 

 

   

 

 

 

 

 

 

.nnJ-n-nr

 

 

 

1.‘

 

 

nu...

 

 

 

.0, .7 “nu"...«yw .v

 

 

 

 

 

 

 

 

 

~' '1.

"H.015

llllllllilHHIIHIIHIHIIHIHJMHlllllllHHlllllllllllll

31293 01701 9187

This is to certify that the

dissertation entitled

Capture anal 1 dreutf camp.4 of tie Intermeol ates
0f Refdolmq anal ReduCtﬂ/e unfoldaaj of Proteins 63
Cyomytati 0M Metkodalpjﬁ and MMS3 SPECtYOMefY‘ﬁ

presented by

Yliaj Ya‘tj

has been accepted towards fulﬁllment
of the requirements for

Doctor of F" (”ﬁligree 1n ($91169ng

 

 

 

 

Date Tune I6 ; 1198

 

LIBRARY

Michigan State
Unlveralty

 

 

 

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

 

    

1
JAN 13 2033

 

 

 

 

 

 

 

 

 

 

 

 

CAPTURE AND IDENTIFICATION OF THE INTERNIEDIATES OF
REFOLDING AND REDUCTIVE UNFOLDING OF PROTEINS BY
CYANYLATION METHODOLOGY AND MAss SPECTROMETRY
By

Ying Yang

A DISSERTATION
Submitted to
Michigan State University
In partial fulﬁllment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
Department of Chemistry

1998

ABSTRACT
CAPTURE AND IDENTIFICATION OF THE INTERMEDIATES OF
REFOLDING AND REDUCTIVE UNFOLDING OF PROTEINS BY
CYANYLATION METHODOLOGY AND MASS SPECTROMETRY
By

Ying Yang

Cyanylation chemistry of cysteine residues combined with mass spectrometry
provides promising methodology for providing structural information of cysteine residues
in the proteins. The sulﬂiydryl groups of proteins can be speciﬁcally cyanylated by a
cyanylation reagent, the N-terminal side of cyanylated cysteine residues can be cleaved in
alkaline condition, and the masses of peptide fragments (one N-terminal side peptide and
a series of iminothiazolidinc-blocked peptides) can be analyzed by mass spectrometry.
Matching the masses of the cleaved peptides can be used to recognize the structural
information of cysteine residues in the proteins.

Two types of mass spectrometry have been used to analyze proteins. One is matrix-
assisted laser desorption ionization time of ﬂight mass spectrometry (MALDI-TOF-MS);
the another is electrospray ionization mass spectrometry (ESI-MS). We found that
MALDl-MS showed frequent, and sometimes complete, suppression of a signal from
iminothiazolidine-blocked peptides; because such a phenomenon can compromise our
analytical strategy, we have investigated the use of ESI for these analyses. The results

indicate that the responses of our blocked peptides to E81 and MALDI are frequently

complementary.

Conventional methodology for determining the pairing in the disulﬁde bonding
structure of proteins relies on the use of appropriate proteases to cleave the peptide
backbone in between the cysteine residues. The possibility of a proteolytic cleavage site
becomes less likely when proteins contain closed or adjacent cysteins. We have
investigated the applicability of a novel methodology involving partial reduction of
disulﬁde bonds, cyanylation of sulﬂiydryl groups, chemical cleavage of partially
reduced/cyanylated protein, and mass mapping of cleavage products to the structural
analysis of proteins containing adjacent cysteines in two model protein, long R3 insulin-
like growth factor-I (LR3IGF-I) and insulin-like growth factor-I (IGF-I). These proteins
contain three disulﬁde bonds, two of which involve adjacent cysteines. The disulﬁde
structures of the proteins were unambiguously determined by the methodology.

A new methodology based on cyanylation chemistry of cysteine residue and mass
spectrometry has been used to capture and identify folding intermediates of long insulin-
like growth factor I (LR3IGF-I) and insulin-like growth factor I (IGF-I). The refolding
was quenched at different time points by adjusting the pH to 2.0-3.0 with a 1N HCl
solution of 1-cyano-4—dirnethylamino—pyridinium (CDAP) which trapped intermediates
via cyanylation of free sulfhydryl groups. The disulﬁde structm'e of the intermediates was
determined by partial reduction/cyanylation/chemical cleavage/mass mapping. The
folding intermediates of LR3IGF-I and IGF-I were successfully trapped and
characterized. The time-dependant distribution and disulﬁde structure of the folding

intermediates of LR3IGF-I and IGF-I were compared based on their structural features.

ACKNOWLEDGMENTS

I wish to thank my mentor, Dr. J. Throck Watson, for his assistance and advice. I
especially would like to thank my daughter, J. Jinjing Yang, my husband, J ie Yang, for

their support and understanding during this project.

iv

 

TABLE OF CONTENTS

LIST OF TABLES ................................................................................. vii
LIST OF FIGURES ................................................................................ viii
CHAPTER 1.

INTRODUCTION .................................................................................... l
I. Introduction .................................................................................... 1
II. Oxidative Folding and Trapping of Intermediates ........................................ 3

A. Disulﬁde Bonds 1n Protein Folding .................................................... 5
B. Trapping of the Folding Intermediates ................................................ 12
III. Recognizing the Location of Cysteines and Cystines ................................... 18
A. Classical Approach ...................................................................... 18
B. Afﬁnity Chromatography ............................................................... 22
C. Cleavage at Cysteine residue .......................................................... 23
IV. Assignment of Disulﬁde Bonds ............................................................ 24
A. Conventional Fragmentation Strategy ................................................ 25
B. Fragmentation of Proteins .............................................................. 25
C. Puriﬁcation of Cystinyl Peptides ...................................................... 27
D. Identiﬁcation of Cysteine-containing Peptides ...................................... 27
B. 8-8 Assignment of Adjacent Cysteines .............................................. 30
V. Matrix-Assisted Laser Desorption/Ionization Mass Spectrometry (MALDI-MS)
and Electrospray Mass Spectrometry (ESI-MS) ........................................ 32
A. Matrix-Assisted Laser Desorption/Ionization Mass Spectrometry
(MALDI-MS) ........................................................................... 33
B. Electrospray Ionization Mass Spectrometry (ESI-MS) ........................... 39
C. The Comparison of MALDI and E81 ................................................ 45
VI. References .................................................................................... 47
CHAPTER 2

A COMPARISON OF MALDI AND ESI FOR ANALYSIS OF
IMINO'I'HIAZOLIDINE-BLOCKED PEPTIDES IN DISULFIDE

MAPPING ........................................................................................... 58
I. Introduction .................................................................................. 58
II. Analytical Methodology ..................................................................... 61

A. Cyanylation and Chemical Cleavage Reaction ...................................... 61
B. Localization of Cysteines and Cystines by ESI-MS ................................ 63
III. Experimental Section ........................................................................ 64
IV. Results and Discussion ..................................................................... 68
A. Recognizing the Location of Cysteine and Cystines of B-Lactoglobulin A. . ...68
B. Recognizing the Location of Cysteines and Cystines in Ovalbumin ............. 76
C. Evaluation of CDAP and NTCB ...................................................... 85
V. Conclusions ................................................................................... 88

VI. References .................................................................................... 90

Chapter 3
DISULFIDE MASS MAPPING IN PROTEINS CONTAINING ADJACENT
CYSTEINES WITH CYANYLATION/CLEAVAGE

METHODOLOGY ................................................................................. 93
I. Introduction ................................................................................... 93
II. Methodology for Assignment of Disulﬁde Linkages ................................... 95
III. Experimental Section ........................................................................ 97
IV. Results and Discussion .................................................................... 100

A. Disulﬁde Mapping of IGF-I .......................................................... 100
B. Disulﬁde Mapping of LR3IGF-I ..................................................... 112
V. Conclusions ................................................................................. 121
VI. Future Work ................................................................................. 121
VII. References ................................................................................. 125

Chapter 4

TRAPPING AND IDENTIFICATION OF INTERMEDIATES DURING THE

REFOLDING OF LR’IGF-I AND IGF-I ...................................................... 128
I. Introduction ....................... 128
H. The strategy for Trapping Folding Intermediates ...................................... 133
III. The experimental Section ................................................................. 135
IV. Results and Discussion .................................................................... 140

A. Refolding and Reductive Unfolding of LR3IGF-I ................................. 140
B. Refolding and Reductive Unfolding of lGF-I ...................................... 157
V. Isolation and Refolding of the Folding Intermediates of LR3IGF-I ................. 176
VI. Comparison of the Folding Process of LR3IGF-I and IGF-1 ......................... 178
VH. Conclusions ................................................................................ 178
VIII. References ................................................................................. 186

vi

Table 1.1

Table 2.1

Table 2.2

Table 2.3

Table 3.1

Table 3.2

Table 4.1

Table 4.2

Table 4.3

Table 4.4

LIST OF TABLES

Comparison of ESI-MS and MALDI-MS .......................................... 46

Expected and observed masses of cleavage product of B-lactoglobulin A

............................................................................................ 69
Expected and observed masses of cleavage product of ovalbumin ............. 77
Comparison of CDAP and NTCB ................................................... 89

Calculated and observed m/z values for possible fragments resulting ﬁom the
cleavage reaction of IGF-I chains at sites of designated cysteine pairs ...... 109
Calculated and observed m/z values for possible fragments resulting from the

cleavage reaction of LR3IGF-I chains at sites of designated cysteine pairs

.......................................................................................... 115
The expected and observed masses of cleavage products of one-disulﬁde
intermediate of LR3IGF-I ............................................................ 148
The expected and observed masses of cleavage products of the
puriﬁed/cyanylated two-disulﬁde intermediate of LR’IGF-I .................. 151

The expected and observed masses of cleavage products of singly
reduced/cyanylated isomers of the three-disulﬁde intermediate of LR’IGF-I
.......................................................................................... 156

The expected and observed masses of the fragments related to the disulﬁde

assignment of the intermediates of IGF-I ......................................... 165

vii

Figure 1.1
Figure 1.2
Figure 1.3
Figure 2.1

Figure 2.2

Figure 2.3

Figure 2.4

Figure 2.5

Figure 2.6

Figure 2.7

Figure 2.8

Figure 2.9

LIST OF FIGURES

Folding pathway of BPTI .......................................................... 13
MALDI-TOF-MS instrument ...................................................... 37
ESI-MS instrument and droplet production in ESI interface .................. 42
Mechanisms of cyanylation and cleavage reaction ............................. 62
Structure of B-lactoglobulin A .................................................... 68
MALDI spectra of cleavage products of B-lactoglobulin A following

reduction/cyanylation/clcavage (panel A) or cyanylation/cleavage/reduction

(panel B) ............................................................................... 71

HPLC chromatogram of cleavage products following procedure
TCEP/CDAP ......................................................................... 72
E81 spectra of cleavage products from procedure TCEP/CDAP of B-
lactoglobulin A ...................................................................... 73
HPLC chromatogram of cleavage products following procedure
CDAP/TCEP ......................................................................... 74
E81 spectra of cleavage products from procedure CDAP/TCEP of B-
lactoglobulin A ...................................................................... 75
Structure of ovalbumin ............................................................. 76
MALDI spectra of cleavage products of ovalbumin following
reduction/cyanylation/cleavage (panel A) or cyanylation/cleavage/reduction

(panel B) .............................................................................. 79

viii

 

Figure 2.10

Figure 2.11

Figure 2.11
Figure 2.12

Figure 2.13

Figure 2.14

Figure 2.15

Figure 3.1

Figure 3.2

Figure 3.3

HPLC chromatogram of mixtures of cleavage product following procedure
TCEP/NTCB of ovalbumin ........................................................ 80

E81 spectra of cleavage products following procedure TCEP/NT CB of

ovalbumin ............................................................................ 81
Continue .............................................................................. 82

HPLC chromatogram of mixture of cleavage products following the
procedure NTCB/TCEP of ovalbumin ............................................ 83
E81 spectra of cleavage products following the procedure NTCB/TCEP of
ovalbumin ............................................................................ 84
HPLC chromatogram of mixture of cleavage product from procedure
CDAP/TCEP of ovalbumin ........................................................ 86
E81 spectra of cleavage product from procedure CDAP/TCEP of ovalbumin
.......................................................................................... 87
Amino acid sequence and disulﬁde structure of IGF-I ....................... 101
Overview of the methodology used to assign the disulﬁde linkages in IGF-I
......................................................................................... 104
HPLC separation of denatured IGF-I and its partially reduced/cyanylated
isomers. Separation was carried out on a Vydac C18 column at a flow rate
of 1.0ml/min with a linear gradient 30-50% B in 45 minutes, where A =
0.1% TFA in water and B = 0.1% TFA in CH3CN. Peaks land 2 represent
singly reduced/cyanylated species. Peak 3 represent doubly
reduced/cyanylated species, as determined by MALDI-TOF

analysis .............................................................................. 106

 

Figure 3.4

Figure 3.5

Figure 3.6

Figure 3.7

Figllre 3.8

The MALDI mass spectra of peptide mixtures resulting from the cleavage
of the two singly reduced/cyanylated IGF-I isomers, corresponding to the
HPLC peaks 2 and 3 in Figure 3.2, respectively. The symbols “itz” and *
represent the iminothiazolidine derivatives and protonated B-elimination
products, respectively ............................................................ 110
Amino acid sequence and disulﬁde structure of LR3IGF-I. The bold letters
at N-terminal sites indicate the 13 amino acid extension. The bold R
represents the replacement of E3 of IGF-I ..................................... 113
HPLC separation of denatured LR’IGF-I and its partially
reduced/cyanylated isomers. Separation was carried out on a Vydac C18
column at a ﬂow rate of 1 ml/min with a linear gradient 30-50% B in 45
minutes, where A = 0.1% TFA in water and B = 0.1% TFA in CH3CN.
Peaks 1 and 2 represent singly reduced/cyanylated species, as determined
by MALDI-TOF analysis ......................................................... 1 14
The MALDI mass spectra of peptide mixtures resulting from the cleavage
of the two singly reduced/cyanylated LR3IGF-I isomers, corresponding to
the HPLC peaks 1 and 2 in Figure 3.4, respectively. The symbols “itz” and
* represent iminothiazolidine derivatives and B-elimination products,
respectively. The peaks with question marks are discussed in the text. . ..1 l7
LC/ESI-MS results of peptide mixture from the cleavage of a singly
reduced/cyanylated LR3IGF-I isomer, corresponding to HPLC peak 1 in

Figure 3.6. Figure 3.8 A is the RTIC of the cleavage products. Figure 3.8 B

 

is the ESI mass spectra of the peptides corresponding to the RTIC peaks in
Figure 3.8 A. The symbol “itz” represents iminothiazolidine derivative...119
Figure 3.9 LC/ESI-MS results of peptide mixture from the cleavage of a singly
reduced/cyanylated LR3IGF-I isomer, corresponding to HPLC peak 2 in
Figure 3.6. Figure 3.9 A is RTIC of the cleavage products. Figure 3.9 B is
the E81 mass spectra of the peptides corresponding to the RTIC peaks in
Figure 3.9 A ........................................................................ 120
Figure 3.10 The overview of proposed non-chromatographic approach .................. 123
Figure 4.1 Amino acid sequences and disulﬁde structures of LR3IGF-I and IGF-1. The
numbers in the parentheses indicate the location of cysteine residues of
IGF-I. The 13 hold letters at the N-terminus indicate the 13-amino acid
extension in LR3IGF-I. The bold R represents the replacement of E3 of
IGF-I ................................................................................ 131
Figure 4.2 The overview of the processes of refolding and reductive unfolding of
proteins .............................................. 134
Figure 4.3A HPLC separation of CDAP-trapped intermediates during the time course of
refolding of LR3IGF-l ............................................................ 141
Figure 4.33 HPLC separation of CDAP-trapped folding products after folding reached
to the equilibrium, increased the concentration of GSSG to 100 mM and
incubated for another 30 minutes ................................................ 143
Figure 4.4 HPLC separation of CDAP-trapped intermediates in the time course of
reductive unfolding of LR3IGF-I ................................................ 144

Figure 4.5 MALDI spectrum of peptide mixture resulting from the cleavage of the

xi

Figure 4.6

Figure 4.7

Figure 4.8

Figure 4.9

puriﬁed and cyanylated intermediate corresponding to peak 3 in Figure 3.4.
The symbol “itz” represents iminothiazolidine derivatives .................. 147
HPLC separation of intact intermediate (peak 1 here) corresponding to
HPLC peak 2 in Figure 4.3 and one of its expected partially
reduced/cyanylated isomers (peak 2). The peak with a question mark did not
correspond to any speciﬁc cleavage fragments to the disulﬁde structure of
the intermediates as analyzed by MALDI ...................................... 150
MALDI spectra of peptide mixtures resulting from the cleavage of the
puriﬁed/cyanylated intermediate corresponding to HPLC peak 2 in Figure
4.3 and peak 1 in Figure 4.6 (A) and one of its partially reduced/cyanylated
isomers (peak 2 in Figure 4.6) (B). The symbols “itz” and * represent
iminothiazolidine derivatives and B-elimination species, respectively.
Itz60-(65)-83 represents that the residue in parenthesis (residue 65) is
cyanylated but remains uncleaved. Itz60—(65*)-83 represents that the B-
elirnination occurs at residue 65 ................................................. 152
HPLC separation of residual intact intermediate corresponding to HPLC
peakl in Figure 4.3 and its partially reduced/cyanylated
isomers .............................................................................. 154
MALDI spectra of peptide mixtures resulting from cleavage of partially
reduced/cyanylated isomers corresponding to HPLC peaks 2 and 3 in Figure
4.8 of the three-disulﬁde intermediate (peak 1 in Figure 4.3). The symbol
“itz” indicate iminothiazolidine derivatives. The question marked peaks are

discussed in the text ................................................................ 155

xii

Figure 4.10

Figure 4.11

The disulﬁde structure of folding intermediates of LR’IGP-I ............... 158
HPLC separation of CDAP-trapped intermediates during the time course of

refolding of IGF-I ................................................................. 160

Figure 4.12A HPLC separation of CDAP-trapped intermediates in the time course of

reductive unfolding of IGF-I ..................................................... 162

Figure 4.123 HPLC separation of CDAP-trapped intermediates in the unfolding after the

Figure 4.13

Figure 4.14

Figure 4.15

Figure 4.16

equilibrium ........................................................................ 164
HPLC separation of residual intact three—disulﬁde intermediate
corresponding to HPLC peak 1 in Figure 4.11 and its partially
reduced/cyanylated isomers. Peaks 2 and 3 each represent two-disulﬁde
species. The question marked peaks are discussed in the text ............... 167
MALDI spectra of peptide mixtures resulting from the cleavage of the two
singly reduced/cyanylated isomers, corresponding to HPLC peaks 2 and 3 in
Figure 4.13, respectively. The symbol “itz” represents iminothiazolidine
derivatives ........................................................................... 168
HPLC chromatogram of partial reduction/cyanylation mixture showing
residual intact intermediate (two-disulﬁde species) corresponding to HPLC
peak 3 in Figure 4.11 and one of the expected partially reduced/cyanylated
species (peak 2). The question marked peak is discussed in the text ...... 169
MALDI spectra of peptide mixtures resulting from the cleavage of the
puriﬁed and cyanylated intermediate corresponding to peak 1 in Figure 4.14
(also to peak 3 in Figure 4.11) (A) and the partially reduced/cyanylated

isomer corresponding to HPLC peak 2 in Figure 4.14 (B). The symbols “itz”

xiii

Figure 4.17

Figure 4.18

Figure 4.19

Figure 4.20

Figure 4.21

Figure 4.22

and * represent iminothiazolidine derivatives and B-elimination peak,
respectively. The peak with a question mark cannot be assigned to any
speciﬁc cleavage fragment of the intermediate .............................. 170
MALDI spectra of peptide mixtures resulting from the cleavage of puriﬁed
and cyanylated two-disulﬁde intermediates corresponding to peaks 4 and 5
in Figure 4.11, respectively. The symbol “itz” represents iminothiazolidine
derivatives ......................................................................... 172
The MALDI spectrum of cleavage products of the puriﬁed/cyanylated one-
disulﬁde intermediate corresponding to HPLC peak 6 in Figure 4.11. The
symbol “itz” represents iminothiazolidine derivatives ....................... 174
The disulﬁde structure of refolding intermediates of IGF-I ................. 175
HPLC separation of CDAP-trapped intermediates after 30-minutes
refolding starting from the puriﬁed 2-disulﬁde intermediate (11) or 3-

disulﬁde intermediate (III) of LR3IGF-I in the same refolding buffer as used

in refolding of reduced/unfolded LR3IGF-I .................................... 177
The predicted folding pathway of LR3IGF-I ................... A ............... 180

The predicted folding pathway of IGF-I ....................................... 183

xiv

 

CHAPTER 1 '

INTRODUCTION

1. Introduction

Recent advances in mass spectrometry have created new technological capabilities
that are applicable to the study of peptides and proteins involved in biological processes.
The discovery and development of ionization techniques for large bio-oligomers at
signiﬁcantly higher ionizing efﬁciencies permits the detection and measurement of very
large intact or ﬁagmented biopolymeric substances (1). These tools are electrospray
ionization (ESI) (2) and matrix-assisted laser desorption/ionization (MALDI) (3). They
have become among the most powerful methods yet available for the macromolecular
characterization of living systems ranging from the measurement of the molecular

weights and purity of large polar biopolymeric substances to delineation of the detailed
sequence and structure of components of the complex mixtures obtained ﬁ'om their
selective enzymatic or chemical degradation. These ionization techniques have been
revolutionary in spurring research in protein biochemistry, glycobiology, and
biotechnology, and most recently in DNA sequence analysis. Over the next few years
these technologies are destined to become as commonplace in the conduct of biomedical

r “catch as HPLC did in the early 19808.

The synthesis of proteins in organisms starts by transcription of the DNA sequences
to. messenger RNA (mRNA). In eukaryotic organisms, mRNA is processed to remove
inn-0m and is then translated on the ribosomal complex to synthesize the protein. To be

b .
lologically active, all proteins must adopt speciﬁc folded three-dimensional structures

after they depart from the ribosome (4, 5). The genetic information for the protein
speciﬁes only the primary structure, the linear sequence of amino acids in the polypeptide
backbone. Many puriﬁed proteins can spontaneously refold in vitro after being
completely unfolded, so the three-dimensional structure must be determined by the
primary structure (6). How this occurs has come to be known as ‘the protein folding
problem’. It had been primarily of academic interest, but with the advent of the
recombinant DNA revolution and its potential for producing therapeutic proteins, often in
an insoluble, unfolded, inactive, and useless form, has made it also of great practical
importance. For many years, scientists have endeavored to understand the process by
which protein folding occurs, that is, the folding pathway.

In order to understand the folding pathway of a protein, it is necessary to trap and
characterize the folding intermediates (7-9). A particularly difﬁcult aspect in the study of
the protein folding is the fact that intermediates may be short-lived and therefore hard to
isolate and analyze structurally and ﬁmctionally. Disulﬁde-containing proteins provide an
opportunity to isolate and characterize trapped intermediates by quenching the disulﬁde
pairing chemically during the time course of folding (4,10-12). Furthermore, the study of
folding process of disulﬁde-containing protein is also important because disulﬁde
formation often confounds the high-level expression and renaturation of recombinant
proteins (5).

The suﬂiydryl groups of cysteine residues are among the most reactive side chains
in proteins. In 1973, Jacobson et al. (13) introduced speciﬁc cyanylation of sulfhydryl
groups of cysteine residues followed by chemical cleavage at cyanylated cysteine

residues. This speciﬁc reaction with the combination of MALDI-MS has been recently

used to characterize the cysteine status of proteins (14, 15). The methodology has the
advantages of being simple, fast, and sensitive.

In this dissertation, several aspects related to the analysis of cysteine status in a
protein, as well as in folding intermediates, based on the combination of cyanylation
methodology and mass spectrometry will be discussed. In chapter 2, two ionization
technologies, MALDI and ESL will be compared for their capabilities of analyses of
iminothiazolidine-blocked peptides resulting from the chemical cleavage reaction at
cyanylated cysteine residues. In chapter 3, assignment of disulﬁde linkages of proteins
containing adjacent cysteines will be discussed. In chapter 4, the trapping method and
disulﬁde structures of intermediates in folding and reductive unfolding processes
including the folding pathway of Long R3 insulin—like growth factor-I will be
investigated. The purpose of this chapter is to introduce basic aspects of MALDI and
ESI-MS, and describe current methodologies for trapping folding intermediates, locating

cysteines and cystines in proteins, as well as recognizing disulﬁde linkages.

II. Oxidative Folding and Trapping of Intermediates

The general rules governing protein folding will be established only by
characterizing for many proteins the conformations and energetics of structures at
particular points along their folding pathways. Description of the properties of a native
state is relatively simple because it is a single, long-lived conformation. Intermediates
and unfolded states are much more difﬁcult to characterize because they may be short-
lived or exist as a distribution of possible conformers. Two useful methods are currently

available to tackle this problem. A newly developed technique of pulsed-label NMR (16-

19) permits trapping and identiﬁcation of amide groups that are engaged in the structured
elements. This technique can, in principle, be applied to all types of proteins. However,
intermediates trapped by this method are not amenable to chromatographic puriﬁcation.
Therefore, the intermediates cannot be further characterized structurally and functionally.

Disulﬁde bonds introduce nonlocal topological restraints in the folding of proteins,
and their formation provides valuable signals in tracing the folding pathway of cysteine-
containing proteins (20, 21) based on two important phenomena. First, it is known that
the formation of disulﬁdes is coordinated with the folding of protein (22). That is, during
folding, disulﬁdcs reshufﬂe (break and reform) readily in order to adapt to more
favorable conformations. Second, disulﬁde-folding intermediates, trapped either
reversibly or irreversibly, can be separated and isolated by chromatographic systems.
This allows examination of the number, diversity, structure, and properties of
intermediate species. Data obtained from these analyses are indispensable for the
construction of protein folding pathways (23). In this method, the disulﬁde-bonded
intermediates which form during the folding process are chemically trapped in a time
course manner. The trapped intermediates are isolated and structurally characterized, the
kinetics of the interconversion is determined, and the different components are
coordinated with a folding pathway. Therefore, the main concerns in a study of disulﬁde-
containing protein folding are trapping of folding intermediates by quenching the

disulﬁde formation as well as the determination of disulﬁde structure of intermediates.

A. Disulfide Bonds in Protein Folding
Disulﬁde bonds are comparatively labile and cleave readily to thiols by reduction,
and thiols are readily oxidized to disulﬁdes. Cystine can be readily formed by oxidation

of cysteine which is reversibly obtained by reduction (Equation 1.1)

 

 

CO H oxidation
2 PIS/Y 2 ‘ 5 HOZCY\S—S/Ycoz

N reduction
”2 NH2 NH2

cysteine cystine

Equation 1.1. Cysteine and cystine

l. Disulfide Formation
Disulﬁdes do not form spontaneously between thiols, even when in close proximity,
unless there is an appropriate oxidant. While molecular oxygen (in the presence of trace
amounts of heavy metals) may adequately serve this role for some proteins, the most
common source of oxidizing equivalents is a low molecular weight disulﬁde. Protein
thiol oxidation to the disulﬁde results from a thiol/disulﬁde exchange process (24) that
transfers oxidizing equivalents from the low molecular weight disulﬁde to the protein
(Equation 1.2).
— — r
'i‘lg Ililg ‘flg
S' 4' f—S ——-—> "IS-
Re

I «— is
Rn Re Rn Re
L _

_a,°?

?

 

 

Equation 1.2. The mechanism of thiol/disulfide exchange

M

Thiol/disulﬁde exchange occurs via direct attack of a nucleophilic thiolate anion on
one of the sulfurs of the disulﬁde bond (25, 26). The transition state is rather symmetrical
with approximately equal bond formation to the nucleophilic sulfur (KS) and the leaving
sulfur (RigS') and with a relatively small amount of negative charge on the central sulfur
(RS) (27). The rate constant for the reaction increases as the basicity of the attacking
thiolate (R..S') nucleophile increases (pKa increases) and as the basicity of the leaving
thiolate (Rth') decreases (pKa decreases). The pKa of the typical cysteine sulﬂlydryl
group is about 8.6; however, this may vary considerably from protein to protein due to
the effects of the local environment. At pH of 8.6, the typical intermolecular step of
thiol/disulﬁde exchange between small molecules or unhindered protein thiols and
disulﬁdes will occur with a second-order rate constant of about 20 M'ls'l (11). Since
thiol/disulﬁde exchange occurs via the attack of a nucleophilic thiolate anion, the rate
will increase with increasing pH until the attacking thiol is predominantly in the thiolate
form.

Upon adding a disulﬁde reagent to a reduced protein, two sequential thiol-disulﬁde
exchange reactions are required to form one protein disulﬁde bond. The ﬁrst is the simple
chemical reaction between the disulﬁde reagent and one of the cysteine thiol groups, to

generate the mixed disulﬁde (Equation 1.3).

'3” -SSR
K1
+ %
RSSR I + RSH
-SH K-1 -311

Equation 1.3. The formation of mixed disulﬁde bond

The formation reaction of a mixed disulﬁde is bimolecular in either direction, so the
rates should be dependent upon the concentrations of the disulﬁde and thiol forms of the
reagent.

The second step is the one in which a second cysteine thiol group reacts with the

mixed disulﬁde to form the protein disulﬁde bond (Equation 1.4).

'33R Kintra
‘ \S + RSH

 

 

-SH

Equation 1.4. The formation of disulﬁde bond

This step is intramolecular thiol/disulﬁde exchange in the forward direction, and its
rate depends primarily upon the tendencies of the thiols of different cysteine residues to
come into proximity of the mixed disulﬁde, which is primarily controlled by the protein
conformation. Therefore, this step provides the most useful information about protein
conformation and folding, and King, can be calculated from the observed rates of disulﬁde
formation and reduction with different reagents (11).

Once a disulﬁde bond has formed in a protein with more than two cysteine residues,
this disulﬁde bond can undergo intramolecular rearrangements via processes analogous to
the intramolecular step in disulﬁde formation. Again, the rate of a rearrangement reaction
depends on the extent to which the conformation of the polypeptide favors an attack of

the thiolate on the existing disulﬁde bond (Equation 1.5).

 

 

Equation 1.5. Intramolecular rearrangement of a disulﬁde bond

The observed rate of disulﬁde formation depends on several factors, including the
degree to which the protein thiols are ionized, the rate of reaction of thiolate with reagent,
the stability of the mixed-disulﬁde intermediate, and the rate of formation of the protein
disulﬁde from the mixed disulﬁde (28). A redox buffer of a mixture of a low molecular
weight disulﬁde and its corresponding thiol usually is used for the formation of disulﬁde
bonds. During oxidative folding, the disulﬁde component of the redox buffer serves to
provide oxidizing equivalents for protein disulﬁde formation. The thiol component of the
redox buffer serves to reduce native or non-native disulﬁde bonds that may otherwise
lock the protein in misoxidized forms or to catalyze thiol/disulﬁde rearrangements. Since
the reaction of disulﬁde formation is reversible, the capacity to form a disulﬁde bond will
depend on the equilibrium and the rate constants for the individual steps and the
concentration and oxidation potential of the redox buffer (24).

