‘F

IV
-.

x .._‘
élndfl‘ '3'

”III";

Iva" ‘

.L‘ifii'finr

.rufn“ :4

s; * - 1»
"9'16 ”my
rt “WI

7R7."

 

 

' w Lin-i-

-—=r.—..:x:.‘.,‘ m' sac—I” _' WW ".

 

.4. '- 4 :A

.,
JV

‘ ){of-I

. 'n‘uvé

I ’ ‘ -
.5 ‘ I ' '3' Hi
." ”I‘IZ’I "III :‘H';11: f? VIII“!

' x’\“'v'r.

‘ %w ‘a‘ i

I’ I
5334 ‘ I. ‘ ‘N V

. . ‘11:}. .I-szI. ?5,LJ‘[.:‘ ’1‘ ‘Ji‘: ‘7' _‘

. ' ' "I. i‘ ‘ H .a' . .41! gigs: Ii 4M, ‘ ‘
I.“ v. ‘. :;l ‘ ‘ “‘1; 3- f:10fifi§]’¥.:

5;.

 

:“(1

n r

I
L..-.:
I "‘ "I’
‘ :2" II 1541".”
II . I“

"EFT; v .1.

13.1?

.u

.I.‘ p:

THEms

9-000

 

IIIIIIIII"!IIIIIIIIIHIIIIIIIIIHIIIIII

I @9695

U .1' varsity

 

 

This is to certify that the

dissertation entitled

ENHANCING SPECTRA IN DESORPTION/IONIZATION
MASS SPECTROMETRY THROUGH THE USE OF
CHEMICAL ADDITIVES

presented by

John M. Asara

has been accepted towards fulfillment
of the requirements for

Ph.D. degree in Chemi Stry

 

 

/Q;A 4/4,;

Major professor

[Mme December 16, 1999

 

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

PLACE IN REI'URN BOX to remove this checkout from your record.
TO AVOID FINE-3 return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

11100 c/CIRCJDuoOmpGS-p.“

ENHANCING SPECTRA IN DESORPTION/IONIZATION MASS SPECTROMETRY
THROUGH THE USE OF CHEMICAL ADDITIVES

By

John M. Asara

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

ENHANCING SPECTRA IN DESORPTION/IONIZATION MASS SPECTROMETRY
THROUGH THE USE OF CHEMICAL ADDITIVES

By

John M. Asara

Multiply-charged molecules have been difficult to detect in desorption/ionization
(D/I) mass spectrometry. Techniques such as fast-atom bombardment mass spectrometry
(FAB MS) and matrix-assisted laser desorption/ionization mass spectrometry (MALDI
MS) produce mainly singly-charged ions in the mass spectrum which causes problems for
analyte molecules that possess a charge greater than one in the condensed phase. D/I
techniques do not deposit sufficient energy into molecules to produce higher charge
states. The rare occasions where multiply-charged ions are observed are usually with
very large molecules such as proteins; in these cases, signals for doubly charged ions are
usually much less intense than those for singly charged ions. Problematic compounds
include transition-metal containing complexes, phosphorylated peptides, and
oligonucleotides. Highly charged transition-metal complexes often do not exhibit mass
spectra, while phosphopeptides and oligonucleotides such as DNA usually result in peaks
of low intensity and poor resolution due to cation adduct attachment. In order to
overcome these limitations, the chemistry of matrix/analyte interactions has been studied,
and matrix additives have been developed in order to reduce the charge of the analyte
molecules. For transition metal-containing salts in the FAB experiment, the addition of
trifluoromethanesulfonic acid (triflic acid) was developed. Triflic acid donates protons in

solution, and the triflate ions bind tightly to the highly-charged complex, thus lowering

its charge and allowing useful spectra to be obtained. In the MALDI experiment,
ammonium salts such as diammonium citrate and ammonium acetate have been used to
reduce the charge of phosphate groups on phosphorylated peptides and displace alkali
cations. In some cases, the phosphorylated peptides form the largest peaks in the MALDI
spectrum, which allows for improved detection in an unfractionated protein digest. The
use of the tetraamine spermine has been developed to improve the spectra of
oligonucleotides in MALDI as well. It has been found that spermine works better than
ammonium salts in displacing alkali cations and substantially increases signal intensity of

DNA.

To my loving and supportive mother, Ann

ACKNOWLEDGMENTS

I would to thank the M.S.U. Department of Chemistry for a very challenging
and rewarding graduate program. I owe much of my gratitude to my advisor John
Allison who always offered his assistance without reservation. His remarkable ability
to explain complicated material in a simple and concise manner is very contagious
and has helped me to become a better presenter. Other faculty members who I would
like to thank are Kim Dunbar whose inorganic compounds were truly a challenge and
Doug Gage who always had an interesting protein project for me to work on. My
family has always supported me and showed pride in all of my endeavors, especially
my mother Ann. Without the values that she instilled upon me at a young age, I
never would have made it this far. My girlfriend Deborah has been with me since my
undergraduate days at Brandeis and I want to thank her for her loving support and
encouragement. I truly enjoyed being a member of the Art’s softball team for several
years and I want to thank the team for some great runs that we made at those
championships, winning two in my final season. I must say thanks to some friends
that I met at M.S.U. such as Al Schwartz, Lamont Terrell, Erik Ruggles, Lee
Kelepouris, and Michael Schall for some great times. I would also like to thank John
Strahler and all of the old and current Allison students that I worked with at Michigan

State.

TABLE OF CONTENTS

LIST OF TABLES ............................................................................ viii
LIST OF FIGURES ............................................................................. ix
LIST OF ABBREVIATIONS .................................................................. xi
CHAPTER 1

Introduction

I. Desorption/Ionization Mass Spectrometry In The Last Two Decades ............... 1
11. Use Of Chemical Additives In Desorption/Ionization Mass Spectrometry ......... 6
CHAPTER 2

Enhanced Detection Of Multiply Charged Transition-Metal

Complexes In FAB MS Using Organic Acids ................................................ 9
CHAPTER 3

Enhanced Detection Of Phosphorylated Peptides And Proteins In

MALDI MS Using Ammonium Salts

I. Phosphopeptide Enhancement Study With B-Casein .................................. 16
11. Determination Of The Sites Of Phosphorylation Of The Protein Bril Kinase. ....23

CHAPTER 4
Enhanced Detection Of Oligonucleotides In MALDI MS Using
The Tetraamine Spermine ...................................................................... 29

CHAPTER 5
Evidence Of Rhodium Bis-Acetate Units Covalently Bound To
1,2 Intrastrand GG And AA Sequences Of DNA .......................................... 35

LIST OF REFERENCES ....................................................................... 58

APPENDIX A

“Analysis of Transition-Metal Compounds Containing Tetrathiafulvalene Phosphine
Ligands by Fast Atom Bombardment Mass Spectrometry: Limitations and the
Development of Matrix Additives for the Desorption of Multiply Charged
Complexes”, Asara, J .M.; Uzelmeier, C.E.; Dunbar, K.R.; Allison, J. Inorg. Chem.
1998, 37, 1833-1840 ........................................................................... 62

APPENDIX B

“Enhanced Detection of Phosphopeptides in Matrix-Assisted Laser
Desorption/Ionization Mass Spectrometry Using Ammonium Salts”, Asara, J .M.;
Allison, J. J. Am. Soc. Mass Spectrom. 1999, 10, 35-44 .................................. 71

vi

APPENDIX C

“Enhanced Detection of Oligonucleotides in UV MALDI MS Using the Tetraamine
Spermine as 3 Matrix Additive”, Asara, J.M.; Allison, J. Anal. Chem. 1999, 71, 2866-
2870 ............................................................................................... 82

vii

LIST OF TABLES

Table 3.1 Lists the phosphorylation sites of the Bril kinase protein found in this
investigation. Note that the last five peptides show ambiguous sites for possible
phosphorylation.

viii

LIST OF FIGURES

Figure 1.1 Description of the fast atom bombardment (FAB MS) experiment.

Figure 1.2 Description of the matrix-assisted laser desorption/ionization
(MALDI MS) experiment.

Figure 1.3 The use of an additive or a third component in a desorption/ionization
(D/I) experiment.

Figure 2.1 Structure of [Ni4(bptz)4(CH3CN)3](BF4)3.

Figure 2.2 (a) FAB mass spectrum of [N i4(bptz)4](BF4)g in a m-nitrobenzyl alcohol
matrix. (b) FAB mass spectrum of [Ni4(bptz)4](BF4)3 in a m-nitrobenzyl alcohol
matrix with the addition of l pL of triflic acid.

Figure 2.3 (a) Expanded portion of the peak at m/z 2223 representing
[Ni4(bptz)4](OTf)7+. (b) Theoretical isotope distribution of the peak at m/z 2223.

Figure 2.4 FAB mass spectrum of [Ni4(bptz)4](BF4)3 in a m-nitrobenzyl alcohol
matrix with the addition of I uL of 48% tetrafluoroboric acid.

Figure 3.1 MALDI mass spectrum of the B—casein digestion mixture before removal
of the peptide at m/z 831.

Figure 3.2 (a) MALDI mass spectrum of the digestion mixture of B-casein using
4-HCCA. (b) MALDI mass spectrum of the same mixture with the addition of 12.5
mM diammonium citrate. (c) 25 mM diammonium citrate. The asterisks represent
nonphosphorylated peptides.

Figure 3.3 (a) MALDI mass spectrum of a portion of one fraction of a tryptic
digestion mixture of Bril kinase protein. (b) PSD spectrum of m/z 1607.6, a tryptic
digestion product containing two phosphate groups. The asterisks represent
incomplete digestion products.

Figure 4.1 The interaction of spermine with DNA in a living cell as well as in the
MALDI MS experiment.

Figure 4.2 (a) MALDI mass spectrum of a single oligonucleotide in the presence of

a high salt concentration buffer in 3-hydroxypicolinic acid with diammonium citrate.
(b) The same sample in 2-aza-2-thiothymine with 15 mM spermine.

ix

Figure 4. 3 (a) MALDI mass spectrum of a mixture six oligonucleotides in the
presence of a typical PCR buffer in 3-hydroxypicolinic acid with diammonium
citrate. (b) The same mixture in 2-aza-2-thiothymine with 15 mM spermine.

Figure 5.1 Dinuclear antitumor-active metal complexes with two possible binding
sites.

Figure 5.2 (a) MALDI mass spectrum of cisplatin bound to single-stranded DNA.
(b) Digestion of the cisplatin-DNA complex using 3' snake venom
phosphodiesterase.

Figure 5.3 A portion of the (a) MALDI mass spectrum of
Rh2(OzCCH3)2(TCTCTAATCTC) using 3-hydroxypicolinic acid with diammonium
citrate and (b) the MALDI mass spectrum of the same compound using 80%
anthranilic acid/20% nicotinic acid with spermine.

Figure 5.4 A portion of the MALDI mass spectrum of

Rh2(OzCCH3)2(CCTCTGGTCTCC) using 80% anthranilic acid/20% nicotinic acid
with spermine.

Figure 5.5 The MALDI mass spectrum of the digestion products of
Rh2(02CCH3)2(CCTCTGGTCTCC) using 3' snake venom phosphodiesterase.

Figure 5.6 The MALDI mass spectrum of (a) Rh2(OzCCH3)2(TCTCTAATCTC),
(b) the digestion products of the same dirhodium-DNA complex using 3' snake
venom phosphodiesterase, and (c) the digestion of the dirhodium-DNA complex
using the double enzyme digestion of 3' snake venom phosphodiesterase and 5' calf
spleen phosphodiesterase (brackets identify the cluster of peaks indicative of
dirhodium acetate bound to DNA).

LIST OF ABBREVIATIONS

A/D ................................................. analog-to-digital

ATP ................................................ adenosine triphosphate

ATT ................................................ 2-aza-2-thiothymine

[A+H]+ ............................................. protonated analyte

bptz ................................................. bis-pyridyltetrazine

Bril ................................................. brassinosteroid 1

BRs ................................................. brassinosteroids

Da ................................................... Dalton, mass unit

DAHC ............................................. diammonium hydrogen citrate
DDP ................................................ diamminedichloroplatinum(II)
D/I .................................................. desorption/ionization

DNA ................................................ deoxyribonucleic acid

ESI ................................................. electrospray ionization

FAB ................................................ fast atom bombardment
[G+H]+ ............................................. protonated glycerol

HCCA ............................................. a-cyano-4-hydroxycinnamic acid
HCit ................................................ hydrogen citrate

HMG ............................................... high-mobility group

HPA ................................................ 3-hydroxypicolinic acid
HPLC .............................................. high performance liquid chromatography
[M-H]' .............................................. deprotonated analyte molecule

xi

[M+H]+ ............................................ protonated analyte molecule

MALDI ........................................... matrix-assisted laser desorption/ ionization
m/z ................................................. mass-to-charge ratio

NGP ................................................ n-octyl-B-D-gluco-pyranoside

NMR ............................................... nuclear magnetic resonance

OAc ................................................ acetate

p ..................................................... phosphate group

o-P2,P4 ............................................. tetrathiafulvalene phosphine ligands
PCR ................................................. polymerase chain reaction

PSD ................................................. post-source decay

PVDF ................................................ polyvinylidene fluoride

SPD ................................................. 5'-calf spleen phosphodiesterase
TBP .................................................. transcription binding protein

TF A .................................................. trifluoroacetic acid

THAP ................................................ 2,4,6-trihydroxyacetophenone

TOF .................................................. time-of-flight

Tm .................................................... melting temperature

TRIS ................................................. tn's[hydroxymethyl]-aminomethane
triflic acid, HOTf ................................... trifluoromethanesulfonic acid

UV .................................................... ultraviolet

VPD .................................................. 3'-snake venom phosphodiesterase
80/20 ................................................ 80% anthranilic acid/20% nicotinic acid

xii

CHAPTER 1 - Introduction

I. Desorption/Ionization Mass Spectrometry In The Last Two Decades

Desorption/ionization techniques in mass spectrometry such as Fast Atom
Bombardment (FAB) and Matrix-Assisted Laser Desorption/Ionization (MALDI) have
been invaluable tools in the analysis of large, non-volatile molecules. FAB was
developed in 1981 by Barber and Bordoli1 and had great success for the structural
determination of organic, organometallic, and biological polymers such as peptides.
Before, FAB’s success, mass spectrometry was limited to the analysis of gas phase
molecules. Thus, only volatile organic molecules or molecules capable of being
thermally vaporized could be analyzed effectively. FAB was a success for a number of
reasons. First, little fragmentation was observed, so molecular weight information was
easy to determine. Second, it has great resolving power since it is used primarily with
magnetic sector mass analyzers capable of resolution up to 100,000. Scientists soon
realized that the disadvantages of FAB were the large amount of sample that must be
used and its mass range limitation. Several nanomoles of material are typically needed
for adequate signal-to-noise (S/N) ratios, and frequently the isolation of proteins from
biological systems only results in femtomoles to picomoles of material. Also, ions are
rarely observed above 4000 Daltons and most biological molecules including proteins
and DNA weigh far more than that. Today, FAB is mostly used for routine service

analyses in mass spectrometry facilities for small synthetic organic and inorganic

molecules but rarely for biological molecules. By the late 1980’s, it became clear that
new desorption/ionization (D/I) techniques were needed in order to compete with other
analytical methods being developed. MALDI MSz, developed by Karas and Hillenkamp
in 1989, has become the standard D/I technique for analyzing biological molecules
including peptides, proteins, and DNA. This is due to the fact that, when used with a
time-of-flight (TOF) mass analyzer, it has a theoretically unlimited mass range and
femtomole sensitivity. The only real disadvantage of MALDI has been low resolving
power, since it is primarily used with TOF mass analyzers. New instrumental designs
such as delayed extraction and reflectron technology have greatly improved this
drawback. Isotope peaks of ionized proteins with masses of up to 15,000 Da can now be
resolved.

A typical desorption/ionization experiment is performed by mixing an analyte
molecule in an appropriate matrix material and depositing this onto a probe or sample
plate which is placed into the vacuum chamber of the source region of the mass
spectrometer. The matrix material is approximately 103-10411 in molar excess of the
analyte. In the fast-atom bombardment experiment, a stream of fast moving inert gas
atoms such as xenon bombards the surface of the liquid matrix solution containing the
analyte. These collisions with the matrix material produce protonated molecules that are

ejected or desorbed as shown in Figure 1.1.

 

Xe +
IA+HI

 

  

 

 

 

 

 

 

 

 

 

 

Mass
+10k glycerol (1-2uL) Analyzer
matrixzanalyte ratio 103- 104: l
G=glycerol matrix =analyte
[G+H]+ Detector

 

 

 

[A+H1* M

 

Relative Intensity

 

 

 

 

 

 

 

 

Figure 1.1 Description of the fast atom bombardment (FAB MS) experiment

It is not clear whether preformed ions are desorbed or whether ions are formed at
the matrix surface or the high-pressure region just above the surface called the selvedge
region.3 Regardless of how the ions are formed, most organic molecules produce singly
charged ions in the protonated form, [M+H]+. For inorganic salts such as NaCl, ions
need only be desorbed and not ionized, resulting in a gas phase ion, such as NaI. The
formed ions are then accelerated through a double-focussing mass spectrometer where
they are sorted according to mass and detected by an electron multiplier.

In matrix-assisted laser desorption/ionization,2 the major differences regarding
sample preparation are that a solid sample is used and the energy is deposited by
absorption of a laser pulse rather than by collisions with fast-moving atoms. A sample is
prepared by dissolving an analyte molecule in a solution of UV absorbing matrix, in a
volatile solvent, and allowing it to air dry on a sample plate for several minutes followed
by introduction to the mass spectrometer. Photons at 337 nm from a nitrogen laser are
absorbed by the matrix which contains analyte molecules, and the analyte is then ionized
and ejected from the sample and presented to a TOF mass analyzer as shown in Figure
1.2. Similar to FAB spectra, analyte ions are mostly singly charged species that result

from protonation, and show little or no fragmentation.

 

sample preparation (solution)

 

 

 

    

 

air dry
M+H +
11V, 337 nm [ ] A+H]+
f organic solid
~5 mm2
matrix/analyte(fmol)
(104/1)

Figure 1.2 Description of the matrix assisted laser desorption/ionization
(MALDI MS) experiment.

11. Use Of Chemical Additives In Desorption/Ionization Mass Spectrometry

Unfortunately, some molecules in mass spectrometric experiments do not yield
representative ions in the gas phase. In these cases, some scientists have turned to adding
a third component, known as an additive, to the matrix/analyte mixture.4 Chemical
additives have been used successfully in the early FAB days and are still making a
valuable contribution to this day. For example, in 1986, Ligon and Dom5 reported the
enhanced detection of phosphate-containing biomolecules such as adenosine triphosphate
(ATP). They noticed that the presence of phosphate groups caused a problem in the FAB
experiment. Phosphate-containing molecules do not ionize well and give rise to low
intensity signals in the mass spectrum. ATP has three phosphate groups and carries a 3-
charge in the common glycerol matrix. The FAB experiment gives rise to mainly singly
charged ions. Glycerol has chemical properties similar to those of water. By simply
adding a cationic surfactant such as hexadecylpyridinium acetate to the matrix/analyte
solution, the ionization efficiency was substantially enhanced and a very abundant peak
representing [M-H]' was observed in the negative ion mode. It can be explained by
arguing that the positively charged surfactant makes the negatively charged ATP
molecules more surface active and that ATP can reduce its charge upon D/I by acquiring
some positive charges from the surfactant. Orlando demonstrated in 1992 that peptides
could be desalted with crown ethers to improve their D/I efficiency in F AB.6 Since
peptides are ionized by protonation, resulting in the formation of [M+H]+ ions, alkali
metal ion impurities such as Na+ and K+ compete with protons resulting in a lower yield

of protonated molecules and cation adduct formation. Metal adducts inhibit the ability to

accurately determine the molecular weight of an analyte by decreasing the sensitivity and
decreasing the resolution by distributing the peak over many m/z values. Crown ethers
such as 18-crown-6'5 have a high affinity for alkali metal cations and remove free cations
from the glycerol solution. Other additives that have been used successfully in FAB
include the addition of metal ions7 such as Ag+ and Li+ to cationize species that yield
inconclusive results or don’t contain efficient protonation sites8 and the addition of
simple acids and bases to increase the solubility of analyte molecules and aid in the
production of protonated/deprotonated species. In MALDI MS, common additives for
synthetic polymers9 include metal ions such as Na+, and LiI. It was found in the early
1990’s that synthetic polymers such as polyethylene glycol and polymethyl
methacrylate10 do not protonate efficiently and usually result in a distribution of low
abundance peaks. Hillenkamp et al. ” found that the addition of alkali metal ions to the
matrix/analyte solution prior to crystallization resulted in cation attachment of each
repeating polymer unit, which increased the ion yield for polymer peaks in the mass
spectrum. It has also been shown that transition-metal ions such as Cu+ and Ag" can be
used for hydrocarbon polymer analysism’l3 Other MALDI additives include acids such
as trifluoroacetic acid (TFA) and sugars such as fucose14 which have been used to
increase analyte solubility and enhance protonation.

Besides their obvious use in spectral enhancement, chemical additives are
convenient due to their simplicity of use. The use of an additive involves doping of the
matrix/analyte mixture with a third chemical compound. This only adds seconds to a
typical FAB or MALDI experiment. The simplicity of use allows for a rapid screening of

many different additives in a short time period. It is important to realize that the sample

solution volumes are only 1-2 uL and experiments can be performed in these very small
volumes. It is well known that reduction of analyte molecules can occur with certain
matrices such as nitrobenzyl alcohol15 in the fast atom bombardment experiment, and
enzymatic digestions have been performed in 2 uL volumes for MALDI analysis directly
on the sample plate before droplet evaporation.16 The use of chemical additives involves
in situ chemical reactions performed directly on the sample stage just prior to mass
spectrometric analysis. Figure 1.3 shows the general use of an additive in the fast atom

bombardment experiment.

matrix/analyte

 

additive a a acquire
P a a a a spectrum’

‘ Intensity

 

 

 

 

 

m/z

 

Figure 1.3 The use of an additive or a third component in a desorption/ ionization (D/I)
experiment.

The use of chemical additives results in an increase in signal intensity of targeted
analyte molecules. When evaluating mass spectra, mass spectrometrists are rarely
concerned with absolute signal intensities, but are only concerned with the relative
intensity of signals. The intensity scale is usually given by data systems in the form of
“counts” or “arbitrary Y”. It is very difficult to relate the number of counts back to the
current generated from ions hitting the detector. Once ions hit the detector, the current is
multiplied by approximately 105 and an analog-to-digital (A/D) converter converts this
current to a voltage in the mV scale. This voltage is then converted to some arbitrary Y
scale by the data acquisition software which can have a range from 1 to 100 or 1 to

60,000 depending on the software used.

