 95“.; h...“ - M -~— A ‘A‘e
3303-7 4414.4
6061.2
2296.6
9293.4 10430.5
WU -— - - L *

5000 10000 15000

Figure 1.16. The MALDI mass spectra of four peptide mixtures resulting ﬂom the CN-
induced cleavage of the singly reduced/cyanylated ribonuclease-A isoforrns. Spectra a-d
correspond to the HPLC peaks 1-4 in Figure 1.15, respectively. The symbols # and "'
represent the doubly charged ﬂagments and protonated B-elimination products,
respectively. The double mark #* in spectrum c shows a doubly charged B-elimination
product (adopted ﬂom reference 16).

40

Figure 1.16a-d are four MALDI—MS spectra of peptide mixtures resulting ﬂom
CN-induced cleavage of singly reduced/cyanylated isoforrns of RnaseA corresponding to
HPLC peaks 1—4, respectively, in Figure 1.15. By comparing the experimentally obtained
m/z values for the protonated peptide ﬂagments (MI-F) with the calculated values in
Table 1.1, assignments of cysteine connectivity can be made. For example, the mass
spectrum in Figure 1.16a (three peaks at m/z 2705.3, 6548.5, and 4527.4) agrees well
with entry No. 5 in Table 1.1, which corresponds to ﬂagments 1-25, itz-26-83, and itz-
84-124, respectively (calculated m/z: 2704.87, 6546.42, and 4526.08) with a relative mass
deviation of less than 0.05%. From these data, one can deduce that peptide chain
cleavages occur at Cys26 and Cys84. Additionally, the MALDI-MS peak at m/z 9176.7
corresponds to an overlapped peptide, 1-[3@26-83 (calculated m/z for MI-I+ = 9174.26),
resulting ﬂom peptide chain cleavage at Cys84, but with B-elimination at Cy526.
Likewise, the MALDI peak at m/z 10998.6 is another overlapped peptide, itz-26-B@84-
124 (calculated m/z for MH+ = 10990.45), with cleavage at Cy326, but B-elimination at
Cys84. Overall, a disulﬁde bond between Cys26-Cys84 can be unambiguously deduced.

Using a similar strategy, three other disulﬁde bond linkages, Cys 65-Cys72,
CysS8-CysllO, and Cys40-Cys95, also can be recognized ﬂom Figure 1.16b, Figure

1.16c and Figure 1.16d, respectively.

5. Advanced Topic: Combining Partial reduction/Cyanylation/CN-induced Cleavage
with Proteolytic Mass Mapping
Proteins that have closely-spaced cysteine residues, are often resistant to

proteolysis; even prolonged incubation may produce only big ﬂagments having two or

41

more intra-disulﬁde bonds (17) or consisting of three or more smaller digested peptides
linked by more than two inter-disulﬁde bonds (27). We can call these big ﬂagrnent
proteolytic fragment assemblies. Analyzing these proteolytic ﬂagrnent assemblies
directly by mass spectrometry may not provide suitable information for deducing the
disulﬁde linkages. In some cases, partial reduction of these large proteolytic ﬂagrnent
assemblies may be used to open one of the disulﬁde bonds; subsequent cyanylation and
CN-induced cleavage of the partially reduced proteolytic ﬂagrnent assemblies followed
by mass mapping could provide useful information about the connectivity of the
constituent disulﬁde bonds.

In one example, a ﬂagrnent of GM2-activator protein (GM2AP, a protein of 162
amino acids, 8 of them are disulﬁde cysteines in the form of 4 disulﬁde bonds), 110-H
FMDVLDMLIP TGEPQ1_25_)PEPLR TYGLPQMHQLIAQPF KE-142, was
obtained by S. aureus V8 digestion (17). Since there are no ﬂee cysteines in the native
GM2AP, the presence of this proteolytic ﬂagrnent indicates that Cys112, Cys125,
Cysl36, and Cys138 form two disulﬁde bonds. This proteolytic ﬂagrnent, GM2AP(110-
142), was subjected to partial reduction by TCEP and cyanylation by CDAP. Two Singly
reduced/cyanylated isoforrns, as determined by MALDI-TOF-MS, were isolated (Figure
1.17). The results of analysis by MALDI-TOF-MS also showed that cyanylation was
incomplete, because species with two nascent ﬂee sulﬂrydryls cyanylated and with only
one cyanylated were both detected (data not shown). Therefore, one would expect singly
cleaved as well as doubly cleaved products, and also products of the side-reaction (I3-
elimination) after CN-induced cleavage in l M NI-I40H and complete reduction of the

residual disulﬁde bond. For example, when Cys125-Cysl36 was reduced to produce one

42

isoform, only Cys125 was cyanylated (Cys136 was not cyanylated because of incomplete
cyanylation), a singly cleaved product, itz-125-l42 (cleaved at cyanylated Cys125, but
not at Cys136 because it was not cyanylated), was produced. This was conﬁrmed in the
analysis by MALDI-TOF-MS (Table 1.2). From the data in Table 1.2, one can deduce

that the two disulﬁde bonds in GM2AP(110-142) are Cys125-Cysl36 and CysllZ-

 

 

 

 

 

 

 

Cys138.
AU 1—
II 3
0.759
0.5—
2
1
0.25%
I I T V 4—1
24 26 28 30 32
ta [min]

Figure 1.17. C18-RP-HPLC separation of singly reduced and cyanylated isoforms of
GM2AP(1 10-142) (adopted ﬂom reference 17).

43

Table 1.2. m/z values of CN-cleavage products of two singly reduced/cyanylated
isoforrns GM2AP(1 10-142) after reduction of the residual disulﬁde bond (adopted ﬂom

reference 17).

 

 

 

 

 

 

 

 

 

Reduced/cyanylated Cys125-Cys136(isoform l) Reduced/cyanylated Cysl lZ-Cysl38(isoforrn 2)
CN-cleavage Calculated Experimental 8 CN-cleavage Calculated Experimental 8
products m/z m/z products m/z m/z

110-124 1686.79 1686.5 llO—lll 303.15 __b
Itz-125-136 1270.63 1269.6 Itz-112-137 2894.33 2893.40
Itz-l36-142 888.35 ,_ b Itz-l38-142 648.28 __ b

110-135c 2913.40 2913.23 110-1370 3153.47 3152.18
Itz-125-142c 2114.96 2114.77 Itz-l 12,142 0 3498.60 3498.04
llO-B@125-135 2879.41 2878.30 110-B@112-142 3119.48 3118.48
Itz-125-B@136-142 2080.97 2080.10 Itz-l 12-B@138-142 3464.62 3464.06

 

 

 

 

 

 

 

a. Average mass deviation < 500 ppm.
b. Not detected possibly due to signal suppression by matrix molecules used in MALDI.
c. Singly cleaved products ﬂom singly reduced but incompletely cyanylated species.

6. Conclusions

The partial reduction/CN-induced cleavage/mass mapping method is simple,
speciﬁc, and applicable to proteins with adjacent cysteines. With appropriate planning,
the cyanylation approach to disulﬁde mapping can be applied to highly knotted compact
systems that are usually resistant to proteolytic attack. Although there has been a
preliminary report on the determination of disulﬁde connectivity ﬂom analysis of an

intact protein by mass spectrometry (51), it was not elaborated here because the new

 

technologies of electron capture dissociation and very high resolution (R = 106) ﬂourier
transform ion cyclotron resonance (FTICR)-MS are either in their infancy or not readily

available in most laboratories.

V. References

1. Zaluzec, E. J ., Gage, D. A. & Watson, J. T. Quantitative assessment of cysteine and
cystine in peptides and proteins following organomercurial derivatization and
analysis by matrix-assisted laser desorption ionization mass spectrometry. J. Am. Soc.
Mass Spectrom. 5, 359-366 (1994).

2. SchrohenIoher, R. & Bennett, J. C. Disulﬁde Bonds. in Practical Protein Chemistry --
A Handbook (ed. Darbre, A.) Vol. 149-163 (John Wiley & Sons, 1986).

3. Smith, D. L. & Sun, Y. Detection and Location of Disulﬁde Bonds in Proteins by
Mass Spectrometry. in Mass Spectrometry of Peptides (ed. Desiderio, D. M.) Vol.
275-287 (CRC Press, 1991).

4. Aitken, A. Analysis of cysteine residues and disulﬁde bonds. in Methods Mol. Biol.
(T otowa, N. .1.) (ed. Walker, J. M.) Vol. 32, 351-360 (Humana Press, 1994).

5. Sun, Y., Bauer, M. D., Keough, T. W. & Lacey, M. P. Disulﬁde bond location in
proteins. in Methods Mol. Biol. (Totowa, N. J.) (ed. Chapman, J. R.) Vol. 61, 185-210
(Humana Press, 1996).

6. Gray, W. R. Disulﬁde bonds between cysteine residues. in Protein Struct. (2nd Ed.)
(ed. Creighton, T. B.) Vol. 165-186 (IRL Press at Oxford University, New York,
1997)

7. Gray, W. R. Disulﬁde structures of highly bridged peptides: a new strategy for
analysis. Protein Sci. 2, 1732-1748. (1993).

8. Gray, W. R. Echistatin disulﬁde bridges: selective reduction and linkage assignment.
Protein Sci. 2, 1749-1755. (1993).

9. Heck, S. D., Kelbaugh, P. R., Kelly, M. E., Thadeio, P. F., Saccomano, N. A., Stroh,
J. G. & Volkrnann, R. A. Disulﬁde Bond Assignment of omega-Agatoxins IVB and
NC: Discovery of a D-Serine Residue in omega-Agatoxin IVB. J. Am. Chem. Soc.
116, 10426-10436 (1994).

45

10.

ll.

12.

13.

14.

15.

l6.

17.

18.

19.

20.

Li, F. & Liang, S. P. Assignment of the three disulﬁde bonds of Selenocosmia
huwena lectin-I ﬂom the venom of spider Selenocosmia huwena. Peptides 20, 1027-
1034 (1999).

Daquinag, A. C., Sato, T., Koda, H., Takao, T., Fukuda, M., Shimonishi, Y. &
Tsukamoto, T. A Novel Endogenous Inhibitor of Phenol oxidase ﬂom Musca

domestica Has a Cystine Motif Commonly Found in Snail and Spider Toxins.

Biochemistry 38, 2179-2188 (1999).

Bures, E. J ., Hui, J. 0., Young, Y., Chow, D. T., Katta, V., Rohde, M. F., Zeni, L.,
Rosenfeld, R. D., Stark, K. L. & Haniu, M. Determination of Disulﬁde Structure in
Agouti-Related Protein (AGRP) by Stepwise Reduction and Alkylation. Biochemistry
37, 12172-12177 (1998).

Leal, W. S., Nikonova, L. & Peng, G. Disulﬁde structure of the pheromone binding
protein ﬂom the silkworm moth, Bombyx mori. FEBS Lett. 464, 85-90 (1999).

Young, Y., Zeni, L., Rosenfeld, R. D., Stark, K. L., Rohde, M. F. & Haniu, M.
Disulﬁde assignment of the C-terrninal cysteine knot of agouti-related protein
(AGRP) by direct sequencing analysis. J. Pept. Res. 54, 514-521 (1999).

Yamashita, H., Nakatsuka, T. & Hirose, M. Structural and ﬁrnctional characteristics
of partially disulﬁde-reduced intermediates of ovotransferrin N lobe. Cystine
localization by indirect end-labeling approach and implications for the reduction
pathway. J. Biol. Chem. 270, 29806-29812 (1995).

Wu, J. & Watson, J. T. A novel methodology for assignment of disulﬁde bond
pairings in proteins. Protein Sci. 6, 391-398 (1997).

Schutte, C. G., Lemm, T., Glombitza, G. J. & Sandhoff, K. Complete localization of
disulﬁde bonds in GM2 activator protein. Protein Sci. 7, 1039-1045 (1998).

Yang, Y., Wu, J. & Watson, J. T. Disulﬁde mass mapping in proteins containing
adjacent cysteines is possible with cyanylation/cleavage methodology. J. Am. Chem.
Soc. 120, 5834-5835 (1998).

Yang, Y., Wu, J. & Watson, J. T. Probing the folding pathways of long R3 insulin-
like growth factor-I (LR31GF-I) and IGF-1 via capture and identiﬁcation of disulﬁde
intermediates by cyanylation methodology and mass spectrometry. J. Biol. Chem.
274, 37598-37604 (1999).

Wu, J ., Yang, Y. & Watson, J. T. Trapping of intermediates during the refolding of

recombinant human epidermal growth factor (hEGF) by cyanylation, and subsequent
structural elucidation by mass spectrometry. Protein Sci. 7, 1017-1028 (1998).

46

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

Billeci, T. M. & Stults, J. T. Tryptic mapping of recombinant proteins by matrix-
assisted laser desorption/ionization mass spectrometry. Anal. Chem. 65, 1709-1716.
(1993)

Chen, X., Chen, Y. H. & Anderson, V. E. Protein cross-links: universal isolation and
characterization by isotopic derivatization and electrospray ionization mass
spectrometry. Anal. Chem. 273, 192-203. (1999).

Adamczyk, M., Gebler, J. C. & Wu, J. A simple method to identify cysteine residues
by isotopic labeling and ion trap mass spectrometry. Rapid Commun. Mass Spectrom.
13,1813-1817(1999).

Zhou, J ., Ens, W., Poppe-Schriemer, N., Standing, K. G. & Westrnore, J. B. Cleavage
of interchain disulﬁde bonds following matrix-assisted laser desorption. Int. J. Mass
Spectrom. Ion Processes 126, 115-122 (1993).

Patterson, S. D. & Katta, V. Prompt ﬂagrnentation of disulﬁde-linked peptides during
matrix-assisted laser desorption ionization mass spectrometry. Anal. Chem. 66, 3727-
373232(1994)

Crimmins, D. L., Saylor, M., Rush, J. & Thoma, R. S. Facile, in situ matrix-assisted
laser desorption ionization-mass spectrometry analysis and assignment of disulﬁde
pairings in heteropeptide molecules. Anal. Biochem. 226, 355-361 (1995).

Jones, M. D., Hunt, J., Liu, J. L., Patterson, S. D., Kohno, T. & Lu, H. S.
Determination of Tumor Necrosis Factor Binding Protein Disulﬁde Structure:
Deviation of the Fourth Domain Structure ﬂom the TNFR/NGFR Family Cysteine-
Rich Region Signature. Biochemistry 36, 14914-14923 (1997).

Jones, M. D., Patterson, S. D. & Lu, H. 8. Determination of Disulﬁde Bonds in
Highly Bridged Disulﬁde-Linked Peptides by Matrix-Assisted Laser
Desorption/Ionization Mass Spectrometry with Postsource Decay. Anal. Chem. 70,
136-143 (1998).

Gorman, J. J ., Ferguson, B. L., Speelrnan, D. & Mills, J. Determination of the
disulﬁde bond arrangement of human respiratory syncytial virus attachment (G)
protein by matrix-assisted laser desorption/ionization time-of-ﬂight mass
spectrometry. Protein Sci. 6, 1308-1315 (1997).

Merewether, L. A., Le, J ., Jones, M. D., Lee, R., Shimamoto, G. & Lu, H. S.
Development of Disulﬁde Peptide Mapping and Determination of Disulﬁde Structure

of Recombinant Human Osteoprotegerin Chimera Produced in Escherichia coli. Arch.
Biochem. Biophys. 375, 101-110 (2000).

47

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

Bean, M. F. & Carr, S. A. Characterization of disulﬁde bond position in proteins and
sequence analysis of cystine-bridged peptides by tandem mass spectrometry. Anal.
Biochem. 201, 216-226 (1992).

Bauer, M., Sun, Y., Degenhardt, C. & Kozikowski, B. Assignment of all four
disulﬁde bridges in echistatin. J. Protein Chem. 12, 759-764 (1993).

Badock, V., Raida, M., Adermann, K., Forssmann, W.-G. & Schrader, M. Distinction
between the three disulﬁde isomers of guanylin 99-115 by low-energy collision-
induced dissociation. Rapid Commun. Mass Spectrom. 12, 1952-1956 (1998).

Yen, T. Y., Joshi, R. K., Yan, H., Seto, N. O., Palcic, M. M. & Macher, B. A.
Characterization of cysteine residues and disulﬁde bonds in proteins by liquid

chromatography/electrospray ionization tandem mass spectrometry. J. Mass
Spectrom. 35, 990-1002. (2000).

Holmes, E. H., Yen, T.-Y., Thomas, S., Joshi, R., Nguyen, A., Long, T., Gallet, F.,
Maftah, A., Julien, R. & Macher, B. A. in J. Biol. Chem. 24237-24245 (2000).

Smith, D. L. & Zhou, Z. R. Strategies for locating disulﬁde bonds in proteins. in
Methods in Enzymology (ed. McCloskey, J. A.) Vol. 193, 374-3 89. (Academic Press,
New York, 1990).

Sechi, S. & Chait, B. T. Modiﬁcation of cysteine residues by alkylation. A tool in
peptide mapping and protein identiﬁcation. Anal. Chem. 70, 5150-5158. (1998).

Hara, S., Katta, V. & Lu, H. S. Peptide map procedure using immobilized protease
cartridges in tandem for disulﬁde linkage identiﬁcation of neu differentiation factor
epidermal grth factor domain. J. Chromatogr. A 867, 151-160. (2000).

Sturrock, E. D., Yu, X. C., Wu, Z., Biemann, K. & Riordan, J. F. Assignment of ﬂee
and disulﬁde-bonded cysteine residues in testis angiotensin-converting enzyme:
functional implications. Biochemistry 35, 9560-9566. (1996).

Benham, C. J. & J aﬁi, M. S. Disulﬁde bonding patterns and protein topologies.
Protein Sci. 2, 41-54 (1993).

Roepstorff, P. & Fohlrnan, J. Proposal for a common nomenclature for sequence ions
in mass spectra of peptides. Biomed. Mass Spectrom. 11, 601 (1984).

Biemann, K. Contributions of mass spectrometry to peptide and protein structure.
Biomed. Environ. Mass Spectrom. 16, 99-111. (1988).

48

43.

44.

45.

46.

47.

48.

49.

50.

51.

John, H. & Forssmann, W. G. Determination of the disulﬁde bond pattern of the
endogenous and recombinant angiogenesis inhibitor endostatin by mass spectrometry.
Rapid Commun. Mass Spectrom. 15, 1222-1228 (2001).

van den Hooven, H. W., van den Burg, H. A., Vossen, P., Boeren, S., de Wit, P. &
Vervoort, J. Disulﬁde bond structure of the AVR9 elicitor of the ﬁrngal tomato
pathogen Cladosporium fulvum: Evidence for a cystine knot. Biochemistry 40, 345 8-
3466 (2001).

Wakselman, M. & Guibe-Jampel, E. 1-Cyanyo-4-dimethylamino-pyridinium salts:
New water-soluble reagents for the cyanylation of protein sulfpydryl groups. J. C. S.
Chem Commun, 21-22 (1976).

Jacobson, G. R., Schaffer, M. H., Stark, G. R. & Vanaman, T. C. Speciﬁc chemical
cleavage in high yield at the amino peptide bonds of cysteine and cystine residues. J.
Biol. Chem. 248, 6583-6591. (1973).

Wu, J. & Watson, J. T. Optimization of the cleavage reaction for cyanylated cysteinyl
proteins for efﬁcient and simpliﬁed mass mapping. Anal. Biochem. 258, 268-276
(1998).

Qi, J. F., Wu, J., Somkuti, G. A. & Watson, J. T. Determination of the disulﬁde
structure of sillucin, a highly knotted, cysteine-rich peptide, by cyanylation/cleavage
mass mapping. Biochemistry 40, 4531-4538 (2001).

Wu, J ., Gage, D. A. & Watson, J. T. A strategy to locate cysteine residues in proteins
by speciﬁc chemical cleavage followed by matrix-assisted laser desorption/ionization
time-of-ﬂight mass spectrometry. Anal. Biochem. 235, 161-174 (1996).

Smyth, D. G., Stein, W. H. & Moore, S. The sequence of amino acid residues in
bovine pancreatic ribonuclease: revisions and conﬁrmations. J. Biol. Chem. 238, 227-
234. (1 963).

Zubarev, R. A., Kruger, N. A., Fridriksson, E. K., Lewis, M. A., Horn, D. M.,
Carpenter, B. K. & McLafferty, F. W. Electron Capture Dissociation of Gaseous
Multiply-Charged Proteins Is Favored at Disulﬁde Bonds and Other Sites of High
Hydrogen Atom Afﬁnity. J. Am. Chem. Soc. 121, 2857-2862 (1999).

49

CHAPTER 2
DEVELOPMENT OF MORE FLEXIBLE PROTOCOLS FOR THE PARTIAL
REDUCTION AND SUBSEQUENT CY ANYLATION/CN-INDUCED

CLEAVAGE/MASS MAPPING METHODOLOGY

1. Introduction

As described in Chapter 1, identiﬁcation of a particular disulﬁde bond can be
achieved by detecting an anticipated array of ﬂagments that derive ﬂom the CN-induced
cleavage of its corresponding singly reduced/cyanylated isoform and complete reduction
of the residual disulﬁde bonds (Figure 1.14 in Chapter 1). In principle, partial reduction
of a protein with n disulﬁde bonds can produce as many as n singly reduced isoforrns. If
the determination of cysteine status is based solely on singly reduced isoforrns, at least n-
1 of them must be isolated and analyzed to deduce the linkage of n-l disulﬁde bonds; the
remaining disulﬁde can be identiﬁed by default. This strategy has been successfully
applied to several proteins, such as ribonuclease-A (1) and long insulin-like grth
factor-1 (LR3IGF-1) (2). However, certain proteins, especially those with a cysteine-rich
knotted arrangement, are resistant to reduction, making it difﬁcult, if not impossible, to
prepare even as many as n-l of the n possible (predicted) singly reduced isoforrns (3). In
these situations, doubly reduced isoforms of the original protein (as well as the available
singly reduced isoforrns) may be used to solve the disulﬁde structure of the protein. In
addition to n singly reduced isoforrns, n*(n-l)/2 doubly reduced isoforms are possible for
a protein containing 11 disulﬁde bonds. The possibility of using both doubly reduced and

singly reduced isoforrns, provides ﬂexibility in applying the chemistry of partial

50

 

reduction, and increases the possibility of successful identiﬁcation of the linkages of all
the disulﬁde bonds in the protein.

In the previous proposed disulﬁde mass mapping methodology based on partial
reduction and cyanylation/CN-induced cleavage (1), the mixture of CN—induced
ﬂagments were analyzed by mass spectrometry only after the complete reduction of the
residual disulﬁde bonds that are either within the ﬂagments or linking some of the
ﬂagments together (Figure 1.14 in Chapter 1). If a CN-induced ﬂagrnent contains one or
more intra-molecular disulﬁde bonds, it is deﬁned as an oxidized CN-induced
fragment; if two or more CN-induced ﬂagments are linked by one or more inter-
molecular disulﬁde bonds, such an assembly is deﬁned as an oxidized CN—induced
fragment assembly. If there are no disulﬁde bonds in a CN-induced ﬂagrnent (all the
cysteine residues within the ﬂagrnent are ﬂee cysteines), the CN-induced ﬂagrnent is
deﬁned as a reduced CN-induced fragment. An additional analysis step of the oxidized
CN-induced ﬂagrnents or oxidized CN-induced ﬂagrnent assemblies by mass
spectrometry prior to the complete reduction step may provide extra information about
the linkages of the residual disulﬁde bonds within various oxidized CN-induced
ﬂagments or CN-induced ﬂagrnent assemblies.

Preliminary data (not shown) indicated that CN-induced cleavage on the N-
terrninal side of the cyanylated cysteine residues might produce CN-induced ﬂagments
with either an amide or a ﬂee acid C-terminus. The mass of a particular CN-induced
ﬂagrnent having an amide C-terminus is one dalton lower that the one with a ﬂee acid C-
terminus. Therefore, the status of the terminus (amide or ﬂee acid) for CN-induced

ﬂagments need to be further investigated.

51

Iminothiazolidine (itz)-blocked peptides constitute the majority of the ﬁnal
products ﬂom the cyanylation/CN-induced cleavage reactions. In the disulﬁde mass
mapping methodology based on cyanylation and CN-induced cleavage, the connectivity
of cysteines in the disulﬁde bonds is deduced ﬂom the mass of cleavage products (itz-
peptides) expected for a particular partially reduced species (a singly, doubly, or triply
reduced species) of the native protein. The identiﬁcation of the CN-induced ﬂagments
(mostly itz-peptides) is usually based on the matching of an experimental mass spectral
peak and the corresponding calculated values. In most cases, this is successful because
each of the possible itz-peptides has a unique mass. However, we may run into cases of
isomass, where two CN-induced ﬂagments have the same or very close masses, but they
have different sequences. In those cases, we need to obtain additional evidence to assign
the correct ﬂagrnent to the experimental mass spectral peak. Collision induced
dissociation (CID) mass spectrometry/mass spectrometry (MS/MS) of itz-ﬂagrnents was
investigated to determine if it can provide some evidence to distinguish a particular itz-
ﬂagrrrent ﬂom its isomass isomers.

All of the aspects introduced above are described in detail in the following

sections.

52

11. Materials and Methods

Chemicals

Bovine pancreatic ribonuclease-A type III-A was purchased ﬂom Sigma. Tris(2-
carboxyethyl)phosphine (TCEP) hydrochloride was purchased ﬂom Pierce Chemical Co.
(Rockford, IL). Guanidine hydrochloride is a product of Boehringer-Mannheim
Biochemicals (Indianapolis, IN). 1-Cyano-4-dimethylamino-pyridinium (CDAP)
tetraﬂuoroborate was purchased ﬂom Sigma. Water, acetonitrile, and triﬂuoroacetic acid
(TFA) were of HPLC grade. TCEP and CDAP solutions were ﬂeshly prepared in 0.1 M
sodium citrate buffer (pH 3.0) prior to use. Because of the instability of CDAP in
aqueous solution, solid CDAP was dissolved in pH 3 sodium citrate buffer just before

1186.

Optimize Partial Reduction to Produce Doubly Reduced/Cyanylated Ribonuclease-
A Isoforms

Approximately one equivalent of TCEP for the total disulﬁde bond content in the
proteins was used for the conditions optimized for producing singly reduced/cyanylated
protein isoforrns (see reference 1 for details). However, a lO-fold equivalent excess of
TCEP over the disulﬁde content in ribonuclease-A was used to increase the yield of
doubly reduced/cyanylated isoforrns. A solution containing 10 mnol ribonuclease-A was
prepared in 10 IIL of 0.1 M pH 3.0 sodium citrate buffer containing 6 M guanidine

hydrochloride. Partial reduction of ribonuclease-A was carried out by adding 400 nmol

53

of TCEP (4 IrL of 0.1 M TCEP solution), followed by incubation at room temperature for

15 minutes.

To the TCEP reaction mixture containing partially reduced ribonuclease-A species
were added 2000 nmol of CDAP (20 IIL of 0.1 M CDAP solution). Cyanylation of the
nascent ﬂee cysteines was accomplished during incubation at room temperature for 15

minutes.

HPLC Separation of Partially Reduced and Cyanylated Protein Species

Mixtures of partially-reduced and cyanylated species were separated by reversed
phase HPLC using a linear gradient elution of 20%-40% B in 90 rrrinutes (B is 90% (v/v)
acetonitrile/O.1% TFA; A is 0.1% TFA in water.) with Waters model 6000 pumps. UV
detection was done at 215 nm. A Vydac C18 (#218TP54, 5-um particle size, 300-A
pores in a 4.6 x 250 mm) column was used. The predominant HPLC ﬂactions were
collected manually and the masses of the collected species in each ﬂaction

(approximately 0.8~1.0 mL) were determined by MALDI-MS.