There are two kinds of redox buffer that can be used for disulﬁde formation, linear
disulﬁde and cyclic disulﬁde. Two of them are most commonly used, glutathione and
dithiothreitol. Glutathione is a tripeptide linear disulﬁde reagent. It exists in two forms:
reduced (GSH) and oxidized (disulﬁde, GSSG). Glutathione universally occurs in the

tissues of animals, plants, and microorganisms at comparatively high concentration (from

0.4 to 12 mM, predominantly in the form GSH) (29). GSI-I/GSSG system attributes an
important role in the formation of disulﬁde bonds in proteins during their synthesis in
vivo (30). Therefore, it is a most pertinent (in viva) electron donor and acceptor to be used
for the formation of disulﬁde bonds in vitro. The formation of disulﬁde with GSSG has
the advantage of being rapid at alkaline pH, simple, extremely speciﬁc, and easy to
control (31). 4

Oxidized dithiothreitol (DTT), a cyclic disulﬁde, has been employed in studies of
the mechanism of protein folding. Disulﬁde formation using DTT is much more
discriminating than with a linear reagent. The rrrixed disulﬁde is energetically
unfavorable and rapidly dissociates intramolecularly, so it should not accumulate
signiﬁcantly (32). Only very favorable protein disulﬁdes are formed with this reagent, at
a rate that should be proportional to the rate of the intramolecular step of the reaction

(33).

2. Oxidative-Folding Problem

In the late 19603 and early 19703, a number of investigators began to determine the
mechanisms by which unfolded proteins regain their native conformations in vitro. These
studies have largely been motivated by the recognition that the number of possible
conformations accessible to a polypeptide chain is so great that ﬁnding the native
conformation by a completely random search would be improbably slow. This suggested
that there might be speciﬁc pathways of protein folding and that characterizing partially

folded intermediates associated with these pathways might provide a key to

understanding the relationship between the amino acid sequence and the three-
dimensional structure of a protein.

In 1962, Haber and Anﬁnsen (34) ﬁrst used the oxidative folding of ribonuclease
and other small, disulﬁde-containing proteins to show that the primary sequence
contained enough information to specify the correct three-dimensional structure of a
protein and that the driving force for protein folding was the thermodynamic stability of
the native structure. Two cysteines must be within covalent bonding distance to form a
disulﬁde bond; therefore, identifying the intermediates that are actually populated during
the folding reaction gives clues about the structural features and interactions that guide
the unfolded protein to its native structure (35). At least for the disulﬁde-containing
proteins that spontaneously refold, the overall driving force is the thermodynamic
stability of the native structure-native disulfide pairings give the most stable protein (7).
The interactions that specify which cysteines will be paired and in what temporal order
could be similar to those in the fully folded protein (native interactions), or they may
represent interactions that are not found in the native protein (non-native interactions).

Compared to the necessary rate of protein synthesis and secretion in viva,
uncatalysed in vitra oxidative folding is often a very slow and inefﬁcient , process.
Consequently, in viva, oxidative folding is catalysed (36, 37) and possibly coupled to
protein translation/translocation (38). The rules that govern the oxidative folding of a
protein in vitra may or may not be the same as those that guide in viva folding.

The folding of even simple disulﬁde-containing proteins illustrates the potential
complexity of oxidative folding. The mechanisms that direct the formation of a three-

dirnensional native structure from an unfolded protein must now accommodate the

10

chemistry of disulﬁde formation. For a protein with six cysteine residues that form three
disulﬁde bonds, there are 15 ways of pairing the cysteines to generate molecules with
three disulﬁde bonds. The complexity rapidly increases when more disulﬁdes are
involved.

After more than 30 years since the ﬁrst study of the oxidative folding of
ribonuclease, the knowledge of the mechanism of protein folding is still incomplete.
There are several central questions concerning the oxidative folding of proteins.

(a) Do disulﬁde bond formation determine what structures will form or do other
interactions in the protein dictate which disulﬁde bonds will form?

(b) Are all of the speciﬁc interactions that guide the folding process present in the native
structure or do non-native interactions play a role?

(c) Does oxidative folding proceed through one pathway or do multiple, alternative
pathways exist?

((1) Why do some disulﬁde-containing proteins fold with high efﬁciency into their native
structures while others do not?

(e) Do catalysts of oxidative folding alter the mechanism of folding or simply increase
the rate and extent of native structure formation?

(1) Is the mechanism of oxidative folding in vivo the same as that in vitra?

3. Folding Pathway of Bovine Pancreatic Trypsin Inhibitor (BPTI)
The most detailed and informative folding pathway elucidated thus far is that of
bovine pancreatic trypsin inhibitor (BPTI), a small single-domain 58-residue amino acid

protein stabilized by three disulﬁde bonds. The folding pathway of BPTI (21) is

11

illustrated in Figure 1.1. The reduced BPTI is a random coil, and it can efﬁciently refold
by sequentially passing through one-disulﬁde intermediates, two-disulfide intermediates,
and ﬁnally to the native protein. The folding experiments of BPTI (4, 11, 39, 40) revealed
two striking features of the BPTI folding mechanism. First, the process is distinctly non-
random: only a small fraction of the possible disulﬁde-bonded intermediates are detected
during folding or unfolding. Second, disulﬁde formation is not a simple sequential
process: the kinetically preferred mechanism includes intramolecular rearrangements
involving intermediates with non-native disulﬁde bonds (i.e, those not present in the
native protein). This ﬁnding raised the possibility that the amino acid sequence might
specify the structures of folding intermediates which have different conformations from
that of the ﬁnal folded protein. This possibility would add new complications to the

already formidable problem of predicting the three-dimensional structure of a protein.

B. Trapping of the Folding Intermediates

To identify disulﬁde-bonded folding intermediates, rapid quenching of the folding
reaction is necessary to trap the intermediates in chemically stable forms that can be
physically separated and analyzed. The criteria for a good trapping agent include that it
block the thiolates of an intermediate rapidly, cOmpletely, and without modifying the
protein at sites other than thiols. Several methods have been used to trap the

intermediates.

12

 

Others H
H
. H
1-1
(30-51. 538) <-<> <=>
- N
C
(54101ng (3e51, 5-55) (30-51. 5-55. 14-38)

/ (30-518. 514)
N<§>

SH

Reduced

C SH
(30.51) (30-51. 14-38)

Figure 1.1. Folding pathway of BPTI

1. Acid Trapping

The folding of proteins are conducted under alkaline condition, usually at around
pH 8.5. The pKa of the typical cysteine sulthydryl group is about 8.6. Therefore, the
thiolate of the intermediates can be rapidly protonated by adding acid in the folding
solution to a pH less than 3 and the further thiol/disulﬁde exchange can be stopped

(Equation 1.6).

   

Equation 1.6. Trapping folding intermediates by acidiﬁcation

Quenching by acidiﬁcation is rapid and occurs at the diffusion control rate (109
M'ls“) (41). Subsequent reversed-phase HPLC at pH 2 is well suited for the separation of
acid-quenched intermediates without introducing signiﬁcant rearrangements during the
separation. A practical advantage of acid quenching is its reversibility. As a result, it is
possible to purify an acid-quenched intermediate and subsequently to allow further
rearrangement or folding to occur so that more detailed information of the folding
pathway can be obtained (4, 42, 43). The folding intermediates of several proteins have
been trapped by this method (4, 44-47). While trapping by acidiﬁcation is rapid, it does
not completely stop thiol/disulﬁde exchange; it just slows it down by reducing the
concentration of the reactive thiolate anion. Assuming that the pKa of the thiol is 8.5, an

intramolecular thiol/disulﬁde exchange occurring with a half-life of lms at pH 8.5 would

14

occur with a half-life of approximately 1600s at pH 2 (41). Therefore, samples must be
analyzed promptly after the acidiﬁcation and the intermediates trapped by acid must be
chemically modiﬁed before the disulﬁde bond structure can be deﬁned. In this respect, it

is a cumbersome procedure.

2. Trapping by Alkylation

Folding intermediates can also be trapped by addition of an alkylation reagent, such
as iodoacetate, iodoacetomide, and 4-vinyl pyridine (48-51). Iodoacetate is the most
commonly used. The thiol groups of the intermediate will be alkylated by the alkylating

reagent and further thiol/disulﬁde exchange will be stopped (Equation 1.7).

-S-CHz-COOf

+ ICHZCOO' ——> -s

 

Equation 1.7. Trapping folding intermediates by alkylation

At a concentration of 0.2 M and pH 8.7, iodoacetate reacts with sterically accessible
sulfhydryl groups with a half-life of 0.3 s (23). This is much faster than many of the
intermolecular thiol/disulﬁde exchange reactions during oxidative folding; at 1mM
GSSG, the half-life for reaction of GSSG with an exposed SH at pH 8.7 should be about
30 8. However, rearrangement of intermediates during trapping with iodoacetate has been
observed for both BPTI (4) and ribonuclease (52). Quenching by 0.2 M iodoacetate is

comparable to or, in some cases, slower than the rate of some of the intramolecular

15

rearrangements that interconvert intermediates of the same oxidation state (4, 53). When
quenching by alkylation is slow compared to rearrangement, and the alkylating reagent
reacts preferentially with one of the intermediates, alkylation will lead to overestimation
of the amount of the intermediate that reacts fastest with the alkylation reagent.
This could be a more severe problem for partially structured intermediates where steric
hindrance would retard the rate of alkylation of some thiols (54). Although a high
concentration of iodoacetate can be applied to mininize such a side reaction, modiﬁcation
of other functional groups by a high concentration of iodoacetate may provoke other

problems (55).

3. Other Trapping Methods

Acid and alkylation trapping described above were attacked by Rothwarf and
Scheraga (56, 57), who insisted that the criteria for a good blocking agent are that it block
quickly, completely, and without modifying the protein at sites other than thiols. They
proposed a new alkylating reagent, 2-aminoethyl methanetlriosulfonate
[(NH2)C2H58S02CH3] (AEMTS) (Equation 1.8). AEMT S is a thiosulfonate reagent (58)
which reacts with free thiols in the solution rapidly; even thiols that are essentially buried
should still be accessible by local or global ﬂuctuations of the protein within the 2 min

blocking time (59).

16

 
  

+
-S-S-CH2-CH2-NH3

f.’ .

+ CH3-ﬁ-S-CH2-CH2-NH3 —>

O

 

Equation 1.8. Trapping folding intermediates by AEMTS

There are several advantages of trapping by AEMTS (57). It can reversibly block
the thiol groups in the intermediates so that it permits the refolding process to be restarted
ﬁ'om isolated intermediates, an invaluable tool in the determination of refolding
pathways. In addition, blocking with a reagent such as AEMTS provides a basis for
separation of intermediates in a predicable manner due to the addition of a positive
charge to a thiol group by AEMTS. It is especially important for very complicated
refolding processes containing large numbers of chemically distinct species. Recently,
Thannhauser et a1. combined AEMTS trapping and amino acid analysis to quantitatively
determine the number of cysteines and cystines in the folding intermediates of hirudin
(60). While trapping by AEMTS is rapid and speciﬁc, disulﬁde scrambling is still
questionable since the blocking of a thiol group by AEMTS is a thiol/disulﬁde exchange
reaction.

Recently, a new trapping reagent was suggested to study the folding of macrophage
colony stimulating factor (M-CSF) (61). The protein-folding intermediates were trapped
by selective modiﬁcation of bis-cysteine suﬂrydryls with phenylarsonous acid derivatives
(PAA) (Equation 1.9). PAA-derivatives form cyclic dithioesters in the presence of two

Snitably spaced sulﬂrydryl groups of the intermediates. Although the trapping method

17

was found to have the advantage of a broad derivatizing pH range, it depends too much
on the structural accessibility of bis-cysteine SH groups, therefore, it is limited in

practical use.

   

Equation 1.9. Trapping folding intermediates by PAA

III. Recognizing the Location of Cysteines and Cystines

Practically, all the methods used to recognize the location of cysteine and cystine
residues in proteins can be used for their location in the intermediates. Localization of
cysteines and cystines in proteins is the oldest and best-studied ﬁeld in protein chemistry.
A large number of reagents and procedures have been developed and used to address this

problem and there is no indication that the search for better methods is over yet.

A. Classical Approach

In the classical approach, one derivatizes free sulfhydryl groups with UV absorbing,
radioactive or ﬂuorescent groups under the condition to prevent sulﬂrydryl-disulﬁde
exchange, and then cleaves the protein by chemical or enzymatic method. The resulting
peptide ﬁagments are separated by and collected from chromatography. The collected

fractions showing speciﬁc UV, radioactive or ﬂuorescent detection are recognized,

18

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

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

Iodoacetate and pyridylethylation are the most extensively used alkylating
derivatization methods for sequencing of peptides (62-68). Recently, Lee et a1. (67)
identiﬁed the position of the sulﬂrydryl group of human glucocerebrosidase by modifying
the sulﬂrydryl group with ["C] iodoacetic acid without reduction of disulﬁde bonds,
digesting the protein with both protease-V8 at pH 4.0 followed by oc-chymotrypsin at pH
7.5, and ﬁnally sequencing the peptide based on detected radioactivity. Krieglstein et a1.
(68) determined the location of sulﬂiydryl groups and disulﬁde bridges of botulinum
neurotoxin type A. They converted ﬁve Cys residues to S-pyridylethyl cysteine residues;
aﬁer mercaptolysis converted all nine Cys residues to the S-pyridylethylated form. After
conﬁrming the predicted number of Cys residues by amino acid analysis, the positions of

the ﬁve Cys residues carrying sulﬂrydryl groups and the four involved in disulﬁde

l9

bridges were determined by comparing the elution patterns in HPLC of the cyanogen
bromide mixtures of the exclusively alkylated and the mercaptolyzed-alkylated
neurotoxin. The chromatographically isolated components were identiﬁed by N-terminal
amino acid sequence analysis.

While iodoacetate has the disadvantages of coelution with PTH-Gln and PTH-Glu
on HPLC aﬁer Edman degradation of modiﬁed protein and makes identiﬁcation
problematic, pyridylethylation results the problem of N-tenninal alkylation of protein
because of the high reactivity of 4-vinylpyridine (69, 70). The other alkylating reagents
have been developed for determination of cysteine residues (69-73). The alkylating
reagent, 3-Bromopropylamine (72, 73), offers the advantage of easy identiﬁcation of its
derivative of cysteine residue by protein sequencing and amino acid analysis over other
alkylating reagents in the modiﬁcation and subsequent identiﬁcation of cysteine residues
by protein sequencing. J ue & Hale have developed the conditions for concomitant on-
sequencer reduction and alkylation of cysteines in proteins with tri-n-butylphosphine and
3-bromopropylamine before sequencing the proteins (71).

While alkylation by iodoacetates continues to be useful, there has been far greater
interest in the use of this chemistry as a mechanism for introducing a larger molecule
which can serve as structural probe. Examples include the ﬂuorescent reagents, 5-
(iodoacetamido)ﬂuorescein (IAF) (74) and 5-[2~((iodoacetyl)amino)-ethyl]naphthalene-
l-sulfonic acid (1,5-IAEDANS) (75-77).

Bishop et at. has used IAF to selectively label two cysteine residues on the surface
of CaI-ATPase of sacroplasmic reticulum (SR), digest with trypsin, purify the tryptic

peptides by size-exclusion and reverse-phase HPLC, and identify the cysteine residues by

20

amino acid sequencing (78). Recently, a combination of 1,5-IAEDANS labeling of
cysteine residues, enzymatic digestion, and MALDI-PSD has been developed to assign
the location of free cysteine residue of testis angiotensin-converting enzyme (tACE) (79).
The ﬁee cysteine residue was located by labeling with 1,5-IAEDAN S, isolating the
ﬂuorescent peptide from enzymatic digests by HPLC, and analyzing its sequence by
MALDI-PSD.

Labeling with IAEDANS or IFA suffers ﬁom less selectivity and reactivity toward
sulﬂrydryls, as well as the fact that the reagents themselves are ﬂuorescent before
reaction with thiols. A highly reactive and speciﬁc reagent for protein thiols, 4-
(aminosulfonyl)-7-ﬂuoro-2,1.3-benzoxadiazole (ABD-F) has been developed (80). The
selectivity of ABD-F is superior to that of other sulfhydryl-reactive ﬂuorophores. In
addition, unlike most other reagents previously utilized, ABD-F is nonﬂuorescent before
reaction with protein thiols. The ABD-Cys adduct absorbs maximally at 385 nm and has
a ﬂuorescence emission maximum at 520 nm, allowing simple and sensitive detection in
the presence of an excess of unlabeled protein.

ABD-F has been utilized for analysis of cysteine residues in the proteins (81-83).
The sulﬂrydryl and disulﬁde bond locations in the proteins were established using
tributylphosphine (Bu3P) as the reducing reagent and ABD-F as the sulﬂlydryl reagent
(83). The reaction was ﬁrst run in the absence of reducing reagent, and all the peaks, both
ﬂuorescent and nonﬂuorescent ones were collected from HPLC of the tryptic peptides.
An aliquot of each of the nonﬂuorescent peaks was then treated simultaneously with
B113P and ABD-F since the two reagents do not react with each other. Sequencing the

ﬂuorescent peptides established the positions of the cysteine residues and disulﬁde bonds.

21

The ability to reduce disulﬁdes and block the resulting thiols simultaneously in a single
reaction, the stability of ABD-Cys to acid hydrolysis, and the recovery, albeit in low
yield, of PTH-(ABD-)Cys on sequencing, along with the characteristic ﬂuorescence of

ABD—thiol derivatives, make it a favorable methodology.

B. Aﬂ'mity Chromatography

The sulﬂrydryl group is the only functional group in a protein for which conditions
are available for the formation of a stable covalent bond which can be split under mild
conditions (84). A thiol-containing proteins can be attached on an afﬁnity column. After
the immobolized protein is cleaved chemically or enzymatically on the afﬁnity column,
the nonsulfhydryl-containing peptides do not bind to the column and can be washed
away. The attached sulﬂrydryl-containing fragments are then removed from the column
and subjected to further structural characterization by Edman degradation and/or mass
spectrometry. The afﬁnity interactions that can be used for the sulfhydryl-containing
proteins include (85): (i) binding to afﬁnity media that contain heavy metals, (ii) binding
to afﬁnity media with reactive disulﬁdes that undergo facile thiol-disulﬁde exchange, and
(iii) binding to afﬁnity media that contain chelated zinc. While these methods were
originally used to purify sulfhydryl-containing proteins, this approach was recently used
to locate cysteine residues in proteins by mass mapping of bound sulﬂrydryl-containing
peptides (86). This approach is powerful for locating sulﬂrydryl groups in high-mass
proteins because the HPLC separation of the digests is unnecessary. But conditions must
be established and optimized under which all the sulﬂlydryls can be attached to the

column.

22

C. Cleavage at Cysteine Residue

The cleavage of the peptide bond at cysteine residue after cyanylation was ﬁrst
introduced by Catsimpoolas and Wood (87). In this reaction, the conversion of an SH
group into an SCN group was achieved in two stages by successive treatment of a protein
with Elhnan’s reagent and then with cyanide. Jacobson and Stark (88, 89) showed that 2-
nitro-S-thiocyanobenzoic acid (NT CB) speciﬁcally cyanylates cysteine sulﬂlydrls, which
subsequently cleave at N—terminal side of the cyanylated cysteinyl residues under mildly
alkaline conditions to form an amino-terminal peptide and a series of 2-
iminothiazolidine-4-carboxylyl (itz) peptides. Walkselman and Guibe—Jampel proposed
another cyanylating reagent, 1-cyano-4-dimethylamino-pyridinimn (90). This reagent has
the advantages of water solubility and reactivity at acidic condition.

The cyanylation/chemical cleavage reaction of cysteine residue can be utilized to
recognize structural information for the protein. If a protein contains 11 cysteine residues,
the cleavage reaction results in the formation of n+1 peptide ﬁ'agments, mass alignment
of which indicates the number and location of cysteine residues. While SDS-PAGE can
be used to mass map the peptide ﬁagments (91 -93), it suffers the poor mass accuracy
(error > 5%) (89, 94). Daniel et al. used electrospray mass spectrometry to determine the
masses of the NTCB cleavage products of E. cali dihydroorotase to identify the location
of the active cysteine residues in the protein (95).

Recently, Wu, et al. (96) combined the cyanylation reaction, speciﬁc chemical
cleavage at cysteine residues, and MALDI-MS to quantitatively determine the number
and location of cysteine and cystine residues in the proteins. This method has the

advantages of being precise, simple and highly sensitive. However, MALDI-MS has a

23

signal suppression phenomenon which may result in incomplete assignments; the
cyanylation reaction with NTCB is performed at alkaline condition, at which sulﬂrydryl-

disulﬁde exchange may proceed for some of the proteins.

IV. Assignment of Disulfide Bonds

Disulﬁde bonds between cysteine residues, to generate cystine, are a common post-
translational modiﬁcation of proteins, principally those that operate in the extracellular
milieu. A full description of the covalent structure demands that the connectivity of the
bridged cysteines be analyzed. Although the amino acid sequences of proteins can be
deduced from the corresponding cDNA sequences, the state and connectivity of the
cysteine residues after the translation cannot be predicted accurately. Characterization of
disulﬁde linkages in proteins has become an essential part of the analysis of recombinant
DNA products, because molecules having incorrect disulﬁde linkages may have much
lower activity than that of the desired product. Therefore, it is important to verify that
recombinant materials have the correct disulﬁde linkages.

Although disulﬁdes are chemically rather simple, two factors can make the analysis
difﬁcult. First, the number of possible isomers grows rapidly as the number of bridges
increases——1, 3, 15, 105, 945, etc. Second, base and reducing agents can catalyse
exchange among partners, obscuring the original connectivity (97). Native proteins under
physiological conditions are deceptively stable, but onCc they are denatured or nicked by

proteases, disulﬁde exchange occurs when even a trace of thiol is present.

24

A. Conventional Fragmentation Strategy
The general strategy for locating disulﬁde bonds in proteins by conventional
methods involves several steps:
(1). Digest proteins by chemical reagents and/or enzymes between the half-cystinyl
residues avoiding disulﬁde exchange.
(2). Separate ﬁagments also under conditions that stabilize the disulﬁde bonds.
(3). Detect disulﬁde-containing peptides.
(4). Identify subﬁ'agments, and hence individual cysteine residues that are connected.
The conventional fragmentation approach has been well established for recognizing
disulﬁde bond structure of proteins (98-100). The characteristic aspects of the strategy

will be discussed below.

B. Fragmentation of Proteins A

The aim of ﬁ'agmentation of a protein for disulﬁde mapping is to produce by
enzymatic or chemical digestion a deﬁnitive set of peptide fragments only containing one
disulﬁde bond. If a protein is not cleaved between every half-cystinyl residue by an
enzyme or a chemical reagent, the peptides obtained by the cleavage must be cleaved
again using another enzyme or chemical reagent. It is easy to identify the peptide
generated by highly speciﬁc cleavages of proteins (101-103), because the N- or C-
terrninal residues of the peptides are known. The general risk of disulﬁde exchange
occurs during digestion; even though the protein, enzymes, and reagents may be initially
the of thiols, the latter can arise by base catalyzed degradation of disulﬁdes. Under any

reasonable conditions very little thiol will be produced, but only catalytic amounts are

25

needed to cause extensive disulﬁde rearrangement during long incubations. Therefore,
the choice of a proper enzyme should satisfy three requirements: the speciﬁcity of the
enzyme; the capability of cleaving between every half-cystinyl residue; and the
conditions to minimize disulﬁde scrambling.

Cyanogen bromide, trypsin, and Staphylococcus aureus V8 protease are available to
cleave proteins speciﬁcally. Speciﬁc chemical cleavage at the C-terminal side of Met by
cyanogen bromide (CNBr) has been widely applied for initial cleavage of proteins due to
its weak acidic reaction condition and volatile reagent and by-products. Since Met is not
a particularly abundant constituent of proteins, cleavage with CNBr usually gives large
peptides. Subsequent digestion by other enzyme(s) is required to produce smaller
peptides (101). Trypsin has also been used ﬁ'equently because it cleaves on the C-
terrninal side of Arg or Lys residues with high speciﬁcity. Unfortunately, trypsin has
maximum activity at pH 8.3, and is not active in acid. As a result, disulﬁde scrambling
may occur during enzymolysis. Despite this shortcoming, trypsin is still the most widely
used enzyme for the digestion of proteins because of its high speciﬁcity. Greatly
increased use is now being made of the V 8 protease from S. aureus. This enzyme
speciﬁcally cleaves on the C-terminal side of Glu or Asp residues in the pH range of 4.0
to 7.8.

Unfortunately, if proteins cannot be cleaved by speciﬁc cleavage reagents, they
must be cleaved by non-speciﬁc reagents. Pepsin, thermolysin and partial acid hydrolysis
are uSeful for this purpose. The major advantage of using pepsin and partial acid
hydrolysis to generate peptides is that disulﬁde bonds in proteins are fairly stable under

mild acidic conditions (104, 105). However, the protein digests using such nonspeciﬁc

26

cleavages are complex mixtures, which make the separation and identiﬁcation much

more difﬁcult.

C. Purification of Cystinyl Peptides

The puriﬁcation of disulﬁde-bonded peptides constitutes a special problem. The
difﬁculties in this area relate to the instability of disulﬁde bonds. Under the condition of
mild acidity, i.e, pH 2.0-6.5, the disulﬁde bond is relatively stable. This requirement
restricts the choice of isolation techniques. Reversed-phase HPLC under acidic
conditions (a gradient of acetonitrile in 0.1% triﬂuoroacetic acid, for example), is ideal
for this purpose, being both rapid and employing conditions favoring stability of disulﬁde
bonds. However, the puriﬁcation of digests from high-mass proteins is still a challenging
task even though microbore or capillary HPLC columns are use because so many peptide
fragments rrright be produced after hydrolysis.

Gel ﬁltration and ion-exchange chromatographic technologies are applicable for
puriﬁcation of digestion mixtures. The techniques are based on the chemical modiﬁcation
of cysteinyl residues to cysteic acids after the reduction of disulﬁde bonds in cystine
peptides. Non-cystine peptides may be separated with great case from the more highly

charged cysteic acid peptides generated from cystine peptides.

D. Identiﬁcation of Cystine-Containing Peptides

The conventional approach for the identiﬁcation of cystine-containing peptides after

the puriﬁcation is very tedious and time consuming. The peptides, which may contain

27

intra- or intermolecular disulﬁde bonds, are reduced, alkylated, isolated by HPLC again,
and identiﬁed by their sequence with Edman degradation (106, 107).

The soﬁ ionization methods in mass spectrometry have proven useful for locating
disulﬁde bonds in proteins. One of the methods is by fast atom bombardment (F AB-MS)
(108-112). Disulﬁde—containing peptides suitable for analysis by FAB-MS can be
produced by treating proteins with enzyme. An aliquot of the digest can be analyzed by
FAB-MS to determine the molecular weights of the peptides which are related to speciﬁc
segments of the native protein; then following reduction or oxidization of the disulﬁde
bonds with chemical reagents, such as the reducing reagent dithiothreitol or the oxidizing
reagent performic acid. The sample is analyzed again by FAB-MS. The disulﬁde
assignment can be completed by comparing the two FAB-MS results. An example is

shown in the following scheme:

 

11111“ 378
Ala Gly Cys Lys Ala Gly Cys Lys
'3 Reduction in
\ >
3 SH
l
Thr Phe Thr Ser Cyi Thr Phe Thr Ser Cys
1111-1" 933 MHI 558

This method soon became an accepted method and was successfully applied to
model proteins such as insulin (109), hen egg-white lysozyme and bovine ribonuclease A
(111), and duck egg-white lysozyme (110). However, it needs speciﬁc enzymes to cleave

the protein to form peptides only containing one disulﬁde bond, and the fragments should

28

not be large since FAB-MS is not suitable for high mass analysis, and it cannot be used
for close or adjacent cysteine analysis.

The another MS method is tandem mass spectrometry (MS/MS) employing high-
energy collision-induced dissociation (CID) or low-energy CID, which produces peptides
by enzyme digestion. The peptides containing a disulﬁde bond are selected as the
precursor ion which is sequenced by MS/MS, and after reduction the peptides are
sequenced again by MS / MS; the disulﬁde linkage can be assigned by comparing the two
results (113-117).

Electrospray ionization (E81) (2) and matrix-assisted laser desorption ionization
(MALDI) (3) have revolutionized the role of MS for the analysis of proteins and
peptides. Peptide-mapping by E81 (118, 119) or MALDI-MS (120, 121) are being used
increasingly in the characterization and identiﬁcation of disulﬁde bonds. The strategy
used here is exactly the same as in F AB-MS. However, they provide much more
sensitivity (low picomole to high femtomole). Besides, the easy interface of ESI to HPLC
has made the puriﬁcation of the disulﬁde-containing peptides an easy task. MALDI-MS
provides the capacity of rapidly analyzing small quantities of mixtures of proteins and
peptides. It is particularly well suited to the determination of peptide mass mapping
without the previous isolation or fractionation.

Recently, MALDI-MS with postsource decay was used to generate fragment ions
from peptide fragments containing heteropeptides linked together by two disulﬁde bonds
(122). Postsource decay analysis of these peptide fragments generates a series of singly
charged fragment ions that, in addition to the peptide sequence ions, provide useful

information for assigning disulfide arrangement in highly bridged disulﬁde-linked

29

peptides. The assignment was made possible by fragmentation at peptide bonds between
two Cys residues in a peptide that constitutes the highly bridged fragment, while retaining
the disulﬁde linkage to the other peptide. Fragmentation using other types of instruments,
such as quadrupole ion-trap mass spectrometry with CID dissociation, usually did not
generate such fragment ions. The method was used to facilitate the assignment of
disulﬁde structure in a tumor necrosis factor binding protein, which contains 162 amino
acids and 13 disulﬁde bonds.

The conventional methodology requires the cleavage between half-cystinyl peptides
by enzyme(s). If proper enzyme(s) cannot he ﬁnd for this purpose, partial hydrolysis by
acid can be used (105). Partial acid hydrolysis is performed under controlled conditions
in which only limited peptide ﬁagments are formed. Partial acid hydrolysis is attractive
as disulﬁde bonds are particularly stable in dilute or weak acids. However, it generates
extremely complex digests. Unless one is dealing with a relatively small molecule, the
problem is almost intractable without complicated chromatographic separation and mass

spectrometry technique.

E. S-S Assignment of Adjacent Cysteines

The major problem of the conventional methodology is that it is unlikely to ﬁnd a
cleavage reagent that can cleave between the two adjacent or close spaced cysteine
residues and produce peptide fragments that only contain one disulﬁde bond. In this case,
alternative methodologies should be used.

Different approaches have been reported to solve the problem, which include the

combination of organic synthesis of the possible isomers of a disulﬁde-containing peptide

30

and high-performance tandem mass spectrometry (123) and multiple steps of Edeman
degradation (124-129).

The selective reduction of disulﬁde bonds can be applied to the problem (130-133).
The selective reduction is based on the observation that disulﬁde bonds in a protein can
often be reduced sequentially to thiols. The disulﬁde bonds of the protein are reduced
stepwise by controlling the reduction conditions (temperature, concentration of reducing
reagent, time, and denaturing conditions). The thiols so formed are labeled with a
reagent, such as alkylating and ﬂuorescent reagents. The protein is then reduced
completely, and subsequently cleaved enzymatically. The location of the disulﬁde bond
is determined by isolating and identifying the labeled peptides. The selective reduction
only can be applied for the proteins containing disulﬁde bonds that have different
stability.

An alternative method for analyzing highly bridged proteins is based on the partial
reduction of disulﬁde bonds by his-(2-carboxyethyl)-phosphine (TCEP) at pH 3,
puriﬁcation of the reduced protein by HPLC, alkylation of free thiols, analysis by
sequencing, providing explicit assignment of disulﬁde bonds that had been introduced
(134,135). This method has the advantages of being suitable for analyzing adjacent
cysteines and tightly clustered cysteines, and preventing the sulfhydryl-disulﬁde
exchange when performing partial reduction. The disadvantage is the limit of the analysis
of the alkylated protein by sequencing. It is time consuming and hard to analyze large
proteins.