CHAPTER 2 - Enhanced Detection Of Multiply Charged Transition-Metal

Complexes In FAB MS Using Organic Acids

Multiply charged molecules have always been a problem in D/l experiments since
these techniques mainly produce singly charged ions. In the fast atom bombardment
experiment, it has been estimated by Takayama17 that the maximum amount of available
energy is ~19.9 eV, which is sufficient only to desorb singly charged ions in most cases.
Doubly charged ions usually have desolvation energies greater than 20 eV.17 Transition
metal-containing compounds can accumulate a high charge since many can be divalent
and trivalent such as Pt2+ and F c”. When several cations form clusters, the charges can
be much higher. These complexes contain outer-sphere counteranions in order to make
neutral salts. In solution for example, a metal complex that possesses an 8+ charge, will
have to acquire seven anions from the matrix solution in order to produce a 1+ ion. It has
been reported in the literature18 that this may occur with certain anions such as BF4’, OTf
, and PF6'. In my experience with similar complexes studied, these anions do not bind to
the complex upon D/I and no spectra are observed. APPENDIX A contains a research
article (Inorg. Chem, 1998, 37, 1833-1840) of a detailed study describing the use of
trifluoromethanesulfonic acid, HOTf or triflic acid, as an additive for the successful
desorption of multiply-charged transition metal complexes. The molecules contain
neutral tetrathiafulvalene phosphine ligands of the type 0-P2 and P4 (shown in
APPENDIX A). This additive works by releasing a proton in solution upon fast atom
bombardment, thus allowing the triflate anion to bind to the positively charged metal

ions. The spectra always result in a singly charged positive ion of the form

[M2+n+L"(,,-1)]+, where M represents the complex containing the transition metals and
neutral ligands and L is OTf. Since the time this work was published in Inorganic
Chemistry in 1998, we have shown that triflic acid can be used successfully in the mass
spectral analysis of other highly charged complexes as well. Specifically, the square
complex [Ni4(prZ)4(CH3CN)3][BF4]3, where bptz represents bis-pyridine tetrazine,
C|2H3N5, shown in Figure 2.1 was subjected to FAB MS studies in the presence of triflic

acid.

 

Figure 2.1 Structure of [Ni4(bptz)4(CH3CN)g](BF4)3.

In solution, this Ni4 square cluster bears an 8+ charge, as it is not coordinated to
the tetrafluoroborate anions. Figure 2.2a shows the FAB mass spectrum of
[NI4(bPIZ)4(CH3CN)3](BF4)3 using nitrobenzyl alcohol as the matrix. Note that no useful
data are observed for this salt under these conditions. Figure 2.2b shows the same salt
with the addition of 1 uL of triflic acid to the matrix/analyte mixture (see APPENDIX A
for experimental details). Notice the peak at m/z 2223, which represents
{[Ni4(bptz)48+](OTf)7}+, is a detectable singly charged ion. The peak at m/z 2174
represents {[Ni4(bptz)47+](OTf)6}+ with one reduced NiI, caused by nitrobenzyl alcohol.
The peaks at m/z 737, 886, and 1035 represent half of the square, Ni2(bptz)2, with one,
two and three triflate ions, respectively. One would only expect to observe the peak at
m/z 1035 representing [Ni2(bptz)2](OTf)3+ since Ni is divalent and the cluster possesses a
4+ charge. The peaks at m/z 737 and m/z 886 result from the reduction of Ni by the
nitrobenzyl alcohol matrix. The data show no evidence for peaks due to Ni3 containing
species. Our initial interpretation of these peaks suggested insight into how the square
forms. It was believed that two halves of the squares come together to form the intact
complex during the synthesis. Alternatively, these data could also suggest that the

addition of HOTf causes the intact square cluster to dissociate into two halves.

ll

 

 

 

3:" 'l 'I

an I I?

a l

3 'I

,E‘: b) 886

o . +

.5 y 737 1N1:%§;Zh](OTD3

.3 /

a)

0!.
[Ni4(thZ)4l(0Tf)7+
1174 2223

 

 

 

 

 

 

 

 

I l l l l l I l l
800 1000 1200 1400 1600 1800 2000 2200 2400
m/z

Figure 2.2 (a) FAB mass spectrum of [Ni4(bptz)4](BF4)g in a m-nitrobenzyl alcohol
matrix. (b) FAB mass spectrum of [Ni4(bptz)4](BF4)3 in a m-nitrobenzyl alcohol
matrix with the addition of 1 uL of triflic acid.

Figure 2.3a shows an expanded portion of the peak at m/z 2223 displaying the isotope
distribution. This agrees with the theoretical isotope distribution for the elemental

composition of C55H32F2187021N24Ni4 shown in Figure 2.3b.

12

 

 

2223
i 2221

2225

 

 

 

 

 

 

 

Relative Intensity

 

 

Relative Intensity

 

 

 

 

 

 

 

 

 

 

 

 

A ,I lllllt..
m/z

 

 

 

 

 

 

2210 w 2220 2230 2240
m/z

Figure 2.3 (a) Expanded portion of the peak at m/z 2223 representing
[Ni4(bptz)4](OTf)7+. (b) Theoretical isotope distribution of the peak at m/z 2223.

It has been shown in our laboratory by Jasminka C. Dragovic that square complexes of
the same form that have been synthesized with replacement of the transition metal ion
with Fe2+ can be successfully detected using triflic acid as an additive.

Interestingly, salts of triflate such as NaOTf did not assist the highly charged
complexes in reducing their charge; only the acid form worked well. We believe that
ion-dipole interactions between the charged metals and the solvent (matrix) molecules
create an environment where counteranions such as OTf cannot penetrate the solvent
shield. However, the acid form, HOTf, can easily become part of the solvent system

surrounding the charged metal and then release a proton upon fast atom bombardment.

l3

. + FAB . + + +
[Nutbptzms (How...) ——> {[N14(bptz)4]8 (OTfm (g) + 7H (son

With BF 4' as the counteranion and no additives, the compounds studied did not
yield mass spectra. The decision was then made to try tetrafluoroboric acid, HBF4, as an
additive. When aqueous 48% HBF4 is used as an additive, a peak appears at m/z 1788
which corresponds to [Ni4(bptz)4(BF4)7]+ (Figure 2.4). Peaks due to the reduction of Ni2+
also appear at m/z 1614 and m/z 1701. Notice that two major peaks also appear at m/z
1018 and m/z 1037. It is not clear what these peaks represent but they are the most
intense peaks in the spectrum. It was originally assumed that these peaks were related to
the half square but that would produce a peak at m/z 850, [Ni2(bptz)2(BF4)3]+, which does
not exist. However the peak at m/z 1037 could represent the half-square complex with
three triflate anions attached. Possible contamination of triflic acid in the source of the
mass spectrometer could explain this peak. However, contamination from triflic acid
would not explain the peak at m/z 1018. Interestingly, the mass difference between m/z
1037 and m/z 1018 is 19, the mass of a fluorine atom that suggests possible F ' abstraction
from HBF4 or BF4'.

The results show the importance of using charge-lowering matrix additives in
FAB MS for characterizing highly-charged complexes. NMR or X-ray crystallography
did not successfully characterize many of the structures discussed in this chapter and the
use of triflic acid in combination with FAB MS was the only successful means for

structural determination.

l4

Relative Intensity

 

1018

 

 

 

 

 

_ 1037
[Ni4(bptz)4(BF4)-;]+
- 1788
1614 1701 /
1000 1:200 1:100 1600 1i300 2000

m/z

Figure 2.4 FAB mass spectrum of [Ni4(bptz)4](BF4)8 in a m-nitrobenzyl
alcohol matrix with the addition of 1 uL of 48% tetrafluoroboric acid.

15

CHAPTER 3 - Enhanced Detection Of Phosphorylated Peptides And Proteins In

MALDI MS Using Ammonium Salts

I. Phosphopeptide Enhancement Study With B—Casein

In a manner similar to positively charged metal complexes, phosphate-containing
biomolecules also cause problems in desorption/ionization mass spectrometry due to their
high negative charge. In Chapter 1, a situation was described where an additive was used
to enhance the spectra for ATP in the FAB experiment. Although a solid sample is used
in MALDI, the situation is similar. In the matrix/analyte solution prior to crystallization,
analyte molecules such as phosphate-containing molecules are negatively charged.
During rapid droplet evaporation, charged analyte molecules become trapped within the
matrix crystals. This can be referred to as a “solid solution”.

Phosphopeptides are very important since protein phosphorylation is probably the
single most common and important reversible intracellular signal transduction event.19
Understanding the regulation of protein function by phosphorylation/dephosphorylation
is a key objective in many areas of biomedical research. Liao et al.20 showed that
phosphopeptides are at least ten times less sensitive than their nonphosphorylated forms.
This can be a problem when trying to identify phosphopeptides in an unfractionated
protein digestion mixture. APPENDIX B contains an article (J. Amer. Soc. Mass
Spectrom, 1999, 10, 35-44) on the enhanced detection of phosphopeptides in MALDI
MS using ammonium salts as matrix additives. Salts such as diammonium citrate and

ammonium acetate can substantially improve the desorption/ionization efficiency of

16

phosphopeptides in a mixture containing nonphosphorylated peptides. In some cases, the
relative intensity of the phosphopeptides increases to the point where the
phosphopeptides become the largest peaks in the spectrum so that they can be identified
in a complicated mixture. In the ideal case, this situation eliminates the need for 32P
radioactive labeling and also eliminates the need for HPLC fractionation. We have
shown that the number of phosphate groups can be determined by performing post-source
decay (PSD) analysis, an MS/MS technique where a peak of interest is selected and its
fragments are analyzed. The ammonium additive works by binding to the negatively
charged phosphate groups during solvent evaporation, thus neutralizing the charge and
displacing cations. Upon laser ablation, neutral NH3 is lost and protonated peptide results

in the mass spectrum as shown below.

+ _ MALDI +
NH4 O'PO3(peptide) D [HO'POBGIeptideflH + NH3(a)

 

The most common method of phosphorylation site mapping involves labeling a protein
with 32F followed by an enzymatic digestion, usually with trypsin.20 Trypsin cleaves
proteins at the C-terminus of arginine (R) and lysine (K) amino acid residues. The
peptides are then separated by reverse-phase HPLC and fractions are collected. Each
fraction is then tested for radioactivity using a phosphoimager. The radioactive fractions
are prepared for MS analysis to determine the location of the phosphate groups along the
amino acid chain. Phosphopeptides can be identified in one of two methods using
MALDI. Liao et al.20 showed that a phosphopeptide could be dephosphorylated using an

enzyme and the loss of 80 mass units (HPO3) can be detected by MALDI. Another

17

method which was first discussed by Annan and Carr19 involves performing PSD analysis
and observing characteristic losses of 98 Da (H3PO4) and 80 Da (HPO3). These
techniques both rely upon the ability to obtain a strong signal for a phosphopeptide in the
mass spectrum.

Unfortunately, the enhancement of phosphopeptides using diammonium citrate
does not work well for all complex peptide mixtures, including mixtures that are
recovered from PVDF membranes or polyacrylamide gels following enzymatic
digestions. As a result, the following two questions need to be answered. Why are
complex peptide mixtures containing phosphopeptides difficult to analyze? What
optimum amount of ammonium salt is needed to enhance the spectra of phosphopeptides
within complex mixtures? In an effort to answer these questions, the decision was made
to investigate several concentrations of a digestion mixture with several concentrations of
the diammonium citrate additive using the common peptide matrix a—cyano-4-
hydroxycinnamic acid (4-HCCA) as the matrix. The experimental procedure was as
follows: A tryptic digestion of B-casein (MW 23,983 Da), a calcium transport protein
found in milk, was performed using 0.25 pg of trypsin at pH=8.2 at 37°C for 20 hours.
The initial concentration of B-casein before digestion was then diluted to l pmol/uL, 100
fmol/uL, and 10 fmol/uL afier digestion. Saturated solutions of the matrix, 4-HCCA,
were prepared in 1:1 acetonitrile/water with several different concentrations of
diammonium citrate including 1 mM, 10 mM, 12.5 mM, 25 mM, 50 mM, and 100 mM.
A 1 11L volume of analyte solution was mixed with l 11L of matrix solution on a 100-well
gold sample plate and allowed to air dry. Positive ion MALDI MS was performed on a

PE Biosystems Voyager STR DE-reflectron mass spectrometer in the linear mode.

18

An initial observation was made regarding mixtures that contained a very
abundant peak. Figure 3.1 shows the spectrum of 1 pmol of a B-casein digestion mixture.
Since the sequence of the protein is known, we can calculate the peptide masses that
result after digestion by mass-mapping.20 Note that in the presence of the
nonphosphorylated peptide at m/z 831, very little information can be extracted about the
remaining peptides. The presence of very intense peptide signals can suppress other
peptide signals in the mass spectrum. The decision was then made to remove the peptide
at m/z 831 in order to obtain useful information about the peptides of interest. In order to
remove the peptide at m/z 831, the digestion mixture was subjected to reverse-phase
HPLC (see Chapter 3, section II for gradient details) and fractions were collected. The
fractions were then checked by MALDI MS to determine which fraction contained the
peptide at m/z 831. Fractions were then recombined except for the fraction containing
the peptide at m/z 831. Samples were prepared by mixing 0.5 uL of analyte with 0.5 uL

of matrix solution on a 400-well Teflon sample plate and allowing the mixture to air dry.

19

 

_ 831

 

 

Relative Intensity

 

 

1000 1500 2000 2500 3000
m/z

Figure 3.1 MALDI mass spectrum of the B-casein digestion mixture before
removal of the peptide at m/z 831.

[BI-casein produces a somewhat complex mixture that contains two major
phosphopeptides; these form protonated molecules at m/z 2063 and m/z 3124. It was
quickly found that concentrations of 50 mM and 100 mM of diammonium citrate are too
high and interfere with the adequate crystallization of the matrix and analyte. This results
in poor signal intensity for all of the digestion products. Also, 1 mM diammonium citrate
had little effect on the phosphopeptides. The most useful concentrations were in the
range from 10 mM to 25 mM of the ammonium salt, which yielded the most intense
phosphopeptide peaks. Upon further investigation with other concentrations, 12.5 mM

ammonium concentration in the matrix produced a spectrum in which the

20

phosphopeptides were nearly the most intense signals in the spectrum. Figure 3.2 shows
the spectrum before and after the addition of diammonium citrate and after the salt was
added. Figure 3.2a shows the MALDI spectrum of 500 finol of the B-casein digest using
the matrix a-cyano-4-hydroxycinnamic acid only. Note that the peaks at m/z 2063 and
m/z 3124 that represent phosphorylated peptides are very small in the presence of
nonphosphorylated peptides. Figure 3.2b shows the same spectrum after the addition of
12.5 mM diammonium citrate. It is clear that the ions for the phosphorylated components
increase substantially upon ammonium ion addition. When the ammonium ion
concentration was increased to 25 mM, however, the peak at m/z 2063 became the largest
in the spectrum but the peak at m/z 3124 that contains four phosphate groups began to
decrease in relative intensity. The results show that it is very difficult to use a single
concentration of the ammonium additive in order to optimize the response. Frequently,
the concentration of recovered peptides is not known, so it is best to try several
concentrations of ammonium ions from 10 mM to 25 mM in order to maximize the signal
intensity of the phosphorylated components. It is important to note that this information
could only be obtained after the peptide that represents the peak at m/z 831 was removed.
This could possibly explain why it is difficult to use additives in complex mixtures where

peptides of many different concentrations are present.

21

Relative Intensity

 

a) No additive
* 33-48 (1p) 1-25 (4p)
W
b) 12.5 mM (NH4)2HCit
ISIWWW
_ 25 mM (NH4)2HCit

 

 

 

 

 

 

 

 

 

 

I I I I
1500 2000 2500 3000
m/z

Figure 3.2 (a) MALDI mass spectrum of the digestion mixture of B-
casein using 4-HCCA. (b) MALDI mass spectrum of the same mixture
with the addition of 12.5 mM diammonium citrate. (c) 25 mM
diammonium citrate. The asterisks represent nonphosphorylated
peptides.

22

11. Determination Of The Sites Of Phosphorylation Of The Protein Bril Kinase

By using our newly developed method of ammonium salt addition followed by
PSD analysis, several phosphorylation sites of a plant protein kinase called
brassinosteroid 1 were determined. Brassinosteroids (BRs) are growth-promoting
natural products found at low levels in pollen, seeds, and young vegetative tissues
throughout the plant kingdom. This project was carried out in collaboration with
Professors Steve Clouse and Steve Huber at North Carolina State University. The sample
was submitted to us after being separated by gel electrophoresis and electroblotted onto a
polyvinylidene fluoride (PVDF) membrane. The experimental procedure was as follows:
The PVDF membrane containing the protein was rinsed with methanol and cut into 1mm2
pieces. The membrane pieces were then placed in a solution containing 1.0 ug of porcine
trypsin in 10% acetonitrile, 1% n-octyl-B-D-gluco-pyranoside (NGP), 100 mM Tris-HCI
(pH=8.4) and incubated at 37°C for 24 hours to digest the protein. The solution
containing the membrane pieces were then vortexed, sonicated for five minutes, and
centrifuged for three minutes. The solution containing digestion products was transferred
to an eppendorf tube. Two washings of the membrane pieces were carried out with 0.1%
TF A and transferred to the same tube. The solution was then injected into a Microm
BioResources, Inc. UMA HPLC system containing a C18 column (1 mm inner diameter).
Solvent A contained 95% HzO/5% ACN/O. 1% TF A. Solvent B contained 10% H20/90%
ACN/O. 1% TFA. The gradient was as follows: 5% B to 65% B from 0 to 50 minutes;
65% B to 95% from 50 to 51 minutes; 95% B from 51 to 53 minutes; 95% B to 5% B

from 53 to 55 minutes; 5% B from 55 to 63 minutes. Fractions were collected by hand

23

and 1 11L of each fraction was checked for 32P incorporation using a phosphoimager. The
radioactive fractions were then analyzed by MALDI mass spectrometry to determine the
sites of phosphorylation.20 Subsequent digestions were performed on specific fractions
for further insight into the exact location of phosphate groups. The incompletely digested
peptide 28-40 was subjected to a second tryptic digest that resulted in peptide 28-35.
Peptide 219-248 was subjected to CNBr cleavage (70% formic acid, 24 hours, 37°C)
which resulted in the phosphopeptide 224-248. CNBr cleaves peptides at the C-terminus
of methionine (M) residues. The peptides 56-85 and 343-357 were subjected to a
digestion using AspN (pH=8.0, 24 hours, 37°C), resulting in the phosphopeptides 56-60,
72-81, and 351-357 AspN is a proteolytic enzyme, which cleaves at the N-terminus of
aspartic acid residues.

MALDI MS spectra were recorded on a PerSeptive Biosystems (Framingham,
MA) Voyager Elite delayed extraction time-of-flight reflectron mass spectrometer
equipped with a nitrogen laser (337 nm, 3 ns pulse). Typically, 128 laser shots were
averaged. The post-source decay (PSD) experiment was performed in the reflectron
mode with an accelerating voltage of 21 kV, a grid voltage of 73.0% of the accelerating
voltage, and a guide wire voltage of 0.15% of the accelerating voltage. The timed ion
selector was set at m/z 1607.6 and three spectra, with mirror voltagezaccelerating voltage
ratios of 1.00, 0.96, and 0.67 were stitched together to obtain the PSD spectrum. The
samples were prepared by mixing 1 uL of analyte solution with 1 11L of saturated 2,5-
dihydroxybenzoic acid containing 25 mM diammonium citrate. The protein was supplied
with a peptide tag attached to the N-terminus that is used for protein identification. The

sequence of the protein, using the standard one letter code for amino acids, is as follows:

24

TAG:

MKRRWKKNFIAVSAANRFKKISSSGALLVPRGSGSGDDDDK

KINASE DOMAIN:
REMRKRRRKKEAELEMYAEGHGNSGDRTANNTNWKLTGVKEALSINLAAFEKP
LRKLTFADLLQATNGFHNDSLIGSGGFGDVYKAILKDGSAVAIKKLIHVSGQGDR
EFMAEMETIGKIKHRNLVPLLGYCKVGDERLLVYEFMKYGSLEDVLHDPKKAG
VKLNWSTRRKIAIGSARGLAF LHHNCSPHIIHRDMKSSNVLLDENLEARVSDF GM
ARLMSAMDTHLSVSTLAGTPGYVPPEYYQSFRCSTKGDVYSYGWLLELLTGKR
PTDSPDFGDNNLVGWVKQHAKLRISDVFDPELMKEDPALEIELLQHLKVAVACL
DDRAWRRPTMVQVMAMFKEIQAGSGIDSQSTIRSIEDGGFSTIEMVDMSIKEVPE
GKL

Proteins can be phosphorylated at serine (S), threonine (T), and tyrosine (Y) residues.
More than 90% of all phosphorylation occurs at serine residues with less than 1%
occurring at tyrosine residues. It was previously determined using radiolabeling that all
phosphorylation in this protein occurs at S and T only.21 The results are shown in Table
3.1. Some of the phosphorylation sites could not be determined unambiguously in the
case where the peptides contained more than one possible 8 or T residue that could be
phosphorylated. Figure 3.3a shows a portion of the MALDI mass spectrum from one
HPLC fraction of the tryptic digest of BrIl kinase protein. The numbers identify the
proteolysis products and p represents a phosphate group. The peak at m/z 1607.6
represents peptide 28-40, which contains two phosphate groups. The peaks marked with
an asterisk result from nonspecific cleavage products. Figure 3.3b shows a portion of the
MALDI PSD spectrum of peptide 28-40. The presence of two phosphate groups is
confirmed by the sequential loss of 98 (H3PO4) and 80 (HPO3) for each phosphate group.

When ambiguous phosphorylation sites are present, additional enzymes can be used in

some cases to further identify the exact location of the phosphate groups. This approach

25

depends upon the presence of a cleavage site in the target peptide. An alternative

approach to directly locating phosphorylation sites by MALDI MS is ladder sequencing.

With this technique, a chemical or enzyme cleaves a peptide at each amino acid residue

from the peptide chain forming a peptide ladder.22a However, the resulting MALDI

spectrum can suffer from significant signal suppression and becomes complicated to

interpret when peptide mixtures are present.22b

Phosphorylation Sites of Bril Kinase

 

 

Seguence (m)
TAG:

8-NFIAVSAANR- l 7

2 l-ISSSGALLVPR-3 l
KINASE DOMAIN:

l l-EAELEMYAEGHGNSGDR-27
28-TANNTNWK-35

4 l -EALSINLAAFEKPLR-55
56-KLTFA-60
72-DSLIGSGGFG-81
l64-LNWSTR- 169

224-DTHLSVSTLAGTPGYVPPEYYQSFR-248

351-DSQSTIR-357

358-SIEDGGFSTIEMVDMSIK -375

[M+H]*

1143.2

1 180.3

1942.0

1108.0

1753.0

659.3

989.2

856.9

2786.0

886.3

2040.2

Sites of Phosphorylation

S-24

T-28, T-32

S-44

T-58

1 site: S-73 or S-77

1 site: S-167 or T-168

3 sites: T-225, S-228, S-230,
T-231, T-235, or S-246

1 site: S-352, S-354, or T-355

1 site: S-358, S-365, T—366,
or S-373

 

Table 3.1 Lists the phosphorylation sites of the Bril kinase protein found in this
investigation. Note that the last five peptides show ambiguous sites for possible

phosphorylation.

