III I I ‘ '7
{III II,I.I.«1III|KI.I" ”erg; ”“1323?“
3 I . . 1' “It?" 3:3”1)’ 3:1}u‘1‘iKIII" ,7 33 39‘“ '93‘11{£ "Ed
I‘fml‘w‘ ””9 .,.c:.§'I5!I' I“? “In? 1;“ 43»: MW- I
33-1‘ I34” 31:“. ‘1“); 1":9 . “331;. “955' 5 1 .
.. IIWIIw
flTwI~¢I‘V%,x
III ‘I‘III II I
III 23¢“ 533$ I'm-i KI MS? .I 113w ‘3
1121;53:13- II?

"t' 393;.

pl
‘, I. It!
3 , '5‘I'III13‘IIMIIIII’W‘II‘!
I:;;~ ,
D

‘J
-«~

    
 
    
 
  
    
  
 

 

is: gLn
s‘IK;y '16:“ I”?! 13.3“
IN IIKIEI
“A; ”.23 III: III-3
3.1133
”H.134” {g}; .;

L ‘ "sulftu‘lflhu
i...‘ " ‘ Ivfoni‘flnjfi

I III:
III

  

u

    
  
   
        
   
  
  
 

  
 
  

 

 

     

|’:6‘

‘ 13%erng

   
 

   
 
  
 

i, 1‘ ' .1"

l'
1

  
 

  

7:17;?

      

       
 
      
 
 
  
   
   
  
 

   

  
 
  
  
  
  
 
   
    

 
 
   

   
         
       
  

      
  

     
 

    
 

 
  
   
    
  
  
   
 
 
 
   
 

  
   
   
 
 
  
  
 
 
 

  
 
  
 
   
   
 
 
 
 
 
  
   

   

 

  

“v.3" z" 'I\
fit}?! 332' 233-33.”: [1313133332
333331;:11399353‘ 3 II‘EI’II'
“’14?" I I33“: I33; IVIII‘}
‘IIIIKIIII...IIII .fIIgI "IIKIIEISII EIIII
I-IIII “III 2
I ' IIIIKIIIIII {II 1.8 II“ .
firs'aim‘i “$1333; 3 3‘s 3 will
3M“ LIN J .:IIII'I?I‘II '.'a. ~
‘I'iII-I-Itgh “‘3‘ SH‘I'II’ ”Egg-3'33 - ’ ‘
. .' 33 3 133'1- 5333 3? if“' 353' 33% 3:3],33E 3H ,; ;;£;§;!§;l,:; : 3 5 . : . I_, :J& .
'3'“ "W . 7135!: II“ III! -. 3.: 433,153“ I
3 @3319“! II: 1?: :II’IIKIQ IIIEKII; - If ' .I

    

.I’ . .. Wig,
» .. I-IIIII»: - an.
‘ . 13mph?'K33:=:—II:I:‘=II..2-‘ .-
IIIIII‘IWI

‘ ii. ‘
“$3533 I!
, .

..IIIII
E‘; 1‘ IIIIIIIIIII’ I ’1

.. I.

’-.,~
'* my...

433.;
., .I
v .N
U:
K I

   
 

It; EIIIWI ifgk. II
II

I.

3.7% I I
III;
3:? ' . 2‘ 1

 
 
 
 

’5'.

I . - K
3‘ “‘53.." I
F. ‘ - "-7.;
. :nlaW»—.y:g

   

      

   

 
  

    

 
 
        
    

E'II'I'.
dfl’g‘th I , “3 'I
IMIIII

“3333.3113“, .K'If

'I
III?

     
   
  
  

  

 
  

   
     
 

 

.r. ;
6 II‘ L
“V."‘§','<;_‘|Ifx
’I’liilf .5
,. .., I ‘1 l? l‘.
I -.

L“;.I;‘(3,i,i I" (

   

       
   

‘: III I”; ',.I
34‘“; 3 311'.)
Ii IK;II~I 3!
II'KKlII‘.”
.534'9'233311',3§

‘03.
3”
33: I5: MIN ' 3:".
' ’[V'

IIII-u'
I'T'i

   
 
   

          
  
  
 
 
 
 
 

  
  
 

”1’ .'- ”I . {,3

‘r‘ ‘m " ,

”PS' ;$"

{av—2L hf gétxi pWLL' "3";
$1); .

  

   
 
  

n~.—.

III “I

   
 

-.
1., .

. ..
. _

 
 
   
     
 
   

£0-
":5.
Qw’h;
«mlr...
IW‘D’“ -
”Mu.
.o_._~—_.-.'D’“
..
“.1 “m"?
"‘6'".-
'w—J
Fw...’..":..r~
-“w

o

v. u...—

M

u I"‘ m- .-....
‘A

1 “lion—o
. ..
. moi-v»...
”Ma
' I—n'--- c . '.
‘ 1......"
' » ‘ pl
- ’52?
' -4.-,
. q.
.-
Ow. “‘1‘
~ ._;_13,_.
.,

a":

..

'. mm
3.

 
 

.4.
~
m..-
V

wfl-fi-

  

ow.» '1‘"
- g u..-

U I.: A‘

$322

was

2900

This is to certify that the

dissertation entitled

ELECTROSPRAY IONIZATION -- COLLISIONALLY ACTIVATED
DISSOCIATION MASS SPECTROMETRIC ANALYSIS OF
CHARGED DERIVATIVES 0F PEPTIDES

presented by

Nalini Sadagopan

has been accepted towards fulfillment
of the requirements for

Ph . D . degree in Chemig try

 

 

Major professor

DateM M69) WW

MS U is an Affirmative Action/Equal Opportunity Institution 0-12771

 

LIBRARY
Michigan State
University

 

 

 

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

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 a/CIRCIDateDuepss-ms

ELECTROSPRAY IONIZATION - COLLISIONALLY ACTIVATED DISSOCIATION
MASS SPECTROMETRIC ANALYSIS OF CHARGED DERIVATIVES OF
PEPTIDES

By

Nalini Sadagopan

A DISSERTATION

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

DOCTOR OF PHILOSOPHY
Department of Chemistry

1999

ABSTRACT

ELECTROSPRAY IONIZATION - COLLISIONALLY ACTIVATED DISSOCIATION
MASS SPECTROMETRIC ANALYSIS OF CHARGED DERIVATIVES OF
PEPTIDES

By

Nalini Sadagopan

Collisionally activated dissociation (CAD) tandem mass spectrometry (MS/MS) is
a useful tool in obtaining protein sequence information. CAD of peptides often yields a
mixture of fragment ions retaining either the N-terminus or C-tenninus; a further
complication usually results from the lack of a continuous series of ions. In contrast,
charged derivatives of peptides, where the charge is fixed on one terminus, yield a series
of ions retaining only one terminus and hence simpler CAD mass spectra.

This dissertation describes electrospray ionization (ESI)-CAD-MS/MS analysis of
peptides derivatized with [tris(2,4,6-tn°methoxyphenyl)phosphonium] acetyl (TMPP+—
Ac), to provide a fixed charge at their N-terminus; ESI-CAD-MS/MS is readily
interfacable with separation techniques like liquid chromatography. Based on the results
obtained from model peptide derivatives, ESI-in-source-fragmentation (ISF)-MS and
ESI-CAD-MS/MS analyses of peptides after TMPP+-Ac-derivatization give a continuous
and simple series of 'afll and wbn ions. Most amino acid residues fragment to form 'an and
'bn ions, however, some specific amino acid residues (e. g., proline, aspartic acid,
asparagine etc.) either do not fragment or fragment to give additional ions, which could

disrupt the continuity in the series. Recognition of these characteristic patterns is useful

when an unknown sequence is to be determined using this methodology. The application
of this derivatizing procedure to tryptic digests of proteins as analyzed by LC-MS is
presented.

The impact of charge-derivatization on methods for determining post-translational
modification such as phosphorylation is described. Results from studying model
phosphOpeptides afier charge—derivatization show simplification of the spectral series.
Phosphoserine and phosphothreonine lose H3PO4 predominantly, while phosphotyrosine
predominantly loses HPO3 when present in the charge-derivatized peptide. Similar results
are also observed in underivatized phosphopeptides.

Studies on the fragmentation mechanisms of charged derivatives of peptides using
deuterium labeling by MALDI-PSD-MS and ESI-CAD-MS/MS show that the migration
of the amide hydrogen via a 1,2 elimination process is the major fragmentation pathway
in the formation of the *an and 'bn ions. However, fragmentation at the site of the proline
residue to yield '11.. during the ESIsCAD-MS/MS studies is possibly due to a B-hydrogen

shifi.

Dedicated in loving memory of my Dad .........

ACKNOWLEDGMENTS

I would like to express my deepest gratitude to Prof. J .Throck Watson. He was a
great mentor and inspired me in many ways. He helped me think through problems
critically, and fundamentally.

Thanks to Prof. Allison my second reader, who read through my thesis and
provided critical insight. I really enjoyed working with him as a TA during my stay at
graduate school.

I also would like to acknowledge the other members of my committee, Professors
Stanley Crouch, Rawle Hollingsworth and Ned Jackson of the Chemistry department.

Many thanks to Charles Ngowe, Jennifer Smith and Jianfeng Qi for their
friendship during my stay at MSU.

Words cannot describe how thankful I am to my sweet daughter Shilpa. I am so
happy that I was able to have her during my graduate studies. Rishi, my husband, has
been so supportive and encouraging during my studies despite his busy schedule. I could
not have come this far without their cooperation.

I would also like to thank my parents-in-law, mother, sister and brother for their

constant encouragement.

TABLE OF CONTENTS

LIST OF TABLES ................................................................................... ix
LIST OF FIGURES .................................................................................. x

CHAPTER 1: INTRODUCTION: ELECTROSPRAY IONIZATION TANDEM MASS
SPECTROMETRY AND ITS USE IN THE ANALYSIS OF CHARGED

DERIVATIVES OF PEPTIDES .................................................................... 1
1. Preface .............................................................................................. 1
II. Analytical Techniques and Chemical Procedures ............................................ 1
A) Desorption/Ionization (D/I) Techniques ................................................... 1

B) Electrospray Ionization ...................................................................... 4

a) Charged-residue Mechanism ........................................................... 5

b) Ion Evaporation Mechanism ............................................................ 6

C) Collisionally Activated Dissociation Tandem Mass Spectrometry .................... 8
D) Triple Quadrupole and Ion Trap Mass Spectrometers ................................. 10

E) Charged Derivatives for Analysis by Mass Spectrometry ............................. 14

a) Quaternary Ammonium Derivatives ................................................. 17

b) Quaternary Phosphonium Derivatives .............................................. 21

c) N-terminal Derivatives With High Proton Affmities ............................. 25

d) The Ideal Charged Derivative ....................................................... 27

III. Conclusions ..................................................................................... 31
IV. References ...................................................................................... 32

CHAPTER 2: ANALYSIS OF TRIS(2,4,6-TRIMETHOXYPHENYL)PHOSPHONIU M
ACETYL DERIVATIVES OF PEPTIDES BY ELECTROSPRAY IONIZATION

FOLLOWED BY QUADRUPOLE OR ION TRAP MASS SPECTROMETRY ......... 37
I. Introduction ....................................................................................... 37
A) Role of Mass Spectrometry in Protein Sequencing ................................ 37

B) Nomenclature of Fragment Ions Formed From Collisionally Activated
Dissociation .............................................................................. 38
C) FAB-CAD-MS and MALDI-PSD-MS Studies of TMPP+-Ac-Peptides. . . . . ...39
11. Experimental .................................................................................... 42
A) Preparation of Charged Derivatives ................................................... 42
B) Mass Spectrometry ....................................................................... 44
a) In-source Fragmentation (ISF) MS ................................................ 44
b) MS/MS ................................................................................. 44
c) MALDI-MS ........................................................................... 45
III. Results and Discussions ....................................................................... 47
A) In-source Fragmentation of the TMPP+-Ac Derivatives of Peptides .............. 47
B) MS/MS of the TMPP+-Ac Derivatives of Peptides ................................. 48
a) Triple Quadrupole Mass Spectrometer ........................................... 48
b) Ion Trap Mass Spectrometer ....................................................... 50

vi

C) Comparison of the Fragmentation of the TMPP+-Ac Derivatives of Peptides by

ISF vs. MS/MS ........................................................................... 55

D) Comparison of the Fragmentation of the TMPP+-Ac Derivatives of Peptides in
Triple Quadrupole vs. Ion Trap ........................................................ 57
E) Speculation of the Fragmentation Pathways of TMPP+-Ac-Peptides ............. 59
F) Characteristic Fragmentation Behavior of Some Amino Acid Residues ......... 61

G) Comparison of the Fragmentation of the TMPPI-Ac Derivatives of Peptides by
ESI-ISF vs. MALDI-PSD and F AB-CAD ............................................ 74
H) ESI-CAD-MS/MS Study of TMPP+-Ac Derivatized Tryptic Digests ............ 75
IV .Conclusions ....................................................................................... 86
V. References ........................................................................................ 87

CHAPTER 3: EFFECT OF CHARGE-DERIVATIZATION IN THE
DETERMINATION OF PHOSPHORYLATION SITES IN PEPTIDES BY ESI-CAD-

MS/MS ................................................................................................ 90
I. Introduction ....................................................................................... 90
II. Experimental .................................................................................... 97
III. Results ........................................................................................... 99
A) Fragmentation of Phosphorylated Tyrosine Residue Containing Peptide
Derivatives by CAD-MS/MS ........................................................ 99

B) Fragmentation of Phosphorylated Serine Residue Containing Peptide
Derivative by CAD-MS/MS ....................................................... 105

C) Fragmentation of Phosphorylated Threoine Residue Containing Peptide
Derivative by CAD-MS/MS ....................................................... 107

D) Fragmentation of Doubly Phosphorylated Peptide Derivative Containing
Threonine and Tyrosine Residues by CAD-MS/MS ........................... 109

E) In-Source-CAD-MS of Phosphopetide Derivatives in an Ion Trap Mass
Spectrometer ......................................................................... 1 1 1

F) Multiply Phosphorylated Peptides From B-casein .............................. 115

IV. Conclusions .................................................................................... 118
V. References ...................................................................................... 120
CHAPTER 4: MECHANISMS OF FRAGMENTATION OF PEPTIDES AFTER
CHARGE-DERIVATIZATION BY ESI-CAD-MS/MS AND MALDI-PSD-MS ....... 123
I. Introduction ..................................................................................... 123
A) Collisionally Activated Dissociation of Protonated Peptides ........................ 123
B) Fragmentation of Charged Derivatives .................................................. 128
C) Current Studies ............................................................................. 132
II.Experimental .................................................................................... 133
A) Peptides Synthesis ......................................................................... 133
B) Derivative Synthesis ....................................................................... 135
C) MALDI-MS ................................................................................. 135
D) ESI-CAD-MS/MS .......................................................................... 136
III. Results .......................................................................................... 137
A) Analysis of the unlabeled TI\/IPP+-Ac-VGVAPG ..................................... 137
B) Results From Peptides with Deuterium Label at the a- Carbon .................... 142

vii

a) Deuterium Label at the a- Carbon of Glycine ...................................... 142

b) Deuterium Label at the a- Carbon of Valine ....................................... 147
C) Results From Peptides with Deuterium Label at the B- Carbons .................. 149
a) Deuterium Label at the [3- Carbon of Valine ...................................... 149

b) Deuterium Label at the [3- Carbon of Alanine .................................... 150

D) Results From Peptides with Deuterium Label on the Amide Nitro gens. . . . . . . . ....151
a) Mechanism of Formation of the an ion due to the Shift of an Amide

Hydrogen ............................................................................... 155

b) Mechanism for Formation of the (.a6+2) Ion Due to the Shifi of C-terrninal
Hydrogen Atom ....................................................................... 156

c) Mechanism for Formation of the b, Ion Due to the Shift of an Amide

Hydrogen ............................................................................... 157

d) Mechanism for Formation of the (.bn+H20) Ion From the ‘bn Ion ............. 159

e) Mechanism for Loss of the C-terminal Residue .................................. 160

E) Genesis of .an and ''b.. ions due to Cleavage at the Site of the Proline Residue...162
V.Conclusions ..................................................................................... 1 67
IV.References ...................................................................................... 1 68

viii

LIST OF TABLES

Table 2.1 Peptides Studied by ESI-CAD-MS After TMPP+-Ac Derivatization ............ 42
Table 2.2 Proteins Studies by LC-MS After Tryptic Digestion and Derivatization ........ 43
Table 3.1 List of Phosphopeptides Used in the Study. . . . . . . . . . . ............................... 97
Table 4.1 List of Labeled Peptides .............................................................. 133
Table 4.2 Ions Observed for d2-Tl\4PP+-Ac-VG°‘VAPG ..................................... 146
Table 4.3 Ions Observed for dz-TMPPI-Ac-VanaAPG .................................... 148
Table 4.4 Ions Observed for d2-TMPP+-Ac-VBGVBAPG .................................... 149
Table 4.5 Ions Observed for d3-TMPPI-Ac-VGVABPG ...................................... 151
Table 4.6 Ions Observed for db-TNIPP+-Ac-V"G“V"A"PG" ........................................... 154

Table 4.7 List of m/z values for .as and 'b5 Ions Observed From the Labeled Variants of
TMPP+-Ac-VGVAPG ............................................................................ 162

LIST OF FIGURES

Figure 1.1 The ESI source ........................................................................... 5
Figure 1.2 Mechanism of ion formation in ESI as proposed by Dole ......................... 6
Figure 1.3 The components of a quadrupole mass spectrometer .............................. 12
Figure 1.4 The ion trap mass spectrometer ...................................................... 13
Figure 1.5 Trimethylammonium derivatiation scheme .......................................... 17
Figure 1.6 Trialkylammonium-acetyl derivatization scheme ................................. 17
Figure 1.7 Quaternary ammonium derivatization using thiocholine iodide ................. 18
Figure 1.8 C5Q derivatization reaction ........................................................... 20
Figure 1.9 Triphenylphosphonium ethyl derivatization reaction .............................. 22
Figure 1.10 TMPPI-Ac derivatization scheme .................................................. 24
Figure 1.11 Reaction schemes of Renner and Spiteller ........................................ 26
Figure 1.12 SPA derivatization reaction ......................................................... 26
Figure 2.1 Nomenclature for peptide fragmentation during CAD ............................ 39
Figure 2.2 Reaction of TMPPI-Ac reagent with the N-terminus of a peptide ............... 40

Figure 2.3 ESI-ISF mass spectrum of TMPPI-Ac-MNYLAFPRM-amide. The inset
shows the portion containing peaks due to fragment ions from cleavage at the
asparagine (N) residue .............................................................................. 49

Figure 2.4 ESI-CAD-MS/MS spectrum of TMPPI-Ac-GGSLYSFGL-amide, doubly-
charged precursor at m/z 736 in a triple quadrupole mass spectrometer at a collision
energy of 65 eV ...................................................................................... 51

Figure 2.5 ESI-CAD-MS/MS spectrum of singly-charged precursor of
GNLWATGHFM-amide in an ion trap mass spectrometer at 30% relative collision
energy ................................................................................................. 52

Figure 2.6 ESI-CAD-MS/MS spectrum of singly-charged precursor of TMPPI-Ac-
GNLWATGHFM-amide in an ion trap mass spectrometer at a relative collision energy of
60% ................................................................................................... 53

Figure 2.7 ESI-CAD-MS/MS spectrum of doubly-charged precursor of TMPP+-Ac-
GNLWATGHFM-amide in an ion trap mass spectrometer at a relative collision energy of
35% ................................................................................................... 54

Figure 2.8 A comparison of the fragmentation of TMPP+-Ac-GGSLYSFGL-amide by A)
In-source fragmentation, B) CAD-MS/MS in a triple quadrupole (doubly-charged
precursor) C) CAD-MS/MS in an ion trap (singly-charged precursor) mass
spectrometers ........................................................................................ 56

Figure 2.9 In-source fragmentation of TMPP+-Ac-GGSLYSFGL-amide followed by m/z
analysis of fragment ions in an ion trap mass spectrometer ................................... 58

Figure 2.10 ESI-source-CAD mass spectrum of TMPP+-Ac-VGVAPG in a single-
quadrupole mass spectrometer. Spectrum shows a series of an ions, except for cleavage

at the site of proline ('as) ........................................................................... 62
Figure 2.11 Structure of proline (P) residue .................................................... 63
Figure 2.12 Structure of threonine (T) residue ................................................. 64
Figure 2.13 Structure of aspartic acid (D) residue ............................................. 64
Figure 2.14a Mechanism for the formation of b. ion due to fragmentation at D .......... 65
Figure 2.14b Mechanism for the formation of c, ion due to fi'agmentation at D .......... 65
Figure 2.14c Mechanism for the formation of ‘an ion due to fragmentation at D. . . . . ....66
Figure 2.15 Portion of the ESI-ISF- mass spectrum of TMPPI-Ac-APGDIRYVHPF,
showing the fragmentation pattern of aspartic acid residue (D) .............................. 67
Figure 2.16 Structure of asparagine (N) residue ................................................ 68
Figure 2.17 Structure of glutamine (Q) residue ................................................. 69

Figure 2.18 Portion ofthe ESI-ISF-MS spectrum of TMPP+-Ap-APSGAQRLYGFGL-
amide. Fragmentation at the glutamine residue (Q) results in a on-1ion. Fragment ions
containing the arginine residue (R) show loss of ammonia ................................... 70

Figure 2.19 Structure of histidine (H) residue .................................................. 71

xi

Figure 2.20 ESI-ISF mass spectrum of TMPPI-Ac-VHLTPVEK showing characteristic
fragmentation of H residue and T residue at positions 2 and 4 respectively ................ 72

Figure 2.21 Structure of arginine (H) residue ................................................... 73

Figure 2.22 A comparison of the fragmentation of TMPP+-Ac-GMDSLASFGGL-amide
by ESI-ISF-MS, MALDI-PSD-MS and FAB-CAD-MS respectively ....................... 76

Figure 2.23 LC-MS results of the TMPP+-Ac-derivative mixture of the tryptic digest of
cytochrome-C. Top panel shows the chromatogram for the separation mixture. Bottom
panel shows the ESI-ISF mass spectrum of a peptide derivative (residues 80-86), TMPP+-
Ac-MIFAGIK ....................................................................................... 78

Figure 2.24 ESI-CAD-MS/MS of TMPP+-Ac-IFVQK, peptide derivative from tryptic
digest of cytochrome-c, in an ion trap mass spectrometer at 55% relative collision
energy ................................................................................................. 80

Figure 2.25 ESI-CAD-MS/MS of a peptide derivative from cytochrome-c tryptic digest
with a precursor mass of 1179 Da at a relative collision energy of 60% .................... 82

Figure 2.26 ESI-CAD-MS of TMPP+-Ac-YLDSR, peptide derivative from the tryptic
digest of glucagon, showing the characteristic fragmentation pattern of aspartic acid
residue D .............................................................................................. 83

Figure 2.27 A comparison of ESI-ISF mass spectrum (top panel) with MALDI-PSD
spectrum (bottom panel) of TMPP+-Ac-GFFYTPK, a peptide derivative from the tryptic
digest of insulin chain-B ........................................................................... 85

Figure 3.1 Structures of amino Acid residues serine (S) , threonine (T), and tyrosine (Y)
respectively .......................................................................................... 90

Figure 3.2 Loss of phosphoric acid via a six-membered transition state in peptides
containing phosphorylated serine, and threonine residues .................................... 99

Figure 3.3 a) Possible elimination of H3P04 from tyrosine phosphate. b) Dehydration of
an aromatic alcohol vs. an aliphatic alcohol ................................................... 101

Figure 3.4 ESI-CAD-MS/MS spectrum of the doubly-charged precursor of LKRAyLG-
amide at 25% (relative collision energy full scale of 5 eV) ................................ 102

Figure 3.5. ESI-CAD-MS/MS spectrum of TMPP+-Ac-LKRAyLG-amide, singly-charged
precursor m/z 1471, at 50% relative collision energy ......................................... 104

Figure 3.6. ESI-CAD-MS/MS spectrum of TMPP+-Ac-LKRASLG-amide, singly-charged
precursor m/z 1395.5, at 50% relative collision energy ....................................... 106

xii

Figure 3.7a ESI—CAD-MS/MS spectrum of LKRAtLG-amide, singly-charged precursor
m/z 837, 30% relative collision energy ......................................................... 108

Figure 3.7b ESI—CAD-MS/MS spectrum of TMPP+-Ac-LKRAtLG- amide, singly-
charged precursor m/z 1409.5, 50% relative collision energy ............................... 108

Figure 3.8 ESI-CAD-MS/MS spectrum of TMPP+-Ac-KIGEthGVVYK, doubly-
charged precursor m/z 1024, 35% relative collision energy. The boxed notations indicate
the fragments retaining the phosphate moiety ................................................. 1 10

Figure 3.9 ESI-source-CAD-MS spectrum of TMPP+-Ac-LKRAyLG—amide, at a
capillary voltage of 70 V and tube lens offset of 50V ....................................... 1 12

Figure 3.10 Abundance of fragment ions formed from TMPP+-Ac-LKRAyLG-amide by
source-CAD-MS ................................................................................... 1 13

Figure 3.11 Abundance of fiagment ions formed from TMPPI-Ac-LKRAyLG-amide by
CAD-MS/MS ....................................................................................... 1 14

Figure 3.12 Structure of c, ions formed from TMPPI-Ac derivatives of peptides. . ....1 15
Figure 3.13 MALDI-TOF mass spectrum of the tryptic digestion mixture of B-casein. The

peak at m/z 3125 represents a protonated phosphopeptide with four phosphates. The
peaks at m/z 3027 and 2939 represent the protonated singly and doubly dephosphorylated

forms of the peptide RELEELNVPGEIVEsLsssEESITR respectively ..................... 1 16
Figure 4.1a Mechanism of fragmentation of protonated peptides to form yn ions ........ 126
Figure 4. lb Mechanism of fragmentation of protonated peptides to form bn ions. . . . . 1 27

Figure 4.2 Protons that could possibly shift during charge-remote fragmentation of a
peptide bond ........................................................................................ 129

Figure 4.3. Plausible structures of fragment ions formed from fixed-charged derivatives..130

Figure 4.4. Mechanism of 'an ion formation by amide hydrogen shift in TMPP+-Ac-
peptides ............................................................................................. 131

Figure 4.5. Structure of the proline residue ................................................... 132

Figure 4.6a ESI—source-CAD mass spectrum of TMPP+-Ac;VGVAPG in a single-
quadrupole mass spectrometer. Spectrum shows a series of an ions, except for cleavage
at the site of proline( a5) ......................................................................... 138

Figure 4.6b MALDI-PSD mass spectrum of TMPPI-Ac-VGVAPG. The spectrum
shows a series of an ions except for cleavage at the site of the proline residue( as). . ..139

xiii

Figure 4.6c F AB-CAD-MS/MS spectrum of TMPP+-Ac-VGVAPG (Adapted from
reference 21). Peak represented by as corresponds to cleavage at the proline residue..140

Figure 4.7 ESI-CAD-MS/MS spectrum of TMPPI-Ac-VGVAPG in an ion-trap mass
spectrometer ........................................................................................ l 4 1

Figure 4.8a ESI-source-CAD mass spectrum of Th/IPP+-Ac-VG°‘VAPG, species B, in an
ion trap mass spectrometer. Fragment ions a2- as are shifted by 2 Da corresponding to
the retention of the deuterium label on the a-carbon during fragmentation ............... 143

Figure 4.8b ESI-CAD-MS/MS spectrum of TMPP+-Ac-VG°‘VAPG, species B, in an ion
trap mass spectrometer at 35% relative collision energy. Fragment ions 'a2- .36 are
shifted by 2 Da corresponding to the retention of the deuterium label on the a-carbon
during fragmentation .............................................................................. 144

F igure 4.8c MALDI-PSD mass spectrum of TMPP+-Ac-VG°‘VAPG, species B. Fragment
ions at; - air are shifted by 2 Da corresponding to the retention of the deuterium label on
the a-carbon during fragmentation .............................................................. 145

Figure 4.9 Full scan mass spectrum of the amide hydrogen labeled species of

TMPPI-Ac-VGVAPG. Completely deuterium exchanged precursor is represented by the
peak at m/z 1077.5 ................................................................................. 153

xiv

Chapter 1

INTRODUCTION

ELECTROSPRAY IONIZATION TANDEM MASS SPECTROMETRY AND ITS
USE IN THE ANALYSIS OF CHARGED DERIVATIVES OF PEPTIDES

1. Preface

The goal of this chapter is to introduce the analytical techniques and concepts that
are primarily used in obtaining the results discussed in this dissertation. Electrospray
ionization (ESI), one of the desorption ionization (D/I) techniques, is presented in section
B. Structural information of the ions formed by ESI is obtained from their collisionally
activated dissociation (CAD). CAD is described in section C. The fragments that are
obtained from CAD are analyzed by triple quadrupole and ion trap mass spectrometers.
Section D of this chapter focuses on these mass spectrometers. The main crux of this
dissertation is to study the fragmentation of fixed charge derivatives; history of charged
derivatives and their study by various DI/—mass spectrometry (MS) is discussed in the

section B of this chapter.

ll. Analytical Techniques and Chemical Procedures

A) Desorption/Ionization (D/l) Techniques:

Mass spectrometry has been used in the structural elucidation of small organic
molecules for several decades [1]. The analyte, which has a significant vapor pressure

(10‘2 mm of Hg) at room temperature, is ionized and fragmented by energy absorbed

from 70 eV electrons. The resultant ionic species are then separated based on their
physical properties such as momenta, or m/z values in the mass analyzer. Thus the
technique itself has two major components, one involves the ionization source and the
other involves the mass analyzer. Hence, if a molecule is non-volatile and thermally
labile it cannot be transferred into the gas phase without degradation. Therefore, it is not
suitable for mass spectrometric analysis. Ingenious scientists realized this shortcoming of
the electron ionization technique in the late 1970’s, when the need to analyze the mass of
biomolecules was identified. These researchers paved way for the D/I techniques. These
techniques are briefly discussed in the following sections with a major focus on
electrospray ionization (ESI). Most of the results presented in this dissertation are based
on E81.