Recently an approach using a combination of the partial reduction by TCEP and

cyanylation of free thiols by 1-cyano-4-dimethylamino-pyridinium (CDAP)

31

tetraﬂuoroborate at acidic condition was reported (136). Following the cleavage of the
peptide bond on the N-terminal side of cyanylated thiols, the mass mapping of peptide
fragments was performed by MALDI-MS. This method has the advantages of minimizing
disulﬁde exchange, the possibility to analyze the proteins containing close or adjacent

cysteines, and being fast, simple, and sensitive.

V. Matrix-Assisted Laser Desorption/Ionization Mass Spectrometry (MALDI-MS)
and Electrospray Mass Spectrometry (ESI-MS)

In the last twenty years, important achievements in organic mass spectrometry have
been obtained by the development of a number of new ionization techniques for the
analysis of large, polar and thermally labile compounds. The general problem to be
solved was how to generate intact gas-phase ions of the intact molecule ﬁ'om compounds
that are usually. solids and degrade or decompose when being exposed to thermal heating.
Progress was achieved with the discovery of the desorption ionization techniques; plasma
desorption mass spectrometry (PDMS) (137) and fast atom bombardment mass
spectrometry (FAB MS) (138). The introduction in 1988 of two new techniques, matrix-
assisted laser desorption/ionization (MALDI) (3) and electrospray ionization (ESI) (139),
has extended the applicability of mass spectrometry in the biological science far beyond
the scope of PDMS and FAB-MS. Although MALDI and E81 utilized very different ideas
and principles, both new techniques have overcome the main difﬁculties with MS
methods, 1. e., the desorption and ionization of large and labile biomolecules, and shown

considerable promise in characterizing biomolecules by accurate determination of

32

molecular mass up to a few hundred thousand daltons. In combination with biochemical

techniques, numerous applications have been achieved (1).

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

The ﬁrst attempts to make use of laser ionization for the analysis of organic
compounds date back to the 1970s. Especially, the desorption of ions containing the
intact molecule of complex compounds, like stachyose and digitonin (140), may be
regarded as a milestone. However, signiﬁcant dissociation occurs in larger molecules
because of the higher energy needed to desorb the large molecules, and the energy goes
directly into the analyte. The major advance in laser desorption was the introduction of
matrices (MALDI) by Hillenkamp and Karas in the late 19803 (141). The key to the
MALDI process is to embed the analyte macromolecule in a high molar excess of a
suitable matrix of molecules having a strong absorption at the laser wavelength. This
induces an efﬁcient transfer of the laser-pulse energy to the analyte and results in a soft
desorption process, i. e., little or no fragmentation. The matrix also provides photoexcited
acid or base sites for ionization of sample molecules during ion/molecule collisions
(142).

Although the mechanism of desorption and ion formation is complicated and
unresolved, the formation of ions to nearly 106 Da and the high sensitivity of MALDI
(subfemtomole) has led to the explosive commercialization of instruments in not only
mass spectrometry laboratories, but also in biochemical facilities and laboratories where

bench scientists are able to use the instrumentation.

33

1. Matrix and Sample Preparation

The matrix is the key to the successﬁrl analysis of high-mass biomolecules. While
several hundred of organic compounds have been investigated (143, 144), only a few of
these are widely applicable for peptide and protein analysis: a-cyano-4-hydroxycinnamic
acid (4-HCCA), 1,5-dihydroxyoxybenzoic acid (DHB), and 3,5-dimethoxy-4-
hydroxycinnamic acid (sinapinic acid) (145). It was considered important that the laser
wavelength match the desorption maximum of the matrix compound. All the matrices
above have strong UV absorption in the 320-350 nm range and can be used with a much
cheaper nitrogen laser (337 nm) or frequency-doubled Nd-YAG laser (355 nm). Other
important matrix characteristics include miscibility with the analyte in the solid phase,
solubility in the same solvents required for the dissolution of the analyte, vacuum
compatibility (low vapor pressure), a chemical composition that promotes the ionization
of matrix substituents that can donate protons to the analyte, nonreactivity with the
analyte, and other physical properties such as the low heat of sublimation and a capacity
to crystallize readily (146).

A number of different methods have been developed for growing analyte-matrix
crystals. The original method described by Karas and Hillenkamp has been named the
‘dried-droplet’ method (141). It entails drying a droplet of a solution containing the
matrix (1-10 mmol) and the protein (1-10 umol). A simple variant on the “dried droplet”
method was proposed by Vorrn, et al. (147). They ﬁrst placed a thin layer of the matrix
compound onto a metal surface and then put a droplet of the analyte containing solution
on top of the layer of matrix compound. As the droplet dries it dissolves some of the

matrix which crystallizes onto the substrate matrix layer when the droplet dries

34

completely. The crystallized matrix layer is very thin, allowing for more reproducible
mass measurements and producing ions for < 10 laser shots on a spot.

It is possible to grow large, analyte-matrix crystals under near equilibrium
condiitons, rather than a rapidly drying droplet (148). Supersaturated matrix solutions
containing analyte will form crystals that can be used directly in an ion source.
Supersaturation can be achieved either by heating and cooling or by slow evaporation.

A recently developed technique of producing crystals for use in MALDI ion sources
involves growing a thick ﬁhn of matrix crystal (149). The substrate for the crystals
(usually a piece of metal) is ﬁrst covered with the matrix material, by rapid drying of a
solution of the matrix dissolved in an organic solvent. The small matrix crystals that
cover the surface are then crushed and smeared over the substrate surface, resulting in a
well-adhered layer of stressed, crystalline matrix. A saturated solution containing the
matrix and analyte is then placed onto the surface and allowed to dry slightly. The
stressed crystalline material on the substrate acts to seed crystal formation at many sites
on the surface, resulting in the rapid growth of a rather uniform polycrystalline ﬁlm of
analyte-matrix crystals over the surface. The liquid drOp can then be removed by blotting.
The resulting ﬁlm is very strongly adhered to the substrate and it can be washed
thoroughly to remove any contaminants that were present in the analyte solution.

The mechanism of “matrix assistance” for the transfer of analyte molecules to the
gas phase and their ionization remains incompletely understood; the choice of matrix and
method application is still empirical. Sample preparation by different matrices and

solvents (150), matrix additives (151), and evaporation rate (147, 152) affects the

35

resolution and sensitivity of MALDI. Optimal results require parallel analyses under

different conditions.

2. Instrumentation

To date, MALDI of large molecules has been obtained mainly in combination with
the ﬁeld-free time-of-ﬂight (TOF) mass analyzer. A schematic drawing of MALDI-TOP
instrument is shown in Figure 1.2.

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

The time-of-ﬂight mass analyzer is simple, cheap, and well-suited to the pulsed
nature of laser desorption in MALDI (153). It has virtually no upper mass range and is

therefore compatible with MALDI, which can produce very high m/z ions. Another

36

EOEEEE wﬁtuOHrarEE .N... 0.59".

 

J. 3509

 

 

   

 

 

 

 

 

 

.28qu .//.._o_ _
/ _ A
I— BE 29.“. 7 #4 , _
._ %% nae % _ _ _ 4./ “noun
‘ x a. Emm
I..— _ \~\
\\
x 528mm _ amour... \\
E239... _ U \\

 

m \

Emmmmwwl Hut

.9325 Emam

37

advantage of TOF mass analyzer is its capacity to generate the entire spectrum from
every single laser shot without losing information, in contrast to a scanning mass
analyzer.

The time required for ions to traverse the ﬂight tube, tof, is dependent on their
massses and is described by the relationship: tof=L/v=L(m/22 eV)”2, where L is the
length of ﬂight tube, v is the ion velocity, m is the mass of the ion, and V is the
acceleration potential. Thus, low-mass ions have a shorter ﬂight time than heavier ions.
They are separated into a series of spatially discrete individual ion packets, each traveling
with a velocity characteristic of its m/z ratio. A detector positioned at the end of the ﬁeld-
free ﬂight-tube produces a signal as each ion packet strikes it. A recording of the detector
signal as a function of time yields a tof spectrum. The difference between the start time,
triggered by the laser pulse and common to all ions, and the arrival time of an individual
ion at the detector is proportional to (m/z)”2 and can be used to convert the x-axis of the
spectrum (ion arrival time, tot) into a m/z ratio axis (a conventional mass spectrlun). The
TOF analyzer is highly efﬁcient because all ions of different m/z ratio arising ﬁ'om a
single laser shot are measured; they simply arrive at the ion detector at different times.
However, the tof recorded for a particular ion reﬂects many different initial conditions
experienced in the ion source (such as the time/location of ion formation and initial
kinetic energy distributions). The poor resolving power is reﬂected by peak broadening
which increases with the mass of the ionized macromolecules. This limits the capacity to
detect certain protein modiﬁcations and protein sequence variations, especially at high

mass. In spite of this, it is possible to achieve a mass accuracy of 0.1-0.01% for proteins

38

with molecular masses between 1 and 40 KDa, and with somewhat poorer accuracy for
proteins above 40 KDa (153).

Two techniques were introduced to solve the low-resolution problem of linear TOF
mass analyzer, reﬂectron-TOF (reTOF) (154) and delayed ion extraction (DE) (155). The
reﬂectron-TOF consists of two linear ﬁeld-ﬁee region and an ion mirror, which
compensates for different, nonzero initial kinetic energies of ions of the same m/z.
Delayed ion extraction allows the ions to disperse in the source region (acceleration
region just above the surface) owing to their initial velocity, while the density of neutrals
is decreased by pumping them away. The number of ion-molecular collisions before
extraction of the ions into the ﬁeld-ﬂee drift region is therefore reduced, which decreases
the width of the translational energy distribution. Simultaneously, ions having higher
initial kinetic energy move further away from the extraction ﬁeld in the source region and
thus are given less “kick”; those of lower kinetic energy experience a higher extraction
ﬁeld. This approach has proven so successful that limitations in mass resolution are now

focused on detection systems (the time resolution).

B. Electrospray Ionization Mass Spectrometry (ESI-MS)

Electrospray ionization-mass spectrometry (ESI-MS) has its origins in research that
long preceded the current ﬂurry of activity. The study of the electrospray phenomena
extends back perhaps over two and one-half centuries to the work of Bose (156), and
certainly to that of Zeleny early in this century (157). The seminal research into the use of
electrospray as an ionization method for macromolecules was due to Malcom Dole and

co-workers (158, 159), who performed extensive studies into the electrospray process and

39

deﬁned many of the experimental parameters. Experimental evidence was presented by
Dole for ionization of zein (Mr. ~50,000) (160) and lysozyme (161). However,
interpretation of their results was problematic because a mass spectrometer was not
available, and only ion retardation and ion mobility measurements were obtained.

In 1984, ESI combined with mass spectrometry was ﬁrst reported, essentially
simultaneously, by both Yarnashita and Fenn (162) and Aleksandrov et al. (163, 164). In
1988, Fenn and co-workers (165) ﬁrst reported ESI-MS spectra of intact multiply
protonated molecules of proteins up to Mr 40,000. Since its introduction in 1988 as a tool
for the ionization of large biomolecules, the practice of ESI-MS in medical,
biotechnological, and pharmaceutical arenas has grown impressively. One obvious reason
for this is the speed with which commercial instrumentation became available, and
particularly the ease of retroﬁting existing quadrupole mass spectrometers for ESI-MS.
As evidenced by the growing number of ESI-MS related publications, routine analysis by
ESI-MS is becoming increasely common in basic research, quality control, and

bioanalytical laboratories.

l. The Electrospray Process

Electrospray ionization requires continuous ﬂow of solution (generally from a small
diameter capillary tube) in the presence of a high electric ﬁeld. Liquid is expelled ﬁ'om
the tube as charged droplets which have a charge-to-volume ratio that can approach the
physical maximum allowed for droplet stability, i. e., the Rayleigh limit (166), which by
deﬁnition is reached when droplet surface charge density exceeds the liquid’s surface

tension (167). Droplet instability is manifested in a “coulomb explosion” or “droplet

40

ﬁssion” event. It is generally believed solvent is lost through evaporation until Coulomb
explosions produce smaller droplets. The small droplets evaporate rapidly and the
protonated molecules are released into the gas phase. It is unclear whether ions escape
from droplets (i. e., ﬁeld ionization) or solvent evaporates to leave ions (1. e., droplet
evaporation), and the two mechanisms will likely be experimentally undistinguishable. In
contrast to all other mass spectrometric ionization methods, the change from the liquid
into gas phase is very gentle and probably occurs with at least one solvation shell still
surrounding the analyte molecule. Fragmentation during this ‘desorption event’ has
hardly been observed, and it is now thought that even aspects of the tertiary structure of a

protein survive into the gas phase, at least under certain circumstances (168).

2. Instrumentation

The generic representation of an ESI source shown in Figure 1.3 incorporates
several features that are found on many ESI sources. The electrospray capillary typically
consists of a small i.d. (<100 um) metal or glass capillary biased at :tl-S KV (positive
potential for cations, negative potential for anions) relative to the desolvating capillary.
Recent studies suggest that signiﬁcant improvements in sensitivity can be realized with
very small diameter (<20 um) electrospray capillaries and the use of very low ﬂow rates
(169). A solution containing the analyte of interest is forced into the electrospray
capillary at a ﬂow rate of 0.01-10 til/min (depending on analyte concentration and
solution conditions) and is nebulized by the resulting electrospray plume. A common
solvent for the analysis of peptide cations is 50/50 MeOH/Hzo (or acetonitrile/H20) with

1-2% HOAc (or formic acid) and a protein/peptide concentration of 10-5 — 10'7 M.

41

Mass Analyzer

Electrospray

 

 

 

 

Plume Skimmer
. Cones
Electrospray , -"
Capillary ‘ —// ”I
‘ D—esolvating\\
Capillary
+High Voltage ion

Optics

__,. ——>[M+nH]n+
_

 

 

 

 

 

 

 

 

 

 

 

 

solvent droplet formation of
evaporation ﬁssion at desolvated Ions
Rayleigh by further ﬁeld
limit desorption and/or
ion evaporation

 

 

 

Figure 1.3. ESI-MS instrument and droplet production in ESI interface

42

Electrospray devices also are used to couple separation techniques, including
capillary electrophoresis (170) and high performance liquid chromatography (171) to
mass spectrometry. Direct coupling of the separation technique and the ionization
technique is accomplished in a number of ways including the addition of a conductive
sheath liquid or the application of an electrically conductive coating to the terminus of the
separation capillary (171).

It has been recognized that the electrospray ionization process is driven by droplet
desolvation and that desolvation at the molecular level is essential to obtain high quality
spectra. Thus, a desolvation capillary often is added to expedite desolvation of the highly
charged droplets generated in the electrospray plume. This capillary typically has an
internal diameter of 400 pm to 1 mm, and a length of 10-25 cm. Glass inlet capillaries are
employed widely in conjunction with a countercurrent ﬂow of heated, dry nitrogen to
enhance the rate of droplet desolvation and prevent large amounts of material from
entering the spectrometer. Both ends of the glass capillary are gold coated and electrically
biased to deﬁne a ﬁnite potential difference between the capillary ends and the
neighboring source elements.

Next, a skimmer cone, or more commonly, a series of skimmer cones with a 300nm
to 1.5mm diameter aperture is used to transmit to the mass spectrometer as many ions
ﬁom the sample as possible while decreasing the pressure in each subsequent vacuum
stage of the chamber. An electrical bias on the skimmer cone(s) serves to focus the [ion
beam and deﬁnes the average kinetic energy of the ion beam. Ion transmission from the
source to the mass analyzer is enhanced greatly by the addition of a few simple ion lenses

that guide and focus the ion beam into the mass analyzer.

43

In conjunction with electrospray ionization a quadrupole mass spectrometer is
commonly used as a result of relatively low cost and ease of interfacing with many
commercial systems. A quadrupole mass analyzer consists of four equidistant rod
electrodes arranged as two pairs to which time-varying ac and dc potentials are applied.
Only ions having a mass-to-charge ratio within a very narrow m/z range are transmitted
by the quadrupole. Ions having a m/z outside this range move in unstable orbits and
collide with the electrodes. Ions striking the detector transfer charge that registers as a
current that is proportional to ion abundance and is ampliﬁed subsequently. Correlating
ion abundance values with the values of applied dc bias and amplitude of ac voltage gives
a mass spectrum. The relatively high sensitivity and rapid scan rates obtainable with

commercial quadrupole mass spectrometers resulted in their widespread use.

3. Interpretation of ESI-MS Spectra

As a majority of ionization techniques today produce single charged ions (and, thus,
mass spectra in which the measured m/z is equivalent to the mass of analyte), the
multiple-charge states common to ESI-MS spectra might seem a bit confusing. However,
their interpretation often is straightforward once the appropriate procedure is applied.

As ﬁrst demonstrated by Mann, Meng, and Fenn (172), a simple algorithm can be
applied to single component spectra based on the assumption that each peak differs from
its neighboring peak by only charge. The value of the charge state of a selected lower m/z
peak (210,.) is determined according to the equation: Ztow = [(m/z)hi]/ {[(m/z)hi-(m/z)tow]/n} ,
where n is the number of peaks between the selected lower m/z peak [(m/z)tow] and a

selected higher m/z peak [(m/z)m] being considered. From the known charge state of a

single peak calculated by the equation, the charge states of all the remaining peaks are
directly calculated by simple addition or subtraction. From measured m/z and the
calculated charge 2, the ionic mass, 111,, is Obtained easily, m1 = m/z x z. The mass of the
corresponding neutral analyte, Mr: Mr = [(m/z) x z] i nMa, Ma is the mass of charged
adduct. As each charge state yields an independent mass measurement, mass accuracy
often is improved considerably by averaging over the entire charge envelope: Mr = (l/N)
am, where Mr is the calculated molecular weight from a single charge state and N is

the number of charge states being averaged.

C. The Comparison of MALDI and ESI

While MALDI and ESI are widely used techniques in analysis of large
biomolecules, they have different and complementary natures (Table 1.1), which has
made it increasingly important that researchers have access to both types of
instrumentation. Independently, E81 and MALDI-MS can help answer many questions;
yet together they present a formidable research tool with new levels of sensitivity,
accuracy, and mass range. Their utility for mass measurement meets the needs of
chemists and biologists alike, facilitating routine characterization in small molecule
synthesis, protein synthesis, and compounds obtained directly from biological matrices.
The use of ESI and MALDI mass spectrometry extends beyond simple characterization.
Noncovalent interactions, protein and peptide sequencing, DNA sequencing, protein
folding, in vitra drug analysis, and drug discovery are among the areas to which ESI and

MALDI-MS have been applied.

45

Table 1.1. Comparison of ESI-MS and MALDI-MS

 

 

ESI-MS MALDI-MS
Mass limit, Da ~200,000 >3.000.000
(practical) (70,000) (1 50,000)
HPLC/MS capable Tolerant of HIM
Advantages Multiple charging concentrations of salts
Capable of observing Highest mass capability
noncovalentcomplexes Tolerant of mixtures
directly from water Femtomole sensitivity
Femtomole-to-picomole Being developed as a 100'
sensitivity for sequence analysis
Disadvantages Multiple charging can be Typically low resolution
confusing with mixtures (4500)

Typically need < "M salt linear TOF without DE
concentrations for 90°d Not amenable to LC/MS
signals

Not tolerant of mixtures

Suitable Peptides Peptides
compounds Proteins Proteins

Carbohydrates Glycoproteins

Nucleotides Carbohydrates

Nucleotides

Oligonucleotides
Phosphoproteins
Small chargeable molecules

Oligonucleotides
Phosphoproteins
Small chargeable
molecules
Heterogeneous
samples

 

46

VI. References

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

2. Fenn, J. B., Mann, M., Meng, C. K., Wong, S. F., and Whitehouse, C. M., Science,
246, 64 (1989).

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

4. Weissman, J. S., Kim, P. S., Science, 253, 1386 (1991).

5. Pain, R. H., Trends Biochem Sci, 12, 309 (1987).

6. Scheraga, H. A., Konishi, Y., 001, T., Adv. Biophys, 18, 21 ( 1984).

7. Kim, P. S., and Baldwin, R. L., Annu. Rev. Biochem, 59, 631 (1990).

8. Matthews, C. R., Annu. Rev. Biochem, 62, 653 (1993).

9. Ptitsyn, O. B., Curr. Opin. Struct. Biol, 5, 74 (1995).

10. Creighton, T. B., J. Mol. Biol., 87, 563 (1974).

11. Creighton, T. B., Goldenberg, D. P., J. Mol. Biol, 179, 497 (1984).

12. Scheraga, H. A., Konishi, Y., 001, T., Adv. Biophys, 18, 21 (1984).

13. Jacobson, G. R., Schaffer, M. H., Stark, G. R., and vanaman, T. C., J. Biol. Chem,
248, 6583 (1973).

14. Wu, J., Gage, D., and Watson, J. T., Anal. Biochem, 235, 161 (1996).

15. Wu, J ., and Watson, J. T., Protein Sci. 6, 391 (1997).

16. Roder, H., Elove, G., and Englander, S. W., Nature, 335, 700 (1988).

17. Balbach, J., Forge, V., Van Nuland, N. A. J., Winder, S. L., Hore, P. J ., and Dobson,
C. M., Nat. Struct. Biol., 2, 865 (1995).

18. Udgaonkar, J. B., and Baldwin, R. L., Nature, 335, 694 (1988).

19. Jennings, P. A., and Wright, P. B., Science, 262, 892 (1993).

47

20. Freedman, R. B., Trends Biochem. Sci., 9, 438 (1984).

21. Creighton, T. E., Biochem. J., 270, l (1990).

22. Haber, E., and Anﬁnsen, C. B., J. Biol. Chem, 237, 1839 (1962).

23. Creighton, T. E., Methods Enzymal., 131, 83 (1986).

24. Gilbert, H. F ., Adv. Enzymal., 63, 69 (1990).

25. Honk, J ., and Whitesides, G. M., J. Am. Chem. Sac., 109, 6825 (1987).

26. Wilson, J. M., Bayer, R. J ., and Hupe, D. J ., J. Am. Chem. Sac., 99, 7922 (1977).

27. Szajewski, R. P., and Whitesides, G. M., J. Am. Chem. Sac., 102, 2011 (1980).

28. Goldenberg, D., Trends Biochem Sci., 17, 992 (1992).

29. Barron, E. S. G., Adv. Enzymal., 11, 201 (1951).

30. Larsson, A., Orrenius, S., Holrngren, A., and Mannervik, B., in Functions of
Glutathione: Biochemical, Physiological, Toxicological, and Clinical Aspects, Raven,
New, York, 1983.

31. Iyer, K. S., and Klee, W. A., J. Biol. Chem, 248, 707 (1973).

32. Creighton, T. B., Methods in Enzymolagy, 107, 305 (1984).

33. Misra, H. P., J. Biol. Chem. 249, 2151 (1974).

34. Anﬁnsen, C. B., Haber, B., Sela, M., and White, F. H., Proc. Natl. Acad. Sci. USA,
47, 1309 (1961).

35. Creighton, T. B., Prog. Biophys. Mol. Biol., 33, 231 (1978).

36. Goldberger, R. F., Epstein, C. J ., and Anﬁnsen, C. B., J. Biol. Chem, 238, 628
(1963).

37. Venetianer,.P., and Straub, F. B., Biochim. Biophys. Acta, 67, 166 (1963).

38. Bergman, L. W., and Kuehl, W. M., J. Biol. Chem, 254, 8869 (1979).

48

39. Darby, N. J., Norin, P. B., Talbo, G., and Creighton, T. B., J. Mol. Biol., 249, 463
(1995)

40. Weissman, J. S., and Kim, P. S., Proc. Natl. Acad. Sci. USA, 89, 9900 (1992).

41. Gilbert, H. F., in Mechanisms of Protein Folding, Pain, R. H. (Editor), Oxford Press,
Oxford, pl 12 (1994)..

42. Chang, J. Y., Biochemistry, 35, 11702 (1996).

43. Chatrenet, B., and Chang, J. Y., J. Biol. Chem, 268, 20988 (1993).

44. Chang, J. Y., J. Biol. Chem, 270, 25661 (1995).

45. Chang, I. Y., Canals, F., Schindler, P., Querol, B., and Aviles, F. X., J. Biol. Chem,
269,22087(1994)

46. Hua, Q. X., Narhi, L., Jia, W., Arakawa, T., Rosenfeld, R., Hawkins, N., Miller, J. A.,
and Weiss, M. A., J. Mol. Biol., 259, 297 (1996).

47. Chatrenet, B., and Chang, J. Y., J. Biol. Chem, 268, 20988 (1993).

48. Wilkins, J. A., Cone, J ., Randhawa, 2., Wood, D., Warren, M. K., and Witkowska, H.
13., Protein Science, 2, 244 (1993).

49. Hober, S., Forsberg, G., Palm, 6., Hartmanis, M., and Nilsson, B., Biochemistry, 31,
1749 (1992).

50. Glocker, M. 0., Arbogast, B., Milley, R., Cowgill, C., and Deinzer, M. L., Proc. Natl.
Acad. Sci. USA, 91, 5868 (1994).

51. Creighton, T. B., Proc. Natl. Acad. Sci., USA, 85, 5082 (1988).

52. Rothwarf, D. M., Scheraga, H. A., J. Am Chem. Sac., 113, 6293 (1991).

53. Konishi, Y., Ooi, T., and Scheraga, H. A., Biochemistry, 21, 4734 (1992).

54. Goto, Y., and hamaguchi, J. Mol. Biol., 146, 321 (1981).

49

55. Torella, C., Ruoppolo, M., Marino, G., Pueci, P., FEBS Letters, 352, 301 (1994).

56. Rothwarf, D. M., Scheraga, H. A., J. Am. Chem. Soc., 113, 6293 (1991).

57. Rothwarf, D. M., and Scheraga, H. A., Biochemistry, 32, 2671 (1993).

58. Bruice, T., and Kenyon, G. L., J. Protein Chem, 1, 47 (1982).

59. Roberts, D. D., Lewis, S. D., Ballou, D. P., Olseon, S. T., and Shafer, J. A.,
Biochemistry, 25, 5595 (1986).

60. Thannhauser, T. W., Sherwood, R. W., and Scheraga, H. A., J. Protein Chem. 17, 37
(1998)

61. Happersberger, P., Cowgill, C., Deinzer, M. L., and Glocker, M. 0., Proceedings of
the 44'" ASMS Conference on Mass Spectrometry and Allied Topics, Portland,
Oregon, May, 1996, p1050.

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

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

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

65. Andrews P. C., and Dixon, J. B., Anal. Biochem, 161, 524 (1987).

66. Kruﬂ, V., Kapp U., and Wittmann-Liebold, B., Anal. Biochem, 193, 306 (1991).

67. Lee, Y.,Kinoshita, H., Radke, G., Weilder, S., Barranger, J. A., and Tomich, J. M., J.
Protein Chem, 14, 127 (1995).

68. Krieglstein, K. G., DasGupta, B. R., and henschen, A. H., J. Protein Chem, 13, 49
(1994).

69. Inc, R. A., and Hale, J. B., Anal. Biochem, 221, 374 (1994).

70. Krutzsch, H., and Inman, J. K., Anal. Biochem, 209, 109 (1993).

71. Brune, D., Anal. Biochem, 207, 285 (1992).

50

72. Inc, R. A., and Hale, J. E., Anal. Biochem, 210, 39 (1993).

73. Hale, J. B., Beidler, D. B., and Jue, R. A., Anal. Biochem, 216, 61 (1994).

74. Seiﬁ'ied, S. B., Wang, Y., and Von Hippel, P. H., J. Biol. Chem, 263, 13511 (1988).

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

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

77. First, E. A., and Taylor, S. S., Biochemistry, 28, 3598 (1989).

78. Bishop, J. B., Squier, T. C., Bigelow, D. J., and Inesi, 6., Biochemistry, 27, 5233
(1988)

79. Sturrock, E. D., Yu, X. C., Wu, Z., Biemann, K., and Riordan, J. F., Biochemistry, 35,
9560 (1996).

80. Toyooka, T., and Imai, K., Anal. Chem, 56, 2461 (1984).

81. Kirley, T., J. Biol. Chem, 264, 7185 (1989).

82. Kirley, T., Anal. Biochem, 180, 231 (1989).

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

84. Jason, J ., and Ryden, L., Protein Puriﬁcation: Principles, High Resolution methods,
and Applications, VCH Press, (1993), p252.

85. Kalef, B., Walﬁsh, P. G., and Gitler, C., Anal. Biochem, 212, 325 (1993).

86. Zaluzec, E. J ., Gage, D. A., and Watson, J. T., in Proceeding of the 42th ASMS
Conference on Mass Spectrometry and Allied Topics, Chicago, (1994), p20.

87. Catsimpoolas, N., and Wood, J. L., J. Biol. Chem, 241, 1790 (1966).

88. Jacobson, G. R., Schaffer, M, H., Stark, G. R., and Venaman, T. C., J. Biol. Chem,
248, 6583, (1973).

89. Stark, G. R., Methods in Enzymology, vol. 47, 129 (1977).

51

90. Wakselman, M., and Guibe-Jampel, B., J. C. S. Chem. Comm, 21 (1976).
91. Mikami, B., and Morita, Y., J. Biochem., 93, 777 (1983).
92. Nika, H., and Hultin, T., Eur. J. Biochem. 142, 521 (1984).
93. Minard, B., Desmadril, M., Ballery, N., Perahia, D., Mouawad, L., Hall, L., and Yon,
J. M., J. Biochem. 185, 419 (1989).
94. Darbre, A., Practical protein Chemistry—A Handbook, John Wiley & Sons, (1985),
p149.
95. Daniel, R, Caminade, B., Martel, A., Goﬂic, F ., canosa, D., Carrascal, M., and
Abian, J ., J. Biol. Chem, 272, 26934 (1997).
96. Wu, J., Gage, D. A., and Watson, J. T., Anal. Biochem., 235, 161 (1996).
97. Ryle, A. P., and Sanger, F., Biochem. J., 60, 535 (1955).
98. Biemann, K., in Methods in Protein Sequence Analysis, Plenum Press, New York,
(1993), p119.
99. Staswick, P. B., Hermodson, M. A., and Nielsen, N. C., J. Biol. Chem, 259, 13431
(1984).
100. Haeﬁher—Gormlcy, L., Parente, L., and Wetlaufer, D. B., Int. J. Pept. Protein
Res., 26, 83 (1985).
101.Yazdanparast, R., Andrews, P. C., Smith, D. L., and Dixon, J. E., J. Biol. Chem,
262, 2507 (1987).
102. Sun, Y., Bauer, M. D., Keough, T. W., and Lacey, M. P., Methods in Molecular
Biology, Vol. 6] : Protein and peptide Analysis by Mass Spectrometry, edited by
Chapman, J. R., Humana press Inc., Totowa, NJ, (1996).

103. Smith, D. L., and Zhou, Z., Methods in Ezymol., Vol. 193, 374 (1990).

52

104. Sun, Y., Zhou, Z., and Smith, D. L., in Techniques in Protein Chemistry, (hugli, T.
B., ed.), Academic, New, York, p176 (1989).