 

Relative Intensity

Relative Intensity

 

8) 28-40 (2p)
‘~1607ri

L

160-171 (1p)

*

164-171(11)) 331-3420.. ,,

_ u \ -w .1,

1'100 1200 1300 1300 1300 1600 1'700

 

 

 

 

 

 

 

 

 

 

 

m/z
D) 1607.6
-1 p
.1 _2p 1509(6 ~H3PO4
1411.6 -H3PO4
(
-HP03 -HPO3
‘—— m I

 

 

 

1'300 1'350 1'400 1'450 1'500 1'550 1'600
m/z

Figure 3.3 (a) MALDI mass spectrum of a portion of one fraction of
a tryptic digestion mixture of Bril kinase protein. (b) PSD spectrum
of m/z 1607.6, a tryptic digestion product containing two phosphate
groups. The asterisks represent incomplete digestion products.

27

In this case, the peptide 28-40 contains three threonine residues that could be
phosphorylated, but only two sites are phosphorylated as determined by PSD analysis.
This peptide resulted from an incomplete digestion since it contains a lysine residue at
position 35. A second tryptic digest of this peptide was performed which allowed for
determination of the exact phosphorylation. Unfortunately, PSD does not work well with
phosphopeptides in determining the specific residue to which the phosphate group(s) is
attached. The most significant fragment is the loss of a phosphate group, which increases
almost exponentially as laser power is increased in order to induce further fragmentation.
This results in a substantial decrease in signal intensity of the precursor ion, which
inhibits fragmentation of the peptide backbone. Phosphate loss seems to be the
kinetically favored fragmentation process. In order to obtain information about the exact
location of the phosphate group, it is necessary that the phosphate group be lost with the
amino acid residue during the fragmentation process.

The determination of phosphorylation sites by MALDI MS can be very
challenging due to the negative charges of the phosphate groups, which inhibit the
successful desorption/ionization of phosphopeptides in mixtures. The use of ammonium
salts as additives can help to overcome this difficulty and in some cases, produce signals

for phosphopeptides that are the most intense in the spectrum.

28

CHAPTER 4 - Enhanced Detection Of Oligonucleotides In MALDI MS Using The

Tetraamine Spermine

Analysis of oligonucleotides such as DNA by mass spectrometry lags far behind
the analysis of peptides and proteins. In order for mass spectrometric techniques such as
MALDI MS to be successful for routine sequencing studies, these obstacles must be
overcome. DNA sequencing by MALDI MS is limited to about 20-30 bases since the
resolution of the experiment decreases with increasing mass. This does not compete with
gel electrophoresis based techniques since some gels allow the reading of sequences as
long as 1200 nucleotides in a few hours.23 The major benefit of using mass spectrometric
based techniques for DNA sequencing is speed, since most MALDI MS experiments take
only seconds to acquire adequate spectra. However, the resolution of the experiment
decreases with larger DNA strands since the number of phosphate groups present
increases with additional nucleotides. Cation adduct formation becomes more
problematic with increasing base numbers of DNA, since the difference between DNA
nucleobases is only 5 mass units in some cases. This has long been recognized as a
limitation of the experiment.24

Since MALDI generates singly charged ions, the high charge that accumulates
along the phosphate backbone of DNA can be very problematic, more so than for
phosphopeptides. This is mainly due to poor ionization efficiency and cation adduct
formation which gives rise to poor resolution. Spermine is a polyamine that is present in
all living cells. Its primary function is to shield the negative charges of DNA phosphate

groups as shown in Figure 4.1.

29

NH 2 _ 0— :0 N
CIH 2 \ FN N
CH 0 N
'2 “if W
. .. N
spermlne C.“ 2 9‘ NW DNA
CIH 2+ 'O—P\:o
CH 2 N
1 7/’
am a). 0
cm, ................................... -O_P\=O
...... / N
0°. 0

Figure 4.1 The interaction of spermine with DNA in a living cell as well as in
the MALDI MS experiment.

30

This reduces charge repulsion within a single DNA strand and is one reason why
condensed DNA structures can easily fit into living cells.

In the MALDI experiment, it has been shown that ammonium salts such as
ammonium acetate and diammonium citrate can be effective in reducing cation adduct
formation and increasing the mass spectral response of oligonucleotides. However,
ammonium ions as additives fail at high salt concentrations and at high mass.
Interestingly, spermine is used in some cases in order to crystallize DNA for X-ray
studies. The decision was made to evaluate spermine as an additive in the MALDI
experiment to help aid in the DH properties of oligonucleotides. The results of using
spermine as a matrix additive in MALDI showed substantial enhancements. APPENDIX
C contains a research article (Anal. Chem, 1999, 71, 2866-2870) where spermine was
used in high salt concentrations and in situations where only low femtomole amounts of
DNA were present. Spermine clearly works better than ammonium salts as a matrix
additive. We have found that spermine is most effective when used with the UV
absorbing matrix 2-aza-2-thiothymine (ATT), but it also works well with 80% anthranilic
acid/20% nicotinic acid (80/20) and 2,4,6-trihydroxyacetophenone (THAP). After this
article was published, a collaboration with Dr. Philip Ross at PE Biosystems in
F rarningham, MA began. The purpose of the collaboration was to test spermine as an
additive in the presence of very high salt and buffer concentrations. These are common
components that are present in polymerase chain reactions (PCR).25 The combination of
PCR and MALDI has the potential to be used as a rapid and accurate method for DNA
diagnostics such as single nucleotide polymorphisms in diseased genes.25 Figure 4.2a

shows 1 pmol of a 15-mer oligonucleotide in a high salt reaction buffer (NaCl, glycerol,

31

triton X-100) including 50 mM KCl using 3-HPA/25 mM diammonium citrate as the
matrix. The exact concentrations of some components are not listed due to proprietary
reasons. Under these conditions, little molecular weight information can be obtained
from the spectrum due to Na+ and K+ adduct formation. Figure 4.2b shows the same
mixture using 2-aza-2-thiothymine and 15 mM spermine as the matrix. Note that a single
peak at m/z 4625 can now be identified in the spectrum, which demonstrates that using
spermine in the matrix allows for a high tolerance of impurities. Figure 4.3 shows an
example of a mixture of single DNA strands in a PCR buffer. PCR is primarily used to
amplify very small quantities of DNA to amounts that can be easily analyzed by gel
electrophoretic methods. Figure 4.3a contains a mixture of six oligonucleotides in a PCR
buffer including small amounts of glycerol, triton X-100, KCl, and the enzymes alkaline
phosphatase and thermosequenase. Usually, very little mass spectral information can be
obtained in the presence of buffers and viscous solvents. Figure 4.3b shows the same
mixture using ATT and spermine as the co-matrix where it can be seen that most of the
components can be observed.

The results of this study indicate that the effectiveness of spermine as an additive
in MALDI MS has the potential for making a major impact on the analysis of
oligonucleotides since PCR reactions usually require time-consuming purification steps
prior to analysis by mass spectrometry. Also, the sequence of longer DNA strands may
be possible by MALDI since spermine allows for more highly resolved peaks than

ammonium ions at higher masses.

32

Relative Intensity

 

 

 

 

 

 

 

 

 

 

 

I a) [M'HL ./[M+Na-2H]'
/[M+K-2H]'
I I
I W, I
III
" b) [M-H]‘

 

 

 

I l I I I I I
4200 4400 4600 4800 5000 5200 5400
m/z

Figure 4.2 (a) MALDI mass spectrum of a single oligonucleotide in the
presence of a high salt concentration buffer in 3-hydroxypicolinic acid with
diammonium citrate. (b) The same sample in 2-aza-2-thiothymine with 15 mM
spermine.

33

Relative Intensity

 

 

 

 

 

 

 

 

1

b)

 

 

 

 

I I I I I I
3000 4000 5000 6000 7000 8000
m/z

 

 

 

Figure 4.3 (a) MALDI mass spectrum of a mixture six oligonucleotides
in the presence of a typical PCR buffer in 3-hydroxypicolinic acid with
diammonium citrate. (b) The same mixture in 2-aza-2-thiothymine with 15
mM spermine.

34

CHAPTER 5 — Evidence Of Dirhodium Bis-Acetate Units Covalently Bound To 1,2

Intrastrand GG And AA Sequences 01' DNA

The following chapter contains the text of a manuscript that has been accepted for
publication in article in Journal of the American Chemical Society. The title reads
“Evidence for Binding of Dirhodium Bis-Acetate Units to Adjacent GG and AA Sites on

Single-Stranded DNA” .

Abstract

Dirhodium tetraacetate has also been found to be a potent inhibitor of cellular DNA
synthesis, but relatively little is known about its interactions with nucleic acids. By using
matrix-assisted laser desorption/ionization mass spectrometry (MALDI MS) and
enzymatic digestions, we show that the dirhodium bis-acetate unit forms an adduct with
AA and GG sites in short DNA strands. Experimental conditions that involve the use of
spermine to stabilize the metal/DNA adduct aid in the MALDI MS detection of the
dirhodium units bound to oligonucleotides. In contrast, no special sample treatment was
required to obtain high quality mass spectra of the corresponding cisplatin-modified

DNA strands.

Introduction
The compound cis-diamminedichloroplatinum(II), also known as cisplatin or cis-DDP,
is a highly successful drug for the treatment of a variety of deadly tumors including

bladder, cervical, ovarian and testicular cancers. While the mechanism of activity of

35

cisplatin is not entirely understood, there is consensus in the community that DNA

binding to cis-DDP is critical to its antitumor activity.26 Alteration of the DNA
structure by coordination to the drug, therefore, is one possible scenario for its cytotoxic
effects. Detailed footprinting studies of cis-DDP bound to DNA reveal a preference for
sequences containing two or more adjacent guanosine nucleosides. These studies imply
that 1,2 and 1,3 intrastrand adducts of the general formulae (NH3)2Pt{d(XpX)} or
(NH3)2Pt{d(GprG)} , where d(XpX) is either GpG or ApG (A = adenosine, G =
guanosine and N = any nucleoside), account for ~90% of all cisplatin/DNA binding
modes, while monofunctional adducts account for the remaining 10% of bound
platinum. Since the estimated number of bound molecules of cis-DDP is several orders
of magnitude lower than the value expected for all of the potentially strong d(GpG)

binding sites in the human cell, it is likely that the drug is acting on a subset of more

27

highly susceptible GpG or ApG sites in viva.
Structural information describing cis-Pt(NH3)2Cl2 adducts with oligonucleotide

duplex sequences is crucial to understanding the antitumor behavior of this important
drug. Recently, Lippard et al. reported an important result in this regard, namely the

crystal structure of the cisplatin dodecamer duplex
d(CCTCTGGTCTCC)-d(GGAGACCAGAGG) which was solved at a resolution of 2.6

A.lg The Pt(II) center forms cis interactions with the N7 positions of the guanines, which
themselves are almost perpendicular. This interrupts the base-stacking and base-pairing
patterns of the DNA helix, which is in accord with the observation that platinated
duplexes exhibit lower melting points (Tm) than unplatinated duplexes. The overall

structure of the duplex is a surprising combination of A and B type helices. Apart from

36

this general distortion, a major feature of this Pt/DNA adduct is the significant bend of
the duplex towards the major groove at the site of the GPtG crosslink. This bend and the
widened minor groove are reminiscent of the structural elements found in high-mobility

group (HMG) domain proteins and transcription binding protein (TBP)/DNA

complexes.28

In the Dunbar laboratories, they focused on a class of antitumor-active transition
metal compounds, the structures of which are often referred to as "lantern-type"
arrangements. These are dinuclear metal-metal bonded compounds of Re, Ru and Rh

that contain at least two bridging carboxylate ligands (Figure 5.1).29

 

 

 

 

R R R
*0 R A0111 GAO/(R
< I) ’ O O/ axial (ax)
1):ng fl... ’ CI "1, m.~‘\\\o Rf _ Fh.‘\\o If 0 _( pOSIUOD
L— (as v I 4| ‘01 6' ¢
Br Br I 07(0 0'Fh 0 I equatoria1(eq)
Br Br R 0Y0 LEO/T O Vposition
R

Figure 5.1 Dinuclear antitumor-active metal complexes with two possible
binding sites.

37

The dirhodium compounds Rh2(OzCR)4L2 (R = Me, Et, Pr; L = solvent) are the most

well-investigated members of this series. Numerous reports emerged in the 1970's that

supported the carcinostatic activity of these dirhodium compounds against Erlich ascites

. . . 30-32 . . . . .
and leukemia L1210 tumors m vzvo. ere crsplatln, d1rhod1um tetraacetate was

found to be a potent inhibitor of cellular DNA synthesis with little effect on RNA or

protein synthesis.30'32'33 In terms of DNA binding, several researchers reported facile
reactions with adenine but no guanine binding was observed. These reactivity
differences were rationalized on the assumption that the purine coordination modes
would be limited to the axial site, a situation that permits favorable hydrogen bonding

contacts between carboxylate O atoms and the NHz group of adenine, but which

. . . . . 34 . .
introduces sterrc repulsrons With the ketone 06 of guanine. F indings from our
laboratories have unequivocally established that adenine and also guanine bases form

thermodynamically stable substitution products with Rh2(OAc)4 that involve an

unprecedented equatorial bridging interaction for the purines.” Contrary to
conventional wisdom, the substitution pathway involves displacement of equatorial
carboxylate ligands rather than axial solvent molecules. Guanine and adenine (as well
as guanosine and adenosine) adducts of Rh2(OAc)4 have been characterized in solution
by NMR spectroscopy and by X-ray crystallography, and a variety of dimetal
compounds containing bridging and chelating 9-ethylguanine and 9-ethyladenine are
now in hand.34 Taken together with our recent studies of amino acid reactions of

dirhodium tetraacetate, these results offer new insights into the possible role of key

. . . . . . 36 .
cellular "ligands" 1n the metabolism of d1rhod1um antltumor complexes. The detection

of “Rh2(OAc)2” units bound to duplex DNA with{GpG} sequences by NMR

38

spectroscopy,37 and the X-ray structures of various M2 complexes containing bridging 9-
ethylguanine and adenine as well as chelating adenines have revealed new potential
binding modes of antitumor complexes with DNA.

This chapter presents matrix-assisted laser desorption/ionization mass
spectrometry (MALDI-MS) data that support the conclusion that the dirhodium bis-
acetate unit binds to small oligonucleotides containing a single GG or AA binding site.
The combination of enzymatic digestion24 experiments with MALDI analysis of the
products provides confirmation of the binding site. This valuable information has not
been easily available through NMR experiments and X-ray crystallographic studies to
date, and are not available from mass spectrometric methods if conventional approaches
and matrices are used. The results presented here suggest that the dirhodium/DNA
complexes are less robust than the platinum adducts but that, under appropriate
experimental conditions, they survive the transition to the gas phase. Most importantly, it
has been found that the dinuclear metal core with two coordinated acetate ligands

remains intact.

Experimental

Materials: Rh2(02CCH3)4(H20)2 and Pt(NH3)2(OH)2 were obtained from Pressure
Chemical Co. (Pittsburgh, PA) and used without further purification. Sodium acetate,
potassium chloride, and TRIS buffer (Tris[hydroxymethyl]aminomethane) were
purchased from Sigma Chemical Co. (St. Louis, MO) and used without fiirther

purification. Acetonitrile (HPLC grade) was obtained from VWR Scientific (Battavia,

39

IL) and the water was purified using a MilliQ purification system.
[Rh2(OzCCH3)2(CH3CN)6][BF4]2 was synthesized according to literature procedures.38
The DNA 12-mer oligonucleotide of sequence 5'-CCTCTGGTCTCC-3' (-GG- strand), 5'-
CCTTCAACTCTC-3' (-AA- strand) and the ll-mer 5'-TCTCTAATCTC-3' were
purchased from the Keck Oligonucleotide Synthesis Facility at Yale University. For
mass spectrometric analysis, 3-hydroxypicolinic acid (3-HPA), anthranilic acid, and
nicotinic acid were purchased from Sigma Chemical Co. (St. Louis, MO). Diammonium
citrate was purchased from J .T. Baker (Phillipsburg, NJ) and spermine was obtained from
F luka (Milwaukee, WI).

Resins: DEAE Cellulose resin was purchased from Sigma Chemical Co. (St. Louis,
MO). Source Q 15 anion exchange resin (HPLC) and Sephadex G-25 were purchased
from Amersham Pharmacia Biotech (Piscataway, NJ). Chelex-IOO chelating ion
exchange resin was obtained from BIO-RAD (Hercules, CA).

HPLC: Dirhodium modified DNA strands were purified via anion exchange
chromatography with a self-packed source Q 15 column (1.6 cm x 10 cm) on a Perkin-
Elmer HPLC instrument equipped with a LC 235 diode array detector. DNA products
were concentrated to dryness using a Centrivap concentrator (LABCONCO, Kansas City,
M0) at 75°C.

Procedures:

Synthesis of Rh2(02CCH 3)2-( T CT C TAA T CT C): A 2 pmol aliquot of the l l-mer

(TCTCTAATCTC) was dissolved in 400 1.1L of degassed, doubly distilled H20. To this

was added 65 uL of a 0.04 M degassed solution of [Rh2(02CCH3)2(CH3CN)6][BF4]2

4O

(1:1.3 ratio). The solution, which immediately turned orange, was maintained at 37°C for
72 h. At the end of the reaction time, the solution was a reddish-purple color.

HPLC Purification: The reacted ll-mer was purified by anion exchange HPLC using an
acetate buffer system monitored at 265 nm and 365 nm simultaneously. Eluent A: 0.2 M
NaO2CCH3, 20% CH3CN; Eluant B: 0.2 M NaO2CCH3, 1.2 M KCl, 20% CH3CN. The
product was purified by the following solvent program: 5 minute gradient 0-28% B, 30
minute gradient 28-33% B, 5 minute gradient 33-100% B, 5 minutes isocratic 100% B, 5
minute gradient 100-0% B, for a total protocol time of 50 min. at a flow rate of 4.0
mL/min.

Concentration and Desalting: Diethylaminoethyl (DEAE) cellulose resin was used to
concentrate the sample. The HPLC fraction containing the product was diluted four
times with 10 mM TRIS pH 7.0. This dilute solution was then loaded onto a 10 cm x 1.5
cm (I.D.) column. The metallated ll-mer product was eluted from the column with 1 mL
aliquots of a 10 mM TRIS pH 7.0, 1.0 M NaCl solution. The fractions were screened for
absorbance at 265 nm. Desalting of the concentrated sample was performed on a G-25
Size Exclusion Sephadex column (25 cm X 2.0 cm) (VWR Scientific, Battavia, IL). The
concentrated sample was loaded onto the column and eluted with doubly distilled H20,
and 1 mL fractions were collected and screened at 265 nm. The sample was then
concentrated to dryness in a Centrivap prior to mass spectrometric analyses. Overall
yield for the reddish colored ll-mer product was 65%.

Synthesis of Pt(NH 3) 2(0H)2~(CCT CT GGT C T CC): The metallated DNA oligomer was
synthesized by reacting the purified 12-mer (CCTCTGGTCTCC) with Pt(NH3)2(OH)2 in

an oligo:Pt ratio of 1:1.3, at 37°C. The metallated strand was purified by ion-exchange

41

HPLC with a 72%A/28%B to 68%A/32%B 40 minute gradient. In DNA purification
HPLC experiments, buffer A is 200mM NaCl and 10 mM NaOH, and buffer B is 1 M
NaCl and 10 mM NaOH. The isolated product was desalted following the same
procedures as described above and dried prior to mass spectrometric analyses.
Synthesis of Rh 2(02CCH 3) 2-(CCTTCAA CTCT C) and

Rh 2(02CCH 3) 2-( CC T C T 06 T C T C C): The metallated DNA oligomer was synthesized
by reacting with [Rh2(O2CCH3)2(CH3CN)6][BF4]2 with the purified 12-mer strands
CCTCTGGTCTCC or CCTTCAACTCTC in an oligoth2 ratio of 1:1.3 at 37°C. Once
the purple solution of the compound was added to the DNA sample, the color
immediately changed to a bright orange hue. Small aliquots of the reaction mixture were
periodically analyzed by HPLC to monitor the progress and extent of the reaction by
simultaneous detection of the DNA and [Rh2(O2CCH3)2(CH3CN)6]2+ chromophores at
260 nm and 365 nm respectively. The best yields for these reactions, obtained after 10
days, are 40% and 60% of Rh2-GG and Rh2-AA adducts, respectively. The metallated
strands were purified by ion-exchange HPLC with a 72%A/28%B to 68%A/32%B 40
minute gradient to give a light green solution. The isolated product was desalted
following procedures described above, and concentrated to dryness for mass
spectrometric analyses.

Digestion of Pt(NH 3) 2(0H) 2-(CC T C T GG T CT CC): A 200 pmol sample of
5'-CCTCTGG{Pt(NH3)2(OH)22+}TCTCC-3' was digested with 2 x 10'3 units of snake
venom phosphodiesterase (VPD) (Sigma Chemical Co., St. Louis, MO) in 110 mM TRIS
buffer (pH = 9.4) dissolved in water. The digestion, which takes place from the 3’ end,

was performed at 37°C for 15 minutes.

42

Digestion of Rh 2(02C CH 3)2°( C C T CT GGT C T CC): A 100 pmol sample of 5'-
CCTCTGG{Rh2(O2CCH3)22+}TCTCC-3' was digested with a mixture of ~2 x 10'3 units
of VPD in 110 mM TRIS buffer (pH = 9.4) at 37°C for 15 minutes.

Digestion of Rh 2(02CCH 3)2-( T C T C TAA T CT C): A 60 pmol sample of 5'-
TCTCTAA{Rh2(O2CCH3)22+}TCTC-3' was digested with a mixture of ~2 x 10'3 units of
VPD in 110 mM TRIS buffer (pH = 9.4) at 37°C for 15 minutes. Another digestion of 60
pmol of the same complex was performed using a mixture of 2 x 10'3 units of VPD and 3
x 10'3 units of calf spleen phosphodiesterase (SPD) (Boehringer Mannheim, Indianapolis,
IN) in 110 mM Tris buffer (pH=9.4) at 37°C for 15 minutes.

Mass Spectrometry of Dirhodium-DNA adducts. Samples were prepared for mass
spectrometric experiments as follows: for the cisplatin-DNA adduct, the digestion of
cisplatin-DNA adduct, and one experiment with a Rh2-DNA adduct, the matrix solution
that was used is saturated 3-hydroxypicolinic acid (3-HPA) in 1:1 acetonitrile/water
containing 25 mM (NH4)2HCit. For experiments with Rh2 adducts of 5'-
TCTCTAATCTC-3', 5'-CCTCTGGTCTCC-3', and 5'-CCTTCAACTCTC-3', and the
digestion products of the Rh2-DNA adducts, the matrix was saturated 80% anthranilic
acid/20% nicotinic acid (hereafter referred to as 80/20) in 1:1 acetonitrile/water
containing 12.5 mM spermine. Linear MALDI mass spectra were recorded on a
PerSeptive Biosystems (Framingham, MA) Voyager Elite delayed extraction, time-of-
flight reflectron mass spectrometer equipped with a nitrogen laser (337 nm, 3 ns pulse).
For all samples, the accelerating voltage was 23 kV, the delay time was 50 ns, the grid
voltage was set to 93.0% of the accelerating voltage, and the guide wire was set to 0.25%

of the accelerating voltage. The mass spectrometer was calibrated using unmetallated

43

5'-CCTCTGGTCTCC-3' as a mass standard, with a resulting mass accuracy of < 21:1 Da
for peaks below m/z 4000. Typically, 64-128 laser shots were averaged for each
spectrum. The sample stage used was a gold sample plate. All spectra were collected in

the negative ion mode.