Biomolecules such as peptides and nucleotides have very low vapor pressure and
are non-volatile. The idea of transferring non-volatile compounds into the gas phase
without decomposition was originally put forth by Beuhler et al. [2]. Based on the rate
theory for unimolecular decomposition they proposed that sufficiently rapid heating
could vaporize complex molecules before decomposition occurred. This led to the
development of “pyrolysis M8” by Meuzelaar et al. [3]. Later, the same group used laser
photons to flash desorb the non-volatile large molecules. These kinds of techniques are
described by F enn et al. as “energy sudden”, since the analyte instantaneously achieves
high-energy density [4]. These experiments can be called the forerunners of the current
D/I techniques.

Fast atom bombardment (FAB) is the first D/I technique used in conjunction with

MS for the analysis of biomolecules up to 5000 Da in molecular weight [5]. Since then

other “energy sudden” techniques such as fast ion bombardment (FIB) [6] and plasma
desorption (PD) [7] have evolved. These techniques are based on generating ions from a
surface that cannot be continuously replenished. However, the development of
continuous-flow—FAB presented a continuously renewed surface to the incident fast
atoms [8].

LD became well known due to the efforts of Karas and Hillenkamp; they showed
the effect of a matrix that assists in the energy transfer from the photons to the analyte
[9]. This technique called matrix-assisted laser desorption ionization (MALDI) is
“softer” than the other “energy sudden” techniques described earlier. The wavelength of
the laser and the composition of the matrix (small organic molecules) play an important
role in this D/I processes. This technique has gained attention in the last few years and is
primarily used with a time-of-flight mass spectrometer. The use of a reflectron with this
technique has also allowed for the analysis of fragment ions produced from metastable
decomposition of the precursor ion during its flight in the field-free region [10]. This
technique is termed post-source decay (PSD) and is widely used in determining the
sequence of peptides and nucleotides.

Field desorption (FD) involves desorbing ions from analytes with the assistance
of very strong electrostatic fields. Inghram and Gamer [11] first used it as an ionization
source for MS. Later on Beckey modified the technique to produce higher currents from
analyte ions [12]. Emitter electrodes are used to desorb analytes into the gas phase in
vacuum. This technique is tedious and can be used only with magnetic sector mass

analyzers since the ions emerging from the ion source have kV’s of kinetic energy.

The other form of field desorption involves production of ions from solutions at
atmospheric pressure rather than in vacuum. It is most generally called atmospheric
ionization (API). It is also called electrospray ionization (ESI) or thermospray ionization
(TSI) depending on the design of the source. However, in both cases the formation of
droplets and the subsequent formation of ions are assumed to be the same. The following
sections will concisely describe the history of E81 and mechanism of ion formation in

ESI.

B) Electrospray Ionization:

Malcolm Dale and his coworkers were the first to report ESI in 1968 [13]. They
pioneered the experiments which explain the mechanism for ion formation in ESI. Figure
1.1 describes ESI in its widely accepted form [14]. The solution, which contains the
analytes in their ionic form, is nebulized into a high electric field gradient from a
capillary (held at 3.5 — 5 kV) with a typical inner diameter of 100 um and a flow rate of
1-10 uL/min. The charged analyte accumulate at the tip of the capillary, and the
electrohydrodynamics of the solution result in the formation of a “Taylor cone”. Droplets
with an initial diameter of lum are then formed from the Taylor cone by the process of
budding [15]. These typical droplets contain numerous charged analytes (150,000 for a

concentration of 500 frnol/ uL) [16].

I
Cylindrical Electrode Ouadrupoe
\Rods

\ Capillary
”(Md ‘ i \ Skimmer H i
53m” I. r: [:
—> 32.3

 

Vacuum Vacuum

Figure 1.1. The ESI source.

The mechanism by which the individual analyte ions are released from the droplet
is explained by two theories.
a) Charged-residue Mechanism:
Dole proposed the charged-residue mechanism. It states that as the charged droplets
progress through the electric field, solvent molecules evaporate due to the effect of the
drying gas, and the droplet shrinks in size. At one point, the number of charges on the
droplet is high and hence there is a very high coulombic repulsion that exceeds the
surface tension that holds the droplet together. This results in the “coulombic explosion”
of the droplet. These explosions occur repeatedly, which result in the ejection of
individual charged analyte molecules [13]. This theory was substantiated by Chapman’s
mobility experiments [17]. Figure 1.2 presents Dole’s mechanism of ion formation in

E81 [14].

M(H+)n

"Coulombic Ion
Explosion' Evaporation Multiply

Droplet After (Disintegration) Protonated
Reduction Analyte
in Size Due to
Evaporation

Figure 1.2. Mechanism of ion formation in ESI as proposed by Dole.

b) Ion Evaporation Mechanism:

Thomson and Iribarne [18] first proposed the field-assisted ion evaporation
mechanism. Later, this was applied for the ion generation for biomolecules by F enn et
al. [19]. This mechanism suggests that a charged part of the molecule penetrates the
surface of a droplet. The coulomb repulsion between this part of the molecule and the
droplet surface would pull the entire charged molecule out of the droplet. The
macromolecule plays an active role in ionization in this mechanism whereas it is passive
in the mechanism proposed by Dole [20].

Although, Dale’s mechanism is widely accepted, the ion evaporation mechanism
has still not been ruled out as a possibility. Whatever the mechanism of ion formation,
the multiply charged individual analytes are then analyzed by the mass spectrometer.
The formation of multiple-charged analyte ions is advantageous since this allows the
usage of a mass spectrometer such as a quadrupole or an ion trap, which has a limited

mass range capability (up to 2000).

There are several reviews and tutorials explaining the process of electrospray and
its advantages over the existing ionization mechanisms [21,22]. One of the advantages is
that the sample need not be volatile; however it should be ionizable. The other advantage
is that ionization occurs at room temperature and at atmospheric pressure. This eliminates
the possibility of sample degradation during ionization, because there is no need for
extreme heating. This aspect has been pursued in using ESI-MS for the study of non-
covalent interactions between molecules [23].

The major advantage is its compatibility with separation techniques such as liquid
chromatography. The effluent fi'om such a separation technique could be directed into
the inlet of the ESI-MS, and samples can be directly analyzed. Modern mass
spectrometers with specially designed sources could tolerate flow rates of even 1
mL/min, at a higher flow of sheath gas. On the other hand, in order to improve the
efficiency of ionization, there has been much focus on the miniaturization of the ESI
device. A report by Wilm and Mann has served as an introduction to this area of
research [16]. Based on theoretical calculations and experimental evidence, they conclude
that by using a capillary of 1-3 um, droplets as small as 180 nm could be formed. The
source was operated at 580-800 V as opposed to 3.5-5 kV in a conventional device.
Using such a construction, a l-uL aliquot of solution was sprayed for ca. 45 min., which
amounts to a flow rate of < 25 nL/min. The overall effect was that the ionization and
transmission efficiencies were improved by several orders of magnitude. It was also
shown that signal could be detected from a solution of concentration as low as 0.05
pmol/uL. Additionally, the micro electrospray device or the nanospray is more stable

than the conventional source and is moderately tolerant to salts. This idea has been

commercialized and there are several vendors who provide nanospray devices. These are
highly useful in analyzing very small amounts (femtomoles to attomoles) of biomolecules

such as protein derived from biological sources [24].

C) Collisionally Activated Dissociation Tandem Mass Spectrometry:

Structure elucidation of an analyte after ionization is possible in MS by
collisionally activated dissociation (CAD). Ions leaving the source can be classified
based upon their lifetime relative to the transient time from the ion source to the detector.
Ions with a lifetime greater than 10'6 sec. are considered stable and reach the detector
before they can fi'agment. Ions with a lifetimes less than 10'7 sec. are unstable and
fragment before leaving the ionsource. Ions with lifetime between 10*3 and 10'7 sec. are
considered metastable as they fragment sometime after leaving the source and before
reaching the detector. The ion with a lifetime longer than 10’6 sec, the stable ion, is
selected and activated before it fragments. The intact ion is energetically activated which
results in its fragmentation. The fragment ions thus formed are mass analyzed. Rules
governing the fragmentation for various molecules have been derived. For example, the
types of bonds that can be cleaved and the ions that could be obtained from a peptide are
well defined [25,26]. Thus, from the fragment ion masses and from using the specific
guidelines for the type of analyte, the structure could be determined. The scientific
community also uses the term collision-induced dissociation (CID) interchangeably with
CAD.

The process of CAD is specific since a precursor is chosen and is selectively

activated. The availability of tandem mass spectrometers allows one to perform MS/MS.

The term MS/MS implies that there are two scans performed. The first process is a full
scan for the entire m/z range to isolate the precursor of interest. The isolated precursor is
activated by collision with a neutral target gas admitted to the reaction region, during
which a portion of its translational energy coverts into internal energy. The excess
internal energy in the analyte is then channeled via the vibrational modes, which results
in the unimolecular decomposition of the activated ion. Only a fraction of the ion kinetic
or translational energy (Elab) is converted into its internal energy. This is given by the
energy of center-of-mass (Ecom) which is dependent on the mass of the target gas (m.) and
the mass of the precursor (mp) [27].
Ecom = Elab[ mt/( mt+ mp)]

Consequently, an increase in the ion kinetic energy (E138) increases the available energy
for ion fiagmentation (Ecom). The second scan involves the mass analysis of the
fragments formed after CAD.

CAD is classified as high-energy (HE) or low-energy (LE) based upon the energy
imparted to the precursor. The precursor is activated with keVs of energy in HE-CAD,
whereas it is activated with less than 200 eV in LE-CAD. HE- CAD is usually performed
in the magnetic sector or hybrid instruments. The primary excitation in this collision
energy region is electronic. The kinetic-to-internal energy conversion occurs most
efficiently when the collision interaction time and the internal period of the mode that
undergoes excitation are comparable [28]. The interaction time of the precursor and the
target in such cases is in the order of 10'” sec. This corresponds to a vertical electronic
transition. Generally, gas such as argon or xenon is used in these studies to enable

efficient internal energy transfer. Although keV’s of energy are supplied to the precursor

ion, only 1-4 eV of energy is acquired by the ion during its stay in the HE collision cell
[29,30].

LE-CAD is usually performed in the triple quadrupole or ion trap mass
spectrometers. The lifetime of the target-precursor complex is only 10'14 sec. and hence
the ions excitation is usually vibrational. The energy that is deposited on the ion is lower
than in HE-CAD. However, the collision yields are higher than in HE-CAD because of
the multiple collisions that can occur in these processes. The effective energy gained by
the precursor ion is about 2.3 eV [27]. Dependent on the mass spectrometer, the tandem
process could occur in time or space. For example, in the triple quadrupole MS, it occurs
in space, while in the ion tap mass spectrometer it occurs in time. There are several
books describing the operational principles of these two mass spectrometers [31,32]. It is
beyond the scope of this dissertation to describe the triple quadrupole and the ion trap
mass spectrometers in their entirety. However, their basic operation and the differences
between these two mass spectrometers will be discussed in the following section.

Based on the above description, the energetics involved in HE- and LE-CAD
processes vary; thus, significantly different fragmentation of the precursor occurs. It has
been reported that while it is possible to generate (1.. ions, formed due to cleavage of side
chains from peptides from HE-CAD, it is not very commonly observed during LE-CAD
[33]. The pathways involved in the formation of the same product ions from a precursor

by these two processes (HE-CAD and LE-CAD) could be different, too.

D) Triple Quadrupole and Ion Trap Mass Spectrometers:
The transmission quadrupole and the ion trap mass spectrometers are dynamic

devices, in which the ion trajectories are influenced by a set of time-dependent forces.

10

Ions in such quadrupole fields experience strong focusing to the center of the device. The
motion of the ions in quadrupole fields is described mathematically by the solutions to a
second-order linear differential equation described by Mathieu [31].

The transmission quadrupole is made of four hyperbolic rods that are electrically
connected as shown in Figure 1.3. RF and DC voltages are superimposed on the rods to
create a hyperbolic electric field. At any one set of voltages, only ions of a particular m/z
value have stable trajectories inside the device and they eventually reach the detector. An
analytical scan is performed by increasing the RF and DC voltages at a constant ratio.
The relationship between the voltages and the m/z values are derived based on the
Mathieu’s equations. Thus, the device could be termed as a filter based on the m/z value
of the ions. In order to obtain structural information of the ions, a triple quadrupole is
used [34]. There are three transmission quadrupoles aligned linearly. The first stage
allows the transmission of a precursor with a particular rrr/z value into the second stage.
The second stage is a RF -only device (non-mass selective) which contains the target gas
such as argon. The RF-only mode allows in better focusing of the precursor ions. The
precursor ions are kinetically accelerated while arriving into the high-pressure collision
cell and collisions with the target gas result in their fragmentation. The third stage
performs a firll scan to analyze the m/z values of the fragment ions resulting from the
CAD process. The fiagrnentation efficiencies of this device are high because of the
considerable distance traveled by the ion in the collision cell. Rapid speeds are
achievable for the precursor and product scan mode in the transmission quadrupole. The

ability to switch between masses makes multiple selected reaction monitoring

11

experiments practical. Thus, these devices are widely used in experiments for

quantitating a single component in highly complex mixture.

 

 

 

 

 

Figure 1.3. The components of a quadrupole mass spectrometer.

The ion trap mass spectrometer is also a quadrupole device composed of three
electrodes (Figure 1.4). There are two end-cap electrodes and a ring electrode.
Application of a RF potential to the ring electrode while maintaining the end-cap
electrodes at ground establishes a 3D quadrupole field in the trap. Helium is usually used
as the gas in the trap, which serves to “cool” the ions as well as to cause CAD. The
analytical scan is performed by operating the trap in the mass-selective instability mode
where the ions are rendered unstable and are ejected out of the trap in the order of
increasing m/z. CAD is performed in the trap by “resonance-excitation”. The precursor
ion of interest is isolated in the trap while all the other ions are ejected. A supplemental
AC voltage (“tickle”) whose frequency corresponds to the fundamental frequency of the
precursor ion is applied through the end-cap electrodes [32]. The ion is thus energetically
activated, which results in collisions with the helium gas. Thus fragments are formed,

which are then mass analyzed by performing a product ion scan in the trap.

12

 

Figure 1.4. The ion trap mass spectrometer.

The reaction times in the transmission quadrupoles and the ion trap quadrupoles
are different. In the trap, the residence times of the ions are in the order of milliseconds
whereas in the transmission quadrupole it is less than 100 usec. Thus, considerably more
collisions will occur in the trap ( which also depend on the pressure in the cell) and hence
higher energies can be deposited but in smaller increments [35]. This is often not the
case because, helium is the collision gas and hence the activating collisions are relatively
inefficient [36]. Moreover, because of the long reaction times, an ion that has
accumulated sufficient internal energy to decompose by a lower-energy process may not
survive to undergo another activating collision. Thus, the ion trap mass spectra are
reflective of fragment ions produced from low-energy processes than those spectra
produced from a transmission triple quadrupole device [37]. These characteristics are also
observed in the CAD-MS/MS results for the charged derivatized peptides as shown in the

following chapters.

E) Charged Derivatives for Analysis by Mass Spectrometry:

The introduction of the new desorption ionization techniques and MS/MS
revolutionized mass spectral analysis of peptides, but two problems remained: lack of
sensitivity and difficulty in data interpretation. Peptide mass spectra can be complicated
due to the many types of fragment ions that can arise during desorption/ionization and
CAD-MS/MS . The peptide fragments observed can be amino-terminal fragments (an,
bn, cn, and (In ions), carboxy-terrninal fiagments (xn, yn, Z“, V", and WH ions), or internal
fragments. The extent of charge localization influences the types of fragment ions
produced [38]. Charge-derivatized peptides have complete charge-localization and
produce fragments with the charge retained at the site of derivatization. Charged
derivatives were developed to simplify and direct the fiagmentation of peptides. Thus,
formation of a series of one ion type from charged derivatives facilitates the
interpretation of their mass spectra. Charged peptide derivatives form fragment ions
through charge-remote mechanisms (i.e., the site of the charge is remote from where the
cleavage of the bond occurs) [39,40]. As a result, specific types of fragment ions are
observed. Several types of derivatives were developed to lower the detection limits of
peptide analytes. Some of these derivatives contained nonpolar functionalities to increase
the surface activities of peptides during analysis by FAB and LSIMS [41-43]. Other
derivatives generated preformed ions, a condition that increased ionization efficiencies of
many analytes during analysis by desorption/ionization techniques [44].

Protonated peptides and charged peptide derivatives differ in mass depending on
the type of derivatization. The nomenclature used here for the fragment ions of charge-

derivatized peptides corresponds to the nomenclature of protonated peptides when the

14

mass difference in (singly-charged) precursor ions is taken into account, but an asterisk is
added to denote the presence of the derivative moiety. For example, the N-tenninal
fragment ion resulting from cleavage of the CH(Rn)-CO bond is called an an ion when it
arises from a protonated peptide and is called an *an ion when it arises from a charged
derivative.

Charge derivatization limits the types of fiagment ions generated from peptides.
Derivatization at the N-terminus causes N-terminal charge—remote fragments to be
observed and derivatization at the carboxy-terrninus (C-terminus) causes C-terminal
fragment ions to be observed. High-energy CAD of N-terrninal charged derivatives
produces primarily *a, and *d, ions plus a few *bn and *c. ions [45] while C-tenninal
charged derivatives produce primarily y.,*-2, y..*, v..* and w..* ions [46, 47]. This section
will focus on derivatives that localize a charge on the N-terminus of a peptide to control
its fi'agmentation during analysis by MS or MS/MS.

The vast majority of charged derivatives were developed for use with FAB or
LSIMS desorption/ionization techniques. Most of these papers were published between
the mid 1980’s and the mid 1990’s. Charge-remote fi'agmentation mechanisms require
greater internal energies than charge-directed fragmentation mechanisms [48] and are
most readily formed by high-energy CAD. Many instruments during this period made
use of high-energy CAD, which can provide the internal energies necessary to induce
charge-remote fiagmentation. Derivatization to increase analyte signal intensity was also
more important when FAB or LSIMS was used, because the response of many peptides

could be increased by increasing their hydrophobicity [41-43].

15

The use of charged derivatives with plasma desorption (PD) has been quite
limited [49,50] mainly because PD was never as widely used as FAB or LSIMS.
However, the high internal energies of ions generated by PD produce abundant fragment
ions without the use of CAD-MS/MS techniques.

Only recently [51-57] have charged derivatives been applied to the analysis of
peptides by matrix-assisted laser desorption/ionization (MALDI). MALDI has detection
limits for peptides and proteins that are several orders of magnitude lower than those
obtained with FAB or LSIMS, hence there was little incentive to develop derivatization
approaches to increase signal intensities of the analytes. Additionally, before the
development of post-source decay (PSD), most MALDI instruments were incapable of
generating and observing fragment ions of peptides. The spectra of underivatized
peptides obtained by MALDI-PSD have been observed to resemble those obtained by
LSIMS low-energy CAD [58]. PSD is an effective technique for forming fragment ions
of charged derivatives, and now charge derivatization to control fragmentation is an
attractive option.

The analysis of charged derivatives by electrospray ionization (ESI) has received
limited attention [59-62]. This is due to the belief that it would be difficult to form
charge-remote fragment ions during analysis by ESI because ions are produced with
lower internal energies than those formed by LSIMS[48], and because the majority of
electrospray instruments use low-energy CAD to form fragment ions for analysis by

tandem mass spectrometry.

16

a) Quaternary Ammonium Derivatives:
In 1984, Kidwell and coworkers reported the development of several
derivatization approaches. The first approach to N-terminal derivatization used methyl

iodide to form a trimethylammonium derivative as shown in Figure 1.5 [63,64].

0 1 1
MHz-Peptide-COOH -—H3—* Jili-Peptide-coou

Figure 1.5. Trimethylammonium derivatization scheme.
This derivatization approach had low yields, so another approach involving
successive derivatization of the peptide N-terminus with chloroacetyl chloride followed
by reaction with triethylarnine to give a triethylarnmonium derivative was proposed as

shown in Figure 1.6A.

. .. R. 0
MHz-Peptide-COOH A+ RaiN-CI-lzg-NH-Pepflde-COOH
R2
. ii .. is
Reagents: 1= Hal-CHZ -R1 11 = RQ-N—R‘

A) l-ial=Cl.R1 =c1, R2=R9=R,=Et
B) Hal=Cl.R1 =01, R2=RG=R4=CH3
C) Hal = '1 R1 = OCIO)CH2I- R2 = Re: CH3, R4 = CH3. CGH13v or C8H17
Figure 1.6. Trialkylammonium-acetyl derivatization scheme.
Both of these derivatization approaches derivatized unprotected lysine side-chains
as well. The derivatized peptides were randomly cleaved with acid or enzymes to
generate a series of derivatives differing by the number of residues present. The cleavage

products were then esterified and acetylated to suppress ionization of the underivatized

fragments. These mixtures were analyzed by secondary ionization mass spectrometry

17

(SIMS), the precursor of FAB and LSIMS. The derivatives had detection limits in the
low nanograrn (picomole) range; detection limits were not given for the underivatized
peptides. Data were presented only for peptides containing three or fewer residues.

Stults et al. presented a selective procedure for attaching a quaternary ammonium
group to the peptide N-terminus [65]. 1—30 nmol of peptide were reacted with iodoacetic

anhydride followed by thiocholine iodide, each at a controlled pH as shown in Figure 1.7.

o
1, 11 l 11
NHz-Pepflde-COOH ———> irii-CHchzs CHZC-NH-Peptlde-COOH

Reagents: 1: (101120020, 11= J+—CH,CH23H I-

Figure 1.7. Quarternary ammonium derivatization using thiocholine iodide.

The pH control allowed selective derivatization of the N-terminus by exploiting
differences between the pK. of the N-terminus and those of the basic amino acid side-
chains. However, protection of fi'ee cysteines was required to prevent derivatization [67].
The reactions were complete in a few hours, but the products had to be purified by HPLC
prior to analysis. Analysis by FAB or LSIMS followed by CAD-MS/MS gave typical N-
terrrrinal charge-remote fragment ions, but the spectrum was complicated by the loss of
trimethylamine from several ions.

Vth and Riemann proposed another procedure for attaching a
trimethylammonium group to the N-terminus of a peptide [67]. The procedure involved
reaction of approximately one nmol of peptide with chloroacetyl chloride followed by

trimethylamine as shown in Figure 1.63. This two-step procedure involved a total

18

reaction time of three hours. The signal intensity of the derivative was approximately
one-half that generated from an equal amount of the underivatized peptide during
analysis by FAB-MS. The derivative gave common charge-remote fragmentation
products when analyzed by FAB-CAD-MS/MS.

Stults et al. developed a similar reaction scheme for attaching a
dimethylalkylammonium (DMAA) group at the peptide N-terminus [68]. The
derivatization involved reaction of at least 100 pmol of peptide with iodoacetic anhydride
followed by reaction with a dimethylalkylamine as shown in Figure 1.6C. The
derivatization required a total reaction time of about two hours and gave overall yields of
60-80%. The pH control of the iodoacetylation gave 70-90% selectivity for the N-
terminus over the s-amino group of lysine, but cysteine residues had to be protected to
prevent derivatization. The presence of a nonvolatile buffer and excess reagents
necessitated HPLC cleanup prior to analysis by mass spectrometry. The derivatives were
analyzed by FAB-high-energy-CAD-MS/MS or by LSIMS-high-energy-CAD-MS/MS.
The dimethylhexylammonium and dimethyloctylarnmonium derivatives of a hexapeptide
yielded signal intensities that were stronger by a factor of 2-5 compared to those from the
underivatized peptide. The increase in signal intensity is presumably the result of
increased surface activity; however, such enhancements were not observed during
analyses of larger peptides (a 14-mer and a 19-mer). The trimethylammonium derivative
gave a slight decrease in signal intensity compared to that for the underivatized peptide.
The derivative gave abundant N-terminal charge-remote fragments and a small loss of the
hexyl group fiom the dimethylalkylammonitun functionality. Stults has applied this

derivatization technique to prepare peptide derivatives with a dimethylhexylammonium

19

group at the N-terminus for analysis by ESI [59]. The signal from the derivative was a
factor of 2-5 less intense than that from the underivatized peptide when analyzed by ESI-
MS. The doubly-charged ions of the underivatized peptide [M + 2H]2+ and the derivative
[C + H]2+ were used as precursors for analysis by ESI-low-energy CAD-MS/MS. In
contrast to the product ion spectrum of the underivatized peptide, the product ion
spectrum of the derivative gave primarily "'an and *b., ion peaks with no C-terminal
fragment ion peaks observed. The presence of abundant *b. ions is in contrast to the
fragmentation patterns observed by high-energy CAD [45]. This derivatization technique
was also used by Dongré et al. to prepare the trimethylammonium acetyl derivative of
leucine enkephalin for analysis by electrospray surface-induced dissociation tandem mass
spectrometry (ESI-SID-MS/MS) [48].

Bartlet-Jones et al. developed another derivatization approach (Figure 1.8) to

attach a trimethylammonium functionality to the peptide N-terminus [51].

O
I O
1N— (CH2)5)LO’N o
MHz-Peptide-COOH —> O —+l|\l— (C H2)5/u\ NH-Peptlde-COOH

Figure 1.8. CSQ derivatization reaction.

This derivative has been named “CSQ”, presumably because it consists of a five-
carbon chain and a quaternary ammonium group. The C5Q derivatization reaction was
complete in 10 minutes and was performed on 20-50 fmol of peptide; 50 frnol of
derivatized peptide was readily observed by MALDI. This derivatization approach was
studied in detail in a later paper [55]. MALDI-PSD analysis of the derivatized peptide

with a C-terminal arginine residue generated several yll ions. Derivatization of Lys and

20

Arg side-chains was recommended to reduce their basicity. The chemistry used to block
lysine side-chains results in cleavage of the N-terminal residue. The blocking of lysine,
derivatization of the N-terminus, and modification of arginine required four reaction steps
and more than four hours with quantities of 5-10 pmol of peptide. HPLC was required to
remove excess reagents. When basic side—chains were modified, only N-terminal
fragment ions were produced, primarily *an and *bn ions. Abundant *bn fragment ions
have not usually been observed for other charge-derivatized peptides analyzed by hi gh-
energy CAD [45] or MALDI-PSD [54,56]. However, *bn ions have been observed in the
MALDI-PSD spectra of charge-derivatized peptides when very high laser powers were
used for desorption [54]. When Hines et al. analyzed peptides derivatized with the C5Q
reagent [52], several *an fragment ion peaks and one *b. fragment ion peak were
observed in the MALDI-PSD spectrum, but the most abundant fragments were
immonium ions and ions presumably resulting fi'om a combination of backbone cleavage
and the loss of trimethylamine from the derivative ([*a.-59] and [*bn-59] ions).
b) Quaternary Phosphonium Derivatives:

Wagner et al. presented an approach for attaching a triphenylphosphonium group
to either the N-terminus or the C-tenninus of a peptide [46]. For the N-terminal
derivative, approximately one nmol of peptide was reacted with 2-bromoethyl-

triphenylphosphonium bromide at pH 9 for 3 hours at 37°C, as shown in Figure 1.9A.

21

@8.

© 9?

MHz-Peptide-COOH ——> ©P—CH20H2-NH-Peptide-COOH

A) R = crichzer
a) R = 01-1=CH2

Figure 1.9. Triphenylphosphonium-ethyl derivatization reaction.

The derivatization reactions gave greater than 75% yield [47], but with some
difficulty in reproducibility. These derivatives provided both charge localization and
enhanced surface activity on the FAB matrix. As a result, the fragmentation was
simplified and the signal intensity increased relative to that obtained from the
underivatized peptides. The detection limit of some peptide derivatives was below 5
picomoles. Both derivatives produced abundant charge-remote fragments arising from the
derivatized terminus. However, when a peptide containing a disulfide bond was
derivatized at the N-terminus, the mass spectrum provided less structural information
than the spectrum of the native peptide [69]. The derivatization reaction used pH control
to prevent reaction with basic side chains, but it required removal of the buffer salts prior
to analysis; this procedure was later modified to eliminate the need for nonvolatile
buffers [70]. With the new procedure, the peptide was reacted with
vinyltriphenylphosphonium bromide in a mixture of acetonitrile and pyridine as shown in
Figure 1.9B. This new procedure eliminated the need for sample cleanup, but the pH

control was uncertain and allowed the possibility of reaction with basic side-chains.

22

Bunk and Macfarlane reported the N-terminal derivatization of bradykinin with
vinyltriphenylphosphonium bromide as shown in Figure 1.98 [49]. The reagents were
combined in a pH 9.0 buffer solution and allowed to react at room temperature overnight.
The derivatives under went metastable decay to produce N-terminal fragments, with *an
ions dominating, when analyzed by plasma desorption mass spectrometry (PD—MS). A
peak for the protonated derivative [C + H]2+ and a few C-terminal fragments were also
observed. Although an increase in N-tenninal fragmentation was observed, the data
obtained by PD-MS analysis of the derivatives gave unsatisfactory results especially
when larger peptides were derivatized and analyzed [50]. The triphenylphosphonium
group was readily cleaved through a reverse Michael-type reaction, which led to a
decrease in sequence-specific fragment ions. Also, the derivative formed primarily
metastable fragments, that produced broader peaks than the fragment ions obtained from
underivatized peptides.

Liao and Allison reported that charge derivatization of a hexapeptide led to signal
enhancement when analyzed by MALDI [71]. They used the procedure of Wagner et al.
to prepare both the N- and C-terminal ethyltriphenylphosphonium derivatives. A one-
pmol sample of the underivatized peptide gave no signal when analyzed by MALDI, but
both the N- and C-tenninal derivatives produced strong signals at the one-pmol level.