CN-induced Cleavage of Partially Reduced and Cyanylated Protein Species

A solution of a dried HPLC ﬂaction was reconstituted in 2 IIL of 6 M guanidine
hydrochloride in 1 M aqueous ammonia to dissolve the partially reduced and cyanylated
ribonuclease-A species; to this solution was added another 8 IIL of 1 M aqueous
ammonia. CN-induced cleavage of the peptide bond on the N-terrninal side of cyanylated
Cysteine residues was accomplished during reaction at room temperature for one hour.

Excess ammonia was removed in a vacuum system. Truncated peptides, still linked in

54

some cases by residual disulﬁde bonds depending on the extent of partial reduction, were
dissolved in 10 ILL of water; 1 LIL of this solution was further diluted with 9 IIL of 50%
(v/v) acetonitrile/O.1% TFA, and analyzed by MALDI-MS. The other 9 IIL of the

solution were dried in a vacuum system for future use.

Complete Reduction of the Residual Disulfide Bonds

Truncated peptides, ﬂom the CN-induced cleavage reaction, some still linked by
residual disulﬁde bonds, were completely reduced by reaction with 200 nmol of TCEP (2
IIL of 0.1 M TCEP solution) at 37 °C for 30 minutes. Samples were then diluted with 90
pL of 50% (v/v) aqueous acetonitrile/0.1% TFA solution prior to analysis by MALDI-

MS.

Production of CN-Induced Fragments from Sillucin
The details of producing the CN-induced ﬂagments (ﬂom sillucin) that are used

in section V can be found in the “Materials and Methods” section of Chapter 3.

MALDI-MS

MALDI mass spectra were obtained on a Voyager Elite or Voyager DE-STR time-
of-ﬂight (TOF) mass spectrometer (Perkin Elmer Biosystems Inc., Framingharn, MA)
equipped with a 337-nm nitrogen laser. The accelerating voltage in the ion source was
set at 25 kV. Except for accurate mass measurement of intact sillucin that was done in
positive reﬂectron mode, all other data were acquired in the positive linear mode of

0Peration. Protein standards of bovine bradykinin (monoisotopic mass of the protonated

55

molecule (MH+): 1060.569 Da; average mass of MH+2 1061.217 Da), bovine pancreatic
insulin chain B oxidized (monoisotopic mass of MH+: 3494.651 Da; average mass of
MH+: 3496.903 Da), bovine insulin (monoisotopic mass of MH+: 5730.609 Da; average
mass of MH+: 5734.6 Da) and Equine Apomyoglobin (average mass of MB“: 16952.6
Da) obtained ﬂom Sigma Chemical Co. (St. Louis, MO). All experiments were
performed using or-cyano-4-hydroxycinnamic acid (Aldrich Chemical 00., Milwaukee,
WI) as the matrix. Saturated matrix solutions were prepared in a 50% (V N) aqueous
solution of acetonitrile/0.1% TFA. Just prior to analysis, 0.5 IIL of matrix solution was
applied to the stainless steel sample plate and allowed to air-dry; then 1 IIL of sample
solution was added to the top of it, and ﬁnally another 1 IIL of matrix solution was added
to the same spot. The mixture was allowed to air-dry before being introduced into the

mass spectrometer.

CID-MS/MS of itz-peptides

The itz-peptides resulting ﬂom the CN-induced cleavage reaction were separated
ﬂom salts and reagents by HPLC, and then analyzed by inﬁrsion on a Finnigan LCQ-
Deca mass spectrometer. The solution of the itz-peptides usually contains 20-40%

acetonitrile and 0.5% acetic acid. CID was performed at 60% relative collision energy.

56

III. Expand Partial Reduction Strategy to Include Doubly Reduced Protein
Isoforms

As discussed in the “Introduction” section, doubly reduced/cyanylated protein
isoforms are need to be explored if not enough number of singly reduced/cyanylated
protein isoforms are obtained. Mass mapping a doubly reduced/cyanylated isoform of a
protein followed by CN-induced cleavage and complete reduction can determine which
four cysteine residues form two disulﬁde bonds (Figure 2.1). This information can be
used along with other information ﬂom the mapping experiment of another doubly or a

singly reduced/cyanylated isoform (Figure 2.1 and Figure 2.2).

In order to identify which four cysteine residues form two disulﬁde bonds (that
had been reduced and cyanylated) in a particular doubly reduced/cyanylated species, the
mass spectral data should be compared against all sets of calculated masses of the CN-
induced ﬂagments expected ﬂom all possible doubly reduced/cyanylated isoforrns after
the chemical treatment represented in Figure 2.1. The strategies for processing the data
ﬂom the CN-induced fragments of a doubly reduced/cyanylated isoform is described in

Chapter 5.

57

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

S
(A) 1 1% 20 3'0 4'0 5'0 6,0 70 tio 90
I s— s
S S
S S
(81) 1 PC” 20 30 40 5o FCN 70 80
10 I I I I 60 I 90
8—8 S
SCN SCN
Cleavage &
Complete Reduction
1—9
112 10 19 $339 1
H SH H appmg (10 20 60 70)
83 , F I § \ I I I _-
( ) "2 2° 30 4o 50 59 ’ form 2 S-S 3 s-s
itz 60 69
SH
itz 70+ 90 (CYS‘I O'CYSGO)
9 T (Cys20-Cys70)
1—
SH Mass
(C3) SH §H ——> form 2 S-S
Itz 30 1 59
50
SH
112 60——7'0—-79
I12 80 90
Complete Reduction
1 9 Mass
20 s—S Mapping 3
itz 1o—I—29 itz 30 1 l 59 ———> Cys40-Cys50
I
S .
itz 60 —I-O— 79 .12 8° 90
I Cleavage
SCN CN
, 20 30 40 50 § 70 80
(or) :0 I g g 60 90
SCN I SCN
S S

Figure 2.1. Overview of chemical reactions involved in the disulﬁde mass mapping of
two doubly reduced/cyanylated isoforrns (B1) and (C1) of a hypothetical protein (A).
(C2): the cleaved ﬂagments ﬂom (Cl) before complete reduction; (B3) and (C3): the
corresponding cleaved ﬂagments from (B1) and (C1) after complete reduction.

58

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

40 50 70 80
(A) 1 1 20 30 l 90
10 I I §_§ 60 I
S s
3 s
C” 20 30 4o 50 CN 70 80
(B1) 1 I0 I . I 6: I 90
S—s
SCN SCN
Cleavage &
Complete Reduction
1 9
"z ‘0 19 Mass
(33) , f“ ~71” fr“ Mapping (10.20.60.70) __
"2 2° 30 4o 50 59 ’ form 2 SS 2 SS
itz 60 69
SH (Cys10-Cys60)
itz 70——80L—— 90
(Cys20-Cys70)
1 9 Mass
SH SH SH SH Mapping
(D3) itz 10 l l I 1 5g ——-) CYSIO-CYSGO J
20 30 4o 50
SH SH
itz 60
70 80
Cleavage 81
Complete Reduction
SCN 4o 50 SC" 70 80
(D1) 1 110 20 3'0 1 l 60 I 90
S—s
S s
S 3

 

 

Figure 2.2. Disulﬁde mass mapping using a doubly reduced (B1) and a singly
reduced/cyanylated (D1) Species of a hypothetical protein (A) containing four disulﬁde
bonds. (B3) and (D3): the corresponding cleaved fragments ﬂom (B1) and (D1) after
complete reduction.

59

In experiments using ribonuclease-A as a model compound, we have found that
the partial reduction conditions can be adjusted principally via changing the
concentration of the TCEP, the reducing agent, to optimize the yield of doubly reduced
isoforms. The expected number of doubly reduced isoforrns of ribonuclease-A, which
has four disulﬁde bonds, is six. (The formula for computing the number of doubly
reduced isoforrns for a protein containing 11 disulﬁde bonds is n(n-l)/2.) Figure 2.3 is the
HPLC chromatogram of singly, doubly, and triply reduced/cyanylated isoforrns of
ribonuclease-A. It is instructive to compare Figure 2.3 with Figure 1.15 in Chapter 1,
which is the HPLC chromatogram of the partially reduced mixture resulting ﬂom
treatment of ribonuclease-A under conditions optimized for production of singly
reduced/cyanylated isoforrns. Under conditions optimized for producing singly
reduced/cyanylated isoforms, four Singly reduced/cyanylated isoforrns were produced for
ribonuclease-A (Figure 1.15 in Chapter 1); whereas under conditions optimized for
producing doubly reduced/cyanylated isoforrns, six doubly reduced/cyanylated isoforrns

were produced in addition to the four singly reduced/cyanylated isoforrns.

60

300 ,

IP
250 1
200 : 1
>
g 2
g 150 f
a 3 4
g ,
100 ‘ A
. BC E F
D
l g h I
50 l j
0 I ' ‘ 7" 7 a I? 7 - '7 T'_ — _ _ I " —
55 60 65 70 75

Retention Tlme Im In

Figure 2.3. The HPLC chromatogram of partially reduced/cyanylated species of
ribonuclease-A: IP represents the intact ribonuclease-A; Peaks 1-4 represent the singly
reduced/cyanylated isoforms (as also shown in Figure 1.15 in Chapter 1); Peaks A-F
represent the doubly reduced/cyanylated isoforrns; Peaks g-j represent the triply
reduced/cyanylated isoforrns, as determined by MALDI-MS.

61

Figure 2.4 is the MALDI mass spectrum of the cleavage products of one of the
partially reduced/cyanylated species represented by HPLC peak B in Figure 2.3. From
Table 2.1, we can see that four MALDI peaks at m/z 4412.2, 2712.3, 2646.7 and 3302.7
are due to ﬂagments 1-39, itz-40-64, itz-72-94, and itz-95-124, respectively. From these
data, one can deduce that CN-induced cleavages occurred at cyanylated cysteines, Cys40,
Cys65, Cys72 and Cys95; and thus, these four cysteine residues form two disulﬁde bonds
in ribonuclease-A. In addition, the MALDI peak at m/z 7050.5 corresponds to the
ﬂagrnent, 1-(B@40)-64, resulting ﬂom cleavage at cyanylated Cys65, but B-elimination
at cyanylated Cys40. Likewise, the MALDI peak at m/z 5874.8 is another such ﬂagrnent,
itz-72-(B@95)-124, resulting ﬂom cleavage at cyanylated Cys72, but B-elimination at
cyanylated Cys95. These two ﬂagments provide additional corroborative information
that the four cysteines, Cys40, Cys65, Cys72 and Cys95, form two disulﬁde bonds in

ribonuclease-A.

62

 

 

 

 

 

 

‘ 6.

.é‘ 5 ‘° 3

3 ‘ 9

,9 e.

5 m,

or 9

.5 ‘ R.

“I N.

'5 a ..

K v are 8 8
9° 00
3 In

3000 4000 5000 6000 7000
Mass (m/z)

Figure 2.4. The MALDI mass spectrum of a peptide mixture resulting ﬂom the cleavage
of the doubly reduced/cyanylated ribonuclease-A isoform corresponding to HPLC peak B
in Figure 2.3.

63

Table 2.1. m/z values for possible ﬂagments resulting ﬂom the cleavage reaction of one
of the doubly reduced/cyanylated isoform of ribonuclease-A represented by HPLC peak
B in Figure 2.3.

 

 

Fragment Experimental Calculated Cyanylated Cysteines in the
m/z MH+ /Da doubly reduced/cyanylated
isoform
itz-72-94 2646.7 2646.9 40, 65, 72, 95
itz-40-64 2712.3 2713.1 40, 65, 72, 95
itz-95-124 3302.7 3302.7 40, 65, 72, 95
1-39 4412.2 4411.9 40, 65, 72, 95
itz-40-83 4839.4 4839.4 26, 40, 84, 95
itz-72-(B@84)-124 5874.8 5872.6 40, 65, 72, 95
itz-40-(B@84)-94 6022.7 6028.7 26, 40, 84, 95
l-(B@40)-64 7050.5 7047.9 40, 65, 72, 95

 

64

Using the same method to analyze the MALDI spectra (data not shown) of the
CN-induced cleavage products ﬂom another doubly reduced/cyanylated isoform
represented by HPLC peak C in Figure 2.3, one can deduce that Cys26, Cys40, Cys84
and Cys95 form two disulﬁde bonds in ribonuclease-A. With the above information
about HPLC peak B and C (Figure 2.3), one can deduce that three disulﬁde bonds,
Cys26-Cys84, Cys40-Cys95 and Cys65-Cys72 exist in a ribonuclease-A molecule.

The HPLC peaks B and C in Figure 2.3 were not resolved completely. Careful
manual collection of HPLC ﬂactions ﬂom one component still contained a small amount
of the other component. This is reﬂected in the MALDI mass spectrum (Figure 2.4) of
the CN-induced cleavage products of HPLC ﬂaction B; two MALDI peaks m/z 4839.4
and 6022.7 are due to CN-induced ﬂagments itz-40-83 and itz-40-([3@84)-94, CN-
induced cleavage products ﬂom another doubly reduced/cyanylated isoform represented
by HPLC peak C in Figure 2.3. The four cyanylated cysteines (ﬂom reduction of two
disulﬁde bonds and the subsequent cyanylation) in this isoform are listed in bold font in
Table 2.1.

Using the strategy described above, we have successfully correlated the disulﬁde
bond connectivity in ribonuclease-A with the masses of the CN-induced ﬂagments ﬂom
cyanylation and cleavage of the doubly reduced isoforrns (peak A in Figure 2.3 represents
the isoform resulting ﬂom reduction of the disulﬁde bonds, Cys26-Cys84 and Cys65-
Cys72; peak B represents the isoform resulting ﬂom reduction of the disulﬁde bonds,
Cys40-Cys95 and Cys65-Cys72; peak C represents the isoform resulting ﬂom reduction
of the disulﬁde bonds, Cys26-Cys84 and Cys40-Cys95; peak D represents the disulﬁde

bonds ﬂom reduction between Cys26-Cys84 and Cys58-Cys110; peak E represents that

65

ﬂom reduction of Cys58-Cys110 and Cys65-Cys72; peak F represents the isoform
resulting ﬂom reduction of Cys40-Cys95 and CysS8-Cys110) by analysis of the CN-
induced cleavage products by MALDI-MS. Whereas, MALDI data corresponding to the
materials represented by peaks g-j indicated (by a mass shift of 156 Da ﬂom the
molecular weight of ribonuclease-A) that these are triply reduced/cyanylated isoforrns (of

which, there are predicted mathematically to be four) of ribonuclease-A.

IV. Addition of an Analysis Step by Mass Spectrometry after CN-induced Cleavage,
but Before Complete Reduction of the Residual Disulfide Bonds

As described in the “Introduction” section, an additional analysis step of the
oxidized CN-induced ﬂagments or assemblies by mass spectrometry prior to the
complete reduction step may provide extra information about the linkages of the residual
disulﬁde bonds within various oxidized CN-induced cleavage ﬂagments or assemblies.
For example, as shown in Figure 2.1, mass mapping analysis 3 is accomplished before
complete reduction of the residual disulﬁde bonds within each ﬂagrnent of the CN-
induced cleavage reaction mixture, and thus the detected presence of CN-induced
cleavage ﬂagrnent itz-30-59-ox (“ox” designates that there are residual disulﬁde bonds
within a CN-induced cleavage ﬂagrnent.) indicates the presence of the disulﬁde Cys40-
CysSO. This observation demonstrated that mass mapping analysis done after CN-
induced cleavage (but before complete reduction) can provide additional useful
connectivity information about the residual disulﬁde bonds (those not reduced during the
partial reduction step). Another important observation is that the disulﬁde information

obtained here is not available through mass mapping analysis alter the complete

66

reduction step. This means that this additional mass mapping step may help to solve the
disulﬁde structure in some situations, where fewer than enough partially reduced
isoforrns are produced. Utilization of this additional analysis by mass spectrometry
before the complete reduction step is described in the determination of the disulﬁde
structure of sillucin (Chapter 3) and in the determination of the cysteine status of the
glyceraldehyde-3-P dehydrogenase subunit-A (Chapter 4).

If there is only one inter-disulﬁde bond in an oxidized CN-induced ﬂagrnent
assembly (deﬁned in the “Introduction” section), detection of such an oxidized CN-
induced ﬂagrnent assembly will also allow us to identify the linkage of the inter-disulﬁde
bond. Such an example is shown in Figure 2.5. The MALDI-MS peak at m/z 6012.2
corresponds to an oxidized CN-induced ﬂagrnent assembly, itz40-64-(58S-S1 10)-itz-95-
124 (“58S-S110” designates the disulﬁde bond between Cys 58 and Cys 110; calculated
mass for MH+ of the assembly is 6013.8 Da). This provided the evidence for disulﬁde
bond Cys58-Cys1 10. In addition, prompt ﬂagrnentation (references 4-8) of the CN-
induced ﬂagrnent assembly into two CN-induced ﬂagments, itz-40-64 and itz-95-124
during analysis by MALDI, further conﬁrmed the existence of the disulﬁde bond CysS8-
Cysl 10.

In other cases, mass mapping analysis done after CN-induced cleavage (but
before complete reduction) may not provide information by which to identify a disulﬁde
bond. However, such results can help eliminate some possible disulﬁde structures

(Figure 2.6). One example of such an application is shown in Figure 2.7.

67

CaIc MI—F: 2713.1 Da

 

 

 

 

 

 

 

 

 

 

 

 

 

f“ CaIc MH‘: 6013.8
itz 4c 58 64 58
Prompt itz 40 l 64
I CaIc MH" : 3302.7 Da Fragmentation I"
+ SH < S
' 124 . I
112 95 110 Itz 95 110
/'
N
6
(D
.3;
(0
C
.9
E -
or
>
'8
713
°‘ 7 ‘0. <0. °°. 0° 0. 0 <0. '5. c», e; v
" " O ‘0 N N or 60
R g 8 B E E 5 8 g g .5
I I L. . A JLI ...“..JI..- I- .L. 2 NWM
4000 6000 8000 10000 12000 14am
m/z

Figure 2.5. The MALDI mass spectrum of the peptide mixture resulting ﬂom CN-
induced cleavage (but before complete reduction) of the doubly reduced/cyanylated
ribonuclease-A isoform represented by HPLC peak B in Figure 2.3. Itz-40-64 represents
the itz-blocked ﬂagrnent of ribonuclease-A ﬂom residue 40 to 64. Details about prompt
ﬂagrnentation (or in-source ﬂagmentation) during analysis by MALDI-MS can be found
in references 4-8.

68

(A)

 

 

 

 

 

 

 

 

S S
I 3° 4° 1
1 10 20 S—S 50 60 7o
1. Partial Reduction;
2. Cyanylation
S S
S S
l :0 4'0 l + Other
1 10 20 SCN SCN 50 5° 70 Products

I CN-induced Cleavage

 

 

 

 

 

 

 

10 20
1 ' ' 29
S S
11230 39 -|- I l
S S
itz ' l
40 50 60 70
IMass Mapping
Eliminates B. but not C.
S—S S—S 8—8
I l l | l L
(B) 1 10 20 30 4o 50 60 70
S S
s 30 S I
I 40 l
(c) 1 10 20 IS g 50 60 70

Figure 2.6. Eliminating a possible disulﬁde structure (B) by mass mapping analysis of the
CN-cleavage products before the complete reduction step. Disulﬁde structure (A) shows
the correct disulﬁde linkages in the hypothetical protein; structures (B) and (C) are two of
15 possible structures of arranging six cysteines as three disulﬁde bonds in such a

protein.

69

CaIc Ml-l'": 8942.2 Da

 

4
“Z 26 '0 {is 64
S S
I I
S S
itz l J

 

84 95 110 124

 

1

89431—9

2704.77

10292.5

/11000.7

—11064.6

11576.0

1

Relative Intensity
12985 3
1 3732.6

 

 

       

 

 

 

4000* 6000 8000 10000 12000 14000
m/z

Figure 2.7. The MALDI mass spectrum of the peptide mixture resulting from CN-
induced cleavage (but before complete reduction) of the doubly reduced/cyanylated
ribonuclease-A isoform represented by HPLC peak A in Figure 2.3. Itz-26-64 represents
the itz-blocked fragment of ribonuclease-A from residue 26 to 64.

70

V. Further Investigation of the Status of the C-terminus (Amide or Free Acid) of
CN-Induced Cleavage Fragments
Jacobson et al. (9) proposed a mechanism for the CN—induced cleavage reaction
that occurs on the N-terminal side of the cyanylated cysteine residue under alkaline
conditions to form an amino-N-terminal peptide (fragment) to the leﬁ of the cleavage site
and a 2-iminothiazolidine (itz) -4-carboxyl blocked peptide (ﬁagment) to the right of the
cleavage site (Figure 2.8). As shown in Figure 2.8, when the CN-induced cleavage
reaction occurs in a buffer that provides OH' to carry out a nucleophilic attack toward the
carbonyl carbon, the resulting fragments, including the amino-N-terminal peptide and the
itz-peptide, have a free acid C-terminus.
The CN-induced cleavage reaction was later optimized by Nakagawa et al. (10)

and Wu et al. (11) by using a stronger nucleophile, NH3, to replace OH'. The CN-

induced fragments produced using this modiﬁed protocol were assumed to have an amide

C-terminus (10, 11). As shown in Figure 2.9, nucleophilic attack by NH3 to the carbonyl

carbon would produce a fragment with an amide C-terminus. However, both OH' (~

0.001 M) and NH3 (~ 1 M) exist in 1 M aqueous ammonia solution, which was used in
the modiﬁed protocol. Although NH3 is far more concentrated than OH', CN—induced

fragments with both a free acid and an amide C-terminus were detected as discussed
below. Therefore, CN-induced fragments with either an amide or a free acid terminus

must be considered in computing the calculated mass of these fragments.

71

.P,

 

 

+ l I II
H3N —-— -NH—CH—C—NH—-—C<
O'
OH"
V
119
in _S
O
a? .. I /
+ /NH CO-NH— —
H3N_ _ U 0'
OH
V
O HN S
O
H3N+—-—c// + u 0
\OH HN C-NH—-—(<
0"

Figure 2.8. A proposed mechanism (9) involving nucleophilic attack by OH' for the CN-
induced cleavage reaction at a cyanylated cysteine residue.

72

N= =C—SEO
0 H
2 u

|
H3+N —-— -NH— —CH—C—NH—-—

1..

NH3

 

 

V
H69
‘N‘g _S
3L
H N+_.j?/NH CO’NH_'—C/\:_
3

@113

v

HN +_._c//:H @1111) C._NH___<OO

Figure 2.9. Production of CN-induced cleavage ﬁ'agments with an amide C-terminus via
a mechanism involving nucleophilic attack by NH3 at a cyanylated cysteine residue.

73

If a CN-induced cleavage fragment is small (usually less than 4,000 Da for an ion
trap mass spectrometer (MS) with an electrospray (ESI) ion source or 6,000 Da for a
matrix assisted laser desorption/ionization (MALDI)-time of ﬂight (TOF)-MS, its
isotopic peaks can be resolved to allow us to determine whether there is an amide C-
terminus, a free acid C-terminus or even a mixture of the two. For example, the CN-
induced cleavage fragment, itz-13-29—red (itz-blocked fragment ﬁ'om residue 13 to 29:
itz-CCGDSGYWCRQCGIKYT; “red” refers to the species in which the remaining
disulﬁde bond have been reduced during the complete reduction step; thus, all the
cysteine residues within the species are free cysteines), produced from sillucin (a peptide
having four disulﬁde bonds), has an amide C-terminus because its ESI spectrum
correlates well with its calculated isotopic distribution (Figure 2.10). Another CN-
induced cleavage fragment from sillucin, itz-12-B@13-23-ox [itz-C(B@C)CGDSGYWC-
RQ; there is a disulﬁde bond between Cys 14 and Cys 21 within the fragment, as
described in Chapter 3] has a free acid C-terminus as indicated in Figure 2.11. In
contrast, two forms of another CN-induced cleavage fragment, 1-12-red
(ACLPNSCVSKGC), resulted from CN-induced cleavage at cyanylated cysteine 13 in
sillucin, one with an amide C-terminus and the other one with a free acid C-terminus
were detected. As shown in Figure 2.12, the isotopic peak series beginning at m/z 1180.4

corresponds to the isotopic distribution of fragment 1-12-red-NH2 (calculated
monoisotopic mass for MH+ = 1180.5 Da; “NH2” designates an amide C-terminus); the

isotopic peak series beginning at m/z 1181.3 corresponds to the isotopic distribution of

fragment 1-12-red-OH (calculated monoisotopic mass for MH+ = 1181.5 Da; “OH”

74

designates a free acid C-terminus); some peaks in the amide series are superimposed on

the corresponding peaks for the free acid series.

 

 

 

 

 

 

 

 

A) Experimental
‘-
3‘ 05 v
._ (D -
(I) 00 E
C " m
93 ‘— v <1
-‘=' E s
<19 .—
._>. 51’
'10—!
£9
a)
0:
m/z
8) Calculated
100.0
a? 72.3
(I)
C
a:
E
— 42.7
(D
.2
q—n
L“ 18.2
<0
6.5
0: l I 20 0.5 01 00
g g 5.2 3 :3. s s :2 s
‘- 00 v to (O I»

Mass (MH+) /Da

Figure 2.10. Isotope peak intensity distribution in expanded ESI mass spectrum of itz-13-
29-red-NH2 (itz-CCGDSGYWCRQCGIKYT-NHZ) produced from sillucin. “M”

represents the molecule and “MHZZ‘L” represents the doubly protonated molecule.

75

A) Experimental

 

 

 

 

 

 

 

 

2* a.
0’ <1-
'5 8 g
C ‘- 00
a) ‘— v
-o-v ‘- V
as :2 g;
.2 ‘-
...—o
L“
(D
0:
m/z
3) Calculated
100.0
a? 72.3
0)
C
.03
5 42.7
OJ
.2
E 18.2
(D
6.5
m I I 2.0 0.5 0.1 0.0
g :2 e s s s s s s
O r- N co <1- u) to N
Mass (MH+) IDa

Figure 2.11. Isotope peak intensity distribution in expanded MALDI mass spectrum of
itz-12-B@13-23-ox-OH (itz-C(B@C)CGDSGYWCRQ-OH) produced from sillucin.

76

A) Experimental
1 181 .3

A :_._,|

1‘ ‘. 1182.3
1 1180.4 “833’

11 4.
8_ 3 1185.3

Relative Abundance

\
. . 1 /
2 \ . - g / . ‘
7 '3‘ I v71 .T- . 344-»: ~. ' m‘ﬂ v -.r. .m‘h-vq—:;;';2¢-¢Q:¢v .7, .... >53,” “:4; “—7 ,.,.V .—. W“'V'vq‘q." "T '7 _ , h: . ‘ 7‘ r ,2 ‘ " -‘
M . . v V., . n .— ﬁm. ﬁr. Fy -w-usﬁ m4 v H..
1 T . i T1 I" 1’ ‘ 7""? I ' 1

1180 1181 1182 1183 1184 1185
mlz

8) Calculated

0.3X100.0

0.3X61.2
0.3X34.8
0.3X13.5

i I 0.3X4.6
1180.53 1182.53 1184.53 1186.53 1188.53

+

ACLPNSCVSKGC-NHZ

Normalized Intensity

 

 

100.0
ACLPNSCVSKGC-OH

.2
2 60.8
.9
E
8
g 34.7
'26
E
z . 13.5

H 4.5

l I .
I l l l l I l l

 

1181.53 1183.53 1185.53 1187.53 1189.53

Mass (MH"') IDa

Figure 2.12. Isotope peak intensity distributions for two forms of the CN-induced
fragment, 1-12-red (ACLPNSCVSKGC; 1-12-red-NH2 and 1-12-red-OH), produced

from sillucin, were detected during ESI infusion on an ion trap. “0.3x” in the middle
panel indicates that the peak intensity is 30% of the intensity for the corresponding peaks

in the bottom panel, respectively.