Results And Discussion
If a heavy metal atom is bound to an oligonucleotide, its presence can be detected using
mass spectrometry (MS). An illustration of the success of mass spectrometry in this area
is the characterization of platinum-oligonucleotide complexes using desorption/ionization
methods that generate gas phase ions for MS analysis directly from condensed phase
targets. Using a combination of electrospray (ESI) MS and matrix-assisted laser
desorption/ionization (MALDI) MS, coupled with enzymatic degradation, kinetic rate
constants for platination of small oligonucleotide chains have been determined.39 Such
approaches work quite well since the enzymes used to degrade the DNA do not recognize
the bound Pt, and therefore the digestion stops when it reaches the platinated site.40 A
number of excellent reviews are available that describe the techniques involved and the
structures of Pt-DNA complexes.26

In the early 1990's, Costello and Lippard reported the characterization of Pt(II)
oligonucleotide complexes by fast atom bombardment (FAB) MS techniques.41 Later it
was demonstrated that the MALDI MS spectrum of a single stranded DNA adduct of
Pt(NH3)2(OH)2 survives intact in the gas phase, with only the loss of the two hydroxyl
groups being observed.42 In order obtain some experience in this area of application of

mass spectrometry, and as a point of reference with which to compare data on dirhodium

44

complexes, we performed a MALDI MS analysis of a 12-mer adduct of Pt(NH3)2(OH)2.
For these experiments, a typical matrix combination of 3-hydroxy-picolinic acid and
diammonium citrate was used, and the negative ion spectrum was obtained (Figure 5.2a).
The intense, high mass peak observed at m/z 3770 represents the anionic form of the
complex [M+Pt(NH3)2-3H]' where M is the neutral form of 5'-CCTCTGGTCTCC-3'.
Note, the formation of this singly charged anionic species requires the loss of three
protons from the neutral species; the resulting -3 species, coupled to a +2 metal, exhibits
an overall -1 charge. This charged state can be generated and detected in the MALDI
experiment. It is important to note that more than 97% of the oligonucleotide is observed
in the platinated form, as there is only a very small peak representing the free strand,
which appears in its anionic form as the [M-H]’ ion at m/z 3546.

After establishing that cisplatin was bound to the oligonucleotide, the next step
was to generate specific structural information. The binding site of cisplatin was
determined by performing an enzymatic digestion from the 3' end of the metallated strand
with the use of snake venom phosphodiesterase (VPD). When the digestion mixture was
analyzed directly using MALDI MS in the negative ion mode, the spectrum shown in
Figure 5.2b was obtained. As before, the highest m/z peak occurred at 3770, which is the
intact metallated strand. In this case, however, the intensity is very low since most of the
original complex was consumed in the digestion process. The spectrum contains a
number of other low intensity peaks due to consecutive removal of mononucleotides,
which occurs until the enzyme encounters the Pt unit on the DNA strand. At this point

the enzyme is inhibited and the digestion stops. The peak at m/z 2294 represents the

45

digestion product [5'-CCTCTGG {Pt(NH3)22+}-3H]‘ which is indicates that cisplatin is

bound to the GG portion of this DNA sequence.

 

a) [M+PI(NH3 )2-3H1

succrcrqprcrcc-y
Pt(NH3)2(OH)2

 

AWN/“AMA
) [M+PI(NH3 )2-TCTCC-3HT

Relative Intensity

I I I I I I
1500 2000 2500 3000 3500 4000
m/z

 

 

 

 

 

Figure 5.2 (a) MALDI mass spectrum of cisplatin bound to single-stranded
DNA. (b) Digestion of the cisplatin-DNA complex using 3' snake venom
phosphodiesterase.

46

Unfortunately the detection of dirhodium-DNA complexes by MALDI MS proved to
be much less straightforward than the analysis of the corresponding platinum/DNA
adducts. A systematic investigation of numerous matrix and additive combinations was
undertaken in an attempt to detect an intact dirhodium DNA adduct in the gas phase.
Figure 5.3a depicts the negative ion MALDI mass spectrum of
Rh2(O2CCH3)2-(TCTCTAATCTC) using the common matrix combination of 3-HPA
with DAHC. The dirhodium-DNA interactions are being destroyed in the course of the
MALDI experiment, as evidenced by our observation of a free DNA peak at m/z 3242
and a complete absence of peaks representing the intact complex. When the matrix was
changed to 80% anthranilic acid/20% nicotinic acid in combination with spermine as the
matrix additive, however, a peak at m/z 3443 was observed in the resulting spectrum
(Figure 5.3b). This peak represents the naked Rh24+ core attached to the DNA strand,
viz., [M+Rh2-5H]'. Apparently these conditions allow for the fundamental
dimetal/oligonucleotide interactions to be retained. Of even greater importance is the
observation of higher mass peaks at m/z 3502 and m/z 3561 that represent the mono- and
di-acetate forms, [M+Rh2(O2CCH3)-4H]' and [M+Rh2(02CCH3)2-3H]’, respectively. The
matrix combination of 80/20 with spermine (a MALDI additive developed in our
laboratories) stabilizes the dirhodium-DNA complex that had been synthesized. We have
found that the low m/z peak in the series with only the Rh24+ unit attached is always the
most intense peak, and the higher m/z acetate-containing components are less abundant.
This cluster of peaks is the signature for the presence of the dirhodium moiety, and they
can be easily observed in the spectra. As an additive for oligonucleotides, spermine has

been shown to greatly reduce alkali cation adducts and enhance sensitivity.43 Spermine,

47

in combination with 80/20, gives a Na+ and K+ free spectrum of the intact metal-DNA
complex. That is, with spermine present, negatively charged phosphates are bound to
protons, rather than to alkali ions. A minor peak is observed at m/z 3242 which
represents [M-H]', an indication that the dirhodium-DNA interactions are disrupted to a
very small extent. Note that the [M-H]' peak does not possess a set of higher mass
“satellite peaks" (as does the m/z 3443 peak), since the m/z 3242 peak does not contain

dirhodium.

48

 

a) [M-H]' 3-HPA
(NH4)2HCit

 

 

Relative Intensity

I
b) _ 80/20
- [M+Rh2-5H] spermine

[M+Rh2(O2CCH3)-4H]'
[M+Rh2(02CCH3)2-3H]-

F[M+Rh2(02CCH3)2+H20-3H]'

/

 

 

 
   

[M-HI

 

 

 

 

| l I I I I I
3000 3200 3400 3600 3800 4000 4200

m/z

Figure 5.3 A portion of the (a) MALDI mass spectrum of
Rh2(O2CCH3)2-(TCTCTAATCTC) using 3-hydroxypicolinic acid with
diammonium citrate and (b) the MALDI mass spectrum of the same compound
using 80% anthranilic acid/20% nicotinic acid with spermine.

49

Experimental results such as those shown in Figure 5.3 lead to the conclusion
that, for these studies, the 80/20 matrix with spermine will allow for dirhodium-DNA
complexes to be detected using MALDI MS. When this matrix is used to analyze
Rh2(O2CCH3)2(CCTCTGGTCTCC), the complex is successfully detected. The high m/z
portion of the MALDI mass spectrum shown in Figure 5.4 shows the intact complex, and
intense peaks representing acetate-containing forms as well. We have also detected the
dirhodium complex of the same strand, with AA replacing the GG segment in the center,
using the same matrix conditions. While the results in Figure 5.4 support the conclusion
that dirhodium binds to a GG-containing strand, it does not prove that GG is the binding

site.

50

 

80/20
[M+Rh2-5HT spermlne

- [M+Rh2(O2CCH3)-4H]'

[M+Rh2(02CCH3)2-3H]'

 

/[M+Rh2(O2CCH3)2+H2O-3H]'

/

 

Relative Intensity

 

 

 

 

 

 

IM-Hl' I II

 

 

 

 

 

 

I I I l I I
3400 3600 3800 4000 4200 4400

m/z

Figure 5.4 A portion of the MALDI mass spectrum of
Rh2(O2CCH3)2-(CCTCTGGTCTCC) using 80% anthranilic acid/20%
nicotinic acid with spermine.

51

In order to locate the position of the dirhodium complex on the oligonucleotide
strand, the sample was subjected to enzymatic digestion followed by MALDI MS
analysis of the resulting mixture. The metallated -GG- strand
(M=5'-CCTCTGGTCTCC-3') was digested using snake venom phosphodiesterase, an
exonuclease that digests from the 3' end. The spectrum in Figure 5.5 shows the result of
the digestion. The peak at m/z 3747 represents the intact Rh2-DNA complex. The peak
at m/z 2272 represents [M+Rh2-TCTCC-5H]'; the digestion stops when the metal is
encountered. Notice the signature peaks for the acetate ligands at m/z 2332 and m/z 2392
which represent [M+Rh2(O2CCH3)-TCTCC-4H]' and [M+Rh2(O2CCH3)2-TCTCC-3H]',
respectively. This is in contradiction to previous claims that dirhodium interactions with
guanine bases do not occur. 303233“ It is unusual that much of the [M+Rh2-5H]' signal
remains. It is most likely due to some inhibition of the enzyme in this particular system.
One possible scenario is that the Rh2-DNA complex exists in multiple forms of
complexation with acetates in the solution. The Rh2(O2CCH3)2-DNA complex can be
digested without metal interactions with the enzyme. However, the Rh2-DNA complex
may lead to metal-enzyme interactions that inhibit digestion. This may explain why, in
Figure 5.5, the digestion product ion [CCTCTGG-Rh2-5Hj' shows the signature for the
dirhodium satellite peaks due to one and two acetates, while the intact complex does not,

suggesting that only the acetate-“protected” Rh2-DNA complex is efficiently digested.

52

Relative Intensity

 

L

 

M= 5’-CCTCTGGTCTCC-3' [M+Rh2-5HI'

[M+Rh2-TCTCC-5H]' 3747

2272
[M+Rh2(O2CCH3)-TCTCC-4H]'

[M+Rh2(02CCH3)2-TCTCC-3H]-
[M+Rh2(02CCH3)-4H]d

 

-1“, -c -T, -c -c

 

 

¢

 

 

 

 

 

 

 

1%00 2500 2§00 3000 300
m/z

Figure 5.5 The MALDI mass spectrum of the digestion products of
Rh2(O2CCH3)2(CCTCTGGTCTCC) using 3' snake venom
phosphodiesterase.

53

 

These experiments are good examples of cases where mass spectrometry alone
cannot be used to provide structure determination. The MALDI instrument used here has
Post-Source Decay (PSD) capabilities for analyzing fragment ions of a selected precursor
ion, but for metal-oligonucleotide complexes, PSD cannot be successfully used. Signal
intensities are frequently too low to perform PSD and the loss of metal is usually the most
energetically favorable fragmentation process in the gas phase.

Figure 5.6a depicts the complete MALDI spectrum of the dirhodium-DNA
complex, a portion of which was shown in Figure 5.3b. In addition to the peak
representing the intact dirhodium-DNA complex, there are lower m/z peaks, notably a
prominent peak at m/z 2056. These peaks may represent fragment ions, formed in the
mass spectrometry experiment, or may represent fragments of the oligonucleotide,
formed in the solution during the MALDI target formation process. The m/z 2056 peak
represents the [M-H]' ion for TCTCTAA, a fragment of the larger analyte. For this
experiment, it is important to define the spectrum, Figure 5.63, before digestion, so the
new peaks formed in the digestion products can be clearly identified. Figure 5.6a sets
this baseline. Figure 5.6a is shown as a reference for the spectrum in Figure 5.6b that
represents the mixture obtained after digestion. The enzyme used was VPD, which
digests from the 3' end. The results of the digestion show a significant loss of the metal
complex, which evidently occurs when the complex interacts with VPD, since the major
series of digestion products are from the free DNA. The [M-H]' peak representing the
free strand is observed at m/z 3242. There are minor peaks at m/z 3154 and m/z 2850

which represent digestion products from the intact dirhodium-DNA complex, but the

54

series is of low intensity and yields no information on the location of the dirhodium
binding site. Thus, while the matrix and additive used here allow the metal core to
remain bound to the oligonucleotide, the interaction with the enzyme appears to displace
the metals at an early point in the digestion, unlike what was observed for GG binding in
Figure 5 .5. This particular approach is, therefore, ineffective for determining the site of
metallation.

After our initial experiments with VPD digestion, a two-enzyme system was
investigated, with the pH optimized to activate only one of the enzymes. The enzymes
VPD and calf spleen phosphodiesterase (which is a 5' enzyme) SPD, were selected and
used at pH = 9.4 which is the optimum pH for VPD. In this manner it was possible to
digest the sequence from the 3' end until the Rh2(O2CCH3)22+ complex is encountered by
the enzyme. As the spectrum in Figure 5.6c shows, the digestion products contain the
intact dirhodium complex. Furthermore, the results indicate that the dirhodium complex
is attached to the AA portion of the DNA strand. Additional confirmation of the
dirhodium position is illustrated by the characteristic pattern of peaks representing acetate
adducts at masses higher than m/z 2257 (indicated by a bracket in the figure). From close
inspection it can be discerned that the peaks in Figure 5.60 all contain the dirhodium core,
as indicated by satellite peaks at higher m/z values due to the presence of one and two
acetate ligands. Clearly, if [M-H]’ is present as in Figure 5.6b, digestion products of
[M-H]' are observed. When [M+Rh2-5H]' is observed, only digestion products of this
metal-DNA complex are observed as in Figure 5.6c. All digestion products in Figure 5.60
contain the metal. It is believed that the digestion with VPD alone results in the removal

of Rh2 from the DNA strand by the enzyme. It is not obvious to us how the addition of

55

 

5’-TCTCTAATCTC—3’
[M+Rh2-5H]'

‘ Rh2(02CCH3)2

b)

 

 

E

1

[M+Rh2-5H]'
[M-HI'

/

<"A “A "T ¢'C 4i +29—
IIIIWIIIW in

C) [M+Rh2-5H]-
[M+Rh2-TCTC-5H]'

L

 

 

 

Relative Intensity

 

 

 

 

-T
4—(0 2 -T (c

IWmI/WIIIIW ' ' I M

I I I r I
1500 2000 2500 3000 3500
m/z

 

 

 

 

 

 

Figure 5.6 The MALDI mass spectrum of (a)
Rh2(02CCH3)2(TCTCTAATCTC), (b) the digestion products of the same
dirhodium-DNA complex using 3' snake venom phosphodiesterase, and (c) the
digestion of the dirhodium-DNA complex using the double enzyme digestion of
3' snake venom phosphodiesterase and 5' calf spleen phosphodiesterase (brackets
identify the cluster of peaks indicative of dirhodium acetate bound to DNA).

56

SPD aids in the digestion of the intact dirhodium-DNA complex, but such results could
not be obtained by using VPD alone. Thus, we assume that SPD, in some manner,
stabilizes the complex while allowing VPD to digest the DNA. We are not necessarily
suggesting the use of two enzymes as a general method; we simply report that this
allowed for the digestion to be completed for this example. The dual enzyme system was

not useful in digesting the GG system (Figure 5 .5).

Conclusions

This study provides the first structural evidence for the binding of dirhodium
complexes to DNA. Furthermore it has been shown that dirhodium bis-acetate binds to
GG as well as the AA containing oligonucleotides in direct contradiction to earlier
claims that guanine adducts of dirhodium complexes are not stable. Although the
dirhodium-DNA interactions are weaker than corresponding cisplatin-DNA adducts,
they are sufficiently strong to be observed in the gas phase when handled in a matrix
that contains spermine. The ultimate importance of these findings is its extension to

double-stranded DNA oligonucleotides, work that is in progress.44

Acknowledgments

Mass spectral data were acquired at the MSU Mass Spectrometry Facility, which
is partially supported by grant # RR-00480 from the Biotechnology Research
Technology Program of the National Center for Research Resources of the NIH. A
grant from the Cancer Center at Michigan State University to KRD is gratefully

acknowledged.

57

10.

11.

12.

13.

14.

15.

16.

17.

18.

REFERENCES

Barber, M.; Bordoli, R.S.; Sedgwick, R.D.; Tyler, A.N. J. Chem. Soc., Chem.
Commun. 1981, 325-327.

Hillenkamp, F.; Karas, M.; Beavis, R.C.; Chait, B.T. Anal. Chem. 1991, 63,
A1193-A1202.

Székely, G.; Allison, J. J. Am. Soc. Mass Spectrom. 1997, 8, 337.

De Pauw, E.; Agnello, A.; Derwa, F. Mass Spectrom. Rev. 1991, 10, 283-301.
Ligon, Jr., W.V.; Dorn, S.B. Fresnius Z. Anal. Chem. 1986, 325, 625-626.
Orlando, R. Anal. Chem. 1992, 64, 332-334.

Musselman, B.D.; Allison, J.; Watson, J.T. Anal. Chem. 1985, 57, 2425-2427.
Dell, A.; Ballou, C.E. Biomed. Mass Spectrom. 1983, 10, 50-56.

Belu, A.M.; DeSimone, J .M.; Linton, R.W.; Lange, G.W.; Friedman, RM. J.
Am. Soc. Mass Spectrom. 1996, 7, 11-24.

Xiong, Y.; Owens, K.G.; Danis, P.O.; Karr, D.E. Proceedings ofthe 42""ASMS
Conference on Mass Spectrometry, 1994; p. 980.

Bahr, U.; Deppe, A.; Karas, M.; Hillenkamp, F. Anal. Chem. 1992, 64, 2866-
2869.

Llenes, C.F.; O’Malley, R.M. Rapid Commun. Mass Spectrom. 1992, 6, 564-
570.

Kahr, M.S.; Wilkins, C.L. J. Am. Soc. Mass Spectrom. 1993, 4, 453-460.
Billeci, T .M.; Stults, J.T. Anal. Chem. 1993, 65, 1709-1716.

Reynolds, J .D.; Cook, K.D.; Burn, J.L.E.; Woods, C. J. Am. Soc. Mass Spectrom.
1992, 3,113-121.

Mechref, Y.; Novotny, M.V. Anal. Chem. 1998, 70, 455-463.

Takayama, M. J. Am. Soc. Mass Spectrom. 1995, 6, 114.

(a) Fleming, J.S.; Mann, K.L.; Carraz, C.-A.; Psillakis, J .C.; McCleverty, J.A.;
Ward, M.D. Angew. Chem. Int. Ed. 1998, 3 7, 1279-1281. (b) Whiteford, J.A.;

58

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

Rachlin, E.M.; Stang, P.J. Angew. Chem. Int. Ed. 1996, 35, 2524-2529. (c)
Didier, P.; Jacquet, L.; Mesmaeker, A.K.-D.; Hueber, R.; Dorsselaer, A.V. Inorg.
Chem. 1992, 31, 4803-4809.

Arman, R.S.; Carr, S.A. Anal. Chem. 1996, 68, 3413-3421.

Liao, P.-C.; Leykam, J.; Andrews, P.C.; Gage, D.A.; Allison, J. Anal. Biochem.
1994, 219, 9-20.

Clouse, 8., North Carolina State University, personal communication, 1999.

(a) Chait, B.T.; Wang, R.; Beavis, R.C.; Kent, S.B.H. Science 1993, 262, 89-92.
(b) Kratzer, R.; Eckerskom, C.; Karas, M.; Lottspeich, F. Electrophoresis 1998,
19, 1910-1919.

Henry, C. Anal. Chem. 1997, 243A-246A.
Limbach, P.A. Mass Spectrom. Rev. 1996, 297-336.
Ross, R; Hall, L.; Smimov, 1.; Haff, L. Nature Biotech. 1998, 16, 1347-1351.

(a) Sherman, S. E.; Lippard S. J. Chem. Rev. 1987, 87, 1153-1181. (b) Reedijk,
J. Pure Appl. Chem. 1987, 59, 181-192. (c) Eastman, A. Pharmac. T her. 1987,
34, 155-166. ((1) Sherman, S. E.; Gibson, D.; Wang, A. H.-J.; Lippard, S. J. J.
Am. Chem. Soc. 1988, 110, 7368-7381. (e) Umapathy, P. Coord. Chem. Rev.
1989, 95, 129-181. (1) Lippert B. Chapter 1 in Prog. Inorg. Chem. Lippard, S. J .,
Ed.; Wiley, New York, 1989, 37, 1-98 and references therein. (g) Farrell N. in
Transition Metal Complexes as Drugs and Chemotherapeutic Agents. Eds., Ugo,
R.; James, B. R. Kluwer: Dordrecht, The Netherlands, 1989; Chapter 2 and
references therein. (h) Kozelka, J.; Chottard, J.- C. Biophys. Chem. 1990, 35,
165-178. (i) Sundquist, W. 1.; Lippard, S. J. Coord. Chem. Rev. 1990, 100, 293-
322. (j) Corda, Y.; Job, C.; Anin, M.-F.; Leng, M.; Job, D. Biochemistry, 1991,
30, 222-230. (k) Keck, M. V.; Lippard, S. J. J. Am. Chem. Soc. 1992, 114, 3386-
3390. (l) McAuliffe, C. A.; Sharma, H. L. and Tinker, N. D. in Chemistry of the
Platinum Group Metals-Recent Developments, Hartley, F. R., Ed., 1992, Volume
11, 546-593. (m) Takahara, P. M.; Frederick, C. A.; Lippard, S. J. J. Am. Chem.
Soc., 1996, 118, 12309. (n) Ohndorf, U.-M.; Whitehead, J. P.; Raju, N. L.;
Lippard, S. J. Biochemistry, 1997, 36, 14807-14815. (0) Reedijk, J. Chem. Rev.,
1999, 99, 2499.

Iwamoto, M.; Mukundan, S. Jr.; Marzilli, L. G. J. Am. Chem. Soc. 1994, 116,
6238.

(a) Ohndorf, U.-M.; Rould, M.A.; He, Q.; Pabo, C.O.; Lippard, S.J. Nature, 1999,

399, 708. (b) Huth, J. R.; Bewley, C. A.; Nissen, M. S.; Evans, J. N.; Reeves, R.;
Gronenbom, A. M.; Clore, G. M. Nat. Struct. Biol., 1997, 4, 657. (c) Werner, M.

59

 

29.

30.

31.

32.
33.

34.

35.

36.
37.
38.
39.

40.

H.; Huth, J. R.; Gronenbom, A. M.; Clore, G. M. Cell, 1995, 81, 705. ((1) Kim, J.
L.; Nikolov, D. B.; Burley, S. K. Nature, 1993, 365, 520. (e) Kim, Y.; Geiger, J.
H.; Hahn, S.; Sigler, P. B. Nature, 1993, 365, 512.

(a) F elthouse, T. R. Prog. Inorg. Chem. 1982, 29, 73. (b) Cotton, F. A.; Walton,
R. A. Multiple Bonds Between Metal Atoms; Wiley, New York, 1982; Chapter 7,
73. (0) Kitchens, J.; Bear J. L. J. Inorg. nucl. Chem, 1969, 31, 2415.

Bear, J. L.; Gray, H. B.; Rainen, L.; Chang, I. M.; Howard, R.; Serio, G.;
Kimball, A. P. Cancer Chemother. Rep., 1975, 59,611.

(a) Bear, J. L.; Howard, R. A.; Dennis, A. M. Curr. Chemother., Proc. Int. Cong.
Chemother, 10th, 1977, 1978, 1321-1323. (b) Rao, P. N.; Smith, M. L.; Pathak,
8.; Howard, R. A.; Bear, I. L. J. Natl. Cancer Inst, 1980, 64, 905.