Huang et al. introduced a derivatization procedure that attaches a tris(2,4,6-
trimethoxyphenyl)-phosphonium (TMPP+-Ac) group to the peptide N-terminus [53]. At
least 10 pmol of peptide were combined with S-pentafluorothiophenyl [tris(2,4,6-
trimethoxyphenyl)-phosphonium]acetate bromide and p-(dimethylamino)pyridine

(DMAP) for 15 minutes at room temperature as shown in Figure 1.10.

23

BF 0
11
MHz-Peptide-COOH —> 1MPP+-CH2-C-NH-Peptide-COOH

OFF

Br'
ll
Reagent = 1MPP+-CHz-C —S-@-F
OCHa
+ +
OCH3 3

Figure 1.10. TMPP+-Ac derivatization reaction.

The reaction gave yields in excess of 90% for most peptides, but a few peptides
have yields as low as 64%. The reaction mixture usually does not require purification
prior to analysis by FAB or MALDI. In addition, peptides composed primarily of amino
acid residues with nonpolar side-chains showed significant signal enhancement upon
derivatization and analysis by FAB or MALDI. When the derivatization is performed at
a controlled pH of 8.2, derivatization will occur at the N-terminus with no derivatization
of lysine side-chains. The spectra of the derivative obtained by FAB-high-energy CAD-
MS/MS show a strong series of charge-directed fragments. When TMPP+-Ac derivatized
peptides were analyzed by LSIMS-CAD-MSn with a quadrupole ion trap instrument,
abundant *b..+H20 ions were produced in addition to a series of *an ions [72].

The TMPP+-Ac derivative (Figure 1.10) has been used to direct the fragmentation
of peptides during analysis by MALDI-PSD [53,54]. The spectra consist of a strong
series of *an ions plus a few *b.., *c.., and *dII ions. The *dn ion peaks in the MALDI-

PSD spectra are less intense than those in the FAB-high-energy CAD-MS/MS spectra.

24

The TMPPI-Ac reagent has been used to derivatize protein digests at the 25 pmol level
[73]. This reagent was also used to derivatize peptides located in an electrophoretic gel
or membrane [56]. The modified peptides were then analyzed by MALDI and MALDI-
PSD directly from the gel or membrane on a modified sample plate.

The TMPP+-Ac derivative has also been applied to peptide analysis by
electrospray using an elevated cone voltage to promote in-source fragmentation [62].
With the cone set at a typical potential of 37 V, the peptide derivatives did not fragment
significantly and the derivatives were observed as the doubly-charged derivative [C +
H]2+. When the cone voltage was raised to 100 V, extensive fragmentation occurred and
the doubly-charged derivative was not observed. The fragmentation produced primarily
*a, and *bII ions with some *cn and *d. ions produced as well. This pattern is similar to
the fragmentation of another charged derivative analyzed by ESI-low-energy CAD-
MS/MS [59].

c) N-terminal Derivatives With High Proton Affinities:

In addition to derivatives with a fixed charge, several N-terminal derivatives have
been developed'with high proton affinities. The basic groups provide some degree of
charge localization, but there is potential for protonation at other sites or for proton
transfer to other sites. These derivatives are analogous to peptides with basic residues
(Arg, His, or Lys) at the N-terrninus. As a result, the nomenclature used for these
derivatives will be the same as that used for underivatized peptides.

Renner and Spiteller [38] reported N-terminal derivatization schemes using
dansylchloride, as shown in Figure 1.11A, and 2-bromo-5-(dimethylamino)benzene

sulfonyl chloride as shown in Figure 1.11B.

25

Ar-SOZCI
MHz-Peptide-COOH ———> Aresoz-NH-Peptide-COOH

/
—-N

A1Ar=> 8 B)Ar= Q»

Br
Figure 1.11. Reaction schemes of Renner and Spiteller.

Although these derivatives did not have a fixed charge, the aromatic amine was
preferentially protonated during analysis by FAB. The derivatives had greater signal
intensities and an increase in fragmentation (without CAD) compared to the
underivatized peptides. The F AB-MS analysis of the dansyl derivative produced only N-
terrrrinal fragments (an and b..), but the second derivative produced a.., b.., and yn ions.
The an and b.I ions arising from the second derivative produced two peaks for each ion, as
a result of the 79Br and 8'Br isotopes. These isotope peaks allowed the N-terminal
fragments be distinguished from the C-terminal fragments. The chemistry employed
would likely also derivatize basic side-chains as well as the N-terminus of the peptide.

Sherman et al. [60] have used N-hydroxysuccinimide-2-(3-pyridyl)acetate (SPA)
to derivatize the peptide N-terminus, as shown in Figure 1.12, in preparation for analysis

by ESI-MS; this chemistry was optimized by Cardenas et al. [61].

0
@ct
“”29 m
MHz-Peptide-COOH ———> 0 NH-Peptide-COOH

Figure 1.12. SPA derivatization reaction.

26

The N-pyridylacetyl functionality is not charged, but it does have a slightly
increased proton affinity compared to that of the peptide N-terminus. The derivatives
were analyzed by ESI-MS/MS on a triple quadrupole instrument [60] or by ESI-MSn on a
quadrupole ion trap instrument [61]. The spectrum of the derivative contains both N-
terrninal fragment (b“ and an) peaks and C-terminal fragment (yn) peaks. However, the bn
fragment ion peaks of the derivative are more intense than the corresponding bn fragment
ion peaks of the underivatized peptides. The optimized reaction occurs in four minutes
and gives yields in the range of 80-100% for peptides that do not contain lysine or
tyrosine. The reagent reacts with the side-chain of lysine and reacts to a lesser extent
with the side-chain of tyrosine. Reactions with the side-chains produce derivatives
having multiple N-pyridylacetyl functionalities. This derivatization has been used on
peptides having as many as 15 residues.

Naven et al. [57] used MALDI-PSD to study peptides derivatized with SPA
(Figure 1.12). The derivatization was performed on as little as 50 frnol of peptide. The
procedure was also applied to tryptic peptides electroblotted onto a PVDF membrane.
The reaction was essentially quantitative at the N-terminus and the lysine side-chains.
The PSD spectra contained peaks for a series of abundant bn ions and some an and yn
ions.

d) The Ideal Charged Derivative:

The ideal charged derivatization approach would have several attributes. The
chemistry should occur rapidly with high product yield without unwanted derivatization
of peptide side-chains. The derivatives should not require purification prior to analysis.

The derivative should be compatible with all mass spectrometry ionization and

27

fragmentation techniques. The ideal derivative would have a lower detection limit than
the underivatized peptide. The derivative should have a fixed charge and produce a
series of charge-remote fragments upon analysis by tandem mass spectrometry. The
derivative moiety should not fragment during analysis by tandem mass spectrometry
because that fragmentation might complicate the mass spectrum and reduce the intensity
of the other fi'agment ion peaks.

Three of the derivatization approaches discussed come closer to the ideal than the
others. They are the dimethylalkylammoniun acetate (DMAA) derivatization reported by
Stults et al. (Figure 1.6C), the C5Q derivatization of Spengler et al. (Figure 1.8) and the
[tris(2,4,6-trimethoxyphenyl)-phosphonium] acetate (TMPP+-Ac) derivatization reported
by Huang et al. (Figure 1.10). These approaches utilize highly specific chemistry to
attach a fixed charge to the N-terminus of the peptide in high yields.

The DMAA derivatization is performed with commercially available reagents.
The C5Q and TMPP+-Ac derivatization reagents must be synthesized, but the TMPP+-Ac
reagent, and possibly the C5Q reagent, is stable for months. The TMPPI-Ac
derivatization occurs in one fifteen-minute step, the DMAA derivatization occurs in two
steps requiring a total of two hours, and the C5Q derivatization between twenty minutes
(without side-chain derivatization) and four hours (with derivatization of Lys and Arg).
The amount of peptide required for the derivatization reactions has been reported as 100
pmol for DMAA, 10 pmol for TMPP+-Ac and 5-10 pmol for the C5Q derivatization.
However, the C5Q derivatization procedure can be performed on sub-picmolar quantities
of peptide if neither Lys nor Arg need derivatization. All three reactions occur with high

yields and are selective for the N-terminus when the pH of the reaction mixture is

28

controlled. The DMAA and C5Q derivatizations usually require purification by HPLC to
remove excess reagent. Such purification usually is not required for the TMPP+-Ac
derivatization when the reaction mixture is analyzed by FAB or MALDI, but it is
required in analyses by ESI.

The TMPPI-Ac derivative has been analyzed successfully with FAB, ESI, and
MALDI techniques. The DMAA derivatives have only been analyzed by FAB and ESI,
but they are probably compatible with MALDI. Likewise, the C5Q reagent has only been
analyzed by MALDI, but should be compatible with FAB and ESI.

DMAA derivatives give a small increase in signal intensity in some conditions
and a small decrease under others. No loss of signal intensity was observed for C5Q
derivatization, but the relative signal intensities of underivatized and derivatized peptides
were not reported. For most peptides, the TMPPI-Ac derivative has only a small effect
on signal intensity, but some hydrophobic peptides have shown a significant signal
enhancement upon derivatization. All of the derivatives have detection limits in the low
picomole to sub-picomole range.

Because the derivative moieties have fixed charges, the fragment ions produced
from these derivatives are believed to form by charge-remote mechanisms. F AB-high-
energy-CAD-MS/MS analysis of the DMAA and TMPP+-Ac derivatives produces
primarily *a. and *d.. plus a few *bII and *cll ions. In addition, the DMAA derivative
loses the alkyl group to generate a peak that can complicate the high-mass end of the
spectrum. When TMPP+-Ac derivatives are analyzed by MALDI-PSD, a dominant series
of *a. ions is produced and a few *b.., *c.. and *dll ions are produced. Peptides

derivatized with C5Q (including modification of Lys or Arg) are reported to form

29

primarily *an and *bll ions, but also *0n and *dn ions to a lesser extent. In a preliminary
study by Hines et al., a C5Q derivative was reported to form primarily *an-59 and *bn-59
ions. The differences in these fragmentation patterns are difficult to explain. Perhaps the
laser power (or some other experimental parameter) is causing the differences in
fragmentation as has been documented in one case of fragmentation of charged
derivatives during analysis by MALDI-PSD. The C5Q derivative moiety has a smaller
head group than the TMPP+-Ac moiety, and the C5Q moiety has a flexible five-carbon
chain. These differences might allow the formation of secondary structures in which the
charge can interact with the peptide backbone to influence fragmentation behavior.

When DMAA and TMPP+-Ac derivatives are analyzed by ESI, primarily *an and

*bn ions are generated plus a few *cn and *d.. ions.

30

lll. Conclusions

Although most of these charged derivatives have been studied by F AB-CAD-MS
and MALDI-PSD-MS, there has been very little focus in using ESI-CAD-MS/MS for the
study of these charged derivatives. The fi'agmentation pattern of peptides in ESI-CAD-
MS/MS is equally complicated as in the other two techniques. Study of the charged
derivatives of peptides by ESI-CAD-MS/MS could provide reliable and easily
interpretable spectra. It is known that charged derivatives require high-energy for their
fragmentation. However, fragmentation of these charged derivatives in the low-energy
regime (< 200 eV) has been observed [59]. These factors have been the major driving
force behind the study of the TMPP+-Ac charged derivatives by ESI-CAD-MS/MS
reported in this dissertation. Chapter 2 presents a systematic study of the TMPPI-Ac
derivaitves of peptides. Attempts are made to derive rules for fragmentation behavior of
various amino acid residues when present in the TMPP+-Ac derivatized peptide
sequence. Results from the fragmentation of these charged derivaitves in a quadrupole
vs. an ion trap mass spestrometer are compared. Chapter 3 presents the application of the
TMPP+-Ac charge- derivatization towards identifying the sites of phosphorylation in a
peptide. Chapter 4 focuses on understanding the basic mechanism of charge-remote
fragmentation under LE conditions at the molecular level using the TMPPI-Ac

derivatives of deuterium labeled peptides.

31

10.

ll.

12.

13.

14.

15.

16.

17.

18.

IV. References

. Budzikiewicz,H.; Djerrasi, C.; Williams, D.H. “Mass Spectrometry of Organic

Compounds” Holden Day, San Francisco, 1967.

Beuhler, R.J.; Flanigan, E.; Green, L.J.; Friedman,L. J. Am. Chem. Soc, 1974, 96,
3990.

Meuzelaar, H.L.C.; Posthumus, M.A.; Kistemaker, P.G.; Kistemaker, J. Anal. C hem,
1973, 45, 1546.

Fenn, J .B.; Mann,M.; Meng,C.K.; Wong,S.F. Mass Spec. Rev., 1990, 9, 37.

Barber,M.; Bordoli,R.S.; Sedgwick, R. D.; Tyler, A. N. J. Chem. Soc., Chem.
Commun., 1981, 325.

Honig, R.E. Int. J. Mass Spectrom. Ion Phys, 1985, 66, 31.
Sundqvist, B.U.R.; Macfarlane, R.D. Mass Spec. Rev., 1985, 42, 421.
Caprioli, R.M. Mass Spec. Rev., 1987, 6, 237.

Karas, M.; Hillenkamp, F. Anal. Chem, 1988, 60, 2299.

Spengler, B.; Kirsch, D.; Kaufinann, R.; Jaeger, E. Rapid Commun. Mass Spectrom.,
1992, 6, 105.

Ingrhram, M.; Gomer, R.J. J. Chem. Phys, 1954, 22, 1279.

Beckey,H.D. Principles of Field Ionization and Field Desorption Mass Spectrometry.
Pergamon, Oxford, 1977.

Dole,M.; Mack, L.L.; Hines, R.L.; Mobley, R.C.; Ferguson,L.D. Alice, MB. .1.
Chem. Phys.,l968, 49, 2240.

Watson, J.T. Introduction to Mass Spectrometry, Lippincott-Raven, Philadelphia,
1997.

Bruins, A.P.; Covey, T.R.; Henion, J .D. Anal. Chem, 1987, 59, 2642.
Wilm, M.; Mann,M. Int. J. Mass Spectrom. Ion Processes, 1994, 136, 167.
Chapman, S. S. Phys Rev., 1937, 52, 184.

Thomson, B.A.; Iribame, J .V. J. Chem. Phys, 1979, 71, 4451.

32

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

Penn, J .B. J. Am. Soc. Mass Spectrom., 1993, 4, 524.

Mann, M. Org. Mass Spectrom., 1990, 25, 575.

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

Loo, J. Bioconjugate Chem, 1995, 6, 644.

Smith, R.D.; Light-Weill, K.J. Biol. Mass Spectrom., 1993, 22, 493.

Wilm, M.; Shevchenko, A.; Houthaeve, T.; Breit, S.; Schweigerer, L.; F otsis, T.;
Mann, M. Nature, 1996, 379, 466.

Roepstorff, P.; Fohhnan, J. Biomed. Mass Spectrom. 1984, I I , 601.

Biemann, K. Nomenclature for Peptide Fragment Ions. Methods in Enzymology, Ed.
JMcCloskey, 1990, 193, 886.

Busch, K.L.; Glish, G.L.; McLuckey, S.A. Mass Spectrometry/ Mass Spectrometry:
Techniques and Applications of Tandem Mass Spectrometry, VCH, New York, 1988.

Benyon, J .H.; Boyd, R.K.; Brenton, A.G. Adv. Mass Spectrom., 1986, 10A, 437.
Neurnann, G.M.; Derrick, P.J. Org. Mass Spectrom., 1984, 19, 165.

Wysocki, V.H.; Kenttamma, H.I.; Cooks, R.G. Int. J. Mass Spectrom. Ion Processes,
1985, 75, 181.

Dawson, P.H. (ed.) Quadrupole Mass Spectrometry and Its Applications, Elselvier
New York, 1976.

March, R.E.; Todd, J.F.J. Fundamentals of Ion Trap Mass Spectrometry, CRC press,
Boca Raton, 1995.

I. A. Papayannopoulos Mass Spectrom. Rev., 14, 49-73 (1995).

Yost, RA; Enke, C.G. J. Am. Chem. Soc. 1978, 100, 2274.

Brodbelt, J .; Kenttamma, H.I.; Cooks, R.G. Org. Mass Spectrom., 1986, 23, 6.
Alexander, A.J.; Boyd, R.K Int. J. Mass Spectrom. Ion Processes, 1989, 90, 211.

Hayes, R.N.; Gross, M.L Collision-Induced Dissociation Methods in Enzymology, Ed.
.ZMcCIoskey, Academic Press, 1990, I93, 337.

Renner, D.; Spiteller, G. Biomed. Environ. Mass Spectrom., 1986, 13, 405.

33

39

40

41.

42

43.

44.

45.

46.

47.

48.

49.

50.

51.

52.

53.

54.

55.

. Gross, M. L. Int. J. Mass Spectrom. Ion Processes, 1992, 118/119, 137.

. Jensen, N. J .; Tomer, K. B.; Gross, M. L. J. Am. Chem. Soc., 1985, 107, 1863
Falick, A. M.; Maltby, D. A. Anal. Biochem., 1989,182,165.

. Ligon, W.V. Anal. Chem., 1986, 58, 485.

Naylor, S. ; Findeis, A. F.; Gibson, B. W.; D. H. Williams J. Am. Chem. Soc., 1986,
108, 6359.

Busch, K. L.; Unger, S. E.; Vincze, A.; Cooks, R. G.; Keough, T. J. Am. Chem. Soc.,
1982, 104,1507.

Zaia, J .; Biemann, K. J. Am. Soc. Mass Spectrom., 1995,6, 428.

Wagner, D. S.; Salari, A.; Gage, D. A.; Leykam, J .; Fetter, J .; Hollingsworth, R.;
Watson, J. T. Biol. Mass Spectrom., 1991, 20, 419.

Watson, J. T.; Wagner, D. 8.; Chang, Y. S.; Strahler, J. R.; Hanash, S. M.; Gage, D.
A. Int. J. Mass Spectrom. Ion Processes, 1991, 111, 191.

Dongré, A. R.; Jones, J. L.; Somogyi, A.;. Wysocki, V. H. J. Am. Chem. Soc., 1996,
118, 8365.

Bunk, D. M. ; Macfarlane, R. D. Int. J. Mass Spectrom. Ion Processes, 1991, 111, 55.

Bunk, D. M.; Macfarlane, R. D. Int. J. Mass Spectrom. Ion Processes, 1993, 126,
123.

Bartlet-Jones, M.; Jeffery, W. A.;. Hansen, H. F ; Pappin, D. J. C. Rapid Commun.
Mass Spectrom., 1994, 8, 737.

Hines, W.; Peltier, J .; Hsieh, F.; Martin, S. Proceedings of the 43"! ASMS Conference
on Mass Spectrometry and Allied Topics, Atlanta, Georgia, May 21-26, 1995.

Huang, Z.-H.; Wu, J .; Roth, K. D. W.; Yang, Y.; Gage, D. A.; Watson, J. T. Anal.
Chem, 1997, 69, 137.

Liao, P.-C.; Huang, Z.-H.; Allison, J. J. Am. Soc. Mass Spectrom., 1997, 8, 501.
Spengler, B.; Leutzenkirchen, F.; Metzger, S.; Chaurand, P.; Kaufmann, R.; Jeffery,

W.; Bartlet-Jones, M.; Pappin, D. J. C. Int. J. Mass Spectrom. Ion Processes, 1997,
169/1 70, 127.

56. Strahler, J. R.; Smelyanskiy, Y.; Lavine, G.; Allison, J. Int. J. Mass Spectrom. Ion
Processes, 1997, 169/1 70, 111.

57. Naven, T. J. P.; Jeffery, W. A.; Bartlet-Jones, M.; Rahman, D; Pappin, D. J. C.
Proceedings of the 45'” ASMS Conference on Mass Spectrometry and Allied Topics,
Palm Springs, California, June 1-5, 1997.

58. Rouse, J. C.; Yu, W.; Martin, S. A. J. Am. Soc. Mass Spectrom., 1995, 6, 822.

59. Stults, J. T. Proceedings of the 40"I ASMS Conference on Mass Spectrometry and
Allied Topics, Washington, DC, May 31-June 5, 1992.

60. Sherman, N. E.; Yates, N. A.; Shabanowitz, J .; Hunt, D. F.; Jeffery, W.; Bartlet-
Jones, M.; D. J. C. Pappin, Proceedings of the 43rd ASMS Conference on Mass
Spectrometry and Allied Topics, Atlanta, Georgia, May 21-26, 1995.

61. Cérdenas, M. S.; van der Heefi, E.; de Jong, A. P. J. M. Rapid Commun. Mass
Spectrom., 1997, 11, 1271.

62. Sadagopan, N.; Wu, J.; Yang, Y.; Huang, Z.-H.; Watson, J. T. Proceedings ofthe 45'h
ASMS Conference on Mass Spectrometry and Allied Topics, Palm Springs, California,
June 1-5,1997.

63. Kidwell, D. A.; Ross, M. M.; Coulton, R. J. J. Am. Chem. Soc., 1984, 106, 2219.

64. Kidwell, D. A.; Ross, M. M.; Coulton, R. J. Springer Ser. Chem. Phys. .° SIMS 4,
1984, 36, 412.

65. Stults, J. T.; Halualani, R.; Wetzel, R. Proceedings of the 3 7"1 ASMS Conference on
Mass Spectrometry and Allied Topics, Miami Beach, Florida, May 21-26, 1989.

66. Wetzel, R.; Halualani, R.; Stults, J. T.; Quan, C. Bioconjugate Chem, 1990, 1, 114.
67. Vath, J. E.; Biemann, K. Int. J. Mass Spectrom. Ion Processes, 1990, 100, 287.

68. Stults, J. T.; Lai, J .; McCune, S.; Wetzel, R. Anal. Chem, 1993, 65, 1703.

69. Chang, Y. S.; Gage, D. A.; Watson, J. T. Biol. Mass Spectrom., 1993, 22,176.

70. Wagner, D. S. Ph.D. Dissertation, Michigan State University, East Lansing, MI
(1992)

71. Liao, P.-C.; J. Allison J. Mass Spectrom., 1995, 30, 511.

72. Lin, T.; Glish, G. Proceedings of the 45'” ASMS Conference on Mass Spectrometry
and Allied Topics, Palm Springs, California, June 1-5, 1997.

35

73. Huang, Z.-H.; Shen, T.S.; Wu, J .; Gage, D. A.; Watson, J. T. Anal. Biochem., 1999,
268, 305.

36

Chapter 2

ANALYSIS OF TRIS(2,4,6-TRIMETHOXYPHENYL)PHOSPHONIUM ACETYL
DERIVATIVES OF PEPTIDES BY ELECTROSPRAY IONIZATION FOLLOWED
BY QUADRUPOLE OR ION TRAP MASS SPECTROMETRY

I. Introduction

A) Role of Mass Spectrometry in Protein Sequencing:

Determining the amino acid sequence of proteins continues to be of great interest
in today’s scientific community especially due to the rapidly progressing human-genome
project [1]. In addition, it is of importance to the pharmaceutical industries where
recombinant proteins are being developed for therapy, diagnosis, and prophylaxis of
human disease [2]. Usually, protein sequences are either deduced from the cDNA
sequence, or determined by Edman sequencing. Although the DNA sequence provides
the primary amino acid sequence, initiated at the methionine codon and terminated at the
stop codon, it does not give any information regarding the post-translational
modifications of the protein, which occur in the cytoplasm [3]. Phosphorylation, disulfide
bond formation, deamidation, and glycosylation are several such important post-
translational events that modify the protein structure outside the nucleus. The Edman
degradation method is very sensitive, but is only unidirectional; it is also difficult to
obtain sequence information if the N-terminus is acylated or blocked due to other
modifications. In cases where results from the previously mentioned techniques are

ambiguous or incomplete, mass spectrometric information can be complementary.

37

B) Nomenclature of Fragment Ions Formed From Collisionally Activated
Dissociation:

The nomenclature for fiagmentation of peptides under mass spectrometric
ionization methods has been clearly defined [4]. When a peptide is subjected to
collisionally activated dissociation (CAD), the fragments produced are dependent on the
position of the most basic residue (if any) in the sequence or protonation of any given
amide group. In the latter case, due to charge-mediated fragmentation, a mixture of ion
types is produced, predominantly b and y ions, and some a, c, x, and 2 ions. This is
depicted pictorially in Figure 2.1. In cases where there is a B-methylene unit in the side
chain, d and w ions are also possible. This multitude of ion types complicates the
spectrum, and requires a great scrutiny to interpret the spectrum. In an effort to obtain a
simple spectrum, charged derivatives of the peptides (as defined in chapter 1) could be
prepared and subjected to fragmentation to generate predominantly one type of ions.
This generally occurs via charge-remote fragmentation, which is also likely when a basic
residue is located on either terminus. Most of the charged derivatives studied so far have
been shown to produce "‘an ions [5]. The results presented in this chapter are those of N-
tenninally derivatized peptides. The fiagment ions contain a charged group at the N-
terrninus, and will be referred to as a.-type ions (*an) instead of all ions. Similarly, the
other charge-derivatized N-terminal containing fragments will be denoted as *b.., *c..,

and *dn ions

38

X3 Y3 23 X2 Y2 Z2 X1 Y1 Z1

 

 

 

 

 

 

 

 

 

r r
R1 0 R2 0 R3 0 R4
II I || I II |
HzN—C—PC--N-fi-C—-C—-N--C--C--N-r-C—COOH
H H H H H H H
__J

 

 

 

 

A1 B1 C1 A2 32 C2 A3 B3 C3

Figure 2.1. Nomenclature for peptide fragmentation during CAD.

C) FAB-CAD-MS and MALDl-PSD-MS Studies of TMPP+-Ac-Peptides:

Although a number of charge—derivatizing reagents have been described, most of
them have disadvantages [5]. One of the charge-derivatizing reagents that was deve10ped
to avoid the disadvantages of other existing reagents, contains a tris[2,4,6-
trimethoxyphenyl]phosphonium acetyl (TMPP+-Ac) cation [6]. The structure of the
TMPP+-Ac reagent and its coupling reaction with the N-terminus of a peptide are shown
in Figure 2.2; the reagent selectively derivatizes the N-terminus of the peptides at the
picomole level, with only a 5-fold molar excess, at room temperature. Studies of
TMPP+-Ac-peptides by fast atom bombardment-CAD (FAB-CAD) [6] and matrix-
assisted laser desorption/ionization —— post-source decay (MALDI-PSD) [7] have
demonstrated that it is possible to correlate the CAD spectrum of a TMPPI-Ac-peptide to
its structure. TMPP+-Ac derivatives of peptides have also been studied by FAB-CAD in

an ion trap mass spectrometer [8].

39

 
 
 
 

+.CH2-CO -SC5F5 B} + NH2-PeptIde-COOH

DMAP,
15'
Room Temperature
“3
H3c P-CHz-CO-NH-Peptlde-COOH
061-13
3
(TMPP-Ac-Peptlde)

Figure 2.2. Reaction of TMPP+-Ac reagent with the N—terminus of a peptide.

TMPP+-Ac derivatization can be applied to protein digests. The lypophilized
tryptic digest of proteins like cytochrome-c can be treated directly with the TMPP+-Ac
reagent, and analysis of derivatized tryptic fragments by MALDI—PSD can be used to
obtain sequence information [9]. Whereas a simple mixture of TMPP+-Ac tryptic
peptides can be analyzed directly by MALDI-PSD, preliminary separation is usually
required for proteins larger than 10 kDa because a larger excess of the TMPP+-Ac reagent
is used for derivatization. This excess reagent may suppress the signal in MALDI, and
hence HPLC is necessary. Such practical issues create a pressing need for the analysis of
these charged derivatives by electrospray ionization (ESI). An obvious advantage is the
possibility of LC-MS based on ESI. Hence, it was of great interest to study the
fragmentation pattern of the TMPP+-Ac-peptides using ESI-MS.

There have been many studies on the fragmentation pattern of charged derivatives
by FAB [6,8,10-18] and MALDI [7,19-23]. However, there are only a few studies based

on ESI of charged derivatives [24-28]. There have been several high-energy studies of

40

peptide fragmentation with ESI. High-energy CAD of multiple-charged ions from ESI
has been used to study the parameters of charge location and charge state [29]. Wysocki
and coworkers have reported ESI-surface induced dissociation (SID) studies of peptides
[30]. Since ESI is more commonly used with triple quadrupole and ion trap mass
spectrometers, it is of interest to study the low-energy CAD of such ESI-generated ions.
This paucity of such studies by ESI prompted our investigation of the fragmentation of
TMPP+-Ac derivatives by ESI via in-source fi'agmentation (ISF) and MS/MS (low-
energy CAD) with a comparison to results from FAB-CAD and MALDI-PSD. Such
studies would also provide insight to the fragmentation pathways of charged derivatives

when ionized by ESI and fragmented under low-energy conditions.

41

II. Experimental

A) Preparation of Derivatives:

Peptides and proteins used in the experiments were purchased from Sigma and
were dissolved in acetonitrile/water (1:1 v/v). Peptides were N-terminally derivatized
using TMPP+-Ac-SC6F5 bromide reagent synthesized as described elsewhere [6]. Table

2.1 shows the list of peptides that were used for this study.

Table 2.1. Peptides Studied by ESI-CAD-MS After TMPP+-Ac Derivatization:

 

 

Peptide Sequence Mass of Derivative
(C+)

Chemotactic Elastin VGVAPG 1071.4
Osteocalcin Fragment FYGPV 1 196.5
Peptide VGGYGYAK 1444.6
Allatostatin III GGSLYSFGL-amide 1472

Hb-S peptide VHLTPVEK 1495.7
Buccalin GMDSLAFSGGL—amide 1626
Lutenizing hormone HWYGLRPG-amide 1644.8
Angiotensin II APGDRIYVI-IPF 1 845
Neuromedin GNLWATGHFM-amide 1704.7
Cardioactive peptide-B MNYLAFRPM-amide 1715
Dynorphin Fragment YGGFLRKYPK 1802.1
Human Renin Inhibitor PHPFHFFYK 1892.1
Allatostatin I APSGAQRLYGFGL-amide 1907.9

 

The peptide solution was mixed with a 5-fold molar excess of the derivatizing
reagent in the presence of dimethylaminopyridine (DMAP). The reaction was allowed to
occur at room temperature for 15 min. After derivatization, the reaction mixture was
analyzed by HPLC to separate the derivatized peptide from any underivatized peptide and

hydrolyzed derivatizing reagent. A C13 column (4.8 x 250m) from Vydac was used.