77

VI. Sequence Analysis of Iminothiazolidine—Blocked Peptides by CID-MS/MS

Papayannopoulos and Biemann (12) showed that high-energy collision induced
dissociation (CID) mass spectrometry/mass spectrometry (MS/MS) on a four-sector mass
spectrometer could be used to deduce the sequence of some itz-peptides. In our
experiments, low-energy CID-MS/MS data from an ion trap mass spectrometer (Finnigan
LCQ) were used to distinguish isomass itz-peptides that may arise in mass computations
of itz-peptide fragments for the many possible disulﬁde structural isomers that must be
considered.

We have studied the itz-peptides that are listed in Table 2.2 by CID-MS/MS on an
ion trap mass spectrometer. The MS/MS data for these itz-peptides are good enough to
allow a partial sequence to be identiﬁed, and a partial sequence is typically enough for us
to distinguish one CN-induced fragment from its isomass isomer. A representative CID
spectrum is shown in Figure 2.13.

In determining the disulﬁde structure of sillucin, a cysteine-rich peptide, a CN-
induced fragment, itz-13-29-red-NH2, was distinguished ﬁom its isomass isomer, itz-l4-

3O-NH2 (see Figure 3.8 in Chapter 3 for details).

78

 

Table 2.2. Some itz-fragments studied by ESI-CID-MS/MS on an ion trap.

 

 

 

 

 

 

 

 

itz-Fragment Calculated Average Mass
(MH‘L) /Da

itz-CKRM 561.7

itz-CFGGR 563.6
itz-CFEKGMNYTVR 1372.6
itz-CVEWLRRYLEN 1505.7
itz-CEGNPYVPHFDASV 1659.8
itz-CAYKTTQANKHIIVACEGNPYVPHFDASV 3302.7

 

 

79

 

 

 

 

 

100
b7 H“ Y9
8 itz-CEGNPYV PVHFDASV
+
3
0) b6
5 b4
:12 Y12
b y y
b2 5 Va 10 91 b y13
O lLInJll Lljid 1111-10.”; 111i Ii y114 .
500 1000 1500
m/z

Figure 2.13. The ESI-ClD-MS/MS of itz-CEGNPYVPVHFDASV (residue 110-124 of
ribonuclease-A), a doubly charged ion (m/z 830.4), derived from cyanylation/CN-induced
cleavage of a singly reduced ribonuclease-A isoform, obtained on an ion trap MS. The
nomenclature of Biemann (13) for b and y ions is used throughout, although for
convenience, lower case letters are also used to indicate the peptide bond that is cleaved
to produced key fragments (b7 and y9) in a manner analogous to that introduced by

Roepstorff et al. (14) using capital letters.

80

VII. Conclusions

Various aspects of the disulﬁde mass mapping methodology based partial
reduction, cyanylation, and CN-induced cleavage are described. Including disulﬁde mass
mapping of doubly reduced isoforrns offers ﬂexibility in the chemistry of partial
reduction and provides complementary data to those obtained from the previous partial
reduction protocol, which utilizes only singly reduced isoforms. An additional analysis
step by mass spectrometry after the CN-induced cleavage reaction of partially
reduced/cyanylated species, but before complete reduction of the residual disulﬁde
bonds, may provide diagnostic information that cannot be obtained aﬁer all the residual
disulﬁde bonds have been reduced. When CN-induced cleavage reactions are conducted
in aqueous ammonia, the resulting CN-induced fragments may have an amide or a free
acid C-terminus. Both cases of an amide and a free acid C-terminus must be considered
in computing the masses for expected CN-induced cleavage fragments from different
disulﬁde structural isomers of an analyte. In cases of isomass iminothialzolidine-blocked
(itz) fragments, sequence analysis of such fragments by low-energy CID-MS/MS can be

used to distinguish one CN-induced fragment from its isomass isomers.

81

1‘.

VIII. References

10.

Wu, J. & Watson, J. T. A novel methodology for assignment of disulﬁde bond
pairings in proteins. Protein Sci. 6, 391-398 (1997).

Yang, Y., Wu, J. & Watson, J. T. Disulﬁde Mass Mapping in Proteins Containing
Adjacent Cysteines Is Possible with Cyanylation/Cleavage Methodology. J. Am.
Chem. Soc. 120, 5834-5835 (1998).

Qi, J. & Watson, J. T. Determination of the Disulﬁde Structure of Sillucin, a
Highly Knotted, Cysteine-Rich Peptide, by Cyanylation/Cleavage Mass Mapping.
Biochemistry 40, 4531-4538 (2001).

Zhou, J ., Ens, W., Poppe-Schriemer, N., Standing, K. G. & Westmore, J. B.
Cleavage of interchain disulﬁde bonds following matrix-assisted laser desorption.
Int. J. Mass Spectrom. Ion Processes 126, 115-122 (1993).

Patterson, S. D. & Katta, V. Prompt fragmentation of disulﬁde-linked peptides
during matrix-assisted laser desorption ionization mass spectrometry. Anal. Chem.
66, 3727-3732. (1994).

Crimmins, D. L., Saylor, M., Rush, J. & Thoma, R. S. Facile, in situ matrix-
assisted laser desorption ionization-mass spectrometry analysis and assignment of

disulﬁde pairings in heteropeptide molecules. Anal. Biochem. 226, 355-361.
(1995)

Jones, M. D., Hunt, J., Liu, J. L., Patterson, S. D., Kohno, T. & Lu, H. S.
Determination of Tumor Necrosis Factor Binding Protein Disulﬁde Structure:
Deviation of the Fourth Domain Structure from the TNFR/NGFR Family
Cysteine-Rich Region Signature. Biochemistry 36, 14914-14923 (1997).

Jones, M. D., Patterson, S. D. & Lu, H. S. Determination of Disulﬁde Bonds in
Highly Bridged Disulﬁde-Linked Peptides by Matrix-Assisted Laser
Desorption/Ionization Mass Spectrometry with Postsource Decay. Anal. Chem.
70, 136-143 (1998).

Jacobson, G. R., Schaffer, M. H., Stark, G. R. & Vanaman, T. C. Speciﬁc
chemical cleavage in high yield at the amino peptide bonds of cysteine and
cystine residues. J. Biol. Chem. 248, 6583-6591. (1973).

Nakagawa, S., Tamakashi, Y., Hamana, T., Kawase, M., Taketomi, S., Ishibashi,

Y., Nishimura, O. & Fukuda, T. Chemical cleavage of recombinant fusion
proteins to yield peptide amides. J. Am. Chem. Soc. 116, 5513-5514 ( 1994).

82

11.

12.

13.

14.

Wu, J. & Watson, J. T. Optimization of the cleavage reaction for cyanylated
cysteinyl proteins for efﬁcient and simpliﬁed mass mapping. Anal. Biochem. 258,
268-276 (1998).

Papayannopoulos, I. A. & Biemann, K. Amino acid sequence of a protease
inhibitor isolated from Sarcophaga bullata determined by mass spectrometry.
Protein Sci. 1, 278-288 (1992).

Biemann, K. Contributions of mass spectrometry to peptide and protein structure.
Biomed. Environ. Mass Spectrom. 16, 99-111. (1988).

Roepstorff, P. & F ohlman, J. Proposal for a common nomenclature for sequence
ions in mass spectra of peptides. Biomed. Mass. Spectrom. 11, 601 (1984).

83

CHAPTER 3
DETERMINATION OF THE DISULFIDE STRUCTURE OF SILLUCIN, A
HIGHLY KN OTTED, CY STEINE-RICH PEPTIDE, BY CYANYLATION/CN-

INDUCED CLEAVAGE MASS MAPPING

I. Introduction

The presence of antimicrobial peptides in growth media of Rhizomucor pusillus
(formerly Mucor pusillus) and Rhizomucor miehei (formerly Mucor miehei) was ﬁrst
observed over thirty years ago (1). Puriﬁcation and compositional analysis of bioactive
peptides isolated from different strains established that Rhizomucor pusillus synthesized a
30-residue peptide designated as sillucin, whereas Rhizomucor miehei strains produced
the antimicrobial peptides designated as mieheins that ranged in size from 56 to 74 amino
acids (2). The range of antimicrobial activity of these cationic peptides was found to be
limited to Gram+ microorganisms.

A common feature of the antimicrobial peptides from Rhizomucor species is the
high cysteine content varying from 8 residues in sillucin, and up to 16 residues in
mieheins (2). Negative analytical tests for reactive -—SH groups in sillucin and mieheins
indicated that most likely, all cysteine residues in these peptides were involved in
disulﬁde bonds, resulting in a highly complex tertiary structure. Although the primary
structure of sillucin was established (3) as ACLPNSCV SK GCCCGBSGYW
CRQCGIKYTC (B indicates either N or D; in our work B is determined to be D by
accurate mass measurement, as described later), the speciﬁc disulﬁde bond pattern

between its cysteine residues has remained unknown.

84

Antimicrobial peptides from eukaryotes have become subjects of intensive research
since they may serve as potential sources of new antibiotics for therapeutic applications
(4, 5). Because minor variations in peptide structure can inﬂuence antimicrobial activity,
precise knowledge of the location of disulﬁde bonds is a prerequisite to future
structure/activity studies of these bioactive peptides.

Conventional methodology for determining disulﬁde linkages involves
enzymatically or chemically cleaving the peptide backbone between the cysteine residues
and identifying the peptide fragments that contain disulﬁde bonds or are linked by them.
Analysis of proteins containing adjacent cysteines is very difﬁcult, if not possible, by
conventional methodology as there is no adequate method to cleave between the adjacent
cysteine residues (6, 7). In special cases such as the one described by Poerio et a1. (8),
proteolytic digests can be analyzed by a combination of Edman degradation and fast atom
bombardment mass spectrometry (FAB-MS) to determine the linkage of two disulﬁde
bonds involving adjacent cysteines. However, success with this approach is only possible
when the residual string of residues on the N-terminal side of the pair of adjacent
cysteines in one peptide of the three-peptide proteolytic fragment (which is linked by two
disulﬁde bonds) is shorter than the residual string of residues on the N-terminal side of
the cysteine residue in each of the other two peptides (8, 9). A similar fortunate situation
occurred during the analysis of insulin (9) and huwentoxin-I (10), where 17 and 6 Edman
degradation steps, respectively, on the intact molecule were required to provide evidence
for the disulﬁde bond pattern involving two adjacent cysteines.

In another case, fast atom bombardment mass spectrometry/mass spectrometry

(FAB-MS/MS) was used to conﬁrm the Cys6-Cys48/Cys47-Cys52 prediction in insulin-

85

like growth factors (IGFs) (11). Recently, matrix assisted laser desorption/ionization
(MALDI) -post source decay (PSD) was found to successfully fragment the peptide
bonds between two adjacent cysteines in a three-peptide proteolytic fragment linked by
two disulﬁde bonds; however, the generality of this procedure remains to be tested (12).

In some cases, if enough sample and good crystals can be obtained, high resolution
X-ray crystallography might be used to determine the disulﬁde structure involving
adjacent cysteines (13). In similar situations, two-dimensional nuclear magnetic
resonance (2D-NMR) could be applied to this analytical problem, if sufﬁcient sample is
available (14). In general, efﬁcient chemical methods are still needed for quality control
of recombinant disulﬁde proteins.

Partial reduction by the water-soluble t1is(2-carboxyethyl)phosphine (TCEP)
hydrochloride to generate an array of partially reduced protein species containing
disulﬁdes and thiols has been widely used to develop new strategies for assigning
disulﬁde bonds (15-18). Such strategies have been especially useful in solving the
disulﬁde structure of peptides and proteins with a cysteine-knotted core or those with
adjacent cysteines, which are very difﬁcult to analyze with traditional strategies based on
enzymatic digestions. For most of these strategies, the nascent free cysteine residues
produced during partial reduction were alkylated and detected by Edman sequencing, as
the basis for deducing the disulﬁde linkages. However, as alkylation is achieved under
basic conditions, artifacts due to disulﬁde scrambling could be a problem. Another
shortcoming of these strategies is that only small peptides can be considered because the

resulting alkylated peptides ﬁnally must be analyzed by Edman sequencing.

86

A novel analytical strategy developed by Wu and Watson (19) uses partial
reduction by TCEP of a multi-disulﬁde protein to produce singly reduced isoforrns,
which are then cyanylated by 1-cyano-4-dimethylamino-pyridinium (CDAP)
tetraﬂuoroborate, chemically cleaved by aqueous ammonia on the N-terminal side of the
cyanylated cysteine residues, and mass-mapped to the sequence for assignment of
disulﬁde bond linkages in proteins. The feasibility of analyzing disulﬁde structures
involving adjacent cysteines by the cyanylation/CN-induced cleavage procedure was
proven by Yang et al. (20).

Sillucin is resistant to various proteases, in part because its three adjacent cysteines
and its tightly folded structure, and attempts to determine its disulﬁde structure using
conventional methods based on enzymatic digestion have not been successful. However,
by mass mapping the CN-induced cleavage products from both singly reduced/cyanylated
and doubly reduced/cyanylated species, we have been able to deduce the disulﬁde
structure of sillucin as CysZ-Cys7, Cyle-Cys24, Cysl3-Cys30, and Cysl4-Cys21. The
chemical processes and array of CN-induced cleavage products used in deducing the

disulﬁde connectivity in sillucin are illustrated in Figure 3.1 and Figure 3.2.

87

l
‘

1i—iii 1i i

(A) 2 7 12 13 14 21 24 30

 

\IJ

 

 

 

1. Partial Reduction: TCEP @ pH 3
2. Cyanylation; CDAP @ pH 3

 

 

 

 

SCN SCN
—s j s s
2 7 12 13 14 21 24 30
HPLC fractionation
|——> MS Analysis

Cleavage; 1M 11114011

 

 

 

 

 

1 ‘17— i’ 12
2 7 [5, +
32 itz _
itz l l
1314 21 24 29
—-> MS Analysis
Complete Reduction; TCEP @pHa
SH SH
1 I 4 TH + “z .—
2 7 12 30
SH SH SH
112
+ 13 14 21 24 29

SH SH SH B SH SH SH
( l I l I I l l
1 2 7 12 1314 21 24 29

 

lMass Mapping

I
. . (Cys13-Cys30)

Figure 3.1. Overview of chemical reactions involved in the disulﬁde mass mapping of a
singly reduced/cyanylated species of sillucin. Recognition of the cleaved ﬁ'agments by
mass spectrometry allows the disulﬁde bond Cys13-Cys30 to be deduced. * The
structure in the parenthesis represents a B-elimination product, which results in this case

from loss of HSCN at cyanylated Cys 13 (ii-elimination) and CN-induced cleavage at
cyanylated Cys30.

88

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

‘ s
H 1 I"
(A) 1 I I i 18 l f
2 7 121314 21 24 30
SCN SCN
——S i S I
1 i I V? l ,,
(B1) 2 7 121314 2124 30
CN-induced Cleavage &
Complete Reduction
SH
TH I 12 + itz—
‘ 2 7 l 30
SH
...
I
SH SH SH M888 . .
(B3) itz I I I Mapping
1314 21 24 29 —> ((3,133,530)—
I
+ u -
SH H SH B SH SH SH
1 F I I I I I (Cys13-Cys30)
1 2 7 12 1314 21 24 29 9
' (Cys12-Cy524)
Mass ' '
B s s Mapplna (12.13.24.21-
112 1 ' ' fonnZS-S "
12 13 14 21 23 ——>
+
i—i
1 2 7 11 + 112 E
+
M888
1 1°» 11......
(CZ) "213 14 21 23 cys '6”
+
itz itz—
24 29 30
T CN-induced Cleavage
(C1) SCN SCN
—3 I s s
1 f l 12 I I 24
2 7 [1314 21 I 30
SCN SCN

Figure 3.2. Overview of chemical reactions involved in the disulﬁde mass mapping of a
doubly reduced/cyanylated species (C1) of sillucin (A) as well as a singly
reduced/cyanylated species (B1). Recognition of the cleaved fragments (C2) from (C1)
before complete reduction allows the connectivity of Cys14-Cys21 to be deduced.
Recognition of the cleaved fragments (B3) from (B2) (shown in Figure 3.1) after
complete reduction combined with the determination of Cys14-Cys21 as described above

allows the connectivity of Cysl2-Cys24 to be deduced; Z could be 2, 7, or 30.

89

II. Materials and Methods

Chemicals

Sillucin was isolated and puriﬁed according to methods previously described (1, 2).
Tris(2-carboxyethyl)phosphine (TCEP) hydrochloride was purchased from Pierce
Chemical Co. (Rockford, IL). Guanidine hydrochloride is a product of Boehringer-
Mannheirn Biochemicals (Indianapolis, IN). 1-Cyano-4-dirnethylamino-pyridinium
(CDAP) tetraﬂuoroborate was purchased from Sigma. Water, acetonitrile, and
triﬂuoroacetic acid (TFA) were of HPLC grade. TCEP and CDAP solutions were freshly
prepared in 0.1 M citrate buffer (pH 3.0) prior to use. Because of the instability of CDAP

in aqueous solution, solid CDAP was dissolved in pH 3 buffer just before use.

Partial Reduction and Cyanylation

A solution containing 20nmol sillucin was prepared in 10 uL of 0.1 M pH 3.0
citrate buffer containing 6 M guanidine hydrochloride. Partial reduction of sillucin was
carried out by adding 1600 nmol of TCEP (16 uL of 0.1 M TCEP solution), followed by
incubation at 50 °C for 10 minutes. Other trials of partial reduction under different
conditions of stoichiometry, temperature, and reaction time were conducted similarly to
optimize production of singly or doubly reduced species; such a search for reasonably
optimum reaction conditions is typical because of the diverse range of stability of
disulﬁde bonds from protein to protein.

To the TCEP reaction mixture containing partially reduced sillucin species were

added 4000 nmol of CDAP (20 uL of 0.2 M CDAP solution; more than a two-fold

9O

equivalent to TCEP because residual TCEP reacts with CDAP). Cyanylation of the
nascent sulﬂrydryl groups was accomplished during incubation at room temperature for

15 minutes.

HPLC Separation of Partially Reduced and Cyanylated Sillucin Species

Mixtures of partially-reduced and cyanylated species were separated by reversed
phase HPLC using a linear gradient elution of 15%—40% B in 50 minutes (B is 90% (v/v)
acetonitrile/0.1% TFA; A is 0.1% TFA in water.) with Waters model 6000 pumps
controlled by a PC. UV detection was at 215 nm. A Vydac C18 (#218TP54, S-um
particle size, 300-A pores in a 4.6 x 250 mm) column was used. The predominant HPLC

fractions were collected manually and the masses of the collected species in each fraction

(approximately 0.8~1.0 mL) were determined by MALDI-MS.

CN-induced Cleavage of Partially Reduced and Cyanylated Sillucin Species

A solution of a dried HPLC fraction was reconstituted in 2 uL of 6 M guanidine
hydrochloride in 1 M aqueous ammonia to dissolve the partially-reduced and cyanylated
sillucin species; to this solution was added another 8 uL of l M aqueous ammonia. CN-
induced cleavage of the peptide chain on the N-terminal side of cyanylated cysteine
residues was accomplished during reaction at room temperature for one hour. Excess
ammonia was removed in a vacuum system. Truncated peptides, still linked in some
cases by remaining disulﬁde bonds depending on the extent of partial reduction, were

dissolved in 10 uL of water; 1 uL of this solution was further diluted with 9 uL of 50%

91

(v/v) acetonitrile/0.1% TFA, and analyzed by MALDI-MS. The other 9 uL of the

solution was dried in a vacuum system for future use.

Complete Reduction of the Residual Disulﬁde Bonds

Truncated peptides, from the CN-induced cleavage reaction, some still linked by
residual disulﬁde bonds, were completely reduced by reaction with 200 nmol of TCEP (2
uL of 0.1 M TCEP solution) at 37 °C for 30 minutes. Samples were then diluted with 90
uL of 50% (v/v) aqueous acetonitrile/0.1% TFA solution prior to analysis by MALDI-

MS.

MALDI-MS

MALDI mass spectra were obtained on a Voyager DE-STR time-of-ﬂight (TOF)
mass spectrometer (Perkin Elmer Biosystems Inc., Framingharn, MA) equipped with a
337-nm nitrogen laser. The accelerating voltage in the ion source was set at 25 kV.
Except for accurate mass measurement of intact sillucin that was done in positive
reﬂectron mode, all other data were acquired in the positive linear mode of operation.
Time-to-mass conversion was achieved by internal calibration using standards of bovine
bradykinin (monoisotopic mass of the protonated molecule (MH+): 1060.569 Da; average
mass of MH“: 1061.217 Da) and bovine pancreatic insulin chain B oxidized
(monoisotopic mass of MH“: 3494.651 Da; average mass of NIH”: 3496.903 Da)
obtained from Sigma Chemical Co. (St. Louis, MO). All experiments were performed
using or-cyano-4-hydroxycinnamic acid (Aldrich Chemical 00., Milwaukee, WI) as the

matrix. Saturated matrix solutions were prepared in a 50% (V/V) aqueous solution of

92

acetonitrile/0.1% TFA. Just prior to analysis, 0.5 uL of matrix solution was applied to
the stainless steel sample plate and allowed to air-dry; then 1 uL of sample solution was
added to the top of it, and ﬁnally another 1 uL of matrix solution was added to the same

spot. The mixture was allowed to air-dry before being introduced into the mass

spectrometer.
III. Results

1. Accurate Mass Measurement of Intact Sillucin

The sequence of sillucin, determined previously by Bradley and Somkuti (3) is
ACLPNSCVSK GCCCGBSGYW CRQCGH<YTC (B stands for N or D). As mass
measurement error in the range of 2,000 to 4,000 Da using internal calibration in the
reﬂectron mode on a MALDI-TOF mass spectrometer (PE Voyager DE-STR) is less than
0.03% (data not shown), the uncertainty in the previously determined sequence of sillucin
can be resolved by accurate mass measurement. The monoisotopic mass of protonated
sillucin (MH+) is determined as 3200.23i0.04 Da (3200.23 is mean of seven
measurements and 0.04 is the conﬁdence value at the 95% level). Figure 3.3 shows one
of the seven measurements by MALDI-MS. The data indicate that sillucin has aspartic
acid (D) in position 16 (with overall calculated monoisotopic mass of 3200.211 Da for
MH+), instead of asparagine (N) (with overall calculated monoisotopic mass of 3199.227
Da for MH+). These experimental data also indicate that sillucin has four disulﬁde bonds

because the observed mass of 3200.23i0.04 D8 is in agreement with the calculated mass

of 3200.211 Da for MH+, which is calculated assuming that all cysteines are involved in

93

disulﬁde bonds. In addition, the data showed that sillucin has a free acid C-terminus
because the experimental data agree with all the corresponding calculated masses with

the assumption of a free acid C-terminus in sillucin.

3202.249
100 ' I 1
3201.250 3203.242

l l

L

on
0

3204.245

0)
O

3200.248

 
  
  
  
   

  
   

 

 

3205.260

.h
C

 

 

% Intensity

3206.21 0
3207.202
3208.230

 

 

 

 

 

 

‘v ‘7

3197.0 3199.8 3202.6 3205.4 3208.2 321 1.0

m/z

Figure 3.3. The MALDI mass spectrum of the intact sillucin.

2. Partially Reduced/Cyanylated Species of Sillucin

Four major components were separated by HPLC after partial reduction and
cyanylation of sillucin (Figure 3.4). Analyses of aliquots of designated fractions by
MALDI-MS indicated (data not shown) that the HPLC fraction IP represents the intact
peptide (The observed MALDI peak at a mass-to-charge ratio (m/z) 3202.8 represents
MH+, whose calculated average mass is 3202.8 Da.). HPLC fractions 1, 2, and 3
represent species that are 52, 104, and 208 Da heavier (MALDI peaks at m/z 3254.9,

3306.6, and 3410.8, respectively), respectively, than the intact peptide. Because two CN

94

 

groups are added to a singly reduced species upon cyanylation, HPLC fraction 1
corresponds to one of the four possible singly reduced/cyanylated isoforms of sillucin.
Similarly, HPLC fraction 2 corresponds to one of the six possible doubly
reduced/cyanylated isoforrns, and HPLC fraction 3 represents the completely

reduced/cyanylated sillucin.

 

 

 

 

400

> 2
E 300 .
8 1 3
5 200 ~
8 IP
2’
> 100 7
D

0 . . .

10 20 30 40

Retention Time [min

Figure 3.4. The HPLC chromatogram of denatured sillucin (IP) and some of its partially
reduced/cyanylated species. Separation was carried out on a Vydac C18 column at a
ﬂow rate of 1.0 mL/min with a linear gradient of 15% to 40% B in 50 minutes, where A
= 0.1% TFA in water and B = 90% (v/v) acetonitrile/0.1% TFA. Peaks IP, 1, and 2
represent the intact peptide, a singly reduced/cyanylated species, and a doubly
reduced/cyanylated species, respectively; peak 3 represents the completely
reduced/cyanylated species, as determined from analysis by MALDI-MS.

95

3. A Single Reduced/Cyanylated Species

Because there are eight cysteine residues in sillucin, 105 disulﬁde structure
isomers are possible. Although there are four possible singly reduced isofonns for each
of the 105 disulﬁde structure isomers, the total number of possible singly reduced
isoforrns is only 28 because some isomers have the same disulﬁde bonds as the others.
The calculated m/z values for MH+ (based on average masses of amino acid residues) of
each of the possible CN-induced cleavage fragments resulting from all 28 possible singly
reduced/cyanylated isoforrns is listed in Table 3.1.

The MALDI-MS spectrum in Figure 3.5 was obtained upon analysis of the mixture
of products resulting from CN-induced cleavage and following complete reduction of
the singly reduced/cyanylated species (HPLC fraction 1 in Figure 3.4); it shows two
major peaks at m/z 1966.8 (monoisotopic mass) and 3072.6 (average mass). As the mass
measurement error in the range of 1,000 to 3,500 Da using internal calibration in the
linear mode on a MALDI-TOF mass spectrometer (PE Voyager DE-STR) is less than
0.03% (data not shown), a certain MALDI peak can be assigned to the CN-induced
cleavage ﬁagment, which has a calculated mass that falls into the range of

massi0.03%*mass. According to calculated values in Table 3.1, the peak at m/z 1966.8
corresponds to itz-13-29-red-NH2 (calculated monoisotopic mass of MH“ = 1966.8 Da).

If the C-terminus of the intact sillucin was an amide, the peak at m/z 1966.8 could have

also been assigned to the fragment itz-14-30-NH2 (calculated monoisotopic mass for

MH+ = 1966.8 D8).

96

Table 3.1. Calculated m/z values for MH+ (based on average masses of amino acid
residues) of possible fragments resulting from CN-induced cleavage of singly
reduced/cyanylated isoforrns of all disulﬁde structural isomers of sillucin.