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

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

107. Thannhauser, T. W., McWherter, C. A., and Scheraga, H. A., Anal. Biochem., 149,
322 (1985).

108. Morris, H. R. and Pucci, P., Biochem. Biophys. Res. Commun., 126, 1122, 1985.

3109. Takao, T., Yoshida, M., Hong, Y., Aimoto, S., and Shimonishi, Y., Biomed. Mass
Spectrom., 11, 549, 1984.

110. Yazdanparast, R., Andrews, P. C., Smith, D. L., and Dixon, J. B., J. Biol. Chem,
262, 2507, 1987.

111. Stults, J. T., Bourell, J. H., Canova—Davis, B., Ling, V. T., Laramee, G. R., Winslow,
J. W., Grifﬁn, P. R., Rinderknecht, B., and Vandlen, R. L., Biochem. Environ. Mass
Spectrom., 11, 549-556, 1984.

112. Vaccaro, A. M., Salvioli, R., Barca, A., Tatti, M., and Pucci P., J. Biol. Chem, 270,
9953-9960, 1995.

113. Yoshida, S., Takamatsu, T., Kanda, F., Rapid commun. Mass Spectrom. 10, 641-

648, 1996.

114. Beam, M. F., and Carr, S. A. Anal. Biochem. 201, 216,1992.

115. Sun, Y., and Smith, D. L. Anal. Biochem. 172, 130, 1988.

116. Hunt, D.F., Yates, J. R., Shabanowitz, J., Winston, S., and Hauser, C. R., Proc. Natl.

Acad. Sci. USA, 83,6233, 1986.

53

117. Bauer, M., Sun, Y., Degenhardt, C., and Kozikoski, B., J. Protein Chem. 12,
759,1993.

118.Krishnamurthy, T., and Hauer, C. C., Biol. Mass spetrom 23, 719, 1994.

119. Bayer, M., Sun, Y., Degenhardt, C., and Kozikowsiki, B., J. Protein Chem. 12, 759,
1993.

120. Stewart, A. B., Rafﬁoni, S., and Chandhery, T., Protein Science, 1, 777, 1992.

121. Patterson, S., and Katta, V., Anal. Chem. 66, 3727, 1994.

122. Jones, M. D., Patterson, S. D., and Lu, H. S., Anal. Chem. 70, 136 (1998).

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

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

125. Serensen, H. H., Thomsen, J ., and Bayne, S., Biomed. Environ. Mass Spectrom. 19,
713 (1990).

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

127. Zhang, D., and Liang, S., J. Protein Chem, 12, 735 (1993).

128. Krishnamurthy, T., Hauer, C. R., Prabhakaran, M., Freedy, J. G., and Hayashi, K.,
Biol. Mass Spectrom., 23, 719 (1994).

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

130.Gray, W. R., Protein Science, 2, 1749 (1993).

131. Haniu, M., Hsieh, P., Rohde, M. F ., and Kenney, W. C., J. Biochem. Biophys., 310,

433 (1994).

54

132. Yamashita, H., Nakatsuka, T., and Hirose, M., J. Biol. Chem, 270, 29806 (1995).

133. Fairlie, W. D., Stanton, P. G., and Hearn, M. T. W., Biochem. J., 314, 449 (1996).

134. Gray, W. R., Luque, F. A., Galyean, R., Atherton, B., Sheppard, R. C., Stone, B. L.,
Reyes, A., Alford, J., Mcintosh, M., Olivera, B. M., Cruz, L. J ., and rivier, J.,
Biochemistry, 23, 2796 (1984).

135. Gray, W. R, Protein Sci., 2, 1732 (1993).

136. Wu, J ., and Watson, J. T., Protein Science, 6, 391 (1997).

137. Thorgersen, D. F., Skowronski, R P., and Macfarlane, R. D., Biochem. Biophys.
Res. Commun., 60, 616 (1974).

138. Barber, M., Bordoli, R. S., Sedgwick, R. D., and Tyler, A. N., Chem. Commun., 325
(1981).

139. Meng, C. K., Mann, M., and F enn, J. B., Proc. 36'” ASMS Conference on Mass
Spectrometry and Allied Topics, San Francisco, CA, (1988), p771.

140. Posthumus, M. A., Kistcmaker, P. G., Meuzelaar, H. L. C., deBrauw, M. C., Anal.
Chem, 50, 985 (1978).

141. Karas, M., Bachmann, D. B., Hillenkamp, F., Int. J. Mass Spectrom. Ion. Proc., 78,
53 (1987).

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

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

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

145. Beavis, R. C., and Chait, B. T., Rapid Commun. Mass Spectrome., 3, 233 (1989).

55

146. Juhasz, P., Costello, C. B., and Biemann, K., J. Am. Soc. Mass Spectrom., 4, 399
(1993).

147. Vorm, 0., Roepstorff, P., and Mann, M., Anal. Chem, 66, 3281 (1994).

148. Xiang, F., and Beavis, R. C., Org. Mass Spectrom., 28, 1424 (1993).

149. Xiang, F., and Beavis, R. C., Rapid Commun. Mass Spectrom., 8, 199 (1994).

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

151. Preston, L. M., Murray, K. K., and Russsell, D. H., Biol. Mass Spectrom., 22,
544(1993).

152.Vorm, O., and mann, M., J. Am. Soc. Mass Spectrom., 5, 955 (1994).

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

154. Mamyrin, B. A., Karateev, V. I., Shmikk, D. V., Zagulin, V. A., Sov. Phys. JETP,
37, 45 (1973).

155. Anderegg, R. J ., Wagner, D. S., Stevenson, C. J ., Borchardt, R. T., J. Am. Soc. Mass
Spectrom., 5, 425 ( 1994).

156. Bose, G. M., Recherches sur le cause et sur la veritable theorie de l'electricite,
Wittenberg, 1745..

157. Zeleny, J., J. Phys. Rev., 10, 1 (1917).

158. Dole, M., Mack, L. L., Hines, R. L., Mobley, R. C., Ferguson, L. D., Alice, M. B., J.
Chem. Phys. 49, 2240 (1968).

159. Mack, L. L., Kralik, P., Rheude, A., Dole, M., J. Chem. Phys., 52, 4977 (1970).

160. Clegg, G. A., Dole, M., Biopolymers, 10, 821 (1971).

161. Gieniec, J ., Mack, L. L., Nakama, K., Gupta, C., Kumar, V., Dole, M., Biomed.

Mass Spectrom., 11, 259 (1984).

56

162. Yamashita, M., Fenn, J. B., J. Phys. Chem, 88, 4451 (1984).

163. Aleksandrov, M. L., Gall, L. N., Krasnov, V. N., Nikolaev, V. I., Pavlenko, V. A.,
Shkurov, V. A., Dokl., Akad. Nauk. SSSR, 277, 379 (1984).

164. Aleksandrov, M. L., Gall, L. N., Krasnov, V. N., Nikolaev, V. I., Pavlenko, V. A.,
Shkurov, V. A., Baram, G. I., Grancher, M. A., Knorre, V. D., Kusner, Y. S., Bioorg.
Kim, 10, 710 (1984).

165. Meng, C. K., Mann, M., Fenn, J. B. Z., Phys. D., 10, 361 (1988).

166. Gomez, A., Tang, K., Phys. Fluids, 6, 404 (1994).

167. Rayleigh, J. W., Philos. Mag, 14, 182 (1882).

168. Siuzdak, G., Proc. Natl. Acad. Sci. USA, 91, 11290 (1994).

169. Gale, D. C., Smith, R. D., Rapid Comm. Mass Spectrom., 7, 1017 (1993).

170. Smith, R. D., Wahl, J. H., Goodlett, D. R., Hofstadler, S. A., Anal. Chem, 65, A574
(1993).

171. Grifﬁn, P. R., Cofﬁnan, J. A., Hood, L. B., Yates, J. R, Int. J. Mass Spectrom. Ion.
Proc., 111, 131 (1991).

172. Mann, M., Meng, C. K., Fenn, J. B., Anal. Chem, 61, 1702 (1989).

57

CHAPTER 2
A COMPARISON OF MALDI AND ESI FOR ANALYSIS OF
IMINOTHIAZOLIDINE-BLOCKED PEPTIDES

IN DISULFIDE MAPPING

I. Introduction

Cleavage at cysteine residues of peptide chains under alkaline conditions after
cyanolysis of disulﬁde bonds was ﬁrst observed by Catsimpoolas and Wood (1).
However, because of the side reactions, such as the reversibility of the cyanolysis and the
elimination of thiocyanylate, the yields were low and the cleavage reaction was not
acceptable to protein chemists. Jacobson et al. (2) showed that 2-nitro-5-
thiocyanobenzoic acid (NT CB) speciﬁcally cyanylates cysteine thiols under alkaline
conditions. Subsequent cleavage occurs on the N-terminal side of the cyanylated
cysteinyl residue under mildly alkaline conditions to form an amino-terminal peptide and
a series of 2-iminothioazolidine (itz)-4—carboxyl peptides. Although the original paper
claimed that the cleavage reaction can come to completion with only minor side reactions
for most of the peptides and proteins tested, Digani and Patchornik (3) ‘ found B-
elimination, occurring also under alkaline conditions, competes with the cleavage as an
adverse reaction. Experimentally, a large excess of NTCB and a low total concentration
of thiol groups must be applied to avoid the side reaction of displacement of CN- from S-
cyanocysteine residues by the unreacted thiol groups (3).

Wu & Watson (4) conducted a systematic study to elucidate the effects of peptide

structure and reaction conditions on the kinetics of the cleavage reactions of cyanylated

58

peptides and proteins and the yields of cleavage products. The optimal results were
obtained in 1M ammonium hydroxide solution in which cleavage is complete within an
hour at ambient temperature (4). Ammonia is a stronger small nucleophile but has smaller
proton afﬁnity than hydroxyl ion so that nucleophilic attack of ammonia on the carbonyl
carbon facilitates the cleavage process giving an a-amidated N-terminal peptide but low
capacity for catalyzing the B-elirnination reaction. Additionally, the excess volatile
ammonia can easily be removed from the mixture alter cleavage, an advantage for
subsequent analysis of the products by MALDI or ESI-MS (4).

Other reagents for cyanylation of SH groups have been suggested (5-8). Among
them l-cyano-4—dimethylamino-pyridinium (CDAP) tetraﬂuoroborate (8) has the
advantage of the minimization of thiol/disulﬁde exchange because of the acidic condition
of cyanylation reaction of sulfhydryl groups.

Cyanylation and cleavage reactions of the cysteine residue have the capacity be
used for its structural analysis. If a protein contains 11 cysteine residues, the cleavage
reaction will result in the formation of n+1 peptide fragments; mass analysis of the
fragments indicates the number and location of cysteine residues. Sodium dodecyl
sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) has been used to mass map the
cleavage products (9-28). However, peptide assignments using this approach are often
complicated by its poor mass accuracy (error>5%). Papayannopoulos and Biemann (29)
used CID tandem mass spectrometry to sequence the NTCB cleavage reaction products
of a protease inhibitor isolated ﬁ'om Sarcophaga bullata. Their work demonstrated that
mass spectrometry could be used to sequence peptides from the NTCB cleavage reaction

in spite of the blocked N-terminus. However, tandem MS is practically limited to low-

59

mass peptides produced from cysteine-rich proteins, because of the effective mass limit
of CID (<3000 Da).

Recently, Wu et al. (30) developed a methodology to recognize the number and
location of both cystines and free sulﬂiydryls in peptides and proteins using the speciﬁc
chemical cleavage reaction described above, followed by mass mapping of the resulting
peptides by MALDI-MS. This approach provides the advantages of fast analysis, easy
operation, high mass accuracy, and high sensitivity.

Both electrospray ionization (ESI) and matrix-assisted laser desorption/ionization
(MALDI) mass spectrometry allow for the analysis of large biomolecules through “mild”
desorption and ionization methods (31, 32), each having unique capabilities, as well as
some fundamental similarities. They are truly complementary in nature, almost a mirror
image being projected between the disadvantages of one technique and the advantages of
the other (33). During our original work, analyses of resulting mixtures of
iminothiazolidine-blocked peptides from the cleavage of cyanylated protein by MALDI
showed ﬁ'equent, and sometimes complete, suppression of a signal for some fragments.
Since such a phenomenon can compromise the structural analysis of cysteine residues by
the methodology based on the cyanaylation chemistry, an alternative mass analysis
technique should be used. ESI-MS will be a reasonable choice because of its capability of
analysis of large molecules and its complementary nature for MALDI.'

In this chapter, we investigated the use of ESI for the analysis of iminothiazolidine-
blocked peptides when MALDI had the signal suppression. We used off-line HPLC to
separate the mixture of iminothiazolidine-blocked peptides and then analyzed the HPLC

fractions by ESI-MS. Two proteins, B-lactoglobulin A and ovalbumin, were used as

60

model proteins to investigate the complementary nature of MALDI and ESI in separate
experiments. The results indicate that the responses of blocked peptides to ESI and

MALDI are frequently complementary.

II. Analytical Methodology

A. Cyanylation and Chemical Cleavage Reaction

Cyanylation of cysteine residue employs a one-step selective chemical reaction
between free sulﬂiydryl and NTCB (3, 30, 34) or CDAP (8). Aﬁer selectively cyanylating
the sulfhydryls, the cysteinyl peptide bonds can be cleaved under alkaline condition to
form an amino-terminal peptide and a series of iminothiazolidine-blocked peptides. The
mechanism of cyanylation and cleavage reactions is illustrated in Figure 2.1.

The side reaction is that the cyanylated cysteine undergoes a base-catalyzed B-
elimination to form thiocyanate and dehydroalanine. This side reaction results in lower
yield of cleavage.

The cyanylation of sulfhydryl groups was traditionally accomplished by 2-nitro-5-
thiocyanobenzoic acid (NT CB) under mildly alkaline conditions (pH 8-10). The reagent
was believed to be speciﬁc to sulﬂiydryl groups, although side reactions, such as the
formation of mixed disulﬁde bonds between the NTCB and protein SH groups, were also
reported (3, 35, 36). Another reagent, 1-cyano-4—dimethylamino-pyridinium (CDAP)
tetraﬂuoroborate, was proposed for the cyanylation of SH groups under slightly acidic
conditions (pH 3-7) (5). CDAP is advantageous over NTCB for speciﬁc cyanylation of
protein SH groups in the presence of disulﬁde bonds, because the acidic condition can

effectively minimize sulfhydryl/disulﬁde exchange. Furthermore, the cyanylated proteins

61

SH
O CH2

— — -NH-CH-CO — -

cyanylation CDAP MezH/_ >1: CN
pH3-7; rt, 10'

N=C-§
ii ‘iH2
- -NH-CH-CO - —

.NH3KO
cleavage
1’“ M44OH B elimination
~1h, It
7°
NH2 NH loo- - - -CO- NH-CI-CO- -
(itz-peptide) (side reaction)

Figure 2.1. Mechanisms of cyanylation and cleavage reactions

62

or peptides are stable under acidic conditions. While complete cyanylation by CDAP can
be carried out using three to ﬁve-fold molar excess of the reagent over ﬁ'ee SH groups
under very mild conditions at room temperature (5, 37, 38), the optimal condition should
be 10 to 20-fold molar excess of CDAP (over sulthydyl groups) in pH 3.0 buffer for 15
min at ambient temperature (4). A large excess of CDAP (~50 fold over SH group) and
excessive incubation time (>2 hours) could result in the formation of an unidentiﬁed side
reaction product (4). Therefore, like cyanylation by NTCB, the cyanylation by CDAP
needs to be performed under controlled conditions to minimize the side reaction.

The another characteristic of the reaction is that the resulting peptide ﬁagments may
vary greatly in size because of the widely variable frequency of occurrence of cysteine in
proteins. Large or small sized ﬁagments may be obtained. Both MALDI and ESI-MS can
be used to measure the masses of fragments. While MALDI-MS usually has the
suppression phenomenon at low mass range, ESI-MS can be used to determine the mass.
Therefore, we can expect that the combination of MALDI-MS and ESI-MS can be

promising for cysteine analysis.

B. Localization of Cysteines and Cystines by ESI-MS

The number and location of cysteines and cystines in a protein can be determined by
a speciﬁc chemical cleavage process, which involves cyanylation of free sulthydryls by
CDAP at pH 3 or NTCB at pH 8.7, then cleavage in 1M ammonia. Two steps can be
used to locate cysteines and cystines (30), one is called NTCB or CDAP/TCEP
(cyanylation/reduction) procedure which involves ﬁrst cyanylating the free sulfhydryls in
non-reducing proteins, then cleaving the cyanylated proteins following reduction of the

remaining disulﬁde bonds. This step can recognize the location of free cysteines. Another

63

step is called TCEP/NTCB or CDAP (reduction/cyanylation) procedure, which ﬁrst
reduces all the disulﬁdes by tris-2-carboxyethyl)phosphine (TCEP), then cyanylates the
reduced protein, and then cleaves the peptide bond on the N-terminal side of cyanylated
cysteine. This step can locate total cysteines. The location of cystines can be deduced by

comparing the results of the two steps (scheme 2.1).

III. Experimental Section

ESI-MS

ESI mass spectra were obtained on a Fisons VG Platform ESI mass spectrometer.
This instrument is equipped with a single quadrupole mass analyzer. The ﬂow rate was
set at 10u1/min for infusion injection mode, and the solvent is 50:50 acetonitrile/P120
containing 1 % formic acid. The capillary voltage is set at 3.25KV, cone voltage at 37V,

and source temperature at 100°C.

MALDI-TOF MS

MALDI mass spectra were obtained on a Voyager Elite time-of-ﬂight (TOF) mass
spectrometer (PerSeptive Biosystems Inc., Framingham, MA) equipped with delayed
extraction and a model VSL-337ND nitrogen laser (Laser Science, Newton, MA). The
accelerating voltage in the ion source was set to 20 KV. Grid voltage and guide wire
voltages were 93.6% and 0.2% of the accelerating voltage, respectively. Data were
acquired in the positive linear DE mode of operation. Time-to-mass conversion was

achieved by external and/or internal calibration using standards of bradykinin (m/z

3:sz bum $5033 “.o c0280. 9: m£~Emooom .FN oEmcow

 

.8an .000
mm\ Jug
31a omm>mm_o w 71%|)
$sz 2\.\/ lea Siege \l: am:
am to 5.2.2 .92 a ..fz 2
couozcom
m0:_0«w>0 Amy
.38 £28. I 08
QTY
255.3 \lw am:
o2“. bacon. As +fz o /\
m_m>_mc< 398.0 w
.mm .o 5.25. co_am_>cm>o
.000\ ong -0001“:
mﬁlzwjomg A 2.\l| 8N:

 

\lije cowosnom 03}

..fz

65

1061.2), bovine pancreatic insulin (m/z 5734.5), and horse skeletal myoglobin (m/z
16952) obtained from Sigma Chemical Co. (St. Louis, MO). All experiments were
performed using a-cyano-4-hydroxycinnamic acid (Aldrich Chemical Co., Milwaukee,
WI) as the matrix. Two types of saturated matrix solutions were prepared. Type A
contains 70% (v/v) solution of acetonitrile/aqueous 1% TFA. Type B contains 50% (v/v)
solution of acetonitrile/aqueous 1% TFA. The protein or peptide samples were applied
with a sandwich mode to a stainless-steel sample plate: the ﬁrst layer was 0.7 ul matrix
Solution A, second layer was 0.6 ul B, the third layer was 1 ul sample, and the last layer
was 0.5 ul matrix solution A again. The sample sandwich was allowed to air dry before

being introduced into the mass spectrometer.

HPLC

HPLC is equipped with Waters model 6000 pumps controlled by a PC computer.
UV detection was at 215nm. The column was a Vydac C 18 (#218TP54, 10pm particle

size, 300- pore, 4.6x250mm).

Chemicals

Tris(2-carboxyethyl)phosphine hydrochloride (TCEP) was purchased ﬁ'om Pierce,
Rockford, IL. Guanidine hydrochloride was a product of Boehringer-Mannheim
Biochemicals (Indianapolis, IN). Proteins, 2-nitro-5-thiocyanobenzoic acid (NT CB), 1-
cyano-4-dimethylamino-pyridinium tetraﬂuoroborate (CDAP), were purchased from
Sigma and used without further puriﬁcation. Acetonitrile and TFA were HPLC grade.

The TCEP solution in 0.1 M citrate buffer at pH 3 or in 0.1 M tris-HCl buffer at pH 8.2

66

was prepared as 0.1 M stock solution and stored under N2 at —20°C for weeks with little
deterioration. The 0.01 M solution of NTCB was prepared in 0.1 M tris-HCL buffer and
the pH of the solution was adjust to pH 8.7. The 0.1 M CDAP solution in pH 3.0, 0.1 M

citrate buffer was freshly prepared prior to use.

Analysis of Free Sulfhydryl Groups in Proteins (CDAP or NTCB/TCEP Procedure)
About 10nmol protein under study is dissolved in lOpl 0.1N tris-HCl buffer (pH
8.0) for NTCB reaction or in 0.1N citrate buffer (pH 3.0) for CDAP reaction, containing
6N guanidine-HC] as denaturing reagent. To the solution, add 5-10 fold molar excess of
NTCB or 10-20 fold CDAP (over the free sulﬂiydryls). Allow the reaction to proceed at
37°C in a water bath for 30min. Adjust pH to 12 by 1N NH3, let it stay in room
temperature for 1 hour. Add 5-10 fold molar excess TCEP (over the disulﬁde bond). The
mixture is put in a water bath at 37°C for 30min to reduce the disulﬁde bond. The masses
of the cleavage fragments were analyzed by MALDI-MS. Comparatively, the mixture of
the cleavage product was separated by reversed-phase HPLC, the peaks were collected

from HPLC and analyzed by ESI-MS.

Conﬁrmation of Total Cysteines (TCEP/CDAP or NTCB Procedure)

The protein solution is allowed to react at pH 8.7 with 5-10 fold NTCB (over free
sulﬂrydryl groups) or at pH 3.0 with 10-20 fold CDAP (over free sulfhydrl groups) at
37°C for 30min after it is reduced with 5-10 fold TCEP (over disulﬁde bonds) at 37°C for
30min, and then adjust pH to 12 by 1N NH3, following reaction at room temperature for 1

hours, analyzed by MALDI-MS or off-line HPLC/ESI-MS.

67

IV. Results and Discussion

A. Recognizing the Location of Cysteine and Cystines of B-Lactoglobulin A
B—lactoglobulin A has 162 amino acids and molecular weight 18368 Da. It contains

two disulﬁde bonds and one free cysteine at Cys 121 (Figure 2.2). The expected masses

of cleavage peptide fragments aﬁer two procedures are shown in Table 2.1.

 

 

 

 

Figure 2.2. Structure of B-Lactoglobulin A

The expected peaks for the fragments 1-65(m/z 7247.4), itle6-118(m/z 1461.6),
itzl 19-120(m/z 274.3), and itzl60-162(m/z 396.5) for the procedure
reduction/cyanylation (TCEP/NT CB) are undetected in the MALDI spectrum (Figure 2.3
A) which clearly indicates the signal suppression in MALDI-MS. The abbreviation “itz”
represents “iminothiazolidine” derivative at the speciﬁed residue. Only two peaks of
itz66-105 and it2121-159 were identiﬁed by MALDI-MS. Because these two peptide
fragments indicate the cyanylation and cleavage only at Cys 66 and Cys 121, the numbers
and location of other cysteine residues couldn’t be recognized. Therefore, MALDI-MS

gave the incomplete mass mapping of cysteine residues of B-lactoglobulin A.

68

Table 2.1. Expected and Observed Masses of Cleavage Product of B-Lactoglobulin A

Reduction/Cyanylation

 

 

 

 

 

 

Fragments Expected Mass (Da) ESI(Da) MALDI(Da)
1-65 7247.4 7243.4 no
itz66-105 4651.6 4651.1 4653.5
itle6-l 18 1461.6 no no
itzl 19-120 274.3 no no
itz121-159 4551.2 4550.4 4548.1
it2160-162 396.5 397.0 no
*1-105 11822.0 11823.0 no
itz*121-162 4870.7 4870.7 no

Cyanylation/Reduction

Fragments Expected Mass(Da) ESI(Da) MALDI(Da)
1-120 13506.7 no no
it2121-162 4905.7 4904.7 4901.1

 

* B-elimination
itz = iminothiazolidine derivative.

69

Instead of the expected two peaks, one peak at m/z 4901.1 (1-120) was identiﬁed by
MALDI-MS for the cleavage product resulting from the procedure cyanylation/reduction
(NT CB/CDAP) (Figure2.3B).

Based on the above results, we used off-line HPLC separation/ESI-MS to analyze
the cleavage products. The HPLC chromatogram of the mixture of cleavage products
resulting from procedure TCEP/CDAP of B-lactoglobulin A is shown in Figure 2.4. The
ESI spectra corresponding to the HPLC peaks in Figure 2.4 are illustrated in Figure 2.5.
ESI-MS identiﬁed peptide fragments of 1-65, itz66-105, it2121-159, and itzl60-162,
indicating that the location of cysteine residues should be at Cys 66, Cys 106, Cys 121,
and Cys 160. The location of one cysteine residue at Cys 119 cannot be recognized
because of two unidentiﬁed fragment peaks of itle6-ll8 and itzll9-120, which is
probably due to the difﬁculty of cleaving the peptide bond at the N-terminal side of
cysteine residue at Cys119 because of the closeness of two cysteines at Cys 119 and Cys
121.

The HPLC chromatogram and corresponding ESI spectra of the cleavage product
resulting ﬁom procedure CDAP/TCEP of B-lactoglobulin A is shown in Figure 2.6 and
Figure 2.7, respectively. One fragment of itz121-162 was identiﬁed, indicating that the
cleavage occurred at Cys 121, thus Cys 121 involves in the pairing of the disulﬁde bond.
However, the cysteine residue paired with Cys 121 cannot be located because of the
unidentiﬁed ﬁ’agment of 1-120.

Comparing the results between MALDI and ESI of B-lactoglobulin A, we can

conclude:

70

 

 

 

 

 

 

 

 

 

 

 

 

 

 

A
TCEP/NTCB
.g itz121-159 _
a; \ /It266-105
'0
cu
66 E
3
m
_=_
1...-__-_L. A. J‘M i... A
' 2000 4000 6000 8000
B
itz121-162 NTCB/TCEP
a. ,1 -- g, 4
' 4000 8000 12000 16000
mlz

Figure 2.3. MALDI spectra of cleavage products of B-lactoglobulin A
following reduction/cyanylationlcleavage (panel A) or
cyanylationlcleavagelreduction (panel B).

itz = iminothiazolidine derivative.

71

.558 s .85. m sense a sense

.5: s 3 9.35:8 e_E_e28e $8 ”m 29.8

uamt. e\e F .0 H< cozow

.n_<ooEmo._. 9:383 0532.8 weaved 0093.0 .6 EEmSmEoEo 04d: .vN 2:9“.

00.0#

loaned:

00.0“

D
O

I
O

 

 

8. r\

 

   

00 Town:

 

 

 

m2-.. $783

«Snow—N:

I

 

 

Illllllllllllllllllll'l

 

 

O
O

I
O

oo.oo«

oo.oov
u
.n

00.000

00.000

00.000d

72

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

4'8 +7 *6
”71% 1m‘ m .
1m 1 1-65 “266-1 05
7243.40a 4651.100
+7
605
11*
+5
001
0‘ 000 no 1000 M
+5 . m ”7 .
100W +6 9" 112121-159 i 1123;768:62
759 4550.40. .
91.1 91
+4
057 | 1139
31° 416
°‘ no coo «co °“’° “eased ester”
+12
+6 .
1m. «seas *1-105 ,1, 1tz*121-162
910 +11 1182:1001: 100‘ 4370.70!
1075
+14
*‘016 91.1
+10 +5
975
an 911
1123
0 Dale
000 100011001200 7m 90° "00

 

 

 

Figure 2.5. ESI spectra of cleavage products from procedure TCEP/CDAP

of B-Iactoglobulin A

73

.vd 9:9“. c_ 035.80 mm 0.5% 05 9m 2826.80 o._n=._

imp—€430 2.6805 0532.8 mSuoa maniac 00 58003825 040: dd 050E

lavﬂnﬂz
00.nﬂ 00.0d 00.m 00.0

 

 

 

00.0w

   

00.00d

NovirNFNz

nu

00.0mﬂ

n:
00.00“

00.0mm

llllllllllllllllllllllllllll

 

 

 

 

 

74

 

 

+6

100 1 8‘8 itz121-162 .ntact

““703 1849300

 

 

 

 

 

700 900 1100

 

Figure 2.7. ESI spectra of cleavage products from procedure

CDAP/TCEP of B—Iactoglobulin A

75

 

1). Every peak in the HPLC chromatogram was identiﬁed by ESI-MS and assigned
to a speciﬁc fragment (Table 1).

2). The peaks for fragments itzl60-l62(m/z 397) and 1-65(rn/z 7243.4) that cannot
be identiﬁed by MALDI-MS were successfully determined by ESI-MS.

3). HPLC and ESI-MS showed a large B-elimination peak *1-105(m/z 11823.0 Da)
and itz*121-162 ( m/z 4870.7 Da ), but those peaks were totally suppressed in MALDI-
MS.

4). MALDI-MS identiﬁed a large peak for ﬁagment itz66-105(m/z 4653.5) and no
peak for fragment 1-65(m/z 7247.4), but HPLC peak belonging to fragment itz66-105 is
smaller than that of fragment 1-65, which clearly indicate the signal suppression
phenomenon of MALDI.

S). The results of the analysis of cysteine residue status in B-lactoglobulin A
showed a promise for the method of HPLC separation/ESI-MS, and this method can also

be further used by on-line LC/MS.

B. Recognizing the Location of Cysteines and Cystines in Ovalbumin

Ovalbumin is a large protein containing 385 amino acids and having a molecular
weight of 42699Da. It contains one disulﬁde bond and four free cysteines (Figure 2.8).
The expected masses of cleavage fragments with the two procedures of

reduction/cyanylation and cyanylation/reduction are shown in Table 2.2.

| 120 I

73 I
30 LS_$_J 367 382 38

Figure 2.8. Structure of Ovalbumin

Is” SH SH SH
1 1

76

Table 2.2. Expected and Observed Masses of Cleavage Product of Ovalbumin

Reduction/Cyanylation

 

 

Fragments Expected Mass(Da) ESI(Da) MALDI(Da)
1-10 1011.1 1011.0 no
itz11-29 2378.7 2377.4 2377 .4
itz30-72 4762.5 4760.2 4761 .4
itz73-1 19 5427.1 5426.9 5426.1
itzl20-366 27581.1 27581.0 no
it2367-381 1715.1 1713.8 1714.8
itz382-385 429.5 430.0 no
*1-29 3312.8 3311.2 3311.2
itz*30-l 19 10112.6 10108.3 no
itz*367-385 2067.6 2066.5 2068.1
itz“ 120-381 29219.2 29198.7 no

 

Cyanylation/Reduction

 

 

Fragments Expected Mass(Da) ESI(Da) MALDI(Da).
1-10 1011.1 1011.0 no

itzl 1-29 2378.7 2377.1 2378.5
itz30-366 37684.7 37684.0 no
itz367-381 1715.1 1714.0 1715.2
itz382-3 85 429.5 430.0 no

 

"' B-elimination
itz = iminothiazolidine derivative.