42

Separation was performed using a gradient elution program from 25% to 60% of
acetonitrile/water/TFA (90°/o/ 10%/0.1% v/v/v) (solvent B) over 30 min at lmL/min using
a Waters Millenium system. Solvent A was water/0.1% TFA (v/v). Fractions
corresponding to the derivatized peptides were collected manually, and dried under
reduced pressure; they were reconstituted in acetonitrile/water (l :1 v/v) when needed.
Three proteins, as shown in Table 2.2 were tryptically digested and the peptide

fragments were TMPP+-Ac derivatized.

Table 2.2. Proteins Studied by LC-MS After Tryptic Digestion and Derivatization:

 

Protein Molecular Weight

 

Insulin Chain -B 3400.9
Glucagon 3482.8

Cytochrome-C 1 1702.6

 

Proteins were digested using 10% (w/w) of trypsin (sequencing grade) to protein,
at 370 C for periods of 30 min to 2 h in a phosphate buffer (SOmM) at pH 8.2. Once
digested, the mixture of peptide fragments was derivatized in situ, using the TMPP+-Ac
derivatizing reagent at a lO-fold molar excess over the estimated number of peptide
fi'agrnents. For example, if 1 nmole of a protein were digested, and 3 fragments were
produced, then there would be 3 nmoles of peptides, so 30 (3 X 10) nmoles of the
derivatizing reagent would be used. The mixture was allowed to react at room
temperature for 15 min. The derivatized tryptic fragments were separated by HPLC

(gradient program used was the same as described previously). The fractions were

43

collected manually (the R,’s of the TMPP+-Ac-peptides were between 28 and 37 min,
while those of the underivatized peptides were between 12 and 20 min), and subsequently
analyzed by ESI-MS on the Fison’s platform. Alternatively, the derivatized digest

mixture was analyzed directly by LC-MS using the LCQ.

B) Mass Spectrometry:
a) In-source Fragmentation (ISF) MS:
The purified TMPP+-Ac-peptides were analyzed by a Fison’s VG platform ESI-
MS. Samples (10 pmol/uL) were flow injected with acetonitrile/water (1:1) as the
mobile phase (10 uL/min). Nitrogen was used as the drying gas, and the capillary was
typically held at 3.5 kV. The quadrupole mass analyzer was Operated at 3 scans/sec over
a range of m/z 200-2000. Increasing the cone voltage in the ESI source to 100 V induced
fragmentation of the peptide derivatives. This is referred to as in-source fragmentation
(ISF). When no fragmentation was desired, the cone voltage was maintained at 37 V.
In-source fiagmentation (ISF) was also performed in the Finnigan LCQ, where
TMPP+-Ac derivatives were fragmented in the atmospheric region by increasing the
capillary and the tube lens offset voltages simultaneously to 59 V and 40 V from initial

values of 25V and 10 V, respectively.

b) MS/MS:
CAD-MS/MS of the TMPP+-Ac derivatized peptides was performed using two
different instruments. A triple quadrupole mass spectrometer, API 2000, manufactured

by PE—SCIEX, with an ESI source was used in one case. The peptide derivative (in

acetonitrile/water/O.1%forrnic acid, 10 pmol/uL) was infused at a flow rate of 5 pL/min

44

into the ion spray source held at 5 kV. The precursor ion selected in the first quadrupole
was collisionally activated in the second stage. Nitrogen was used as the collision gas.
The third quadrupole was scanned to detect product ions, in steps of 0.3 m/z for a dwell
time of 1 msec. Optimum conditions for efficient fragmentation were determined by
using a “ramp parameter” feature of the software. Fragmentation of singly-charged ions
of the TMPP+-Ac derivatives occurred at accelerating potential differences
(corresponding to lab frame collision energies in eV) of 50—70 V.

A Finnigan LCQ, an ESI-ion trap mass spectrometer, was also used to obtain
CAD information on the TI\/IPP+-Ac derivatives (acetonitrile/water/ 1%acetic acid, 10
pmol/pL) both by infusion of the sample (3 uL/min) and by LC-MS. The source was
held at 4.3 kV, and helium was used as the CAD gas. MSn was performed, where
n=1,2,or 3. Fragmentation of singly-charged TMPP+-Ac derivatives occurred at relative
collision energies of 60-70% of a maximum of 5 eV (LAB). LC instrumentation included
the Alliance system from Waters corporation, and the Symmetry (from Waters) HPLC
column (2.1x150 mm). Solvents were pumped at a flow rate of 200 pUmin. The solvent
composition and gradient are the same as described earlier. Approximately 10 pmol of

the derivatized, digested protein were loaded on the HPLC column for LC-MS.

c) MALDI-MS:

MALDI spectra were obtained with a Voyager Elite reflectron time-of-flight mass
spectrometer (Perseptive Biosystems) using an accelerating voltage of 22 kV. The mass
spectrometer was equipped with a 337-nm nitrogen laser. A saturated solution of a-

cyano-4-hydroxycinnamic acid, prepared in acetonitrile/water (1:1,v/v), was used as the

45

matrix. A l-uL aliquot of the sample (5 picomoles/uL) was mixed with an equal volume
of the matrix solution on the sample plate, and the mixture dried in air. PSD data were
obtained when the mass spectrometer was operated in the reflectron mode. Several PSD
spectra, each optimized for a different range of m/z values for the fragment ions, were

obtained and "stitched" together to yield the composite spectrum.

46

III. Results and Discussions

A) In-source Fragmentation of the TMPP*-Ac Derivatives of Peptides:

Infusion of the TMPP+-Ac derivatives of the peptides into the E81 source, at a
cone voltage of 37 V, usually produced doubly-charged ions. This is not unusual because,
along with the fixed charge on the TMPP+-Ac head group, there could be protonation on
a basic residue in the peptide. In larger peptides, triple-charged species are also observed.
However, in the case of TMPP+-Ac-VGVAPG, only a singly-charged ion is observed (the
charge is the one carried by the phosphonium moiety), due to the presence of amino acids
containing non-polar side chains. In the case of the VG-platform instrument, which has a
single quadrupole mass analyzer, the only means to obtain fragmentation and hence
structural information is by CAD in the source. In-source fragmentation (ISF) of the
charged derivatives of peptides is observed at elevated cone voltages. A recent article
elaborately describes the in-source fragmentaion of various peptides in a similar
instrument, now called the micromass platform (29). By elevating the cone voltage, the
potential difference between this electrode (the cone) and the counter electrode at ground
is increased. Thus, the species travelling through this region undergo a drastic increase in
kinetic energy, approximately tens of eV’s for a singly-charged species, which results in
fragmentation of the precursor ions. In our studies, a cone voltage setting of 100 V was
sufficient to fragment peptide derivatives with 6 - 12 amino acids to yield discernible *an
ions. Peptide derivatives with fewer amino acids fragmented at a lower cone voltage, and
those with more than 12 amino acids required a higher cone voltage. ISF of the TMPP+-

Ac derivatized peptide produces a series of *an ions, with a few *bn ions and *cn ions.

47

Derivatized peptides containing amino acid residues with side chains tend to yield "‘dn
ions in some cases. As an example, Figure 2.3 shows the ESI-ISF mass spectrum for
Th/IPP+-Ac-MNYLAFPRM-anride at a cone voltage of 100 V. The spectrum consists of
peaks corresponding to an almost complete series of “’an ions, along with a low
abundance series of *1)” ions.

While most peptides always give a complete series, it cannot be guaranteed. The
intensity of the peaks due to fi'agment ions of high-m/z seems to be lower than that of the
peaks due to low-m/z fragments. This may be either due to the mass discrimination effect
inherent to the quadrupole mass analyzer, or simply that the low-mass fragments are

formed preferably over high-mass fragments.

B) MS/MS of the TMPP*-Ac Derivatives of Peptides:
a) Triple Quadrupole Mass Spectrometer:

The ESI source of the API/SCIEX instrument produced both doubly- and singly-
charged ions of the intact peptide derivative under the given conditions, and CAD was
performed on both types of ions. Regardless of whether the singly- or the doubly-
charged species was selected as the precursor, the product ion spectrum was the same,
i.e., in both cases peaks were observed at the same m/z values, although the peak
intensities differed by as much as 40% at a given m/z value. Since the ion count for the
doubly-charged species was higher than that for the singly-charged species, the MS/MS
spectrum of the former had better S/N. Fragmentation of the precursor ions occurred at
lab-frame collision energies of 60-70 eV. A series of "'an ions was observed from
fragmentation of either the singly- or doubly-charged ions, along with a few *bn and *c.,

ions.

48

.2668 A70 ofimfiamm 05 8 omega—o Bob 82 Eon—wet 8 26 833 9:58:00
Exxon 05 950% 8.0.5 2.2. .oEEm-2Mmm<A>ZE-o<-L%.—. no 500% $9: mmzmm .m.N 3:me

cos F com: com. v 83

come OONF core

. coo?

 

 

ea 4?... a 3.

L I?
1
Ir C.-
9 Z .7
a .0 we.
9 .
D 8
+ e 3...
a”. r
m
c.

and own 05 08 can own 0: own own 9R oak on~ o7~n och cam 08 osm 80 000

l 4 3K .3
«.2:

‘

own quz 34

q -

Z

N
H

 

44# it. < 1.. I 117
*—
H 0.1.
9 0
9 6
m s m
8. Ir
a. r .. 7.
S 8 be
L 1%.. 0...
8.7V .V £8
a... . ...
.7 B
s

4% was OK.

0. F2.

3...

I

C
2662 ocmwflag 05 3 0933—0 8 26 82 Eofiwfim

T

 

E

.08 own 2: 08

 

 

 

 

 

tiwwldrtlislhlhrl
,_ _
d...
mm m. g
.0 .9 1.
SI. 0 .8
aw 7v 8
7w» .0 e a a
8 q l 1
Ir fill flu
.1 a N ..u.
8 I .d M
9 e .d . I
9 a. + u
8 z c I?
. V a
...v u
3 S
3 le
+ ..M
m
88
+9
.1.
8 Q:
0.
m
+

49

A representative spectrum, for the CAD of TMPP+-Ac-GGSLYSFGL-NH2 in the triple
quadrupole, as observed from the doubly-charged precursor (m/z 836) is shown in Figure
2.4. It is readily noticeable that a complete series of *an ions is formed, along with some
*bn ions.

b) Ion Trap Mass Spectrometer:

Figure 2.5 shows the ESI-CAD-MS/MS spectrum of the underivatized peptide
(singly-charged) Neuromedin-B. Figure 2.6 shows the MS/MS spectrum for the TMPP+-
Ac-derivative of the same peptide, obtained from the singly-charged species (1704 Da).
The spectrum from the underivatized species shows both C-tenninal ( e. g., m/z 433, m/z
592, m/z 662 etc.) and N-terminal ions ( e.g., m/z 682, m/z 700, m/z 837 etc.) along with
some internal fragments (e.g., m/z 514). The spectrum from the derivatized peptide
shows peaks corresponding to N-terminal fragments. The chemistry during the
fragmentation in the trap is rich, as represented by many more fragment ions than seen
during fragmentation of similar peptides in the quadrupole mass spectrometer. There are
peaks corresponding to loss of water and ammonia from the precursor and the fragment
ions, which are low energy processes that occur commonly in the trap.

The type of fragment ions produced from the TMPP+-Ac charged derivatives of
peptides in the LCQ mass spectrometer are different, depending on selection of either the
singly- or the doubly-charged intact species as the precursor. An example spectrum for
the fragmentation of the doubly-charged TMPP“-Ac-GNLWATGHFM-NHZ in the ion
trap is shown in Figure 2.7. This can be compared with the MS/MS spectrum from the

singly-charged precursor in Figure 2.6.

50

.>e we go 385 5658 a “a
36888on Bee 2833:: 035 a E ems NE. «a c8585 3&2?»ng

.oEEm-qOmm>quO-o<-+mm—2Hmo 8.2.50on mng-Q<UAmm .vN PBwE

 

 

 

 

 

N}:
. 8.: 89 BE 8.: 8.2 8m 8»
N: A q ‘tqq “*1 . 1.. . lttfl J “nag
v
+0 NM D...
a . an.
a. mm.
on
em

 

com com

3......
0

 

 

MrsuerUI annalau

 

.00 F

51

c355 5658 2538 $9,“ an HeHoEoboeam mme nab :2 an E
oEEw-2EOH<>>AZO me 88:85 unmeasurnfifim Co .550on mngAH<0Amm .m.N 0.5me

N E
ooFFomowoooF 0mm com omw cow cm» 00% one com com com omv oov 0mm

can. A— irttifrhrjfii arthrttrr C: . .1; Fri. alias leriJCfiL L, rafirFLifLa‘trfij J... .2“ .r h. is. .1 4:... O

9
8
z

.L

c.
I.
I C I

0

'7

..q\

9

 

 

”a 8'608
002

 

.A
c.

 

ozH-wr rm

0'269

‘11 6'28?
5
N

:IHOLV any";

o‘H-‘q
9x 2299
“A

 

 

17881 L ___-
E:
«a

+HW
if OZHHHW E'SLLI

l

l

l

l

l

l

l

l

”E
w

 

sq eves
lb
‘0

 

Mrsuerul engines

52

{coo mo 38:... ”53:8 0338 a 8

888288on 82: an: :2 an E 02892mmmvh<3gzce<fimm29
me 88:85 vowsno-bw£m me 8:50on m§m2-Q<U-_mm 6N onE

N}: cot. 000—. comp. Go: 009. oomw 00—. Door com com con

5.9.1.1.th >r~ : F

. _
:3: r:

 

 

 

  

 

 

 

 

 

 

 

 

 

w .£.h,t.r..t..._..1 ..L 8:4.— _ :7 Ta. 7: .11. _.::._.r.
j g .3 a: z . 2 .
_
Ir Ir Ir Ir
I. Sly .7 Ir Iv Ir
a 9m. m n u a o ..
+0 I an n I .Vlmo owe. M 0 9 .l. I.
9 .3 on 1 Z on Let r ...3 l 0 7cm"
8 e e .7 I. 08 0? n N m3]
.9 T. .1. an . en a.
8 z .7 e e "a N“
o a“ I C 7?.
. H o
+ 8 .
- ne “0 a m
mu 7. c
O
O ”a.
.7
.
3
H
8
3
H
O

17'099 L

Nx

53

hirsuerul extrema

 

Z'ELS
9*“!de

 

 

53mm 40 35:0 53:8 02:28 a 3

8.308880% mmaE nab :2 :w E 028e12mmck<34206<¢mmgh
mo 88:er BMBBABBB 40 83.58% mgm—Z-Q<U-Hmm SN 8=wE

 

 

 

 

 

 

 

a}:
con— comw Comp 003. com? 89. Do: 000— ooa 08 cos
_ 8i» 4F_htr _ L 4P. 4 1-.-... 1.1..4LrLif4LJLirlriiiJagt 4841‘)? 1817.34 .J {al.—:jfjn‘dlj‘ #1333
w. l w . -H
..oo m n/ w
a _ 1.1. m> M
m u w. . .s e a.
a e m 3. «w m 1 .mn
. ne mg + W o
_ . H
We +N e w
5 ./
e +N
m
.9
a
+« n.

 

l "811

Aug.

T—

'ITT'

Tflrrr.r11

Kirsuarul 3.1911133

r1 1hr *fif

 

l

It can be seen that fragmentation of the singly-charged precursor yields an
uninterrupted series of *an ions along with *bn ions, as opposed to the case of the
fragmentation of the doubly-charged precursor, which gives a series of *bn ions, and
some doubly-charged (*bfi-H") ions, along with a few C-terminal ions noticeable in the
lower m/z region of the mass spectrum. It may be that the fragmentation pathways
available to the doubly-charged ions are charge-mediated by the proton, which is mobile,
while fragmentation of the singly-charged ion (the fixed charge) occurs by a charge-
remote mechanism. It has also been reported by Lin and Glish [8] that a series of "‘an
ions is observed for TMPP+-Ac derivatized peptides when a singly-charged precursor

generated by FAB is fragmented in an ion trap.

C) Comparison of the Fragmentation of the TMPP*-Ac Derivatives of Peptides by
ISF vs. MS/MS:

Although fragmentation of the TMPP+-Ac derivatives of peptides by MS/MS in triple
quadrupole and ion trap mass spectrometers is more specific due to precursor selection,
essentially the same information can be obtained from in-source fragmentation (ISF).
Figure 2.8 shows the results from analyzing TMPP+-Ac-GGSLYSFGL-NH2 by ISF and
MS/MS. A complete series of *an ions is obtained in each case. Signal due to the intact
molecular cation (C) is not completely diminished with ISF (Figure 2.8a), indicating that
this process is not as efficient for fragmentation as conventional CAD. However,
MS/MS studies provided the opportunity to compare the fragmentation pattern
originating from a singly-charged ion with that fi'om a doubly-charged ion. The MS/MS
technique also allowed us to observe the fragmentation of the various precursors at

various collision energies.

55

8

- A A

an

Relative Intensity

 

 

 

600'

 

 

aas as c...
.37 .38

. .,.| . ...l ..l -,
1000 1200 1400

 

 

 

 

 

 

 

 

tbs
. Ill .U WILL
1200 400

 

 

Figure 2. 8. A comparison of the fragmentation of TMPP+—Ac-
GGSLYSFGL-amide by A) In-source fiagmentation, B) CAD-MS/MS in
a triple quadrupole (doubly-charged precursor) C) CAD-MS/MS in an
ion trap (singly-charged precursor) mass spectrometers.

56

D) Comparison of the Fragmentation of the TMPP*-Ac Derivatives of Peptides in
Triple Quadnipole vs. Ion Trap:

It appears that the TMPP+-Ac derivatives of peptides fragment through the
charge-remote pathways (based on the formation of a-type ions) when subjected to CAD-
MS/MS in the triple quadrupole mass spectrometer both as doubly- and singly-charged
precursor, and as a singly-charged precursor in the ion trap. However, during the
fragmentation of a doubly-charged precursor (one charge on the derivatized moiety, the
second due to a proton along the backbone) in the trap, some C-terminal ions were
detected along with the N-terminal ion types. These results suggested to us that perhaps
due to the long (tens of msec) residence time in the trap, the large number of low-energy
collisions allow observation of both charge-mediated and charge-remote fragmentation.
During the fragmentation of the doubly-charged precursor, doubly—charged fragment ions
were also commonly observed unlike fragmentation in the triple quadrupole and by ISF.
Indirect evidence for these phenomena was acquired by in-source fragmentation where
the trap was used only as an m/z analyzer (in which case the residence time of the ions
was on the order of tens of microseconds), rather than as a CAD device with subsequent
m/z analysis. The result of ISF with m/z analysis by the trap is shown in Figure 2.9,
where the mass spectrum essentially represents "‘all. ions, since processes leading to
charge-mediated fragmentation did not have sufficient time to occur. However, the
doubly-charged ion did not seem to fragment by charge-mediated pathways during
MS/MS in the triple quadrupole mass spectrometer. The quadrupole ion trap and
transmission quadrupole mass spectrometers are distinctly different, i.e., in the

transmission quadrupole, the residence time of the ions in the collision cell is short (<100

psec).

57

c+

888‘8
I

Relative Intensity
a a 8 83 8 8% _

“80(8-

 

 

mlz

Figure 2.9. In-source fragmentation of TMPP+-Ac-GGSLYSFGL-amide
followed by m/z analysis of fragment ions in an ion trap mass
spectrometer.

58

In the ion trap considerably more collisions will occur and, in principle, higher internal
energies can be deposited in an ion, but in smaller increments. In addition, due to the
longer reaction times, an ion that has accumulated sufficient internal energy to
decompose by a low-energy process may not survive to undergo another activating
collision [32]. Thus, it is not surprising that the mass spectra for the fragmentation of
doubly-charged ions in the trap have peaks corresponding to fragment ions produced
from charge-mediated processes. The presence of peaks corresponding to neutral loss
such as water and ammonia from both the precursor and fragment ions also suggest the

predominance of lower energy pathways.

E) Speculation on the Fragmentation Pathways of TMPP”-Ac-Peptides:

The major pathway by which an aII ion can be formed during fragmentation of the
peptide backbone is by loss of CO from a bn ion [13]. In the case of charged derivatives,
1,2-elimination is responsible for *an ion formation [33], which is a charge-remote
mechanism. Recently, a mechanism for ”'11.. ion formation in the case of TMPP+-Ac
derivatives of peptides has been proposed by Liao et al. [7], through the shifi of an amide
hydrogen.

Although in FAB-CAD (high-energy), *bn ions were not commonly observed,
they are observed during our low-energy CAD studies by ESI-MS. Higher laser powers
in MALDI-PSD also produced "‘bl| ions. Charge-remote mechanisms for bn ion formation
in multiple-charged ions by low-energy CAD have been proposed earlier [34]. Wagner
proposed charge-remote mechanisms for *b. ion formation in charge-derivatized peptides
when studied by FAB-CAD [35]. According to this, formation of *b.. ions by charge-

remote mechanisms in TMPP+-Ac-peptides could be from (at-hydrogen shifts.

59

During the fragmentation of doubly-charged ions of TMPP+—Ac-peptides (hence
protonated along the backbone) by low-energy CAD (other than in the ion trap), if
charge-mediated processes occur predominantly, there should be some characteristic
evidence in the spectrum. For example, if a "'b.I ion is formed by a charge-mediated
process, it should carry a charge on its carbonyl end (C-terminus). This process will
make it a doubly-charged ion (as there is already a fixed charge on the N-terminus),
which would be represented by a peak in the lower m/z range. This was not observed
during CAD of any of the model peptides (ranging from 5 to 15 residues) that we
investigated. Further, no complementary y-ion signals were widely observed. Hence,
charge-mediated pathways can be ruled out as producing the *bn ions. Conceivably,
prior to the fragmentation of the doubly-charged species, charge-stripping (removal of the
additional proton) may occur yielding a species with a fixed charge (C), which
undergoes fiirther fragmentation (presumably via charge-remote pathways) both in ISF
and MS/MS. Since charge-stripping is the process that requires the least amount of
energy, this assumption seems reasonable. However, during fragmentation of the doubly-
charged ions in the ion trap mass spectrometer, a mixture of N-terminal and very few C -
terminal ions was observed which led us to conclude that charge-mediated pathways may
also be available. Also doubly-charged fragment ions were commonly seen during
fragmentation of the doubly-charged precursor. This suggests that charge-stripping may
not be complete in the ion trap mass spectrometer, which allows some possibility for
charge-mediated fiagmentation. This does not suggest that in all cases where *bn ions
are formed there is some charge-mediated fi'agmentation. In fact, the predominant

pathway is probably charge-remote. However, the origin of the shifting proton during

60

their formation is not clear at this point. Labeling studies using deuterium at various sites

of the peptide would help identify this mechanism clearly.

F) Characteristic Fragmentation Behavior of Some Amino Acid Residues:

There are 20 amino acids and it is important to understand if the fragmentation of
all the amino acids is the same or different when present in the sequence of the TMPP+-
Ac derivative. For most amino acids, present in the sequence of a charge-derivatized
peptide, the predominant ion formed is *a... In the case of the quadrupole mass
spectrometers the high-mass fragment ions are less abundant than the low-mass fragment
ions; and low-mass *an fiagment ions were accompanied by *bll and *c.. ions. However,
in the ion trap such discrimination was not observed. For residues such as serine,
threonine, valine, and leucine, *1!“ ions were often observed. But this was not consistent.
So it cannot be generalized or used as a rule in the identification of these residues. Some
characteristic fragmentation patterns were observed for the following residues and an
effort to generalize this observation is made in the following text, which will aid in
identifying the presence of the residue in unknown sequences.

a) Proline (P):

In our study of a variety of derivatized peptides, whenever proline (P) was present
in the amino acid sequence, there was no conspicuous signal due to cleavage at that
position during ISF in the VG platform MS in most cases. Figure 2.10 shows the E81-
ISF mass spectrum for TMPP+-Ac-VGVAPG; it is clear that a series of "an ions is
produced except for *as, which corresponds to cleavage at the position of the proline

residue.

61

.Ammb cam—ea mo 88 2: Ha omega? no.“
300$ .22 pm. we 85% a 959% 838on .568onon $42: EomEcgcofifim
a E Om<>0>-o<-+mm§.:o 82.58% mme Q<U-ooSem-Hmm .ofim oSmE

 

NE 82 25 com o2 o8
c
.< :4 1 <4 4 :4 Id
\ 35 4 L42 28
\ x. x.
.7. $532 emu N... 2.8 7%.. 98m
2:2; 3 3..
8
9s .+ .
D «Mum
mg a
.. L m
Do
I w ...A...
.d 2;”
.m m.
. a
w m.
AM
m
H
JV
[T
32

 

 

62

This provides a significant clue on the mechanism that is probably operative in
the formation of the "'all ions during low-energy CAD. Since proline lacks an amide
hydrogen when present in a peptide sequence, it is probably the amide hydrogen shift that
results in the formation of the *an ions. It has been already been proposed by Liao et al.
[7] that the lack of an amide hydrogen on a proline residue, as shown in Figure 2.11, does
not allow the mechanism for *an ion formation in the TWP+-Ac derivatives when

analyzed by MALDI-PSD.

HVH

H2 H2
WN- H-CON

Figure 2.1]. Structure of proline (P) residue.

However, a weak signal corresponding to cleavage at the proline residue is
observed during fragmentation of certain peptides like TMPP+-Ac-FYGPV and TMPP-
PHPFHFFVYK during ISF-MS. During our ESI-CAD-MS/MS study of the peptide
derivative of Th/IPP+-Ac-VGVAPG a weak signal is observed for *as and *b5 ions. Also,
a weak signal was obtained in MALDI-PSD for cleavage at proline at higher laser powers
[7]. Thus, the mechanistic aspects are unclear. These observations suggest that, if during
the analysis of an unknown sequence there is a large mass difference in between two
peaks that cannot be filled by one amino acid residue, it probably implies the presence of

proline in the corresponding portion of the sequence.

63

b) Threonine (T):

A threonine residue has a hydroxyl containing side chain. Loss of water from the
precursor is a common observation due to the presence of such residues during CAD-
MS/MS. Other noticeable features due to fragmentation at T were the formation of a *dn
ion and the formation of an *a.-44 ion. The formation of the *an-44 could be due to the
loss of the complete side chain as CH3CHO. The structure of T is shown in Figure 2.12.
This can be seen from the MS/MS spectrum of TMPP+-Ac-GNLWATGHFM-amide
(Figure 2.6). This is also observed in the ISF—CAD-MS spectrum for TMPP+-Ac-

VHLTPVEK ( Figure 2.20).

H
H3C- H
-NH-CH-CO-

Figure 2.12. Structure of threonine (T) residue.

c) Aspartic Acid (D):

Aspartic acid (D) is an amino acid with an acidic side chain, as shown in Figure

f0

CH2
W‘NH-CH-CO

Figure 2.13. Structure of aspartic acid (D) residue.

2.13.
HO\

 

64

Fragmentation of a peptide derivative containing aspartic acid, does not yield an
*all ion; it yields a *dn ion, due to the loss of the side chain at the site of D. It has been
proposed that the *d.., *bn, and *c...1 ions are formed via cyclic intermediates involving
the side chain, and hence are formed preferentially over the *all ion [7]. Mechanism for
formation of *b.., *d.., and *c..-l ions as adopted from this reference are shown in Figures

2.14 a, b and c, respectively.

 

”\7_ /°
NH-CH-CO—NH-é‘l O
O;
NH
i \
“\7' //0
C—C

NNH'CH’CO-Nfl'éi B
\

if.

Figure 2.14a. Mechanism for the formation of ""bI| ion due to fragmentation at D.

”\7'_ /0

A—n)
NH-CH-CO-NH-CH O

 

l”’
NH-CH-CO—NH H + C<OO

 

+3

,./Z—O
:1: x
0

Figure 2.14b. Mechanism for the formation of *dn ion due to fragmentation at D.

65

"for:

NNH-CH-CO-NH" ,C_.H
\ H

W“

/o
NNH-CH-co—NH, + C
\
er] 0
+
O=C~CH=CH2

Figure 2.14c. Mechanism for the formation of *c..-1 ion due to fragmentation at D.

Figure 2.15 is a portion of the ESI-ISF mass spectrum of TMPP+-Ac-
APGDIRYVHPF. Fragmentation at D, shows a prominent peak for the *d3 ion (no *a3
ion) along with peaks for a *b3 ion and a *cz ion. Similar observations have been made
for other peptide derivatives containing D. Thus, cleavage at the site of the aspartic acid
residue yields a signature pattern (*cM, *b.., *d..), which, even though it disrupts the "an
series, should not be difficult to identify due to the characteristic presence of the three

peaks as a cluster.

66

.64 2668 Ben 2:33 40 F533 :ouficoEwEm 05 wagon,“
Emiaaemifiamzto 5888. was 652% 05% SEE in can

a}:
com com com ova own com om“ ooh _ ova . ow» 005

L .1! 40"”‘1 'Ivl I." . vi
4! 11< 11% 1 it

”ill" LOH.

41.14 1.. -l .4144 411

  

.. [OLA -<

6
Iv
.9
8
I.
q
v

in 9'9L8

8

Se
Kuwaiti] analog

i’p‘ 9' WQ

 

67

d) Asparagine (N):

Asparagine is the amide version of aspartic acid ( residue structure shown in
Figure 2.16), where the —OH on the side chain is replaced by —NH2. Thus, a similar

behavior as that of aspartic acid could be expected for asparagine.