 

Cx-Cj a 1-(x-1)-red-NHZ b itz-xiy-lirecLNHz itz-y-30-red—OH 1-(B@r)-(y-l)-red-NH2 C itz-x-(B@y)-30—red—OH

 

 

C2-C7 89.1 557.6 2650.1 569.6 3130.7
C7-C12 603.7 517.6 2175.5 1044.3 2616.1
C2-C12 89.1 1032.2 2175.5 1044.3 3130.7

C12-C13 1078.3 146.2 2072.4 1147.4 2141.5
C7-C13 603.7 620.7 2072.4 1 147.4 2616.1
C2-C13 89.1 1135.3 2072.4 1147.3 3130.7

1181.4 146.2 1969.2 d 1250.6 2038.4

C13-C14 1967.8

C12-C14 1078.3 249.3 1969.2 1250.6 2141.5
C7-C14 603.7 723.9 1969.2 1250.6 2616.1
C2-C14 89.1 1238.5 1969.2 1250.6 3130.7

C14-C21 1284.6 811.8 1200.4 2019.4 1935.2

C13-C21 1181.4 915.0 1200.4 2019.4 2038.4

C12-C21 1078.3 1018.1 1200.4 2019.4 2141.5
C7-C21 603.7 1492.7 1200.4 2019.4 2616.1
C2-C21 89.1 2007.3 1200.4 2019.4 3130.7

C21-C24 2053.4 430.5 813.0 2406.8 1166.4

C14-C24 1284.6 1199.3 813.0 2406.8 1935.2

C13-C24 1181.4 1302.4 813.0 2406.8 2038.4

C12-C24 1078.3 1405.6 813.0 2406.8 2141.5
C7-C24 603.7 1880.1 813.0 2406.8 2616.1
C2-C24 89.1 2394.7 813.0 2406.8 3130.7

C24-C30 2440.8 708.8 147.2 3072.5 779.0

C21-C30 2053.4 1096.3 147.2 3072.5 1166.4

C14-C30 1284.6 1865.1 147.2 3072.5 1935.2

1 181.4 1968.2 147.2 3072.5 2038.4

C13’C3O (1180.5) (1966.8)

C12-C30 1078.3 2071.4 147.2 3072.5 2141.5
C7-C30 603.7 2545.9 147.2 3072. 2616.1
C2-C30 89.1 3060.5 147.2 307;; 3130.7

 

a Cx-Cy designates the disulﬁde bond that is reduced in a given singly reduced/cyanylated
isoform. b 1-(x-l)-red-NH2, itz-x-(y—1)-red-NH2, and itz-y—30-red-OH designate the CN-
C 1-(B@x)-(y—1)-red-NH2 and itz-(B@y)-30-red-OH
designate the B-elimination products. d Bold and underlined m/z values indicate matches
(or close values) with experimental data. The corresponding monoisotopic masses for

average masses 1181.4 Da, 1968.2 Da, and 1969.2 Da are 1180.5, 1966.8 Da and 1967.8
Da, respectively, as shown in the parenthesis.

induced cleavage fragments.

97

 

  
 
 

 

 

 

 

120001 g
5:? N
5 10000- 3
IS
8 8000- IS 09 \ ,0
£- 8 oi 0’.
g 60001 <6 2 8 8
v 8 .4 09 *0. <2 3
.2 4000- o .4 ,2 1;,
2 ‘- 8 53, a
.9. 2000‘ ‘- \
E.
0
1000 1500 2000 2500 3000 3500
m/z

Figure 3.5. The MALDI mass spectrum of the peptide mixture resulting from CN-
induced cleavage and subsequent complete reduction of the singly reduced/cyanylated

sillucin species represented by HPLC peak 1 in Figure 3.4. “IS” designates an internal
standard.

98

1966.8 Da (monoisotopic)

itZ\
NH
observed X 14 30 2

 

 

m/z 1966.8
1966.8 08 (monoisotopic)
itZ\
13 29 NH? \
3072.5 Da (average) / .-. Cys13 - Cys30
Observed B
—'> 1 NH

m/z 3072.6 1 X 29 2

Data are from a singly reduced!
cyanylated isoform.

Figure 3.6. Assignment of the disulﬁde Cysl3-Cys30 in sillucin.

99

However, the accurate mass measurement of intact sillucin (Section A) showed its
C-terminus is a free acid, instead of an amide. Therefore, the MALDI peak at m/z 1966.8

should be assigned itz-13-29-red-NH2. The peak at m/z 3072.6 can be attributed to 1-
(B@X)-29-red-NH2 (calculated average mass of MH+ = 3072.5 Da; B@X refers to

cyanylation, but subsequent B-elimination (no CN-induced cleavage) at cyanylated Cys
X, X can be 2, 7,12,13,14, 21, or 24).

Which of the 28 candidate singly reduced isoforms in Table 3.1 corresponds to
HPLC peak 1 in Figure 3.4? According to Table 3.1, only the singly reduced/cyanylated
isoform represented by Cysl3 - Cys30 has a match at MALDI peaks m/z 1966.8 as well
as at m/z 3072.6. This coincidence also indicates that the MALDI peak at m/z 3072.6

corresponds to the fragment l-l3@X-29-red-NH2. The basis for assignment of the

disulﬁde bond, Cys 13-Cys 30, is summarized in Figure 3.6.

Two expected fragments, l-lZ-red-NHZ and itz-30-red-OH (Table 3.1), were not

detected. Fragment itz-30-red-OH (calculated average mass of MH+ = 147.2 Da) falls
into the matrix-peak region and cannot be detected. Another expected fragment, 1-12-
red, (With a free acid C-terminus, the calculated average mass of MH+ is 1181.4 Da and
the calculated monoisotopic mass of MI-I+ is 1180.5 Da.) was not detected because of
signal suppression during the initial analysis by MALDI. It was detected in MALDI after
it was separated from other species in the mixture by HPLC (Figure 3.7). As shown in
Figure 3.7, the MALDI peak at m/z 1181.5 corresponds to 1-12-red-OH; the MALDI
peak at m/z 1179.5 represents l-lZ-ox-OH, which resulted from oxidization of two of the
three free cysteines, Cys 2, Cys 7, and Cys 12, into a disulﬁde bond. The oxidation

occurred because HPLC separated the reducing agent, TCEP, from the fragment 1-12-

100

red-OH. However, slightly different from the result by ofﬂine HPLC-MALDI-MS,
another two forms of the CN-induced fragment 1-12-red, 1-12-red-NH2 and 1-12-red-
OH, were detected when the whole mixture including two other fragments, itz-l3-29-red-
NH2 and 1-B@l3-29-red-NH2, was subjected to analysis by infusion on an ESI-ion-trap-
MS (Figure 2.12 in Chapter 2). The sources of these two forms of the fragment l-12-red
were discussed in detail in chapter 2. Fragment 1-12-red-NH2 was not detected in
ofﬂine-HPLC-MALDI-MS because the presence of 1-12-ox-OH, resulted from oxidation,
masked the presence of 1-12-NH2 (Figure 3.7). To the contrary, fragment 1-12-red-NH2
was detected by ESI infusion (Figure 2.12 in Chapter 2) because the reducing reagent,
TCEP, was still present in the mixture, preventing 1-12-ox-OH from forming by
oxidation of 1-12-red-OH.

The experimental peak at m/z 1966.8 (monoisotopic mass) in Figure 3.5 is the same
as the monoisotopic mass (1966.8 Da for MW“) of itz-14-30-NH2 (if the C-terminus of
the intact sillucin is an amide); it is close to the calculated monoisotopic mass (1967.8 Da
for MH+) of itz-14-30-OH (if the C-terminus of the intact sillucin is OH). As shown in
Figure 3.8, Table 3.2, and Table 3.3, electrospray (ESI) tandem mass spectrometry
(MS/MS) data by an ion trap further showed that the MALDI peak at m/z 1966.8 should
be assigned to itz-13-29-NH2 rather to itz-14-30-NH2 or to itz-l4-30-OH. Almost a

complete series of b ions from itz-13-29-NH2, rather than those from itz-14-30-NH2 or

itz-14-30-OH, were detected (Figure 3.8).

101

 

1181.5
1182.5

1183.5

1179.5

Relative Intensity
1 180.5

 

 

 

11.74 11.76 1178 1180 1182 1184 1186 1188
m/z

Figure 3.7. Detection of the CN-induced fragment 1-12-OH (two forms: 1-12-ox-OH and
l-12-red-OH) by MALDI-MS after it was separated from itz-12-29-red-NH2 and 1-

B@13-29-red-NH2.

102

itz-13-29-NH2 from Sillucin

a) Lower-Mass Portion

 

 

 

 

ID132+ : Ib16 - H2012+
31°07 2+ .. 3 [MHz - H20 - NH312+
‘ b12 .— b 2+
5 807‘ o .- 0, ,_ 2+ 16
3 601:5 5.)- is \i: 914 \\ [MI-12 - H20]2+
‘ 3 b 2+ 7 b 2+ /
401 v 11 a 14 2+
, y. .. .. \ 7. \ .. I
. I
a 170.1 _: '-.. 21-‘l-311431-l.2 --.~. =--L 3.31 Mill-ILL? -.Ll ullhl LIL-ALL' I: 3._.i
400 500 600 700 800 900 ‘ 1000
ml:
b) Higher-Mass Portion
1005 Y1o NH3 b" ‘7 Sr? b“: 1557.5
8 .ye - NH3 1/284. 8 L; g N.
i 80" :3 '- 8
c (D
601 1256 a ‘-
3 I 0'70. 5 b10: b1” 14
2 40 1156. 3 1315
i°°II.L. l I
I 0.1, ;_ i I 11 III 1 :1 .tI ‘11.; l l 3.451411 1 Iii J '_ i LLB 3...; mi...
1 100 1200 1300 1400 500 1600 1700
m/z

Figure 3.8. The ESI-CID-MS/MS spectrum of CN-induced fragment itz-13-29-NH2 from
sillucin. The precursor is a doubly protonated molecule; [MHZ-H20]2+ designates a

dehydrated and doubly protonated fragment of the precursor; “B13” indicates the peptide

bond that is cleaved in the peptide backbone between residue G and I (see deﬁnition of
the nomenclature in reference 21) to form an amino terminal ion represented by “b13”

(singly charged) and b132+ (doubly charged); the deﬁnition of a, b, and y ions can be
found in reference 22.

103

Table 3.2. Calculated m/z values (monoisotopic) for b and y ions of CN-induced
fragments itz-13-29-NH2 and itz-14-30-NH2 from sillucin.

 

 

bi 11243-29-an itz-14-3O-NH2 yi 11243-2913112 itz-14.30-NH2
61 129.0 129.0 y1 119.1 121.0
b2 232.0 186.0 y2 282.2 222.1
b3 289.0 301.1 y3 410.2 385.2
b4 404.1 388.1 y4 523.3 513.3
b5 491.1 445.1 y5 580.4 626.3
b6 548.1 608.2 y6 683.4 683.4
67 711.2 794.3 y7 811.4 786.4
b3 897.3 897.3 y8 967.5 914.4
b9 1000.3 1053.4 y9 1070.5 1070.5
610 1156.4 1181.4 yio 1256.0 1173.5
b11 1284.4 1284.4 y11 1419.7 1359.6
612 1387.4 1341.5 y12 1476.7 1522.7
b13 1444.5 1454.6 y13 1563.7 1579.7
b14 1557.6 1582.6 y14 1678.8 1666.7
b15 1685.6 1745.7 yls 1735.8 1781.8
b16 1848.7 1846.8 y..S 1838.8 1838.8

 

104

Table 3.3. Calculated m/z values (monoisotopic) for b and y ions of CN-induced
fragments itz-13-29-NH2 and itz-14-30-OH from sillucin.

 

 

bi 112-13-29-an itz-14-30-OH yi 11243-29-an itz-14-30—OH
b1 129.0 129.0 y1 119.1 122.0
b2 232.0 186.0 y2 282.2 223.1
b3 289.0 301.1 y3 410.2 386.1
b4 404.1 388.1 y4 523.3 514.2
b5 491.1 445.1 yS 580.4 627.3
b6 548.1 608.2 y6 683.4 684.3
b7 711.2 794.3 y7 811.4 787.4
b3 897.3 897.3 y8 967.5 915.4
b9 1000.3 1053.4 y9 1070.5 1071.5
610 1156.4 1181.4 on 1256.0 1174.5
611 1284.4 1284.4 y11 1419.7 1360.6
612 1387.4 1341.5 ylz 1476.7 1523.7
b13 1444.5 1454.6 y13 1563.7 1580.7
b14 1557.6 1582.6 y14 1678.8 1667.7
b15 1685.6 1745.7 y,5 1735.8 1782.7
b16 1848.7 1846.8 y..S 1838.8 1839.8

 

105

4. A Doubly Reduced/Cyanylated Species

The MALDI-MS spectrum in Figure 3.9 corresponds to the fragments resulting
from CN-induced cleavage (before complete reduction) of the doubly
reduced/cyanylated species (HPLC fraction 2 in Figure 3.4). The MALDI peak series

beginning at m/z 1299.4 and at m/z 1300.4 can be attributed to itz-l3-23-ox-NH2

(calculated monoisotopic mass of MH+ = 1299.4 Da) and itz-13-23-ox-OH. As this
analysis by MALDI-MS was done before complete reduction of the sample, the
presence of itz-13-23-ox-NH2 and itz-13-23-ox—OH implies the existence of a disulﬁde
bond, Cys14-Cys21 (see summary in Figure 3.10). With the disulﬁde bond Cys14-Cys21
identiﬁed, the MALDI peak series at m/z 1369.4 can be inferred as representing itz-12-
(B@13)—23-ox-OH (calculated monoisotopic mass of MH+ = 1369.5 Da); thus, Cys 12,
Cys 13, Cys 24, and Cys Z form two disulﬁde bonds because two disulﬁde bonds had to
be reduced to form the doubly reduced/cyanylated species (HPLC fraction 2 in Figure
3.4). Because the disulﬁde bond Cys13-Cys30 was identiﬁed in the last section, Cys Z
must be Cys 30. Thus, another disulﬁde bond, Cys12-Cys24, can be deduced (Figure
3.10). In addition, MALDI peak at m/z 1331.4 can be attributed to 1-SCN@12-

SCN@13-ox-NH2 (calculated monoisotopic mass of MH+ = 1331.5); and thus, CysZ-
Cys7 can be deduced because 1-SCN@12-SCN@13-ox-NH2 contains a disulﬁde bond

(see summary in Figure 3.10).

Analysis by MALDI-MS (data not shown) shows that denatured intact sillucin does
not react with CDAP, the cyanylation reagent. This implies that there are no free
cysteines in sillucin and that all eight cysteine residues form four disulﬁde bonds. This

observation is in agreement with the accurate mass measurement described above and

106

 

literature results (3). With the three disulﬁde bonds, Cyle-Cy524, Cysl3-Cys30, and
Cys14-Cys21, having been deduced as described above, the fourth disulﬁde bond, Cys2-
Cys7, can be deduced by default. This deduction is consistent with the presence of a
MALDI-MS peak at m/z 1331.4 (Figure 3.9), representing 1-SCN@12-SCN@13-ox-

N112. These assignments are summarized illustratively in Figure 3.10.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

v 5r.
o 0’ <3
8 8 o
V ‘- "' N
as i ‘9
a: ‘- V
‘- CO
,.
'5 OLD
3 5000- IS ; .58
2‘ ‘9. ‘0 ‘3‘“ Is
10 40001 o 00 N
u- .1 o o .
‘9 300° *- 8 '0 o: co 5’ in 8 co
<( 9" 00 N V V - O_ co ' N 0)
v ' 0 V ' O) 0') w m 1 (D
2000- 00 ‘— "' v- ‘- 00 N (O °° 3
34 00 <- co o: o 38 g} o
C 1000 ‘— , ‘— \ \
a)
E 0
1000 1500 2000 2500 3000 3500
m/z

Figure 3.9. The MALDI mass spectrum of the peptide mixture resulting from CN-
induced cleavage (before complete reduction) of the doubly reduced/cyanylated sillucin
species represented by HPLC peak 2 in Figure 3.4.

107

1299.4 Da (monoisotopic)

 

 

A observed peak itz\ Tc’ 18
' at m/z 1299.4 —‘ 13 14 21 23 ””2\
1300.4 Da (monoisotopic) Cys14-Cy321
observed peak itz\ § §

 

atm/z1300.4 —> 1314 2123OH

Data were from a doubly reduced/cyanylated isoform. \
Cys12,13,24.Z

 

 

 

1369.5 Da (monoisotopic)
B Cys14-Cys21 it2\ ‘1 s 19. OH / Form 2 s-s Cys12-Cy824
observed peak 12 13 14 21 23 Cys13-Cys30
at m/z 1369.4
1331.4 Da (monoisotopic)
Cys12 & Cys13 z z
I . 0
were cyany ated ,3 J3 (f ('03
observed peak 1 2 7 12 13 NHZ - CysZ-CysT
at m/z 1331.5

Cys12-Cy524 —

Cys13-Cys30 -

D- L» c 2-c s7
Cys14-Cys21 4 y‘ y

 

Cys2,7,12,13,14,21, _L
24,30 Form 4 S-S

Figure 3.10. Assignment of the disulﬁde bonds Cys14-Cyle, Cys12-Cy824, and Cys2-
Cys7.

108

 

The MALDI-MS spectrum in Figure 3.11 was obtained from analysis of the
products from CN-induced cleavage and subsequent complete reductions of the
doubly reduced/cyanylated species (HPLC fraction 2 in Figure 3.4). The MALDI peak
series at m/z 1301.5 (Figure 3.11) is 2 mass units higher than the corresponding peak at
m/z 1299.4 (Figure 3.9), which represents the oxidized form itz-l3-23-ox-NH2
(calculated mono isotopic mass of MH+ = 1299.4 Da); thus, the peak at m/z 1301.5 can
be attributed to itz-13-23-red-NH2 (calculated monoisotopic mass of MH+ = 1301.5 Da),
in which the disulﬁde bond Cysl4-Cy521 had been reduced during the complete
reduction step. This mass spectrometric evidence of a shift of 2 Daltons upon reduction
of itz-13-23-ox-NH2 to itz-l3-23-red-OH conﬁrms the existence of the disulﬁde bond
Cysl4-Cy521. Similarly, the MALDI peak series at m/z 1371.5 can be assigned as
representing itz-12-B@13-23-red-OH (calculated monoisotopic mass of NIH" = 1371.5
Da), which results from reduction of the disulﬁde bond Cysl4-Cy821 in itz-12-B@13-23-

ox-OH (observed at m/z 1369.4; calculated monoisotopic mass of MH’r = 1369.5 Da).

In summary, the linkage of four disulﬁde bonds in sillucin is identiﬁed as Cys2-
Cys7, Cys12-Cy324, Cysl3-Cys30, and Cys14-Cys21, by partial reduction, cyanylation,
and CN-induced cleavage with or without complete reduction, followed by mass mapping

analyses.

109

 

 

 

 

 

 

 

 

 

 

 

  
 

 

 

 

 

 

 

1 l0
‘- 3 {B N to
,9, co '0. ‘— N «5
‘- ‘_ (‘0 L0 2 N
o - J co
m g ‘—
‘—
CO
‘—
A \
“
$120001” '3 '2“.
‘_
(c5 ‘9 0 9°.
E‘ LO O CO ‘- IS
(U O) CO ‘- LO CD
*1: 8000 2 o . O)
3 ‘- “). ‘0 O3. ‘— w 0).
L v- O N ' (D
s. ('7, $- 60 0°. 8 o:
2‘ 4000? ‘- / 93 5 a g
.7) \ \ O /
c N L
1000 1500 2000 2500 3000 3500

m/z
Figure 3.11. The MALDI mass spectrum of the peptide mixture resulting from CN-

induced cleavage and subsequent complete reduction of the doubly reduced/cyanylated
sillucin species represented by HPLC peak 2 in Figure 3.4.

110

IV. Discussions
1. Partial Reduction

Obtaining several partially reduced species of a multi-disulﬁde peptide is essential
for successful determination of the disulﬁde linkages. All the disulﬁde linkage
information is lost by generating the completely reduced peptide species. The rationale
for optimizing partial reduction conditions is to generate a reasonable yield of partially
reduced peptides with at least one disulﬁde bond opened, while at least one remains
closed (23). For cysteine-rich, knotted peptides, it is best to use a large excess of reducing
reagent (TCEP) and control the extent of reduction by varying the time and temperature
(23). Partial reduction can be accomplished in a relatively short time (in the range of 2-15
minutes) and stopped by removing the TCEP by HPLC when the reaction mixture is
injected while still under kinetic control. As a result, a fair amount of the starting material
will be left untouched, while a useful amount of partially reduced peptide species is
obtained. The general guidelines for optimization of partial reduction have been
established by Gray (23) and Wu (l9); usually, only a few experiments are needed (23).

Because eight of the thirty amino acid residues in sillucin are cysteine, including

three adjacent cysteines, the peptide appears to be in a knotted arrangement, which offers
great resistance to chemical reduction. For this reason, a large excess (IO-fold equivalent
excess over total disulﬁde content) was used in the initial trial for optimizing the partial
reduction reaction. For the same reason, the temperature of the partial reduction reaction
was raised from room temperature to 40°C, and eventually to 50°C. A good yield (25%+)
of doubly reduced species of sillucin was produced during the third attempt to optimize

the partial reduction reaction by increasing the temperature to 50°C while using a 20-fold

111

equivalent excess over total disulﬁde content for 10 min; the relative distribution of
products in this reaction mixture is represented by the HPLC chromatogram in Table 3.4.
The reaction mixture from the ﬁrst two optimization trials gave similar chromatograms
(data not shown), except with considerably smaller peaks for the singly- and doubly-
reduced species.

Information leading to the identiﬁcation of one disulﬁde bond (Cysl3-Cys30) can
be gleaned from mass mapping analysis of the reaction mixture following CN-induced
cleavage of a singly reduced species of sillucin after complete reduction of residual
disulﬁde bonds; the key reactions and data are shown in Figure 3.1. Ideally, one would
hope to isolate a singly reduced isoform of the intact peptide that corresponds to each of
the disulﬁde bonds; this was the case for ribonuclease A (consisting of eight cysteines in
the form of four disulﬁdes), where nearly equal amounts of four singly reduced isofonns
were available for mass mapping (19). However, in the case of sillucin, its knotted
conformation apparently prevents equal access of the chemical reducing agent to each of
the disulﬁde bonds, and thus only one of the four possible singly reduced isoforrns can be
readily generated.

In situations where an insufﬁcient number of singly reduced isoforrns of the
analyte is available, attempts to produce and isolate doubly reduced isofonns of the
analyte may solve the problem. Such is the case with sillucin, where as indicated in the
HPLC chromatogram in Figure 3.4, a substantial amount of one of the six possible
doubly reduced isofonns is available for analysis. As summarized in Figure 3.2, mass
mapping analysis of the reaction mixture resulting from CN-induced cleavage before and

after subsequent complete reduction to reduce the residual two disulﬁde bonds allows the

112

 

connectivity of three more disulﬁde bonds (CysZ-Cys7, Cysl4-Cy321 and Cyle-Cy324)

to be deduced.

2. Incomplete Cyanylation, Incomplete CN-induced cleavage, and Disulﬁde
Scrambling

Although the efﬁciency of the cyanylation reaction is usually 95%+ (24), there are
various imperfections associated with the cyanylation and CN-induced cleavage
reactions, especially when there are three adjacent cysteines to offer extreme steric
hindrance for the cyanylation and cleavage reagents to access the middle cysteine
residue. In Figure 3.1, a product from a competing side reaction, B-elimination of HSCN
from a cyanylated cysteine residue, was shown; depending on the sequence, variable
amounts of this B-elimination product are generated in competition with the desired CN-
induced cleavage on the N-terminal side of cyanylated cysteine residues. Other side
reactions include incomplete cyanylation and incomplete CN-induced cleavage at
cysteine residues. Figure 3.12 and Figure 3.13 show some examples of incomplete
cyanylation and incomplete CN-induced cleavage reactions. These side reactions,
including B-elimination, incomplete cyanylation, and incomplete CN-induced cleavage,
can occur in combination. Furthermore, these side reactions contribute additional peaks
to the MALDI spectra of the CN-induced cleavage reaction mixture. Sometimes these
peaks are minor, indicating little side reaction; in other cases, these peaks dominate as a
result of super sensitivity of some of these minor products in the MALDI process or high

yields for the side reactions.

113

 

Another complication is that the partially reduced/cyanylated species of the peptide
are sometimes similar, and carry-over or co-elution of such species with others is possible
during HPLC separation. This aspect can also contribute to additional MALDI peaks that
are not related to the CN-induced cleavage products of a speciﬁc partially
reduced/cyanylated species.

Most of the MALDI peaks in Figure 3.5, Figure 3.9, and Figure 3.11, except those
described for determination of the disulﬁde structure of sillucin (“Results” section), can
be attributed to the side reactions discussed above. The identities of those peaks are
listed in Table 3.4, Table 3.5, and Table 3.6. Because these “extra” peaks can be
accounted for, they do not preclude correct data interpretation, but simply complicate it.
Some representative MALDI peaks related to side reactions are described below.

Some of the MALDI peaks in Figure 3.5, other than those identiﬁed in the
"Results" section, can be interpreted as follows. The peak at m/z 1991.7 corresponds to

itz-13-SCN@24-29-red-NH2 (calculated monoisotopic mass of MH+ = 1991.8 Da;

SCN@24 is used to designate a cyanylated cysteine residue at position 24). Itz—13-
SCN@24-29-red could result from incomplete CN-induced cleavage (CN-induced
cleavages occur at CyslBSCN, and CySBOSCN, but NOT at Cys24SCN) and complete
reduction of minor "carry-over" of the doubly reduced/cyanylated species,
[sillucin(13SCN-3OSCN, IZSCN-24SCN)], in which disulﬁde bonds Cys13-cys30 and
Cys12-cy824 have been reduced and cyanylated (Another partially reduced/cyanylated
species is designated similarly in the following discussions.) as represented by HPLC

peak 2 in Figure 3.4. Formation of itz-13-SCN@24-29-red-NH2 is illustrated in Figure

3.12.

114

 

Table 3.4. Identities of the MALDI-MS peaks in Figure 3.5.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Experimental Calculated CN-induced Fragment S-S Reduced Side
m/z m/z during Partial Reactions
Reduction involved
MALDI Peaks Identiﬁed in the “Results” Section
1966.8 (M) "' 1966.8 (M) itz-13-29-red-NH2 13-30 N/A
3072.6 (A) 3072.5 (A) 1-B@13-29-red-NH2 13-30 N/A
3202.8 (A) 3202.8 (A) Intact Sillucin N/A N/A
MALDI Peaks Identiﬁed in the “Discussion” Section
1991.7 (M) 1991.8 (M) itz-13-SCN@24-29-red- 13-30, 12-24 No cleavage
NHZ at SCN@24
2070.8 (M) 2070.8 (M) itz-13-30-red-OH 13-30 No
cyanylation
at SH@3O
2097.5 (A) 2097.4 (A) itz-13-SCN@30-red-OH 13-30 No cleavage

at SCN@30

 

" “M” designates monoisotopic mass; “A” designates average mass.