77

The spectra of MALDI-MS with the two procedures of CDAP/TCEP and
TCEP/CDAP are presented in Figure 2.9A and 2.9B, respectively. MALDI-MS identiﬁed
peaks of itzll-29 (m/z 2377.4), itz30-72 (m/z 4761.4), itz73-119 (m/z 5426.1), itz367-
381 (m/z 1714.8) and B-elimination peaks of *1-29 (tn/z 3311.2), itz*367-385 (m/z
2068.1) resulting from procedure TCEP/CDAP (Figure 2.9B). The expected peaks for
fragments 1-10 at m/z 1011.1, itz382-385 at m/z 429.5 and it2120-366 at m/z 27581.1 in
procedure TCEP/CDAP were undetected. Instead of ﬁve expected fragment peaks
resulting from procedure CDAP/TCEP, only three peaks were identiﬁed by MALDI-MS,
the expected peaks at m/z 1011.1 and m/z 37684.7 were undetected. Therefore, MALDI-
MS gave an incomplete mass-mapping for protein ovalbumin. The location of cysteines
and cystines in ovalbumin cannot be recognized unambiguously.

The HPLC chromatogram and corresponding ESI spectra of cleavage mixture
resulting from procedure TCEP/NT CB are shown in Figure 2.10 and Figure 2.11
respectively. Fragments of 1-10, itzll-29, itz30-72, itz73-119, itzl20-366, it2367-381,
an itz382-3 85 have been all identiﬁed. Therefore, total numbers and locations of cysteine
residues in ovalbumin can be recognized unambiguously by I-IPLC/ESI-MS.

The HPLC chromatogram and . corresponding ESI-MS spectra of cleavage products
resulting from procedure of NTCB/TCEP are presented in Figure 2.12 and Figure 2.13,
respectively. The fragments of 1-10, itzll-29, itz30-366, itz367-381, and it2382-385
were identiﬁed, indicating the location of free cysteine residues in ovalbumin at Cys 11,
Cys 30, Cys 367, and Cys 382. By comparing the results of procedures of TCEP/NT CB

and NTCB/TCEP, the location of cystine in ovalbumin should be at Cys 73 and Cys 120.

78

 

 

 

 

 

 

 

ID
3 A
“2357-331 \ 5 TCEPICDAP
1."
£5
112367-381\ B
a CDAP/TCEP
Ea
/
O3
‘7‘
R
1o'oo zo'oo ao'oo 40'00 so‘oo who who so

MIZ

Figure 2.9. MALDI spectra of cleavage products of ovalbumin following
reduction/cyanylationlcleavage (panel A) or

cyanylation/cleavagelreduction (panel B).
itz = iminothiazolidine derivative.

79

 

ES 00 E .505. m $0hi$0 000580

em: s8 8.53:8 25298 #8 ”m .828 a“: a 3 .< .828

.c_E:n_m>o S

mueziwoe 2880.5 0532.8 8:85 898.0 00 8538 00 88008525 ohm: .oﬁm 050E

 

 

 

 

 

 

 

 

 

 

 

 

lob—dude
8.3. 8.3 86
p — — F — r—
1: oo.oom
3083 M
89858. I 8.88
«83 H m
_ 1 a
l 8.88
m i
11 oo.ooo«
mzﬁa\ m.— g u

 

 

 

 

 

 

 

 

80

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

+3
100 790
1011 1-10 ‘°° 11211-29
10110:: 237mm
,9
% %
13
846
s +2
111 II 1190
o 900 1000 1100 D” 0 000 ‘1000 '1mvaa/e
+5 +5
100 95“ itz30-72 1m “”5 itz73-119
4760.209 5426.9oa
+6
905
% 11.
+6
:1 i ..
08001000121000. o1100100012001400Dale
+4211
100 1°15 lntact(A)+ilz120-366(B) 5:; itz367-381
+41A +2313 W275810a 100 1713306
+2
959
v.
409
.11:
o 500 700 900 110043“9

 

Figure 2.11. ESI spectra of cleavage products following procedure TCEP/NTCB of ovalbumin

81

 

 

 

 

 

+3

 

 

 

 

 

 

 

+12 . m .
100 844 Itz*30-119 100 1 Itz*367-385
i 10108.30a m5);
1
1 +9
i 1124
l
9‘ 4 % ‘
I
. +2
134 $7 1034
l 794
ﬂ
0 Dale 0 ‘
900 1100 1!!) 500 700 91) 1100
26 10° ‘ 400
4.. 0
1m .. +29 [12. 1tz*120-381 “2332-385
1 1009 291mm 4m
%
412 452
I 468
o "7' 73.2.:1 LATATA: We

 

 

 

 

 

 

8;

 

0 ‘

1045

 

111!) 1211

 

Figure 2.11. Continue

82

 

.oﬁu 0.59”. 5 8888 mm 088 05 em 80288 E285 new Eczow

£23.96 a $0882 9:885 ea

0532.2 3885 31.58.20 232E do ESmBmEoEo 04d... .mﬁw 059“.

laps
00.0N

_ r

 

 

 

 

3

 

 

 

 

 

000.33

 

 

 

00.0
00.0
1. 8.8»
a u
I .n
I. 00.000."
I 00.003

 

 

83

553.95 so awotmopz 2:885 8. 9.52.2 32.85 898.0 .6 808» .mm new 2:9...

 

 

 

 

 

 

 

 

 

 

 

RNF 90. Ru 2!. RN. Rx... R0
030 _ooio§p§p8=§8~0§8$8vaowvao
v9 _
$9
v+ «3
0»
8m .x. a
0+ 8
3:.
0+
8.8% 88$... 888. 884
0+ 3. 8. 30 . 8p . 8..
0... NF...

 
   

 

 

 

 

 

 

 

 

e8 o88v.§.89o8oe8o§ o§ 83:88:05.8 a
one
8..
..x 82
8m .x. 8: «8.x. 8 ..x
N. (8+ <91
838.8884
80.3: 3500-005? 8.8M. Raﬁ 8:2 /§
8.33 Asses. 8.28 e. .8. o: =2 is.
«B 8. as. .8.

 

 

 

 

 

0... En...

 

 

Because ovalbumin only contains one disulﬁde bond, the linkage of disulﬁde bond must
be between Cys 73 and Cys 120.

Comparing the results of MALDI and ESI, several aspects could be concluded:

I). All the undetected peaks by MALDI-MS were identiﬁed by HPLC/ESI-MS
(summarized in Table 2.2 ). Therefore, HPLC / ESI-MS gave the complete mass mapping
for ovalbumin.

2). Peak of ESI at m/z 37684.7 in procedure NTCB/TCEP and peak of ESI at m/z
27581 in procedure TCEP/NT CB are very noisy which are due to the broad HPLC peaks
that include the intact ovalbumin.

3). Compare the two HPLC chromatograms for procedure TCEP/NT CB and
NTCB/TCEP. The procedure NTCB/'1‘ CEP also forms the peaks for ﬁ'agments it230-72
and itz73-119 at m/z 4760.8 and m/z 5426.2 which should not appear in the procedure of
NTCB/'1‘ CEP, although they have lower intensity. Therefore, the cyanylation reaction

with NTCB at pH 8.7 induced the thiol-disulﬁde exchange of ovalbumin.

C. Evaluation of CDAP and NTCB

Based on the results of ovalbumin in Section B, NTCB induced the thiol/disulﬁde
exchange of ovalbumin in procedure of cyanylation/reduction. In order to evaluate the
performance of NTCB and CDAP. We used CDAP to perform the same procedure of
CDAP/TCEP as that of NTCB described in section B. The HPLC chromatograms and
corresponding ESI spectra of the cleavage product from procedure CDAP/TCEP of ‘

ovalbumin are illustrated in Figure 2.14 and Figure 2.15, respectively.

85

 

.05 8:9... 5 contomou mm 058 05 2o; 282260 £0.36 new Eczow

5:50.90 do emote/Bo 23809.8 Eat 8:33 mug—No.0 yo «SEE *0 £832.95 0:: .EN 239".

O
O

I
O

 

 

  
   

www-monng

O
O

I
O

86

oo.ood

oo.ooN

m

oo.oon

 

 

 

 

oo.oo#

mwnémﬂri.

 

 

‘11IIIIIIII1TIIIIIIIIIIT—I

 

 

 

 

 

 

 

 

 

 

 

. +36A
. l "92 |ntact(A)+itz30-366(B)
10° Itz11-29 ‘ +368
2377.60a 1045 1212 W3768408
96
733
Q 2
i it 49°
0 J ‘L we
a” 1000 1200
+3
52 _ 100 ‘
100 ItZ367-381 “0 itz382-385
1714.503 m
% %
439 $1; 4&
m8 412 J a
0 We 0 . J ..l. 41.. IA IL think-Dale
5(1) 700 900 1100

 

 

 

 

 

420440460480

 

 

Figure 2.15. ESI spectra of cleavage product from procedure CDAP/TCEP of ovalbumin

87

The results are summarized in Table 2.3. Several aspects can be concluded:

1). No peaks for fragments itz30-72 and itz73-ll9 at m/z 4762.5 and m/z 5427.1
resulting from the procedure CDAP/TCEP were identiﬁed. Therefore, CDAP reaction did
not induce the thiol-disulﬁde exchange.

2). CDAP has the lower capacity to cyanylate free cysteines. The fragment 1-10 of
ovalbumin cannot be detected by HPLC by using 20-fold excess of CDAP.

3). By comparing the results from NTCB and CDAP, we can obtain clear

assignment for cysteines and cystines in ovalbumin.

C. Conclusions

0 The use of NTCB or CDAP is unique in that it speciﬁcally targets the site being
analyzed.

0 Since CDAP reaction is performed at pH 3, it eliminates the thiol-disulﬁde exchange in
the analysis.

0 HPLC separation / ESI-MS overcomes the signal suppression phenomenon of MALDI-

MS; it can become a complementary method to MALDI-MS. The ESI method can be

ﬁrrther used with on-line LC/MS.

88

Table 2.3. Comparison of CDAP and NTCB

 

 

 

Fragments CDAP/TCEP(Mass: Da) NTCB/TCEPOVIass: Da)
1-10 no 101 1.0

itzl 1-29 2377.6 2377.1
itz30-366 37684.0 37684.0
it2367-381 1714.5 1714.0
itz382-385 430.0 430.0

#itz‘30-1 19 no 10112.6
#itz73-l 19 no 5426.2
#itz30-72 no 4760.8

#: fragments due to thiol—disulﬁde exchange. "': B-elimination

itz = iminothiazolidine derivative.

89

References:

I—l

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

2. Jacobson, G. R., Schaffer, M. H., Stark, G. R., and Vanaman, T. C., J. Biol. Chem,
248,6583(1973)

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

4. Wu, J ., and Watson, J. T., to appear in Anal. Biochem. (1998).

5. Wakeselman, M., Guibe-Jarnpel, E., Raoult, A., and Busse, W. D., J. Chem. Soc.
Chem. Commun., 1, 21(1976).

6. Brockleehurst, K., Malthouse, J. P. G., Baines, B. S., Blenkinsop, R. D., Churcher, J.
A., Mushiri, M. S., and Ormerod, F ., Biochem. Soc. Transactions, 6, 261 (1978).

7. Brocklehurst, K., Malthouse, J. P. G., Biochem. J., 175, 761 (1978).

8. Walkselman, M., and Guibe-Jampel, B., J. C. S. Chem. Comm, 21 (1976).

9. Casey, R., and Lang, A., Biochem. J., 145, 251 (1975).

10. Price, N. C., Biochem. J., 159, 177 (1976).

11. Schaffer, M. H., and Stark, G. R., Biochem. Biophys. Res. Comm, 71, 1040 (1976).

12. Lu, H. S., and Gracy, R. W., Arch. Biochem. Biophys., 212 (2), 347 (1981).

13. Thomas, M. L., Janatova, J ., Gray, W. R., and Tack, B. F., Proc. Natl. Acad. Sci.
USA, 79, 1054 (1982).

14. Roberts, D. D., and Goldstein, I. J ., J. Biol. Chem, 259, 909 (1984).

15. Matsudaira, P., Jakes, R., Cameron, L., and Atherton, B., Proc. Natl. Acad. Sci. USA,
82, 6788 (1985).

16. Mahboub, S., Richard, C., Delacourte, A., and Han, K., Anal. Biochem., 154, 171

(1986)

90

17. Bahler, M., Benfenati, F ., Valtorta, F., Czemik, A. J ., and Greengard, P. J., Cell Biol.,
108, 1841 (1989).

18. Minard, P., Desmadril, M., Ballery, N., Perahia, D., and Mouawad, L., Eur. J.
Biochem., 185, 419 (1989).

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

20. Eshdat, Y., Chapot, M. P., and Strosberg, A. D., FEBS Letters, 1-2, 166 (1989).

21. Katayama, B., J. Biochem., 106, 988 (1989).

22. Sutherland, C., and Walsh, M. P., J. Biol. Chem, 264, 578 (1989).

23. Tang, C., Yuksel, K. U., Jacobson, T. M., and Gracy, R. W., Arch. Biochem. Phys.,
283(1), 12 (1990).

24. May, J. M., Buchs, A., and Carter-Sn, 0, Biochem. J., 29, 10393 (1990).

25. Wines, B. D., and Easterbrook-Smith, S. B., Molecular Immunology, 28 (8), 855
(1991).

26. Som, S., and Friedman, S., J. Biol. Chem, 266, 2937 (1991).

27 . Altamirano, M. M., Plumbridge, J. A., and Calcagno, M. L., Biochem. J., 1153
(1992).

28. Goold, R., and Baines, A. J ., Eur. J. Biochem., 224, 229 (1994).

29. Papayannopoulos, I. A., and Biemann, K., Protein Sci., 1, 278 (1992).

30. Wu, J., Gage, D. A., Watson, J. T., Anal. Biochem., 235, 161 (1996).

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

32. Fenn, J. B., Mann, M., Meng, C. K., Wong, S. F., and Whitehouse, C. M., Science,

246, 64 (1989).

91

33. Muddiman, D. C., Bakhtiar, R., Hofstadler, S., and Smith, R. D., J. Chem. Edu., 74,
1288(1997)

34. Stark, G. R., Methods in Enzymology, vol. 47, 129 (1977).

35. Denslow, N. D., and Nguyen, H. P., in Techniques in Protein Chemistry VII,
Academic Press, Inc., (1996), p.241-248.

36. Price, N. C., Biochem. J., 159, 177 (1976).

37. Nakagawa, S., Tamakashi, Y., Harnana, T., Kawase, M., Taketomi, S., Ishibashi, Y.,
Nishimura, 0., Fukuda, T., J. Am. Chem. Sac., 116, 5513 (1994).

38. Wu, J., and Watson, J. T., Protein Sci., 6, 391 (1997).

92

CHAPTER 3
DISULFIDE MASS MAPPING IN PROTEINS
CONTAINING ADJACENT CYSTEINES WITH

CYANYLATION/CLEAVAGE METHODOLOGY

1. Introduction

Disulﬁde bond formation between cysteine residues is a posttranslational event that
commonly occurs in proteins synthesized in the endoplasmic reticulum. In many cases,
disulﬁde bond formation contributes to the stability of the tertiary structure of the folded
protein molecule (1 ). A full description of the covalent structure of proteins demands that
the connectivity of the bridged cysteines be analyzed. Although there are good methods
for quantifying the number of disulﬁde bonds in proteins, the unambiguous determination
of the location or pairing of disulﬁde bonds continues to be a challenging task for the
protein chemists.

The general strategy for locating disulﬁde bonds in proteins by conventional
methods involves several steps (2, 3). First, a protein is cleaved by enzymes or chemical
reagents between half-cystinyl residues to obtain peptides that contain only one disulﬁde
bend. Second, the amino acid composition, amino terminal sequence or molecular masses
of these peptides are determined using an amino acid analyzer, manual or automated
Edman degradation, and/or mass spectrometry. Finally, peptides identiﬁed using these

data are related to speciﬁc segments of the protein (3). There are two major problems are

93

related to the methodology. First, if a protein contains a pair of adjacent or closely spaced
cysteines in its primary structure, it is impossible to obtain peptides containing a single
disulﬁde bond (3). Second, the conditions most frequently used in speciﬁc enzymatic
digestions may promote disulﬁde bond scrambling (4). Although supplementary
methodologies such as non-speciﬁc fragmentation by partial acid hydrolysis (2) have
been proposed to avoid disulﬁde bond scrambling, it is difﬁcult to deal with the data
obtained by these non-speciﬁc techniques.

Diﬂ‘erent approaches have been made to assign the disulﬁde linkages of proteins
containing adjacent or closely spaced cysteine residues (5-17). Among them, the
methodology proposed by Gray (16, 17) has the capacity to analyze proteins containing
adjacent or closely spaced cysteine residues. In his experiments, peptides were partially
reduced, the nascent free thiols alkylated, and their positions recognized from the results
of sequence analysis. However, it is obviously tedious, if not impractical, to sequence an
alkylated high-mass peptide or protein using this approach.

Recently, Wu and Watson described a novel approach for the assignment of
disulﬁde bonds in proteins of known sequence (18). In this approach, the denatured
protein was subjected to limited reduction by tris(2-carboxyethyl)phosphine (TCEP) in
pH 3.0 citrate buffer to produce a mixture of partially reduced protein isomers; the
nascent sulthydryls were immediately cyanylated by 1-cyano-4-dimethylamino-
pyridinium tetraﬂuoroborate (CDAP) under the same buffered conditions. The cyanylated
protein isomers, separated by and collected from reversed-phase HPLC, were subjected
to cleavage of the peptide bonds on the N-tenninal side of cyanylated cysteines in

aqueous ammonia to form truncated peptides that were still linked by residual disulﬁde

94

bonds. The remaining disulﬁde bonds were then completely reduced to give a mixture of
peptides that can be mass mapped by MALDI-MS. The masses of the resulting peptide
fragments were related to the location of the paired cysteines that had undergone
reduction, cyanylation, and cleavage. This strategy minimizes the disulﬁde bond
scrambling and is simple, fast, and sensitive. Furthermore, it is possible to assign the
disulﬁde linkages of proteins containing adjacent cysteines by this new methodology.

In this chapter, the applicability of partial reduction, cyanylation, chemical
cleavage, and mass mapping approach to the disulﬁde structure analysis of proteins
containing adjacent cysteine residues will be demonstrated. Two model proteins, long R3
insulin-like growth factor-1 and insulin-like growth factor-I, will be used to show the
capability of the new approach for the determination of disulﬁde structure involved in

adjacent cysteine residues.

II. Methodology for Assignment of Disulfide Linkages

The methodology utilized to determine disulﬁde linkages of proteins containing
adjacent cysteines are based on the strategy proposed by Wu & Watson (18). The
approach is illustrated in scheme 3.1. The disulﬁde bonds of a denatured protein are
partially reduced with TCEP by controlled TCEP concentration and reducing time in a
pH 3.0 buffer. Then the resulting nascent sulﬂ'rydryls will be cyanylated by CDAP in the
same buffer solution. After the separation of partially reduced/cyanylated protein
isomers, HPLC fractions can be analyzed by MALDI-MS or ESI-MS to determine which
isomers are singly reduced and cyanylated. Those shifted by +52 Da correspond to a

singly reduced/cyanylated species; +104 Da correspond to doubly reduced/cyanylated

95

 

 

 

Denatured Protein

 

partial TCEP, pH 3,
reduction 10', rt

 

Partially Reduced Protein Isoforms

 

 

 

cyanylation CDAP. pH 3,

610', rt

Cyanylated Protein Isoforms

 

 

 

 

 

HPLC

 

 

 

 

A Separated Protein lsomer —> Identiﬁed by MS

 

 

1N NH4OH,
1hr, rt

cleavage

 

Cleavage Products

 

 

 

reduction TCEP

 

Mass Spectrometry

 

Scheme 3.1. Conceptual overview of the methodology for
assignment of disulﬁde bond pairings in proteins

96

species, etc. Those isomers with a 52-Da mass shift from the mass of the intact protein
are dried and subjected to speciﬁc chemical cleavage in aqueous ammonia. The cleaved
peptides, which may be linked by residual disulﬁde bonds, are then completely reduced
to give a mixture of peptides that can be mass mapped by either MALDI-MS or LC/ESI-
MS. The masses of the resulting peptide fragments are related to the location of the

paired cysteines that had undergone reduction, cyanylation, and cleavage.

III. Experimental Section

MALDI-TOF MS

MALDI mass spectra were obtained on a Voyager Elite time-of-ﬂight (TOF) mass
spectrometer (PerSeptive Biosystems Inc., Framingham, MA) equipped with delayed
extraction and a model VSL-337ND nitrogen laser (Laser Science, Newton, MA). The
accelerating voltage in the ion source was set to 20 kV. Grid voltage and guide wire
voltages were 93.6% and 0.2% of the accelerating voltage, respectively. Data were
acquired in the positive linear DE mode of operation. Time-to-mass conversion was
achieved by external and/or internal calibration using standards of bradykinin (m/z
1061.2), bovine pancreatic insulin (m/z 5734.5), and horse skeletal myoglobin (m/z
16952) obtained from Sigma Chemical Co. (St. Louis, MO). All experiments were
performed using a-cyano-4-hydroxycinnamic acid (Aldrich Chemical Co., Milwaukee,
WI) as the matrix. Saturated matrix solutions were prepared in a 50% (v/v) solution of

acetonitrile/aqueous 1% TFA, and mixed in equal volumes with peptide or protein

97

samples, and applied to a stainless-steel sample plate. The mixture was allowed to air dry

before being introduced into the mass spectrometer.

LC/ESI-MS

The LC system used for LC/ESI-MS analysis was Perkin-Elmer API Applied
Biosystems 173. The separation was carried out on a C18 capillary column. Solvent A
was 0.085% TFA in water and solvent B was 0.085% TFA in CH3CN. The linear
gradient was 5% B to 65% B in 135 minutes at a ﬂow rate of 7 ul/min. ESI mass spectra
were obtained on-line on a Micromass Platform-LC mass spectrometer. This instrument
was equipped with a single quadrupole mass analyzer. The capillary voltage was set at

3.25KV, cone voltage at 30V, and source temperature at 80°C.

HPLC
The HPLC is equipped with Waters model 6000 pumps controlled by a personal

computer. UV detection was at 215nm. The column was a Vydac C18 (#218TP54, lOum

particle size, 300-A pore, 4.6x250mm).

Chemicals

Tris(2-carboxyethyl)phosphine hydrochloride (TCEP) was purchased from Pierce,
Rockford, IL. Guanidine hydrochloride was a product of Boehringer-Mannheim
Biochemicals (Indianapolis, IN). Human recombinant long R3 insulin-like growth factor I
(LR3IGF-I), 2-nitro-5—thiocyanobenzoic acid (NT CB), l-cyano-4-dimethylamino-

pyridinium tetraﬂuoroborate (CDAP), were purchased from Sigma. Human recombinant

98

insulin-like growth factor I was obtained from Austral Biologicals (San Ramon, CA). The
proteins were puriﬁed before using. Acetonitrile and TFA were HPLC grade. The TCEP
solution in 0.1 M citrate buffer at pH 3 was prepared as 0.1 M stock solution and stored
under N2 at —20°C for weeks with little deterioration. The 0.1 M CDAP solution in pH

3.0, 0.1M citrate buffer was freshly prepared prior to use.

Partial Reduction of Proteins

10nm01 Protein samples (IGF-I and LR3IGF-I) were solubilized in 4on1 of 0.1M
citrate buffer (pH 3.0) containing 6 M guanidine-HCI. Partial reduction of disulﬁde bonds
were carried out by adding 10 molar fold of TCEP for the cystine content of the proteins
(300nmol of TCEP is reacted with lOnmol of IGF-I or LR3IGF-I as lnmol of IGF-I or
LR3IGF-I contains 3nmol of cystine). The reduction was conducted at room temperature

for 10 minutes.

Cyanylation of Nascent sulﬂrydryls
To the partially reduced protein mixture was added a 30—fold excess of CDAP over
the cysteine content of IGF-I or LR3IGF-I. Cyanylation of the nascent sulﬂrydryl groups

was accomplished by incubation at room temperature for another 10 minutes.

HPLC Separation of Partially Reduced and Cyanylated Protein Isomers
Partially reduced and cyanylated species were separated by reversed-phase HPLC
_with linear gradient of 30% B to 50% B in 45 minutes, where solvent A was 10% CH3CN

containing 0.1% TFA and solvent B was 90% CH3CN containing 0.1% TFA. The

99

predominant HPLC fractions were collected manually and the masses of the collected
protein isomers were determined by MALDI-MS or ESI-MS. Appropriate ﬁactions were

then dried for further use.

Cleavage of Singly Reduced and Cyanylated Protein Isomers

To the dried HPLC fractions was added 2ul of 6M guanidine-HCI in 1M NH40H
aqueous solution to dissolve the protein residue and then 5 ul of 1M NILOH. Cleavage
of the peptide chain was performed at room temperature for one hour. Excess ammonia

was removed in a vacuum system.

Complete Reduction of Remaining Disulfide Bonds

Truncated peptides, still linked by residual disulﬁde bonds, were completely
reduced by reacting with 2 ul of 0.1M TCEP solution at pH 3.0, 0.1M citrate buffer at
37°C for 30 minutes. Samples were diluted with 100 pl of a 50% (v/v) CH3CN/ 1% TFA
for MALDI-MS analysis or 100 pl of 5% (v/v) CH3CN/0.085% TFA for LC/ESI-MS

analysis.
IV. Results and Discussion
A. Disulfide Mapping of IGF-I
Insulin-like growth factor I (IGF-I) (Mr = 7648.6) is a single-chain polypeptide of

70 amino acid residues containing three intramolecular disulﬁde bonds, two of which

involve adjacent cysteines (19) (Figure 3.1). IGF-I is postulated to be the mediator of

100

@0
0 $33 96
<9 0 o o
e ‘9 e e
e g .9er o o
is 479152 ”e 8 8
69 4863 me Q Q
© 6 G)
09 61 96
60 1a

Figure 3.1. Amino acid sequence and disulfide structure of lGF-l

101

growth hormone action on skeletal tissue as well as mitogenic activity on several cell
types (20, 21). While the complete amino acid sequence of IGF-I was determined in 1978
(19), very little experimental data is available describing the secondary and tertiary
structure of the molecule. A three-dimensional model of human IGF-I has been suggested
ﬁrst by Blundell et al in 197 8 (22). This model is based on the proposed tertiary structure
of porcine insulin as determined by X-ray crystallography (23). From this model, a
disulﬁde bond arrangement between Cys6-Cys48, Cys47-Cy352 and Cysl8-Cys6l was
predicted for human IGF-I. However, since IGF-I contains three disulﬁde bonds, several
isomeric forms can theoretically be formed, especially in the region with the two adjacent
cysteine residues, i.e., Cys47 and Cys48.

Several approaches have been used to identify the disulﬁde linkages of IGF-I (24-
27). While the conventional method of enzymatic digestion did not completely assign the
disulﬁde linkages because of the two adjacent cysteines (24), chemical synthesis and the
combination of enzymatic digestion and fast-bombardment mass spectrometry have been
utilized to try to assign the linkages. The method based on the chemical synthesis
determined disulﬁde linkages of IGF-I by comparing the retention time during HPLC
chromatography of the enzymatic digested peptide fragments with these of chemically
synthesized peptides. Two types of peptides related to two adjacent cysteines, Type I
with Cys6-Cys47 and Cys48-Cys52 as well as Type H with Cys6-Cys48 and Cys47-
Cy552, have been synthesized (25). The disulﬁde bond system of IGF-I was determined
to be the Type 11 form. This method is tedious and only provided indirect experimental
evidences. The other approaches utilized a combination of multiple steps of enzymatic

digestion, Edman degradation, and mass spectrometry (26, 27). Raschdorf et al. assigned

102

the disulﬁde linkages of IGF-I by a three-step mass spectrometric analysis (26). Firstly,
the correct molecular weight of the intact protein was determined by fast atom
bombardment (FAB) mass spectrometric analysis. Secondly, two-fold enzymatic
degradation (chymotrypsin followed by V3 protease, FAB mapping of the cleavage
products) was employed to assign the disulﬁde linkage of CyslS-Cys6l. Thirdly, FAB
tandem mass spectrometry and manual Edman degradation were further used to analyze
the peptide fragment, which contained two other disulﬁde bonds and two adjacent
cysteines. Axelsson et al. (27) utilized the combination of enzymatic digestion, reversed-
phase HPLC, FAB mass spectrometry, and one-cycle Edman degradation. They ﬁrst
digested the protein with V3 protease, then analyzed it by FAB-MS to determine the
disulﬁde linkage unrelated to adjacent cysteine residues. The peptide ﬁagment involved
in adjacent cysteine residues was further degraded by one-cycle Edman degradation,
followed by trypsin digestion and FAB-MS analysis. The methodology consisting of
multiple steps of enzymatic digestion, Edman degradation, and mass spectrometry is
tedious and time consuming.

The disulﬁde linkages of IGF-I can be determined unambiguously by the new
approach based on the partial reduction, cyanylation, chemical cleavage, and mass
spectrometry (18). The overview of the methodology used to assign the disulﬁde linkages

of IGF-I is illustrated in Figure 3.2.
Partial Reduction and Cyanylation of I GF-I

The partial reduction of IGF-I was carried out in citrate buffer (pH 3.0) containing 6

M Guanidine-HCl. Guanidine-HCl was utilized to completely denature the protein.

103

 

 

1II r—I—il 70

 

 

 

 

 

6 18 47 48 52 61
Partial Reduction
TCEP, pH=3
3" i“
1 I l—l—l n 70
18 47 48 52 61
(+other isomers)
Cyanylation
CDAP, pH=3
SCN lSCN
1 I I I I 1 70
6 18 47 48 52 61

(+other isomers)

HPLC Separation

Chemical Cleavage
1 M NH30H

 

1—5 G6 1| IQ‘IBI I 70

8 47 52 61

Reduction
TCEP

SH

I Ind I”
6 8 i
1_5 G 13 47 52 61 7°

 

 

Figure 3.2. Overview of the methodology used to assign
the disulﬁde linkages in lGF-l

104

Tris(2-carboxyethyl)phosphine hydrochloride (TCEP) was used as a reducing reagent
since it has proved to be an excellent reducing agent for disulﬁde bonds (12, 17, 18, 28).
Reduction by TCEP can be carried out at pH 3.0 to suppress disulﬁde bond scrambling.
Furthermore, at pH 3.0, the reduction of disulﬁde bonds is kinetically controlled which
makes partial reduction possible (17 ).

After the partial reduction of the protein, the nascent sulfhydryls can be cyanylated
by CDAP in the same citrate buffer (pH 3.0). There are several advantages of using
CDAP as the cyanylation reagent (18, 28). Firstly, cyanylation by CDAP can be carried
out at pH 3.0, which is compatible with the partial reduction by TCEP. It is therefore
unnecessary to remove excess TCEP, change buffer, or readjust the pH prior to using
CDAP. Secondly, CDAP reacts instantly with TCEP even at pH 3.0, and the excess of
TCEP in the reaction system will be consumed upon addition of CDAP, and no further
reduction during cyanylation is possible. This is deﬁnitely an advantage for the control of
reduction in an unknown protein. While a larger amount of CDAP has to be applied in
case more TCEP is required for partial reduction of proteins with a tight structure, a large
excess of CDAP (>50 fold molar excess over peptide sulﬂrydryls) and excessive
incubation time (> 2 hours at room temperature) will result in some side reactions.
Finally, CDAP is water soluble, which is compatible to the reaction system.

Because IGF-I consists of three disulﬁde bonds, use of the partial reduction
technique should result in the production of three isoforrns, each of which consists of a
singly reduced isomer emanating from a diﬁ‘erent disulﬁde bond. However, as shown in
the chromatogram (Figure 3.3) obtained after partial reduction and cyanylation, besides a

major peak for the intact unreduced protein there are only two discernible peaks (peak 1

105

 

Intact

 

 

 

 

 

_J -

 

 

 

 

1'5 2'5 35 45
minutes

Figure 3.3. HPLC separation of denatured IGF-I and its partially
reduced/cyanylated isomers. Separation was carried out on a Vydac C18
column at a ﬂow rate of 1.0 mllmin with a linear gradient 30-50% B in 45
minutes, where A = 0.1% TFA in water and B = 0.1% TFA in CH3CN. Peaks
1 and 2 represent singly reduced/cyanylated species. Peak 3 represent doubly
reduced/cyanylated species, as determined by MALDI-TOP analysis.