H2N\ /0

CH2

 

\/\/‘NH-CH-CO

Figure 2.16. Structure of asparagine (N) residue.

An example of a peptide containing N, TWP+-Ac-MNYLAFPRM-amide is shown in
Figure 2.3. It can be noticed that as in the case of aspartic acid, prominent peaks due to
*c..-1, *dnfbn ions are seen. Along with these, *all and *an - NH; ions are also seen. Thus
a similar mechanisms as shown in Figures 2.14a, b and c can be assumed operative even
for asparagine. However, the mechanism for the formation of the "an ion is also
operative as seen from the *az peak in the spectrum along with the *az-NH3 peak.
Although this is a cumbersome pattern of fragmentation with the appearance of 5 peaks
due to one single residue, it is still characteristic and can be used as a bench mark in the
case of unknowns. The presence of the two additional peaks (*aIn and *an — NH3) in the
case of asparagine along with the signature as seen for D (*c..-1,*d..,*b..) , will help

distinguish between the aspartic acid and asparagine residues.

68

e) Glutamine (Q):

The structure of Glutamine is shown in Figure 2.17. Glutamine has a structure
similar to asparagine with the side chain longer by one methylene unit. Hence, a similar

fragmentation pattern to that of aparagine can be expected for glutamine.

H2N\f¢0

CH2
CH2
\/\/\NH-CH-CO

 

Figure 2.17. Structure of glutamine (Q) residue.

However, the results from two peptides ( TMPP+-Ac-APSGAQRLYGFGL-amide and
TMPP+-Ac-IFVQK) containing Q suggests that fiagmentation at the site of glutamine
results in *c..-1 and *an ions. There is no conspicuous signal for *dn or "’bn ions. Loss of
ammonia from the *an ion is not observed either. Thus, the presence of one additional
methylene unit drastically alters the fragmentation pattern of this residue because it
disrupts the intermediates required to form the expected ions. The spectrum of TMPP+-
Ac-APSGAQRLYGFGL-amide is shown as an example to demonstrate the
fragmentation pattern of Q in Figure 2.18. Here Q is the 6th residue, and peaks for *c5
and *a6 ions are seen in the spectrum. Similarly, in Figure 2.23, the spectrum of TMPP+-

Ac-IFVQK, peaks for *c; and *a4 ion are observed.

69

$5988 .«o 82 305 9C 2662 2:59.» 2.:

@8528 £84 £2:me .5420. a 5 83%.: 9% 2662 oEESEw 05 “a 5:855me
.ov_§-AOm0>AMO<Omm<-o<-+mg me 858% mzimzmm 05 .«o newton— .w_.N 2&5
E88? 1 8m. . 8: . 8mm. 89 . 83 . 82 89. -81- F881

 

o

 

 

 

 

 

 

 

 

C1.
‘ t. .. .1 4. _‘__,.....;_......§
Ir a 9 Ir Ir
6 8 U. 9 L 8 H
9 e _. z 7v 9 WW
8 B e Ir 0 lo
a u ..m .m. ... A
+0 .N .N 0. he . a OD; #3
H H m . m an I
c. C N 9 NM .96 m
c ...H O ..9 3
.6 a. 9 L u
0 L. a. W «h.
. MJ ”.9 .M
on c N 0 I. 6.
N H Ir A
c . 0a 8
H 8 ..
s “a 8 8—
. no

«I

70

A similar result can also be expected for the residue glutamic acid, which has the
acidic functionality instead of the amide functionality in glutamine. In the list of peptides
studied, only one has the residue E, TMPP+-Ac-VHLTPVEK, and the fragmentation at E
for this peptide is not very intense to draw any conclusions. However, there are weak
signals observed for *c6 and *a7 ions (Figure 2.20). Roth has observed *dn and *c..-1 ion
formation for fragmentation at E during a MALDI-PSD study of the peptide TMPP+-Ac-
AEKAA and has proposed mechanisms similar to that shown in Figure 2.14a, b and c for

their formation [36].
f) Histidine (H):

Histidine is one of the three residues (Structure shown in Figure 2.19), that

contains a basic side chain.

 

WNH-CH-CO

 

Figure 2.19. Structure of histidine (H) residue.

The fragmentation at histidine results in both ’"an and *bn ions. The intensity of the peak
due to the *b. ion is higher if not equal to that of the peak due to the *an ion. An
example spectrum is shown in Figure 2.20 for the peptide derivative TMPP+-Ac-

VHLTPVEK.

71

b scam

>I'+3 asset;
371221 «1

4
q .

,Jz'eszug

391611....

20171 L “‘1

 

 

P §'196-+

Viv-’8 J

.. 7'768 4,-

.“ 1

,5“ . 4

. 7608 .
«V'LQL

 

.~ “nu-Zn. 6'992 ’

.5 921.9

 

..-
9'1779 ' \

9'92; °V'+J<IW.I. 4

 

14

J. o

Kitsuaml aAtmlau

100'

72

1300

1000 '

m/z

1500

Figure 2.20. ESI-ISF mass spectrum of TMPP+-Ac-VHLTPVEK showing

1400

1200

1100

1.700 300 900

600

characteristic fragmentation of H residue and T residue at positions 2

and 4 respectively.

The peak for *bz is more intense than the peak for *32, This observation was also made
for TMPP+-Ac-PHPFHFFYK and TMPP+-Ac-HWSYGLRPG-amide. The structure of
*b. is probably stable when H is the terminal residue in the fragment and hence it is

possibly formed preferentially.
g) Arginine (R):

Arginine is another amino acid (Structure shown in Figure 2.21) with a basic side

chain along with histidine and lysine.
H2N\C/NH

NH

(CH2)3
NH- H-CW

 

Figure 2.21. Structure of arginine (R) residue.

Cleavage at arginine for TMPP+-Ac-Peptides results in an *an ion with loss of ammonia;
all the fragment ions that contain arginine seem to lose ammonia, too. This behavior is
demonstrated in the CAD-MS/MS spectrum (Figure 2.18) for TMPP+-Ac-
APSGAQRLYGFGL—amide. There is a series of *an — NH3 peaks after residue 7, which
is R. The other feature noticeable is the *c...1 ion which is *c6_ A similar observation is
made for TMPP+-Ac-YGGFLRKYPK. Formation of the *c..-1 ion is not dependent on the
residue present before R, because the residue L does not produce a characteristic *cn

fragment when present in other peptide sequences. It is due to the influence of R being

73

the successive residue in the sequence; a mechanism similar to the one shown in Figure
2.140 can be assumed. In the case of arginine, the proton transfer for the formation of the
*c..-1 ion probably originates from the —NHz group of the side chain. These patterns
suggest the presence of R in the sequence. Although these characteristics are not very

diagnostic, they do provide some clues to identifying the presence of R in the sequence.

G) Comparison of the Fragmentation of TMPPi-Ac-peptides by ESI-ISF vs. by
MALDl-PSD and FAB-CAD:

TMPP+-Ac-peptides have been thoroughly studied for their fragmentation
behavior by MALDI-PSD [7] and FAB-CAD [6]. Hence, it is meaningful to compare the
results from ESI-ISF with those from the other two methods. The results from ESI-ISF
match more closely with those from MALDI-PSD than those from FAB-CAD. For
example, analysis of a derivatized peptide containing proline by F AB-CAD always yields
an *an ion due to cleavage at the site of the proline residue, while in ESI-ISF the signal
for this ion is weak and, in MALDI-PSD, the signal is laser power-dependent [7]. The
formation of a *dll ion from fragmentation exclusively at the position of aspartic acid is
observed with both ESI-ISF and MALDI-PSD. In FAB-CAD, although a *dn ion is
observed, there is also an "'a.I ion without the loss of side chain.

These variations in results can be attributed to the differences inherent in these
methods of analysis; high-energy vs. low-energy. High-energy CAD spectra are usually
the result of single collisions between precursor ions and the inert gas-phase atoms, while
low-energy CAD results from multiple collisions and reflects multi-step cleavage
reactions [37]. In FAB-CAD ( high-energy) , dissociation is induced by imparting excess
energy to a selected precursor ion, promoting high-energy fragmentation. In MALDI-

PSD, although a precursor can be selected, CAD is not promoted intentionally. The

74

fragments formed either during desorption (prompt fragmentation) or from metastable
decay of the precursor during flight are identified using the reflectron. Since it is possible
to control the fragmentation by adjusting the laser power, most probably fragmentation
occurs during the desorption/ionization (D/I) process. Therefore, it cannot be called a
true collisional process. Although the energy imparted in the process is quite enough to
promote fragmentation, it is not like the high-energy CAD in FAB. Similarly, in ESI-
ISF, the increase in cone voltage promotes more energetic ion/molecule collisions, but
the amount of energy transferred to the fragmenting ion is unknown. However, because
ISF is a low-energy process, fragmentation is not expected to be similar to that in high-
energy FAB-CAD. Figure 2.22 compares the spectra of TMPP+-Ac-GMDSLAFSSGL-
NH; as obtained by all the three ionization techniques. All three spectra show a complete
series of "'a.| ions; a few *bn ions and *cn ions are observed in all of them especially in
the lower m/z region. Note that due to cleavage at the aspartic acid residue, *a3 is seen in
the FAB-CAD spectrum in addition to *d3, while in MALDI-PSD and ESI-ISF spectra,

only *d; is seen.

H) ESI-CAD-MS/MS Study of TMPP”-Ac Derivatized Tryptic Digests:

We have extended the study of derivatized synthetic peptides to derivatized
peptide fragments produced during tryptic digestion of proteins. Advantages due to
TMPP+-Ac derivatization are explicit during the chromatographic separation of digestion

products.

75

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

100 . 'a,
da C“
.86
C2 b3 35
a1 ‘
a2
a .
7- a10
J as ’a 'a
11
1 Lr 1-: .147 '1‘] ‘7: . :4. x; r: I: :- 3:1 -—.—. ‘2‘; h
>. 600 800 1000 1200 1400 1600
at: 100 (3+
"’ 'dl
C 3
3 .ba 6.4 g
5 a11
@ i
.2 a ‘ . .
*5 a2 ' as . a7 a8
— . ‘ a o . '
g 31 . 6 39 a1O ]
600 800 1000 1200 1100 . 1600
. C+
100 811
. . .a I
1 . .a . I
Ia 82 .a 4 a5 . . '1
1 ,.
‘h 36 as I.
3 . . 1
- 11»
r 51 31
600 800 1000 1200 1400 1600

Figure 2.22. A comparison of the fragmentation of TMPP+-Ac-GMDSLASFGGL-
amide by ESI-ISF-MS, MALDI-PSD-MS and FAB-CAD-MS respectively.

76

Certain peptides, which are of short length, have very short retention times; after TMPP+—
Ac derivatization, the retention times are increased, thereby permitting better separation
and detection of these species. Also, the detection limit for fragments in protein digests
on a conventional reverse phase column can be increased by derivatization because the
molar absorptivity of the TMPP+-Ac derivatized peptides is much higher than that for
underivatized peptides. Hence, even if picomoles of the protein are digested, TMPP+-Ac
derivatized tryptic peptides can be separated and detected by conventional HPLC using
an absorbance detector.

It should be noted that the TMPP+-Ac reagent undergoes rapid hydrolysis to
produce TMPP-CHzCOOH during its reaction with the peptides. TMPP itself is another
by— product. These products are represented by major peaks in the chromatogram. Often
a major peak for TMPP-CHZCOOH occurs amidst the peaks corresponding to derivatives
of some of the peptides. Fortunately, there were no cases of co-elution of the reagent-
related species and derivatives of the peptides reported in this paper. Problems due to co-
elution could be overcome by LC-MS using the mass spectrometer for precursor ion
selection.

a) Cytochrome-C:

The chromatogram for the HPLC separation of TMPP+-Ac derivatized tryptic
peptides of cytochrome-c is shown in the top panel of Figure 2.23. The peak at ~35 min
corresponds to TMPP-CHzCOOH. It was possible to identify 15 TMPP+-Ac derivatized
tryptic fragments from cytochrome-c, whereas in the underivatized mixture only 13 were

identified.

77

 

TMPPf-CH2000H
TMPP-Ac-MIFAGIK

/

 

 

 

 

Absorbance (mV)

 

 

 

 

 

 

 

 

 

 

 

 

 

l
Time(min)
0 111}
o
C.
1‘; .
1: a2
3 . 573 *a1
2 i i *a3
0 . w
..>.
H
U . 1 g 1
2 Oidlluhnm

  

 

 

 

Figure 2.23. LC-MS results of the TMPP+-Ac-derivative mixture of the tryptic
digest of cytochrome-C. Top panel shows the chromatogram for the
separation of the mixture. Bottom panel shows the ESI-ISF mass spectrum

of a peptide derivative (residues 80-86), TMPP+—Ac-MIFAGIK.

78

While some tryptic fragments were unidentified (by MALDI-MS or ESI-MS) before
derivatization, it was possible to detect them after derivatization, especially the ones with
masses less than 500 Da.

The ESI-ISF mass spectrum of TMPP+-Ac-MIFAGH( (see labeled peak in the
chromatogram), representing the TMPP+-Ac derivative of the tryptic fragment containing
residues 80-86, is shown in the bottom panel of Figure 2.23. A series of “'an ions, from
which the sequence can be deduced, is seen. In some cases, the signal due to
fragmentation of the last residue in a tryptic fragment is not detectable. However,
because the C-terrninal residue can only be lysine or arginine, absence of said peak
should not preclude identification of the complete sequence. Note also that most of the
fragments observed in bottom panel of Figure 2.23 are "'aII ions, and that no *bn or "'cn
ions are present.

Analysis of the same sample by online LC-MS was performed using a Waters
chromatographic system and LCQ mass spectrometer. During the LC-MS run, the LCQ
was operated in the dual scan mode (double play). A full scan was first performed, and a
subsequent MS/MS scan was performed at relative collision energy of 65% on a list of
precursor ions. The ESI-CAD-MS/MS spectrum of TMPP+-Ac-IFVQK (residues 9-13)
is shown in Figure 2.24. A series of *an ions is seen in the spectrum along with an
intense peak at m/z 1078, corresponding to the *c4 ion or loss of the terminal K residue
from the precursor (C+).

The following example shows that it is possible to obtain simple and interpretable

spectra from TMPP+-Ac derivatives of peptides whose sequences are not known a priori.

79

 

 

 

 

 

50

 

 

1
i
la "1 *
:20; 65%1
0 ‘ a
E : a2
- 3 005 *33
g i 904
'13 i 573
g : .
10: 9
1
‘ l
1
llllillllj iIIIJIH—iiliTiI]IAIII]llll] rlTr
600 700 800 900
mlz

1000

1032
*b4i

1061

 

 

 

1078

 

1100

(C*- K)

1162

Figure 2.24. ESI-CAD-MS/MS of TMPP”-Ac-IFVQK, peptide derivative from
tryptic digest of cytochrome-c, in an ion trap mass spectrometer at 55% relative
collision energy.

80

While analyzing the derivatized mixture of tryptic fragments of cytochrome-c by LC-MS,
MS/MS was performed on a precursor ion of m/z 1179.0. From the calculated masses of
possible tryptic fragments and their derivatives, 1179.0 Da corresponds to TMPP+-Ac-
GITWK. However, the peaks in the MS/MS spectrum did not match the hill of any ions
that would be produced from fragmenting this derivatized peptide. Upon closer
examination, it was possible to deduce the sequence of the derivatized peptide as being
MIFAGI, corresponding to residues 80-85; the mass of TMPP+-Ac derivative of
MIFAGI (where the C-tenninus is not an acid) is also 1179 Da. This peptide probably
originated from the loss of K from the tryptic fragment MIFAGIK. Thus, this example
demonstrates the applicability of correlating the MS/MS spectra of the TMPP+-Ac
derivatives of peptides to their amino acid sequence, due to a major series of *an ions, as
shown in Figure 2.25.

b) Glucagon:

Another protein studied was glucagon, HSQGTFTSDY SKYLDSRRAQ
DFVQWLMNT, 29 residues in length, with three cleavage sites. Three tryptic fragments
were obtained corresponding to residues 1-12, 13-17, and 18—29. The TMPP+-Ac
derivatized fragments were prepared, and then separated by HPLC. Figure 2.26 shows
the ESI-ISF MS (on the VG-Platform MS) of the TMPP+-Ac derivative of YLDSR
(residues 13 — 17). Note that there is a series of mostly "'an ions, corresponding to the
first four residues. Cleavage at the aspartic acid residue results in a *d3 ion rather than an
*a3 ion, along with a *b3 ion and a *c; ion. This result is expected from our earlier

observations.

81

100 82
1 a
1 a1
mi
5? .
...3 a3
5:
03
j: a
vo-I:
.1! E F: t
“’1 I <——> a4*a
0:; <—> 5

M

a <—>

3 TMPP-Ac I C
573 1179

 

 

 

 

 

 

 

1111.1 2.1 4 11.1.11. liLiLLllJinniqur
lrlil ITIW reriIIIII Illii'l'l [III

600 700 800 900 1000 1 1 00

mlz
Figure 2.25. ESI-CAD-MS/MS of a peptide derivative from cytochrome-c
tryptic digest with a precursor mass of 1179 Da at a relative collision energy
of 60%.

 

82

100* 0 .

 

 

 

 

 

 

 

a1 0,,
'5
c
3 'a2
5
s 62
I;
E
o
0: 'd3
cm

I ha

Jll ,I, I WI. VI. 1 Y II II

700 800 900 1000 1100 1200

mlz

Figure 2.26. ESI-CAD-MS of TMPP+-Ac-YLDSR, peptide derivative from the
tryptic digest of glucagon, showing the characteristic fi'agmentation pattern of
aspartic acid residue D.

83

c) Insulin Chain-B:

The oxidized B-chain of insulin (30 residues and 2 cleavage sites) was digested
and the resulting fragments derivatized with the Tl\/IPP+-Ac reagent. The ESI-ISF and
MALDI-PSD spectra of one of the derivatized fragments (residues 23-29) are shown in
Figure 2.27. Again, a series of *an ions is observed for the first five residues in the top
panel. Residue 6 is proline, and the signal due to flu is not discernible; this, again, is
consistent with our earlier observations. The ESI-ISF spectrum in Figure 2.27 compares
well with that of the MALDI-PSD spectrum of the same derivatized peptide, in that *an
ions are observed in each. The difference is that *ag, the ion due to cleavage at the
proline residue is observable in MALDI—PSD, as is a signal due to fragmentation of the
terminal lysine (*a7). Previous studies have shown that the signal due to cleavage at
proline in MALDI-PSD is laser power-dependent [7].

Although analysis of enzymatic digests of proteins by LC-MS/MS is becoming
routine, preliminary TMPP+-Ac derivatization would help detect the smaller peptide
fragments, resulting in better coverage. In addition, the relatively simple fragmentation
pattern of the TWP+-Ac-peptides, even under low-energy CAD, would help one to

deduce the amino acid sequence without the aid of extensive computation.

 

 

 

 

 

 

4..
2
100 d .
o. 32
2
'—
. I33
a,
la ‘84
5 CI
“
c Y le w i L] JYLV % l l _L 7L .1 L J. l ,_.,.,-.,.__.~_~.J.J-
— 600 800 1000 1200 1409
Q3 3 C I
.2 10000 37
fl
2 i
d) 3000 I,
m ; II
+
0. a2 'a _ III
I 3 +-
4000* 02. ' C K I;
I- I I I
2000 '31 II T5 '6a I
'I I:
“In; , WW WW I '
mlz

Figure 2.27. A comparison of ESI-ISF mass spectrum (top panel)
with MALDI-PSD spectrum (bottom panel) of TMPP+-Ac-GFFYTPK,
a peptide derivative fi'om the tryptic digest of insulin chain-B.

85

IV. Conclusions

This ESI-ISF/CAD-MS study of the Tl\/IPP+-Ac derivatives of peptides shows that
a series of *an ions is consistently produced. The fact that *all ions are observed suggests
that the fragmentation is likely via charge-remote pathways. Our results from ESI-CAD-
MS/MS are also similar to those observed in FAB-CAD-MS and MALDI-PSD-MS for
TMPP+-Ac-peptides. Comparison of the results of the ESI-ISF-MS and the MS/MS study
of the TMPP+-Ac-derivatives shows that precursor-selected CAD-MS/MS is similar to
ISF-MS. Fragmentation of these charged derivatives differs in an ion trap depending on
whether a singly- or doubly-charged precursor is selected; charge-mediated processes
also seem to be present in the latter case.

The results from analyses of derivatized protein digests indicate that the behavior
of individual TMPP+-Ac derivatives is maintained even in the case of complicated
mixtures such as those derivatized from protein digests. Sequence information is readily

obtainable based on the predominance of a simple series of *an ions.

86

10.

11.

12.

13.

14.

15.

16.

17.

18.

V. References

. Yates, J. R. J. Mass Spectrom. 1998, 33, l.

Carr, S.A.; Hemling, M.E.; Bean, M.F.; Roberts, G.D. Anal. Chem. 1991, 63, 2802.
Biemann, K. Biomed. Environ. Mass Spectrom. 1988, I6, 99.
Biemann, K. Methods in Enzymology, Ed. .LMcCloskey, 1990, 193, 886.

Roth, K.D.W.; Huang, Z.-H.; Sadagopan, N.; Watson, J .T. Mass Spectrom. Rev.
1998, 1 7, 255.

Huang, Z.-H.; Wu, J .; Roth, K. D. W.; Yang, Y.; Gage, D. A.; Watson, J. T. Anal.
Chem. 1997, 69, 137.

Liao, P.-C.; Huang, Z.-H.; Allison, J. J. Am. Soc. Mass Spectrom. 1997, 8, 501.

Lin, T.; Glish, G.L. Proceedings of the 45th ASMS Conference on Mass Spectrometry
and Allied Topics, Palm Springs, California, June 1-5, 1997.

Huang, Z.-H.; Shen, T.S.; Wu, J .; Gage, D. A.; Watson, J. T. Analytical Biochemistry,
1999, 268, 305.

Kidwell, D. A.; Ross, M. M.; Colton, R. J. J. Am. Chem. Soc. 1984a, 106, 2219.
Renner, D.; Spiteller, G. Angew. Chem. Int. Ed. Engl. 1985, 24, 408.
Vath, J. E.; Zollinger, M.; Biemann, K. Fresenius 2. Anal. Chem. 1988, 331, 248.

Johnson, R. S.; Martin, S. A.; Biemann, K. Int. J. Mass Spectrom. Ion Processes. 1988,
86, 137.

Wetzel, R.; Halualani, R.; Stults, J .T.; Quan, C. Bioconjugate Chem. 1990, 1, 114.
Vath, J. E.; Biemann, K. Int. J. Mass Spectrom. Ion Processes. 1990, 100, 287.

Wagner, D. S.; Salari, A.; Gage, D. A.; Leykam, J .; Fetter, J .; Hollingsworth, R.;
Watson, J. T. Biol. Mass Spectrom. 1991, 20, 419.

Burlet,O.; Orkiszewski, R.S.; Ballard, K.D.; Gaskell,S.J. Rapid Commun. Mass
Spectrom. 1992, 6, 658.

Stults, J. T.; Lai, J .; McCune, S.; Wetzel, R. Anal. Chem. 1993, 65 , 1703.

87

19. Jeffery, W.A.; Bartlett-Jones, M.; Pappin, J .C. Proceedings of the 43m ASMS
Conference on Mass Spectrometry and Allied Topics, Atlanta, Georgia, May 21 ~26,
1995.

20. Spengler, B.; Luetzenkirchen, F .; Metzger, S.; Chaurand, P.; Kaufinann, R.; Jeffery, W.;
Bartlet-Jones, M.; Pappin, D. J. C. Int. J. Mass Spectrom. Ion Processes. 1997, 169/1 70,
127.

21. Hines, W.; Peltier, J.; Hsieh, F.; Martin, S.A. Proceedings ofthe 43"I ASMS Conference
on Mass Spectrometry and Allied Topics, Atlanta, Georgia, May 21-26, 1995.

22. Naven, T. J. P.; Jeffery, W. A.; Bartlet-Jones, M.; Rahman, D.; Pappin, D. J. C.
Proceedings of the 45m ASMS Conference on Mass Spectrometry and Allied Topics,
Palm Springs, California, June 1-5, 1997.

23. Liao, P.-C.; Allison, J. J. Mass Spectrom. 1995, 30, 511.

24. Stults, J. T. Proceedings of the 40m ASMS Conference on Mass Spectrometry and
Allied Topics, Washington, DC, May 31-June 5, 1992.

25. Cardenas, M. S.; van der Heefi, 13.; de Jong, A. P. J. M. Rapid Commun. Mass
Spectrom. 1997, 11, 1271.

26. Griffiths, W.J.; Lindh, I.; Bergman, T,; Sjovall, J. Rapid Commun. Mass Spectrom.
1995, 9, 667.

27. Lindh, I.; Griffiths, W.J.; Bergman, T.; Sjovall, J. Int. J. Mass Spectrom. Ion
Processes, 1997, 164, 71.

28. Lindh, I.; Sjovall, J .; Bergman, T.; Griffiths, W.J. J. Mass Spectrom. 1998, 33, 988.
29. Downward, K.M.; Biemann, K. J.Am.Soc.Mass.Spec. 1994, 5, 966.

30. Dongre,A.R.; Jones,J.L.; Somoygi, A.; Wysocki, V.H. J.Am.Chem.Soc. 1996, 118,
8365.

31. Van Dongen, W.D.; van Wijk, J .I.T.; Green, B.N.; Heerrna, W.; Haverkamp,J. Rapid
Commun. Mass Spectrom. 1999, 13, 1712.

32. Hayes, R.N.; Gross, M.L. Methods in Enzymology, J. McCloskey, Ed.1990, 193, 886.

33. Watson, J .T.; Wagner,D.S.; Chang, Y.S.; Strahler, J .R.; Hanash, S.M.; Gage, D.A. Int.
J. Mass Spectrom. Ion Processes, 1991, 111, 191.

34. Tang, X.-J.; Thibault, P.; Boyd, R.K. Anal. Chem, 1993, 65, 2824.

88

35. Wagner, D. S. Ph.D. Dissertation, Michigan State University, East Lansing, MI
(1992)

36. Roth, K.D.W. Ph.D. Dissertation, Michigan State University, East Lansing, MI
(1999)

37. Papayannopoulos, I. A. Mass Spectrom. Rev. 1995, 14, 49.

89

Chapter 3

EFFECT OF CHARGE-DERIVATIZATION IN THE DETERMINATION OF
PHOSPHORYLATION SITES IN PEPTIDES BY ESI-CAD-MS/MS

I. Introduction

Phosphorylation is a post-translational modification of proteins important for signal
transduction and enzyme activity. Proteins are phosphorylated in the post-transcriptional
stage by enzymes known as kinases using adenosine-tri-phosphate (ATP). Of the twenty
essential amino acids, only the amino acids tyrosine, serine and threonine can be
phosphorylated; these contain hydroxyl groups in their side chain. The structures of these

three amino acid residues are given in Figure 3.1.

H H
H2 H30— H H2
-NH-CH-CO- -NH-CH-CO- -NH-CH-CO—
Serine Threonine Tyrosine

Figure 3.1. Structures of amino Acid residues serine (S) , threonine (T), and
tyrosine (Y) respectively.

90

Enzymes called phosphatases are usually involved in dephosphorylation of a
protein. Many proteins have multiple phosphorylation sites, which are phosphorylated by
different kinases. A cascade of processes eventually results in an active protein. For
example, receptor tyrosine kinases phosphorylate specific tyrosine residues on
intracellular signaling proteins (receptors). On the other hand, receptor tyrosine
phosphatases remove phosphate groups from tyrosine residues of specific intracellular
signaling proteins. Signaling through tyrosine kinase domains of such receptor proteins
is involved in a variety of diverse biological processes such as cell growth, transcription
and apoptosis. The cytosolic region of the receptors has a tyrosine kinase domain of
250-300 amino acid residues, which bind to different proteins when phosphorylated [1].
Thus, in order to understand the activity of proteins, in many cases it becomes vital to
locate the sites of phosphorylation.

The conventional technique used for amino acid sequencing, Edman sequencing,
uses harsh chemical conditions under which the phophoester bonds in serine and
threonine are unstable [2,3]. The phosphoester bond is usually hydrolyzed or the moiety
is eliminated as inorganic phosphate. However, phosphotyrosine can be converted to the
phenylthiohydantoin derivative without degradation, and can be analyzed by HPLC. But
very low recovery continues to be a problem. The other techniques for determining sites
of phosphorylation include, radiolabeling with the 32F isotope [4]. The cells are
incubated with inorganic radiolabeled phosphate until a steady state is achieved in cells,
at which point the radiolabel is incorporated into the proteins. Further enzymatic
digestions of the protein and HPLC fractionation followed by Cherenkov counting of the

radioactivity in the HPLC fractions help identify the site of phosphorylation. However,

91

the validity of the assmnption of achieving a true steady state has never been tested, and
this poses a problem in estimating the stoichiometry of phosphorylation at a particular
site.

Mass spectrometry has been used as a tool for determination of the site of
phosphorylation since fast atom bombardment (FAB) or liquid secondary ionization
(LSI) was adopted as a desorption/ionization (D/I) technique. Theoretically,
phosphorylation of a single site on a peptide increases the mass by 80 Da. This increase
in mass can be used as a method of determining if a particular peptide is phosphorylated.
Also, if a peptide is originally phosphorylated it can be dephosphorylated using alkaline
phosphatase and a consequent decrease in mass by 80 Da will indicate the presence of a
single site of phosphorylation. If a protein contains phosphate and if the sequence of the
protein is known a priori, the mass mapping approach is used. The protein is
proteolytically digested and the smaller pieces are analyzed. Thus, if there is a mass
increase of 80 Da for a peptide containing a phosphorylatable site, then it is concluded
that the peptide sequence derived from the given protein is phosphorylated. However, if
there is more than one possible site for phosphorylation, direct sequencing of the peptide
by collisionally activated dissociation (CAD) alone can exactly locate the site, provided
the phosphate remains bound and intact. During CAD, loss of H3PO4 from the precursor
results in a peak that is 98 Da lower than the precursor and loss of HPO3 results in a peak
80 Da lower than the precursor. Loss of 80 Da is common when the phosphotyrosine is
present, and loss of 98 Da is commom for phosphoserine and phosphothreonine.
Biemann and Scoble have reported the determination of phosphorylation at the N-

terminal serine residue of troponin T, a regulatory muscle protein by FAB-CAD-MS [5].