115

 

Table 3.5. Identities of the MALDI-MS peaks in Figure 3.9.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Experimental Calculated CN-induced Fragment S-S Reduced Side
m/z m/z during Partial Reactions
Reduction involved
MALDI Peaks Identiﬁed in the “Results” Section
1299.4 (M) 1299.4 (M) itz-13-23-ox-NH2 12-24, 13-30 N/A
(contains Cys14-Cys21)
1300.4 (M) 1300.4 (A) itz-1 3-23-ox-OH 12-24, 13-30 N/A
(contains Cysl4-Cy321)
1369.4 (M) 1369.5 (M) itz-12-B@13-23-ox-OH 12-24, 13-30 B-elimination
(contains Cysl4-Cys21) at SCN@13
MALDI Peaks Identiﬁed in the “Discussion” Section
1331.4 (M) 1331.5 (M) 1-SCN@12-SCN@13- 12-24, 13-30 No cleavage
ox-NHZ at SCN@12
(contains Cys2 -Cys7) and SCN@13
1989.9 (M) 1989.8 (M) itz-l3-SCN@24-29-ox- 12-24, 13-30 No cleavage
NHZ at SCN@24
(contains Cys14-Cys21)
3066.6 (A) 3066.5 (A) 1-B@13-29-ox-NH2 13-30 B-elimination
at SCN@13
3202.8 (A) 3202.8 (A) Intact sillucin N/A N/A

 

116

 

 

Table 3.6. Identities of the MALDI-MS peaks in Figure 3.11.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Experimental Calculated CN-induced Fragment S-S Reduced Side
m/z m/z during Partial Reactions
Reduction involved
MALDI Peaks Identiﬁed in the “Results” Section
1301.5 (M) " 1301.5 (M) itz-l3-23-red-NH2 12-24, 13-30 N/A
1302.5 (M) 1302.4 (M) itz-l3-23-red-OH 12-24, 13-30 N/A
1371.5 (M) 1371.5 (M) itz-12-B@13-23-red-OH 12-24, 13-30 B-elimination
at SCN@13
MALDI Peaks Identiﬁed in the “Discussion” Section
1405.5 (M) 1405.5 (M) itz-12-23-red-OH 12-24, 13-30 No
cyanylation
at SH@13
1932.9 (M) 1932.8 (M) itz-13-B@24-29-red- 12-24, 13-30 B-elimination
NH2 at SCN@24
1991.8 (M) 1991.8 (M) itz-13-SCN@24-29-red- 12-24, 13-30 No cleavage
NH2 at SCN@24
2061.8 (M) 2061.8 (M) itz-13-B@24-SCN@30- 12-24, 13-30 B-elimination
red-OH at SCN@24,
no cleavage at
SCN@3O
itz-13-SCN@24-B@30- 12724, 1330 B-elimination
red-OH at SCN@30,
no cleavage at
SCN@24
2120.8 (M) 2120.8 (M) itz-13-SCN@24- 12-24, 13-30 no cleavage at
SCN@30-red-OH SCN@24 and
SCN@3O

 

It . . . .
“M” des1gnates mono1sotopic mass; “A” desrgnates average mass.

117

 

sillucin(13SCN-3OSCN, 12$CN-24SCN)

 

 

Z Z Z Z
0 O 0
s s w 8 s—s w w
I I J I I I l 10H
1 2 7 12 13 14 21 24 30

CN-induced Cleavage at Cys12$CN,
Cys13SCN, and CysSOSCN,

 

but NOT at Cy824SCN.
itz-1 3-SCN@24-29-ox-th
itz s ——s SCN
\ l I I 0th P od t
13 14 21 24 29"”2 'I' 8' ' ”cs

Complete Reduction

itz-13-SCN@24-29—red-Ni-Q

itz SH SH SCN

\ I I l
13 14 21 24 29NHz

 

Figure 3.12. Formation of itz-13-SCN@24-29-ox-NH2 and itz-l3-SCN@24—29-red—NH2
because of no CN-induced cleavage at SCN@24.

118

The peak at m/z 2070.8 could be attributed to itz-13-30-red-OH or itz-12-29-red-
OH (calculated monoisotopic masses of MI-I+ ions for both fragments are 2070.8 Da).
Itz-13-30-red-OH could result from CN-induced cleavage and complete reduction of the
singly reduced/incompletely cyanylated species, [Sillucin(13SCN-3OSH)] ("SH"
designates a free sulﬂiydryl at Cys 30), which is a minor component in the collected
HPLC fraction 1 in Figure 3.4. [Sillucin(13SCN-30$H)] (calculated average mass of
MH+ = 3229.8 Da; 27 Da heavier than that for intact sillucin.) is represented by a minor
MALDI peak at m/z 3229.7 (data not shown). Figure 3.13 illustrates the formation of itz-
13-30-red-OH.

Itz-12-29-red-OH can result from CN-induced cleavage and complete reduction of
scrambled/cyanylated species, [Sillucin(1 ZSCN-308CN)], which results from scrambling
between Cys 12 and Cys 13 (Figure 3.14).

Further experiments showed that the MALDI peak at m/z 2070.8 disappeared after
additional HPLC separation removed the minor component [Sillucin(13SCN-30SH)]
from HPLC fraction 1 in Figure 3.4; the puriﬁed species [Sillucin(13SCN-3OSCN)] was
cleaved and the remaining disulﬁde bonds were completely reduced (data not shown).
This infers that no scrambling occurred; the peak at m/z 2070.8 represents itz-13-30-red-
OH, which is a CN-induced cleavage product of the singly reduced/incompletely
cyanylated species, [Sillucin(13SCN-SH)]. Disulﬁde scrambling does not occur despite
the existence of three adjacent cysteines, probably due to the fact that partial reduction
and cyanylation are carried out at pH 3, and, thus, the concentration of thiolate anion is

minimal.

119

 

S

S S
s s J
4

 

 

 

 

N1-m
N—m

, l l
12 13 14 21 2

'OH

1. Reduction of Cys13-Cy630;

2. Incomplete Cyanylation:
Cyanylation at Cys138H,
but NOT at Cya30$H

 

V

sillucin(13SCN-3OSH)

SCN SH

 

S
l 2 I2
12 13 14 21 24 30OH

 

 

 

NI—a)
NJ—(D

1.CN-induced Cleavage at Cys13SCN
2. Complete Reduction

itz-1 3-30-red-OH

itz SH SH SH SH +
L I I I 1 OH
13 14 21 24 30

Other Products

Figure 3.13. Formation of itz-13-30-red-OH due to incomplete cyanylation at Cys 30;
Cys3OSH is used to represent the free cysteine residue.

120

I

<3 35, ‘3,\ Sillucin
<05; 12 -

\13 J

3'1 #1
2 A 2 A
I
a: 24® 21 ‘ 24x® 21
(10’ I 3 in (:5? I 30 1
) 3 P a” + )1 (900 g ‘3, + Others
(3.1 as“ \ @I H \

12 6‘ N” 12 '0’ 5‘"
K13 J K18 -/
Sillculn(13SH-308H) l Scrambling

l Cyanylation
-1
2 A
.41 l 21
2 A S” 24 |
l 0)
24 21 C l
((3’), /® 1m ,1 -S\3£H (31
é. ((1)) ¢3ECN 2? Q4: \ J1“
l
12 d“ J01 K13
1 3 Cyanylation
Silicuin(1asCN-303c~) ﬂ 1
2
«I 20821
I“ \s 3'0 <12
@3811 \ SCN “i
9 N
12 K13 J

Sillculn(1ZSCN-30$CN)

Figure 3.14. Didactic illustration of possible scrambling in a singly reduced species,
[sillucin(l3SH-3OSH)], illustrated for the case in which the disulﬁde Cys13-Cys30 has
been reduced, and a trace amount has dissociated to form the thiolate ion, [sillucin(13S'-
3OSH)], which scrambles to [sillucin(1287-3OSH)] before cyanylation is accomplished.

121

V. Conclusions

In summary, the cysteine-rich and apparently highly knotted anti-microbial peptide
sillucin, which is resistant to the proteolytic mapping approach to disulﬁde mapping, is
amenable to partial reduction/cyany]ation/CN-induced cleavage mass mapping
methodology. One of four possible singly reduced isoforrns and one of six possible
doubly reduced isoforrns of sillucin generated sufﬁcient information from mass mapping
of the cyanylation/CN-induced cleavage products to assign the disulﬁde connectivity
(Cys2-Cys7, Cyle-Cys24, Cysl3-Cy530, and Cysl4-Cys21) and rule out the other 104

isomeric possibilities.

VI. References

1. Somkuti, G. A. & Walter, M. W. Antimicrobial polypeptide synthesized by
Mucor pusillus NRRL 2543. Proc. Soc. Exp. Biol. Med. 133, 780-785. (1970).

2. Somkuti, G. A. & Greenberg, R. Antimicrobial peptides of thermophilic Mucor
ﬁmgi. Dev. Ind. Microbiol. 20, 661-672 (1979).

3. Bradley, W. A. & Somkuti, G. A. The primary structure of sillucin and
antimicrobial peptide ﬁom Mucor pusillus. FEBS Lett. 97, 81-83. (1979).

4. Hancock, R. E. W. & Lehrer, R. Cationic peptides: a new source of antibiotics.
Trends Biotechnol. 16, 82-88 (1998).

5. Ganz, T. & Lehrer, R. I. Antibiotic peptides from higher eukaryotes: biology and
applications. Mol. Med. Today 5, 292-297 (1999).

6. Smith, D. L. & Zhou, Z. R. Strategies for locating disulﬁde bonds in proteins. in
Methods in Enzymology (ed. McCloskey, J. A.) Vol. 193, 374-389. (Academic
Press, San Diego, CA, 1990).

7. Maeda, K., Wakabayashi, S. & Matsubara, H. Disulﬁde bridges in an alpha-
amylase inhibitor from wheat kernel. J. Biochem. 94, 865-870. (1983).

8. Poerio, E., Caporale, C., Carrano, L., Pucci, P. & Buonocore, V. Assignment of
the ﬁve disulﬁde bridges in an alpha-amylase inhibitor from wheat kernel by fast-

122

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

atom-bombardment mass spectrometry and Edman degradation. Eur. J. Biochem.
199, 595-600. (1991).

Morris, H. R. & Pucci, P. A New Method For Rapid Assignment of S-S Bridges
in Proteins. Biochem. Biophys. Res. Commun. 126, 1122-1128 (1985).

Zhang, D. & Liang, S. Assignment of the three disulﬁde bridges of huwentoxin-I,
a neurotoxin from the spider selenocosmia huwena. J. Protein Chem. 12, 735-740.
(1993)

Raschdorf, F., Dahinden, R., Maerki, W., Richter, W. J. & Merryweather, J. P.
Location of disulﬁde bonds in human insulin-like growth factors (IGFs)

synthesized by recombinant DNA technology. Biomed. Environ. Mass Spectrom.
16, 3-8 (1988).

Jones, M. D., Patterson, S. D. & Lu, H. S. Determination of disulﬁde bonds in
highly bridged disulﬁde- linked peptides by matrix-assisted laser
desorption/ionization mass spectrometry with postsource decay. Anal. Chem. 70,
136-143 (1998).

Oda, Y., Matsunaga, T., Fukuyama, K., Miyazaki, T. & Morimoto, T. Tertiary
and quaternary structures of 0.19 alpha-amylase inhibitor from wheat kernel
determined by X-ray analysis at 2.06 angstrom resolution. Biochemistry 36,
13503-13511 (1997).

Hrabal, R., Chen, Z. G., James, S., Bennett, H. P. J. & Ni, F. The hairpin stack
fold, a novel protein architecture for a new family of protein growth factors. Nat.
Struct. Biol. 3, 747-752 (1996).

Gray, W. R. Disulﬁde structures of highly bridged peptides: a new strategy for
analysis. Protein Sci. 2, 1732-1748. (1993).

Gray, W. R. Echistatin disulﬁde bridges: selective reduction and linkage
assignment. Protein Sci. 2, 1749-1755. (1993).

Yamashita, H., Nakatsuka, T. & Hirose, M. Structural and Functional-
Characteristics of Partially Disulﬁde-Reduced Intermediates Ovotransferrin N-
Lobe - Cystine Localization by Indirect End-Labeling Approach and Implications
for the Reduction Pathway. J. Biol. Chem. 270, 29806-29812 (1995).

Li, F. & Liang, S. P. Assignment of the three disulﬁde bonds of Selenocosmia
huwena lectin-I from the venom of spider Selenocosmia huwena. Peptides 20,

1027-1034 (1999).

Wu, J. & Watson, J. T. A novel methodology for assignment of disulﬁde bond
pairings in proteins. Protein Sci. 6, 391-398 (1997).

123

20.

21.

22.

23.

24.

Yang, Y., Wu, J. & Watson, J. T. Disulﬁde mass mapping in proteins containing
adjacent cysteines is possible with cyanylation/cleavage methodology. J. Am.
Chem. Soc. 120, 5834-5835 (1998).

Roepstorff, P. & Fohlman, J. Proposal for a common nomenclature for sequence
ions in mass spectra of peptides. Biomed. Mass. Spectrom. 11, 601 (1984).

Biemann, K. Contributions of mass spectrometry to peptide and protein structure.
Biomed. Environ. Mass Spectrom. 16, 99-111. (1988).

Gray, W. R. Disulﬁde bonds between cysteine residues. in Protein Structure: A
Practical Approach (ed. Creighton, T. 13.), 164-186 (IRL Press at Oxford
University, New York, NY, 1997).

Wu, J. & Watson, J. T. Optimization of the cleavage reaction for cyanylated
cysteinyl proteins for efﬁcient and simpliﬁed mass mapping. Anal. Biochem. 258,
268-276 (1998).

124

CHAPTER 4
DETERMINATION OF THE CY STEINE STATUS OF THE CHLOROPLAST

GLYCERALDEHYDE-3-P DEHYDROGENASE SUBUNITS

I. Introduction

The NADP (nicotinamide adenine dinucleotide phosphate)-linked chloroplast
glyceraldehyde-3-P dehydrogenases (GAPD) (EC 1.2.1.13) of ﬂowering plants consist
of two subunits. The A subunit is similar in length and in sequence to the eucaryotic and
eubacterial NAD (nicotinamide adenine dinucleotide) -linked enzyme (EC 1.2.1.12)
protomers. The B subunit has a 27- to 29-residue C-terrninal extension that contains two
invariant cysteines (Figure 4.1). B subunits have only been found in angiosperrns. In
species ranging from cyanobacteria to angiosperrns, NADP-linked glyceraldehyde-3-P
dehydrogenases are light activated (1, 2). Activation involves reduction of an inhibiting
disulﬁde bond or bonds. While modeling indicates that two Cys residues in the A subunit
are almost certainly responsible for the redox-sensitivity of the enzyme in green algae
and non-ﬂowering plants (3), the presence of the conserved Cys pair in the extended
angiosperrn B subunit suggests that it, too, might be involved in redox-regulation. The
purpose of the present experiments was to determine the status of the cysteines in the A
and B subunits of pea chloroplast glyceraldehyde-3-P dehydrogenase and which disulﬁde

bond(s) is responsible for the redox-sensitivity of the enzyme.

125

20

l gdl AVKVG INGFGRIGRNVFRAAL - - KNPD I WVAVNDLTDANTLAHLLKYDSVHGRLDAEVSVNGNN -
pea a KQLKVA I NGFGR I GRNFLRCWHGRKDS PLDVI AI NDTGGVKQASHLL KYDSTLG I FDADVKPVGTDG
pea b KLKVAINGFGRIGRNFLRCWHGRKDS PLEV I VVNDSGGVKNASHLLKYDSMLGTFKAKVKI LNNET

")7 g MVNVAVNGYGTIGKRVADAI I - - KQPDMKLVGVAKTS P NY ELF I AHRRG -----------------

 

19
1 gdl LVVNGKEIIVKAERDPENLAWGEI ------------ GVDIWBST-GRFTKREDAAKKLEAGAKKVI
pea a I SVDGKVI KVVSDRNPANLPWKEL ------------ GI DLVI EGT- GVFVDREGAGRHITAGAKKVL
pea b I 'I'VDGKPIKVVSSRDPLKLPWAEL ------------ GIDIVIEGT- GVFVDGPGAGKHI QAGAKKV I
lb7g ------- IR I YVP QQS I KKFEESG I PVAGTVEDLI KTSDIWD'IT PNGVGAQY - KP I YLQL - QRNAI

 

154 158
I
lgdl ISAPAK- - NEDI - TIVMGVNQDKYDPKAI-IIIVISNASCTI'NCMPFAKVLHBQFGIVRGWVHSYT
pea a I TAPGK- - G - D I PTYWGVNADAYTH - ADDI I SNAS CTTNCLAPFVKVLDQKFGI I KGTMTTTHSYT
pea b ITAPAK - - GADI PTYV I GVNEQDYGH EVADI ISMSCWNCLAPPAKVLDBEFGIVKGWSYT
lb7g FQGGEKAEVADI - S FSALCNYNEALG - - KKYIRWSCNTTALLRTI CTVNKVSKVEKVRATIVRRAA

155 159

 

l gdl NDQRI LDLPHKDLRRARAMES I I PTT - — — - MVALVLPELKWVPTPNVBVVDLV
pea a GDQRLLDASHRDLRRARAAALN IVPTS - - - - TGAAKAVALVLPTLKGKLNGIALRVPTPNVSWDLV
pea b GDQRLLDASHRDLRRARMALNIVPTS - - - -TGMVSLVLPQLKGKLNGIALRVPTPNV8WDLV

 

 

 

lb7g DQKEVKK ---------- GPINSLVPD PATVPSHHAKDVNS VIRNL-DIAMVIAPTTLMHMHFIN
27 8 2 8 9
l l
lgdl AELEKEV- TVEEVNAALKAMEGELKGILAYSEE - - - PLVSRDYN ----------- GSTVSSTIDAL
pea a VQVS KKTF - AEEVNEAFRESAAKELTGI LSVCDE - - - PLVSVDFR ----------- CTDVSSTVDSS
pea b VNVAKKG I SAEDVNMFRKAABGPLKGILDVCDV- - -PLVSVDFR ----------- CSDVSTTIDSS
1b7g ITLKDKV - EKKDI LSVLEI N ------ TPRIVLI SS KYDAEATAE LVEVARDLKRDRNDI PE - VMIFSD
| I
280 2 91

l gdl STMVI DGKMVKVVSWYDNETGYS HRVVDLAAY I ASKGL

pea a LTMVMGDDLVKVI AWYDNEWGYSQRWDLADIVANNW K

pea b LTMVMGDDHVKVVAWYDNEWGYSQRVVDLAHLVANKWPGTPKVGSGDPLEDFCETNPADEECXVYE
1b7g S I YVK-DDEVMLMYAVHQES IVVPENIDAI RASMK LMS --------- AEDSMRITNESLGI LKGYLI

354 363

 

Figure 4.1. The Sequence alignment of the pea chloroplast glyceraldehyde-3-P
dehydrogenase A (gi [“gi” is the National Center for Biotechnology Information protein
identiﬁcation number] 66025) and B (gi120663) subunits (pea a and pea b), the
Sulfolobus solfataricus glyceraldehyde-3-P dehydrogenase (Protein Data Bank (PDB) ﬁle
1b7g), and the Bacillus stearothermophilus glyceraldehyde-3-P dehydrogenase (PDB ﬁle
lgdl). Identical residues in the B subunit (pea b) and in one or both of the sequences
from the Protein Data Bank are bolded. Residues that are superimposable in the two
crystal structures are underlined. Cysteines in pea a and pea b are numbered, above the
residue for pea 8, below for pea b. Dashes indicate gaps introduced to optimize the

alignment.

126

Determining the status of cysteines in proteins includes locating free cysteines
that have sulﬂiydryl groups and disulﬁde cysteines that are involved in forming disulﬁde
bonds. The conventional methodology usually involve several steps: 1) Modify free
sulﬂiydryl groups in free cysteine residues, usually by an irreversible reaction, such as
alkylation; 2) Cleave the protein backbone by enzyme(s) or chemical reagent(s); 3)
Separate and identify the cysteine-containing protein fragments by Edman sequencing or
mass spectrometry; 4) From the protein fragments that contain modiﬁed cysteines,
deduce the location of free cysteines; and from the protein fragments that contain intra-
molecular disulﬁde bonds or are connected by inter-molecular disulﬁde bonds, deduce
the disulﬁde structure (Figure 4.2).

While this methodology was widely used in the past, there are several limitations
associated with it. First, since most alkylation’s reactions have to be done under alkaline
conditions (pH >7), disulﬁde/thiol exchange (Figure 3.15 in Chapter 3 and reference 4)
might occur and lead to artifactual results. In addition, deducing the location of a certain
free cysteine requires the successful identiﬁcation of the protein fragment containing the
modiﬁed cysteine. Thus, nearly all of such protein fragments need to recovered. This
may not be always possible due to chemical suppression (Cysteine-containing pieces are
typically a very small percentage of a large number of cleaved protein fragments.) and

signal suppression during analysis such as mass spectrometry.

127

SH

 

 

 

 

 

 

SH S-—-S S S
1 I I l I I ,0
10 20 30 4O 50 60
1)Alkylation
SR
SR S—S S S
MIMI IT TIM HIM
E E E E E E E E E E E E
2) Enzymatic Digestion

Many Non-Cysteine-Containing Protein Fragments

 

+ 40
T
SR SR SL_ ? s\
I S
10 50 20 30 _X_
60
3) Analysis by
Mass Spectrometry
l 4) 4) Data Processing

Location °f Disulﬁde Structure

Free Cysteines

Figure 4.2. The conventional methodology for determining the status of the cysteines in a
hypothetical protein of 70 amino acid residues. “SR” designates an alkylated cysteine
residue; and “E” together with an arrow indicates where the enzyme cleaves.

128

An alternative mass mapping methodology based on cyanylation and the
following CN-induced cleavage was introduced to determine the location of the free
cysteines in proteins (reference 5 and Figure 4.3). Since cyanylation by reagent l-cyano-
4-dimenthylamino-pyridinium (CDAP) tetraﬂuoro-borate can be carried out under acidic
conditions (pH 3), disulﬁde/thiol exchange (reference 4, Figure 3.15) is minimized.
Disulﬁde/thiol exchange usually occurs only after the protein is denatured and free
cysteines move more “freely” towards disulﬁde bonds. Therefore, denaturing the protein
with a large excess of the cyanylation reagent CDAP may further minimize disulﬁde/thiol
exchange.

As shown in Figure 2.9 (in chapter 2 and reference 6), cyanylation of the
sulfhydryl group of a single free cysteine residue in a peptide followed by CN-induced
cleavage in aqueous ammonia will produce two fragments (6-8). Recovery and
identiﬁcation of only one of the two fragments will be sufﬁcient to determine the location
of the free cysteine. Furthermore, cyanylation and the following CN—induced cleavage
lead to a much simpler mixture compared to that resulting ﬁ'om enzymatic digestion. If
there are n free cysteines in the peptide, only n+1 CN-induced fragments will be
produced by the cyanylation and CN-induced cleavage process. Therefore, the mass
mapping methodology based cyanylation and CN-induced cleavage may be less

vulnerable to chemical and signal suppressions.

129

SH

 

 

 

 

SH S —- S S S
10 20 30 40 50 60
1) Cyanylation
SCN
SCN S — S S S
I I I I I ,0

 

 

1 0 20 30 4O 50 60
l 2) CN-induced Cleavage

 

 

 

 

S S s s
1— 9 itz-10 I I I 49 itz-50 —I-—70
20 30 40 60
3) Reduction of SS
SH SH SH SH
1 — 9 itz-10 I I L 49 itz-5O -I——70
20 30 40 60
4) Analysis by
Mass Spectrometry
V
5) Data Processing
Location of

Free Cysteines

Figure 4.3. Cyanylation and CN-induced cleavage process to identify ﬁee cysteines in a
hypothetical protein of 70 amino acid residues.

130

II. Materials and Methods

Enzyme Isolation

The wild type NADP-linked glyceraldehyde-3-P dehydrogenase used in the mass
mapping experiments was puriﬁed from pea plants by a modiﬁcation of the method of
Anderson et al. (9). Prior to the acetone fractionation step, the suspended MgClz-
Polyethylene Glycol (PEG) fraction was passed through a Diethylarninoethyl (DEAE)
Sepharose column (10 mL Sigma Fast Flow DEAE) in 10 mM KHPO4, pH 7.8 buffer.
The glyceraldehyde-3-P dehydrogenase in this ﬁ'action was not retained on the column.
It emerged in a milky fraction. Aﬁer acetone ﬁactionation, the cloudy 50% acetone
supernatant was applied to a second 10-mL DEAE column in 10 mM
Tris(hydroxymethyl)aminomethane hydrochloride (Tris-HCl), pH 7.8, the column was
washed overnight with the Tris-HCl buffer and eluted with a linear phosphate gradient

(200 mL IO-mM KZHPO4, pH 7.8 and 200 mL 0.4-M KZHPO4, pH 7.0). In some

experiments, including those represented by Figure 4.4 and Figure 4.5, 10 mM
mercaptoethanol was included in the buffers. In other experiments, including the
experiments represented by Figure 4.6 and Figure 4.7, mercaptoethanol was omitted.
Essentially identical results were obtained with or without inclusion of mercaptoethanol
in the buffers during isolation. Chromatography on a Phenyl Sepharose column (10) was
added as a ﬁnal step in the puriﬁcation of the enzyme used in the mass mapping
experiments reported here. This isolation procedure resulted in the co-puriﬁcation of the
glyceraldehyde-B-P dehydrogenase and an 8.3-kDa protein, which is apparently the small

chloroplast protein designated CP12 (11, 12).

131

Cyanylation, HPLC separation and CN-induced cleavage

To 150 pg of enzyme in 50 pL of 50-mM pH 7.0 ammonium acetate buffer
(approximately 3.8 nmol of A and B subunits in total) was added 2,400 nmol of l-cyano-
4-dimethy1amino-pyridinium (CDAP) tetraﬂuoroborate (in 24 uL of 0.1-M pH 3.0
sodium citrate buffer, 12.0 M guanidine hydrochloride; this represents approximately a
120-fold equivalent excess of CDAP to the free sulﬂiydryl content). Cyanylation of the
free sulfhydryl groups was accomplished during incubation at room temperature for 30
minutes.

The mixture of cyanylated A and B subunits of GAPD was separated by reversed
phase high performance liquid chromatography (HPLC) at a ﬂow rate of 1.0 mL/min
using gradient elution (10% B @0-5 min -) 40% B @20 min -) 80% B @60 min)
(solvent B was 90% (v/v) acetonitrile/0.1% triﬂuoroacetic acid (TFA); solvent A was
0.1% TFA in water) with Waters model 6000 pumps. UV detection was done at 215 nm.
A Vydac C18 (catalog number 218TP54, S—pm particle size, 300-A pores, 4.6 x 250 mm)
column was used. Fractions (approximately 0.8~1.0 mL) were collected manually. The
masses of the collected species in each fraction were determined by matrix-assisted laser
desorption/ionization (MALDI) time of ﬂight (TOF) mass spectrometry (MS).

Aﬁer reduction of the volume (to less than 5 uL) under partial vacuum, the HPLC
fractions containing the cyanylated A and B subunits of GAPD were reconstituted in 20
uL aliquots of 1.0 M aqueous ammonia. CN-induced cleavage of the peptide chain on
the N-terminal side of the cyanylated cysteine residues (8, 6) was accomplished during

reaction at room temperature for one hour. Excess ammonia was removed under

132

vacuum. After addition of 40 uL of 30% (v/v) acetonitrile/0.1% TFA to the dried

cleavage products, the samples were analyzed by MALDI-TOF-MS.

MALDI-TOF-MS

MALDI mass spectra were obtained on a Voyager DE-STR or DE-Elite mass
spectrometer (Perkin-Elmer Biosystems Inc., Framingham, MA) equipped with a 337-nm
nitrogen laser. The accelerating voltage in the ion source was set at 25 kV. Mass spectra
were acquired in the positive linear mode of operation. For analysis of the intact enzyme
subunits, the mass-to-charge ratio (m/z) scale was calibrated externally using bovine
ribonuclease A (average mass of singly protonated molecule (MH+): 13683.2 Da) and
bovine serum albumin (average mass of MH+z 66431 Da) obtained from Sigma Chemical
Co. (St. Louis, MO). To analyze mixtures of CN-induced cleavage fragments, bovine
ribonuclease A was added to the sample and the m/z scale was internally calibrated on the
singly- and doubly- protonated ribonuclease A peaks (m/z values are 13683.2 and 6842.1,
respectively). The m/z values of the two most intense peaks were then determined.
Because mixing ribonuclease A with the sample caused signal suppression, several less
sensitive components in the sample were not detected when the sample was mixed with
ribonuclease A. These less sensitive components were detectable if the sample alone was
analyzed by MALDI-MS. In order to determine the accurate masses of those less
sensitive components, the mass spectrum of the sample alone was acquired and calibrated
with the accurate m/z values of the two most intense peaks that were determined as

described above.