Howard, R. A.; Spring, T. G.; Bear, J. L. Cancer Res. 1976, 36, 4402.

(a) Yu, B. S,; Choo, S. Y.; Chang, I. M. J. Pharm. Soc. of Korea, 1975, 19, 215.
(b) Farrell, N. J. Inorg. Biochem. 1981, 14, 261. (c) Waysbort, D.; Tarien, E.;
Eichhom, G. L. Inorg. Chem. 1993, 32, 4774 and references therein.

Rubin, J. R.; Haromy, T. P.; Sundaralingam, M. Acta. Crystallogr. 1991, C4 7,
1712.

(a) Dunbar, K. R.; Matonic, J. H.; Saharan, V. P.; Crawford, C. A.; Christou, G.
J. Inorg. Biochem., 1993, 51, 401. (b) Dunbar, K. R.; Matonic, J. H.; Crawford,
C.; Christou, G. J. Am. Chem. Soc., 1994, 116, 2201. (0) Day, E. F.; Crawford,
C. A.; Folting, K.; Dunbar, K. R.; Christou, G. J. Am. Chem. Soc., 1994, 116,
9339. ((1) Crawford, C. A.; Day, E. F.; Saharan, V. P.; F olting, K.; Huffman, J.
C.; Dunbar, K. R.; Christou, G. Chem. Commun., 1996, 1113. (e) Dunbar, K. R.;
Catalan, K. V. McGraw Hill Encyclopedia of Science &Technology 1997,
McGraw-Hill, 1996, 40-43. (1) Catalan, K. C.; Dunbar, K. R.; Mindiola, D. J.
Inorg. Chem, 1997, 36, 2458. (g) Prater, M. E.; Mindiola, D. J.; Ouyang, X.;
Dunbar, K. R. Inorg. Chem. Commun., 1999, 1, 475. (h) Catalan, K. V.; Hess, J.
S.; Maloney, M. M.; Mindiola, D. J .; Ward, D. L.; Dunbar, K. R. Inorg. Chem,
1999, 38, 3904.

Jennifer Simone Hess, M. S. thesis, 1998, Michigan State University.
Elizabeth U. Lozada Carrasco, M. S. thesis, 1998, Michigan State University.

Pimblett, G.; Garner, C. D.; Clegg, W. J. Chem. Soc. Dalton Trans. 1986, 1257.

(a) Gonnet, F .; Kocher, F .; Blais, J .C.; Bolbach, G.; Tabet, J .C.; Chottard, J .C. J.
Mass Spectrom. 1996, 31, 802. (b) Murdoch, P. del S.; Guo, Z.; Parkinson, J .A.;
Sadler, R]. J. Biol. Inorg. Chem. 1999, 4, 32.

(a) Cohen, G.L.; Ledner, J .A.; Bauer, W.R.; Ushay, H.M.; Caravana, C.; Lippard,
SJ. .1. Am. Chem. Soc. 1980, 102, 2487. (b) Tullius, T.D.; Lippard, S]. J. Am.

60

 

41.

42.

43.

Chem. Soc. 1981, 103, 4620. (c) Chottard, J.C.; Girault, J.P.; Chottard, G.;
Mansuy, J. J. Am. Chem. Soc. 1980, 102, 5565. (d) Inagaki, K.; Kasuya, K.;
Kidanai, Y. Chem. Lett. 1983, 1345. (e) Inagani, K.; Kidani, Y. Inorg. Chim. Acta
1985, 106, 187.

Costello, C.E.; Comess, K.M.; Plaziak, A.S.; Bancroft, D.P.; Lippard, S. J. Int. J.
Mass Spectrom. Ion Processes, 1992, 122, 255.

Costello, C.E.; Nordhoff, E.; Hillenkamp, F. Int. J. Mass Spectrom. Ion
Processes, 1994, I32, 239.

Asara, J.M.; Allison, J. Anal. Chem, 1999, 71, 2866.

Lozada, E. U.; Bishop, K.; Dunbar, K. R., manuscript in preparation.

61

 

APPENDIX A

62

Inorg. Chem. 1998, 37. 1833-1840 1833

Analysis of Transition-Metal Compounds Containing Tetrathiafulvalene Phosphine Ligands
by Fast Atom Bombardment Mass Spectrometry: Limitations and the Development of
Matrix Additives for the Desorption of Multiply Charged Complexes

John M. Asara, Calvin E. Uzelmeier, Kim R. Dunbar,‘ and John Allison‘

Department of Chemistry. Michigan State University, East Lansing. Michigan 48824
Received September 3. 1997

A series of new complexes incorporating the functionalized teuathiafulvalene ('ITF) ligands ortho-(Q-Ig)1(P(C¢H;)2)r
Tl'F (o-PZ) and (P(C¢115)2)1'1TF (P4) have been preparedand studiedby fast atom bombardmentmass specuonreny
(FABMS). The mononuclear o-P2 complexes [M(o-P2)2](BF.)2 (M = Fe, Pd. Pt) and [Co(o-P2)2(NCCH;)2](BF4)2
were synthesized from reactions of the free ligand with the fully solvated BFs‘ salts [M(NCCH3).](BF4)2 (M =
Fe, Co. n = 6: M = Pd. Pt. n = 4). The dinuclear P4 complex, [P12(P4)(NCCH3).](BF4)4. was produced by
reacting the free ligand with PtCl2(NC7H5)2 followed by abstraction of the chlorides with AgBFs in acetoninile.
Reaction of [Pd(NCCI—13)](BF4)2 with 1 equiv of P4 produces the polynuclear compound formulated as
[Pd(P4)],.(BF.)2. which was characterized by infrared. 'H. and 31PM!) NMR spectroscopic!» and elemental analysis.
The use of FABMS in this study was undertaken in order to elucidate the chemical options of multiply charged
cations in the desorption process from the liquid matrix to the gas phase. The use of additives to the FAB matrix
(m-niuobenzyl alcohol) was demonstrated for the P4 complexes which do not give spectra in this medium due to
a high positive net charge. The addition of triflic acid (HOTf or CF3803H) to the FAB matrix/analyte solution
was shown to assist in the MS analysis of cationic complexes with up to six charges and a mass range up to mlz
4000. When HOTf is added. bound 0T1“ anions are formed and are attached to the cationic complex. lowering

the net charge to +1. The method of using matrix additives as opposed to chemically synthesizing the 011‘
compounds is a convenient in situ method which produces species that are capable of being analyzed by FABMS.

Introduction

The design of molecule-based materials with tunable con-
ducting. optical, and magnetic materials is the focus of much
research activity in synthetic organic and inorganic chemistry.L2
In one approach to preparing materials with paramagnetic metal
centers in the same structural framework as open-shell n-organic
radical cations. researchers are investigating salts of teu'athi-
afulvalene (I'm-based radical cations with hi gh-moment metal
cluster anions.3 Organic donor molecules such as 'ITF, bis-

 

‘To whom correspondence should be addressed. E-mail: dunbaro
cemvaxcemmsu. edu: allisonGcemvax. cem. msu edu Fax: (517) 3531793.
(I) (a) Fourmigue. M.; Uselmeier. C. 5.: Boubekeur. K.: Bartley.S. L.:
Dunbar. K. R. I. Orgonomer. Chem 1997. 529. 343 (b) Uzelmeier,
C. E: Fourmigue. M.; Grandinetti, 6.; Dunbar. K. R. manuscript in
preparation. (c) Mann'quez. J. M.; Yee. G. T.; Mcbean. S.: Epstein,
A. J.; Miller. J. S. Science 1991. 252. 1415. (d) Pei. Y.; Kahn. 0.:
Nahum. K.; Codjovi. E.; Mathoniere. C.; Slenen. .I. J. Am. Chem.
Soc. 1991. 113. 6558. (e) Cornelissen, .I. P.; LeLoux. R.; Jansen. J.;
Hannoor, J. 6.: Reedin 1.: Horn, 5;: Spek. A. L.; Pomarede. 8.:
begins. 1.; Reefman, D. J. Chem. Soc., Dalton Trans. 1992. 2911. (f)
Senoni. 8.: Dean. 0.; Campagna. 8.: Juris. A.; Ciano. M.; Balzani.
V. Angew. Chem. Int. Ed. Engl. 1992. 31. 1493. (g) Kollmar. C.:
Kahn. 0. Acc. Chem. Res. 1993. 26. 259. (h) Miller. J.; Epstein. A.
Angew. Chem. Int. Ed. Engl. 1994. 33. 385. (i) Kuhn. 0. Molecular
Magnetism: VCH: New York, 1993. (j) Pomarede. 8.: Garreau. 8.:
Malfant. 1.; Valade. L.: Cassoux. P.; Legros. J.; Audouard. A.;
Brossard. L.: Ulmet. 1.: Doublet. M.; Canadell. E. Inorg. Chem 1994.
33. 3401. (k) Stumpf. H. 0.; Pei. Y.; Michaul. C.: Kahn. 0.; Renard.
J.; Ouahab. L. Chem. Mater. 1994. 6. 257. (1) Miller, J.; Epstein. A.
Angew. Chem. Int. Ed. Engl. 1994. 33. 385.

(2) (a) Saito, G. S.. Kagoshima. 8.. Eds. The Physics and Chemistry of
Organic Superconductors: Springer: Berlin. 1990. (b) Wudl. F. Acc.
Chem Res. 1984. I7. 227. (c) Williams. J. M.; Wang. H.; Emge. T.
J.; Geiser. U.: Beno. M. A.; Leung. P. C. W.: Carlson. 1C 0.: Thom.
R. J.; Schultz, A. J.; Whangbo. H. H. In Progress in Inorganic
Chemistry; Lippard. S. 1., Ed; Wiley: New York. 1987; Vol. 35. pp
51-218.

(ethylenedithio) tetrathiafulvalene (BEDT-‘ITB. and tetra-
methyltetraselenafulvalene (TMTSF) form salts with polychal-
cogenide. halide. and mixed chalcogenidelhalide cluster anions
of Re with remarkable variations in properties. differences that
have been attributed to changes in the size. shape. and redox
properties of the organic donor and inorganic acceptors.

An entirely different philosophy for the construction of hybrid
inorganic/organic arrays involves the direct coordination of
metals to organic radicals through a heteroatom.‘ In this vein,
it has been shown that the use of Open-shell oxygen donor
ligands such as semiquinones,’ nitroxides. and nitronyl niuox-
ides‘” contribute interesting variations to the electronic and
magnetic properties of a metal-containing chain. The groups
of Batail and Fourmigue et a1. first demonstrated in 1992 the
synthesis of a series of molecules such as 3,4-bis(diphenylphos-
phino)-3’.4’-dimethyltetratluafulvalenc (o-P2) and tetra(diphe-

 

(3) (a) Davidson. A: Boubekeur. K.; Pdnicaud. A.; Anbln. P.: Lenoir.
C.: BataiI.P .:.Herv€.6 J. Chem. Soc., Chem. Com.19l9.1373.
(b) Penicaud. A.; Boubekeur. K.; Batail. P. Canadell. 5.: Arthur-
Senzier. P.: Jerome. D. J. Am. Chem Soc. 1993. 115. 4101. (c) Coulee.
C.: Livage. C.: Gonzalvez. L.; Boubekeur. K.: Bani]. P. J. Phys. I
1993. 3, 1. (d) Coronado. E.: Gomez-Garcia. C. J. Comments Inorg.
Chem. 1995, I7, 255. (e) deez-Gareia. C. 1.; Gimme-Sail. C; Triki.
8.: Coronado. 5.: Magueres. P. L: Ouahab. L.: Dueasse. L.: Sean's-
seau. C.; Delhaes. P. Inorg. Chem. 1995. 34. 4139.

(4) (a) Fox. M. A.; Chandler. D. A. Adv. Mater. 1991, 3. 381. (1))ng
P.-W.; Fox. M. A. Inorg. Chem 1994. 33. 2938. (e) Anmllller. A.;
Erk. P.; Klebe. (1.; Hunig. 8.: von Schiltz. 1.: Werner. H. Angew.
Chem. Int. Ed. Engl. 1986. 8. 740. (d) Kato. K: Kobayashi. 11.:
Kobayashi. A. I. Am. Chem. Soc. 1989. III. 5224. (e) Manriquez..1.
M.; Yes. G. T.:Mc1.ean. 8.; Epstein. A. J.; Miller. .I. S. Science 1991.
252. 1415. (1') Miller. J. 5.: Calabrese. J. C.; McLean. R. 8.: Epstein,
A. J. Adv. Mater. 1992. 4, 498.

(5) Benelli. C.; Oct. A.; Ganeschi. D.; Gildel. H. U.; M L. Inorg. Chem.
1989. 28. 3089.

SOO20-1669(97)01123-3 CCC: $15.00 © 1998 American Chemical Society
Published on Web 03/19/1998

63

 

'F

1834 Inorganic Chemistry. Vol. 37. No. a 1998

nylphosphino)teuathiafulvalene (P4).‘u This series comprises
stable redoxcactive ligands that form strong metal interactions
enhanced by chelation through their phosphine functionalization.
Work in our laboratories has established that cationic complexes
of the type [M(o—l’2)z]2+ and [M(P4)].2"+ can be prepared by
reactions of the solvated precursors [M(1*<1CCI~13)...5]2+ with the
appropriate phosphine ligand.

In lieu of X-ray structures. a convenient tool for the
characterization of these molecules, particularly with respect
to nuclearity. is desired Mass spectrometry techniques provide
such a tool for detailed structural analysis. with fast atom
bombardment (FABMS) being the method of choice for the
analysis of nonvolatile. thermally labile compounds with mo—
lecular weights in the 500- 3000 range.’ However. approaches
used for obtaining and interpreting mass spectra for organo-
metallic complexes are not nearly as developed as those for
organic molecules. For organic molecules. atom bombardment
of a matrix/analyte solution typically yields protonated analytes
as the most abundant species due to “desorption/ionization". if
the species present in solution is already ionic such as Na+ or
Kt the species need only be desorbed. In most cases. only
singly charged ions are observed in FAB so the situation
becomes more complex when an analyte in the matrix solution
exists as a multiply charged species. The desorption of multiply
charged uansition-metal-containing species in FAB has long
been recognized as a limitation of the experiment. '° For organic
molecules that are multiply charged anions in solution. matrix
additives have been identified that assist analytes in lowering
their net charge. to a single negative charge. during the
desorption event.”

In the course of our studies on the tetrathiafulvalene phos-
phine complexes. we noted with interest. recent reports on the
characterization of cationic “molecular squares". It appears that
considerable promise exists for the use of FABMS to character-
ize cyclic inorganic ring structures with charges ranging from
+4 to +8 when the counteranion is trilluoromethanesulfonate
(tritlate).l2 Given this precedence for the usefulness of the
FABMS tool in these charged systems. considerations for the
successful analysis of phosphinecontaining cationic transition-
metal species are presented here. The chemical options of
highly charged ionic species in the FAB experiment will be
considered. Specifically. metal-containing complexes of the
form [M(o-P2)2](BF.); will be evaluated. The ntass spectro-
meuic analyses are being developed simultaneously with
synthetic protocols for these compounds The goal of this work
is to develop an understanding of the FAB mass spectra of such
complexes and to identify approaches for obtaining structurally
significant mass spectra. Also. the analysis of molecules
containing the P4 ligand. such as [Ptz(P4)(NCCH3)4](BFi)4. with
a higher ionic charge in solution. will be discussed. The FAB
specuum of this, and of larger P4 complexes. will be considered

 

(6) Canneschi. A.; Ganeschi. D.; Rey. P. Prog. Inorg. Chem. 1991. 331.

(7) lnoue. K.; 1wamttra. H. J. Am Chem Soc. 1994. 116. 3173.

(8) (a) Fourrrngue M.; Batail. P. Bull Soc. Chim Fr. 1992. 129. 829. (b)
Jar-chow. 5.; Fourrnigue. M.; Batail. P. Acta Crystallogr.1993.C49.
1936. (c) Fourrnigue. M.; Huang. Y. -S. Organometallics 1993. [2.
797. (d) Gerson. F.; Larnprecht. A.; Fourmigue. M. J. Chem. Soc.,
Perkin Trans. 1996. 2. l. (c) Fourmigue. M.; larchow. S.; Batail. P.
Phosphorus. Sulfur Silicon Relat. Elem. I993. 75. 175.

(9) De Pauw. 5.; Caption. R. M.; Moore. W. T.; Hayes. R. N: Gross. M.
L. In Methods in Etzymlogy: Mass Spectrometry; McCloskey. J.
A.. Ed; Academic Press: Boston. 1990; Vol. 193. pp 201-263.

(10) Miller. .1. M. Mass Spectrom. Rev. 1990. 9. 319.

(11) Huang. Z.-H.; Shyong. B.-.l.; Gage. D. A.; Noon. K. 11.; Allison. J. .I.
Am. Soc. Mass Spectrom. 1994. 5. 935.

(12) Whiteford, l. A.; Rachltn. E. M.; Stang, P. .1. Angew. Chem. Int. Ed
Engl. 1996. 35. 2524.

metal.

as well as a simple approach to obtaining meaningful spectra
of multiply charged u-ansition-metal species.

Experimental Section

Synthesis of o-P2 Complexes. The [M(o-P2);](BF.), (M = Pd.
Pt. Fe) and [Co(o-P2)2(NCCH3)z](BF4)2 complexes were synthesized
by reacting 2 equiv of the a-P2 ligand. dissolved in 5 mL of Cl-lzCla.
with [M(NCCH3).](BF4)2 (M - Pt. n == 4; M 8 Fe. Co. Pd. n :- 6)
dissolved in 5 ml. of CHgCN. The resulting solutions were stirred fa
thutdtheproductswcreprecipitatedbytheadditionoftoluene.
This method ts very similar to the previously reported synthesis of [Rh-
(o-P2);](BF.). which was structurally characterized by X-ray methods.“
Allproceduresfortheironcomplex wereperformedunderanuobic
conditions. whereas the other metal products were handled tn air. A
more detailed example of these procedures is described as follows.

[Co(o-P2)1(NCCHi)zl(BI-‘t)a. in separate flasks. 0 040 8 (0 084
mmol) of [Co(NCCH,).](BF.)1 was dissolved' tn 5 ml. of acetonitrile.
and 0.103 g (0.171 mmol) of o-(HuPhMeng‘F was dissolved tn 5 ml.
of dichloromethane. The cobalt solution was added to the Tl'F solution.
which effected a color change to dark brown. and the reaction solution
was stirred for 12 h. The solution volume was reduced by ~50%.
treatedwichOmLoftoluenetoaffordabtown solid.andftlteredln
air. 1heremainingbrownsolidwaswashedwithtoluene.3 x10mL.
to remove unreacted o-P2 and with diethyl ether. 4 x 5 mL. until the
washings became colorless. Finally. the solid was dried in vacuo. The
yield was 0.098 g (77%).

[Pt;(P4)(NCCHi)41(3Fc)a One equivalent of the P4 ligand and 2
equiv of PtCMNC-Illg); were combined in CHzCl; to form the insoluble
complex Ptz( P4)C1.. The solid was then treated with 4 equiv of AIBFa.
and in CH3CN. refluxed for 2 days. The resulting mixture was filtered
toremovetheAgCl byproductandthedesiredptoductwasprecipitamd
from solution with diethyl ether.

[Pd(P4)].(BI-‘.);.. Solid samples of P4 (0.207 g. 0.213 mmol) and
[Pd(NCCH,).](BF4)z (0.116 g. 0.261 ntrnol) were loaded into separate
flasks and dissolved in 6 mL of CH2C12 and CH3CN. respectively. The
P4 solution was then added to the metal solution. and the resulting
mixturewasstirred for 12h.resultinginadarkbrownsolution. The
solution was concentratedto5 mL. treatedwith 20mLofdiethyletbcr
to yield a brown precipitate. and removed by cannula. The resulting
brown solid was washed with diethyl ether 3 x 10 mL. and dried in
vacuo. The yield was 0.243 g (91% based on [Pd(P4)][BFc]1). Anal.
Calcd for PdPtSsCuHafirFt: C. 53.12; H. 3.31. Found: C. 53.67; H.
3.69. "PH-1} (6 ppm CD;CN): 44.2. IR (Nujol. cm"): 1049 (v.4).
689 (o-P2).

Mus Spectrometry. Mass spectra were obtained on a JEOL 10(-
110 double-focusing mass spectrometer (JEOI. Ltd. Tokyo. Japan)
operated in the positive ion mode. lons were produced by fast atom
bombardment (FAB) with a beam of 6 kV Xe atoms and an emission
currentomeA. 'l'hemassspecu'ometerwasoperatedwithan
accelerating voltage of 10 kV at resolutions between 1500 and 3000.
In the experiments investigating the desorption of Ba“. the solution
used was BaClz (EM Science. Gibbstown. NJ) dissolved in methanol
(.1. ‘1‘. Baker. Phillipsburg. NJ) at a concentration of 5 areal/pl. One
microliter of analyte was mixed with l pL of glycerol (Sigma Chemical
Co., St. Louis. MO) matrix on the FAB direct insertion probe. and
spectra were obtained. The o-P2 and P4 metal complexes were
dissolved in acetonitrile (J. ‘1'. Baker. Phillipsburg. NJ) at communion
of 5 nmollaL; 1 pL of this solution was mixed with 1 pl. of
tit-nitrobenzyl alcohol (Aldrich Chemical Co., Milwaukee. WI) and
introduced into the FAB ion source. Volatile solvertts qtnckly evaporate
and are removed when the target is placed into the vacuum system of
the mass spectrometer. For the Bad; and o-P2 complexes. mass spectra
were obtained by scanning over the mlz range of 1-1500 Da in 15 s;
for the P4 complexes. the m/z range of 1-6000 was scanned in 1 min.
Multiple spectra were collected. In some experiments triflic acid was
added; 1 uL of trit'lic acid (Easunan Kodak Co., Rochester. NY) was
mixed with the analyte/matrix solution immediately prior to analysis.

Analysis of TransitionaMetal Compounds

Mm: +2 +1 . +1

I
. mm :
W (Baa)? MOW. (ENGHW

a.
Bah

 

 

 

 

Glycerol Droplet Containing user,

figure 1. Desaptionfronization possibilities for Ba2+ from BaClz in
glycerol by +FAB.

Resultsanlescusslon

A. FAB Mass Spectra of Species That Are Multiply-
Charged In Solution. To consider the options of multiply
charged ionic species in the FAB experiment. we will discuss
BaClz as an analyte using glycerol (G) as the matrix. As a
solvent. glycerol is in many ways similar to water. ‘3 and BaClz
exists in ionic form in this matrix. A variety of desorption
pathways yielding positive ions are shown in Figure 1. The
formation of doubly charged ions. either bare. [Ba2*](g). or
partially solvated. [Ba2++G](g). is not expected. In FAB.
insufficient energy is provided to condensed phase species to
result in desorption of multiply charged ions. Ionic desolvation
energies are proportional to the square of the ionic charge.
according to the Born equation.“ The Born equation was
developed to estimate solvation energies of atomic and small
molecular ions. A form of the equation. to describe the energy

required for ion desolvation during desorption. is shown in eq
1.
v 8350’: 6r

In eq 1. an ion of charge z.e and radius r. is considered. being
desorbed from a solution with a dielectric constant of e... where
so is the permittivity of a vacuum. The dielectric constant for
glycerol is 42.5. The energy required for the complete desol-
vation of a Ba2+ ion from glycerol is 20.98 eV. The maximum
energy available for molecular desorption in FAB has been
estimated to be no greater than 19.9 eV.'5 Thus. one would
not expect to form Ba2+(g) in the FAB experiment. for
introduction into the mass analyzer. although it is the only form
of barium that exists in the solution. If barium existed as Ba"
in solution. only 4.6 eV would be required for its desolvation!
desorption. Thus. singly charged species such as Na+ are readily
desorbed; multiply charged ions usually cannot be desorbed
directly upon fast atom bombardment of a solution containing
the ion. As will be seen in subsequent spectra. +2 ions can be
desorbed in some cases. When charge is distributed throughout
a molecular framework. the Born equation does not accurately
represent salvation energies. However. for similar species.
solvation energies will have a strong dependence on overall
charge.