92

The presence of an ion 98 Da less than the precursor ion, in the positive ion mode, for
one of the peptides from the protein, due to the loss of H3PO4, helped determine
phosphorylation. The sequence of the peptide contained only one such phosphorylatable
residue (S) at the N-terminus. In the case of FAB, a peptide containing a phosphate
moiety, and thus charged, would have a reduced surface activity [6]. Surface activity of a
peptide is important for its desorption and ionization in FAB. However this is not a
serious problem. For example, proteins such as glycogen synthase have been studied by
continuous-flow LSIMS and sequenced by CAD-MS [7]. There are several other
examples where FAB-CAD-MS has been used for determination of phosphorylation sites
such as in riboflavin binding protein, spinach chloroplasts, and MAP kinase [8,9,10].
MALDI-MS has also been used in such determination of sites of phosphorylation
in proteins from biological sources using picomolar quantities by mass mapping
approaches [ll-l4]. Wang et al. used 32F labeling of Jurkat leukemic T-cells. The
phosphoprotein OP18 isolated by preparative 2D electrophoresis was eluted from the
nitrocellulose membrane. Its tryptic digest after HPLC fractionation was analyzed by
MALDI-MS. Although there were three isomass peptides present in the mixture analysis
that would correspond to the phosphopeptide, use of phosphatase decreased the mass of
one of the peptides (15-27) by 80 Da, proving that it was phosphorylated. Further
enzymatic digestions with chyrnotrypsin and endoproteinase Glu-C helped exactly
determine the site of phosphorylation, serine-25. Annan and Carr (1996) have shown
the use of MALDI-PSD-TOF in the determination of sequence as well as phosphorylation
information using a- casein and calmodulin as examples [15]. Qin and Chait have also

shown the use of MALDI combined with an ion trap as an analytical tool for

93

determination of posttranslational modifications including phosphorylation in model
proteins [16]. Following these, the MALDI-trap instrument has been used to obtain
phosphorylation information on biological samples such as myosin kinase and Sendai
virus protein [17,18]. MALDI-MS has been used in combination with ESI-MS in certain
cases as a complementary tool [19,20]. Burlingame and coworkers have presented a study
using IR-MALDI on (it-casein [21]. They find that the elimination of labile phosphate
groups is minimized in IR-MALDI as opposed to using a UV laser, and thus the signal is
improved. Recently, a matrix additive, diammonium citrate, has been shown to enhance
the detectability of phosphate containing peptides in the positive ionization mode in
MALDI-MS [22].

Electrospray ionization (ESI) mass spectrometry has served as an important tool
in the identification of phosphorylation sites in many proteins as well. Several studies
using this technique are reported because it can be interfaced with separation techniques.
Enzymatic-digestion/peptide mass mapping has been routinely used for this purpose
[23,24], and MS/MS with CAD has been employed to obtain unambiguous results [25-
30]. In order to improve the sensitivity of the technique, neutral loss monitoring with
orifice-potential stepping in a triple quadrupole was developed by Carr and coworkers in
1993 [31]. The loss of HPO3 or H3PO4 moieties from phosphopeptides, due to increasing
cone voltage, is monitored by scanning the first stage of the triple quadrupole mass
spectrometer. This helps selectively identify the phosphopeptides. This procedure has
been used by Hunter et al. (1994) in studying a-casein [32]. Similarly, orifice potential
stepping and detection of negative ions was simultaneously developed by Kassel and

coworkers for phosphopeptide monitoring [33]. Knapp and coworkers studied the effect

94

of various charge states on peptides during MS/MS in a triple quadrupole instrument
[34]. Carr and coworkers also developed a reaction monitoring method, in which the
peptide was collisionally activated in the second stage of a triple quadrupole mass
spectrometer and the m/z difference between the first and third stage was used for
monitoring the neutral loss originating from a phosphopeptide [35]. This enabled
obtaining information selectively on phosphopeptides. Lehmann and coworkers have
systematically studied the fragmentation mechanisms of various phosphopeptides and
their analogues [36]. Almost all of the studies mentioned above are based on the use of
triple quadrupole instruments. However, recently there is a growing interest in
investigating phosphopeptides by ion trap mass spectrometers equipped with an ESI
source [37]. Degnore and Qin have reported detailed study on the behavior of
phosphopeptides in an ion trap mass spectrometer and the effects of charge state of a
phosphopeptide on its fragmentation during CAD [38]. Although the loss of H3PO4 or
HPO3 has been used as a signature to determine the presence of phosphorylation in most
studies, they show that phosphorylation on S, T, or Y cannot generally be distinguished
based solely on the observed loss of H3PO4 in an ion trap mass spectrometer. In addition,
they have shown that this loss is charge-state dependent, and is not observed when higher
charge-states of the same peptide are subjected to MS/MS in most cases.

CAD-MS/MS of a peptide often yields a variety of fragment ions retaining either
the N- or C-terminus, which can make spectral interpretation tedious and sometimes
uncertain. Another complication relates to dephosphorylated fragment ions obtained
during CAD-MS/MS of a phosphorylated peptide [8]. Thus the ion current is distributed

over multiple channels of information and hence the sensitivity is diminished. In some

95

cases, peptides do not fragment completely during CAD; charge-derivatization of
peptides ofien produces better fragmentation yielding a series of ions and simpler spectra
during CAD-MS/MS. In the case of underivatized peptides and those lacking basic side
chains where charge fixation is not possible, fragmentation is explained by the “mobile-
proton model” [39]. Here, the amide bonds are protonated preferentially and localization
of charge at one particular site is minimal. A recent approach to minimizing charge-
localization involves addition of a sulfonic acid group to the N-terminus [40]; this is an
alternative to the fixed charge-derivatization approach. Conversion of the N-terminus to a
sulfonic acid derivative has been used in the determination of phosphorylation in a model
peptide by MALDI-PSD. However, in the case of charge-derivatized peptides, the
charge is fixed on either terminus and charge-remote fragmentation pathways enable the
production of a series of either only N-terminal or only C-terminal fragment ions, thereby
simplifying the spectrum. There are very few studies describing the impact of charge-
derivatization on the CAD-MS/MS of phosphopeptides [41].

The advantage of tris-[(2,4,6-trimethoxyphenyl)phosphonium]-acetyl (TMPP+-
Ac) charge-derivatization of peptides prior to analysis by FAB-CAD-MS and MALDI-
PSD has been demonstrated [42-44]. The previous chapter describes a detailed ESI-
CAD-MS/MS study of TMPP+-Ac-peptides and their behavior in various mass
spectrometers [45]. This chapter focuses on the utility of TMPP+-Ac derivatization, for
preparation of an N-terminal charge-derivative, in facilitating the determination of the

site(s) of phosphorylation in a peptide by CAD-MS/MS (using an ion trap).

96

“Experimental

Phosphopeptides are purchased from the University of Michigan Peptide
Synthesis Facility, Ann Arbor, MI. The following is a list of phosphopeptides used in this

study:

Table 3.1. List of Phosphopeptides Used in the Study:

 

 

Peptide Mass of Peptide Mass of TMPP+-Ac
Derivative
KRtIRR 908.5 1481.5
LKRAyLG-amide 899.5 1471 .7
LKRAsLG-amide 823 .5 1396.6
LKRAtLG-amide 837.4 1409.5
KRPsQRHGSKY-arnide 1423.5 1996.2
KIGEthGVVYK 1474.5 2047.3

 

TMPP+-Ac derivatization of the peptides was performed by adding a 5-fold molar
excess of the derivatizing reagent to the peptide (1 nmol) in the presence of a lO-fold
excess of the base (dimethylaminopyridine) at room temperature. The reaction was
allowed to proceed for 15 min. The reaction mixture was separated by HPLC, and the
fraction corresponding to the TMPP+-Ac-peptide (diluted to 10 pmol/uL) was infused

into the mass spectrometer at a rate of 3 uL/min.

97

In order to obtain the phosphopeptide containing multiple phosphorylation sites
from a biological sample, tryptic digestion was attempted on the milk protein B-casein.
The protein (100 pg) (purchased from Fluka) was digested with trypsin (5% w/w) in 0.05
M tris hydrochloride (pH 8.6) for 4 h at 35°C. The resultant digestion mixture was
analyzed by MALDI-MS. An aliquot of the digest mixture was reacted with a 10-fold
molar excess of the TMPP+-Ac reagent with the resultant number of moles of peptide
from the digest. The mixture was allowed to react for 30 min at room temperature. The
products were then separated by reverse-phase HPLC. An Immobilized metal affinity
column (IMAC), made from FeCl3 in sodium acetate, was used to enhance recovery of
the phosphopeptide from the digestion mixture prior to HPLC [46].

Mass spectrometry was performed in a Finnigan-LCQ instrument; CAD was
performed using MS/MS in the trap with helium as the collision gas. The spectra were
collected at 3 uscans/sec. Similarly, CAD-MS/MS of underivatized phosphopeptides
was also performed in the trap. The collision energy for the phosphopeptides was about
30% (of a total scale of 5V) while it was 60% for the derivatized phosphopeptides.

In-source fragmentation of the phosphopeptides was performed by elevating the
capillary voltage to 90 V from a regular value of 25 V. This causes an increase in the
kinetic energy of the ionic species travelling between the capillary and the counter-
electrode. Consecutive collisions with the gas and other species in that region result in

fragmentation of the ions.

98

 

III. RESULTS

A) Fragmentation of Phosphorylated Tyrosine Residue Containing Peptide
Derivatives by CAD-MS/MS:

Of the three phosphorylated residues, phosphorylated tyrosine (Y) is distinguished
from phosphorylated serine (S) and threonine (T) by a signature loss of just 80 Da as
opposed to the loss of 98 Da from the latter residues during CAD-MS/MS. The recent
article on the study of mechanisms of fragmentation of phosphopeptide, as studied in a
triple quadrupole mass spectrometer, suggests a reasonable mechanism explaining this

observation as shown in Figure 3.2 [36].

 

 

 

 

SERINE
‘12.” _. H °T¢

 

 

 

 

 

 

 

 

H2C p/
OH 0 \0H
THREONINE
H —>
wNH-p(3 NH-C OH
HO y/OH H0 III I OH
P/
ch/ \ \ | /
H30 0 \OH
TYROSINE 0 \/\
i 7%»
/OH
\OH

Figure 3.2. Loss of phosphoric acid via a six-membered transition state in peptides
containing phosphorylated serine, and threonine residues.

99

The presence of a bulkier aromatic group in tyrosine prevents the formation of a
six-centered transition state required for the loss of H3PO4 (98 Da). The mechanism as
proposed by Lehmann and coworkers is shown in Figure 3.2. The authors suggest that
the C-0 bond in tyrosine phosphate is stronger than the corresponding CO bond in serine
phosphate due to the direct neighborhood of the aromatic ring. This also results in a
weaker O-P bond in the phosphotyrosine. Formation of H3PO4 requires cleavage of the
C-0 bond whereas formation of HPO3 requires fission of the P-O bond. Thus tyrosine
phosphate preferentially eliminates HPO3. Although this is a reasonable mechanism, loss
of H3PO4 from tyrosine phosphate residue is possible since the hydrogen from the phenyl
ring could be involved via a six-membered transition state as shown in Figure 3.3a.
However, the AH for this reaction is probably higher than elimination of H3PO4 from an
alipahtic system. This can be assumed to be similar to a dehydration reaction. A simple
comparison between dehydration in an aliphatic and aromatic alcohol is shown in Figure
3.3b. The AH for the formation of benzyne is 80.2 kcal/mol, whereas it is only 10.8
kcal/mol for ethene [47]. Therefore, benzyne is formed less preferentially over ethene.
A similar extension can be made for the elimination of phosphoric acid. Thus, the loss of
HPO; rather than H3PO4 is preferred in the tyrosine phosphate containing peptides.

Similar observations have been made during studies in an ion trap as well [38,48].
The same results are als true for TMPP+-Ac-derivatized peptides. Figure 3.4 shows the
ESI-CAD-MS/MS spectrum of doubly-charged precursor of LKRAyG-amide (m/z 450)
(the phosphorylated residues are denoted as y, s and t while the non-phosphorylated

residues are denoted as Y,S and T).

100

a) oc\/\
WNH-C/H —)>

O \/\
\/\NH-C/H

|
II/OH ‘l +
\OH IIH
b) / /OH
\
CH3 AH=10.8kcal/mol H + H20 0) OH
HO—CH2

AH=80.2kcal/mol .I + H20
OH

Figure 3.3. a) Possible elimination of H3PO4 from tyrosine phosphate.
b) Dehydration of an aromatic alcohol vs. an aliphatic alcohol.

101

 

-Ixf_
\O
..o
‘9.
to
N
co
cut

5 -951692-————«4
HN S tow—4:
qzzu—é
92 21799—3

Snags—.2.
EHN-SK awe—j

x5

9 1789 —.-E
Adm I

m o ace :2
; OZH'z+HW 9' ”77

18W—
‘zocII—I- HIAl /

 

sq 2:669 ———-3;
m 9999 .——_—5;
Adv ewe é
iI

zq Lava/f;

 

'x 5%. I

:4 o 991—"

zz ion-1

1
’ *TT""" T'TT‘T‘IH'TUWT‘TTT“ 7“,”? fiFTTfi'T—T" ' ' ' I'T ' '

 

,8 MisuewIaAtteIea °

102

udvu 3 669 *5?

IIq 6 897 :3

300

700

IIGOOIII

mlz
Figure 3.4. ESI-CAD-MS/MS spectrum of the doubly-charged precursor of LKRAyLG-amide

500

400

200

at 25% ( relative collision energy full scale of 5 eV).

The precursor loses 80 Da, which corresponds to loss of HPO; (peak at m/z 413).
Fragmentation of the singly-charged precursor of the same peptide, show peaks
corresponding to loss of both 80 Da and 98 Da. Figure 3.5 shows the ESI-CAD-MS/MS
of the TMPP+-Ac derivatized version of the same peptide (CIr peak at m/z 1471). There is
a peak corresponding to the loss of P03 at m/z 1391. There is also a peak corresponding
to loss of 98 Da (m/z 1373), which can be assigned as loss of water along with loss of
HPO3. This is similar to what is observed for the single-charged precursor of the
underivatized peptide. These observations are consistent to those observed by Qin and
DeGnore, that the charge state of a phosphopeptide dictates its fragmentation pattern.

In charge-derivatized peptides more information is obtained from backbone.
However, elimination of the phosphate moiety seems to be preferred whether the peptide
is derivatized or not. Minimization of the loss of the phosphate group in the charged
derivative might be due to a salt-bridge type of interaction between the positively-
charged moiety and the phosphate group to some extent. A similar hypothesis is proposed
for larger peptides that get multiply protonated and undergo more backbone cleavages
than phosphate loss [38]. It is suggested in these cases that probably the charged amino
acid residue ( a basic residue) interacts with the phosphate group.

Figure 3.5, the CAD-MS/MS spectrum for TMPP+-Ac-LKRAyLG-amide, shows
peaks corresponding to N-terrninal ions corresponding to cleavage at residues 1-5. The
peak at m/z 1314 corresponds to loss of HPO; from the *b6 fragment. Thus the spectrum
gives evidence for the presence of a tyrosine residue, due to the occurence of -80 Da
peaks both from the precursor and the fragment ions. Also a series of peaks is evident

which aids in better spectral interpretation.

103

 

 

 

 

 

 

 

 

 

 

 

z I E
0 H20 9 99v1____ I1
171 ‘

. 0JH‘+3 9 1681I ”—7359:
9 ELEL / _.» '-
OZH' OdH'+I) _ o

IEOJH'gq‘V 7191/; 8
39931 '?
SB!- .4L 8
431‘
i 8
if F
”(1.17 wow :3: 8
¢ “La 01.6 i: 9
’f ‘HM-‘B, 9936 _::{§
3.?
z t:
‘L 9'719 8
“I w
I v- ' I._:

, [Bq“ 39%88——-—+ §
2 4 V 899 j
; aV'HIJWJ. 5
I \ ---§

9 as -----T_ --.--

g KJISUOJUI ”new 0

104

mlz
Figure 3.5. ESI-CAD-MS/MS spectrum of TMPP+-Ac-LKRAyLG-amide, singly-charged

precursor m/z 1471, at 50% relative collision energy.

The spectrum is shown for CAD-MS/MS at a relative collision energy of 50%.
Higher collision energies up to 70% result in the same pattern of fragmentation, but the

S/N is poor.

B) Fragmentation of Phosphorylated Serine Residue Containing Peptide
Derivative by CAD-MS/MS:

Derivatized peptides containing phosphorylated serine expel H3P04 from the intact
cation (C*), resulting in a dehydroalanine residue in place of the phosphorylated
residue, when analyzed by CAD-MS/MS. This can be seen in Figure 3.6, the CAD-
MS/MS mass spectrum for T MPP+—Ac-LKRAsLG-amide, as represented by the peak
labeled as C+-98 in the mass spectrum. In the case of phosphotyrosine containing
peptide, there is loss of HPOa, followed by loss of water from C +.

CAD-MS/MS of the derivatized phosphopeptides in the trap yield N-terminal
fragment ions such as 'b.. from which the sequence of the peptide could be deduced;
these included fragment ions providing information on the site of phosphorylation.
Predominantly IIbn ions are formed with occasional Ian ions, as seen in Figure 3.6. This is
also seen in Figure 3.5, the CAD-MS/MS spectrum of TMPP+-Ac-LKRAyLG-amide.
However, CAD-MS/MS of derivatives of peptides without phosphorylation yield
abundant Ian ions along with *bn ions. These observations imply that the presence of
phosphorylation in the peptide seems to favor the formation of Ibn ions. It may be that
the structure of the .bn ions with a phosphate group is more stable, or the conformation of
the derivatized peptides with phosphogroups in the gas-phase has a greater feasibility for
the formation of the 'bn ions. The mass difference between 'b4 and 'b5 is 167 Da, which

confirms the presence of a phosphorylated serine residue.

105

.3520 83:8 3:28 e\eom an 6.32 NE
69:86 8929395.”. 62890933656598 6.58% mgmzdgamm ea 2%:

 

 

 

 

 

 

NE.
83 com? com? o3 P coop com com com com
fiféq . . -thfttlkrIl...-..t[zitiv‘rtitIL-ILIIIFILLILIE-Y.;”.LIrILL-L... .IrrtEtLIIFc era-6.1.4... LL)... 4, p. ...I». t. . .4 ILL IL: wraip r......1I.ylt?rtIrJLtttrLJ.
: t m ,. __ .9 m. .1. f
U a. l .L a Z 8 9 m}
I. u L . m 9 a. h
n _ m a .9 6.. m w h. b. u.
no. .9 w r .7 Z r r.
L a 8 7 q C. 9 .H-
m .+ m Du a: cum 3
.9 N .e c. .
09 H 09%... 9 9 new m 01% 303. Him
8 o . N Ir?
+ W .7 We own \ m H
8 8 l m. R.
m . ....
9 ..w M.
w... -. n
I
o m m
T a
.w u
I S
- m.
T
g
u f
T
r
0 3+ .r.
z
m ..r.
9

ovx L

 

p

106

The peptide KRPsQRHGSKY-amidc, which also contains a phosphorylated
serine residue, could not be TMPP+-Ac derivatized. The reaction procedure was
attempted multiple times, and neither the HPLC separation nor the MALDleMS
analysis of the mixture showed the formation of the derivative. The presence of more
than one residue with bulky side chains in the peptide could be inhibiting the reaction of

the N-terminus with the TMPP-Ac reagent.

C) Fragmentation of Phosphorylated Threonine Residue Containing Peptide
Derivative by CAD-MSIMS:

Figure 3.7a shows the ESI-CAD-MS/MS of the peptide LKRAtLG—NHZ
(underivatized). Upon comparison with Figure 3.7b, the ESI-CAD-MS/MS spectrum of
TMPP+-Ac-LKRAtLG-NH2 (derivatized), it is apparent that the latter is a simpler
spectrum than the mass spectrum of the underivatized peptide (Figure 3.7a). Fragments
such as Idis, .36 — NH3, *b6, and Icc, also contain the phosphate group intact. The same
information could be obtained from Figure 3.7a, but it is relatively complicated. There
are several other peaks between those representing b4 and b5 in Figure 3.7a, some of
which correspond to C-terminal fragments and their dephosphorylated variants. If the
sequence of this model peptide were unknown, it appears that it would be easier to derive
the sequence information from the MS/MS spectrum of the derivatized peptide (Figure
3.7b) than from that of the underivatized peptide (Figure 3.7a).

Similar information regarding phosphorylation is obtained for the peptide KPtIRR
afier TMPP+-Ac derivatization. The difference between this peptide and LKRAtLG-
amide is the position of the phosphorylated residue. However, the results do not seem to

have a large impact due to the position of the phosphate moiety,

107

—-‘x10 —

 

x10

 

0

o I‘quIl

1

+12 Ea
099:: :5

800

2. 32

750

2a
vOdmIJIs.

700

0! 56¢

v n o v.30
omega»
0 12.68 M NMO u

A

I
e 9: .1: 33 i...
A

A

v n 3, . .
Om ...wo N.Mwwvovh
vs v63 1.

n IZ-Ca mdvvu
no 33.1

n Izsnnmdwmlll-

3 $3....

500 650

a» meow

v v—VTWW

Ifigbbfissd

3'50

vr'vvv

350 400

r

3100' I

no 2% ..n

 

259'

A '

5.1135 {aw—em . .

m/z

.25)

t 0.52.

 

 

 

xiG

 

b

 

r 3...!)

 

 

0.» -Iu...... 1 1.1.11}-
are; m. a e e s

Emu—.8:— 9523—

1200

two

a.
m/z

Figure 3.7a. ESI-CAD-MS/MS spectrum of LKRAtLG-amide, singly-

790

charged precursor m/z 837, 30% relative collision energy.

Figure 3.7b. ESI-CAD—MS/MS spectrum of TNIPP+-Ac-LKRAtLG— amide,

singly-charged precursor m/z 1409.5, 50% relative collision energy.

108

D) Fragmentation of Doubly Phosphorylated Peptide Derivative Containing
Threonine and Tyrosine Residues by CAD-MS/MS:

Figure 3.8 shows the CAD-MS/MS spectrum of TMPP+-Ac-KIGEthGVVYK. Note
that there are two phosphorylated sites in this case. The doubly-charged ion of the above
species was fragmented at 35% relative collision energy, which produced a nearly
complete series of ''b., fragment ions. The fragment ions 'b-I through *bm contain both
phosphate groups; some of these also appear as doubly-charged ions, as seen in the lower
m/z range of the spectrum. There are a few peaks that represent loss of the phosphate
moiety from these fragment ions. Some of the fragment ions lose 80 Da, which implies
that the loss of the phosphate group is from the tyrosine residue rather than from the
threonine residue. A few very weak C-terminal fragments (yn ions) and their
dephosphorylated variants are observed in this spectrum, possibly because the doubly-
charged precursor is fragmented, which has a mobile proton in addition to the fixed-
charge moiety. In this context, it should be mentioned that fragmentation of a doubly-
charged precursor of a TMPP+-Ac-peptide usually results in less abundant C-tenninal
fragment ions. This has been discussed in Chapter 2 of this dissertation. Thus, it is
preferable to use the singly-charged precursor to obtain CAD-MS/MS data. Use of the
singly-charged precursor does require higher collision energy, but it does not result in C-
terminal ions. In this particular case the mass (2047 Da) of the singly-charged precursor
is higher than the mass range of the mass spectrometer and hence the doubly-charged

precursor is chosen for CAD-MS/MS.

109

 

.3208 82323 on. wEEEE mEoEwwb on. 836:. £83.30: @825 2: .385 53:8 03.22 3mm
.48. 4E. .858... Emsfiéeé £>>>obumu§¢<fimze .8 6.58% m§m2-m<o-.mm w m 2a,...

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

hysuaml M93193

 

N): 000
ooom 002. com? 00: now? 2 .o.oo..”:..: 2.... 1 _. ..4... as... ...,?jfi....a.. o
.L . F»; J J J: «J. :JJJ . JJJ .532. J.J.4JJJJ..J..4 JJJZJJ...J$J.J (JJJ. J JJJ. J.. J. J. JJJzE... 4J4 .... €sz1 _ J.J JJ;
4 J J J. JJJ. JJJJ: JJJJJ J.
L. \ J J . I. ..
9 q. ..JJ J , . JJJ J m J J
o 8 Ir 7v 0 Ir _ .00.?
l C. 7. 9 m8 J Ir \ J;
m... JHW mm... .M x .. . 4 ... _
I! IV .0 J 8 I
a. W
Jo. R m J 6 .d I
8 .u .7 __ J u _+ m-
00 0 . _J z m V W:
- W .. . o
w . L M. n
H CH m J... W
C d I\ m .% Z W
.W 0 w . r. m
6 0..
a. .7 T
.7 . w
4,. m .4
+ o
m 6 w
m w. J1
+ 6
.m+ .4
. . fl
% . JJT.
[luv . cow
L _ mx _
— ONK _ opx

 

 

110

E) ln-Source-CAD-MS of Phosphopetide Derivatives in an Ion Trap Mass
Spectrometer:

In-source-CAD of the phosphopeptide derivatives were performed by elevating
the voltages of the capillary and the skimmer in the ion source of the instrument. Figure
3.9 shows the in-source-CAD-MS spectrum for TMPP+-Ac-LKRAyLG-amide. This and
in-source-CAD-MS spectra for other peptide derivatives show that the fragment ion
abundances are higher than those derived from MS/MS. The graph represented by Figure
3.10 compares the peak intensities due to the fragment ions derived from source-CAD-
MS. The intensity of the fragment ion peaks at high m/z is lower than of the fragment ion
peaks at low m/z, when produced by source-CAD-MS. Efficiency of fragmentation of
the precursor is less during in—source-CAD, since the signal of the intact precursor is still
significant (Figure 3.9). There is considerable dephosphorylation of the precursor, but
backbone cleavage is equally observed which produces the fragment ions. The fragment
ions include *cn ions that are predominant (Figure 3.10). Typically, the c. ion requires
higher energy for formation than the 1in“ or 'bn ions [49]. The energy imparted during
source-CAD is high enough to result in the formation of these ions.

The graph shown in Figure 3.11 compares the abundances of the fragment ions
derived from MS/MS. In MS/MS, at least one type of ion ('an or 'bn) is abundant at both
high and low m/z values. This is helpful in obtaining complete sequence information.
The other major feature in the case of MS/MS is that the intensity of the *c,. ions is very
low throughout the spectrum. These types of ions are not predominantly observed during

MS/MS.

111

 

+3 91m

ll
l

l

is»; [WML‘begurgJquA‘i K.» L; *L..- Ar. J“. L. l L
' I

com” user—.7..—
o‘n-‘oan-o [VLSL "

EM” 9 6€ZL

OHH- OJH-s' . 9'————69H

'9.
9 LVOL

c “N" . 9 vzor
r
"N" 9966

Ih_anJ‘N—l] WWW

‘.q 9 0L6:
‘HN~‘ 9.9. 996

CHN'C‘” 9'986 “‘4

Cp

, 9 1.98
3
' 9199

'.q 148—"-
9 9 981.—-——-

1'.

L-nl LA-..

Kahu-

 

 

.l
. JJLJJUVJILa ca .Lmduw

bud

ll

 

(+H++3) [991
‘3. 9.901.

:0,

 

l,

1
ant. «xx-«W. ...rdw- ‘w ~

 

l

‘3. 9999

 

9V'+<I¢IW.L "5

100}
0

MisuewI eAiteIeu

112

1300 1400

1200

mlz

Figure 3.9. ESI-source-CAD-MS spectrum of TMPP+-Ac-LKRAyLG—amide, at a

capillary voltage of 70 V and tube lens offset of 50V.

E 3_§-oi<§q-o<-

L

.m2-o<o-838
a): see Bane 89 “Spas co guessed. .03. flea

IIEEquE

 

113

.mgmzdfi E
oeaaoosafiéig ace 8&8 82 eases .8 855?. ._ 2 29mm

£E§u3£

 

114

These observations once again imply that phosphorylation of the peptide seems to
favor the formation of ion ions during in-source-CAD-MS. This may be attributed to the
stability of these ions or the conformation of the derivatized phosphopeptides in the gas-
phase that may aid in their formation preferentially. Additionally, the energetics involved
during in—source-CAD is higher than during MS/MS. The proposed structure of a *cn ion

is shown in Figure 3.12 [50].

R
(+)—CH2-co-NH-cH-co-NH2

Figure 3.12. Structure of 'cn ions formed from TMPP+-Ac derivatives of peptides.

F) Multiply Phosphorylated Peptides From B-casein:

Analysis of the digest from B-casein shows the presence of a phosphopeptide with
four adjacent phosphates, Figure 3.13. The sequence of the peptide with four
phosphorylated residues is RELEELNVPGEIVEsLsssEESITR. The original
experimental design was to isolate this peptide fragment and further enzymatically
process it with chymotrypsin (to cleave after L and F), to yield a peptide fragment
(sssEESITR) that can be used for sequencing experiments. However, due to recovery
problems of the precursor peptides from the digest mixture, this could not be done.
Attempts to derivatize the mixture with the TMPP+-Ac reagent failed. Although there
were signals due to TMPP+-Ac derivatives of other tryptic peptides of the protein, no

signal is observed for the derivatized version of this precursor peptide.

115

503809.“: m._._mm_m_wmm._mm>_w01>z.._mm._m_m cocoon 2:
mo mate.“ Bum—boimosaov >358 can bwfim @825on 05 E3038 ammm we“ Rom NE.