133

Analysis of the intact and cyanylated enzyme subunits was performed using
sinapinic acid (Aldrich Chemical Co., Milwaukee, WI); all other analyses were
performed using a-cyano-4-hydroxycinnamic acid (Aldrich Chemical Co.) as the matrix.
Saturated matrix solutions were prepared in a 50% (VN) aqueous solution of
acetonitrile/0.2% TFA. Just prior to analysis, 2 1.1L of matrix solution and 2 uL sample
solution were mixed in a vial; then 2 uL of the mixed solution were applied to a stainless
steel sample plate and allowed to air-dry before being introduced into the mass

spectrometer.

III. Results
1. Mass Mapping Based on Cyanylation and CN-induced Cleavage

When the enzyme puriﬁed as described in ‘Materials and Methods’ was subjected
to MALDI-TOF-MS, three species were detected: mass spectral (MS) peaks at m/z
36218.9, 18101.8, and 12125.2 corresponding to the singly, doubly, and triply protonated
A subunit of GAPD (calculated mass: 36185.4), and MS peaks at m/z 39341.3, 19691.8
and 13122.5 corresponding to the singly, doubly, and triply protonated B subunit of
GAPD (calculated mass: 39304.1 Da) [Figure 4.4]. Proteins with molecular weights of
8.0 kDa, 8.3 kDa, and 16.5 kDa were also detected. After cyanylation and HPLC
separation on a C18 column, four proteins were found, two corresponding to cyanylated
products of an 8.3-kDa protein (HPLC peak 1 and 2), one corresponding to the
cyanylated B subunit of GAPD (HPLC peak 4) and one corresponding to the cyanylated
A subunit (HPLC peak 5) [Figure 4.5 ] as determined by MALDI-MS analysis. The ratio

of the area under the HPLC peaks corresponding to the A and B subunits was 1:17

134

 

100

 

 

 

 

 

m.
90 Q g
80‘ ‘9 a co
5‘ 70' 8 8
M w ‘-
C I
3 6° N. no 3
.5 so g ‘5 37
o N I“ m. “ 8
3‘ 40 ‘°° to' N <1- 22
N 53 N
3°‘ 3 :2 §
20
10 ‘
o v v - v
7381 .0 14592.0 21 824.2 29055.8 38287.4 43519.0

mlz

Figure 4.4. The MALDI mass spectrum of the chloroplast glyceraldehyde—B-P
dehydrogenase. luL of a solution of 9 pmol enzyme subunits in 12 mM pH 7-
ammonium acetate buffer was used.

135

 

300
250- 3
200- 1 I

 

 

150- 2

 

100‘
501

 

 

UV Response ImV

 

 

 

 

O I l ﬁ l T

20 25 30 35 40 45 50
Retention Time [min

Figure 4.5. The HPLC chromatogram of the cyanylated A and B subunits of chloroplast
glyceraldehyde-3-P dehydrogenase (GAPD). The cyanylated proteins (approximately 3.8
nmol total of the A and B subunits) were separated by reverse phase HPLC on a Vydac
C 18 column. UV detection was done at 215 nm. For details see ‘Methods’. Peaks 1 and
2 represent an 8.3k-Da protein; peaks 4 and 5 represent the cyanylated B and A subunits
of GAPD, respectively, as determined by MALDI-TOF-MS. Nothing was detected from
the fraction (represented by peak 3) by MALDI-TOF-MS. The arrows indicate the
starting and stopping points for collecting fractions represented by peaks 4 and 5,
respectively.

136

 

As shown in Figure 4.6 and Table 4.1, mass-mapping analysis of the cleavage
products of the cyanylated A subunit of GAPD indicated the presence of CN-induced
fragments 1-19, itz-20-153, itz-158-277, itz-278-288, and itz-289-337 (where itz =
iminothialzolidine-carboxyl blocked amino terminus). According to the CN-induced
cleavage reaction shown in Figure 2.9 in Chapter 2, 1-19 resulted ﬁ'om CN-induced
cleavage on the N—terminal side of cyanylated Cys 20; and thus, in the original A subunit
of GAPD, Cys 20 was a free cysteine that was not involved in a disulﬁde bond.
Similarly, Cys 154, Cys 158, Cys 278, and Cys 289, were identiﬁed as free cysteines in
the original A subunit. In summary, all ﬁve cysteines in A subunit are free cysteines.

Mass mapping of the CN-induced cleavage fragments of the cyanylated B subunit
indicated that there were two different subunit B species in the original sample, namely
the oxidized form and the reduced form. The presence of 1-18, itz-19-154, itz-159-279,
itz-280-290, and itz-291-367 (Figure 4.7 and Table 4.2) in the cleavage reaction mixture
indicates the presence of the oxidized form of subunit B, which has a disulﬁde bond
between Cys 354 and Cys 363 and free sulthydryls at ﬁve other cysteines residues. The
presence of itz-291-367 provides evidence for the Cys 354 - Cys 363 disulﬁde bond;
otherwise, the fragment itz-291-367 would not exist. Cyanylation and CN-induced
cleavage would have occurred at Cys 354 and Cys 363 had the cysteines been free. The
mass of the singly-protonated ﬁ'agment, itz-291-367, resulting from CN-induced cleavage
at cyanylated Cys 291, was determined to be 8570.5i0.3 (8570.5 is the average of 13
measurements using calibration by internal standards; 0.3 is the conﬁdence interval at the
level of 95%). Further, the data conﬁrm that itz-291-367 (see Figure 4.8 for the structure

of itz-291-367) contains a disulﬁde bond, namely Cys 354 - Cys 363 (with overall

137

 

 

 

 

 

 

 

 

 

 

 

g Portlon1
7:40000- “.2
‘E N
a /
2‘30000«
E ..
m A
Ezoooo- v E;
a éo § §
.3 N I'D N N
= 100001 ,5 8
8 R "
c 1-
0.1L- 0. A I- -II In. I.
1400 1600 - 1800 2000 2200 2400
mlz
5000* 2 Portlon2
c 8
.2 ‘ m
a 4000 /
E;
E 30001 A 1‘
~‘-‘ 3 as:
2 a 0. 1‘ g t a Is
v2°°°':‘e 8:35?» 1‘ 3:
IE. I)" m a; N N E E 1" : 3 I" to
31000‘83 I: “8'8 3 g "6'3 SQ
3 g 3 l 3 ./ g ‘:9‘: 31‘ ﬂ 9 8%
C ‘ I Q 1- ‘. ‘-
0
6000 8000 10000 12000 14000 16000
mlz

Figure 4.6. The MALDI mass spectrum of the CN-induced cleavage ﬁ'agments of the
cyanylated A subunit of chloroplast glyceraldehydes-3-P dehydrogenase represented by
HPLC peak 5 in Figure 4.5. Portions l and 2 are the low-mass and high-mass portions,
respectively. The mark “B” in parenthesis indicates that the peak represents a CN-
induced cleavage fragment of the cyanylated B subunit because HPLC peaks 4 and 5
(Figure 4.5) were not completely resolved. The mark “++” designates a peak
representing a doubly protonated (also doubly charged) protein fragment. All other peaks
without such marks are singly protonated protein fragments. The mark “?” indicates that
the source of the peak is not identiﬁed.

138

Table 4.1. The experimental and calculated m/z values for the protonated CN-induced

cleavage fragments of the cyanylated GAPD subunit A.

 

 

CN-induced Fragment Calculated m/z Experimental m/z
1-19 2188.6 (+1)a 2188.6
itz-20-153 14087.9 (+1) 14084.7

7044.5 (+2) 7042.1
itz-154-157 463.5 (+1) Not detected b
itz-158-277 12824.0 (+1) 12821.1

6412.5 (+2) 6411.9
itz-278-288 1305.5 (+1) 1305.0
itz-289-337 5537.2 (+1) 5537.2

 

3‘ ‘+1’ designates singly-protonated (also singly-charged) species and ‘+2’ designates
doubly protonated (also doubly charged) species.

1’ Due to signal suppression by matrix molecules in MALDI.

139

 

 

 

 

 

 

 

 

 

g 50000. g Portion1
E 40000- §
8 / 2
= 30000- VA
2 :3 s
‘5 200001 “I 59 °°' 8?
'5 ‘9. K’ "g 3;
§ 10000I g 9 * “I g
s - “:3
DIN g A II, III ,
1000 1200 1400 1600 1800 2000 2200 2400
mlz

5000 Q
:7 g Portion2
“g 4000 1 A a g
E‘ I A +' 1‘
g 30004 3; 2 3 Tag, /
'E I" : F. m 4‘ A
$20001§ 3.0352: .— $5
'a “’8 ‘° N/ a a 8 8 N 3 ..- 2
g 1000 . co g ,3 s: 8 3 E g g g
o-I ‘- v (p
5 / I. I \\ I .- ‘- ‘-

0 Tara‘wwf: ;..

 

 

 

4000 6000 8000 10000 12000 14000 16000
mlz

Figure 4.7. The MALDI mass spectrum of the CN-induced cleavage fragments of the
cyanylated B subunit of chloroplast glyceraldehyde-S-P dehydrogenase represented by
HPLC peak 4 in Figure 4.5. Portions l and 2 are the low-mass and high-mass portions,
respectively. The mark “A” in parenthesis indicates that the peak represents a CN-
induced cleavage fragment of the cyanylated A subunit because HPLC peaks 4 and 5
(Figure 4.5) were not completely resolved. The mark “++” designates a peak
representing a doubly protonated (also doubly charged) protein fragment. All other peaks
without such marks represent singly protonated protein fragments. The mark “7”
indicates that the source of the peak is not identiﬁed. The peak at m/z 6962.1 represents
both a singly protonated CN-induced cleavage fragment (itz—291-353, marked as “+, B,”;
calculated m/z for MH+: 6960.8) from the cyanylated subunit B-reduced and a doubly
protonated CN—induced cleavage/B-elimination ﬁ'agment (itz-159-[3@280-290) [Table
4.2] formed by B-elimination at the cyanylated Cys280 instead of cleavage (marked as
“++, B”; calculated m/z for MHZ”: 6962.6).

140

Table 4.2. The experimental and calculated m/z values for the protonated CN-induced
cleavage fragments of the cyanylated GAPD subunit B-oxidized

 

CN-induced Fragment L
1-18

itz-19-1 54

itz-155-158

itz-159-279

itz-280-290

itz-291-367 C

Calculated m/z

" 2060.5 (+1)a "

14426.5 (+1)
7213.8 (+2)
463.5 (+1)
12726.8 (+1)
6363.9 (+2)
1275.5 (+1)
8570.5 (+1)

4285.3 (+2)

Experimental m/z

 

"2060.5

14426.7
7212.9

Not detected b
12725.1
6362.1

1275.2

8570.5

4285.9

 

a ‘+1’ designates singly-protonated (also singly-charged) species and ‘+2’ designates

doubly-protonated (also doubly-charged) species.
b Due to signal suppression by matrix molecules in MALDI.

C There is a disulﬁde bond (Cys 354 - Cys363) in itz-291-367.

141

calculated mass of 8570.5) rather than two ﬁee sulﬂiydryls, at Cys 354 and Cys 363 (with
overall calculated mass of 8572.5). Detection of 1-18, itz-19-154, itz-159-279, itz-280-
290, itz-291-353, and itz-354-362 (Figure 4.7 and Table 4.3) indicates the presence of the
reduced form of subunit B, which has free sulthydryls at all seven cysteine residues. The
CN-induced cleavage reactions of the oxidized and reduced forms of subunit B are
illustrated in Figure 4.8.

Because GAPD was simultaneously denatured in 4 M guanidine hydrochloride
and cyanylated at acidic pH (pH 3), the possibility of scrambling between thiolate ions
(from ionization of the free cysteine sulﬂiydryl groups) and a disulﬁde bond was
minimized. The experiments reported here demonstrate that it is feasible to determine the
cysteine status of various forms (such as partially oxidized and reduced forms of GAPD
subunit B) of a protein using methodology based on cyanylation, CN-induced cleavage,
and mass mapping, even when these different forms cannot be separated by HPLC.

As is apparent from examining Figure 4.6 and Table 4.1, the expected cleavage
fragment itz-154-157 from the cyanylated A subunit was not detected by MALDI-MS
because of suppression by matrix molecules used in MALDI. However, determination of
the cysteine status was still possible because there were two CN-induced cleavage
fragments related to cleavage at one cyanylated cysteine residue: one to the N-terminal
side and the other to the C-terminal side of the cleavage site. Detecting only one of the
two CN—induced cleavage fragments will allow for the identification of the cyanylated
cysteine residue. This methodology, then, does not require complete detection of all the
CN—induced cleavage fragments. Similar cases of signal suppression are shown in Figure

4.7, Table 4.2 and Table 4.3.

142

Table 4.3. The experimental and calculated m/z values for the protonated CN-induced
cleavage fragments of the cyanylated GAPD subunit B-reduced.

 

CN-induced Fragment

Calculated m/z

Experimental m/z

 

1-18

itz-19-154

itz-155-158

itz-159-279

itz-280-29O

itz-291-353

itz-354-362

itz-363-367

2060.5 (+1 ) a
14426.5 (+1)
7213.7 (+2)
463.5 (+1)
12726.8 (+1)
6363.9 (+2)
1275.4 (+1)
6960.8 (+1)
1033.0 (+1)

666.8 (+1)

2060.5
14426.7
7212.9

Not detected b
12725.1
6362.1

1275.2

6962.1

1032.6

Not detected b

 

a ‘+1’ designates singly-protonated (also singly charged) species and ‘+2’ designates

doubly-protonated (also doubly charged) species.
'0 Due to signal suppression by matrix molecules in MALDI.

143

A) The Cyanylated GAPD-Subunit-B-Oxldized-Form
SCN SCN SCN SCN SCN S—S

1 1 1 1 1 1 1 | 367
19 155 159 280 291 354 363

CN-induced Cleavage

 

 

1—18 02-19—154 "2455—158
1 _i’
'tz-1 — 279 itz-280 —- 290 itz-291 367
' 59 354 363
B) The Cyanylated GAPD-Subunit-B-Reduced-Form
SCN SCN SCN SCN SCN SCN SCN
. l I l l l | I
19 155 159 280 291 354 363
CN-induced Cleavage
1—18 itz-19—154 itz-155—158
itz-159 —- 279 itz-280 —— 290
itz-291 — 353 itz-354 — 362 itz-363 — 367

Figure 4.8. The CN—induced cleavage fragments from the oxidized and reduced forms of
the cyanylated subunit B of GAPD.

144

In Figure 4.5, HPLC peaks 4 and 5 (corresponding to the cyanylated B subunit
and A subunit, respectively) were not completely resolved. This is reﬂected in the
MALDI spectra [Figure 4.6 and Figure 4.7] of the CN-induced cleavage products of these
two fractions, each of which contains minor peaks representing fragments corresponding
to the CN-induced cleavage products of the other fraction. Since fractions that are not
completely separated do not interfere with the identiﬁcation of the cysteine status, care
was not taken to avoid collecting some of the fraction represent by HPLC peak 5 into the
fraction represented by HPLC peak 4 and some of the fraction represented by HPLC 4
into the fraction represented by HPLC peak 5 (Figure 4.5).

Most of the other MALDI peaks in Figure 4.6 and Figure 4.7 that were identiﬁed,
but not described above, are from side reactions such as B-elimination (Figure 1.13 in
Chapter 1), incomplete CN-induced cleavage (cleavage does not occur at all cyanylated
cysteines), and incomplete cyanylation (not all free cysteines are totally cyanylated). The
identities of those peaks are listed in Table 4.4 and Table 4.5. Formation of some
representative fragments in Table 4.4 is described below. Itz-154-SCN@158-SH@278-
288 (Table 4.4) resulted from CN—induced cleavage at 154 and 289 but no cleavage at
158 of an incompletely cyanylated species of subunit A of GAPD (Figure 4.9). 1-
SCN@20-B@154-157 (Table 4.4) resulted from CN-induced cleavage at 158, no
cleavage at 20, and B-elimination at 154 (Figure 4.9). As we can see ﬁom Table 4.4 and
Figure 4.9, identiﬁcation of these fragments does not interfere with the determination of
the cysteine status of GAPD subunits A and B because the data are consistent with the
data that we have discussed early in this section. Details about these side reactions can

be found elsewhere (chapter 3 and reference 13).

145

Table 4.4. The experimental and calculated m/z values for the fragments from B-
elimination and incomplete CN-induced cleavage of subunit A of GAPD.

 

 

CN—induced Fragment Calculated m/z Experimental m/z
itz-154-SCN@158-SH@278-288 31 14529.8 (+1) 14528.5
(or itz-154-SH@158-SCN@278-288) 7265.4 (+2) 7265.1
1-SCN@20-153 16257.5 (+1) 16256.5

8129.3 (+2) 8128.5
1-SCN@20-B@154-157 b 16643.0 (+1) 16644.6
(or 1-B@20-SCN@154-157) 8322-0 (+2) 8320.1

 

3‘ itz-l54—SCN@158-SH@278-288 designates the protein fragment from amino acid
residue 154 to 288, in which cysteine 158 is cyanylated (SCN) and cysteine 278 is free
(SH).

b B@154 designates that B-elimination (loss of HSCN) happened on the cyanylated
cysteine 154.

146

Table 4.5. The experimental and calculated m/z values for the fragments from [3-
elimination and incomplete CN-induced cleavage of subunit B of GAPD.

 

 

CN-induced Fragment Calculated m/z Experimental m/z
itz-159-B@280-290 a " 13924.2 (+1) 13922.6

6962.6 (+2) 6962.1
itz-19-SCN@155-158 b 14871.0 (+1) 14869.6

7436.0 (+2) 7435.3
1-SCN@19-154 16468.0 (+1) 16468.1
1-SCN@19-B@155-158 16853.4 (+1) 16857.5

(or l-B@19-SCN@155-158)

 

a itz-159-B@280-290 designates the protein fragment from iminothialzolidine-blocked-
residue 159 to residue 290, in which B-elimination (loss of HSCN) occurred on the
cyanylated cysteine 280.

b SCN@155 designates that cysteine 155 is cyanylated.

147

A) An lncompletely Cyanylated GAPD-Subunlt-A
(Cys 278 Was Not Cyanylated)

 

SCN SCN SCN SH SCN
1 1 1 1 1 1 33,
2O 1 54 1 58 278 289

CN-induced cleavage at 154 & 289;
no cleavage at cyanylated Cys 158

itz-154-SCN@158-SH@278-288
SCN SH

itz-1 54 I I 288
1 58 278

 

B) The Cyanylated GAPD-Subunit-A

 

SCN SCN SCN SCN SCN
1 1 1 1 1 1 337
20 1 54 158 278 289

CN-indueed cleavage at 158;
B-elimination at 154;
no cleavage at cyanylated Cys 20

1-scu@20-13@154-157

 

SCN 0
1 1 I 157
20 154

Figure 4.9. Formation of some ﬁ'agments resulted from B-elimination, incomplete CN-
induced cleavage, and incomplete cyanylation of subunit A of GAPD.

148

2. Subunit Composition

It has been suggested that the chloroplast enzyme is present in vivo as a mixture
of the A4 tetramer, the Asz tetramer and higher oligomeric forms of the A2B2 tetramer.
The recombinant B4 tetramer forms in the absence of the A subunit in Escherichia coli
(10, 14). Our results indicate a 1.7 to 1 ratio of the B to the A subunit in the pea leaf
enzyme preparation used in these experiments. There is no indication in our experiments
of the presence of degradation products of the B subunit or that the shorter subunit
corresponds to a mixture of A subunit and degraded B subunits. We suggest that the
wild-type tetramer represents at least one additional species containing at least three B
subunits. More likely, it represents the ﬁve possible combinations of A and B subunits in

a tetramer (B4, A1B3, A232, A313, A4).

IV. Discussions

The mass mapping analysis based on cyanylation and CN-induced cleavage
clearly indicates that the regulatory disulﬁde bond in the light-activated pea chloroplast
glyceraldehyde-3-P dehydrogenase links the two Cys residues in the C-terminal extension
of subunit B. Recently, Hutchison et al. (15) determined the redox-potential for the
inactivating disulﬁde bond in the tomato leaf chloroplast enzyme. That value then
represents the redox-potential of the C-terminal disulﬁde bond. The recombinant pea B4
glyceraldehyde-3-P dehydrogenase expressed in Escherichia coli is redox-sensitive (10).
Consistent with the present results, when the C-terminal extension is removed
genetically, the resultant recombinant mutant enzyme is signiﬁcantly less redox-sensitive

(10, 14). The regulatory disulﬁde bonds in the redox-regulated chloroplast malate

149

dehydrogenase (16, 17) and in the chloroplast fructose bisphosphatase (18) are also
located in extensions or insertions. Modeling suggests that the two Cys residues in the C-
tenninal extension of the chloroplast glyceraldehyde-B-P dehydrogenase are located in
different domains (19). If there is inter-domain movement during catalysis (20) in the
NADP-linked chloroplast enzyme, crosslinking the domains would be expected to be
inhibitory. In contrast, the disulﬁde bonds in malate dehydrogenase and ﬁ'uctose
bisphosphatase are intra-domain disulﬁde bonds. A crystal structure will be required to
conﬁrm the prediction that the C-terminal disulﬁde bond forms an inter-domain crosslink
in the chloroplast glyceraldehyde-3-P dehydrogenase (19).

There is no experimental evidence here for a second disulﬁde bond in the
angiosperm chloroplast enzyme, although modeling (19, 21) suggests a disulﬁde bond
between CysZO and Cys289 in the A subunit and another one between the corresponding
Cys residues in the B subunit. Similarly, in both malate dehydrogenase (21, 22) and
fructose bisphosphatase (23), at least one additional disulﬁde bond not seen in the
available crystal structures seems possible. Homology modeling, based on the
coordinates of the NAD-linked Bacillus stearothermophilus enzyme in PDB ﬁle lgdl
(24), implicates the Cys residues corresponding to Cys20 and Cy3289 in the dark
inactivation of the NADP-linked Chlamydomonas reinhardtii enzyme. The algal enzyme
has no C-terminal extension, and, as modeled, no other Cys residues positioned to form
disulﬁdes (reference 3; PDB ﬁles lnlg and lnlh). As lower plant and algal chloroplast
glyceraldehyde-3-P dehydrogenases are not known to contain B subunits, it seems likely
that this disulﬁde bond can form and does affect activity in many species. Presumably,

when the enzyme acquired a more effective disulﬁde bond in the C-terminal extension,

150

formation of the obsolete CysZO-Cy3289 disulﬁde was eliminated. This could have been
accomplished by a change in structure and/or charge distribution. Interestingly, the
Sulfolobus enzyme also has an inter-domain disulﬁde bond (25). This non-regulatory
disulﬁde bond is thought to stabilize the enzyme in the high-temperature environment
where the organism exists in nature.

Consistent with the mutational experiments of Baahnann et al. (14) and of Li and
Anderson (10), these results indicate that the Cys residues in the extension are
responsible for the redox-sensitivity of the higher plant chloroplast glyceraldehyde-B-P

dehydrogenases that contain B subunits.

V. References

1. Anderson, L. E. Light Dark Modulation of Enzyme-Activity in Plants. in
Advances in Botanical Research Incorporating Advances in Plant Pathology (ed.
Callow, J. A.) Vol. 12, 1-46 (Academic Press, New York, NY, 1986).

2. Buchanan, B. B. Role of Light in the Regulation of Chloroplast Enzymes. Annu.
Rev. Plant Physiol. Plant Molec. Biol. 31, 341-374 (1980).

3. Li, A. D., Stevens, F. J ., Huppe, H. C., Kersanach, R. & Anderson, L. E.
Chlamydomonas reinhardtii NADP-linked glyceraldehyde-3- phosphate
dehydrogenase contains the cysteine residues identiﬁed as potentially domain-
locking in the higher plant enzyme and is light activated. Photosynth. Res. 51,
167-177 (1997).

4. Smith, D. L. & Zhou, Z. Strategies for locating disulﬁde bonds in proteins. in
Methods Enzymol. (ed. McCloskey, J. A.) Vol. 193, 374-389 (1990).

5. Wu, J ., Gage, D. A. & Watson, J. T. A strategy to locate cysteine residues in
proteins by speciﬁc chemical cleavage followed by matrix-assisted laser
desorption/ionization time-of-ﬂight mass spectrometry. Anal. Biochem. 235, 161-
l 74. (1996).

151

10.

11.

12.

l3.

14.

15.

Wu, J. & Watson, J. T. Optimization of the cleavage reaction for cyanylated
cysteinyl proteins for efﬁcient and simpliﬁed mass mapping. Anal. Biochem. 258,
268-276 (1998).

Jacobson, G. R., Schaffer, M. H., Stark, G. R. & Vanaman, T. C. Speciﬁc
chemical cleavage in high yield at the amino peptide bonds of cysteine and
cystine residues. .1. Biol. Chem. 248, 6583-6591. (1973).

Wakselman, M. & Guibe-Jampel, E. 1-Cyano-4.dimethylamino-pyridinium salts:
New water-soluble reagents for the cyanylation of protein sulphydryl groups. J.
Chem. Soc., Chem. Commun., 21-22 (1976).

Anderson, L. E., Goldhabergordon, I. M., Li, D., Tang, X. Y., Xiang, M. H. &
Prakash, N. Enzyme-Enzyme Interaction in the Chloroplast - Glyceraldehyde- 3-

Phosphate Dehydrogenase, Triose Phosphate Isomerase and Aldolase. Planta 196,
245-255 (1995).

Li, A. D. & Anderson, L. E. Expression and characterization of pea chloroplastic
glyceraldehyde—3-phosphate dehydrogenase composed of only the B-subunit.
Plant Physiol. 115, 1201-1209 (1997).

Wedel, N., Soll, J. & Paap, B. K. CP12 provides a new mode of light regulation of
Calvin cycle activity in higher plants. Proc. Natl. Acad. Sci. U. S. A. 94, 10479-
10484 (1997). -

Wedel, N. & 8011, J. Evolutionary conserved light regulation of Calvin cycle
activity by NADPH-mediated reversible
phosphoribulokinase/CP12/glyceraldehyde-3-phosphate dehydrogenase complex
dissociation. Proc. Natl. Acad. Sci. U. S. A. 95, 9699-9704 (1998).

Qi, J. & Watson, J. T. Determination of the Disulﬁde Structure of Sillucin, a
Highly Knotted, Cysteine-Rich Peptide, by Cyanylation/Cleavage Mass Mapping.
Biochemistry 40, 4531-4538 (2001).

Baalmann, E., Scheibe, R., Cerff, R. & Martin, W. Functional studies of
chloroplast glyceraldehyde-3-phosphate dehydrogenase subunits A and B
expressed in Escherichia coli: formation of highly active A4 and B4
homotetrarners and evidence that aggregation of the B4 complex is mediated by
the B subunit carboxy terminus. Plant Mol. Biol. 32, 505-513. (1996).

Hutchison, R. S., Groom, Q. & Ort, D. R. Differential effects of chilling-induced

photooxidation on the redox regulation of photosynthetic enzymes. Biochemistry
39, 6679-6688 (2000).

152

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

Carr, P. D., Verger, D., Ashton, A. R. & Ollis, D. L. Chloroplast NADP-malate
dehydrogenase: structural basis of 1i ght-dependent regulation of activity by thiol
oxidation and reduction. Struct. Fold. Des. 7, 461-475 (1999).