Figure 2 shows the +FAB mass spectrum of BaClz. using a
glycerol matrix. Barium contains seven isotopes. in the range
mBa -- m13a. The most abundant form is ‘3'Ba. with a natural
abundance of 71.6%. Each of the other isotopes has an
abundance of less than 15%. If formed. Ba3*(g) ions would
appear in the spectrum with an mlz value of 138/2 = 69; none

 

(13) Noon. K. R. PhD. Thesis. Michigan State University. 1995.
(14) Atkins. 1’. Physical Chemistry. 5th ed.; W. H. Freeman: New York.
1994

(15) rm...“ M. J. Am. Soc. Mass Spectrom. 1995. 6. 114.

65

Inorganic Chemistry. Vol. 37. No. 8. I 998 1835

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

macaw
229
[m1 .rIS/ ‘
t
321
g3 122 _
2
g (tumor .
. 2‘5\ +92 +92 1‘9
a‘ J \. a
0-4 :‘JL . . . . . h. an]. L} 4‘, -‘T
6 too 200 300 400 '
m

Figurez. +FAB massspechumofBaChinglycerolJeaksrepesent-
ing ions evolving from the glycerol matrix are labeled with an asterisk
(‘l-

are observed. Also. no peak representing Ba+(g) is present (HI/Z
= 138). Thus. reduction accompanying desorption. to lower
the charge on the metal. does not occur. Typical of a FAB
spectrum. most of the low-mass ions (below mlz = 200) are
related to the glycerol matrix. The peaks at rot/z = 93. 185.
277. 369. and 461 are all proton-bound glycerol clusters.
[G.+ 1—1]+.n=l-°5.'l'liepeaksatlowermass(mlz=45.
57. 75) are fragment ions of glycerol.“

Interestingly. two types of ions are generated during FAB of
BaCl; in G. An intense peak is observed at nrlz = 229. which
is not due to the glycerol matrix. From its isotope pattern. it
must contain one Ba atom. and it does not contain C1. The ion
must be singly charged since the separation between isotope
peaks is l mu. The ion corresponds to [Ba + 911*. Since
themassofGis92.theaddcdmassof9l Damustbe(G-
H)‘. the glyceride ion. Therefore the major peak in the
spectrum at m/z = 229 represents [Ba2"'(G - 11)‘]*. Appar-
ently. Ba“ reacts with a glycerol molecule to form a singly
charged ion. leaving a proton in solution. as suggested by
reaction 2. A bound glycerol molecule forms a negatively

[Ba2+ + G](solv) L“- [Ba2+(G - H)'l*(g) + H+(solv) (2)

charged ligand (G - H)‘. and the resulting complex has a single
chargeandcanbedesorbed. ‘Ihisisoneexampleofthesolution
chemistry that may occur during the desorption process in order
to reduce a high charge (greater than unity). Solvated forms
of the W: = 229 ion are observed at m/z = 321 and 413 in
FigureZ. ’l‘hesethreepeaksareseparatedby92u.correspond-
ing to the addition of neutral glycerol molecules.

If a Ba“ ion reacts with G to form an anion and complexes
with that anion to desorb with a single net charge. one would
expect that the +2 metal ion could also bind to a Cl“ ion during
desorption. A peak representing BaCl" (m/z = 173) is not
observed. However. an ion is generated of the form [BaCl + 01*
(m/z = 265). representing a solvated BaCl" ion. Since glycerol
forms clusters with glycerol ions. clusters of glycerol attached
to ions evolving from the Bad; are not unexpected. In the
FAB target in this experiment. [Ba2+] < [0"] 4! [0], so it is
not unreasonable that Ba“ reacts with glycerol in solution as
well as CI‘ in order to lower its charge. Similar systems have
been studied by Miller.” Instead of describing the mlz = 265
ion as a solvated BaCl+ species. it should be considered as

 

(16) Székely. 0.; Allison. I. J. Am. Soc. Mass Spectrom. 1997. 8. 337.
(17) Saraswathi. M.; Miller. J. M. J. Chem. Soc., Dalton Tm. 1997. 341.

1836 Inorganic Chemistry. Vol. 37. No. 8. I998
93 @© 2+
{>41 H21:

Flgare 3. Monomrclear structure law-92);”.

n,cI:)—<:

31C

an incompletely desolvated ion. Thus. its formation requires
less energy than that for the completely desolvated form. Note
that. similar to the sequence of peaks labeled mlz = 229. 321.
and 413. addition of 1 and 2 6': to mlz = 265 occurs. leading
to mix = 357 and 449.

lfcations from ionic compounds of the form [M'l'(L‘)..] have
available pathways. they can desorb from solution with a low
(single) charge in the positive ion FAB experiment. If they
cannot. no analyte-related peaks will be observed in the +FAB
spectrum. When analyzing an ionic compound by FAB. the
chemist should not expect to detect abundant doubly or triply
charged ions. although doubly charged ions are observed for
some complexes. ‘0'" Even in cases where doubly charged ions
are observed. the singly charged ion peaks are usually of greater
intensity with a better signal-to-noise (SIN) ratio. Observation
of triply charged ions by FAB is extremely rare for ionic
species.”

B. FAB Mass Spectra of [M(o-P2)1](BF4); Complexes. 1.
Chdce of Matrix. While BaCh. in the example above. is
soluble in glycerol. the o-P2 complexes studied here are not. A
matrix used very successfully in the analysis of inorganic and
organometallic compounds by FABMS is nr-nitrobenzyl alcohol
(NBA).‘°~m‘ These complexes are soluble in this matrix. and
representative ions are generated. There is a significant
difference between NBA and glycerol in the FAB experiment.
NBA is known to participate in redox reactions and oxygen-
transfer reactions with analyte molecules. rather than just acid!
base chemisz. NBA has both oxidative and reductive capa-
bilities.“ This provides an opportunity to investigate other
chemical options available to multiply charged complexes. to
lower their charge during FAB.

2. FAB Mass Spectra of o-P2 Complexes. The BaCh FAB
example represents a simple system to show how a multiply
charged atomic ion can be converted into a singly charged
molecular gas-phase species. Of course. the situation changes
for the FAB analysis of these complexes. There are reduced
direct metal—solvent interactions due to the presence of the o-P2
ligands. and the counterions are molecular and more complex.
Also. the NBA solvent offers other options for desorption of
charged complexes in lower charged states. A set of complexes
of the form llv‘1(o-l"'2)22+ was made (M = Fe“. Co”. Pd“. Pt“)
and isolated as the tetrafluoroborate salts. The structure of the
him-1’2):2+ complex is shown in Figure 3. Since the ligands
are neutral. the metal determines the charge on the complex.

a. The FAB Mass Spectrum of [Fe(o-P2);](BF4);. Figure
4a shows the +FAB spectrum of [Fe(o-P2)1](BF.); using NBA

 

(18) Miller. J. M.; Balasmmugam. X.; Nye. J.; Deacon. G. 3.: Thomas.
N. C. Inorg. Client. 1987. 26. 560.

(19) Milt. M.; Tamas. J.; Maho. 8.; Przybylslti. M.; Tuba. 2. Anal. Chins.
Acta 1990. 24!. 289.

(20) Barber. M.; Bell. 0.; Eckersley. M.; Morris. M.; Tetler. 1.. Rapid
Cbnunan.filass5becvowt 1988.2.18.

(21) Reynolds. I. D.: Cook. K. 0.; Burn. 1. L. E: Woods. C. J. Am. Soc.
Mass Spectrom. 19,1. 3. 113.

 

 

 

 

 

 

 

 

 

 

 

Asara et a1.
a
if» I a fl
1m? r. r- rm,
...» 1
i
Vi ’
0 ‘ in“ VI
8
i /m \
o _‘ Ll_ _
t ' T T r x 1 1
o 500 1
unit
b ms
1276
as 10" 1277
i ...
i 3‘
0.. T a - e
1270 1275 12”
av:
C 1275
1m
1271
so
1279
rare
I l l 7 l .
1273 m use

Flgure 4. (a) +FAB mass spectrum of [Fe(o-P2);](BF.); in nt-
nitrobenzyl alcohol. (b) Enlargement of a portion ofthe mass spectrum
about the nth. = 1275 peak. (c) Theoretical isotopic distribution for an
ionwith the[Fe(o-P2)z + Freompositioo. Portionsofthespectraare
enlarged-multiplicationfactorsareindicated(intheseandother
spectra).

asthematrix. All ionsobservedare+l ions;nopeakatm/z
= 628. representing the doubly charged. intact Fe(o-P2)23+
species.isobserved. Themostintensepeakobservedrepresent-
ing ions containing the intact Fe(o-P2)2 core is at nr/z = 1275.
This is 19 u greater than the core mass and represents
[Fe(o-P2)1 + 171*. Figure 4b shows an enlargement of the mlz
= 1275 peak. The elemental composition of the ions repre-
sented by this peak is FeCuHuPtSf. 1f the lowest mass isotope
of each atom is used. 1273 would be the nominal mass. This
peak is very small. The most intense peak in the isotopic
distribution. III/Z = 1275. will be used to represent [Few-1’2);
+ 171*. Figure 4c shows the theoretical isotopic distribution
for this elemental composition as calculated by Isotopic Profiler
from WindowChem Software. Inc. (Fairfield. CA). The "P-
{'I-I} NMR spectrum of [Fe(o-P2)2](BF4)2 displays two signals
at d = 56.1 and 75.5 ppm. respectively. indicating the presence
of two different phosphorus-containing species in solution. The
resonance at 56.1 ppm is shifted downfield from the free ligand
and falls in the range previously reported for M(o-P2)g
complexes.““ The second resonance is shifted ~20 ppm

66

Analysis of Transition-Metal Compounds

downfield from the first. which has been reported by others to
signify the presence of coordinated BF.‘ or F‘ ligands?"c
The 19F NMR Spectrum is in accord with this hypothesis. with
singlets appearing at d = - 148.6 and -503.0 ppm. respectively.
The resonance at d = -l48.6 ppm is in the range for values
reported for uncoordinated BF.” and for the terminal fluorines
of a coordinated BFr" anion.22d The singlet at -503.0 ppm is
indicative of bound fluorine from coordinated BFi" or an
abstracted F“ ionm‘ In any case. it is obvious that the cation
is associated with a F source in some form. which leads to a
fluoride ion being associated with the expected Fe(o-P2)22+
complex. reaction 3. Such a phenomenon has been seen

Fe(o-P2)22+ + BF." fl [Fem-i=2)2 + F]+ + BF,(g) (3)

previously in the detection of [LW(CO)..(l~<lO)F]+ (n = 0. 1. 2)
ions in the mass spectra of [l.W(CO)(NO)SbF..]+ (L = PMe3.
PMCzPhP).m This shows an additional option that multiply
charged cations have. when the counterion is molecular and
bonds within the counterion can be broken. With both F-
coordinated and uncoordinated species present prior to analysis.
it may be important to determine if correlations exist between
the relative intensity of the mlz = 1275 peak and variations in
the synthetic procedures. Figure 4a also shows a singly charged
ion. Fe(ooP2)2+. at m/z = 1256. Apparently. in NBA. a le‘
reduction of the complex occurs. either prior to or during the
FAB process. to generate a singly charged ion.

Interestingly. the most intense analyte-related peak for this
complex is m/z = 600 which represents the ionized free ligand.
o-P2'*. Free o-P2 ligand could have been present in the sample.
and the ion could have been formed directly. However. the
complex was washed several times in order to remove any free
ligand present. so it is unlikely that o-P2 was in the sample.
The absence of a peak at -l8.8 ppm in the 3'I’{‘H} NMR
spectrum confirms this result. The other option is that the peak
at nrlz == 600 is a fragment ion. If the primary ion, Fe(o-P2)22“’.
was desorbed but was unstable in the gas phase. fragments may
appear at mlz = 656 and m/z = 600 representing Fe(o-P2)+
and o—P2"'. respectively. However. Fe(o-P2)* is not observed.
In a complex such as Fe(o-P2)2“. considerable charge delo-
calization would be expected. and the formation of charged
ligands during metal-ligand bond cleavage is reasonable. It
is expected that the o-PZ ligand is easily ionized. due to the
presence of the aromatic groups. For example. while the
ionization energy (IE) of phosphine is 9.9 eV. the IE of triphenyl
phosphine is more than 2 eV lower, 7.4 eV.23 Thus. formation
of ionized o-P2 ligands in the spectra should not be unexpected

There are a number of other analyte-related peaks in the m/z
= 600-700 region of the spectrum shown in Figure 4a. The
peaks at m/z = 616 and 632 represent the addition of 1 and 2
oxygen atoms to the ligand. The oxidation sites are presumably
at the phosphorus atoms since the o-P2 molecule appears to be
capable of incorporating no more than 2 oxygen atoms. Also.
+FAB spectra (not shown) of the free o-P2 ligand have been
obtained and show the same two oxidation products. The peaks
at m/z = 616 and 632 may arise from oxidation of the ligand

 

(22) (a) Honeychuck. R. V.; Hersh. W. H. Inorg. Chem. 1987. 26. 1826.
(b) Honeychuck. R. V.; Hersh. W. H. Inorg. Chem 1989. 28. 2869.
(c) Lundquist. E. 0.; Folting, K; Hoffman. I. C.; Caulton. K. G.
Organomerallics 1990. 9. 2254. (d) Fernandez. I. M.; Gladysz. l. A.
Inorg. Chem 1986. 25. 2672.

(23) Lias. S. G.; Bartmess. J. E.; Liebman. J. F.; Holmes. 1. 1...; Levin. R.
D.; Mallard. W. G. Gas-Phase Ian and Neutral Thermochemtstry.
American Institute of Physics and American Chemical Society: New
York. 1988.

Inorganic Chemistry. Vol. 3 7. No. 8. I 998 1837

when it is still bound to the metal. Clearly. P-oxidation occurs
at the expense of P—M complexation. This has been seen
previously through the isolation of dioxidized P4 as a result of
the reaction between CoClz and 1’4.‘e Other small peaks are
present which show that the intact complex can undergo
oxidation. For example. the peak at ntlz = 1291. 16 it above
1275. represents [Fe(o-P2)1F + O]+. This should be more
accurately described as [Fe2+(o-P2)(ooP2-l-O) + F‘]+. Finally.
consider the peak at mlz = 675. which is 75 u above that for
the free ligand. This corresponds to addition of an iron and a
fluorine. This is the only peak which represents the metal with
a single ligand attached. as [Fe(o-P2)+F]"'.

The BaClz example showed that multiply charged metals lave
a variety of options in generating singly charged species in the
desorption process. The [Fe(o-P2)1](BF4)2 data show that
additional options are available when the counterion is BF.’.
In this case. the spectrum is dominated by ions representing
the free ligand. and relatively small peaks establish the molecular
structure.

h. The FAB Mass Spectrum of [Pt(o-P2);](BF4);. On the
basis of the spectrum of the Fe(o-P2)22+ complex. we would
expect certain peaks to be observed for the platinum complex.
Few of these are present in the spectrum. which is shown in
Figure 5a. On the basis of data obtained using Fe(o-P2)zz*.
one would expect no peak for the doubly charged l’t(o-P2)z2+
ion (tn/z = 698). The spectrum shown in Figure 5a displays a
relatively intense peak at m/z = 698. Isotopic peaks for the
doubly charged ion displayed a separation of 0.5 m]: units
between peaks and an oxygen adduct (nu/z = 706) at only 8
ml: units higher than the peaks for Pt(o-P2)22+. confirming the
desorption of doubly charged ions. In contrast to the spectrum
of the iron complex. no peak representing the ionized o-P2
ligand is present at m/z = 600. This may suggest that there is
a metal-dependent relationship controlling the relative abun-
dances of the M(0-P2)22+ and o-P2"' ions. The M(o-P2)1“'
species should have different charge distributions when M =
Fe and Pt. The sum of the first and second ionization energies
of Fe (24 eV) is less than that for Pt (27.5 eV).“ Thus. the
iron complex may have more L‘+-M*-L6* character and the
Pt complex more L-M2+-L character. Delocalized charge can
weaken the metal-ligand interaction and would result in an
ionized ligand during bond cleavage of the desolvated complex
in the case of iron. If a larger fraction of the total charge resides
on the metal in the platinum complex. then the Pt(o-I’2)22+
species may survive the desorption process. without undergoing
fragmentation.

For the iron complex. a small peak was seen for Fe(o-P2)1+
due to a le‘ reduction, possibly during the desorption process.
Figure 5a shows that Pt(o-P2)2+ (m/z = 1396) is an intense peak
in the spectrum. The second IE ofPt (18.6 eV) is greaterthan
that of Fe (16.2 eV);2-" thus reduction by reaction with NBA is
more exothermic for the Pt complex. favoring formation of the
singly charged Pt(o-P2)2‘* species. Figure 5. parts b and c.
shows the isotopic peaks for mlz = 1396 and the theoretical
isotopic distribution. confirming its elemental composition and
its net charge.

The spectrum of the iron complex showed that the doubly
charged species could react with BF.‘ to form a singly charged
gas-phase species. via fluoride extraction. No ions are forer
containing a Pt-F bond. There is. however. a peak at m/z =

 

(24) Huhecy. J. E. Inorganic Chemistry: Principles of Structure and
Reactivity. Harper & Row: New York. 1972.

(25) Miller. I. M.; Jones. T. R. B.; Deacon. G. B. Inargan. China Acta
1979. 32. L75.

67

1838 Inorganic Chemistry. VoL 3 7. No. 8. I 998

 

 

 

 

 

 

 

 

 

 

 

 

 

 

I | :75
a 1001 14°!“ 1'
. 698 1396
9 1
’3
g 1
r ”i
i + 706 1483
"‘ . ( 1412\
l
0'1 r t ' TL t r ' to r r r V'T'H
0 :00 11!!) 15(1)
1396
1.04 13
b 397
— J
2 0.5
3
o'o-‘UA'I'T' t'r'r'r 'r rUfIvTfij
1380 1390 1400
N:
c 1396
too
1398
1394
50
rm 7 1402
ml!

Figure 5. (a) +FAB mass spectrum of [Pt(o-PZH(BF.)2 in rn‘
nitrobenzyl alcohol. (b) Enlargement of a portion of the mass spectrum
about the mlz - 1396 peak. (c) Theoretical isotopic distribution for an
ion with the Pt(o-P2)2* composition.

1483 which represents the addition of the BF? anion. [Pt(o-
P2); + BF.]+. not unlike the complexation of Ba2+ with CI‘ to
form a singly charged gas-phase species.

In the spectrum of the iron complex. evidence for oxidation
was observed. The free ligand. as the mono- and di-oxide. was
observed. Oxygen addition to the complex was also directly
observed. In the spectrum of the Pt complex. there is a peak
16 u above m/z = 1396. representing oxygen addition to the
singly charged complex. Thus. the ion with M]: = 1412 in
Figure 5a is described as [Pt(o.P2)(o-P2 + O)]*.

The complexes of [Pt(o-P2)2](BF4)2 and [Fe(o-P2)2](BFa)2
show that there is a strong metal dependence on the +FAB
spectra of o-P2 complexes. Hard/son acid/base (HSAB) theory
may be a useful tool in predicting metaLdependent differences.”
In reactions involving F‘ extraction from BFa". HSAB theory
would suggest the process to be more likely for Fe than Pt; this
is observed Additional data for other metal complexes is
provided in Table l. The iron and cobalt complexes were the
only ones observed to extract P. while the larger divalent metals
form adducts with the intact BF4‘ ligand. supporting this
correlation.

Asaraetal.

Table 1. ion intensities for [mo-mums.» Complexes by
+FAB‘

metal C’ CH [C +0]+ [Ci-Fl" P2" M(o-P2)'* [C-erFa]+

 

Fe” W W I I
Co’+ l M W W W W

* I M W M M
Pt” 1 M W M

‘C-[M(o-P2);].i-intensepesk:M-mediurnintatsitypeak;
W-lowintensitypeak.

.... “M

Q Q
>~<§H©JES
w ©< t».

Figure 6. Polynuelesr strucntre [M(P4)]."*

l

\
Y7

 

 

The M-ligand bonds are more easily cleaved in the Fe
complex than in the Pt complex. consistent with HSAB
interactions. Again. the Pd complex yields a spectrum similar
to that observed for Pt. from the data in Table 1. Table 1 also
further supports the proposal that the intensity of the peak
representing the ionized. free ligand is proportional to the sum
of the first and second 153 of the metal. Of the four metals.
the sum is lowest for Fe. and in Figure 4a the o-PZ'+ peak is
most intense. The sum is higher for Co. and a weak signal for
the ionized ligand is observed; the sum is highest for Pt and
Pd. and no ionized ligand peak is present.

C. FAB Mass Spectra of Transition-Metal Complexes of
the P4 Ligand. Structurally similar to o-P2. the P4 ligands
contain four phosphines. It is apparent that the use of P4 can
lead to oligomeric complexes of the type [M(P4)].3'"". Figure
6. These represent an important aspect of the analytical
challenge of this work. Due to the unavailability of single
crystals at this stage. it is unclear precisely what oligomeric
species are being formed. Consequently. there is a keen interest
in being able to identify them using FABMS. Unfortunately.
with each additional metal. the charge increases by +2. and
the complexes become increasingly refractory to FAB analysis.
unless the species in solution can access chemical routes to
sufficiently lower their charge during the desorption process.
The work presented for the o-P2 complexes was undertaken in
an attempt to develop an understanding of how the structures
of such larger species correlate with the FAB spectra obtained.

1. The FAB Mass Spectrum of [Pt;(P4)(NCCHJ)4](BFa)4.
The first P4 complex studied was [Pt2(P4)(NCCH3)4](BF¢)..
which was synthesized specifically to terminate with one P4
ligand and two metals (see Experimental Section). Each square-
planar Pt” is also bound. initially. to two acetonitrile molecules.
Figure 7. In solution. the core cationic complex is quadruply
charged. [Ptz(P4)]‘*. There is no evidence in the spectra that
the acetonitrile ligands are retained. Figure 8a shows the +FAB
Spectrum of [Pt2(P4)](BF4)¢ using NBA as the matrix. Due to
the high desolvation energy for this highly charged ion and the
limited amount of energy provided by the fast atom beam. the
ions must access a pathway to reduce their charge. It apparently
cannot do so; desorption does not occur. and no gas-phase ions
representing this complex. in any charge state. are formed.