3 38¢ 23. .moumfiwonm So.“ 53, ovsaoqumoa—a 33:88.“ a 35388 mm: NE. we gave
2? Someod to 2:328 cosmowzo 03%.: 05 mo 85.58% 938 mOHAQé—Z .26 oSwE

ea

comm coon comm ooom
F . »

lliliiiii. it. —. viii.

 

6€6Z

LZOS

 

 

9Z1

.IEtmmmmamamfiwamzmoa>z._mm:mm_ .

8
O
N
v

116

Arginine (R) is the N-terminal residue of this peptide; the bulky side chain of arginine
probably offers steric hindrance to the reactive amino terminus which prevents it from
reacting with the incoming TMPP+-Ac group. Additionally, several attempts to try to
separate the phosphopeptide before derivatization failed. Although, an IMAC column
was used as a precolumn to help improve the recovery of the multiply phosphopeptide,
the results were not reproducible. Thus, it was not possible to pursue further experiments

using the multiply phosphorylated peptide from B-casein.

117

IV. Conclusions

TMPP+-Ac derivatives of phosphopeptides predominantly lose the phosphate
moiety during CAD-MS/MS along with other fragments due to weaker backbone
cleavage. The fragment ions formed in the examples shown are useful in predicting the
sequence as well as the phosphorylation information. Loss of HPO; is preferred from the
singly-charged precursor for tyrosine phosphorylated peptides, whereas loss of H3PO4 is
preferred from serine and threonine phosphorylated peptides. This is similar to what has
been reported by other researchers for underivatized peptides. However, there are no
distinguishing features in terms of the fragment ion formation when any one of the three
phosphorylated residues is present. Usually a series of b, ions is observed. The ’bn ions
seem to be preferentially formed over .21.. ions in the case of phosphorylated peptides.
During in—source-CAD-MS of the same peptide derivatives, the fragment ion abundances
(at the low m/z range) are higher and more backbone cleavage is observed. However,
fragment ions with intact phosphate moieties are not clearly observed. Thus one should
use CAD-MS/MS if the sequence is known a priori and the goal is to obtain
phosphorylation information. However, if the sequence is unknown, source-CAD-MS
can also be used in conjunction with CAD-MS/MS to obtain complementary information
on sequence, as the backbone cleavage is higher in the former method. Additionally, *cn
ions are formed in the case of the in-source-CAD-MS of phosphopeptide derivatives.
These ions are usually not observed as predominantly in non-phosphorylated peptide
derivatives during source-CAD-MS. This suggests that the TMPP+-Ac head group of
phosphorylated peptide derivatives allows for special conformation of these gas phase

precursors, enabling the formation of the 'cn type ions.

118

These results show promise that the TMPP+-Ac derivatives of phosphorylated
peptides, when subjected to CAD, provide direct evidence for the site of phosphorylation.
The type and abundance of fragment ions appear to depend on the sequence of the
peptide. Another encouraging feature of these preliminary data is the presence of a
nearly complete series of peaks corresponding to N-terminal fragment ions; such a
feature is important to confirm the sequence in cases of isomass peptides and in the

analysis of unknowns.

119

 

10.

ll.

12.

13.

14.

15.

16.

17.

18.

19.

V. References

LN. Johnson, M.E.M. Noble, D.J. Owen Cell, 85, 149 (1996).
D.B.Rylatt, P.Cohen FEBS Lett., 98, 71 (1979).

C.G.Proud, D.B. Rylatt, S.J.Yeaman, P.Cohen FEBS Lett, 80, 435 (1977).
J .C. Garrison Methods in Enzymology, 99, 20 (1991).

K.Biemann, H.A.Scob1e Science, 237, 992 (1987).

B.W. Gibson, P.Cohen Methods in Enzymology, 193, 488 (1991).

L.Poulter, S.-G. Ang, B.W.Gibson, D.H.Williams, C.F.B.Holmes, F.B.Caudwell,
J .Pitcher, P.Cohen Eur. J. Biochem., 175, 497 (1988).

C.Fenselau, D.Heller, M.S.Miller, H.B.White Anal. Biochem., 150, 74 (1994).
M. Hanspeter, D. F. Hunt, J. Shabanowitz, J.Bennet J.Biol. Chem, 263, 309 (1985).

A.K.Erickson, D.M. Payne, P.A.Martino, A.Rossomando, J. Shabanowitz, M.
Weber, D. F. Hunt, and T. W. Sturgill J.Biol. Chem, 265, 19728 (1990).

Y.K.Wang, P.-C.Liao, J. Allison, D.A. Gage, P.C.Andrews, D.M. Lubman, S.M.
Hanash, J.R. Strahler J.Biol. Chem, 268, 14269 (1993).

P.-C.Liao, J. Leykam, P.C.Andrews, D.A. Gage, J. Allison Anal. Biochem., 219, 9
(1994)

W.Zhang, A.J. Czemik, T.Yungwirth, R.Abersold, B.T. Chait Prat. Sci., 3, 677
(1994)

T.-T.Yip, T.W. Hutchens FEBS Lett, 308, 149 (1992).
RS. Annan, S.A.Carr Anal. Chem, 68, 3413 (1996).
J.Qin, B.T.Chait Anal. Chem, 69,4002 (1997).

J. Szczepanowska, X. Zhang, C. J. Herring, J. Qin, E. D. Korn, H. Brzeska
Proc. Natl. Acad. Sci. USA, 94, 8503 (1997).

KR. Jonscher, J .R. Yates J. Biol. Chem, 272, 1735 (1997).

R. S. Annan, and S. A. Carr .1. Prot. Chem, 16, 391 (1997).

120

 

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

X. Zhang, C.J. Herring, P.R. Romano, J. Szczepanowska, H. Brzeska, A.G.
Hinnebusch, J. Qin Anal.. Chem, 70, 2050, (1998).

R.Cramer, N.J.Richter, E.Stimson, A.L. Burlingame Anal. Chem, 70, 4939 (1998).
J .M.Asara, J .Allison J.Am.Soc. Mass. Spectrom., 10, 35 (1999).

A. Rossomando, J. Wu, M. Hanspeter, J. Shabanowitz, D. F. Hunt, M. Weber, T. W.
Sturgill Proc. Natl. Acad. Sci. USA, 89, 5779 (1992).

K. Palczewski, J. Buczylko, P. Van Hooser, S. A. Carr, M. J. Huddleston, J.
W. Crabb J. Biol. Chem, 267, 18991 (1992).

H. Taniguchi, M. Suzuki, S. Manenti, K. Titani J. Biol. Chem, 269, 22481 (1994).
M. Affolter, J. D. Watts, D. L. Krebs, R. Aebersold Anal. Biochem., 223, 74 (1994).
C. R. Lombardo, T. G. Consler, D. B. Kassel Biochem., 34, 16456 (1995).

K. A. Resing, R. S. Johnson, K. A. Walsh Biochem., 34, 9477 (1995).

R. D. Ladner, S. A. Carr, M. J. Huddleston, D. E. McNulty, S. J. Caradonna J. Biol.
Chem, 271, 7752 (1996).

A. Weijland, G. Neubauer, S. A. Courtneidge, M. Mann, R. K. Wieranga, G.
Superti-Furga Eur .J. Biochem., 240, 756 (1996).

M. J. Huddleston, R. S. Arman, M. F. Bean, S. A. CarrJ. Am. Soc. Mass Spectrom.,
4,710(l993)

A.P.Hunter, D.E. Games Rapid Commun. Mass Spectrom., 8, 559 (1994).

J. Ding, W. Burkhart, D. B. Kassel Rapid Commun. Mass Spectrom., 8, 94
(1994)

M. Busmann, K.L. Schey, J .E. Oatis Jr., D.R. Knapp J. Am. Chem. Soc., 7, 243
(1996).

S.A.Carr, M.J.Huddleston, R.S. Annan Anal. Biochem., 239, 180 (1996).
A. Tholey, J. Reed, and W. D. Lehmann J. Mass Spectrom., 34, 117 (1999).

Y.V.Lyubraska, S.A.Carr, D. Dunnington, W.P. Pritchett, S.M. Fischer, ER.
Applebaum, C.S.Jones, B.L. Karger Anal. Chem, 70, 4761 (1998).

J. P. DeGnore, J. Qin J. Am. Soc. Mass Spectrom., 9, 1175 (1998).

121

 

 

39

. A.R. Dongre, J .L. Jones, A.Somoygi, V.H.Wysocki J. Am. Chem. Soc., 118,
8365. (1996).

40. T.Keough, R.S.Youngquist, M.P.Lacey Proc. Natl. Acad. Sci, 96, 7131, (1999).

41

42

. Wagner, D.S. Ph.D. Dissertation, Michigan State University, East Lansing, MI,
(1992)

. Z. —H. Huang, J. Wu, K. D. W. Roth, Y. Yang, D. A. Gage, J. T. Watson

Anal. Chem, 69, 137 (1997).

43

44.

45.

46.

47.

48.

49.

50.

. P. —C. Liao, Z. —H. Huang, J. Allison J. Am. Soc. Mass Spectrom., 8, 501 (1997).

Z. -H. Huang, T. L. Shen, J. Wu, D. A. Gage, J. T. Watson Anal. Biochem., 268,
305-317 (1999).

N. Sadagopan, J. T. Watson J. Am. Soc. Mass. Spectrom., 1 1, (2000).
L.M. Nuwaysir, J .T. Stults J. Am. Soc. Mass Spectrom., 4, 662 (1993).

S.G.Lias, J .F .Bartmess, J.F.Liebman, J .L.Holrnes, R.D.Levin, W.G.Mallard J. Phys.
Chem. Ref Data, 17 (l), (1988).

N.Sadagopan, M.Malone, J .T. Watson J. Mass Spectrom., 34, 1279 (1999).
I. A. Papayannopoulos Mass Spectrom. Rev., 14, 49 (1995 ).

K.D.W. Roth, Z.-H. Huang, N. Sadagopan, J .T. Watson, Mass Spectrom. Rev., 17,
255(1998)

122

Chapter 4

MECHANISMS OF FRAGMENTATION OF PEPTIDES AFTER CHARGE-
DERIVATIZATION BY ESI-CAD-MS/MS AND MALDI-PSD-MS

I. Introduction

A) Collisionally Activated Dissociation of Protonated Peptides:

Structural information for a molecule after ionization and desorption into the gas
phase by electrospray is typically obtained by m/z analysis of fragments following
collisionally activated dissociation. In the case of peptides, the protonated molecule
fragments by charge-mediated pathways. There are several mechanisms proposed
explaining these pathways, and the “mobile-proton model” is widely accepted [1]. The
random localization of the proton has been supported by deuterium exchange studies
performed by Johnson et al. [2]. Proton transfers occur very fast, in the nanosecond time
scale, and since the lifetime of the ions in the mass spectrometer is considerably larger
than this, the ionizing proton gets randomly distributed throughout the peptide. In
solution, the pK. values of the N-terminus and the basic side-chains (lysine, arginine and
histidine) are significantly higher than those of the intervening amides, however the gas-
phase proton affinites are comparable [3]. During electrospray ionization, the singly- or
multiply-charged ions are ejected with the protons initially located on the same position
as in solution , the amino terminus or the basic side chain. However, in the gas phase this
proton is flee to transfer to the amides where it may induce cleavage [4]. Also, in the gas

phase, solvent molecules are absent and charge delocalization can occur by internal

123

 

solvation, a folding process that brings one or more of the amide linkages in close
proximity to the protonated amino terminus [5]. Since the proton affinity of a single
peptide amide linkage (217 kcal/mol [6]) is comparable to that of a three-carbon primary
amine in the gas phase (218 kcal/mol [7]), internal solvation provides a convenient
pathway for proton-transfer reactions that distribute the charge onto the various amide
linkages of the peptide backbone. Intramolecular hydrogen bond formation involving
other amide linkages in the peptide liberates at least 7 kcal/mol and provides the
additional driving force for preferential charge localization on the various amide bonds in
the gas phase [5]. In cases where there is no basic side chain, protonation occurs at the
amino terminus and it has been shown that fragmentation of peptides with the free amino
terminus requires more energy than those with an acetylated N-terminus [8].

When the collection of peptides protonated at various amide locations is subjected
to CAD, kinetic energy is converted to vibrational energy and many of the ions suffer
fragmentation at one of the amide bonds. During CAD, the proton transfers from the
more basic sites to less basic sites and initiates cleavage. The Ab initio and MNDO bond
order calculations have indicated that the amide N—protonated form can cleave readily
[9]. High energy pathways involve cleavage of single bonds whereas low-energy
pathways involve mechanisms of simultaneous bond breaking and bond making.

When proline is present in any of the protonated peptides, fragmentation is highly
favored at that site. This is called the “proline effect” [5]. This is because the secondary
amine functionality, the imino group of proline, is part of a five-membered ring and in an

amide linkage, it is devoid of hydrogen. The imino group has a higher gas phase basicity,

124

 

several kcal/mol higher than the other amide bonds in the peptide backbone, and is
preferentially protonated [10].

The nomenclature of ions formed from protonated peptide fragmentation was first
defined by Roepstmfl' and Fohlmann in 1984 [11]. CAD studies were originally
performed by FAB-MS. In 1986, Hunt et al. proposed mechanisms and structures for
various fragment ions formed from protonated peptides based on their FAB-low energy
CAD MS studies [5]. They suggested that the YI| + 2 ion receives one of the hydrogens
whose origin is from the FAB matrix, while the other hydrogen is the one shifting from
the a—carbon of the amino acid residue (B with respect to the protonated amide nitrogen).
Biemann proposed modified structures for various ions formed during these studies in
1987, and hence the YII + 2 ion is termed the y.ll ion [12]. In order to verify the structure
of the y" ion, Mueller et al. performed the first deuterium labeling studies in 1988, by
FAB-CAD-MS (low-energy CAD in a triple quadrupole) to pinpoint the origin of the
shifting protons resulting in the formation of the yII ions [13]. This is a hydrogen/
deuterium exchange study in which only the hydrogens on the amide nitrogen and the
terminal groups are labeled. The results prove that the exchangeable amide hydrogen
rather than the hydrogen from the or—carbon is transferred to the yn fragment. Later,
Kenny et al. (1992) performed labeling studies that involved labeling of the a—carbon
using FAB-high-energy CAD-MS [14]. Results from their studies show that it is
migration of the hydrogen from the amide nitrogen, rather than from the a—carbon, that
results in the formation of the y“ ion. Tang and Boyd in 1992 proposed a mechanism for
formation of the yn ion that involves the shift of an amide hydrogen, not necessarily to

the protonated amide hydrogen but to the amide nitrogen at the site of cleavage is

125

proposed by [15]. Formation of the b. ion, the N-terrninal counterpart of the yn ion,
resulting from cleavage of the amide bond is described as a charge-site induced on—

cleavage mechanism [5].

 

H+ +

Figure 4.1a. Mechanism of fragmentation of protonated peptides to form yn ions.

126

 

Figures 4.13 and 4.1b depict the previously proposed mechanisms for the

formation of the y.. and the bII ions, respectively.

Figure 4.1b. Mechanism of fragmentation of protonated peptides to form bn ions.

These mechanisms suggest that the localization of the proton is random and since
the loocation of the proton dictates the fragmentation, a complete series of fragment ions
necessary for sequence interpretation cannot be guaranteed. In order to overcome this
problem of not being able to obtain a complete series of ions, two approaches have been

developed. One approach is to fix the charge on one end of the peptide, so that the

127

fragmentation will not depend on an incoming proton, but rather will be due to intra-
molecular hydrogen shifts within the precursor ion. The result is often termed “charge-
remote fragmentation” because the cleavage of the peptide bond occurs remote from the
charge-site [16]. This type of fragmentation mechanism also seems to be operative in
tryptic peptides where the C-terminal residue is a lysine or arginine. Because of the
increased basicity of their side-chains, these residues are preferentially protonated in the
gas-phase.

The second approach intentionally randomizes the charge among the various
amide bonds thus resulting in a population of protonated species in which any one of the
amide bonds could be charged. Recently, Keough et al. reported the sulfonation of the N-
terminal end to aid in increased mobility of the proton [l7 ]. This follows from similar
approaches where the cysteine in the peptide, which also contains a basic side chain
containing a histidine residue, is oxidized to form cysteic acid. This results in the specific
interaction between the acidic and basic sites. Hence, fixation of the charge at a basic site
is minimized, which results in the formation of a precursor population with a

heterogeneous charge distribution [18].

B) Fragmentation of Charged Derivatives:

Fixed-charge derivatives were initially developed for FAB studies for purposes
of increasing the hydrophobicity of the analyte [19]. Studies of these charged derivatives
continue by MALDI-PSD and ESI-CAD-MS/MS for purposes of facilitating the
formation of a complete series of ions and simplifying spectral interpretation, because
fragment ions retaining just one terminus are formed. Recent reports of the [tris (2,4,6-

trimethoxyphenyl) phosphonium] acetyl (TMPP+-Ac) N-terminal derivatives of peptides

128

show that charge-remote fragmentation is indeed responsible for formation of a series of
'an ions in FAB-CAD [20], MALDI-PSD [21] and ESI-CAD [22]. The fragmentation
mechanisms of a related charge derivative the triphenylphosphonium (TPP) derivatives
have been studied by deuterium labeling [23]. Although typically it is deemed that only
F AB-CAD is suitable for studying such charged derivatives because higher energy (kV)
is deposited in the precursor ion, MALDI-PSD and ESI-CAD studies which are not hi gh-
energy techniques have also been successful in obtaining fragmentation of these charged
derivatives. In ESI-CAD, the charged derivatives require higher energy for
fragmentation than their underivatized counterparts, but still can be fragmented well
within 120 eV. This is fortunate from the peptide sequencing standpoint because
MALDI-PSD and ESI-CAD-MS/MS techniques are increasingly applied for the
sequencing of proteins/peptides. The energetic requirements in MALDI-PSD are still not
very clear, but our studies indicate that it is intermediate between the high-energy and
low-energy techniques.

During fragmentation of the charge-derivatized peptide there are three protons
(or, B or the amide (11) hydrogen as seen in Figure 4.2) that could undergo intramolecular

migration to yield the fragment ions.

 

1.
i
i
l

B
(R')2— -H
NH-CH-CO

i
i

 

 

x
n
l
r
u
i

 

Figure 4.2. Protons that could possibly shift during charge-remote fragmentation of
a peptide bond.

129

The FAB-CAD studies of TPP+-Ac derivatives of peptides suggest that the
[3 hydrogen migrates resulting in fi'agrnentation [23]. However, if this were the case,
fragmentation at the site of glycine should not be observed. This is true for the TPP+-Ac-
peptides or TMPP+-Ac-peptides when studied by FAB-CAD. However, when the
TMPP+-Ac-peptides are studied by MALDI-PSD or ESI-CAD fragmentation at glycine is
not observed [21,22]. Plausible structures of tan, *bn ions and related ions are given in
Figure 4.3 (R, R', R" correspond to the different side chains in the amino acid residue,

and ‘n’ is the number of residues in the peptide).

 

*an (formed from shift of B-Hydrogen) *bn (formed from loss of (Jr-Hydrogen) i
MH-iH-CON‘IJH- H WH-:H-CO¥\IIH_ =C=O
n-

 

 

*bn (formed from cleavage at aspartic acid)

1

*an (formed from shift of amide - Hydrogen)

l

 

 

i o
('D—(‘NH-iH-Comill Rn @‘l‘NH' ““903“?
*an+2 (formed from 1,2-elimination *bfiHZO
at the C-terminal end)
i l i l l“
GH—NH- H-Conx'IH-CHRn ©—+NH- H-comu- H-COOH
n-

 

 

 

 

 

 

Figure 4.3. Plausible structures of fragment ions formed from fixed-charged
derivatives.

130

 

There are signals observed due to fragmentation at the glycine residue (MALDI-
PSD and ESI-CAD-MS/MS) Hence, the previously proposed mechanism (B-hydrogen
migration) is seemingly not responsible for fragmentation in MALDI-PSD and ESI-CAD.
These observations imply that the mechanisms of fragmentation are different for high-
energy and low-energy CAD. Also important is the observation that there is not a
noticeable signal due to cleavage at the site of proline. This is not the case in F AB-CAD
studies where there is a signal due to cleavage at proline. This observation suggests that
in the case of MALDI-PSD and ESI-CAD the amide proton shifts and causes
fragmentation, because proline is a residue that lacks an amide hydrogen. In some cases
under MALDI-PSD fragmentation at proline was dependent on laser power [21]. Also in
the CAD-MS/MS studies of the TMPP+-Ac peptides in the ion trap, there is a very weak
signal due to cleavage at the site of proline. The previously proposed mechanism
involving an amide hydrogen shift for fragmentation is shown in Figure 4.4 [21]. The
validity of these previously proposed mechanisms was tested in deuterium-labeling

studies performed on a model peptide.

o—W

(+)\/\/\/CO-N VW

\n—fia +=

(+)\/\/\/CO-N + %V\/y
o

* an
Neutral Fragment

Figure 4.4. Mechanism of ''an ion formation by amide hydrogen shift in
TMPP+-Ac-peptides.

131

C) Current Studies:

The following is a report based on the ESI-CAD and MALDI—PSD studies of the
peptides with deuterium labeling at various sites ( a—, B—, and n—) and also with the
fixed charge due to the TMPP+-Ac moiety at the N-terminus. The peptide chosen for this
study is VGVAPG. This is a Chemotactic peptide with no acidic or basic side chains.
Hence fragmentation is simple. It contains proline, a residue without an amide hydrogen
when present in a peptide bond, which would provide us with additional insight regarding
the mechanisms of fragmentation. The following figure shows the structure of proline,
Figure 4.5. In order to label the amide hydrogens deuterium exchange studies need to be
performed, and the efficiency of deuterium incorporation decreases with increasing size
of the peptide. Hence this relatively smaller peptide, a hexamer (VGVAPG), which will

allow for complete exchange of the acidic hydrogens with deuterium is our choice.

 

HVH

H2 H2
\A/‘N- H-CO

 

 

 

 

Figure 4.5. Structure of the proline residue.

132

A) Peptides Synthesis:

The following are

ll. Experimental

the deuterium-labeled peptides used in the study:

Table 4.1. List of Labeled Petides:

 

 

Name Position of the Peptide
Deuterium Label

A No label VGVAPG

B or— Carbon ( Glycine) VGaVAPG

C a— Carbon (Valine) VQGVGAPG

D [1— Carbon ( Valine) VBGVBAPG

E B— Carbon ( Alanine) VGVABPG

F Amide Nitrogen "VnGnVnAPr‘G

 

 

The non-labeled species, a— and the B— labeled peptides were synthesized using

solid-phase peptide synthesis. The amino acids with deuterium labels were purchased

from CDN isotopes (Canada). The solid—phase peptide synthesis [24] involves three

steps: chain assembly, cleavage from resin and purification. These steps are briefly

described in the following text.

The peptide is synthesized from the C-terminal end to

the N-terminal end. The C-terminal amino acid of the desired peptide, in this case

glycine, linked to the resin (a polymer of 4-(oxymethyl)-phenylacetoamidomethyl) via a

benzyl ester, is used as the starting point.

133

The amino acids have their N-terminal

protected by the F MOC (F luorenylrnethoxycarbonyl) group before they are assembled on
the peptide chain, thereby minimizing interaction with the resin. The custom amino acids
with the deuterium label are FMOC protected ‘in house’ and the product characterized by
NMR. All the other amino acids were purchased with the FMOC protection. The first
cycle is to deprotect the amino acid, to remove the FMOC group, using a base (ususally
piperidine). The next cycle involves activation of the carboxy terminus of the incoming
amino acid (proline in our case), to ensure better reactivity of the carboxy terminus.
HBTU (2-(1H-benzotriazol-1-yl)-1,1,3,3tetramethyluroniumhexafluoro phosphate) is
used as the activating reagent. The activated amino acid is now coupled to the
deprotected amino terminus of the amino acid linked to the resin. Four equivalents of the
activated amino acid are added per one equivalent of the growing peptide chain. The
efficiency of coupling is over 99%. Coupling is the time consuming step of the automated
synthesis. About 2 hours are allowed for each coupling step. It is usually performed in
polar, aprotic solvent like dimethylforrnamide (DMF), to minimize peptide-peptide
hydrogen bonding. After each step the resin is usually washed with dichloromethane
(DCM) and N-methylpyrrolidine (NMP). A similar repetition of deprotection and
coupling results in the desired peptide, which is still attached to the resin. This is an
automated synthesis where the FMOC-protected amino acids are used in the synthesis of
the peptide using the Fastmoc cycles (HOBt/DCC) in the AP-Biosystems peptide
synthesizer. The peptide is cleaved fiom the resin using a TFA/water (95/5, v/v) mixture
and purified by preparative HPLC. The mass of the product peptide is checked by

MALDI-MS.

134

The peptide with deuterium labels at the amide nitrogen is prepared by
incubating in DzO/deuterated acetonitrile (1:1 v/v) for 30 min and allowing for free
exchange of hydrogen with deuterium. In the case of the derivatized peptide, the peptide
derivative is incubated using a DzO/deuterated acetonitrile mixture. In other cases, the
labeled synthetic peptide is first derivatized with the TMPP+-Ac reagent and HPLC

purified to obtain the TMPP+-Ac derivative of the labeled peptide.

B) Derivative Synthesis:

The TMPP+-Ac derivatives of peptides are prepared by reacting 1 nmol of the
peptide with 5-fold molar excess of the derivatizing reagent (TMPP+-Ac-SC(,F5;
synthesized as described in ref. 20) in the presence of dimethylaminopyridine at room
temperature for 15 min. The peptide derivative is separated from the byproducts and

reagents using HPLC.

C) MALDl-MS:

MALDI spectra are obtained with a Voyager Elite reflectron time-of-flight mass
spectrometer (Perseptive Biosystems) using an accelerating voltage of 22 kV. The mass
spectrometer is equipped with a 337-nm nitrogen laser. A saturated solution of a-cyano-4-
hydroxycinnamic acid, prepared in acetonitrile/water (1:1,v/v), is used as the matrix. A 1-
11L aliquot of the sample (picomoles/uL) was mixed with an equal volume of the matrix
solution on the sample plate, and the mixture dried in air. PSD data are obtained with the
mass spectrometer operated in the reflectron mode. Several PSD spectra, each optimized for
a different range of m/z values for the fragment ions, are obtained and "stitched" together to

yield the composite spectrum.

135

D) ESI-CAD-MS/MS:

A Finnigan LCQ, an ESI-ion trap mass spectrometer, is used to obtain CAD
information on the TMPP+-Ac derivatives of the peptides (acetonitrile/water/ 1%acetic
acid,10 pmol/uL) by infusion of the sample (3 uL/min). The source is held at 4.3 kV, and
helium is used as the CAD gas, for MS/MS. Fragmentation of singly-charged TMPP+-Ac
derivatives occurred at relative collision energies of 60-70% of a maximum of 5 V. For in-
source CAD, the source CAD option is turned on, and typically 50% is used for
fragmentation. The fragrnentaion occurs in the source in the octapole region. This is still part
of the source which is used to focus the ions into the mass spectrometer. The voltages at the
entrance and exit of the octapole are varied to increase the kinetic energy of the incoming

ions, which results in fragmentation.

136

III. Results

A) Analysis of the unlabeled TMPPI-Ac-VGVAPG:

The spectral results of CAD-MS of TMPP+-Ac-VGVAPG as analyzed by ESI (in-
source CAD), MALDI-PSD and FAB are shown in Figures 4.6a, 4.6b, and 4.6c
respectively. It can be seen that in the case of the FAB spectrum (adopted from Ref. 22) a
signal is present for cleavage at proline, whereas in MALDI and ESI there is no peak
representing ‘35. This suggests that different mechanisms are operative for the genesis of
these ions from these charged derivatives under high-energy and low-energy conditions.

The ESI—CAD-MS/MS spectrum of TMPP+-Ac-VGVPAG, in an ion trap, is shown
in Figure 4.7. This spectrum is different from the spectrum in Figure 4.4a, obtained from
source-CAD-MS. Here *bn ions are also seen along with the 'an ions. The most abundant
ions are the 'an and *bn ions, except for cleavage at proline where the signal is very weak
compared to the signal due to other ions. There is a signal due to *a6+2 (m/z 1028). The
other ions observed are those that are 18 Da higher in mass from the *b.. ions, which are
denoted as (‘b., + H20) ions.

In the following sections, results such as these, obtained from labeled derivatives
will be summarized in the tables for both 'an and 'bn ions, along with MALDI—PSD

results.

137

 

 

 

Ana V cam—Pa mo 02m 2: we awakens? col. E088 .32 em. mo moron a 36% 2:50on .coaoEoboeam
mane 283339326 m E Om<>0>é<£mm2h «o 838% $2: Q<erSomAmm .36 oSwE

 

 

 

 

 

a}:
83 com com com coo . l
4411 m 14 l i - I W. C
D
0sz 32 28
a a 08m 4
mass emu no use 3%
WON?“ ”on.
.L 3 S
n .+ emm
B
a s a m
.. w m...
B .d .A.
I ...m u
w a a
u
.6.
F”
+
2.:

 

 

138

Ammbosflme 05—05 05 mo 85 2: an owm>e2o com 8088 32 em. we grow
a 950% 500% BE. .0m<>0>-o<-+%=zh .8 85.58% mama Qmméaqaaz bod Semi

 
   

 

  
   

 

 

 

 

 

 

as
ooow com 8 con
1 old 88
0 9 m"
I I S
O .7 e
7e D. L
L 9 c. M W ooov
8.. . Po 8 I. .d
9+ 9 I 0 B .d
Z .0 z W 88
'8 CB 3 Mu
.. n
. ooom w.
/ W s
a V
H w . L 82:
. .I. [B _
D
M +

M 08$

 

139

.osflmoc eczofi 2: an omnibus 8 mvaommoboo we .3 coucemocqfi xmem .CN oococomoc
See 358$ 92>o>d<uaaze do sateen m§m2-o<o-m<a d3. edema

 

 

 

 

"\E
2:: 8n 2:. 8.. can a
—p-.-_.-.> -._“_..—...-_.--—-Ap<-._.~.._...._h-. °

1.4—1 4: c . <44 «4 aim—A . 1 1.. a

A _ u. :_ H
on 3 6 H
a. an L
we .lu
e. 3. m .
a .
..o 10lon<>e>l8£olh on: .. n
a: .
“a A

 

 

salmon

Azlsuezul

140

 

. / 9
9 nor 3 q. -
+ 9 399m

93.3 ’ uzm , ”if;
0 H- a. a 600i ——
Sq, 7'966 ,i

 

1000 1950

Figure 4.7. ESI-CAD-MS/MS spectrum of TMPP*-Ac-VGVAPG in an ion—trap mass spectrometer.