J ohansson, K., Ramaswamy, S., Saarinen, M., Lemaire-Chamley, M., Issakidis-
Bourguet, E., Miginiac-Maslow, M. & Eklund, H. Structural basis for light
activation of a chloroplast enzyme: The structure of sorghum NADP-malate
dehydrogenase in its oxidized form. Biochemistry 38, 4319-4326 (1999).

Chiadmi, M., Navaza, A., Miginiac-Maslow, M., Jacquot, J. P. & Cherﬁls, J.
Redox signaling in the chloroplast: structure of oxidized pea ﬁ'uctose-1,6-
bisphosphate phosphatase. Embo J. 18, 6809-6815 (1999).

Qi, J ., Isupov, M. N., Littlechild, J. A. & Anderson, L. E. Chloroplast
Glyceraldehyde-3-P Dehydrogenase Contains a Single Disulﬁde Bond Located in
the C-terminal Extemsion to the B Subunit. J. Biol. Chem. in press (2001).

Skarzynski, T. & Wonacott, A. J. Coenzyme-induced conformational changes in
glyceraldehyde-3-phosphate dehydrogenase from Bacillus stearothennophilus. J.
Mol. Biol. 203, 1097-1 118. (1988).

Li, D., Stevens, F. J ., Schiffer, M. & Anderson, L. E. Mechanism of Light-
Modulation - Identiﬁcation of Potential Redox-Sensitive Cysteines Distal to
Catalytic Site in Light- Activated Chloroplast Enzymes. Biophys. J. 67, 29-35
(1994)

Muslin, E. H. & Anderson, L. E. Identiﬁcation of one pair of domain-locking,
redox-sensitive cysteine residues in maize chloroplast malate dehydrogenase. in
Photosynthesis, Mechanisms and Eﬂects (ed. Garab, G.) Vol. 5, 3583-3586
(Kluwer, Boston, MA, 1999).

J acquot, J. P., Lopez-Jaramillo, J ., Miginiac-Maslow, M., Lemaire, S., Cherﬁls, J .,
Chueca, A. & Lopez-Gorge, J. Cysteine-153 is required for redox regulation of
pea chloroplast fructose-l,6-bisphosphatase. FEBS Lett. 401, 143-147 (1997).

Skarzynski, T., Moody, P. C. E. & Wonacott, A. J. Structure of holo-
glyceraldehyde-3-phosphate dehydrogenase from Bacillus stearothennophilus at
1.8 .ANG. resolution. J. Mol. Biol. 193, 171-187 (1987).

Isupov, M. N., Fleming, T. M., Dalby, A. R., Crowhurst, G. S., Boume, P. C. &
Littlechild, J. A. Crystal structure of the glyceraldehyde-3-phosphate
dehydrogenase ﬁom the hyperthermophilic archaeon Sulfolobus solfataricus. J.
Mol. Biol. 291, 651—660 (1999).

153

 

CHAPTER 5
COMPUTER ALGORITHMS FOR DISULFIDE MASS MAPPING OF
PROTEINES BASED ON PARTIAL REDUCTION AND CY ANY LATION/CN-

INDUCED CLEAVAGE

I. Introduction

The number of possible disulﬁde structural isomers of a given protein containing
several cysteines increases rapidly with the number of cysteines. For example, in the
case of ribonuclease-A, which has eight disulﬁde cysteines (“Disulﬁde cysteines” are
cysteines that are involved in disulﬁde bonds), 105 isomers are possible. For a protein
having ten disulﬁde cysteines, 945 isomeric forms are possible. In general, the total
number of all possible disulﬁde structures (A disulfide structure is deﬁned as a
particular arrangement of all the disulﬁde bonds in a protein.) is (2n-1)*(2n-3)...3*1 for
a protein that has 2n disulﬁde cysteine residues that are involved in forming n disulﬁde
bonds. Clearly, in proteins containing multiple disulﬁde cysteines, computational
assistance is essential to calculate the mass of CN-induced cleavage fragments from all
possible disulﬁde structural isomers (of a given sequence) that have been processed by
partial reduction, cyanylation and CN-induced cleavage.

In principle, for a given cyanylated protein having two free sulﬂiydryls on two
free cysteines (“Free cysteines” are cysteines that have free sulﬂiydryls and are not
involved in disulﬁde bond formation; reducing a disulﬁde bond in a protein will

produced two nascent free cysteines), after cyanylation and CN-induced cleavage, only

154

three cleavage products are expected for the protein because two cyanylated cysteines
provide two cleavage sites. In practice, analysis of a protein of unknown disulﬁde
structure (but known sequence) requires the computation of the mass of each of all
possible CN-induced fragments of each of all possible disulﬁde structural isomers of the
analyte. For example, if the disulﬁde bonding in ribonuclease-A (containing 8 disulﬁde
cysteines residues) were unknown, the analyst would need to compute the masses of the
expected three cleavage fragments for each of all four possible cyanylated singly reduced
isofonns of each of the 105 possible disulﬁde structural isomers. Mass spectrometric
analysis of a cyanylated and cleaved singly reduced isoform of the protein generates three
mass spectral peaks that must be compared with the array of computed masses
corresponding to the three ﬁagments expected from each of the four possible singly
reduced isoforrns of each of the 105 different disulﬁde structural isomers of the unknown
protein. Since there are four possible singly reduced isoforrns of each of the 105 different
isomers obtained in the process of partial reduction, 105 x 4 x 3 = 1260 computations
must be made. A match between the three experimentally observed mass spectral peaks
with the mass values of three computed fragments would constitute identiﬁcation of one
of the connectivities involved in one of the four disulﬁde bonds in one of the 105 possible
isomeric forms involving four disulﬁde bonds. Clearly, the development of a suitable
algorithm to compute these masses, to coordinate a given mass set with a given
connectivity for a given isomer, and to match the experimental mass set with a calculated
one to determine the disulﬁde connectivity is essential to the success of applying the
partial reduction/cyanylation/CN—induced cleavage/mass mapping methodology to

proteins containing several cysteines.

155

 

II. A Naive Algorithm

In the computational analysis phase, the computer program must determine the
connectivity of the disulﬁde bonds (disulﬁde structure) given the mass spectral data, the
location of the original free cysteines, and the sequence of the protein. A naive
algorithmic approach to solving this problem is the following. First, the program will
compute all possible sets of masses of cleaved fragments; that is, for each disulﬁde
structure in the complete disulfide structure list A (such a list can be computed given
the sequence of the protein to be analyzed), the program computes the entire set of
masses of all the CN-induced ﬁ'agments that such a protein would be cleaved into using
our partial reduction, cyanylation, and CN-induced cleavage methodology. These are
then compared against the experimental mass spectral data using a disulfide structure
search module. Given the accuracy of the mass spectrometric measurement, we should
be able to match each experimentally measured mass spectral peak with its corresponding
calculated value. The correct disulﬁde structure is then established by determining which
set of computed fragments matches the actual data. For the correct disulﬁde structure,
there exists a perfect matching between its set of calculated values of m/z and the list of
experimentally obtained mass spectral peaks (represented by m/z values); that is, each
calculated m/z value will match exactly one experimentally obtained peak (unless the m/z
values for two fragments differ by less than the mass spectrometric error), and each
element in both the correct set of calculated values and the list of experimental peaks will
be matched. For other disulﬁde structures, there will not exist a perfect matching

between the set of calculated values and the list of experimentally obtained peaks. Thus,

156

 

the correct disulﬁde structure can be easily identiﬁed because it is the one with the
highest number of CN-induced ﬂagments matched in the ranked disulﬁde structure list.

The information ﬂow for such an algorithm is illustrated in Figure 5.1.

 

Structure m =
(2n-1)(2n-3) 1
l

Structure1 + Structure2 + ... +

l
l
I I
E r l
I

Calculated m/z for + Calculated m/zfor + ... + Calculated m/z for
Fragment Set 1 Fragm nt Set 2 Fragment Set m

 

 

Disulﬁde Structure
Search Module

 

 

Accuracy

I _________________ 'l
| Experimental m/z Set 1 1 Output I
| 4 l I ARanked Disulﬁde Structure List I
I I M I | Ordered by :
11s + m” ......m: lit‘ﬂﬂii’itaﬂﬂeﬂtiﬂaffﬂn

I

Figure 5.1. Information ﬂow for a naive algorithm for disulﬁde mass mapping based on
partial reduction, cyanylation and CN-induced cleavage. The sequence of the protein to
be analyzed and the locations of ﬂee cysteines are used as input 1; MS spectra of the CN-
induced ﬂagments and m/z measurement accuracy are used as input 2. The output is a
ranked disulﬁde structure list, in which candidate disulﬁde structures are ranked
according to the number of CN-induced ﬂagments matched (Experimental m/z values and
calculated m/z values are compared).

157

III. A Basic Algorithm

As discussed in the ﬁrst section of this chapter, the disulﬁde structure list A is
extremely long as the number of candidate structures, ((2n-1)*(2n-3)*...*3*1)), is
exponential in the number of disulﬁde cysteines (Zn) in the original protein containing n
disulﬁde bonds. Thus, the matching step in the naive algorithm will take considerable
time as the experimental data are compared against each structure in list A. Furthermore,
this is not a preprocessing step as this matching phase cannot occur until the experimental
data have been gathered. In addition, many repeated calculations and comparisons take
place in this algorithm because some of all the CN-induced ﬂagments from the chemical
processing (partial reduction, cyanylation, and CN-induced cleavage) of protein isomers
of different disulﬁde structures will be the same. An example of this is shown in Table
5.1. Therefore, it is desirable to develop alternative matching strategies that do not
require us to compare the experimental data against the entire list A.

One observation shows that while the number of possible disulﬁde structures is
large, the number of possible disulﬁde linkages is much smaller (Table 5.2). Thus, a
better strategy is to deal with disulﬁde linkages instead of disulﬁde structures.

Many times, partially reduced/cyanylated species of the original protein are
isolated by reversed phase HPLC (l, 2). Considering this, we can develop a basic
algorithm (Figure 5.2) to analyze the mass spectral data of the CN-induced ﬂagments
ﬂom each isolated species to ﬁnd out which disulﬁde bond(s) in the original protein is
reduced to form the species. Then, we can combine all the individual results to obtain the
correct disulﬁde structure. Next, the algorithms for singly reduced, doubly reduced, and

triply reduced isofonns are discussed in detail.

158

Table 5.1. Some of the same CN-induced fragments ﬂom two singly reduced isofonns of
two hypothetical protein isomers with different disulﬁde structures. The hypothetical
protein has 90 amino acid residues, among which 8 disulﬁde cysteines are located at
positions 10, 20, 30, 40, 50, 60, 70, and 80.

 

 

Disulﬁde Structure (Cyle-CysZO, (Cys10-Cy320,

Cys30-Cys40, Cys30-Cys40,

CysSO-Cys60, CysSO-Cys70,

Cys70-Cys80) Cys60-Cys80)

S-S Opened in Singly Cys10-Cy820 Cyle-Cys20
Reduced Isoform l

CN-induced Fragments ﬂom 1-9, 1-9,

Singly Reduced Isoform 1 itz—10-19, itz-10-19,

itz-20-90 itz-20-90

S-S Opened in Singly Cys30—Cys40 Cys30-Cys40
Reduced Isoform 2

CN-induced Fragments ﬂom 1-29, 1-29,

Singly Reduced Isoform 2 itz-30-39, itz-30-39,

itz-40-90 itz-40-9O

 

159

 

 

 

Table 5.2. The number of possible disulﬁde linkages and disulﬁde structures for a protein

 

 

having 11 disulﬁde bonds.
Number of Disulﬁde Number of Possible Number of Possible
Bonds in the Protein Disulﬁde Linkages Disulﬁde Structures
2 6 3
3 15 15
4 28 105
5 45 945
6 66 10,395
7 91 135,135
8 120 2,027,025
n n*(2n-1) (2n-1)*(2n-3)...3*1

 

160

 

 

I.
l Singly Reduced + Doubly Reduced + Triply Reduced |
I lsofonns Isoforms Isoforms |
1 . 1 . 1
I 1’ v V l
1 Calculated m/z of Calculated m/z of + Calculated m 0f I

l

I Fragments from Singly Fragments from Doubly Fragments for Triply

h————

 

 

Partially Reduced Species
Search Module

 

I Experimental m/z for an Isolated

|
| Partially Reduced Species I I OUtPUt- 1I
I ’1 I l A Ranked Partially Reduced Species |
l | + m/z MeasurementI l Ordered by |
I Accuracy 1 l the Number of Fragments Matched l
|

MS Spectra n ut 2

Figure 5.2. A basic algorithm for disulﬁde mass mapping based on partial reduction,
cyanylation and CN-induced cleavage. The sequence of the protein to be analyzed and
the locations of ﬂee cysteines are used as input 1; MS spectra of the CN-induced
ﬂagments and m/z measurement accuracy are used as input 2. The output is a ranked
partially reduced species list, in which candidate partially reduced species are ranked
according to matching percentage = (the number of CN-induced ﬂagments matched/the
number of such ﬂagments expected for a partially reduced species).

161

 

1. Consideration of Singly Reduced Isoforms

A given singly reduced/cyanylated isoform of the protein followed by CN-
induced cleavage and complete reduction of the residual disulﬁde bonds can provide
information leading to identiﬁcation of one particular disulﬁde bond, which was reduced
during partial reduction. In order to identify this disulﬁde bond, the mass spectral data of
the CN-induced cleavage ﬂagments should be compared against the calculated masses of
the CN-induced ﬂagments expected ﬂom each of the possible singly reduced isoforrns,
which has been cyanylated, then subjected to CN-induced cleavage followed by complete
reduction of the residual disulﬁde bonds. Although there are [(2n-1)*(2n-3)*...*3*1]
possible disulﬁde structures for a protein containing n disulﬁde bonds, only n(2n-1)
singly reduced/cyanylated isoforrns need be considered because of the degeneracy
associated with completely reducing the n-I residual disulfide bonds in these isofonns
dining chemical processing. For example, there are only 28 (=4*(2*4-1)) possible singly
reduced/cyanylated isoforrns (Table 5.3) for a protein having four disulﬁde bonds, such
as bovine ribonuclease-A, in spite of the fact that there are 105 (=(2*4-l)*(2*4-
3)*. . .*3*1) possible disulﬁde structures for the protein.

All the singly reduced isoforrns can be computed as follows. A hypothetical
protein is used to illustrate the computation process. The protein has 100 amino acid
residues, among which 2n disulﬁde cysteines (forming n disulﬁde bonds) are at positions
C1, C2, ..., C(2n) [“C” represents a cysteine residue].

Step 1. Compute all singly reduced isoforrns in which the disulﬁde bond reduced

is Cl-CG) (i = 2, 3, ..., or 2n). There are 2n-l such singly reduced isoforrns.

162

 

Step 2. Compute all singly reduced isofonns in which the disulﬁde bond reduced

is C2-C(j) (j = 3, 4, ..., or 2n). There are 2n-2 such singly reduced isoforrns.

Step i. Compute all singly reduced isoforms in which the disulﬁde bond reduced

is C(i)-C(j) (j = i+l, i+2, ..., or 2n). There are 2n-i such singly reduced isoforrns.

Step 2n-1. Compute the singly reduced isoform in which the disulﬁde bond
reduced is C(2n-l)-C(2n). There is only one such singly reduced isoform.
Counting ﬂom Step 1 to Step 2n-1, there are a total of 2n*(2n-1)/(2*l) singly

reduced isoforrns.

Once all the singly reduced isoforrns are computed, all the CN-induced cleavage
ﬂagments ﬂom such singly reduced/cyanylated isoforrns can be easily computed. For
example, when the disulﬁde bond C(i)-C(j) is reduced to form a singly
reduced/cyanylated isoform, its corresponding CN-induced fragments are l—[C(i)-l], itz-
C(i)-[C(i)-l], itz-C(j)—100 [“1-[C(i)-l]” designates the ﬂagment ﬂom residue 1 to residue

C(i)-1.].

As an example, all sets of CN-induced ﬂagments ﬂom 28 (=4*(2*4-l)) possible
singly reduced/cyanylated isofonns of ribonuclease-A (consisting of 124 amino acid
residues, 8 disulﬁde cysteines at positions 26, 40, 58, 65, 72, 84, 95, and 110 forming 4

disulﬁde bonds) are represented in Table 5.3.

163

 

Table 5.3. The CN-induced ﬂagments ﬂom all the possible singly reduced/cyanylated
isoforrns of ribonuclease-A.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

S-S S-S reduced * Fragment 1 F ragrnent 2 Fragment 3
No. Cys i - Cys j l-(i- 1) itz-i-(j- 1) itz-j-124
l Cy526—Cys40 1-25 itz-26-39 itz-40-124
2 Cys26—CysS 8 1-25 itz-26-57 itz-58-124
3 CysZ6—Cys65 l-25 itz-26-64 itz-65-124
4 Cy526—Cys7 2 1 -25 itz-26-71 itz-72-124
5 Cy526—Cys84 1-25 itz-26-83 itz-84-124
6 CysZ6—Cys95 1-25 itz-26-94 itz-95-124
7 CysZ6—Cys1 10 1-25 itz-26-109 itz-l 10-124
8 Cys40-Cy558 1-39 itz-40-57 itz-58-124
9 Cys40—Cys65 1-39 itz-40-64 itz-65-124
10 Cys40—Cys72 1-39 itz-40-7l itz-72-124
l 1 Cys40—Cys84 1-39 itz-40-83 itz-84-124
12 Cys40—Cys95 1-39 itz-40-94 itz-95-124
l 3 Cys40-Cysl 10 1-39 itz-40-109 itz-1 10-124
14 CysS 8—Cys65 1-57 itz-58—64 itz-65-124
15 Cys5 8—Cys72 1-57 itz-5 8-71 itz-72-124
l6 CysS8—Cys84 1-57 itz-58-83 itz-84-124
17 Cys58—Cys95 1-57 itz-58-94 itz-95-124
18 CysS 8—Cysl 10 1-57 itz-58-109 itz-1 10-124
19 Cys65—Cys72 1-64 itz-65-71 itz-72-124
20 Cys65—st84 1-64 itz-65-83 itz-84-124
21 Cys65—Cys95 1-64 itz-65-94 itz-95-124
22 Cys65—Cysl 10 1-64 itz-65-109 itz-1 10-124
23 Cys72—Cys84 l-71 itz-72-83 itz-84-124
24 Cys72—Cys95 1-71 itz-72-94 itz-95- l 24
25 Cys72—Cysl 10 1-71 itz-72-109 itz-l 09-124
26 Cys84—Cys95 1-83 itz-84-94 itz-95-124
27 Cy884—Cysl 10 1-83 itz-84-109 itz-95-124
28 Cys95—Cysl 10 1-94 itz-95-109 itz-l 10-124

 

* A singly reduced/cyanylated isoform is designated by a numeric representation of the
two disulﬁde cystines residues constituting the disulﬁde bond that was reduced.

164

 

Once the masses (m/z values) of the CN—induced fragments ﬂom all singly
reduced/cyanylated isoforrns are calculated, mass spectral data ﬂom analysis of the CN-
induced fragments ﬂom singly reduced/cyanylated isofonns can be compared against the
calculated values to identify the disulﬁde bonds in the original protein. For example, the
MALDI-MS spectrum of the CN-induced ﬂagments ﬂom a singly reduced/cyanylated
isoform of ribonuclease-A shows major peaks at m/z 2705.3, 4527.4, and 6548.5 (1).
These three peaks match with the calculated m/z values of singly protonated CN-induced
fragments, 1-25 (m/z 2704.9), itz-84-124 (m/z 4526.1), and itz-26-83 (m/z 6546.4),
respectively, within the mass measurement error of 0.1%. According to Table 5.3, the
disulﬁde bond Cy526-Cys84 was reduced to produce the singly reduced/cyanylated
isoform of ribonuclease-A.

Note that while seven different singly reduced/cyanylated isofonns (No. 1 to No.
7 in Table 5.3) would produce the CN-induced ﬂagrnent 1-25 and ﬁve singly
reduced/cyanylated isoforrns (No. 5, 11, 16, 20, and 23 in Table 5.3) would produce the
CN-induced ﬂagrnent itz-26-83, only one (No. 5 in Table 5.3) of these singly reduced
isoforrns would generate 1-25, itz-26-83 and itz-84-124, a pattern that matches the
experimental data. What if some of the expected CN-induced ﬂagments were not
detected because of chemical instability of the ﬂagrnents or signal suppressions in mass
spectrometry? Will we be able to identify the disulﬁde bond that was reduced to produce
the singly reduced isoform? The answer is ‘yes’ in some cases, ‘no’ in other cases. We
introduce a term “matching percentage” to determine when a singly reduced isoform
can be identiﬁed with the detection of an incomplete set of CN-induced cleavage

ﬂagments. Matching percentage (MP) is deﬁned as the number of detected CN-induced

165

 

cleavage ﬂagments divided by the total number of the expected such ﬂagrnents for a
certain partially reduced/cyanylated species of the original protein. When a complete set
of CN-induced ﬂagments was detected, MP is 100%; otherwise, MP is less than 100%.
A candidate isoform with a MP of 0 is no longer considered. The candidate isoform with
the highest MP is the correct one for the experimental data.

Case 1. 1-25 and itz-26-83 were detected. According to Table 5.3, the matching
percentages for all the singly reduced isofonns are 2/3 for No. 5; 1/3 for No. 1, 2, 3, 4, 6,
and 7; and 0/3 for all other singly reduced isoforrns. In this case, MP for No. 5 is the
highest. The correct isoform, No. 5, is identiﬁed.

Case 2. Itz-26-83 was detected. According to Table 5.3, the matching
percentages for all the singly reduced isofonns are 1/3 for No. 5; and 0/3 for all other
singly reduced isoforrns. In this case, NIP for No. 5 is the highest. The correct isoform,
No. 5, is identiﬁed.

Case 3. 1-25 was detected. According to Table 5.3, the matching percentages for
all the singly reduced isoforms are 1/3 for No. 1, 2, 3, 4, 5, 6, and 7; and 0/3 for all other

singly reduced isoforrns. In this case, the correct isoform, No. 5, is not identiﬁed.

If there are some ﬂee cysteines in the protein, they need to be pre-modiﬁed
(alkylated or cyanylated) and their positions need to be determined before the protein is
subjected to the partial reduction/cyany]ation/CN-induced cleavage/complete reduction
protocol. The data processing strategies can be easily adjusted for the calculated masses

of the species affected by the pre-modiﬁcations.

166

2. Consideration of Doubly Reduced Isoforms

As described in chapter 2, the possibility of using doubly reduced isoforrns, in
addition to singly reduced isoforrns, provides ﬂexibility in applying the chemistry of
partial reduction, and increases the possibility of successful identiﬁcation of the linkage
of all the disulﬁde bonds in the protein. Mass mapping a doubly reduced/cyanylated
isoform of the protein followed by CN-induced cleavage and complete reduction can
determine which four disulﬁde cysteine residues form two disulﬁde bonds (Figure 2.1 in
Chapter 2). This information can be used along with other information from the mapping
experiment of another doubly or a singly reduced/cyanylated isoform (Figure 2.1 and
Figure 2.2 in Chapter 2).

In order to identify which four disulﬁde cysteine residues form two disulﬁde
bonds (that have been reduced and cyanylated) in a particular doubly reduced/cyanylated
isoform, the mass spectral data should be compared against all sets of calculated masses
of the CN-induced ﬂagments expected ﬂom all possible doubly reduced/cyanylated
isoforrns. For a protein consisting of n disulﬁde bonds, there are [(2n-1)*(2n-3)*...*3*1]
possible disulﬁde structures and n*(n-1)/2 possible doubly reduced/cyanylated isoforrns
for each structure, but only 2n*(2n-1)*(2n-2)*(2n-3)/24 doubly reduced/cyanylated
isoforrns need be considered because of the degeneracy associated with completely
reducing the n-2 residual disulﬁde bonds in these species during chemical processing.
For example, only 70 (=(2*4)*(2*4-1)*(2*4-2)*(2*4-3)/24) possible doubly reduced/
cyanylated isoforrns for a protein having four disulﬁde bonds, such as bovine
ribonuclease-A, even though there are 105 (=(2*4-l)*(2*4-3)*. . .*3*1) possible disulﬁde

structures for the protein.

167

 

All sets of calculated masses of the CN-induced ﬂagments ﬂom all possible
doubly reduced/cyanylated isoforrns can be computed and tabulated in an analogous
manner to those ﬂom all possible singly reduced/cyanylated isoforrns, as listed in Table
5.3. We use a hypothetical protein to illustrate the computation process. The protein has
100 amino acid residues, among which 2n disulﬁde cysteines (forming n disulﬁde bonds)
are at positions C1, C2, ..., and C(2n).

Step 1. Compute all doubly reduced isoforms in which the four nascent ﬂee
cysteines (ﬂom reduction of two disulﬁde bonds) are C1, C(i), C(k), and C(m) (1 < j, k,
m < 2n). We can choose C(j) ﬂom C2, C3, ..., C(2n) ﬁrst; then choose C(k) and C(m)
ﬂom C(j+1), C(j+2), ..., and C(2n) for each C(j) using the strategy described for
computing singly reduced isofonns in the last section. There are (2n-1)*(2n-2)*(2n-
3)/(3*2*1) such doubly reduced isofonns.

Step 2. Compute all doubly reduced isoforrns in which the four nascent ﬂee
cysteines are C2, C0), C(k), and C(m) (2 < j, k, m < 2n). We can choose C(j) ﬂom C3,
C4, ..., C(2n) ﬁrst; then choose C(k) and C(m) ﬂom C(j+1), C(j+2), ..., and C(2n) for
each C(j) using the strategy described for computing singly reduced isoforms in the last

section. There are (2n-2)*(2n-3)*(2n-4)/(3*2*1) such doubly reduced isoforrns.

Step i. Compute all doubly reduced isoforrns in which the four nascent ﬂee

Cysteines are C(i), C(j), C(k), and C(m) (i < j, k, m < 2n). We can choose C(j) ﬂom C(i),

C(i+l), ..., C(2n) ﬁrst; then choose C(k) and C(m) ﬂom C(j+1), C(j+2), ..., and C(2n)

168

for each C(j) using the strategy described for computing singly reduced isofonns in the

last section. There are (2n-i)*(2n-i-1)*(2n-i-2)/(3*2*1) such singly reduced isoforrns.

Step 2n—3. Compute the doubly reduced isoform in which the four cysteines that
are involved in two disulﬁde bonds are C(2n-3), C(2n-2), C(2n-1), and C(2n). There is
only one such doubly reduced isoform.

Counting ﬂom Step 1 to Step 2n—3, there are a total of 2n*(2n-1)*(2n-2)*(2n-

4)/4*3*2*1 doubly reduced isofonns.

Once all the doubly reduced isoforms are computed, all the CN-induced cleavage
ﬂagments ﬂom such doubly reduced/cyanylated isoforrns can be easily computed. For
example, when a doubly reduced/cyanylated isoform contains four cyanylated cysteines
C(i), C(j), C(k), and C(m), its corresponding CN-induced ﬂagments are l-[C(i)-l], itz-

C(i)-[C(j )- 1 ], itz-C(i)-[C(k)-1], itz-C(k)-[C(m)- l ], itz-C(m)-100.

Once the masses (m/z values) of the CN-induced fragments ﬂom all doubly
reduced/cyanylated isoforms are calculated, mass spectral data ﬂom analysis of the CN-
induced ﬂagments ﬂom doubly reduced/cyanylated isoforrns can be compared against the
calculated values to identify the four cysteines that are involved in two of the disulﬁde
bonds in the original protein. We can use the concept of matching percentage in the same
way as we did for the singly reduced isoforrns. The candidate doubly reduced isoform

with the highest matching percentage is the correct one for the experimental data.