In FAB. some analytes show a strong dependence on the
matrix used. However. for organometallic and inorganic

68

Analysis of Transition-Metal Compounds

 

 

Q Q Q M
rt,ccu\ s /Nccn,
B,CCN/ “21“]: :ll:\/:[ / PKNccrt,

Figure 7. Dinuclear structure (Pt1(P4)(NCCH,).]“.
a too """°°
E-
E
.2 5°
E
&
we
0 I l V '—r I V T l I V Y
0 2000
rnlz
b
1794
E‘
“' mo
5 1152\9 1927
2 5°
:3.

 

Figure I. + FAB mass spectrum of [Pt1(P4)(NCCH3).l(BFa)i in (a)
Iii-nitrobenzyl alcohol and (b) nt~nitrobenzyl alcoho lwith addition of
1111' he" acid.

complexes. NBA is preferred. When organic molecules cannot
be readily detected in FAB. "additives" can be used. These
are frequently as simple as acids (to increase the amount of
protonated molecules formed) or salts (to allow species such
as [M + Na]+ to he formed). In considering possible additives.
we noted that some highly charged organometallic complexes
have been successfully characterized by FAB AB.”3‘ For example.
Stang et al. ‘2 used FAB to characterize a complex containing
four PW“ ions. introduced as a [core)‘+(L ). salt They observed
singly charged species in the mass spectrum such as loom“ +
7L‘]". Unfortunately. similar species are not formed in the
[Ptz(P4)](BF.). experiment One possibly important variable
is that. in the Stang experiment. the counteranion used was
triflate (CF3503‘ or (71‘1“). Rather than have the P4 compounds
made again as the triflate salts. the decision was made to
evaluate NaOTf u a matrix additive in these experiments. This
a proach did not assist in generating ions related to the [Pt1-
(P4)]‘+ complex. However. success was achieved by using
triflic acid (HOT f) as a matrix additive.

D. HOT! as a Matrix Additive in FAB. When 1 pL of
HOTf is added to the NBA matrix/analyte solution used to
obtain the spectrum shown in Figure 83. that shown in 8b is

 

(26) Didier, P.. lacquet. L; Me sermaek A. K. -D.; Hueber, R.: van
Dorsselaer. A Inorg. Client. 1992. JI 4803.

Inorganic Chemistry. Vol. 37. No. 8. I998 1839

obtained. High-mass ions in the mlz = 1500-2000 range are
now observed. The mass of the Pt1(P4)“’ core is 1331 it (based
on the most abundant isotope of the metal. MR). The most
intense peak in Figure 8b in the expanded region of the spectrum
is at m/z = 1778. which is 447 u greater titan the core mass.
Since uiflic acid molecules have a mass of 150 u. this
corresponds to the addition of three triflate ions. each with a
mass of 149 u. reaction 4. That is. the addition of triflic acid

MM)" + snort]... L“. [mmnromftsi (4)

allows for the detection of +4 species that were present in the
NBA solution and that could not be successfully converted into
lower-charged species without this additive. As previously
reported. triflic acid does not serve as a source of fluoride ions.
which explains why F“ abstraction was not observed.”

Oxygen adducts are seen at m/z = 1794. 1810. and 1826 with
each Peak Elmsmfins IPlz(P4)(0T 034-01”. [PlzINXOTDi
+20]"‘. and [Pt2(P4)(OTf)3+30]+. respectively. Recall that.
in the case of o-P2 complexes. no more than two oxygen atoms
could be added. Other in the spectrum include rrrlz =
1629. which represents [Ptz(P4)(0Ti)z]*. and a peak at rrllz =
1927. which represents [Ptz(P4)(OTf).]+. Presumably. the mlz
= 1629 peak is a partially reduced form. [Ptz+Pt+(P4)(OTl):]*.
Reduction of Pt2+ in the o-P2 complex was also observed. More
interesting is the m/z = 1927 peak. which could represent a
mixed valence species [Pt1*Pt’*(P4)(OTf).]+. Oxidation of a
Pt“ would require a considerable amount of energy. even in a
solvent in which oxidation has been reported to occur. A more
reasonable possibility is that oxidation of the ligand has occurred
to give (Pt1*Pt“(P4)+(0TI).]+. In any case. the ion observed
contains four triflates; the increased electrostatic interactions
of the charged ligands with the charged metals may effectively
stabilize the higher-charge state in this gas-phase complex. These
results show that addition of uiflic acid allows for the detection
of ions present in the FAB matrix that are undetected in the
normal FAB experiment

To test the general utility of triflic acid addition. the
oligomeric complex [Pd(P4)].(BF4)z. was analyzed by FAB
using NBAas ematrix. Again. no peaks representing the
analyte are observed. Figure 9a. Triflic acid was then added t
the matrix/analyte solution. and the spectrum shown in Figure
9b was obtained. In this case. ions with m/z values as large as
4000 were detected. When expanded. each peak in Figure 9b
contains a cluster of peaks representing various oxygen-
containing forms. For example. the peak at ml: = 940
represents the ionized P4 ligand. Higher mus peaks at nr/z =
956. 972. 989. and 1015 indicate the addition of up to 4 oxygen
atoms (1 to each phosphorus atom). The dominant peaks
fewest!!! [Pd(P4)](0Tf)" ("I/z = 1194). IPd2(P4)](0Tf)3"’ (In/z
= 1601). [l’dz(i’4)zl(0'l‘f)i+ (In/z = 2543). [Pd3(P4)z](0Tt)5"
(mlz = 2945). and [li'd;(1’4)3](0'l'1)5+ (m/z = 3890). Clearly.
reduction and triflate addition occurs to yield singly charged
ions which are desorbed. The important result is the observation
that an oligomeric rrtixture of complexes was formed with n z
3. Larger oligomers may be present but not detected. The P4
complexes could not have been characterized by FAB without
the chemical assistance of an additive such as triflic acid.

Having demonstrated the utility of triflic acid addition. the
question remains as to why it is effective in enhancing detection
of highly charged species in the FAB experiment and why
addition of sodium triflate is not. We will assume that. for the

 

(27) Slang. P. J.; White. M. R. Aldricht'micn Acta 1983. I6. 15.

69

1840 Inorganic Chemistry. Vol. 37. No. 8. I998

 

 

 

 

 

 

 

a 1001 r‘”
.é‘ ‘
a 4
g 4
.2 ”‘
i
3
4 W: .a. “Vi-J‘- ......uu
0- .--, ..---s,
icon 2000 3000 4000
all
h .001 I—«zs /“94 r-tso r-zso
h 4
i ‘ / ' 1601 2543 2945 3890
3- /
.9. so
3.
a?
o- ' firfi ' ' V I ' ' ' ' I ' V ' ' I
1000 2000 3000 4000
W1

Figure 9. +FAB mass spectrum of [Pd(P4)].(BF4)z. in (a) m-
nitrobenzyl alcohol and (b) Ira-nitrobenzyl alcohol with addition of triflic
acid.

amount of triflic acid used. the NBA solvent used. and the
dielectric constant of NBA. most of the triflic acid is present as
HOTf. We will also assume that one or more triflic acid
molecules are present in the salvation sphere of ionic species
formed. In the liquid FAB target. ions and their salvation spheres
are more accurately described as [Pt2(P4)]‘*(NBA).(1-l0'l‘f),. for
example. From such species. desorbed. desolvated ions are
formed. In the desorbed species. whatever their final fornt. all
NBA solvent molecules are removed; when HOTf is added. all
are removed as well. However. the triflic acid appears to be
converted into triflate ions during the atom bombardment
process. The neutral triflic acid may easily bind to the core
metal complex. while counteranions are solvated and are farther
from the cations. Thus. when energy is deposited and desorption
can occur. one or more uitlic acids are converted into triflate
ions. releasing protons to the solution, reaction 5. similar to the
reaction of Ba2+ with glycerol.

[MP4)1“(Hom.(sotv) L”- wm + aorrm) +
3H+(solv) (5)

Conclusions

Mass spectrometry is a commonly used tool in the develop-
ment of synthetic methodology in organic chemistry. since
meaningful spectra can usually be obtained in the period of less
than an hour. However. inorganic and organometallic chemists
turn far less frequently to MS, in part because of the situation
encountered in the analysis of the metal tetrathiafulvalene
phosphine complexes described in this report. Highly charged
states negate the possibility of generating analyte-related gas-
phase species for subsequent mass spectrometric analysis. The
conversion of multiply charged cationic complexes to singly
charged species by acquiring anions from neutral molecules in
the salvation sphere appears to be a viable pathway. For
compounds that are soluble in glycerol. glycerate generation
can occur. but when monitrobenzyl alcohol alone is used. the

Asaraetal.

analogous process does not occur. Triflic acid as an additive.
however. seems to be able to serve in this capacity. The neutral
HOTf molecule complexes with the multiply charged species.
it forms an artionic ligand that strongly binds to the metal
center(s). and it ejects a stable cation (proton) into the solution.
Overall. the process can occur with the energy available in the
FAB experiment. As synthetic procedttrcs develop and larger
oligomeric species of the tetraphosphine ligand (P4) in this work
are generated. we will be in a position to determine the range
of charge states over which FAB can be applied when matrix
additives are employed.

It should be noted in closing that. while the work reported
here used FAB. a variety of desorption/imitation techniques
are now available in mass spectrometry. Newer methods such
as electrospray ionization (1381) and matrix-assisted laser des~
orption/ionization (MALDI) are rapidly becoming more avail-
able. while FAB continues to find new areas of application.”
Each method has its strengths. weaknesses. successes. and
failures. which is why few MS laboratories solve structural
problems with just one technique. MALDI has captured the
attention of the MS community by generating predominantly
singly charged ions from analytes with molecular masses greater
than 100 000. The role of additives in the MALDI experiment
is not well-understood. In making the target for laser irradiation.
matrix. analyte. and additives must first be cocrystallized. Triflic
acid is not a candidate additive in MALDI. since it is a liquid
at room temperature. Electrospray is an obvious method to turn
to when the analyte is multiply charged in solution. since it is
capable of generating multiply charged gas-phase ions. How-
ever the mass spectrometer of choice in ESIMS is usually a
quadrupole mass filter or an ion trap. These have limited mlz
ranges. relative to time-of-flight or magnetic sector instruments.
Also. they do not provide high-resolution data as would a
double-focusing instrument. There are advantages to using the
FAB-based approach developed here, since singly charged forms
of the analyte are generated. This facilitates the interpretation
of MS/MS spectra which can provide structural information.
While MS/MS spectra can be generated for multiply charged
species. interpretation of the resulting spectrum can be very
difficult. If a +4 ion is subjected to collisional~activation and
fragmentation. it will typically form +4, +3. +2. and +1
fragment ions. This is particularly difficult to deal with when
the analytes are metal-containing compounds such as those
discussed here.

Acknowledgment. 'IheauthorswouldliketothankDr.Mare
Fourrnigué and co-workers for supplying the o-P2 and P4
ligands. We also gratefully acknowledge the NSF International
Programs (INT-9217191) and ACS PRF for financial support.
We thank Michigan State University (MSU). for providing the
Herbert T. Graham and Carl H. Brubaker fellowships for C.E.U..
and the Sigma Xi chapter of MSU for a graduate student
research grant. The NMR equipment was supported by a grant
from the National Institutes of Health (NIH). Grant No. 1-510-
RR!475001. The mass spectrometry work was performed in
the MSU Mass Spectrometry Facility. which is partially
supported by Grant No. RR-00480 from the Biotechnology
Research Technology Program of the National Center for
Research Resources of the NIH.

IC971123W

 

(28) Henry. C. Anal. Chem. 1997. 625A.

70

APPENDIX B

71

 

Enhanced Detection of PhosphOpeptides in
Matrix-Assisted Laser Desorption/Ionization
Mass Spectrometry Using Ammonium Salts

John M. Asara and John Allison
Department of Chemistry, Michigan State University, East Lansing, Michigan, USA

 

Matrix-assisted laser desorption/ ionization mass spectrometry (MALDI MS) has been used
successfully to detect phosphorylation sites in proteins. Applications may be limited by the
low response of phosphopeptides compared to nonphosphorylated peptides in MALDI MS.
The addition of ammonium salts to the matrix/analyte solution substantially enhances the
signal for phosphopeptides. In examples shown for equimolar mixtures, the phosphorylated
peptide peaks become the largest peaks in the spectrum upon ammonium ion addition. This
can allow for the identification of phosphopeptides in an unfractionated proteolytic digestion
mixture. Sufficient numbers of protonated phosphopeptides can be generated such that they
can be subjected to postsource decay analysis, in order to confirm the number of phosphate
groups present. The approach works well with the common MALDI matrices such as
a-cyano-4-hydroxycinnamic acid and 2,5«dihydroxybenzoic acid, and with ammonium salts
such as diammonium citrate and ammonium acetate. (J Am Soc Mass Spectrom 1999, 10,
35—44) © 1999 American Society for Mass Spectrometry

 

 

rotein phosphorylation is probably the single
Pmost common and important reversible intracel-

lular signal transduction event [1]. Understand-
ing the regulation of protein function by phosphoryla-
tion/dephosphorylation is a key objective in many
areas of biomedical research, including studies involv-
ing cell cycle regulation, enzyme activation/deactiva-
tion, and protein-protein association. Phosphorylation
serves other functions as well in protein chemistry.
Protein-bound phosphate may serve to maintain the
structure of an enzyme without regulating its activity
[2]. Also, as in the classic cases of those found in egg
yolk and milk, phosphoproteins serve as stores of
phosphate and bound metal ions [2].

Direct analytical methods to identify phosphoryla-
tion sites of proteins of known sequence, at or below the
picomole level, are critical to these research areas. Mass
spectrometry has played an essential role in mapping
posttranslational modifications of proteins such as
phosphorylation [3]. Both matrix-assisted laser desorp«
tion/ ionization mass spectrometry (MALDI MS) [1,
3—6] and electrospray ionization mass spectrometry
(ESl MS) [3, 7, 8] have been used effectively, each with
its strengths and limitations.

Suppose a protein P is multiply phosphorylated. One
approach to locate the phosphorylation sites is to first
subject P to a proteolytic digestion using, typically,

 

Address reprint requests to Dr. John Allison, Department of Chemistry,
Michigan State University, East Lansing, Ml 48824. E-mail:
allisonecemvaxcemmsuedu

© 1999 American Society for Mass Spectrometry. Published by Elsevier Science Inc.

1044-0305/99/51900
Pll SlO44-0305(98)00129-9

trypsin. The set of degradation products [DJ may be
analyzed in a "batch" method. That is, all of the
products can be analyzed simultaneously using MALDI
[4, 6]. If the mixture is complex, high-performance
liquid chromatography (HPLC) can be used to fraction-
ate the mixture. Fractions can be collected and each
subjected to MALDI analysis [1, 6, 9], or the HPLC
effluent can be analyzed directly using ESI MS. Elegant
methods for locating phosphorylated fragments and
determining the site of the modification using HPLC/
ESI MS have been described in the literature [IO-14]. If
HPLC fractions are collected for subsequent MALDI
analysis, incorporation of 32P labeling can be used to
indicate which D,'s are phosphopeptides [9]. Mass
mapping techniques [6, 15] are commonly used to
predict the mass-to-charge ratio values of the proton-
ated proteolytic digestion products. If a peak is ob—
served in a MALDI mass spectrum which is 80 Da
higher than a predicted peak, it may indicate phosphor-
ylation. If multiple phosphorylations are present, a
mass shift of 80 for each will be observed. The 80 Da
shift corresponds to addition of HPO3 to the -OH group
of a serine, threonine, or tyrosine residue. To confirm
the presence of the phosphate group, a portion of the
sample can be treated with an enzyme such as alkaline
phosphatase [6]. If the peak observed truly represents a
protonated phosphopeptide, its mass will decrease (by
80 u) upon enzymatic dephosphorylation.

In MALDI MS, an analyte peptide M will usually
yield a peak representing [M + HP ions. However, in
the direct analysis of a mixture of several peptides, such

Received May 6, 1998
Revised September 7, 1998
Accepted September 23, 1998

72

36 ASARA AND ALLISON

as those encountered in digestions, peaks for all com-
ponents are usually not observed—and certainly not
with equal intensities for equally abundant compo-
nents. One reason for this has been attributed to sup-
pression effects [16, 17], in which the presence of one
peptide inhibits the response for another. Thus, some
form of fractionation would seem reasonable to use,
although time consuming, prior to MALDI MS analysis
of digestion mixtures. The situation is particularly prob-
lematic when phosphopeptides are present. Liao et al.
[6] have shown that the desorption/ ionization efficien-
cies for phosphopeptides in MALDI MS are approxi-
mately an order of magnitude lower than for the
nonphosphorylated forms [5]. The situation is worse
when more than one phosphate group is present [6].
These facts limit the utility of MALDI M5 for the
identification of phosphorylation sites. If one uses the
trypsin/ alkaline phosphatase combination, and a pep-
tide containing three phosphates is encountered, for
example, no signal may be detected for the intact
analyte. When the phosphates are removed, the de-
phosphorylated form may be detected, but without
knowing the mass-to—charge ratio values before and
after dephosphorylation, the number of phosphoryla-
tion sites cannot be determined.

Why do phosphorylated peptides exhibit low re-
sponse in MALDI? Presumably the presence of the
phosphate group is exhibiting a direct influence. This
group exists in solution in an anionic form. If the
peptide skeleton is considered in its neutral form, each
phosphate group can carry up to two charges. If a
multiply charged analyte is trapped in the MALDI
matrix crystals in ionic form, insufficient energy is
available in the MALDI experiment for the direct de-
sorption of multiply charged species. For most analytes.
singly charged ions are predominantly formed [18]. A
tetraphosphorylated protein fragment could hold a
charge of -8; certainly the -8 form of the ion would not
be generated in the gas phase in the MALDI experi-
ment. If a low energy chemical pathway does not exist
that allows the analyte to lower its charge, then no
signal will be observed [18]. Similar considerations have
been discussed in the analysis of multiply charged
anionic [l9] and cationic [20] analytes in the fast atom
bombardment (FAB) MS experiment.

This paper discusses the use of matrix additives in
the MALDI experiment to allow for improved detection
of phosphopeptides. Oligonucleotides have been suc-
cessfully analyzed in MALDI MS [21-24]; these analytes
contain a high density of negatively charged phosphate
groups, and can carry a substantial accumulated charge
in solution. In this field of research, ammonium salts
such as ammonium acetate or diammonium citrate are
routinely added, and appear to lower the charge on the
analyte [22-24], making detection possible. A basic
model describing the influence of ammonium ions has
been proposed [25]. Ammonium cations complex with
negatively charged groups to form (NFL,+ ),, (phosphate"').
Either prior to or during the desorption/ ionization

73

I Am Soc Mass Spectrom 1999. 10. 35-44

(D/ I) process, ammonia is lost, Ibaving the charged
group in a neutral form containing protons as the
counter ions, (H*),, (phosphate‘"). It should be noted
that ammonium ions have been implicated as being
protonating (ionizing) agents in the MALDI experiment
as well [24]. We will show here that such additives,
when used correctly, dramatically improve the desorp-
tion/ionization efficiency for phosphOpeptides in
MALDI. The enhanced response in both positive and
negative ion mode suggests that citrate ions scavenge
alkali ions [23], to limit their access to phosphate
groups, whereas the ammonium ions serve to ulti-
mately generate the analyte in neutral form. With the
phosphopeptide available in neutral form, it can be
protonated or deprotonated as are most unmodified
peptides in the MALDI experiment.

This paper demonstrates the utility of matrix addi-
tives with common MALDI matrices for the analysis of
single and multiply phosphorylated peptides, both as
pure analytes and in complex mixtures. Conditions are
presented that can lead to the preferential detection of
phosphorylated peptides in a digestion mixture. With
multiply phosphorylated peptides, signals can be en-
hanced to levels that allow for MS/MS experiments to
be performed.

Experimental
Materials

Bradykinin, B-casein, equine myoglobin, angiotensin I,
and angiotensin III were purchased from Sigma Chem-
ical (St. Louis, MO). The phosphopeptides LKRApSLG-
amide and LKRApTLG-amide were purchased from the
University of Michigan Protein and Carbohydrate Fa-
cility (Ann Arbor, MI). Peptides and proteins were used
as supplied without further desalting or purification.
Trypsin was purchased from Promega (Madison, W1).

A four component equimolar mixture of the peptides
LKRApSLG-amide, LKRApTLG-amide, angiotensin III,
and bradykinin was prepared in 1:1 acetonitrile(ACN)/
water at a concentration of 1 pmol/uL (mixture I) and
diluted to 500 and 100 fmol/uL. Another four compo-
nent mixture of the peptides LKRApSLG-amide,
KIGEGprYGVVYK, angiotensin I, and angiotensin III
was prepared in 1:1 ACN/l-IZO at a concentration of l
pmol/iii. (mixture II) and further diluted to 100 fmol/
51.1.. A two component mixture of B—casein and equine
myoglobin was prepared at a concentration of 10
pmol/pl. in 1:1 ACN/HZO (mixture III). The matrix
used for mixtures I and II was a saturated solution of
a-cyano-4-hydroxycinnamic acid (4-HCCA) (Sigma
Chemical) in 1:1 ACN/HZO. For mixture III, a saturated
solution of 2,5-dihydroxybenzoic acid (DI-IB) (Sigma
Chemical) was employed. In cases where diammonium
citrate (J .T. Baker, Phillipsburg, NJ) or ammonium ace-
tate (Sigma Chemical) was used as an additive, a 1 mM
solution and a 25 mM solution, respectively, was pre-
pared in 1:1 ACN/Hzo. In the text, citric acid will be

J Am Soc Mass Spectrom 1999, 10, 35-44

referred to as H3Cit, so diammonium citrate will be
designated as (NI-I4)2HCit. An incomplete tryptic digest
of B—casein was performed in 40% ACN with a 20:1
excess of B—casein to trypsin. The digest was performed
in a 60 mM NHJ-ICO, (Sigma Chemical) buffer (pl-l =
8.2) and allowed to incubate at 40°C for 24 h. The
reaction was terminated using formic acid (J.T. Baker).
A 10 mL 1:1 ACN/H20 solution was saturated with
4-I-ICCA and the pH measured with an Accumet model
15 pH meter, yielding an average value of 2.45. '

Mass Spectrometry

Linear MALDI mass spectra were recorded on a Per-
Septive Biosystems (Framingham, MA) Voyager Elite
delayed extraction time—of-flight reflectron mass spec-
trometer equipped with a nitrogen laser (337 nm, 3 ns
pulse). For mixtures I and II and the tryptic digest, the
accelerating voltage in the ion source was 20 kV, the
delay time used was 50 ns, the grid voltage was set to
93.0% of the accelerating voltage, and the guide wire
voltage was set to a magnitude of 0.25% of the acceler-
ating voltage. For mixture III, the accelerating voltage in
the ion source was 21.5 kV, the delay time used was 100
ns, the grid voltage was 92.0% of the accelerating
voltage, and the guide wire voltage was set at 0.25%.
Typically, 128—256 laser shots were averaged for each
spectrum. The sample stage used was a gold sample
plate capable of holding 100 targets. The postsource
decay (PSD) experiment was performed in the reflec-
tron mode with an accelerating voltage of 21 kV, delay
time of 100 ns, grid voltage of 70.0%, and guide wire
voltage of 0.15%. The timed ion selector was set at m / z
3124 and three spectra, with mirror voltagezaccelerating
voltage ratios of 1.00, 0.96, and 0.67, were stitched
together to obtain the PSD spectrum.