106

V and peak 2) each of which represents a singly reduced isofonn as determined by mass
analysis showing that the molecular weight of each species shifted by 52 Da from that of
the intact protein (a 2-Da shift for conversion of a cystine to two cysteines followed by a
50-Da shift for the replacement of a hydrogen on each of two free sulﬂrydryls with a
cyano group). Peak 3 resulted from the reduction of two disulﬁde bonds as indicated by a
mass shift of 104 Da from the intact protein. This result is not unexpected, as the some
disulﬁde bonds are much more stable than others as indicated in a recent study of the
stability of the disulﬁde bond of Cysl8 to Cys6l in IGF-I (29). This CyslS-Cys61
disulﬁde bond is preferentially formed in the folding process, a feature that may explain
its resistance to reduction under our reaction conditions. In any case, in dealing with a
protein known to contain three disulﬁde bonds, the determination of the pairing of two of

the disulﬁde bonds allows the remaining one to be determined by default.

Cleavage of Peptide Chains

After the reversed-phase HPLC separation of partial reduced/cyanylated protein
isomers, the peptide bonds at N-terrninal side of cyanylated cysteine residues can be
cleaved in 1N NHiOH at room temperature for 1 hour in the presence of guanidine-HCl
(18). These modiﬁed cleavage conditions have the advantage over the conventional
cleavage method by using the mildly alkaline conditions (pH~9) (18, 30). Ammonia can
be easily removed from the reaction system and the cleavage reaction is fast (in one
hour).

After cleavage at cyanylated cysteinyl residues, the truncated peptide chains, still

linked by the remaining disulﬁde bonds, can be easily reduced by excess TCEP. Finally,

107

the peptide mixture is diluted to minimize the adverse effect of guanidine on the

subsequent analysis by MALDI-MS.

Interpretation of MALDI Data

Because cleavage of the peptide chain takes place only at cyanylated cysteinyl sites,
there will be three ﬁ'agments for each singly reduced/cyanylated protein isomer (and
sometimes two overlapped fragments corresponding to B-elimination at one cysteine
site). The mass of each fragment is related to the position of the two-cyanylated cysteinyl
residues, which in turn can be used to deduce the disulﬁde bond linkage. B-Elimination,
an alternative to peptide chain cleavage, provides mass spectral data corresponding to
overlapped peptides (that otherwise would have cleaved) and serves as a conﬁrmation for
the disulﬁde bond pairing assignment.

Table 3.1 lists the calculated m/z values for possible fragments due to cleavage of
the peptide chains at different sites depending on which disulﬁde bond was reduced and
cyanylated. Figure 3.4A and B are two MALDI spectra of peptide mixtures resulting
ﬁ'om cleavage of isomers of singly reduced/cyanylated IGF-I corresponding to HPLC
peaks 1-2 in Figure 3.3, respectively. HPLC peak 3 is due to the reduction of disulﬁde
bonds of Cys47-Cy852 and Cys6-Cys48 according to the MALDI spectra of the cleavage
mixture of HPLC peak 3.

The mass spectrum in Figure 3.4A corresponds to the cleavage products represented
by HPLC peak 1 in Figure 3.3. Two peaks at m/z 4910.4 and m/z 2192.1 are due to
fragments 1-46 and it252-70, respectively (expected m/z 4909.3 and 2189.6). From these

data, one can deduce that peptide chain cleavages occur at Cys47 and Cy352. The

108

Table 3.1. Calculated and observed mlz values for possible fragments

resulting from the cleavage reaction of IGF-I chains at sites of designated
cysteine pairs

 

 

Reduction
of Disulﬁd e Fragment Calculated mlz Observed mlz
Cys47-Cys52 1-46 4909.3 4910.4
47-51 638.7 nd
52-70 2189.6 2192.1
CysG-Cys48 1—5 516.5 nd
6-47 4539.0 4539.2
48-70 2682.1 2533.3
*1-47 4976.5 4977.4

Cys15-cys61 Unreduced

109

 

 

 

 

 

 

 

 

 

 

 

A
/.t252-70 Peak 2
Cys47-Cy352
I
(D
.15
,r 1-46
Ex
._ JL. .JIAAA __L_A
e
Itz48-70 Peak 3
Cys6-Cys48
itz6-47-103 \itz/647
*147
JL . i _
1000 2000 3000 4000 5000 6000 7000 8000

Figure 3.4. The MALDI mass spectra of peptide mixtures resulting from the
cleavage of the two singly reduced/cyanylated lGF—l isomers, corresponding
to the HPLC peaks 2 and 3 in Figure 3.2, respectively. The symbols "itz'"

and * represent the iminothiazolidine derivatives and protonated B—elimination

products, respectively.

110

MALDI peak marked itz47-70-SH in Figure 3.4A indicates the cleavage of the peptide
bond at Cys47, and the cleavage of SH group from Cys52. This further conﬁrms the
cleavage site at Cys47. The fragment itz47-51 (m/z 638.7) was not determined by
MALDI, which is due to the signal suppression at low-mass range of MALDI. Overall, a
disulﬁde bond linkage between Cys47-Cy552 can be unambiguously deduced.

With similar strategy, the other disulﬁde-bond linkage, Cys6-Cys48 also can be
recognized from Figure 3.4B. The MALDI spectrum in Figure 3.4B, corresponding to the
cleavage product represented by HPLC peak 2 (Figure 3.3), shows peaks at m/z 2683.8
and m/z 4539.2, corresponding to ﬁ'agment itz48-70 and itz6-47, respectively. Although
another expected fragment 1-5 (m/z 516.5) is missing, it is still possible to deduce Cys6-
Cys48 ﬁom these two fragments because no other combination gives such masses. One
minor B-elimination product, residues 1-47 (m/z 4977.4), is particularly informative for
conﬁrmation of the assignment in this case. It should be pointed out that the question-
marked peak in Figure 3.4B corresponds arithmetically to itz6-47 minus 103 Da; we have
no rational explanation for its origin at this time. Whether this peak corresponds to the
expulsion of a cysteine residue from itz6-47 awaits further study with other peptides or
proteins containing adjacent cysteines.

After the two disulﬁde linkages of Cys47-Cys52 and Cys6-Cys48 are assigned
unambiguously, the third one can be reasonably deduced as Cysl 8-Cys6l because there
are three disulﬁde bonds in IGF-I, even though the singly reduced/cyanylated isomer

could not obtained for this disulﬁde bond.

lll

B. Disulﬁde Mapping of LR’IGF-I

Recombinant LR3IGF-I is a variant of human IGF-I that contains arginine replacing
glutamate-3; it also has an amino terminal extension of 13 amino acids (30). LR3IGF-I
shows higher biological activities than its analog, IGF-I. Like IGF-I, LR3IGF-I contains
adjacent cysteines at positions 60 and 61. The disulﬁde-bonding scheme, assigned on the
basis of homology to the insulin (or IGF-I) sequence and shown in Figure 3.5, has never
been veriﬁed experimentally. In preparing authentic LR3IGF-I from recombinant sources,
it is important to conﬁrm that the disulﬁde bond linkage is the same as for IGF-I isolated
from natural sources, since a mismatching of disulﬁde bonds could have a major
inﬂuence on any biological activity, as has been observed with disulﬁde-bonded isomers
of insulin (31). Figure 3.6 shows the HPLC separation of LR3IGF-I and isomers of its
partially reduced/cyanylated species. The chromatogram shows similar results as those of
obtained for IGF-I. There are only two discernible peaks, each of which represents a
singly reduced isoforrn as determined by mass analysis. Stronger reducing conditions
(increase the ratio of reducing reagent, apply higher temperature, and prolong the
reduction time) result in the formation of doubly reduced/cyanylated protein isomers, but
the third disulﬁde bond still refused to reduce, indicating the third disulﬁde bond is very
stable.

Both MALDI-MS and on-line LC/ESI-MS were used to analyze the cleavage

products of singly reduced/cyanylated isomers. The results are shown in Table 3.2.

112

Figure 3.5. Amino acid sequence and disulﬁde structure of LR3IGF-l
The bold letters at N-terminal site indicate the 13 amino acid extension.

The bold R represents the replacement of E3 of lGF-l.

113

 

Intact

A]

 

 

 

 

 

 

 

Figure 3.6. HPLC separation of denatured LR3IGF-I and its partially

reduced/cyanylated isomers. Separation was carried out on a Vydac C18
column at a ﬂow rate of 1 mllmin with a linear gradient 30-50% B in 45

l .
10 1'5 20 2'5 3'0 3'5 4'0 45
min

minutes, where A = 0.1% TFA in water and B = 0.1% TFA in CH3CN. Peaks 1
and 2 represent singly reduced/cyanylated species, as determined by MALDI-

TOF analysis.

114

 

 

0.88 0.88 .88 8-8

0.084 0.88 0.88 8-8

0.58; 0.22 4.82 0: 880-280

e: .88 00:0 8-85

588 0.88 0.88 8-8

ammo 0: Esme voéo

~88 0.88 «:8 8-. 880-808
m2-_mm\o.. >0 5.2.2 B 005305 00
NE. 0020000 NE. 0020000 N)». 0003200 2050.000 0000:00m

 

0:00 050005 0000:9000 .6 00:0 00 050.0 700.0”... 00 000000. 0m0>00_0 05

E0: 000.0000 00:09.00: 0.90000 00.. 0029 N0: 0020000 000 0000.30.00 NM 030...

115

Interpretation of MALDI Data

The two HPLC fractions (HPLC peak 1 and 2 in Figure 3.6) were subjected to
cleavage under the described conditions. The corresponding MALDI mass spectra are
shown in Figure 3.7A and B. The peaks at m/z 2191.6 and 6373.3 are due to fragments
itz65-83 and 1-59, respectively. From these data, one can deduce that the cleavage sites
are at positions 60 and 65. The fragment itz60-64 was missing from the MALDI
spectrum due to the low-mass range signal suppression of MALDI-MS. However, the
peaks at m/z 2755.1, representing overlapped peptide fragment itz60-83, where B-
elimination occurs at Cys60, conﬁrm that position 60 is the cleavage site. Therefore,
disulﬁde linkage between Cys60 and Cys65 can be recognized.

The cleavage products from HPLC peak 2 show m/z values attributable to
fragrnentsl-l8, itz19-60, and itz61-83, respectively, suggesting Cysl9 and Cys6l are
cleavage sites (Figure 3.7B). The peak with a single question mark showed a mass
increase of 103 Da ﬁom the fragment 1-18, suggesting a cysteine attached to the
fragment, whereas the peak with double question marks showed a mass decrease of 103
Da from the fragment 19-60, suggesting a cysteine removed from the fragment as
occurrence in IGF-I. Like IGF-I, further study should be conducted to address the
problem. On the other hand, the third disulﬁde bond in LR3IGF-I can be easily assigned
to Cys31-Cys74 based on the linkages of other two disulﬁde bonds. Therefore,
unambiguous assignment of the disulﬁde structure is evident even if such undesired
fragments are present, because the fragments from speciﬁc cleavage at the native

disulﬁde bonds are sufﬁcient to make a positive conclusion.

116

 

A

Peak 1
CysGO-Cys65

/it265-83

 

bum-83
l
i

- /rt260-83 /1_59

Peak 2
Cys1 9-Cys61

/itz61-83

Q
‘-

.L 7
/ 7? itz19-60
\ /
..ll-J. n

2000 £00 4000 5000 6000 7000 8000 9000
Mass (mlz)

 

 

 

 

Figure 3.7. The MALDI mass spectra of peptide mixtures resulting from the

cleavage of the two singly reduced/cyanylated LR3IGF-l isomers,

corresponding to the HPLC peaks 1 and 2 in Figure 3.4, respectively. The
symbol "itz" and * represent iminothiazolidine derivatives and B—elimination
products, respectively. The peaks with question marks are discussed in the text.

117

Interpretation of L C/ESI-MS Data

In order to conﬁrm the disulﬁde structure of LRJIGF-I assigned by MALDI-MS,
especially for the low-mass range fragments and question marked peaks, LC/ESI-MS was
utilized to analyze the cleavage products of singly reduced/cyanylated isomers. Figure
3.8A and B as well as Figure 3.9A and B show the results.

The ﬁrst isomer (peak 1 in Figure 3.6) was analyzed by LC/ESI-MS to give the
reconstructed total ion current (RTIC) chromatogram in Figure 3.8A and the
corresponding mass spectra in Figure 3.8B. The peaks in the RTIC in Figure 3.8A can be
assigned to speciﬁc fragments 1-59, itz60-64, itz65-83, itz65-83 minus the mass of H20,
and the uncleaved protein (IP in Figure 3.8A) according to the corresponding masses
from ESI spectra (Figure 3.83). The ESI spectra in Figure 3.8B give masses 6372.7Da,
639 Da (protonated peptide), and 2189.7 Da which are due to fragments 1-59, itz60-64,
and itz65-83, respectively. The missing peak (mlz 638.0 corresponding to ﬁagment itz60-
64) in MALDI spectrum was identiﬁed by LC/ESI-MS. From these data, the cleavage
sites can be easily deduced, which are at positions 60 and 65. Therefore, the disulﬁde
linkage should be between Cys60 and Cys65.

The peaks in the RTIC in Figure 3.9A resulting from analysis of cleavage products
of the second singly reduced/cyanylated isomer (peak 2 in Figure 3.6) can be assigned to
fragments 1-18 (1977.8 Da), itz19-60 (4540.98 Da), and itz6l-83 (2684.93 Da) according
to corresponding ESI spectra in Figure 3.9B, which indicate Cysl9 was linked to Cys 61.
While the peak corresponding to 1-18 plus 103 Da that was identiﬁed by MALDI-MS
was not detected by LC/ESI-MS, the ‘redundant’ peak B in the BSI spectrum in the

middle panel of Figure 3.9B that corresponds arithmetically to itzl9—60 minus 103 Da

118

002002.00 050__0N0_£o5:._

€000.00. N... 60.50 0.... <00. 0.090 5 00.000 0:. 05 0. 05000000000 0000000 05 .0 0.0.0000 000E

_8 e... 0. 8.0 0.8.. 08:85 8880 0...... 0E. 0. <00 0.8.. 0.0 0.8.. e. . 0.8. 0.5: e. 8.88088
000.05 _-00_0m._ 00003005000000. >650 0 .0 00.0800 05 So... 0.35:. 000000 .0 00:00. 020090.. .0.0 0.390

 

 

NE... . 8.0.. . 8.0,. 08 08 8.. 8: 8m. 80. 8... 08 80 80. 8-0. 8.. 8. 08 08

 

 

 

 

 

 

 

 

 

 

 

 

 

 

a: i 0 NE .................. 1. o NEV ......... 1.14.1 .13....- . . L . L —O
0.. 0 4 A .81.
000.
00+.- ~+ 08
0mm. ..x. ..+ 0.. ..x.
+
000.
0+
. . . 0 .. 038.88. 2.
cm 008 «.8 .< .
8.. . a 0.. .8. 8.8 u. .2 .8. 8.8 u. 08 8.
m h... . 0+
88 88 800 88 08. 08. 08
cmowptlil-tt-.-.--b--.»--l-r--i-f-[-i.ti-rl-r-- ---.-i-_°
/ ._.
- /.._ one M
00 A“ 6 ... .x.
0. m
a a
8 o
0.0
6
.0 .
< .

 

 

 

uoow

 

119

 

.<0.m 0.090 5 00.000 0.... 0.0 00 05000000000 0000000 000 00 0.00000 0000. .ww 0.0 0.
00.0 0.090 08:00.0 000>00.0 0.0 00 0...”. 0. 4.0.0 0.090 0.0 0.090 5 N 0.000 040... 00 05000000000 000.00.
700.00.. 00003003000000. 2050 0 00 000>00.0 000 E00 0.02:. 000000 00 00.000. ms.-.wm<o._ 0.0 0.090

 

co: com? ooor com com...

 

 

 

00.0 «00.0000 H<
m 2.-..

 

 

cow? 009 009. com com

 

 

 

 

 

vmd 000.0030 Hm
and 0 00.0000 H4.
«00-0.0.0 FM: .0 00.0 05.

 

 

 

00..-... ...... _ .. .0... g 00 ; 00mm3000f0000 000 .000. -.-0
_ i__ 2.00 :0 .44 .
0000 0+ 0.000
0 0
m...
..x. .00 .00
wmm ,000
00:. 0:. 0+
V+
.000 0+ 000
000 000 .000 000 0+ .000
m... 0+ 0+

0 0.0 0 m0.mmmm H4.
nu. we”.

 

 

 

 

000m

.0000.

Mama . .oowm F .oomm .

.80.. . 2.0

 

 

 

 

be

 

 

120

was still present. Overall, the disulﬁde bond structure of LR3IGF-I can be easily assigned
by the combination of MALDI and ESI-MS. These results provide direct evidence for the
disulﬁde-bonding pattern in LR3IGF-I that has hitherto only been presumed by homology

to that in IGF-I (30).

V. Conclusion

In this chapter, the disulﬁde linkages of IGF-I and LR3IGF-I have been determined
by the methodology of partial reduction, cyanylation, chemical cleavage, and mass
mapping. Both proteins contain three disulﬁde bonds and two adjacent cysteines. We
have demonstrated that the methodology provides a simple and effective method for the
assignment of disulﬁde-bond pairing in proteins containing adjacent cysteine residues, a
challenging problem for the conventional approach. Furthermore, the new approach
offers an important advantage of minimizing disulﬁde bond scrambling. The combination
of MALDI and LC/ESI-MS are very powerful tools to analyze a variety of cleavage

products for disulfide mapping.

VI. Future Work

The approach used in this chapter to assign the disulﬁde linkages in proteins
involved in partial reduction, cyanylation, HPLC separation, chemical cleavage, and mass
mapping. Theoretically, it is not necessary to separate the mixture of singly
reduced/cyanylated protein isomers if only the singly reduced isomers are produced
during the partial reduction process. Controlling the partial reduction conditions (ratio of

the reducing reagent to disulﬁde content, reduction time and temperature) can satisfy this

121

condition. Omitting the step of chromatographic separation of partial reduced/cyanylated
protein isoforms could make the analysis faster and less sample consumption. An
overview of a proposed new approach is illustrated in Figure 3.10 based on a protein
containing two disulﬁde bonds.

In the case of two disulﬁde bonds in a protein, there are three possible disulﬁde
linkage patterns (Case I, II, III in Figure 3.10). The partial reduction can be conducted
at pH 3.0 in the controlling conditions to produce only singly reduced protein isoforms,
two isoforms (1, H, or III) for each case shown in the Figure 3.10. Cyanylation can be
performed in the same buffer (pH 3.0). Right after the cyanylation, the pH of the solution
can be adjusted to pH 12.0 by ammonia, followed by incubation for 1 hour at room
temperature. After evaporating ammonia from the solution, the remaining disulﬁde bonds
can be reduced by TCEP. MALDI-MS or LC/ESI-MS can be used to analyze the
cleavage mixture. Because the sequence of a protein is known as well as the numbers and
location of cystines can be determined by the method described in Chapter 2, the

disulﬁde linkages of a protein can be determined based on the combination of different
Peptide fragments for each case (Figure 3.10). The masses of peptide fragments for each
case can be predicted according to its disulﬁde linkages. By matching the predicted and
Observed masses, the disulﬁde-bond-pairing pattern can be determined.

It should be pointed out that the numbers of peptide fragments in the cleavage
Inixture would be increased with the increasing number of disulﬁde bonds in the protein.
IVIA‘LDI-MS sometimes has the signal suppression phenomenon as described in Chapter
2’ which will result in the incomplete identiﬁcation of the peptide fragments in the

cleavage products, providing uncertainty in the disulﬁde linkage determination. While

122

zomoanm ofamﬁoﬁEoEoéo: nomoaoa Co 39205 of. .35 0.59...

 

n. N. n. n.
omuom .muuow .m Tr .omév dang. d-..

8|,“qu 8:80 ELHIL
Iw I
8 cm
3'0 mi all?
2. 1 :mo

:otozbomw

Slomumc $.30 EJSIP
+
3.90 malomwda all?

«NFuISougmoG»

om o¢om owe. _.

, -
zom 20%
+
8 3 on cm S

20 20
8 .mquFoamicmxoa
to.co:bom 3..th
on 3. on cm 2 F
=_

_.

 

 

.. w. n. n.
oméw .mnuou .m 7.. .092.” .omuce .m-—.

on or
$.90 8% 2.le
:w 5
av
3|...de awbmlqc all
:w m
cozoaboKJ

9019.0 mmﬁmlob ﬁlm—Fl?
+
Slomlomhv 3% ml;

~9ut$oum>m2o+
oz ov on cm or

 

w

20% 20w
+
co ov on em or

u _.
20% 20W

«oﬁulqyzozmicg
Sensuom REE
on S. on cm 2

al..

an am:

also 8'30 anomﬂml
mlﬁPbWIMh v Bi V Ill
O I” I o a o Q _.

tOtOBbmm

cm 2
84:0 mmlqmc mudlv
+
8% Elev mIF

«usuISoumSmoG
ov on ow or

on _.
23, 5% _L

+
co CV omoN or

u w
7 o.._ 5%

a .mutscoﬁxcgo
counsnmm Ream

om ov. on em 3...

 

 

 

 

\ x \
omév .mnucn .muur .cmuou .232. .m-..

L

 

123

LC/ESI-MS might be used to solve the problem, disulﬁde linkages may still be identiﬁed
even though the peptide fragments cannot be completely determined by mass
spectrometry. The peptide fragments indicated in bold numbers in Figure 3.10 for each
case can be used to identify the disulﬁde linkages because they have unique masses from
the set of peptide fragments that result from the other cases. Alternatively, the
combination of the masses of several not complete peptide fragments can also be used to
determine the disulﬁde linkages. For example, the combination of ﬁagments 1-9, 10-19,
and 20—50 in case I of Figure 3.10 can be used to determine the disulﬁde linkage pattern
of Cyle-Cys20 and Cys30-Cys40 because the same combination cannot be obtained for
the other two cases (11 and III). An effective way to do the calculation and matching is by
the computer soﬁware. The combination of partial reduction, cyanylation, mass
spectrometry and computer software matching would provide a powerful technique to

assign the disulﬁde linkages in proteins.

124

VII. References

l. Creighton, T. B., Bioessays, 8, 57 (1988).

2. Smith, D. L., Zhou, Z., Methods Enzymal., 193, 374 (1990).

3. Hirayama, K., Akashi, S., in Biological Mass Spectrometry: Present and Future,
Matsuo T., Caprioli, R. M., Gross, M. L., and Seyama, Y., eds., pp 299 (1994).

4. Ryle, A. P., Sanger, F ., Biochem. J., 60, 535 (1955).

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

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

7. Serensen, H. H., Thomsen, J ., and Bayne, S., Biomed. Environ. Mass Spectrom. 19,
713 (1990).

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

9. Zhang, D., and Liang, S., J. Protein Chem, 12, 735 (1993).

10. Krishnamurthy, T., Hauer, C. R., Prabhakaran, M., Freedy, J. G., and Hayashi, K.,
Biol. Mass Spectrom., 23, 719 (1994).

11. Tang, Y., and Selsted, M. B., J. Biol. Chem, 268, 6649 (1993).

12. Gray, W. R., Protein Science, 2, 1749 (1993).

13. Haniu, M., Hsieh, P., Rohde, M. F., and Kenney, W. C., J. Biochem. Biophys., 310,
433 (1994).

14. Yamashita, H., Nakatsuka, T., and Hirose, M., J. Biol. Chem, 270, 29806 (1995).

15. Fairlie, W. D., Stanton, P. G., and Hearn, M. T. W., Biochem. J., 314, 449 (1996).

125

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

17. Gray, W. R., Protein Sci., 2, 1732 (1993).

18. Wu, J ., and Watson, J. T., Protein Science, 6, 391 (1997).

19. Rinderknetcht, B., and Humbel, R. B., J. Biol. Chem, 253, 2769 (1978).

20. Froesch, E. R., Burgi, H., Ramseier, E. B., Bally, P., and Labhart, A., J. Clin. Invest.,
42, 1816 (1963).

21. Rinderknecht, B., and Humbel, R. B., Proc. Natl. Acad. Sci., USA, 73, 2365 (1976).

22. Blundell, T. L., Bedarkar, S., Rinderknecht, B., and Humbel, R. B., Proc. Natl. Acad.
Sci. USA, 75, 180 (1978).

23. Blundell, T. L., Dodson, G. G., Hodgkin, D. C., and Mercola, D. A, Adv. Protein
Chem, 26, 279 (1972).

24. Forsberg, G, Palm, G., Ekebacke, A., Josephson, S., and Hartmanis, M., Biochem. J.,
271, 357 (1990).

25. Iwai, M., Kobayashi, M., Tamura, K., Ishii, Y., Yamada, H., and Niwa, M., J.
Biochem., 106, 949 (1989).

26. Raschdorf, F., Dahinden, R., Maerki, W., and Richter, W. J ., Biomed. Environm.
Mass Spectrom., 16, 3(1988).

27. Axelsson, K., Johansson, S., Eketorp, G., Zazzi, H., Hemmendorf, B., and Gellerfors,
P., Eur. J. Biochem., 206, 987 (1992).

28. Wu, J., Gage, D. A., and Watson, J. T., Anal. Biochem., 235, 161 (1996).

126

29. Hober, S., Forsberg, G., Palm, G., Hartmanis, M., Nilsson, B., Biochemistry, 31, 1749
(1992)

30. Tomas, F. M., Knowles, S. B., Chandler, C. S., Francis, G. L., Owen, P. C., and
Ballard, F. J ., J. Endocrinology, 137, 413 (1993).

31. Rinderknetcht, B., Humbel, R. B., FEBS Lettt., 89, 283 (1978).

127

CHAPTER 4
TRAPPING AND IDENTIFICATION OF INTERMEDIATES

DURING THE REFOLDING OF LR3IGF-I AND IGF-I

I. Introduction

Considerable insight to the folding and unfolding pathways of a protein can be
obtained from trapping and characterizing intermediate structures involved in the
dynamic process (1-4). A particularly difﬁcult aspect in the study of protein folding is the
fact that intermediates may be short-lived and therefore difﬁcult to isolate and analyze
structurally and functionally. Disulﬁde-containing proteins provide an opportunity to
capture sulfhydryl-containing intermediates by chemical reaction during the time course
of folding or unfolding (5-8). The folding pathways of several proteins (9-11) have been
studied in this way; among them, bovine pancreatic trypsin inhibitor (BPTI) (6, 8, 12,
13), and ribonuclease A (14-18) have been extensively characterized.

In order to isolate and characterize the intermediates that are involved in folding or
unfolding of proteins containing disulﬁde bonds, it is necessary to stop both the inter- and
intramolecular thiol/disulﬁde exchange reactions that convert one intermediate to
another. Two traditional trapping methods have been used. One method traps all
sulthydryls irreversibly by alkylation with iodoacetate or iodoacetamide. However,
rearrangement of intermediates during trapping with iodoacetate has been observed for
both BPTI (8) and ribonuclease A (19), since quenching by 0.2M iodoacetate is
comparable to or, in some cases, slower than the rate of some of the intramolecular

rearrangements that interconvert intermediates of the same oxidation state (20). Another

128

method quenches thiol/disulﬁde exchange by lowering the pH to < 2 by adding acid. An
advantage of acid trapping is its reversibility; intermediates trapped at low pH can be
transferred to high pH to allow further rearrangement or folding in experiments designed
to more completely characterize particular pathways (8, 10). While quenching by
acidiﬁcation is rapid and occurs at the diffusion-controlled rate, it does not completely
stop thiol/disulﬁde exchange (21). Therefore, structural characterization of a given
trapped intermediate must be done promptly.

We have developed methodology to trap and identify folding intermediates based
on cyanylation methodology and mass spectrometry (22). Our approach has two
advantages. Firstly, acidiﬁcation quenches the thiol/disulﬁde exchange rapidly, and
cyanylation of free suﬂ'iydryls precludes further exchange because of irreversible
chemical modiﬁcation. Secondly, trapping of sulfhydryls is already part of our procedure
for structural elucidation of the intermediates.

After trapping and isolation of folding or unfolding intermediates, the disulﬁde
structure must be determined. Conventional methodology for determining disulﬁde
structures of folding intermediates involves cleavage between every cysteine residue in
the intermediates by enzymatic/chemical digestion, which is tedious, cumbersome, and
may risk artifact formation due to disulﬁde exchange. Furthermore, conventional
approaches may have difﬁculty in analyzing intermediates containing adjacent cysteines
since proteolytic and/or chemical degradation may not achieve cleavage between cysteine
residues (23).

1 Our methodology involves partial reduction of disulﬁde bonds in a protein,

cyanylation of suthydryl groups, HPLC separation of partially reduced/cyanylated protein

129

isoforms, chemical cleavage of the peptide bonds at the N-terminal side of cyanylated
cysteine residues, and mass analysis of cleavage products by mass spectrometry (24).
This new developed methodology has the advantages of preventing disulﬁde scrambling
and cleaving the peptide backbone between adjacent cysteines; furthermore, it is fast,
simple, and sensitive (25).

Insulin-like growth factor-I (IGF-I) is a single-chain polypeptide of 70 amino acid
residues containing three intramolecular disulﬁde bonds, two of which involve adjacent
cysteines (26) (Figure 4.1). IGF-I is postulated to be the mediator of growth hormone
action on skeletal tissue as well as mitogenic activity on several cell types (27, 28).
Several research groups have studied refolding pathway of IGF-I (29-33). Hober, et al
(29, 30) trapped the folding intermediates of IGF-I by pyridylethylation at pH 8.7. Five
major forms of IGF-I were detected, which includes a one-disulﬁde intermediate, two
two-disulﬁde intermediates, and a mismatched three-disulﬁde intermediate (29). A
different folding pattern was obtained by trapping the folding intermediates with
acidiﬁcation (31-33). Instead of ﬁve intermediates detected by pyridylethylation, six
intermediates were identiﬁed by trapping with acidiﬁcation, which include a mixed
disulﬁde intermediate (3 2, 33). A nonnative two-disulﬁde intermediate was. identiﬁed
with acidiﬁcation, but not identiﬁed with pyridylethylation. Furthermore, while a one-
disulﬁde intermediate (Cysl 8-Cysl6) was identiﬁed as a major form in the folding
process of IGF-I by trapping with pyridylethylation (29), it only appeared as a minor
component when trapped by acidiﬁcation (32, 33). “Therefore, a further study of the

folding intermediates of IGF-I needed to be conducted.

130

 

 

Figure 4.1. Amino acid sequences and disulﬁde structures of LR3IGF-l
and lGF-l. The numbers in the parentheses indicate the location of
cysteine residues of lGF—l. The 13 hold letters at the N-terminus indicate

the 13-amino acid extension in LR3IGF-l. The bold R represents the
replacement of E3 of lGF-l.