—1,

Jr
‘1

o‘H+”9. 91*6 *4:

 

 

 

 

600 650 700 750 800 850 900 950

 

 

"1,0669

'8, 311.9 :5;

Dims“. c9179 —:_

‘q, real; i

8

3. 9009 j

:

O‘H+‘q\ t

Z.LVL -—l'

q. real ..._...l:

‘8 j

‘ V’LOL _

'q. l'Zl9 .:

‘9. 9999 w —— — ;
3V'+<IJW.L Z'SLS—
‘Y—VT'*’*""—T‘Y—TY rl'llll'ljr‘rfijrwn V—r—‘q fi“— *‘o

AllsuelUI eAlleleu

100‘}

141

 

mlz

 

8) Results From Peptides with Deuterium Label at the a-Carbon:
a) Deuterium Label at the a-Carbon of Glycine:

Table 4.2 compares the m/z of the ions observed from the TMPP+-Ac-VGVAPG
(A) (the unlabeled species) with TMPP+-Ac-VG°‘VAPG (B) (the or—labeled species) by in-
source CAD, MS/MS in the trap and MALDI-PSD-MS. Figures 4.8a, 4.8b and 48C
show the spectra for the fragmentation of species B by source-CAD-MS/MS, trap-CAD—
MS/MS and MALDI-PSD-MS, respectively.

In species B, only one of the glycines ( only or—hydrogens) is labeled and hence the
precursor mass shifts by 2 Da (The C-terminal G is not labeled as the resin was
commercially available only with the unlabeled residue). Thus, a mass shift of 2 Da will be
observed for fragment ions 2-6, when compared to the fragments from the unlabeled species
A, if the label is retained during the fragmentation. A summary of peaks observed for *an
and *b.. ions are listed in Table 4.2 for each of three different conditions. In the MALDI-
PSD spectrum very weak peaks due to id“ ions for valine are observed . From the
standpoint of fragmentation, these are formed from 1'a“ ions. To focus more on the

aspects of the formation of the major ions, the "(1" ions are not listed in Table 4.2.

142

 

808080883 886 8800-5 05 co $8— 88888 08 .8 888088
05 8 mam—808000.80 «D N t3 8%.? 08 ea. .3. 82 80898..“ .8088o88% $3:
98 88 an 8 .m 860% .Om<>e0>-0<-+mm2._. mo 83.58% 808 Q<U-00.88-Hmm .awé 08wE
£8
omow coo? omm com 0mm com can con omm coo

 

 

 

 

a... L . is L . a ..a o
l .2:
m a
t w 9 cm
0 .. w m w a
u om. .8 8 .2 one
.. .. .. 2......
o. ..e a. A a
.. “em
a Lee...
W... m: n.
.. 7.2.2.0
B
e. 1 A
w flS
M M8
.. A
.. w 2:

B
€‘t7179 \

143

822880888 886 88200-20 05 so 3902 82208020 05 Ho 8280202
0E 02 8688888 «Q N .3 82.22% 08 ea. he. 82 808wfim .3880 828:8 032202 $3 20 00208828%
as: 822 82 ea 222 .m acne .om<>so>-a<-.82e 0e 688% 0282-30-28 .223. 282.2

NH:— omo F coop 0mm com omm com own 02. one cow

 

 

 

 

 

 

 

 

 

 

 

 

 

rjflilHJ... T... . 4-1. a . 2.1.1.2.... - .49. H . . ...l ._ . ..lj. Jialllalln O

_ __m d— _ 2 ‘ a ._ L ‘1 .r

_ .u av _.

._ 6 I. _ .6 .

m l. 9 m m . m m. 2

8 0 S 8 .7 + ..Iv [— .2

e .lrlrlOr e e m UH a. e Z ..
681 + + 70 an 2. a
9 9 emu % zH l A B
. o o . 2- u.

I. 8 Ir q H
II' B 9. o t T A
O 9 H CO a
i Q 7- D I
.8 . O a. W
8 9 v ; 9
3 8 .H w
+ m 9 _ m.

n A .

Z .8 m

£8: Wu 7v _.

.7 l r
7v
.9 . a 2
’3 u ..mu ...m
a w

144

 

.aouficofiwmb watsc 28928-5
05 no Bam— EstoSou 05 Ho coca—88 05 3 wEHEOQmotoo «Q N 3 vocEm 0.8 95. .. Na. 32
2358.2 .m 8.8% .02<>...o>-o<-.&25o 2.558% .38 3.2.5.2: 8.2. 2&3

EB 82 com 8w 8.2 o8

'1. -LI-

Eiijjgéaiéié.

 

L

0

18cm

 

92:) 9 9L
saunoo

9
1+ ”2 9'630L

7w. \ -83

 

 

..oooo

 

I) 8 SLOL
Z 7179
3V'+<I<IW.I.

+

 

 

 

145

Table 4.2. Ions Observed for d2-TMPP+-Ac-VG°VAPG (B):

 

 

-H 1|) -H
(+)—CH2CONH-C tic NH-CDéCO-NH—CH O-NH- Ergo-111% O—NH CH? CO-OH
*a1 b *az *a3 *36 .3
644 672 703 802 873 9a570 1029
Ion Type Unlabeled a-G-labeled a-G-labeled (B) a-G-labeled (B)
(A) (B) (Source-CAD- (MALDI-PSD-
(CAD- MS) MS)
MS/MS)
a1 644 644 644 644
bl 672 672 none none
32 701 703 703 703
b2 729 731 none none
a; 800 802 802 802
b3 828 830 none none
34 871 873 873 873
be 899 901 none none
.asm 968 970 none none
.bsw) 996 998 none none
‘a6 + 2 1027 1029 1029 1029
.bé 1053 1055 none none

 

#- Only observed during CAD-MS/MS in the ion trap mass spectrometer.

In the above table it can be seen that the expected mass shifts are observed
corresponding to the retention of the hydrogen on the a—carbon. These results indicate
that the mechanism of ‘an ion and 1b.. ion formation does not involve the shift of
hydrogen away from the a—carbon.

There is a significant difference in fragmentation of species B during in-source
CAD or MALDI-PSD from that observed during MS/MS in the trap. In MALDI-PSD,

only 'an ions are observed , as opposed to fragmentation in the trap (MS/MS) where b"

146

ions are also observed. This is in agreement with what has been reported in Chapter 2 ,
that fragmentation in the trap is different than fragmentation in the source.

In the case of source-CAD and MALDI-PSD there is no evident signal
representing the '35 ion, whereas a very weak signal (<10% relative intensity) is observed
during MS/MS in the trap. This suggests that there may be other mechanisms ( not
significant) that are operative along with the major pathways that lead to this type of ions.

In the MS/MS spectrum of TMPP+-Ac-VGVAPG in the trap, ions that are 18 Da
higher than the b, ions are observed, which is consistent with the structure for (”bn
+H20); as shown in Figure 4.3. The corresponding mass shifts are observed for those ions
formed from the deuterium-containing peptide derivatives. This suggests that the proton
on the oc—carbon is retained in these fi'agment ions. Such ions ("bII +H20) have also been
observed by Lin et al. during the study of TMPP+-Ac-peptides by FAB-ion trap-MS,
which they refer to as (b..-.+OH) ions [25]. This type of ion is only observed for

fragmentation of peptides with a free C-terminus.

b) Deuterium label at a-Carbon of Valine:

In species C, TMPP+-Ac- VaGVuAPG, both valines are labeled at the a—carbon
and hence the precursor mass shifts by 2 Da. Fragment ions 1 and 2 will have a mass shifi
of 1 Da, whereas ios 3-6 will have a mass shift of 2 Da if the deuterium is retained on the on-

carbon during fragmentation.

147

 

 

 

Table 4.3. Ions Observed for dz-TMPP‘iAe-vacvaAPG (C):

 

(+)—CH2CONH CDECb ogNu-C Héco ”NH cnéco-NH cuzco-JEC 0 NH CHzg CO-OH
*8] *a2 *36};
645 673 702 802 873 951570 1029

 

Ion Type Unlabeled a-V-labeled a-V-labeled

 

 

(A) (C) (C)
(CAD- (MALDI-
.. MS/MS) PSD-MS)
a1 644 645 645
‘b. 672 673 none
‘n2 701 702 702
.bz 729 730 none
'a. 800 802 802
‘b. 828 830 none
‘a. 871 873 873
b; 899 901 none
'215 968 970 none
‘bs 996 998 none
‘am 1027 1029 1029
‘b. 1053 1055 none

 

The mass shifts for all the ions are observed as expected. Also, as observed in
Table 4.1, during MALDI-PSD only iran ions are observed. These results support the
conclusions obtained from analyses of species B (Table 4.2), that the a—hydrogen does
not shift during the fragmentation and formation of the 'an and 1b., ions.

On the other hand, one need not consider all possible protons that can shift. For
example, one would not consider breaking two bonds to the same carbon as this would
leave a carbene as a resultant product. Such a consideration must be given to the
a—carbon where migration of the hydrogen in question and cleavage of the backbone at

that site would involve breaking two bonds attached to the same carbon.

148

 

C) Results From Peptides With Deuterium Labels at B-Carbons:

a) Deuterium Label at the B-Carbon of Valine:

In order to investigate the fate of hydrogen on a B-carbon atom in the TMPP+-Ac
derivatized peptide, a variant of the peptide with deuterium on the B-carbon of valine was
synthesized and studied by MALDI—MS and ESI-CAD-MS. Although our observations
suggest that the migration of this B-hydrogen is not the most plausible mechanism (since
signal due to cleavage at G which lacks a B-hydrogen is observed), it is yet to be proven.

Table 4.4. lons Observed for dz-TMPPtAe-VBGVBAPG (D):

 

 

Yb .1) HH -H
(+ )——CH2CONH CHécb (:an inéco NH CH2CO—NH CH§C0-*al:§C O NH CH23CO OH
*a1*a
645 673 702 802 873 98570 1029
Ion Type Unlabeled B-V-labeled B-V-labeled
(A) (D) (D)
(CAD (MALDI-
. MS/MS) PSD-MS)
al 644 645 645
‘b, 672 673 none
’az 701 702 702
‘bz 729 730 none
‘213 800 802 802
'b; 828 830 none
‘21.. 871 873 873
‘b. 899 901 none
‘a5 968 970 none
‘bs 996 998 none
‘a6 + 2 1027 1029 1029
’b6 1053 1055 none

 

149

The glycine residue, which lacks a B-hydrogen, does undergo fragmentation and a
signal is observed for cleavage at glycine during analysis by MALDI-PSD and ESI-C AD.
However only a very weak signal is Observed during FAB-CAD.

In this case, since the B-hydrogen is labeled in both the valines, the fragment ions
1-2 are expected to have a shift of 1 Da, while the fragment ions 3-6 are expected to have
a shift of 2 Da, if the deuterium label is retained on the B-carbon. As listed above in Table
4.4, the mass shifts indicate retention of the deuterium label on the B-carbon.

b) Deuterium label at the B-Carbon of Alanine:

Similarly, d3-analogue of VGAVPG containing three deuterons on the methyl
group of alanine was synthesized and its TMPP+-Ac derivative analyzed by MALDI-PSD
and ESI-CAD-MS/MS. The precursor is increased in mass by 3 Da, but the fragment ions
1-3, do not have shift whereas the fragment ions 4-6 shift in mass by 3 Da, provided the
label is retained. The results ( Table 4.5) show that the fragment ions do retain the
deuterium label on the B-carbon of the alanine residue. These results corroborate those
obtained from the study of the derivative where the B-carbon of valine is labeled. Thus,
the major fragmentation pathway does not involve a [3— hydrogen shift when the TMPP+-

Ac-peptides are analyzed by MALDI-PSD and ESI-CAD-MS/MS.

150

 

Table 4.5. Ions Observed for d3-TMPP+-Ac-VGVABPG (E):

Yaw/no-1?

(+)——CH2CONH CHECO OzNH-LI-léco NH CH§CO-NH Cl-léco-‘ralgc O NH-CH CO-OH
*a1*a *a6+
644 672 701 800 874 92671 1030

 

Ion Type Unlabeled B-A-labeled B-A-labeled

 

 

(A) (E) (E)
(CAD- (MALDI-
, MS/MS) PSD-MS)
al 644 644 644
‘b. 672 672 none
‘a2 701 701 701
‘bz 729 729 none
’a. 800 800 800
b; 828 828 none
‘a4 871 874 874
‘b. 899 902 none
ias 968 971 none
‘bs 996 999 none
‘am 1027 1030 1030
'b. 1053 1056 none

 

D) Results For Peptides with Deuterium Label on the Amide - Nitrogens:

The acidic hydrogens of a peptide (-CONI_I_, C003, and NI_I;_) are readily
exchangeable with those of the surrounding solvent molecules; similarly, these hydrogens
can be exchanged for deuterons in a deuterated solvent. For this experiment, the TMPP'i-
AC-VGVAPG is synthesized and dried under vacuum to remove any solvent. The residue
is reconstituted with a mixture of deuterated water and deuterated acetonitrile. The
mixture is allowed to equilibrate for 30 mins to ensure complete exchange of all six

acidic hydrogens ( five amides plus one carboxylic hydrogens).

 

151

The infusion/ sample delivery line in the ESI mass spectrometer is flushed with
the deuterated solvents to ensure that the exchanged deuterium on the peptide remains
intact. During MS/MS one can select ions of m/z 1077.4 +/- 0.2 for the deuterated
precursor. Figure 4.7 shows the full scan mass spectrum of the dG-TMPP+-Ac-VGVAPG
or species F. The overall deuterium incorporation at all six expected sites is relatively
high (relative intensities (%) of the other species d5,d4 are only 30% and 10%

respectively).

152

6.5.2 ES 2a 236; 65 3 32:8an

22 282365 comgnoxo 82533 3863800 .0m<>0>-o<-+m22h
Ho 88on Hoe—692 cowePE 2:82 65 20 89.58% 3228 58m :25 dd 823m

 

 

 

 

 

 

NHE
mwov owe. m2: Ohm: mmow 0.0.9.: ..mrmmvro
. a 2i... an... n. m... n...
m a e o .. .. .2... 2 9 9.2
8 9 .1. 8 (’22....0 0 .L Z .9
.9 .9 9 WL__...:..W.7 8L 9 Qmom
m _ Vim. Hon.
.6 0 W. m H
9 _ m2. 0 wow 9
.. .9 m: 2 WI
mw _ ...... ...oo 9
I. s 2 .ul.
9 : m. ..
V _ 0 Hon M
I. w mowm.
_ A23 :Om=<=>:0:>-o<-+mm2H. cor

9110i“

153

Table 4.6. Ions Observed for d6-TMPP*-Ae-V“G"V"A"PG" (F):

*31 b1 *8 *33 "W
645 673 703 803 875 1032

(+)—CH2COND-Cl‘§.C ND-Cl?CO—ND-CI%CO—ND§:§CO— algco-Nncng €0.01)
2
972

 

Ion Type Unlabeled Amide hydrogen labeled Amide hydrogen labeled

 

(A) (F) (F)
(CAD-MS/MS) (MALDI-PSD-MS)

‘a1 644 644 644
’131 672 672 none
".12 701 702 702
‘bz 729 730 none
“.13 800 802 802
"b3 828 830 none
‘a. 871 874 874
‘b. 899 902 none
~‘as 968 972 none
"b5 996 999 none
'21... 1027 1033 1033
"‘o6 1053 1057 none

 

The tabulated results differ from those indicated in the structure diagram at the
top of the Table 4.6 which shows the expected mass of the fiagments if the deuterium
label on the amide hydrogen is retained. But the actual results (Table 4.6) show fragment
ions one mass unit less than the expected value for ions *aI through '34, These results
suggest that in the formation of these ions, there is a shift of the amide hydrogen. If there
were no shift of the amide hydrogen, a peak at the m/z values as shown in the structure

should have been observed. Also, the results from analysis of the peptides labeled at the

01— and B— carbon complement this information indirectly.

154

a) Mechanism of Formation of the '11.. ion due to the Shift of an Amide Hydrogen:
As mentioned at the outset, there are only three local protons that could shift. The
proton located on the (it-carbon, B—carbon or the n-nitrogen. The results shown earlier in
this chapter suggest that there is no observed shifi of the former two protons. The results
from analyses of the species F, where the amide nitrogen is labeled with deuterium
suggest that it is the amide hydrogen that is involved in the formation of the '3" fragment
ions when the TMPP+-Ac analogs of a peptide are analyzed by MALDI-PSD or ESI-
CAD-MS/MS. These observations serve as experimental evidence for the plausible
mechanism proposed by Liao et al. for the fragmentation of the TMPP+-Ac-peptides by
MALDI-PSD [23]. The pathway involving the shift of the amide hydrogen was shown in
the Figure 4.4. Evidently, the same mechanism is also responsible for fragmenation by

ESI-CAD-MS/MS for most situations.

155

b) Mechanism for Formation of the (n+2) Ion Due to the Shift of C-terminal
Hydrogen Atom:

During the fragmentation of the TMPP+-Ac-VGVAPG, a peak appears
corresponding to the mass of an .116 + 2 ion instead of just an '35 ion. A mechanism that
migh be responsible for the origin of the .a6 + 2 ion has been proposed earlier [23]. The
results from our studies serve as evidence for such a mechanism as explained in the
follosing paragraph.

For the *as + 2 ion the expected mass is (1027 + 5) 1032 Da, but the peak appears
at m/z 1033. The structure of the 'a5 + 2 ion is given in Figure 4.2. These types of ions
are formed only in cases where there is a free carboxy terminus. The fact that there is a
larger than expected mass shift (1033 Da instead of 1032 Da) suggests that one of the two
additional protons on the ('a6 + 2) is not a hydrogen, but a deuterium which has its origin

from the carboxy terminus.

t
(+)—\/\/\/\/NH-CH———C =0

¢ H O
i
W ' =
(+ HEH + (If 0
O
*a6+2

Figure 4.10. Mechanism for formation of the terminal *a6 + 2 ion.

156

c) Mechanism for Formation of the .bn Ion Due to the Shift of an Amide Hydrogen:

The results from this study suggests that 1'b.. ions are formed from the shift of
amide hydrogens as well because the mass of the *bn ions is 28 Da higher than that of the
*an ion. The previously proposed structure for a b. ion is a ketone as shown in Figure
4.8 (The structure proposed for the 'bn ion due to cleavage at the aspartic acid residue is

different from this [18]).

*bn
l? R“
(+)-——(NH- ngO-NH- =C=O

(formed from loss of a-hydrogen)

Figure 4.11. Structure of a *bn ion.

If this structure is accurate, then the results from the current studies should show
that the deuterium label on the nitrogen is retained and the mass shift of the *bn ion from
an 'an ion would be 29 Da. Also, the results from the labeling on oc-carbon (Tables 4.1
and 4.2) suggest that the cat-hydrogen is retained during the formation of 'bn ion, which
according to Figure 4.8 is lost. Thus the structure of a ''b.. ion is possibly an aziridinone,
formed through a mechanism as shown in Figure 4.12, and not a ketone as pr0posed in

Figure 4.8.

157

Figure 4.12. Mechanism for formation of the .bn ion during ESI-CAD-MS/MS in an
ion trap mass spectrometer.

158

The mechanism in Figure 4.12 shows that the 'bn could be formed with the loss of
an amide hydrogen. An alternate structure, as in Figure 4.16, is reasonable only if there is
a rearrangement in the gas phase following fragmentation involving the loss of the amide
hydrogen, but later shifting the hydrogen from the a-carbon to the amide nitrogen.

However, this multiestep process seems less plausible from the mechanistic point of view.

(1) Mechanism for Formation of the ('b..+H20) Ion From the 'bn Ion :

The 'b..+H20 ions observed in all of the CAD-MS/MS spectra from the ion trap
correspond to a mass shift of 18 Da from their corresponding 'bn ions. However, these
are observed at a mass shift of 20 Da from the 'b.. ions in the case of the amide
deuterium labeled species (F). This can be explained as the addition of D20, instead of
H20 since there are no external water molecules available in this case. This may be
merely hydrolysis of the bond between the N and C0 of the 'bn ion as shown in Figure

4.13.

159

(+)—\/\/\/\/N_ H

¢ D20

R

(4.2/W222...)

D H
( 1urbn+D20)

Figure 4.13. Mechanism for formation of the 'b..+D20 ions from the deuterated
TMPP+-Ac-peptide during CAD-MS/MS in an ion trap mass spectrometer.

e) Mechanism for Loss of the C-terminal Residue:

Loss of the terminal glycine residue is observed in data obtained by both MALDI-
PSD and ESI-CAD-MS/MS. This has been typically observed in cases where the C—
terminus is a free carboxy functionality [26], from which the authors suggest that the -OH
group shifts to the neighboring amide carbonyl, resulting in the loss of the terminal
residue. If this mechanism were true, the resulting fragment from the amide deuterium

labeled species would be shifted by 5 Da, correspondingly.

160

| H
(+ c/
|/\H
H
H (Kg
(+ ‘ Cl/H
)__\/\/W
Rab/\H
H

H?
(+WVWC=O + ii +
C/H
0”; N/ \H
J.

Figure 4.14. Mechanism for formation of the C+-G fragment from
TMPP+-Ac-VGVAPG.

161

This is observed, and the fragment has a mass of 1019 Da from the deuterium exchanged
species (F), while the mass of the fragment arising from the unlabeled species is only

1014 Da. The mechanism adopted for the TMPP+-Ac-VGVAPG ( Species F) for the loss

of C-terminal glycine is shown in Figure 4.14.

E. Genesis of 'an and '13.. ions due to Cleavage at the Site of the Proline Residue:

The CAD-MS/MS and MALDI—PSD results from analysis of the TMPP+-Ac
derivatives of the labeled peptides suggest that the major fragmentation pathway in the
formation of 'an and ‘'b.. ions is due to the shift of the amide hydrogen over to the neutral
fragment of the peptide as shown in Figure 4.4. Yet, in the ESI-CAD-MS/MS spectra,
there is always a low intensity signal at the m/z values corresponding to the ’an and *b..
ions. The following table concisely presents the m/z values of the 'as and 'b5 peaks

observed for the various labeled and unlabeled species.

Table 4.6. List of mlz values for 'as and 'b5 Ions Observed From the Labeled
Variants of TMPP+-Ac-VGVAPG:

 

i i

 

 

Peptide a5 [)5
(+)-VGVAPG 968 996
(+)-VG°‘VAPG 970 998

 

(+)-V°‘GV°APG 970 998

 

(+)-VEGVBAPG 970 998

 

(+)-VGVABPG 971 999

 

(+)-"V"G“V“AP"G 972 999

 

162

The shift in mass for the two ions in the case of the species with deuterium labels
at 01— or [3— carbon suggests that these protons do not shift during the formation of the
'as and “b5. But, the data from the amide hydrogen—labeled species is different (compare
row 7 with row 2). For the 1'ag, ion, if there were no shift of amide hydrogens, then the
m/z value of the fragment ion (m/z 972) should be 4 Da higher than that for the fragment
ion formed from the unlabeled species (m/z 968). This expectation is in agreement with
the observed result, which suggests that the fragmentation does not involve any of the
neighboring residues, but the proline residue itself. This type of B-hydrogen shift is
primarily responsible for fragmentation in high-energy CAD analysis. Such a mechanism
is not the most predominant in low-energy CAD, and may not be favored energetically.
Hence, a very weak signal is observed, which is <10% of the relative intensity of the base
peak in the mass spectrum. However, the fact that there is a signal is intriguing and the
results from mass spectrometric analysis of the labeled peptides support that this signal is
real. The following scheme shows a plausible pathway for forming an 'as ion involving

the proton from the cyclic side chain of proline.

163

Figure 4.15. Mechanism for formation of the 'as ion due to cleavage at
proline residue.

The data from the above table also indicate that in the case of the formation of
the 'b5 ion, the shift in mass in the case of the amide labeled peptide and the unlabeled
peptide is not 4 Da (as observed for the 'as ion), but only 3 Da. This suggests that one of
the deuterons on the amide nitrogen fi'om the neighboring residue is possibly shifted
away resulting in the formation of the 'bg, ion. A reasonable mechanism depicting this
process is shown in Figure 4.16. This pathway involves a six-centered intermediate, and

forms a six-membered ring, which is a stable system. This is a more stable structure than

164

the structure of the 'as ion, which may explain why the signal due to the .b5 ion is

slightly larger than the signal due to .as( ca. 12% of the base peak).

("7‘\\\\\\\\
(N W412

i
l 62

Figure 4.16. Mechanism for formation of the 'b5 ion due to cleavage at
proline residue.

The above discussions indicate that the formation of the '.115 ion does not involve a

shift of the amide hydrogen whereas the formation of the 1'b5 does.

165

IV. Conclusions

The results from the mass spectrometric study of the TMPP+-Ac derivatives of the
deuterium-labeled forms of the peptide VGVAPG show that the major pathway during
the formation of the 'an and the 'bn ions involves the shift of the amide hydrogen of the
residue at which the cleavage occurs. This suggests that the structure of the *bn ion is not
a ketone as suggested previously, but is an aziridinone. Minor signal due to the cleavage
at the proline residue suggests that ions do form due to cleavage at this site during CAD-
MS/MS in a trap, although there is no amide hydrogen present. The shift in masses for
the fragment ion *as implies that the ion formation dues to cleavage at proline involves a
proton shift from the ring structure of the side chain in proline. However, the *b5 ion
forms due to the shift of a neighboring amide hydrogen with a resultant structure
including a six-membered ring. The data also serve as evidence for previously proposed

mechanism for the loss of C-terminal residue.

166

V. References

1) Dongre, A.R.; Jones, J.L.; Somoygi, A.; Wysocki, V.H. J. Am. Chem. Soc., 1996,
118, 8365.

2) Johnson, R.S.; Krylov, D.; Walsh, K.A J. Mass Spectrom., 1995, 30, 386.
3) Lias, S.G.; Liebman, J.F.; Levin, R.D. J. Phys. Chem. Ref Data, 1984,13,695.
4) Tang, X.-J.; Thibault, P.; Boyd, R.K. Anal. Chem, 1993, 65, 2824.

5) Hunt, D.F.; Yates, JR; Shabanowitz, J.; Winston, S.; Hauer, C.R. Proc. Natl. Acad.
USA, 1986, 83, 6233.

6) Meot-Ner, M. J. Am. Chem. Soc., 1984, 106, 278.

7) Aue, DH. and Bowers, M.T. Gas Phase [on Chemistry, Bowers, M.T. ed. (Academic,
New York), 1979, pp.l —51.

8) Nair, H.; Wysocki, V.H. Int. J. Mass Spectrom. Ion Processes, 1998, 1 74, 95.

9) Jones, J .L.; Dongre, A.R., Somogyi, V.H.; Wysocki, V.H. J. Am. Chem.Soc., 1994, 116,
8368.

10) Schwartz, B.L.; Bursey, M.M. Biol. Mass Spectrom., 1992, 21, 92.

1 l) Roepstorff,P.; Fohlman, J. J. Biomed. Mass Spectrom., 1984, 11, 601.

12) Biemann, K.; Martin, S.A. Mass Spectrom. Rev., 1987, 237, 992.

13) Mueller, D.R.; Eckersley, M. Richter, W.J. Org. Mass Spectrom., 1988, 23, 217.
14) Kenny, P.; Nomoto, K.; Orlando, R. Rapid Commun. Mass Spectrom., 1992, 6, 95.
15) Tang, X.; Boyd, R.K. Rapid Commun. Mass Spectrom., 1992, 6, 651.

16) Roth, K.D.W.; Huang, Z.-H.; Sadagopan, N.; Watson, J .T. Mass Spectrom. Rev., 1998,
1 7, 255.

17) Keough,T.; Youngquist, R.S.; and Lacey, M.P. Proc. Natl. Acad. Sci. USA, 1999, 96,
7131.

18) Burlet,O.; Orkiszewski, R.S.; Ballard, K.D.; Gaskell,S.J. Rapid Commun. Mass
Spectrom., 1992, 6, 658.

167

l9) Ligon, W.V. Anal. Chem, 1986, 58, 485.

20) Huang, Z.-H.; Wu, J.; Roth, K. D. W.; Yang, Y.; Gage, D. A.; Watson, J. T. Anal.
Chem. 1997, 69, 137.

21) Liao, P.-C.; Huang, Z.-H.; Allison, J. J. Am. Soc. Mass Spectrom. 1997, 8, 501.
22) Sadagopan, N.; Watson, J .T J. Am. Soc. Mass Spectrom. 2000, 11, (in press).
23) Wagner, D.S. Ph.D. Dissertation, Michigan State University, East Lansing, MI, 1992.

24) Barany, G.; Merrifield, RE. The Peptide Analysis, Synthesis, Biology (Gross, E. and
Meinhofer, J ed.), Academic Press, New York, 1980.

25) Lin, T.; Glish, G.L. Proceedings of the 45th ASMS Conference on Mass Spectrometry
and Allied Topics, Palm Springs, California, June 1-5, 1997.

26) Thome, G.C.; Ballard, K.D.; Gaskell, S.J. J. Am. Soc. Mass Spectrom. 1990, 1, 249.

168