169

3. Consideration of Triply Reduced Isoforms
A very similar strategy can be used to compute all the triply reduced isofonns and
to search the correct candidate isoform for the experimental data. Therefore, its details

are not described here.

The algorithms discussed in the last three sections assume that we know how
many disulﬁde bonds were reduced and cyanylated in the partially reduced/cyanylated
species. Namely, we know whether the species is a singly reduced, doubly reduced or
triply reduced isoforrns. However, we may not know this in some cases. We can still use
the algorithms discussed above, but we need to compare the experimental data against the
calculated values for singly reduced, doubly reduced, and triply reduced isofonns. A
candidate singly reduced isoform list ranked by descending matching percentage, a
candidate doubly reduced isoform list ranked by descending matching percentage, and a
candidate triply reduced isoform list ranked by descending matching percentage will be
produced. The top candidate in each group (singly reduced, doubly reduced, or triply
reduced) needs to be considered by the analyst to determine the disulﬁde structure of the
original protein.

Only up to triply reduced isoforrns are considered because normally only singly,

doubly, and triply reduced isoforrns are used in the disulﬁde mass mapping strategy.

170

IV. Negative Signature Mass and an Improved Algorithm

1. Negative Signature Mass

As discussed above, it is desirable to deal with disulﬁde linkages rather than
disulﬁde structures. Partially reduced/cyanylated species of the original protein need to
be separated to allow application of the basic algorithm described in the last section. An
alternative algorithm based on negative signature mass (deﬁned later) that does not
require this condition is described as follows.

The presence of a negative signature mass in the experimental data insures that
a candidate disulﬁde bond cannot be present in the correct disulﬁde structure. For
example, a MALDI-MS peak at m/z 6546.4 indicates the presence of the fragment itz-26-
83 ﬂom ribonuclease-A (contains 8 disulﬁde cysteine residues at positions 26, 40, 58, 65,
72, 84, 95, and 110); and thus, this ﬂagrnent shows that cysteine 26 cannot be connected
to cysteine 40 for example; that is, Cys 26 and Cys 40 cannot form a disulﬁde bond
(linkage) in the protein. Let us see why this is true. If the disulﬁde bond Cys26-Cys40
had been in the original protein, it would have been reduced; and thus, itz-26-39 would
have been produced by the following cyanylation and CN-induced cleavage reactions
(Figure 5.3). Similarly, the presence of itz-26-83 allows us to reject the existence of
seven other disulﬁde linkages: Cys26-Cy858, Cys26-Cys65, Cy826-Cys72, Cys40-Cy584,

CysS8-Cy884, Cys65-Cys84, and Cys72-Cys84 (Figure 5.3).

171

 

Detection of m/z 6546.4

1

8H lSH SH 811
”Z 26 40 58 65 72

l

The following disulﬁde linkages could not have been
present in the original protein:

Cys26 +Cys40 + Cy526—-)(- Cy358 +

83

 

"’ Cy526-—X— Cy565 1’ Cy526 —X-Cys72
+ Cys40—X-Cy384 + Cy558 +Cy384

+ Cy565+ Cy384 + Cys72 —x—Cys84

For example, if
had been present in
Cys40 the original protein

 

Cysze

SCN SCN
| J 58 65 72 84 95 110 124

126 461811111

S S S S S
SH SH SH SH SH SH
1 r 1 I

l

— ' — + '
1 25 + It226 39 1tz40 58 65 72 84 95 110

124

Figure 5.3. Using a negative signature mass (m/z 6546.4) ﬂom ribonuclease-A to reject

the existence of 8 disulﬁde linkages.

172

2. An Algorithm Based on Negative Signature Masses

We use the list of negative signature masses as follows. Rather than comparing
the experimental data against list A, we scan the experimental data looking for the
presence of the negative signature masses. Then we can use the presence of some of the
negative signature masses to reject certain disulﬁde linkages in list B to obtain a
shortened disulﬁde linkages list C. A shortened disulﬁde structure list D can then be
produced ﬂom list C. If enough negative signature masses are found in the experimental
data, the list D will only contain the correct disulﬁde structure. Such an improved
algorithm can be illustrated by Figure 5.4.

All the negative signature masses (ﬂagments) can be computed as the following.
We use a hypothetical protein to illustrate this computation process. The protein has 100
amino acid residues, among which 2n disulﬁde cysteines (forming n disulﬁde bonds) are
at position C1, C2, ..., C(2n).

Step 1: Compute negative signature ﬂagments beginning ﬂom position 1. They
are l-(C2-l), l-(C3-1), ..., 1-(C(2n)-1) [“1-(C2-1)” designates the CN-induced
ﬂagments ﬂom residue 1 to residue C2-1.]. There are 2n-1 such ﬂagments.

Step 2: Compute negative signature ﬂagments beginning ﬂom position Cl. They
are itz-C1-(C3-l), itz-C1-(C4-1), ..., itz-C1-(C(2n)-1), itz-C 1-100 [“itz-Cl-(C3-1)”
designates the CN-induced ﬂagments ﬂom residue C1 to residue C3-l .]. There are 2n-l

such ﬂagrnents.

Step 2n: Compute negative signature ﬂagments beginning ﬂom position C(n-l).

There is only one such ﬂagrnent, itz-C(2n-l)-100.

173

Counting ﬂom Step 1 to Step n, there are a total of (2n-l)(n+1) such ﬂagments.

------------------ -| —--———--———-—————--—-I
I "IN“ 2 l I Input 1: l
I MS Spectra + m/z Measurement: I Sequence of a Protein Containing n S-S I
1 l Accuracy I I + I
I I l 1 Location of Free Cysteines l
I . I ——— —————————————— J
I Expenmental m/z Set _1 I

    
 

The Complete
Disulﬁde Linkage

/ /st 8

The Negative Signature Mass List

[Filtering Modulel

U

The Shortened Disulﬁde
Linkage List C

Figure 5.4. An algorithm based on negative signature masses.

174

Once negative signature fragments are computed, their mass (negative signature
masses) can be easily calculated, as the sequence of the original protein is known.

Next, we are going to use ribonuclease-A as an example to see how this algorithm
can be applied. Suppose the chemical process using the partial reduction/cyanylation/
CN-induced cleavage/complete reduction protocol for ribonuclease-A yields only three
MALDI-MS peaks, m/z 6548.5, 5766.8, and 6061.2, which correspond to ﬂagrnents itz-
26-83, itz-58-109, and itz-40-94 with calculated m/z value 6546.4, 5766.4, and 6062.7,
respectively. Actually, instead of three MALDI-MS peaks, the experimental data consist
of eleven identiﬁed peaks from four singly reduced/cyanylated isoforrns (1). Only a
subset of the experimental data is selected in order to illustrate the effectiveness of the
improved algorithm. As ribonuclease-A has 8 disulﬁde cysteine residues, which are
involved in 4 disulﬁde bonds, there are 28 possible disulﬁde linkages in its list B. The
three experimental peaks can be used as negative signature masses to shorten list B into
list C. Please see Table 5.4 for details.

After the ﬁltering process using the negative masses, there are only seven possible

disulﬁde linkages leﬁ in list C (Table 5.5).

175

Table 5.4. Filtering the List B for ribonuclease-A using three negative masses

S-S S-S in the List B Identiﬁed Negative Masses
No.

6548.5 5766.8 6061.2
1tz-26-83 1tz-58-109 1tz-40-94

X

X
X
X

l
2
3
4
5
6
7
8
9

110
5

110

110

 

x means that a certain disulﬁde linkage cannot exist in the analyte because of the
presence of a particular negative mass .

176

Table 5.5. The shortened disulﬁde linkage List C for ribonuclease-A after ﬁltering using
three negative masses.

 

 

 

 

 

 

Disulﬁde 1st Cys in S-S with the Same First Disulﬁde Cys
Linkage Group i the S-S S-S No. i.l S-S No. i.2 S-S No. i.3
1 Cys58 CysS 8-Cys1 10
2 Cys65 Cys65-Cys72
3 Cys40 Cys40-Cys95 Cys40-Cys110
4 Cy526 Cy326-Cy584 Cy526-Cys95 CysZ6-Cys110

 

 

 

 

 

 

 

The shortened disulﬁde structure list D can be computed as follows.

Step 1. The disulﬁde linkages in list C are classiﬁed and sorted as in Table 5.5.
The disulﬁde linkages are ﬁrst divided into groups according to the ﬁrst disulﬁde
cysteine in each disulﬁde linkage (bond) (The ﬁrst disulﬁde cysteine in a disulﬁde
linkage is the one with the lower numeric value for the position of the residue in the
protein sequence.) The groups are then sorted according to the number of disulﬁde
linkages in each group in an ascending order.

Step 2. Compute disulﬁde structure list D from the disulﬁde linkages in list C. In
the case of ribonuclease-A, a disulﬁde structure consists of four disulﬁde bonds
(linkages). Thus, four disulﬁde linkages need to be chosen from list C to form a disulﬁde
structure. Note that we can only select one disulﬁde linkage ﬂom each of the four groups
for forming a disulﬁde structure. As shown in Table 5.6, the positions of disulﬁde
cysteines in the original protein are ﬁlled with numbers according to the constituent
disulﬁde linkages. For example, if the ﬁrst disulﬁde linkage ﬁlled for a disulﬁde

structure is Cy58-Cysl 10, position 58 and 110 will be both ﬁlled with number 1. For one

177

disulﬁde structure, if all positions of disulﬁde cysteines are ﬁlled and each of them is
only ﬁlled once, then that structure is a valid structure. Otherwise, if any position is
ﬁlled twice with two different numbers, the corresponding structure is not valid. The
process of forming disulﬁde structures continues until all possible combinations are
enumerated. Such a typical process is illustrated in Table 5.6.

As shown in Table 5.6, there is only one disulﬁde structure, (CysS8-CysllO,
Cys65-Cys72, Cys40-Cys95, Cys26-Cys84), in the list D because only one disulﬁde
structure can be formed by the allowable disulﬁde linkages in list C (Table 5.6).

One of the advantages of this algorithm is that the number of candidate disulﬁde
linkages is quadratic in the number of disulﬁde cysteines in the original protein, so the
matching phase should be more efﬁcient than that of the naive algorithm where the
number of disulﬁde structures is exponential in the number of disulﬁde cysteines. The
other advantage of this algorithm is that, in some cases, even if only an incomplete data
set is obtained, the correct disulﬁde structure can be deduced. However, there is a
condition for using this algorithm. That is, cyanylation needs to be complete for the
nascent ﬂee cysteines aﬁer partial reduction of the original protein. Otherwise, incorrect
deduction may result ﬂom an incorrectly identiﬁed negative signature mass. Let us
examine what would happen if the condition of complete cyanylation cannot be met for
the previous example of using itz-26-83 as a negative signature ﬂagrnent for
ribonuclease-A. Without the condition of complete cyanylation, any of the eight
previously rejected disulﬁde linkages may be wrongly ruled out. For example, Cy526-
Cys40 could be valid because itz-26-83 could result ﬂom the CN-induced cleavage of a

doubly reduced and incompletely cyanylated isoform of ribonuclease-A, in which two

178

 

disulﬁde bonds Cy526-Cys40 and Cys84-Cys95 were reduced, but only Cys 26, Cy584,

and Cys95 were cyanylated.

Table 5.6. Forming the disulﬁde structure list D ﬂom the shortened disulﬁde list C (Table
5.5) for ribonuclease-A.

S-S No. Disulﬁde ° Positions

C26 C40 C58 C65 C72 C84 C95 Cl 10
1.1: 1 1
2.1: 2 2
3.1:
4.1:

 

179

V. Dealing with Imperfect Data

Assuming we have perfect experimental data, there will be a perfect matching
between the list of experimental peaks for a particular partially reduced/cyanylated
species and the list of calculated masses of ﬂagments for that species. However, we
cannot assume that the list of experimental peaks will be perfect. It may include extra
mass spectral peaks and/or some expected mass spectral peaks are missing. This leads to
a situation where we will not get a perfect matching between the experimental peaks and
the correct set of the calculated masses.

One possible way to deal with the imperfect data is to rank candidate partially
reduced/cyanylated species by their matching percentage using the basic algorithm
(Section III). An incomplete experimental data set could be obtained due to signal
suppression and interference ﬂom matrix peaks (1, 2). As demonstrated in section III, in
some cases, ranking the candidate partially reduced species may allow us to identify the
correct species and thus, the correct disulﬁde bond(s).

Utilizing negative signature masses (Section IV) is another way to deal with
incomplete experimental data sets. The algorithm based on negative signature masses
does not even require partially reduced/cyanylated species of the original protein to be
separated prior to CN-induced cleavage and complete reduction of the residual disulﬁde
bonds.

Other sources of imperfect data are products of side reactions, such as incomplete
cyanylation, B-elimination, and incomplete CN-induced cleavage (2). CN-induced
fragments due to B-elimination and incomplete CN-induced cleavage can normally be

identiﬁed by mass mapping; and sometimes provide complementary information for

180

determining disulﬁde connectivities (l, 2). Therefore, theses side reactions need to be
considered in computing the expected values. On the other hand, incomplete cyanylation
may prevent us ﬂom using negative signature masses (Section IV). If incomplete
cyanylation happened, the algorithm based on negative masses may reject all candidate
disulﬁde structures of the original protein; the results from the basic algorithm based
matching percentages may contradict each other for identiﬁed partially
reduced/cyanylated species. These can then be used as an indication that incomplete
cyanylation occurred. In such cases, additional chemical or mass spectrometric analyses
may be needed to determine where incomplete cyanylation occurred if they cannot be

deduced using the described algorithms.

VI. Implementation and Testing of the Algorithms
The algorithms are coded in C++. The program code and its test results are listed

as follows.

1. Implementation of the Basic Algorithm

A. C++ Program Files:
SSbyCyanl .cpp, the source ﬁle for the main program’s source code
InputOutput.cpp, the source ﬁle for the Input and Output library functions to input

the protein sequence, output results, etc.

InputOutput.h, the header ﬁle for the Input and Output library functions

181

 

DisulﬁdeProtein_extern.h, the header ﬁle for declaration of all the extem variables

All these C++ ﬁles are available electronically.

B. Preliminary Testing Using Data from a Model Protein: Ribonuclease-A

Input from Keyboard:

Standard Error Limit: 2

Text Files for Input:
ExpMassl.txt, the text ﬁle that stores the experimental m/z values of the CN-
induced cleavage ﬂagments (MH+) from a partially

reduced/cyanylated species of ribonuclease-A
ProteinSequence.txt, the text ﬁle that stores the name of the protein, its sequence,

etc.

Text Files for Output:
Result1.txt, the text ﬁle that stores the results: ranked lists of candidate singly,

doubly, and triply reduced species of ribonuclease-A

182

C. The Content of Input and Output Text Files

The Content of ExpMassl.txt:

2705.3 4527.4 6548.5

The Content of ProteinSequence.txt:

A template for input ﬁle protein sequence, by J ianfeng (Frank) Qi, 6/27/2001.

 

NOTE: ALL ENTREES BEGIN WITH "*" AND TERMINATE WITH "#".

 

 

1) Name of the Protein

 

*Ribonuclease A, Bovine#

 

2) Sequence of the Protein (Amino Acids, Capital Single Letter Representation
except lower case "c" for ﬂee cysteines and upper case "C" for disulﬁde
cysteines that are involved in disulﬁde bonds)

NOTE: Space and return can be used within the sequence.

 

183

*KETAAAKFER QHMDSSTSAA SSSNYCNQMM KSRNLTKDRC KPVNTFVHES
LADVQAVCSQ KNVACKNGQT NCYQSYSTMS ITDCRETGSS KYPNCAYKTT

QANKHIIVAC EGNPYVPVHF DASV#

 

3) User-deﬁned Amino Acids
(Format: "AA_user monoisotopic_residue_mass average_residue_mass)
NOTE: Up to Num__AA__User (6 in this case) user-deﬁned amino acids

(B, J, O, U, X, and Z) can be used.)

 

*BOO

J00

OOO

U00

X00

ZOO#

 

4) Status of Cyanylation of Free Cysteines (1 (yes) or 0 (no), default is 1)

 

*l#

184

 

5) Status of Non-cyanylation Modiﬁcation of Free Cysteines (1 (yes) or 0 (no),

default is 0)

 

*0#

 

6) Added Mass to Free Cysteines by the Non-Cyanylation Modiﬁcation

(Format: "monoisotopic_mass average__mass", default is "0 0")

 

*0 0#

 

The Content of Resultl.txt:

The Protein Name is: Ribonuclease A, Bovine. 124 AAs.

The AA sequence is:
KETAAAKFERQHMDSSTSAASSSNYCNQMMKSRNLTKDRCKPVNTFVHESLAD
VQAVCSQKNVACKNGQTNCYQSYSTMSITDCRETGSSKYPNCAYKTTQANKHII

VACEGNPYVPVHFDASV.

185

 

It has 4 Disulﬁdes.

It has 8 CysS.

The positions of them are:

26 40 58 65 72 84

The experimental data are:

2705.3 4527.4 6548.5

The standard error limit for searching is: 2 Da

type 0: monoisotopic with -OH as C-terminus,

95 110

type 1: monoisotopic with -NH2 as C-terminus,

type 2: average with -OH as C-terminus,

type 3: average with -NH2 as C-terminus

186

Search Result from Singly Reduced Isoforms:
The following is the name of each of the items in each of the data entries and their

appearing order in the results.

NO: Cysi-Cysj has (N) matches: Mass] FragID; MatchedMassTypes Mass_N

FragID; MatchedMassTypes

l: 26 — 84 has 3 matches: 2705.3 1 to 25 ; 0 1 2 3 6548.5 26 to 83 ; 2 3 4527.4
84 to 124; 2 3

2: 26 - 40 has 1 matches: 2705.3 1 to 25 ; 0 l 2 3

3: 26 - 58 has 1 matches: 2705.3 1 to 25 ; 0 l 2 3

4: 26 - 65 has 1 matches: 2705.3 1 to 25 ; 0 1 2 3

5: 26 - 72 has 1 matches: 2705.3 1 to 25 ; 0 1 2 3

6: 26 - 95 has 1 matches: 2705.3 1 to 25 ; 0 1 2 3

7: 26 - 110 has 1 matches: 2705.3 1 to 25 ; 0 1 2 3

8: 40 - 84 has 1 matches: 4527.4 84 to 124; 2 3

9: 58 - 84 has 1 matches: 4527.4 84 to 124; 2 3

10: 65 - 84 has 1 matches: 4527.4 84 to 124; 2 3

Search Result from Doubly Reduced Isoforms:

1: 26 - 40, 58 - 84 has 2 matches: 2705.3 1 to 25 ; 0 l 2 3 4527.4 84 to 124; 2 3

2: 26 - 40, 65 - 84 has 2 matches: 2705.3 1 to 25 ; 0 1 2 3 4527.4 84 to 124; 2 3

3: 26 - 40, 72 - 84 has 2 matches: 2705.3 1 to 25 ; 0 1 2 3 4527.4 84 to 124; 2 3

187

4: 26 - 58, 65 - 84 has 2 matches: 2705.3 1 to 25 ; O 1 2 3 4527.4 84 to 124; 2 3
5: 26 - 58, 72 - 84 has 2 matches: 2705.3 1 to 25 ; 0 1 2 3 4527.4 84 to 124; 2 3
6: 26 - 65, 72 - 84 has 2 matches: 2705.3 1 to 25 ; 0 1 2 3 4527.4 84 to 124; 2 3
7: 26 - 84, 95 - 110 has 2 matches: 2705.3 1 to 25 ; 0 l 2 3 6548.5 26 to 83 ; 2 3
8: 26 - 40, 58 - 65 has 1 matches: 2705.3 1 to 25 ; 0 1 2 3
9: 26 - 40, 58 - 72 has 1 matches: 2705.3 1 to 25 ; O 1 2 3

10:26 - 40, 58 - 95 has 1 matches: 2705.3 1 to 25 ;0 1 2 3

Search Result from Triply Reduced Isoforms:

1:26 - 40 , 58 - 65 , 72 - 84 has 2 matches: 2705.3 1 to 25 ; O 1 2 3 4527.4

84 to 124; 2 3

2: 26-40,58-65,72-95 haslmatches:2705.3 1t025;01 2 3
3:26-40,58-65,72-110 haslmatches:2705.3 lt025;012 3
4: 26-40,58-65,84-95 haslmatches:2705.3 1t025;0 l 2 3
5:26-40,58-65,84-110 haslmatches:2705.3 1t025;012 3
6: 26-40,58-65,95-110 haslmatches:2705.3 1t025;012 3
7: 26-40,58-72,84-95 haslmatches:2705.3 lt025;01 2 3
8: 26-40,58-72,84-110 haslmatches:2705.3 1t025;012 3
9: 26-40,58-72,95-110 haslmatches:2705.3 lto25;012 3

10: 26-40,58-84,95-110 haslmatches:2705.3 1t025;0 1 2 3

188

 

2. Implementation of the Algorithm Based on Negative Signature Masses

A. C++ Program Files:
SSbyCyan2.cpp, the source ﬁle for the main program’s source code
InputOutput.cpp, the source ﬁle for the Input and Output library ﬁmctions to input

the protein sequence, output results, etc.

InputOutput.h, the header ﬁle for the Input and Output library functions

Si gnaturecpp, the source ﬁle for the signature library functions that deals with
computations related to negative signature mass

Si gnature.h, the header ﬁle for the signature library ﬁrnctions that deals with

computations related to negative signature mass

All these C++ ﬁles are available electronically.

B. Preliminary Testing using Data from a Model Protein: ribonuclease-A

Text Files for Input:
ExpMass2.txt, the text ﬁle that stores the experimental m/z values of the CN-

induced cleavage ﬂagments (MH+) ﬂom one or more partially

reduced/cyanylated species of ribonuclease-A

ProteinSequence.txt, the text ﬁle that stores the name of the protein, its sequence,

SIC .

189

Text Files for Output:
Result2.txt, the text ﬁle that stores the results: found negative signature masses, a
list of bad disulﬁde bonds that were rejected by the found negative
signature masses, a list of good disulﬁde bonds that are still

possible, the possible good disulﬁde structures, etc.

C. The Content of Input and Output Text Files

The Content of ExpMass2.txt:

A template for input ﬁle ExpMass2.txt for the algorithm based on negative

signature masses, by Jianfeng (Frank) Qi, 9/8/2001.

 

NOTE: THE MASS DATA BEGIN WITH "*".

 

190

 

Exp. Mass Mass Deviation

(MH+) (/Da)

 

6548.5 2

5766.8 2

6061.2 2

The Content of ProteinSequence.txt:

A template for input ﬁle protein sequence, by J ianfeng (Frank) Qi, 6/27/2001.

 

NOTE: ALL ENTREES BEGIN WITH "*" AND TERMINATE WIIH "#".

 

 

1) Name of the Protein

 

*Ribonuclease A, Bovine#

191

 

2) Sequence of the Protein (Amino Acids, Capital Single Letter Representation
except lower case "c" for ﬂee cysteines and upper case "C" for disulﬁde
cysteines that are involved in disulﬁde bonds)

NOTE: Space and return can be used within the sequence.

 

*KETAAAKFER QHMDSSTSAA SSSNYCNQMM KSRNLTKDRC KPVNTFVHES
LADVQAVCSQ KNVACKNGQT NCYQSYSTMS IT DCRETGSS KYPNCAYKTT

QANKHIIVAC EGNPYVPVHF DASV#

 

3) User-deﬁned Amino Acids
(Format: "AA_user monoisotopic_residue_mass average_residue_mass)
NOTE: Up to Num_AA_User (6 in this case) user-deﬁned amino acids

(B, J, O, U, X, and Z) can be used.)

 

*BOO

J00

000

U00

X00

ZOO#

192

 

4) Status of Cyanylation of Free Cysteines (1 (yes) or 0 (no), default is 1)

 

*1#

 

5) Status of Non-cyanylation Modiﬁcation of Free Cysteines (1 (yes) or 0 (no),

default is 0)

 

*O#

 

6) Added Mass to Free Cysteines by the Non-Cyanylation Modiﬁcation

(Format: "monoisotopic_mass average_mass", default is "0 0")

 

*0 O#

193

 

The Content of Result2.txt:

1) Name of the Protein

Ribonuclease A, Bovine

2) Sequence of the Protein
KETAAAKFERQHMDSSTSAASSSNYCNQMMKSRNLTKDRCKPVNTFVHESLAD
VQAVCSQKNVACKNGQTNCYQSYSTMSITDCRETGSSKYPNCAYKTTQANKHII

VACEGNPYVPVHFDASV

3) The Positions of Disulﬁde Cysteines in the Protein: Ribonuclease A, Bovine

2640 58 65 72 8495110

4) A list of bad disulﬁde linkages that can be rejected by each of the matched

negative signature masses for the protein: Ribonuclease A, Bovine

type 1: monoisotopic with -OH as C-terminus,
type 2: monoisotopic with -NH2 as C-terrninus,

type 3: average with -OH as C-terminus,

194

type 4: average with -NH2 as C-terminus

Frag_Type = 0: N-terrninal ﬂagrnent, residue 1 is not Cys, l-(Ci-l)
F rag_Type = l: itz-fragrnent between two cysteines, itz-Ci-(Cj-l)

Frag_Type = 2: C-terrninal ﬂagrnent, residue Num_AA is not Cys, itz-Cj-

 

Num_AA
Frag_Type = 3: N-terminal ﬂagrnent, residue 1 is CysSH, 1-(Ci-1)
Frag_Type = 4: N-tenninal ﬂagrnent, residue 1 is CysSCN, 1-(Ci-1)

F rag_Type = 5: C-terminal ﬂagrnent, residue Num_AA is CysH, itz-Cj-Num_AA

The following is the name of each of the items in each of the data entries and their

appearing order in the results.

ExpMass

MatchedCalMass; MatchedMassType; MatchedSigFragID (LeftAA,

RightAA, Frag_Type);

BadSS_linkages (1 st_CysS, 2nd_CysS)

195

6548.5

6547.47; 3; 26 83 1

2640 26 58 26 65 26 72 40 84 58 84 65 84 72 84

5766.8

5767.49; 3;

5766.49; 4; 58 109 1

5865 5872 5884 5895 65110 72110 84110 95110

6061.2

6059.72; 1;

6062.77; 4; 4O 94 1

40 58 4O 65 40 72 40 84 58 95 65 95 72 95 84 95

5) The good disulﬁde linkage list for the protein: Ribonuclease A, Bovine

26 84

26 95

196

 

26110

40 95

40110

58110

65 72

6) The good disulﬁde structures for the protein: Ribonuclease A, Bovine

CysZ6-Cys84 Cys40-Cys95 Cys58-Cys110 Cys65-Cys72

VII. Conclusions

Computer-aided data processing for disulﬁde mass mapping based on partial

reduction and cyanylation/CN-induced cleavage is possible. The concept of using

negative signature masses facilitates the disulﬁde assignment process and, in some cases,

allows identiﬁcation of the disulﬁde structure even when an imperfect data set, resulting

ﬂom side reactions and/or signal suppression, is obtained.

197

VIII. References

1. Wu, J. & Watson, J. T. A novel methodology for assignment of disulﬁde bond
pairings in proteins. Protein Sci. 6, 391-398 (1997).

2. Qi, J. & Watson, J. T. Determination of the Disulﬁde Structure of Sillucin, a

Highly Knotted, Cysteine-Rich Peptide, by Cyanylation/Cleavage Mass Mapping.
Biochemistry 40, 4531-4538 (2001).

198