131

Recombinant human long R3 insulin-like growth factor-I (LR3IGF—I) is a variant of
human insulin-like growth factor-1 (IGF-I) in which glutamate 3 is replaced by arginine
and a 1.3-residue extension appears at the N-terminus (Fig. 4.1). It contains three disulﬁde
bonds and two adjacent cysteines. The disulﬁde linkages have been determined by partial
reduction/cyanylation/mass mapping (25). LR3IGF-I is substantially more potent than
IGF-I in affecting carbohydrate metabolism and in stimulating the growth of fetal tissue
in animals (34). The results of a folding study of LR3IGF-I (31) were signiﬁcantly
different from those published for IGF-I refolding (29, 30, 32-33). However, the disulﬁde
structure of the folding intermediates of LR3IGF-I has not been studied by the authors
(31).

In this chapter, we integrate the different structural features of IGF-I and LR3IGF-I
into the observed folding and refolding dynamics by using the same methodology to
study the behavior of both proteins. Furthermore, our methodology permits direct
determination of the disulﬁde pattern in each of the intermediates, which were trapped
without artifact formation. We have trapped the intermediates of refolding of IGF-I and
LR3IGF-I and determined disulﬁde structures of the intermediates by cyanylation
methodology and mass spectrometry. The disulﬁde structure of all the refolding
intermediates of both IGF-I and LR’IGF-I were identiﬁed by our methodology. It is the
ﬁrst direct experimental evidence for the disulﬁde structure of the nonnative three-
disulﬁde intermediate, which was only predicted previously by the structural constraints
of disulﬁde bonds in the three-disulﬁde intermediates (29). The refolding pattern of IGF-I

obtained by our methodology was similar to that of obtained by trapping with

132

acidiﬁcation. Furthermore, the results of our investigation provide integrated information

for the folding of insulin-like growth factor families.

11. The Strategy for Trapping Folding Intermediates

The procedures for trapping and analyzing refolding and reductive unfolding
intermediates of proteins are shown in Figure 4.2. Reduced-unfolded protein was
refolded at pH 8.7 in the presence of redox buffer consisting of oxidized glutathione
(GSSG) and reduced glutathione (GSH). The folding intermediates were trapped by
cyanylating the sulﬂrydryl groups of the intermediates with 1-cyano-4-dimethylarnino-
pyridinium (CDAP) at pH 2~3. The trapped cyanylated intermediates were separated by
reversed-phase HPLC. The numbers of disulﬁde bonds were determined by MALDI-MS
based on the mass shift in MW between native protein and that of intermediates. The
mass shiﬂ should be 104 Da between native protein and the one-disulﬁde intermediate,
52 Da between native protein and the two-disulﬁde intermediates, and no mass shift
between the intact protein and the three-disulﬁde intermediates. The 52-Da mass shift
between intact protein and the two—disulﬁde intermediate arises from 2 Da for reduction
of one disulﬁde bond and 25 Da for replacement of a hydrogen atom on each of the
sulﬂrydryl groups with CN, etc. Aﬁer HPLC separation and puriﬁcation, the disulﬁde
structure of the intermediates was determined based on the methodology of partial
reduction, cyanylation, chemical cleavage, and mass mapping (24).

The reverse direction of refolding involving disulﬁde formation in a protein is
reductive unfolding involving the reduction of disulﬁdes by a free-thiol reagent such as

cysteine. A study of reductive unfolding can provide insight to protein refolding since

133

 

2:205 ho 9.5.8:: 9:032 ucm @5282 no $3803 2: .6 263.05 on... .Né 2:9“.

 

. mOwEUOr—tmuc— ho Ghauuzbw 005330 0.3 05559.00;

 

 

_ 855852. 85.5 _

 

 

 

mecca $535 _ 5:838
Co .353: mos—E030 _ .

Dull

 

 

 

_moﬁﬁoctcg 9.6.8:: cognac...—

 

moﬁeoctous 9.6.8 339...

 

n_<oo\miNuIa

Econom 1w.

4

 

v

 

 

d<ooat~uze 358:0

 

 

953530
28530
Sure unisex
_chi wou_£c:-uoo:uom _
8.233%
maneuver

61-50 uawe... :80:on

 

 

 

_ seen. 352 _ seen. 362 _

 

134

reductive unfolding is the reverse of refolding. The disulﬁde bonds of a native/folded
protein are reduced in the presence of a thiol reagent, cysteine. The unfolding
intermediates are trapped, separated, and analde by the same methods as those used in
the study of folding intermediates.

Compared to the conventional approach, our methodology provides several
advantages. First, the trapping reaction, occurring in acidic solution, greatly reduces the
risk of sulfhydryl/disulﬁde exchange that bothers alkylation trapping. Secondly, the
CDAP quantitatively cyanylates free sulﬂrydryl groups to stop further folding. Unlike
acid quench, which is totally reversible and requires further chemical modiﬁcation in
order to determine the disulﬁde structure of the folding intermediates, the cyanylated
intermediates can be directly subjected to characterization of disulﬁde structure by the
approach described in chapter 3. Thirdly, the disulﬁde-mapping technique itself provides

unique advantages over conventional techniques, as demonstrated in previous chapter.

111. The Experimental Section

MALDI-TOF-MS

MALDI mass spectra were obtained on a Voyager Elite time-of-ﬂight (T OF) mass
spectrometer (PerSeptive Biosystems Inc., F ramingham, MA) equipped with delayed
extraction and a model VSL-337ND nitrogen laser (Laser Science, Newton, MA). The
accelerating voltage in the ion source was set to 20 KV. Grid voltage and guide wire
voltages were 93.6% and 0.2% of the accelerating voltage, respectively. Data were

acquired in the positive linear DE mode of operation. Time-to-mass conversion was

135

achieved by external and/or internal calibration using standards of bradykinin (m/z
1061.2), bovine pancreatic insulin (m/z 5734.5), and horse skeletal myoglobin (mlz
16952) obtained from Sigma Chemical Co. (St. Louis, MO). All experiments were
performed using a-cyano-4-hydroxycinnamic acid (Aldrich Chemical Co., Milwaukee,
WI) as the matrix. Saturated matrix solutions were prepared in a 50% (v/v) solution of
acetonitrile/aqueous 1% TFA, and mixed in equal volumes with peptide or protein

samples, and applied to a stainless-steel sample plate. The mixture was allowed to air dry

before being introduced into the mass spectrometer.

Materials

Recombinant human long insulin-like growth factor-I (LR3IGF-I) was purchased
from Sigma Chemical Co. (St. Louis, MO). Insulin-like growth factor-I was obtained
from Austral Biologicals (San Ramon, CA). The proteins were puriﬁed before use.
Oxidized glutathione, reduced glutathione, and cysteine were obtained from Sigma
Chemicals (St. Louis, MO). Tris(2-carboxyethyl)phosphine (T CEP) hydrochloride was
purchased from Pierce Chemical Co. (Rockford, IL.). Guanidine hydrochloride was
obtained from Boehringer-Mannheim Biochemicals (Indianapolis, IL.). Citric acid,
sodium citrate, hydrochloric acid, and l-cyano-4-dimethylamino—pyridinium
tetraﬂuoroborate (CDAP) were purchased from Sigma and used without further
puriﬁcation. Acetonitrile and triﬂuoroacetic acid (TF A) were of HPLC grade. TCEP
solution in 0.1M citrate buffer at pH 3.0 was prepared as 0.10 M stock solution and
stored under N; at 40°C for weeks with little deterioration. The 0.10M CDAP solution in

0.1 M citrate buffer at pH 3.0 was freshly prepared prior to use.

136

 

Refolding Experiments

LR3IGF—I or IGF-I (0.1mg) was dissolved in 0.5ml citrate buffer, pH 3.0, containing
6 M guanidine-RC1 and 0.1M TCEP reducing agent. The denaturation and reduction of
the proteins were carried out at 37°C for 2 hours. HPLC was used to separate the reaction
mixture and the fraction corresponding to the reduced LR3IGF-I or IGF-I was collected,
dried and stored at -80°C.

The refolding of reduced and denatured protein was initiated by diluting the
reduced/unfolded protein sample with 0.10 M Tris-HCl buffer (pH 8.71), containing
1mM oxidized glutathion (GSSG), 10mM reduced glutathion, 0.2M KCl, 1 mM EDTA,
to a ﬁnal protein concentration of 0. lmg/ml. Folding intermediates were trapped in a time
course manner by removing aliquots (0.1ml) of protein solution and mixing them with
1.0M HCl solution containing freshly prepared 0.2 M CDAP to pH 2~3. Cyanylation of
free sulthydryl groups by CDAP proceeded at room temperature for 10 minutes. The
trapped intermediates were immediately separated by HPLC. The ﬁactions were
collected manually and analyzed by MALDI-TOP MS. Those with 0-Da, 52-Da, 104-Da,
156-Da increases over the mass of the intact protein corresponding to the three-disulﬁde
(non-native or native N), 2-disulﬁde, l-disulﬁde, and O-disulﬁde (R) species,
respectively. The collected intermediates were dried in a speedvac and stored in a —80°C
freezer.

Reductive Unfolding Experiments

Native LR’IGF-I and IGF-1 (0.1mg/ml) was reduced/unfolded in 0.10 M Tris-HCl

buffer (pH 8.71), containing 0.25mM cysteine, 0.20 M KCl, 1 mM EDTA. The reduced-

137

 

unfolded intermediates were trapped, separated, analyzed with the same method as
described for folding intermediates.

After unfolding reached equilibrium, 0.1ml of unfolding solution was removed and
cysteine was added to a ﬁnal concentration of 100 mM. After incubate at different time

points, the intermediates were trapped, separated, and analyzed by the method describe

above.

Disulfide Mapping of Refolding and Unfolding Intermediates

One-disulﬁde intermediate (1): Aﬂer capture by CDAP, puriﬁcation by HPLC, and
drying by speedvac, the one-disulﬁde intermediate (1) was subjected to chemical cleavage
by adding 2 ul of 1 M aqueous NH40H containing 6 M guanidine-HCI to dissolve the
protein residue in the bottom of a 500-141 Eppendorf tube. Then an additional 5 p1 of 1 M
NILOH were added, and cleavage of the peptide chain was performed at room
temperature for one hour. Excess ammonia was removed in the speedvac. Truncated
peptides, which may still be linked by residual disulﬁde bonds, were completely reduced
by reacting with 2 pl of 0.1 M TCEP solution at 37°C for 30 minutes at pH 3-5. Samples
were diluted with 20 pl of a 50% (v/v) acetonitrile/ 1% TFA solution prior to analysis by
MALDI-MS.

Two-disulﬁde intermediate: Two experiments were applied for recognizing
disulﬁde pairing in the two-disulﬁde intermediate. One portion of the puriﬁed cyanylated
intermediate was subjected to the same treatment as the one-disulﬁde intermediate
described above. This approach was applied to deﬁne the location of the two remaining

sulfhydryl groups (that had not yet formed a covalent disulﬁde bond) by mass mapping of

138

 

 

the resulting cleaved peptides. Another aliquat of the same cyanylated intermediate was
subjected to partial reduction/further cyanylation/chemical cleavage/mass mapping to
determine the disulﬁde bond structure (24, 25). Partial reduction of the disulﬁde bonds in
the two-disulﬁde intermediate was performed by adding a 20-molar excess of TCEP
which was estimated from the ratio of the HPLC peak area for two-disulﬁde intermediate
to the total of all peak areas representing various forms of the protein; the total protein in
the mixture was ﬁxed. The reduction was performed at room temperature for 10 minutes.
After cyanylating the sulfhydryl groups of the partially reduced intermediate, HPLC was
used to separate the partially reduced/cyanylated protein isoforms. The fractions
corresponding to the singly reduced/cyanylated isomers were subjected to chemical
cleavage, reduction of the remaining disulﬁde bond, and mass analysis of the cleavage
products.

Three-disulﬁde intermediate: the three-disulﬁde intermediate contains three
disulﬁde bonds based on no observed mass shift from the MW of the intact protein. The
assignment of disulﬁde bonds in three-disulﬁde intermediate was performed by partial
reduction/cyanylation/cleavage/complete reduction according to the mee described

above of the assignment of disulﬁdes in two-disulﬁde intermediate.

HPLC Separation

The puriﬁcation and separation of reduced LR3IGF-I and IGF-I, the trapped
(cyanylated) intermediates, and the partial reduction/cyanylation products were
performed by reversed-phase HPLC with linear gradient elution using Waters model

6000 pumps controlled by a PC. UV detection was at 215 nm. The column was a Vydac

139

 

C18 (#218TP54, IO-pm particle size, 300-A pore, 4.6X250 mm). Solvent A was water
containing 0.1% TFA. Solvent B was acetonitrile/water (9:1, v/v) containing 0.1% TFA.
The gradient was 30%-50% solvent B linear in 45 minutes. The ﬂow rate was lmllmin.
The HPLC ﬁactions were collected manually and the contents were then dried for further

use.

IV. Results and Discussion

A. Refolding and Reductive Unfolding of LR’IGF-r
Trapping and HPLC Separation of Folding Intermediates of LR3 IGF -I

The distribution of intermediates at any time during the refolding process can be
visualized by a chromatogram of cyanylated and other (i.e., three-disulﬁde species)
intermediates. Figure 4.3A shows chromatograms of an array of intermediates trapped by
CDAP at acidic condition at various times showing the appearance and disappearance of
certain species during the refolding process. In addition to native LR’IGF-I (N) and
completely reduced LR3IGF-I (R), a total of 3 peaks were identiﬁed at different time
points. The structural information of intermediates was determined by MALDI-MS based
on the mass shift between the MW of intermediate and that of intact protein. Peak 1
contains two species, a three-disulﬁde intermediate and a mixed two-disulﬁde
intermediate, according to the mass shift between the MW of intermediates and that of
intact protein. While no mass shift was observed for the three-disulﬁde intermediate, a
612-Da mass shift was determined for mixed two—disulﬁde-intermcdiate. Peak 2 is a two-

disulﬁde intermediate because a 52-Da mass shift was determined. Peak 3 represents a

140

 

 

 

 

 

2min R
l 3
10min 2
N R
1
W
30min N
2
l R
. 3.7 .
5 1'5 2'5 3“ 3'5 35 55
Minutes

Figure 4.3A. HPLC separation of CDAP-trapped intermediates during

the time course of refolding of LR3IGF-I

141

one-disulﬁde intermediate based on a mass shift of 104 Da between the intermediate and
the intact protein.

Reduced/unfolded LR3IGF-I is ﬁrst converted to a one-disulﬁde intermediate (peak
3) and a two-disulﬁde intermediate (peak 2) after 1 minute in the presence of 10mM GSH
and 1mM GSSG. A three-disulﬁde intermediate and a mixed two-disulﬁde intermediate
(coeluted in peak I) appeared by 2 minutes. Native protein (N) was formed by 10
rrrinutes. The ratio of the components changed after 10 minutes. However, the ratio of the
components after several hours was almost the same as that at 30 minutes, indicating that
the refolding of LR3IGF-I reached equilibrium by 30 minutes. Therefore, the folding
intermediates of LR3IGF-I appeared in the order of one-disulﬁde intermediate, two-
disulﬁde intermediate, three-disulﬁde intermediate, and native protein in the folding
process of LR3IGF-I.

Alter equilibrium was reached, if the concentration of GSSG was increased to
100mM and the mixture incubated for another 30 minutes, almost all of the two-disulﬁde
and three-disulﬁde intermediates were converted to native/folded protein (Figure 4.3B).
This result implies that the normative three-disulﬁde intermediate can convert to the

native protein by the pathway(s) involving the two-disulﬁde intermediate.
Trapping and HPLC Separation of Unfolding Intermediates of LRj [CF -I

Figure 4.4 shows the time-dependant distribution of the intermediates of reductive

unfolding of LR3IGF-I in the form of several HPLC chromatograms.

142

 

 

 

 

 

W

 

 

 

10 2'0 3ro 4'0 5'0

Minutes ' 1

Figure 4.3B. HPLC separation of CDAP-trapped folding products after
folding reached to the equilibrium, increased the concentration of GSSG
to 100 mM and incubated for another 30 minutes.

143

 

 

 

 

 

 

 

 

 

0.25mM Cysteine N
3h
N

0.25mM Cysteine
24h

. N
0.25mM Cysteine
71h

2

JUL-

a) Increase to 100mM Cysteine
b) Incubate for an additional 0.5h

 

 

 

 

 

 

10 20 3'o"4'o"5'0
min

Figure 4.4. HPLC separation of CDAP-trapped intermediates in the time
course of reductive unfolding of LR’IGF-I

144

 

In order to avoid the formation of the mixed-disulﬁde intermediate, cysteine is
used as a thiol reagent in the reductive unfolding. The pattern of the intermediates of
reductive unfolding was similar to that of the refolding of LR3IGF-I. A total of three
intermediates was identiﬁed in addition to native (N) and reduced (R) forms. Peak 1 is
the three-disulﬁde intermediate, peak 2 is the two-disulﬁde intermediate, and peak 3 is
the one-disulﬁde intermediate. Reductive unfolding of LR3IGF-I reached equilibrium at
approximately 71 hours. However, if the concentration of cysteine was increased to 100
mM and incubated for another 30 minutes after reaching the 71 -hour equilibrium point
under the initial conditions, the one—disulﬁde intermediate (peak 3) and a large amount of
reduced/unfolded protein (R) were formed. The intermediates of reductive unfolding of
LR3IGF-I appeared in the order of three-disulﬁde intermediate (peakl ), two-disulﬁde
intermediate (peak 2), one-disulﬁde intermediate (peak 3), and ﬁnally reduced/unfolded
protein, which is the reverse of the folding process of LR3IGF-I. It should be pointed out
that some of the two-disulﬁde interrnediate(s) should exist in the pathway of the
conversion of native protein to the three-disulﬁde intermediate. However, no two-
disulﬁde intermediates were trapped and identiﬁed even at early time of unfolding,
indicating that the two-disulﬁde intermediates had too short a life time for our

methodology.

Disulﬁde Structure of One-Disulﬁde Intermediate @eak 3) of LR’IGF-I
The one-disulﬁde Intermediate (peak 1) contains only one disulﬁde bond, thus four
sulfhydryl groups should be available (free) in the intermediate 1. After intermediate I is

trapped by CDAP at acidic condition, four sulﬂrydryl groups should be cyanylated. The

145

chemical cleavage reaction can be conducted in 1M aqueous NH40H right after the
separation of the intermediates by HPLC, and MALDI-MS can be used to analyze the
masses of the cleavage products. The location of the sulﬂrydryls can be determined by
identifying the cleavage sites of the intermediate because only free sulfhydryl groups will
be cyanylayed and the peptide bond at the N-terminal side of the cyanylated sulfhydryls
will be cleaved.

Figure 4.5 shows the MALDI spectrum of the cleavage products of the cyanylated
one-disulﬁde intermediate corresponding to HPLC peak 3 in Figure 4.3A. The expected
and observed masses of the cleavage products of the one-disulﬁde intermediate are
shown in Table 4.1. Fragments 1-18, itzl9-59, itz61-83, and itz65-83 were detected.
Ideally, ﬁve cleavage products were expected; however, because the adjacent cysteines
were free, cleavage between them yields a single residue (as itz) and it is too low in mass
to be detected (interference from matrix ions). Observing a cleavage product terminating
at residue 59 implies that Cys60 was cyanylated. As a whole, these data indicate that
peptide bonds at the N-terminal side of cyanylated cysteine residues at positions 19, 60,
61 , and 65 were cleaved. Thus four free cysteine residues were located at positions 19,
60, 61, and 65. Obviously, no cleavage occurs at Cys31 and Cys74; otherwise, ﬁagments
related to cleavage at cyanylated Cys31 and Cys74 should be identiﬁed by MALDI-MS.

Therefore, the Cys31-Cys 74 linkage can be deduced. It is notable that this intermediate

has a native disulﬁde structure.

146

 

it265-83
/ . Peak 3
CysS1-Cys74

93*
e

.g' .
1-1a\ .\ /t261-83

\ /'t219-59
MMJLUL 1114....» A ELLA

2600 3000 4000 5000 6000 7000
Mass (mlz)

 

 

 

 

Figure 4.5. MALDI spectrum of peptide mixture resulting from the cleavage
of the puriﬁed and cyanylated intermediate corresponding to peak 3 in Figure 3.4.
The symbol "itz" represents iminothiazolidine derivatives.

147

Table 4.1. The expected and observed masses of cleavage products of
one-disulﬁde intermediate of LR3IGF-l

 

 

F rag men ts Expected masses Observed masses
(Da) (Da)
1-18 1978.4 1g7g_5
itz19-59 4435.8 4436.7
it260 145.2 nd
it261-64 535.5 nd
itz65-83 2188.6 2190]

it261-83 2681.1 2684.0

 

148

...«1‘

Disulfide Structure of T wo-Disulﬁde Intermediate @eak 2)

The intermediate corresponding to HPLC peak 2 in Figure 4.3A contains two
disulﬁde bonds. The disulﬁde structure of the intermediate can be determined by the
methodology of partial reduction/ cyanylation/chemical cleavage described in Chapter 3.
The isolated fraction of HPLC peak 2 (Figure 4.3A) was subjected to partial reduction
and cyanylation. Figure 4.6 shows the chromatogram of the intact intermediate and its
partially reduced/cyanylated products. Two one-disulﬁde bonded species are expected
from partial reduction of the intermediate (a two-disulﬁde species). However, only one
one-disulﬁde species arises from the partial reduction and cyanylation. That is, the HPLC
peak 1 (Figure 4.6) has one disulﬁde opened that has the retention time and MW of the
intact intermediate, and peak 2 has two disulﬁdes Opened (its MW is 52 Da greater than
that of the intact intermediate). Only one of the two expected one—disulﬁde isoforms is
detected. after cyanylation (peak 2 in Figure 4.6) because of the high stability of one of
the two disulﬁde bonds in the intermediate.

Although the disulﬁde structure of the intermediate can be directly determined ﬁ'om
the cleavage products of the residual cyanylated intermediate and partially reduced
cyanylated isoforrn of the intermediate, it is much easier to identify the location of the
two free sulﬂrydryls ﬁrst. Table 4.2 lists the expected and observed masses of the
cleavage products of HPLC peak 2 in Figure 4.3A. The MALDI spectrum (Figure 4.7A)
of the cleavage products of the species corresponding to the HPLC peak 2 in F igru'e 4.3A
shows ﬁagments 1-59, it265-83, b-elimination product itz60—(74*)-83, and uncleaved
product itz60—(74)-83, indicating that cleavage occurred at cysteine residues 60 and 65.

Therefore, Cys60 and Cys65 are the two free suﬂrydryls in the intermediate.

149

 

 

 

 

 

 

 

Intact
Intermediate\ 1
2
rd Lin/It
5 "' 1'5 " is ‘ i 5 ‘ 45
min

Figure 4.6. HPLC separation of intact intermediate (peak 1 here)
corresponding to HPLC peak 2 in Figure 4.3 and one of its expected
partially reduced/cyanylated isomers (peak 2). The peak with a question
mark did not correspond to any speciﬁc cleavage fragments to the
disulﬁde structures of the intermediates as analyzed by MALDI.

150

 

Table 4.2. The expected and observed masses of cleavage products of
the puriﬁed/cyanylated two-disulﬁde intermediate of LR3IGF-l

 

 

Frag men ts Expected masses Observed masses
(Da) (Da)
1-59 6371.2 6373.3
“260.64 637.7 nd
it265-83 2188.6 2191.6
2755.1

itz60-(65*)-83 2748.3

151

 

/i1265-83 A

0)
a
é’

a? .
.~\6\ /I1260-(65)-83 ”-59

 

 

 

.314 ° 3 l/ /itz19-59,
. A

2000 3000 4000 5000 6000 7000 8000 9000
Mass (mlz)

Figure 4.7. MALDI spectra of peptide mixtures reSulting from the cleavage

of the puriﬁed/cyanylated intermediate corresponding to HPLC peak 2 in

Figure 4.3 and peak 1 in ﬁgure 4.6 (A) and one of its partially reduced/cyanylated
isomers (peak 2 in Figure 4.6) (B). The symbol "itz" and * represent
iminothiazolidine derivatives and B-elimination species, respectively.

|tz60-(65)-83 represents that the residue in parenthesis (residue 65) is

cyanylated but remains uncleaved. lt260-(65*)-83 represents that the B—elimination
occurs at residue 65.

 

 

 

152

 

Figure 4.7 B shows the MALDI spectrum of cleavage products of the cyanylated
species represented by HPLC peak 2 in Figure 4.6. Fragment 1-18, itzl9-59, itz61-83,
and itz65-83 are identiﬁed, implying that another disulﬁde pair, Cysl9-Cys6l , must have
been reduced, cyanylated and cleaved. Therefore, cysteine 19 is connected to cysteine 61
as a disulﬁde bond in the two-disulﬁde intermediate, and the second disulﬁde bond must
be linked between the two remaining cysteine residues, 31 and 74.

In summary, the two-disulﬁde intermediate corresponding to HPLC peak 2 in

Figure 4.3A contains a native structure, Cysl9-Cys61 and Cys3 l-Cys74.

Disulﬁde Structure of Three-Disulﬁde Intermediate weak 1)

Using the similar methodology of partial reduction/cyanylation described above one
can recognize the disulﬁde structure of the three-disulﬁde intermediate. The HPLC
chromatogram of the residual intact three—disulﬁde intermediate and its partially
reduced/cyanylated isoforms is shown in Figure 4.8. Mass analysis showed that peaks 2
and 3 represent singly reduced/cyanylated species. Three singly reduced isoforms were
expected from partial reduction, but the HPLC chromatogram shows only two singly
reduced/cyanylated isoforms, indicating that one of three disulﬁde bonds is sufﬁciently
stable to resist to reduction. Figure 4.9A and B show MALDI spectra of cleavage
products of the species represented by peak 2 and peak 3, respectively. Table 4.3 lists the
observed and expected masses of cleavage products of the species represented by peaks 2
and 3 in Figure 4.8. The MALDI spectrum of peak 2 (Figure 4.9A) identiﬁed fragment
peaks consisting of residues 1-60, it261-83, and itz65-83, indicating that Cys6l is linked
to Cys65 as a disulﬁde bond.

153

 

 

 

Intact
Intermediate

\

 

 

 

J Latent -
N—JL:___5 - - -. i-

15 ' ' 2'5 35 45
min

 

 

 

Figure 4.8. HPLC separation of residual intact intermediate corresponding to
HPLC peak 1 in Figure 4.3 and its partially reduced/cyanylated isomers.

154

 

 

 

 

 

 

 

 

 

' -83 A
”“265 Peak 2
Cy561-Cys65
7? ,
/It261-83
‘ 7‘ 1-60
3.. A r .L A
. B
1-1 8 /l’(260-83 peak 3
\ Cys19-Cys60
? o
l|/?? 2? /ltz19-59
.w a. _n ___
3000 5000 7000 9000
mlz

Figure 4.9. MALDI spectra of peptide mixtures resulting from cleavage

of partially reduced/cyanylated isomers corresponding to HPLC peaks 2

and 3 in ﬁgure 4.8 of the three-disulﬁde intermediate (peak 1 in Figure 4.3).
The symbol "itz" indicates iminothiazolidine derivatives. The question marked
peaks are discussed in the text.

155

Table 4.3. The expected and observed masses of cleavage products of singly
reduced/cyanylated isomers of the three-disulﬁde intermediate of LR3IGF-l

 

 

Fragments ExPectgda)masses observgia)masses
1-60 6473.4 6476.3
itz61-83 2681.1 2684.9
11265-83 2188.6 2191.4

1-18 1978.4 1978.6
itz19-59 4435.8 4437.4
11260-83 2783.3 2788.1

 

156

The MALDI spectrum of the cleavage products of the species represented by peak
3 (Figure 4.9B) shows fragment peaks corresponding to residues 1-18, itzl9-59, and
itz60-83, indicating that Cysl9 is connected to Cys60 as a disulﬁde bond. The third
disulﬁde bond must be linked between the two remaining cysteines residues, Cys3] and
Cys74. It should be pointed out that in addition to the mass spectral peaks corresponding
to the speciﬁc cleavages related to reduced disulﬁdes, some of the peaks corresponded to
a mass of the MW of the expected ﬁagrnent i103 Da or $185 Da were identiﬁed. How
these ions were produced is still not clear. It may result from impurities of recombinant
LR3IGF-I or some side reactions. Further work should be done to obtain a reasonable
explanation. A MS/MS experiment might be used to analyze the ilO3Da and :l:185Da
species to provide sequence information. In spite of the existence of the i103Da and
21:185Da peaks, assignment of disulﬁde linkages is still possible, since the speciﬁc
cleavage related to the cyanylated nascent sulfhydryls ﬁ'om reduced disulﬁdes provides
enough information.

To summarize, three intermediates in the refolding of LR3IGF-I have been captured
and characterized, which include a one-disulﬁde intermediate (I), a two-disulﬁde

intermediate (II), and a three-disulﬁde intermediate (111) (Figure 4.10).

B. Refolding and Reductive Unfolding of IGF-I

LR’IGF-I is an analog of IGF-I with an N-terminal extension of 13 amino acids and
replacement of glutamate 3 of IGF-I with arginine. The effect of structural features on the
folding of LR3IGF-I and IGF-1 can be obtained by comparing their refolding and

unfolding processes.

157

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

19 31 60 61 65 71%
1 C C CO C v ------ 83
I !
Native LR3IGF-l (N)
SH
19 ‘ 6 61 65 '
1 P 0 80 “ dunuas
; 31 g g 74
H H H
One-disulﬁde intermediate (l)
l 1‘ .
19 31 60 61 65 74
C C C r‘ C ------ 83
l 1
5H 6H
Two-disulﬁde intermediate (II)
I l
19 31 60 61 65 7A
C C C C V ------ 83
I J !

 

 

Nonnative three-disulﬁde intermediate (lll)

Figure 4.10. The disulﬁde structure of folding intermediates of LR3IGF-l

158

Trapping and HPLC Separation of Folding Intermediates of [CF -1

Folding intermediates of IGF-I trapped at 10 seconds, 1 minute, 2.5 minutes, and 30
minutes, respectively, were separated by HPLC (Figure 4.11). The number of disulﬁde
bonds was determined from mass shift between the MW of intermediates and intact IGF-I
based on analysis by MALDI-MS. A total of 6 peaks were identiﬁed in addition to the
native (N) and reduced (R) IGF-I. Peak 1 represents a three-disulﬁde species. Peaks 3, 4,
5 represent two-disulﬁde species. Peak 6 represents a one-disulﬁde specie. Peak 2
represents a mixed disulﬁde species according to the mass shift between the MW of the
intermediate and that of the intact protein.

One of the two-disulﬁde intermediates (peak 3) and the one-disulﬁde intermediate
appeared after 10 seconds. The other two two-disulﬁde intermediates (peaks 4 and 5) and
the mixed-disulﬁde intermediate (peak 2) showed up after 1 minute. After 2.5 minutes
the three-disulﬁde intermediate (peak 1) and native IGF-I formed. The folding between 1
minute and 2.5 minute was fast. The pattern of the folding intermediates of IGF-I at 30
minutes was similar to that at 2.5 minutes. The only difference is a small change in the
ratio of the intermediates; the peak corresponding to native IGF-I was the largest one at
30 minutes, while the two-disulﬁde intermediate (peak 3) was the largest one at 2.5
minutes. Therefore, the appearance of the folding intermediates of IGF-I is in the order of
a onerdisulﬁde intermediate, the two-disulﬁde intermediates, and the three-disulﬁde
species including native IGF-I.

The distribution of the folding intermediates of lGF-I trapped by cyanylation at
acidic condition is similar to the published results that were obtained by acidic trapping

(32, 33). Six intermediates, including a one-disulﬁde intermediate, three two-disulﬁde

159

 

 

 

 

10 seconds

1 minute

 

 

 

 

 

 

Figure 4.11. HPLC separation of CDAP-trapped intermediates during

3
2.5 minutes 2
N
R
1 4
5 6
2 N
30 minutes 1
4 5
5 1'5 2'5 3'5
Minutes

the time course of refolding of IGF-I.

160

 

intermediates, a three-disulﬁde intermediate, and a mixed disulﬁde intermediate, were
captured here by our methodology and as well as by acidic trapping by others (32, 33),
while only four intermediates were trapped by pyridylethylation under basic condition
(29). Furthermore, the one-disulﬁde intermediate was the major form captured by
pyridylethylation (29), while it was the minor form in the distribution of the folding
intermediates trapped by our cyanylation methodology as described herein and as well as

by acidiﬁcation as reported elsewhere (32, 33).

Trapping and HPLC Separation of Unfolding Intermediates of IGF -I

Figure 4.12A shows the time-dependent distribution of the intermediates of
reductive unfolding of IGF-I in the form of several HPLC chromatograms. Normative
three-disulﬁde intermediate 111 (peak I) appeared after 30 minutes in the presence of 0.25
mM cysteine. The ratio of the three-disulﬁde intermediate III and native IGF-I (N, peak
2) was almost the same at the time points of 2, 20, and 71 hours. This indicates that an
equilibrium was reached at 2 hours. Two different features can be revealed between the
unfolding processes of LR3IGF-I (Figure 4.4) and IGF-1 (Figure 4.12A). Firstly, The
conversion of native IGF-I to the nonnative three-disulﬁde intermediate of IGF-I is faster
than that of native LR3IGF-I to the normative three—disulﬁde intermediate of LR’IGF-I.
This implies that the normative three-disulﬁde intermediate of IGF-I is more stable than
the normative three-disulﬁde intermediate of LR3IGF-I. This also can be seen from the
refolding processes of IGF-I and LR’IGF-I, in which the normative three-disulﬁde

intermediate (peak 1 in Figure 4.11) of IGF-I has a higher yield (to total amount of IGF-I)

161

 

 

 

0.25 mM cysteine

 

 

 

 

 

 

 

30 min
111
N
0.25 mM cysteine
2 h
111
N

0.25 mM cysteine l

20h
L

0.25 mM cysteine N
71 h

 

 

 

 

Wm EC.

111
5 10 1'5 2'0 35 3'0 3' 5 40

Minutes

Figure 4.12A. HPLC seperation of CDAP-trapped intermediates in the time course of
reductive unfolding of IGF-I

 

 

 

 

162

 

than that of the normative three-disulﬁde intermediate (peak 1 in Figure 4.3A) of
LR3IGF-I. Secondly, while there was no detectable two-disulﬁde intermediate 11;, (see
disulﬁde structure of HA in Figure 4.19) in the unfolding process of IGF-I in the presence
of 0.25 mM cysteine until even 71 hours, the two-disulﬁde intermediate (peak 2 in Figure
4.4) was appeared at 24 hours. This indicates that disulﬁde bond Cys47-Cy852 in IGF-I is
more stable than that of corresponding disulﬁde bond Cys60-Cys65 in LR3IGF-I.

Using the similar process to the unfolding of LR3IGF-I, the concentration of
cysteine was increased to 100 mM and incubated for another 1 and 30 minutes after the
equilibrium of the reductive unfolding of IGF-I. Figure 4.12B shows the chromatograms
of the unfolding products of lGF-I at 1 and 30 minutes after the equilibrium. The one-
disulﬁde intermediate (1) and the two-disulﬁde intermediate (le) were formed at 1
minute. After 30 minutes, two-disulﬁde intermediate 1111 was disappeared and one-
disulﬁde intermediate I almost completely converted to reduced-unfolded IGF-I (R). The
second-stage unfolding (after the equilibrium) of IGF-I is faster than that of LR3IGF-I
(Figure 4.4), which indicates that the disulﬁde bond Cys6—Cys48 in native IGF-I is
reduced rapidly after the reduction of disulﬁde bond Cys47-Cys52. This result may be

due the cooperative effect between these two disulﬁde bonds.

Disulfide Structure of Folding and Unfolding Intermediates of IGF-I
The disulﬁde structure of the folding intermediates of IGF-I were determined by
using the same methodology as described earlier for identifying the disulﬁde structure of

the folding and unfolding intermediates of LR3IGF-I. Table 4.4 lists the expected and

163

 

 

 

A) Increase to 100 mM cysteine
B) Incubated for an additional 1 min

 

 

:11A l

A) Increase to 100 mM cysteine
B) Incubated for an additional 30 min

 

 

 

 

Ur

 

10 i5 20 25 30 35 40

Minutes

Figure 4.12B. HPLC separation of CDAP-trapped intermediates in the unfolding
alter the equilibrium.

164

Table 4.4. The expected and observed masses of the fragments related
to the disulﬁde assigment of the intermediates of IGF-I

 

 

Fragments Expected masses Observed masses
(Da) (Da)
1-5 515.54 nd
1—47 5010.5 5012.2
1-46 4908.5 4909.1
it26-46 4435.8 4433.9
11247 145.2 nd
itz47-70 2784-1 2786.7
itz47-51 535.5 nd
itz48-51 637.7 nd
“243-70 2681.1 2684.4
itz52-70 2188.6 2191.0

 

165

observed masses of the fragments related to the disulﬁde assignment of the intermediates
of IGF-I.

Figure 4.13 represents the chromatogram of partially reduced/cyanylated isomers of
the three-disulﬁde intermediate (peak 1 in Figure 4.11). Peaks 2 and 3 indicate the singly
reduced/cyanylated isomers of the intermediate based on the mass shift between the MW
of isomers and intact intermediate. The fractions of peaks 2 and 3 were collected and
incubated in aqueous ammonia. From analysis of the cleavage products of the cyanylated
isomer represented by peak 2, fragments consisting of residues 1-47, itz48-70, it252-70,
and itz48-70 minus the mass of an SH group were identiﬁed (Figure 4.14A), indicating
cleavage sites at residues 48 and 52. Therefore, the connection between Cys48 and CysS2
can be deduced. The MALDI spectrum of the cleavage products of peak 3 (Figure 4.14B)
shows the fragments itz6—46 and itz47-70, indicating cleavage sites at residues 6 and 47.
Thus, Cys6 is connected to Cys47 as a disulﬁde bond in the intermediate. Although a
singly reduced/cyanylated isomer was not obtained for the remaining disulﬁde bond due
to its high stability, its linkage between Cys18 and CysGl can be easily deduced by
default. Thus, the three-disulﬁde intermediate (peak 1 in ﬁgure 4.11) has disulﬁde bonds
between residues 48-52, 5-47, and 18-61.

The intermediate corresponding to HPLC peak 3 in Figure 4.11 is a two-disulﬁde
intermediate. The chromatogram of its partial reduction/cyanylation reaction mixture is
shown in Figure 4.15. Instead of two partially reduced/cyanylated isomers, only one
isomer was identiﬁed (peak 2) apparently because of the stability of one of the disulﬁde
bonds. The MALDI spectrum (Figure 4.16A) of the cleavage reaction mixture for the

species represented by peak 1 in Figure 4.16 (also peak 3 in Figure 4.11) shows

166

 

l

/Intact Intermediate

 

 

 

 

 

 

2
7 3
d 9
5 1'5 2'5 3'5 4
Minute

Figure 4.13. HPLC separation of residual intact three-disulﬁde intermediate
corresponding to HPLC peak 1 in Figure 4.1 land its partially reduced/cyanylated

isomers. Peaks 2 and 3 each represent two-disulﬁde species. The question marked
peaks are discussed in the text. .

167

 

 

 

 

 

 

 

 

A
' 52-70
/112 Peak 2
:1: Cys48-Cy352
"9
O
[x
2'?
"E ' 48 70 147
...Juamaa A L
B
/ "Z474" Peak 3
8 Cys6-Cys47
$2
28'
g
. 1126-46
2000 3000 4000 5000 6000 7000 8000

mlz

Figur 4.14. MALDI spectra of peptide mixtures resulting from the
cleavage of the two singly reduced/cyanylated isomers,
corresponding to HPLC peaks 2 and 3 in Figure 4.13, respectively.
The symbol ”itz" represents iminothiazolidine derivatives.

168

 

MW}

1

Intact
/ Intermediate

 

 

 

 

(in

 

i5 25 35 45
min

Figure 4.15. HPLC chromatogram of partial reduction/cyanylation mixture
showing residual intact intermediate (two-disulﬁde species) corresponding
to HPLC peak 3 in Figure 4.11and one of the expected partially
reduced/cyanylated species (peak 2). The question marked peak is discussed
in the text.

169

 

 

 

 

 

 

 

 

 

 

 

A
/i1252-70
I
”9
8 1-46
"7 um 70 ‘
#LEDAM an n
/itzSZ-70 '3
8
a
itz48-70 av
10
° 47-70 ‘1' )26-46
7 ' 3\ 146
J- n EJLLLWK 1
2000 3000 4000 5000 6000 7000 8000
mlz

Figure 4.16. MALDI spectra of peptide mixtures resulting from the
cleavage of the puriﬁed and cyanylated intermediate corresponding
to peak 1 in Figure 4.14 (also to peak 3 in Figure 4.11) (A) and the
partially reduced/cyanylated isomer corresponding to HPLC peak 2
in Figure 4.14 (B). The symbol "itz" and * represent iminothiazolidine
derivatives and B-elimination peak, respectively. The peak with a
question mark cannot be assigned to any speciﬁc cleavage fragment
of the intermediate.

170

 

fragments consisting of residues 1-46, itz47-70, it252-70, and itz47-70 minus the mass of
SH group, indicating that the cleavage sites occur at cysteine residues 47 and 52, which
represent two ﬁee sulﬂrydryl groups in the two-disulﬁde intermediate. The MALDI
spectrum (Figure 4.16B) of the cleavage products represented by peak 2 in Figure 4.15
shows fragments consisting of residues itz6-46, 1-46, itz47-70, itz48-70, and it252-70,
indicating that another disulﬁde bond, Cys6-Cys48, must have been reduced, cyanylated
and cleaved. Therefore, Cys6 is connected to Cys48 as a disulﬁde bond in the two-
disulﬁde intermediate. The remaining two cysteine residues, Cys18 and Cys6l, must be
linked together as a disulﬁde bond as deduced by default. Thus, the two-disulﬁde
intermediate corresponding to peak 3 in Figure 4.11 contains a native disulﬁde structure,
Cys6-Cys48 and Cys18-Cys61.

It should be pointed out that the question marked peaks in the chromatograms in
Figure 4.13 and Figure 4.15 did not correspond to any speciﬁc cleavage fragments related
to the disulﬁde structures of the intermediates as analyzed by MALDI. We thought at
ﬁrst that the peaks corresponded to species resulting from a side reaction of CDAP with
the intermediates. However, these species also showed up in the chromatogram when
only partial reduction with TCEP was performed but without cyanylation with CDAP.
Therefore, the question-marked peaks apparently result from the process of partial
reduction, which may be due to some impurity that coeluted with the intermediates
during the HPLC separation of the folding products.

Figure 4.17A and B show the MALDI spectra of cleavage products of the
cyanylated two-disulﬁde intermediates corresponding to peaks 4 and 5 in Figure 4.11,

respectively. The ﬁ’agments 1-47, itz48-70, and it252-70 were identiﬁed (Figure 4.17A)

171

 

 

 

 

A

it252-70
/ Peak 4
Cys6-Cys47
Cys1 8-Cy661

 

itz48-7O
IA 1 1-47
- _44 2 2L_ ; _x

itz48-70
/ Peak 5
Cys6-Cy352
Cys1 8-Cys61

 

 

1-46 .

I/1-47
n—n—L—JLA—urr ~~- ruLeC A.“

2000 3000 4000 5000 6600 7000 8000
mlz

Figure 4.17. MALDI spectra of peptide mixtures resulting from the
cleavage of puriﬁed and cyanylated two-disulﬁde intermediates
corresponding to peaks 4 and 5 in Figure 4.11, respectively.

The symbol "itz" represents iminothiazolidine derivatives.

 

 

 

 

172

 

 

ﬁ'om the cleavage products of the intermediate represented by peak 4, indicating that
Cys48 and Cys52 are two free sulfhydryl groups; thus Cys6, Cysl8, Cys47, and Cys6]
must be linked together as two disulﬁde bonds. The disulﬁde linkages can be reasonably
deduced as Cys6-Cys47 and CyslS-Cys6l based on the known high stability of the
disulﬁde bond between Cysl 8 and Cys6].

Fragments 1-46, 1-47, and itz48-70 were identiﬁed in the MALDI spectrum (Figure
4.17 B) of the cleavage products of peak 5 in Figure 4.11, indicating cleavage sites at
residues 47 and 48. Because residues 47 and 48 are free sulﬂrydryls in this intermediate,
the linkages of the two-disulﬁde bonds should be between Cys6-Cys52 and Cys18-Cys6l
considering the known high stability of the disulﬁde bond between Cysl8-Cys6].
Therefore, the two-disulﬁde intermediates corresponding to HPLC peak 4 in Figure 4.11
contains the normative disulﬁde structure, Cys6-Cys47 and Cysl 8-Cys61; the
intermediate corresponding to HPLC peak 5 in Figure 4.11 has the normative disulﬁde
structure of Cys6-Cy552 and Cysl 8-Cys61.

The intermediate corresponding to HPLC peak 6 in Figure 4.11 contains one
disulﬁde bond. Figure 4.18 shows the MALDI spectrum of the cleavage products of this
puriﬁed and cyanylated one-disulﬁde intermediate. The fragments itz6-46, 1-46, itz48-
70, itz47-70, and it252-70 were identiﬁed, indicating that four free sulfhydryl groups are
located at positions 6, 47, 48, and 52. Therefore, Cys18 is linked to Cys6l as a disulﬁde
bond.

To summarize, six folding intermediates were captured and characterized by our
methodology (Figure 4.19). The one-disulﬁde intermediate (1) contains disulﬁde bond

between Cysl 8-Cys6 l . Three two-disulﬁde intermediates were identiﬁed. One two-

173

 

/it252-70

 

Peak 6
Cys1 8-Cys61

 

 

 

 

 

a;
2
itz48-70 ~r
‘s‘i
° 47-70 646
7 .3\ 7
‘L ..A Mus
2000 3000 4000 5000 6000 7000 8000
mlz

Figure 4.18. The MALDI spectrum of cleavage products of puriﬁed/cyanylated
intermediate corresponding to HPLC peak 6 in Figure 4.11. The symbol

"itz" represents iminothiazolidine derivatives.

174

 

1

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

. I I I'_I_l l
NativelGF-l 1 6 18 4748 52 61 70
SH SIH SH SH
I I 1
One-disulﬁde 1 g I I I I 70
intermediate (I) 6 18 47 48 52 61
7 SH SH
Native two-disulﬁde I ' ' ‘
. . 1 l l | | I 70
intermediate (HA) 6 18 4748 52 61
c i“ i”
Nonnative two-disulﬁde l 4| l I I I
intermediates ("3) 1 6 18 47 48 5-2 61 70
SH SH
Normative two-disulﬁde I I J i i 1
Intermediates (lie) 1 6 18 4., 48 52 61 70
Normative three-disulﬁde
interrnediate(lll) 1 6 18 47 48' 52' 61 70

Figure 4.19. The disulﬁde structure of refolding intermediates of lGF-l.

175

disulﬁde intermediate (11A) contains native disulﬁde structure Cys6-Cys48 and Cysl8-
Cys6l. The other two two-disulﬁde intermediates have nonnative disulﬁde structures,
Cys6-Cys47 and Cys18-Cys61 (113); Cys6-Cys52 and Cysl8-Cys6] (He)- The nonnative
three-disulﬁde intermediate (111) contains disulﬁdes Cys6-Cys47, Cys48-Cy352, and

Cysl 8-Cys6l . A mixed disulﬁde intermediate was also identiﬁed.

V. Isolation and Refolding of the Folding Intermediates of LR’IGF—I

Additional detailed information on the refolding of a protein can be obtained ﬁom a
stop/go experiment. The folding intermediates are isolated and puriﬁed at pH 3, and then
subjected to a second-stage refolding by increasing the pH to 8.7 in the presence of
GSSG/GSH.

Figure 4.20 shows the distribution of the intermediates after 30-minutes of
continued refolding starting from the puriﬁed two-disulﬁde intermediate (11) or the
puriﬁed three-disulﬁde intermediate (III) of LR’IGF-I. The three-disulﬁde intermediate
(III) can convert to the two-disulﬁde intermediate (11) and native LR3IGF-I. The
distribution of species after 30 minutes of second-stage refolding is similar to the
equilibrium distribution shown in Figure 4.3A, especially for the continued refolding of
the puriﬁed two-disulﬁde intermediate (11) as shown in Figure 4.20 (top panel).
Continued refolding of the puriﬁed two-disulﬁde intermediate (11) gave a similar result to
that of the puriﬁed three-disulﬁde intermediate (III) as shown in Figure 4.20. These
results imply that the three species (11, III, and N) can interconvert to each other by
breakage/reformation of disulﬁde bonds.

176

 

Starting with
2-disulﬁde intermediate

III+Mix
\

 

Starting with
3-disu1ﬁde intermediate

 

 

 

 

10 20 3'0 40

' Minutes
Figure 4.20. HPLC separation of CDAP-trapped intermediates
after 30-minutes of refolding starting from the puriﬁed 2-
disulﬁde intermediate (II) or 3-disulﬁde intermediate (HI) of
LR3IGF-I in the same refolding buffer as used in the refolding of
reduced/unfolded LR3IGF-I.

177

vr. Comparison of the Folding Process of LR’IGF-I and IGF-1

LR3IGF-I is an analog of IGF-I with the differences at N-terminal side. There are
several different aspects between the folding processes of IGF-I and LR3IGF-I.

1). Fewer intermediates in the folding pathway of LRslGF-I were identiﬁed than in
the folding pathway of IGF-I. The extension of 13 amino acids at N-terminus of LR3IGF-
I might result in the reducing the options of forming different disulﬁde bonds due to the
lower ﬂexibility in LR3IGF-I.

2). In the folding process of IGF-I, two nonnative two-disulﬁde intermediates were
detected. However, there were no detectable nonnative two-disulﬁde intermediates in the
folding or reductive unfolding process of LR3IGF-I.

3). The disulﬁde structures of the normative three-disulﬁde intermediate, the native
two-disulﬁde intermediates, and the one-disulﬁde intermediate of LR’IGF-I and IGF-1
have similar disulﬁde-linkage patterns.

4). The folding process of IGF-I is much faster than that of LR3IGF-I in the early
folding stages, which may be due to the lower ﬂexibility of LR3IGF-I since the protein is

a more extended random coil at the early folding stage.

VII. Conclusions

The folding and reductive unfolding intermediates of LR3IGF-I were captured and
characterized by cyanylation methodology, which include a native one-disulﬁde
intermediate, a native two-disulﬁde intermediate, and a normative three-disulﬁde
intermediate. The pathways of refolding and reductive unfolding of LR3IGF-I can be

proposed on the basis of the time courses of the refolding and unfolding (Figure 4.3A,

178

4.3B and 4.4) and on the structure of disulﬁde bonds of the intermediates. The proposed
pathways are illustrated in Figure 4.21.

In the pathway of the refolding of LR3IGF-I, the reduced/unfolded LR3IGF-I was
converted to one-disulﬁde intermediate (I) with the formation of the ﬁrst disulﬁde bond,
31-74. Then intermediate I was converted to two-disulﬁde intermediate (11) with the
formation of the second disulﬁde bond, 19-61. Three-disulﬁde intermediate (111) can be
formed ﬁom the two-disulﬁde intermediate 11 based on the results of the stop/go
experiment starting with the two-disulﬁde intermediate 11 (Figure 4.20). Some transient
nonnative two-disulﬁde intermediates should exist in the pathway of two-disulﬁde
intermediate 11 to three-disulﬁde intermediate 111 since two-disulﬁde intermediate 11
cannot convert to the three-disulﬁde intermediate 111 without the breakage/reformation of
disulﬁde bonds. However, no nonnative two-disulﬁde intermediates were trapped even at
an early stage of the unfolding because of the apparent short lifetime of the intermediates.
Alternatively, it is possible that the three-disulﬁde intermediate III is formed by the
pathway of the one-disulﬁde intermediate (1) to nonnative two-disulﬁde intermediates,
and then to the three-disulﬁde intermediate (Figure 4.21).

Further experiments should be conducted to answer whether only one pathway
(two-disulﬁde intermediate to three-disulﬁde intermediate) or two pathways (either the
two—disulﬁde intermediate (11) or the one-disulﬁde intermediate (I) to the three-disulﬁde
intermediate (111)) exist. Unfortunately, these two pathways are unlikely to be
distinguished by using our folding methodology (stop/go) since the one-disulﬁde
intermediate (1) is the precursor of the two-disulﬁde intermediate (11). Therefore, we

cannot determine whether the three-disulﬁde intermediate (III) is formed from the

179

SIH S'H HSI .T.H SH SIH
1 ------- c ------- c -------- cc -------- t': -------- c----83
19 31 6061 65 74
§H ,
,_ 19 I 60 1 65 I
_. 1 ----—--c--------c-------cc -------- c ------- C----83
| 31 | i 74
SH SH H
19 ' 60 65 7'
31 61 4
1 ------ c-----c ------ off-mm- ------c---83 S
' I
SH SH

 

 

 

 

 

 

17

(31-74, 61—65) or(and) (31-74, 19-60)

11

 

 

 

I I
19 31 60 61 65 74
1------(l:---—c-------<l: -------c|:------c----83
ll'(?) or 1(7)
I i
19 31 60 61 65 74
1 -------CI--“---C--m--.C? ------ CI -------- C----83

 

 

 

Figure 4.21. The predicted folding pathway of LR3IGF-I

180

 

 

<—

pathway of the two-disulﬁde intermediate (11) to the nonnative two-disulﬁde intermediate
or from one-disulﬁde intermediate (1) to the normative two-disulﬁde intermediate. The
technique of mutating Cys60 and Cys65 (such as mutate the cysteines to alanines) may be
used to solve the problem because the two-disulﬁde intermediate (II) will not formed
from the one—disulﬁde intermediate (1) after the mutation, and the only precursor of the
three-disulﬁde intermediate (III) would be the one-disulﬁde intermediate (1).

The folded/native LR3IGF-I can ﬁnally be formed by two pathways (Figure 4.21).
One pathway between the three-disulﬁde intermediate (III) and native LR’IGF-I can be
reasonably suggested according to the time course of the refolding and unfolding of
LR3IGF-I. In this pathway, some transient two- or one-disulﬁde intermediates (1’ and/or
11”) chould be involved through the breakage/reformation of disulﬁde bonds. It is also
possible that an alternative pathway between the two-disulﬁde intermediate (11) and
native LR3IGF-I may exist based on two reasons. Firstly, the two-disulﬁde intermediate
(11) already contains two native disulﬁde bonds, one can reasonably suggest that
intermediate 11 could convert to native protein with the formation of the third disulﬁde
bond, 61-65. Secondly, increasing the oxidized glutathione concentration in the
equilibrium mixture of refolding species (previously established with lower GSSG
concentration) resulted in all the intermediates (including intermediate 11 and
intermediate 111) being converted to native protein (Figure 4.33) also suggests the
existence of this pathway. The possible conversion of two-disulﬁde intermediate II to
native protein will increase at higher concentration of GSSG so that the equilibrium will
move to the formation of native protein from the pathway of the three-disulﬁde

intermediate III to two-disulﬁde intermediate II, ﬁnally to native protein.

181

 

By comparison, the folding intermediates of IGF-I were also identiﬁed, which
include a native one-disulﬁde intermediate (I), a native two-disulﬁde intermediate (11A),
two nonnative two-disulﬁde intermediates (113 and He), and a normative three-disulﬁde
intermediate (III). The disulﬁde structures of the native intermediates contain the similar
linkage patterns in LR3IGF-I and lGF-I. Two nonnative two-disulﬁde intermediates were
identiﬁed in the folding process of IGF-I, but not in the refolding of LR3IGF-I. The
folding of IGF-I was much faster than that of LR3IGF-I in the early folding stage.

A folding pathway of IGF-I can be suggested based on the time course of the
refolding and the disulﬁde structure of the intermediates (Figure 4.22). The one—disulﬁde
intermediate is formed from the reduced/unfolded lGF-I by the formation of the disulﬁde
bond between 18 and 61. The two-disulﬁde intermediate HA is formed from the one-
disulﬁde intermediate by the formation of the second disulﬁde bond between 6 and 48.
The two nonnative two-dimrlﬁde intermediates (113 and He) can be formed from the one-
disulﬁde intermediate by the formation of the disulﬁde bond between 6 and 47 (as in 113)
or between 6 and 52 (as in He). Alternatively, it is possible that the two nonnative two-
disulﬁde intermediates could be formed from the two-disulﬁde intermediate (11A) by
breakage/reformation of disulﬁde bonds.

The nonnative three-disulﬁde intermediate could be formed from the nonnative
two-disulﬁde intermediate (113) by the formation of the disulﬁde bond between 48 and
52. Finally, native IGF-I could be formed either from the native two-disulﬁde
intermediate (IIA) by the formation of the third disulﬁde bond between 47 and 52 or from

the normative three-disulﬁde intermediate (III) by the breakage/reformation of disulﬁde
bonds.

182

 

?H .?H ?H iH

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 4.22. The predicted folding pathway of lGF-l.

183

47 46 52 61 7°
SH H SHSH SH
i l l i i i
F3 6 16 4746 52 61 7°
1 l l | l l ﬁ‘ 70 ,
6 16 4746 52 61 ‘—
3 H W f“
l i 1 l | l
1 6 1‘6 47 46 5'2 61 7°
__ +
—> SH SH
I F i i i l
1 6 16 H 47 46 52 61 7°
W
l T l l‘_l
1 6 16 47 46 52 61 7°
. l l .—l—. l 7....
6 16 47 46 52 61

 

It should be pointed out that the folding pathways of LR3IGF-I and IGF-1 described
in Figure 4.21 and Figure 4.22 are predicted only based on the time dependent
appearance and disulﬁde structure of the folding intermediates. Two problems must be
solved before the actual pathway can be deﬁned. Firstly, the well-populated species are
not necessarily the productive species that account for the ﬂow of intermediates. One way
to identify the productive intermediates is to perform stop/go folding experiments of all
species of the intermediates, both major and minor, that present at the same stage of
equilibrium. The dilemma faced by this approach is that productive intermediates may
exist as minor species. Finding all these minor species will be a daunting task. The
predicament can be further complicated by the argument that an undetected species is not
necessarily a non-existing species. Secondly, protein folding may occur through multiple
pathways and these pathways may not be distinguishable. The other techniques such as
mutation of the cysteine residues might solve the problem.

The trapping methodology, based on the cyanylation of he thiol groups under
acidic conditions, showed similar results to those of acid trapping (31-33), but different
ﬁom those of pyridylethylation trapping at alkaline condition (29) in terms of the
distribution of intermediates during the folding processes of LR’IGF-I and IGF-1. The
methodology for the determination of disulﬁde structure by chemical cleavage and mass
mapping of the fragments can be used to determine the disulﬁde linkages of the folding
intermediates containing adjacent cysteines. The direct experimental evidences for the
disulﬁde linkages of the nonnative three-disulﬁde intermediates of LR’IGF-I and IGF-1

were obtained for the ﬁrst time by our methodology, which cannot be solved by the

184

conventional approach. Furthermore, the cyanylation methodology is much faster and

sensitive.

185

VIII. References

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

Kim P. S., Baldwin R. L. (1990) Annu. Rev. Biochem. 59, 631-660.

Matthews CR. (1993) Annu. Rev. Biochem. 62, 653-683.

Ptitsyn QB. (1995) Curr. Opin. Struct. Biol. 5, 74-78.

Li Y.J., Rothwarf D.M., Scheraga H.A. (1995) Nature Struc. Biol. 2, 489-494.
Creighton, T.E. (1974). J. Mol. Biol. 87, 563-577.

Creighton, T.E., Goldenberg, D.P. (1984) J. Mol. Biol. 179, 497-526.
Scheraga, H.A., Konishi, Y., Ooi, T. (1984) Adv. Biophys. 18, 21-41.
Weissman, J.S., Kim, RS. (1991) Science 253, 1386-1393.

Creighton, T.E., Ewbank 1.]. (1994) Biochemistry 33, 1534-1538.

Chang, J .Y. (1996) Biochemistry 35, 11702-11709.

Glocker M.O., Arbogast B., Milley R., Cowgill C., Deinzer M.L. ( 1994) Proc. Natl.

Acad. Sci. USA 91, 5868-5872.

Darby, N.J., Norin P.E., Talbo, G., Creighton, T.E. (1995) J. Mol. Biol. 249, 463-
477.

Weissman, J.S., Kim, RS. (1992) Proc. Natl. Acad Sci. USA 89, 9900-9904.
Xu, X., Rothwarf, D.M., Scheraga, H.A. (1996) Biochemistry 35, 6406-6417.
Rothwarf, D.M., Scheraga, H.A. (19936) Biochemistry 32, 2671-2679.
Rothwarf, D.M., Scheraga, H.A. (1993b) Biochemistry 32, 2680-2689.
Rothwarf, D.M., Scheraga, H.A. (1993c) Biochemistry 32, 2690-2697.
Rothwarf, D.M., Scheraga, H.A. (1993d) Biochemistry 32, 2698-2703.
Rothwarf, D. M., Scheraga, H. A. (1991) J. Am. Chem. Soc. 113, 6293-6294.

Creighton, T.E. (1992) Bioessays 14, 195.

186

 

 

 

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

Pain, RH. (1994). In mechanisms of protein folding. Oxford University Press, New
York, p112.
Wu, J., Yang, Y., Watson, J .T. (1998) Accepted by Protein Science.
Smith, D.L., Zhou, Z.R. (1990) Methods Enzymol. 193, 374-389.
Wu, J., Watson, J.T. (1997) Protein Sci. 6, 391-398.
Yang, Y., Wu, J ., Watson, J .T. (1998) Accepted by J. Am. Chem. Soc..
Rinderknetcht, B., and Humbel, R. B., J. Biol. Chem., 253, 2769 (1978).
Froesch, E. R., Burgi, H., Ramseier, E. B., Bally, P., and Labhart, A., J. Clin.
Invest., 42, 1816 (1963).
Rinderknecht, B., and Humbel, R. B., (1976) Proc. Natl. Acad Sci., USA, 73, 2365.
Hober, S., Forsberg, G., Palm, G., Hartmanis, M., and Nilsson, B. (1992)
Biochemistry, 31, 1749-1756.
Hober, S., Forsberg, G., Palm, G., Hartmanis, M., and Nilsson, B. (1992)
Biochemistry, 31, 1749-1756.
Milner, S.J., Francis, G.L., Wallace, J.C., Magee, B.A., Ballard, F.J. (1995)
Biochem J. 308, 865-871.
Hua, Q.X., Narhi, L., Jia W., Arakawa, T., Rosenfeld, R., Hawkins, N, Miller, J .A.,
Weiss, M. (1996) J. Mol. Biol. 259, 297-313.
Miller, J.A., Narhi, L.O., Hua, Q.X, Rosenfeld, R., Arakawa, T., Rohde, S.P.,

Lauren, S., Stoney, K.S., Tsai, L., Weiss, M.A. (1993) Biochemistry 32, 5203-5213.

34. Tomas, F.M., Knowles, S.E., Chandler, C.S., Francis, G.L., Owens, P.C., Ballard,

F.J. (1995) J. Endocrinology 137, 413-421.

187

 

 

"‘llllllllllllllllllll“