Sample Preparation

Targets were prepared by mixing 1 iii. of analyte and 1
iii. of matrix on a sample stage and allowing them to air
dry. When ammonium additives were used, 1 uL of the
ammonium salt solution was also mixed with analyte
and matrix solutions on the sample stage.

For the experiment involving a PVDF membrane
(Millipore, Bedford, MA), 1 [LL of the 1 mel/uL
mixture I was spotted onto the membrane and allowed
to air dry. The saturated matrix was then introduced by
the perfusion method described by Strahler et al. [26].
Two pieces of filter paper cut in rectangular strips were
loaded into a shallow well and sealed on the bottom
and all four sides. The filter paper was saturated with
matrix. A piece of PVDF membrane carrying analyte,
identical in size to the filter paper, was then placed on
the filter paper stack such that the matrix solution could
only evaporate through the membrane surface. The
membrane containing matrix crystals was then intro-
duced to the MALDI instrument by placing a fine
conducting grid on t0p, as described previously [26].

PHOSPHOPEI’TIDE DETECTION BY MALDI 37

When an ammonium salt was used, it was introduced
again by the perfusion method—a mixture of saturated
4-HCCA and 25 mM diammonium citrate (1:1 ratio)
was used to saturate the filter paper in this experiment.

Results and Discussion

We show here that the addition of (NI-IozHCit to the
MALDI matrix/analyte target solution just prior to
crystal formation leads to enhanced detection of phos-
phopeptides. The enhancements shown were observed
upon addition of other ammonium salts such as ammo-
nium acetate, but the influence of the citrate salt is
greater. Also, experiments were performed at a variety
of matrix:additive:analyte ratios. Although the effects
shown can be realized with addition of diammonium
citrate using millimolar to micromolar concentrations,
we believe that introduction of 1 [1L of a 1 mM
diammonium citrate solution to the MALDI target
solution (1 [LL matrix solution-+1 jiL analyte) works
well when the matrix is 4-I-ICCA, whereas use of a 25
mM additive solution works well when the matrix is
DHB. These are suggested ammonium salt concentra-
tions for the matrices tested, and for the analysis of
phosphopeptides in the femtomole to low picomole
range. The fact that different concentrations are re-
quired for different matrices may be related to the
varying matrix compound solubilities.

Figure 1a shows the positive ion MALDI spectrum of
a 4-component peptide mixture (mixture I) using
4-HCCA as the matrix. Each component is present at a
level of 500 frnol. Two components are clearly detect-
ed—the peak at m / 2 917.5 represents protonated angio-
tensin III molecules, and the peak at m /2 1060.6 repre-
sents protonated bradykinin molecules. All mass-to-
charge ratio values represent monoisotopic masses
unless otherwise indicated. The other two components
present are singly phosphorylated peptides, and exhibit
substantially smaller signals. The peak for protonated
LKRApSLG-amide is observed at m/z 823.5 and that
for LKRApTLG-amide is at m/z 837.5; both are barely
detectable above the noise. This example shows typical
relative responses for singly phosphorylated peptides
relative to nonphosphorylated peptides of similar mass.
When the target crystals are grown with inclusion of
(NI-lgzl-ICit, the same sample yields the spectrum
shown in Figure 1b. The increase in intensity and
signal-to-noise ratio for the peaks representing the
phosphorylated peptides is typical of our experience
using this additive. The phosphate-containing peptides
now yield the most intense peaks in the spectrum.
Figure 1c shows the spectrum following addition of
(NHQzl-ICit, where each analyte is present at the 100
fmol level. Again, the phosphorylated components are
clearly detectable, whereas they are not detectable
when the additive is not used.

Although not the point of this paper, a comment
should be made on the influence of this additive in
negative ion mode. When analytes are negatively

74

 

38 ASARA AND ALLISON

 

 

.) [w "n". m w
W
W
mm
i)
M was
r W

 

 

 

e) I“ m
t) .
.4
l r I l
m M I” I IN

d8

Figure 1. MALDI mass spectra of mixture 1 in (a) a-cyano-4-
hydroxycinnamic acid and (b, c) a-cyano-4-hydroxycimamic acid
with addition of diammonium citrate.

charged in solution, mass spectrometrists frequently
turn to negative ion modes of mass spectrometry anal-
ysis when desorption/ ionization techniques such as
FAB and MALDI are used. Small signals are observed
for phosphorylated peptides in the negative ion mode
of MALDI MS, however their intensities greatly in-
crease when diammonium citrate is added, similar to
that observed in positive ion mode. This suggests a
common, neutral form of the analyte, as a starting point
in the ionization chemistry, which can be protonated or
deprotonated to yield cationic and anionic gas phase
ions.

Multiply phosphorylated peptides exhibit low re-
sponses in MALDI, and the effect of diammonium
citrate addition is also dramatic. This is shown using
mixture 11. Figure 2a shows the positive ion MALDI
mass spectrum of a four component equimolar mixture
of peptides (1 pmol/ peptide). The peaks representing
protonated angiotensin [II (m/z 917.5) and angiotensin
I (m/z 1296.7) are the only clearly visible peaks in the
spectrum. For comparison with results from Figure 1,
LKRApSLG-amide is a component (77: / 2 823.5). The
fourth component is the diphosphorylated peptide
KICEGprYGVVYK (m/ 2 1474.7). Figure 2b shows the
MALDI spectrum of the same mixture after the addition
of diammonium citrate. The monophosphorylated and
diphosphorylated peptide components are now detect-

I Am Soc Mass Spectrom 1999, 10, 35-44

 

 

 

 

 

 

 

 

 

., [mmlfl‘ Isa-Impala
(An-W
(Ant-Hf
. .
/ lA-slllO- (“mmm’
l Milk //
L v ._4___ w a_;'l A A WA
5 4
i)
l . IW
W
smac-
«murmur
w {4.1-.11 a. AA..- -A JALL..- LL-
1 1 I T
II um I” I“

Figure 2. MALDI mass spectra of mixture II in (a) a-cyano-4-
hydroxycinnamic acid and (b) a-cyam-ll—hydroxycinnamic acid
with addition of diammonium citrate.

able, and these peaks are the most intense. A compari-
son of pairs of spectra, with and without the additive,
clearly indicates which components are phosphory-
lated. Also of note is the decrease in the relative
intensities of the metal ion adducts. The peak represent-
ing protonated angiotensin III is accompanied by an
[M + Na]+ peak (m/z 939.5) and an [M + Cu]+ peak
(m / 2 979.4) in Figure 2a. The 4-HCCA matrix has been
cited as one that ”seems to encourage copper attach-
ment to some peptides" [27]. Metal ion adducts are also
present for angiotensin III. The intensities of the metal
ion adducts of these species, relative to the protonated
forms, decrease upon ammonium ion addition. Similar
effects have been observed in DNA fragment analysis
using MALDI, when ammonium salts are added [22-
24] or when ammonium-charged ion exchange beads
are used [21]. Similar to the data in Figure 1, if the
analysis of this mixture is performed at the 100 fmol
level, the diphosphorylated peptide does not yield a
discernible peak. However addition of (NI-UZHCit re-
sults in a peak at m/z 1474.7 with a signal-to-noise
(S/ N) ratio greater than 6:1.

Membranes such as PVDF are frequently used in the
analysis of proteins and peptides. Samples can be
directly applied, and the membrane then washed to
remove salts, buffers, contaminants, etc. Also, proteins

75

I Am Soc Mass Spectrom 1999, 10, 35—44

or digested protein fragments can be electroblotted
from polyacrylamide gels onto membranes. MALDI
mass spectra can be obtained directly from membranes
[26, 28-31], and (NHJZHCit has the same effect. When
the peptide mixture used to generate the data shown in
Figure 1 was applied to a PVDF membrane, similar
enhancements and displacement of metal ions were
observed (data not shown). In work with membranes, a
perfusion approach is used. The membrane is placed on
a stack of wet filter paper containing the matrix solu-
tion. A mask prohibits solvent evaporation from any-
where except the membrane surface. As matrix and
solvent pass through the membrane, solvent evaporates
from the “top" side, where matrix crystals grow. This
perfusion approach appears to increase analyte avail-
ability for incorporation into the matrix crystals. When
ammonium salts are used, they can be introduced
simultaneously with the matrix by this perfusion ap-
proach. '

The influence of (NH.)2HCit as a MALDI matrix
additive is also observed when larger numbers of
phosphates per peptide are encountered, and for spe-
cies of higher mass. The pentaphosphorylated peptide
B—casein has an average mass-to-charge ratio for [M +
l-I]+ of 23,9843. Although it can be detected using
MALDI [32], without the use of additives, the response
is low. Also, its response can be suppressed by the
presence of other high molecular weight analytes. For
example, even at a level of 10 pmol, B-casein cannot be
detected in the presence of an equimolar amount of
equine myoglobin (avg. mass-to-charge ratio for [M +
H]* 16,9517). This is shown in Figure 3a, using DHB as
the matrix, for mixture 111. 01-18 was chosen because it
gives good results for high mass peptides and proteins
in our laboratory. When 1 [4L of a 25 mM solution of an
ammonium salt is added, the spectrum in Figure 3b is
obtained. The pentaphosphorylated peptide is now
clearly detectable. To obtain the data in Figure 3b,
ammonium acetate was used. In many situations, the
acetate is as effective as the citrate salt.

A comment should be made on changes in ion
intensities when ammonium salts are added. The data
shown here are typical of our experiences to date, in
that the relative intensities of peaks representing phos-
phorylated peptides increase upon addition of the ad-
ditive. In some cases, this represents a real increase in
peak intensity. In other cases, the absolute intensities of
all of the peaks may increase or decrease. We note that
Woods et al. reported the observation that ammonium
salts can influence the response for nonphosphorylated
peptides in MALDI as well [33]. It is not our intention to
suggest that absolute intensities for phosphorylated
peaks alone will always increase upon addition of
ammonium salts.

At this point a comment should be made on a
common additive in MALDI, trifluoroacetic acid (TFA).
TFA is frequently added to increase the solubility of
polypeptides. TFA presumably changes the pH of the
droplet from which crystals are formed, however it is

PHOSPHOPEPTIDE DETECTION BY MALDI 39

 

a) [Mtotbhfllm

 

i MWMJ MW
in

 

 

 

 

+ NW
Mill"
Mimi WAD—Ash.
|Tm I zfim I z‘lu I
all

Figure 3. MALDI mass spectra of 10 pmol/ component of mix-
ture [II in (a) 2,5«dihydroxybenzoic acid and (b) 2,5-dihydroxy-
benzoic acid with addition of ammonium acetate.

also volatile. Acid hydrolysis is always possible when
acidic solutions are used, and it is known that TFA
reacts with amino groups of peptides [27], so it should
be used with caution. However, studies which mea-
sured relative responses for phosphorylated and non-
phosphorylated forms of peptides, showing lower D/ I
efficiencies for the phosphorylated forms, were per-
formed using TFA [5, 6]. Although TFA was not used in
the experiments presented here, it should be made clear
that the addition of diammonium citrate or ammonium
acetate to the matrix/analyte sample also works well in
the presence of TFA.

If the products of a proteolytic digestion of a phos-
phopeptide are analyzed as a mixture in a single
MALDI experiment, the phosphorylated fragments
yield small peaks [5, 6]. Addition of ammonium citrate
will enhance their relative intensities, confirming their
assignments as containing one or more phosphate
groups. This is very helpful in extracting targeted
information. The mass spectrometrist need not spend
time interpreting peaks representing fragments which
do not contain phosphorylation sites, when posttrans-
lational modifications are the structural information
desired. It is critical that this additive enhances the
response for multiply phosphorylated polypeptides,

76

40 ASARA AND ALLISON

because phosphoserine resides frequently are found in
chains [2]; thus, multiply phosphorylated tryptic frag-
ments of large phosphoproteins are to be expected. For
example, shown here is the sequence of B-casein [2, 6].
Its five phosphorylation sites are indicated (all serine
residues), and sites of cleavage by trypsin are desig-
nated by slashes (\).

1 R\ELEELNVPG EIVEpSLpSpSpSE ESITR\
26 INK\I<\I EK\FQpSEEQQQ TEDELQDK\IH
51 PFAQTQSLVY PFPGPIHNSL PQNIP

76 PLTQT PVWPPFLQP EVMGVSK\VK\E
101 AMAPK\HK\EMP FPK\YPVEPFI‘ ESQSL
126 TLTDV ENLHLPLPLL QSWMHQPHQP
151 LPPTVMFPPQ SVLSLSQSK\V LPVPQK
176 K\AVPY PQR\DMPIQAF LLYQEPVLGP
201 vmcprrnv

This peptide was digested using trypsin at a level of 40
pmol, and the crude digest analyzed using MALDI,
with DHB as the matrix. Each spectrum was obtained
from 1/ 8 of the total digestion solution (corresponding to
digestion products from 5 pmol of the protein). A relevant
portion of the resulting mass spectrum, in which frag-
ments containing approximately 25 residues are observed,

 

 

 

 

 

 

 

 

 

a) [in-zoom‘
[no-ml“.
“-25“.
1
KIM-Mir (I-ruu‘
. i '
I " b
l
g . ’(NflthCit
cl
I 1 T I
3000 3m 4000 4300

Figure 4. MALDI mass spectra of a tryptic digestion mixture
from 5 pmol of B-casein in (a) 2,5—dihydroxybenzoic acid and (b)
2.5-dihydroxybenzoic acid with addition of diammonium citrate.
The peaks marked by an asterisk represent phosphorylated com-
ponents.

I Am Soc Mass Spectrom 1999. 10. 35-44

is shown in Figure 4a. The largest peaks in the spectrum
are (average values are listed for all) m/z 3723.5, which
represents the protonated form of fragment [177-209], and
m/z 4485.4 representing [170-209]. Neither is phosphory-
lated. The minor peaks at m/z 3124 [1-25] and m/z 3479.4
[1-28] both represent peptides that contain the four phos-

horylated serine residues at positions 15, 17, 18, and 19.
Figure 4b shows the MALDI analysis of the same tryptic
digest after the addition of (NI-{ozliCit Note the substan-
tialincreaseinthepealsrepresentingthephosphoryiated
components. They become the most intense peaks in the
spectrum.

It should be noted that ammonium bicarbonate was
used as a buffer in the digestion. This volatile buffer
was removed by application of a vacuum to the sample
for 12 h before the MALDI matrix was added. This
allowed the spectrum in Figure 4a to be obtained. Then,
additives were introduced to obtain the spectrum in
Figure 4b. This raises an interesting observation. One
approach following digestion is separation using
HPLC, and analysis of the fractions using MALDI. If
this is done to simplify the mixtures analyzed and to
avoid possible suppression effects, the ammonium
buffer salt is separated from the peptides, and the
response for phosphopeptides is low. If the digestion
products are analyzed directly, and ammonium salts
are used as buffers, one may be more likely to detect the
phosphorylated components because of the influence of
ammonium ions.

An additional advantage of enhanced response, as
displayed in Figure 4b, is that the phosphopeptide
peaks are now sufficiently intense that PSD mass spec-
tra [1, 34, 35] can be obtained with a good 5/ N ratio.
Annan and Carr [1] have shown that the PSD spectra of
singly modified peptides phosphorylated at serine and
threonine eliminate H3130, (loss of 98 u), with a rela-
tively minor loss of HPO3 (-80 ii). For phosphorylated
tyrosine, loss of I-IPO, dominates. If the ions at m/z
3124, [1-25] H“, in the linear spectrum of the unfrac-
tionated tryptic digest mixture are subjected to PSD
analysis, the spectrum that results is shown in Figure 5.
The Spectrum shows four consecutive losses of phos-
phate groups, confirming its assignment.

The fragmentation processes leading to the peaks
observed in Figure 5 warrant comment. Loss of multiple
phosphates must be sequential, and possible scenarios
for forming the observed ions are presented in Figure 6.
The tetraphosphorylated precursor ion is at m/z 3124;
all four phosphorylation sites are serine residues. Loss
of 98 yields the peak at 3M6. There is no peak corre-
sponding to l-‘IPO3 loss at m/z [3124 - 80], and there
could be two reasons for this. The first may be that,
when the internal energy content of the ion is high, the
process leading to loss of H3PO. has a higher rate than
that leading to loss of HPO3, so the former is observed
exclusively. Alternatively, the ion with m/z 3044 may
be formed but may dissociate completely, to undergo
further eliminations. We suggest that the former expla-
nation is correct. Analysis of the entire spectrum shows

77

I Am Soc Mass Spectrom 1999, 10, 35—44

 

.I'
——3124

 

 

 

 

 

 

 

 

 

 

Figure 5. PSD spectrum of m/z 3124, a tryptic digestion product
of B-casein containing four phosphate groups (p).

that phosphate elimination via loss of 98 dominates,
which is consistent with Carr’s observation for phos-
phoserine [1]. However, all four phosphates are obvi-
ously not lost as H,PO.. Three forms of the completely
dephosphorylated ions are present. The peak at m/z
2732 corresponds to [3124 - 4(98)]*, m/z 2749 corre-
sponds to [3124 - 3(98) — 80]* and m/z 2768 to
[3124 - 2(98) - 2(80)]*. As successive eliminations oc—
cur, loss of 80 becomes more competitive with loss of

lent-fi-

Figure 6. The mass-to-charge ratio values of observed fragment
ions resulting from losses of 80 and 98 Da from the tetraphospho-
rylated peptide at m/z 3124. The mus-to—charge ratio values in
parentheses represent peaks not detected. The loss of 98 Da
corresponds to elimination of l-IJPOg loss of 80 indicates loss of
”N3.

PHOSPHOPEPTIDE DETECTION BY MALDI 41

98. With each phosphate loss, internal energies drop
and relative rates change. For the completely dephos-
phorylated forms, relative intensities and mass-to-
charge ratio values suggest that 82% of the phosphates
were eliminated as H3PO‘ and 18% as I-IPO3. Also of
note, there is a small peak at m/z 3106, which corre-
sponds to [3124 - Hp)". This ion also appears to lose
phosphates, and is responsible for the low intensity
peaks at mass-to-charge ratio values below those of the
[3124 - n(98)]* peaks, such as m/z 3008, [(3124 -
H20) - 981*. In the context of the PSD experiments, we
note a publication by Gorman et al. [36], introducing a
mixture of 2,6—dihydroxyacetophenone and diammo-
nium citrate as a MALDI matrix. This matrix was
shown to decrease fragmentation because of phosphate
loss for a phosphopeptide, relative to what is observed
using a-cyano-4-hydroxycinnamic acid as the MALDI
matrix. The ammonium salt may have been playing a
key role in this experiment.

Phosphate elimination occurs at a higher rate than
skeletal bond cleavage, and the available internal en-
ergy is apparently used to eliminate multiple phosphate
groups rather than from sequence-specific fragment
ions. This has been observed previously for a triphos-
phorylated peptide by MALDI [1] and a pentaphospho-
rylated peptide using electrospray [37]. Thus, whereas
the PSD spectrum indicates the number of phosphates
present, it does not also yield the sequence. The sample
should be treated with alkaline phosphatase, and the
sequence determined from PSD analysis of the non-
phosphorylated form. Alternatively, the sample may be
treated with base resulting in fl—elimination of H3PO4.
This may aid in identifying the location of the phos-
phate group as well as obtaining sequence specific
fragment ions [37].

Influence of (NHJZHCit on the MALDI Mass
Spectra of Phosphopeptides: Mechanistic
Considerations

To understand the role of diammonium citrate in this
experiment, consider the order in which compounds
precipitate from the matrix target solution as the vola-
tile solvents evaporate. To do so, initial concentrations
were defined, and the volume decreased in regular
steps. At each step, new concentrations were computed,
and the products of concentrations were computed to
determine the volumes at which solubility limits were
reached. This exercise contains a number of approxima-
tions and assumptions, and is meant to serve a simple
purpose rather than to be a rigorous description of the
system. In the experiment described, the initial solution
has a volume of 3 al., and is described in the sidebar of
Figure 7. It contains matrix molecules, in which it is
nearly saturated. It contains diammonium citrate. As-
sume that NaCl is present. To show the role that the
additive can play, we will assume that NaCl is intro-
duced with the matrix because it is a common impurity

78

42 ASARA AND ALLISON

CONCETIRATIONS VS. VOLUME

 

 

9.1;] m
i: ii l—uh-ur-w’u
outbursts“:
Dim-SOVIU‘I
'0 .. Mun-I‘-
Ma‘i'w'w'l

g In “fir-l.

 

 

 

[NOW A - -h
a - ns,»:-
M391 C - which
Mai-m o - MM.

 

 

 

 

 

 

 

 

 

O . . __/ L
-! 4 -1 cl 4 4. -ll '13 J!
A I C D
U!) MUN! ll)
[VARIATION ————>

 

Figure 7. Plot of concentrations of ions vs. volume of an evapo-
rating sample droplet containing matrix, a phosphopeptide, a
sodium salt, and diammonium citrate. The volumes at which
precipitation of each solid would occur is indicated by points A, B,
C, and D.

in matrices and peptides [38]. The 4-HCCA matrix is
provided from the manufacturer as 97% pure, so we
will make the extreme assumption that the remitting
3% is NaCl, leading to the initial [Na‘] of approxi-
mately 0.7 mM. There is also a phosphopeptide present.
Information on the solubilities of phosphopeptides is
not readily available, so for simplicity we will consider
the phosphopeptide as a phosphate ion. A picomole of
phosphopeptide in a solution of 3 uL results in a
solution which is initially 0.3 uM in phosphopeptide
(phosphate).

Figure 7 suggests how concentrations change as
solvent evaporates and volume decreases. When the
volume decreases by a factor of 3, the solution becomes
saturated in matrix, and precipitation of matrix crystals
begin. This is indicated as point A in Figure 7. Further
reduction in volume should not yield increases in
matrix concentration—it will remain constant and for-
mation of solid matrix will continue. Of note is the pH
of the solution from which other compounds will
precipitate. The pH of a solution saturated in the matrix
4-I-ICCA was measured, yielding a value of 2.45. At this
value, most of the phosphopeptide is expected to be in
the form of R—OPO3I-I‘; hence, the form of phosphate
used in the calculation is HZPO: .

As solvent evaporates further, the species with the
highest concentrations become Na”, NHL and HCitz'.
For the initial concentrations used, and considering the
I<SP [39] of Nazi-ICit to be 43 (in water) and that for
(NH,)2HCit to be larger (345), Nazi-[Cit will precipitate
first, at the point indicated by B in Figure 7. This process
has two consequences. Once the point of saturation is
reached, [Na*] and [HCit‘z'] will remain constant as

] Am Soc Mass Spectrom 1999, 10, 35-44

further solvent evaporates. Also, it indicates that citrate
is effective in ”capturing" sodium ions and forming a
precipitate. Thus, Na+ ions can be effectively removed
during the solvent evaporation process by the HCit”
ions. Further evaporation increases the ammonium ion
concentration to the point where diammonium citrate
precipitates, indicated by C in Figure 7, limiting further
increases in ammonium ion concentration. The last
ionic species in solution is the phosphate (phosphopep-
tide), which can precipitate in the high concentration of
ammonium ions as the ammonium salt, which degrades
at some later point in the experiment via ammonia loss as
explained previously. In this calculation, concentratiors
are never achieved to allow precipitation of sodium phos-
phate, the phosphate only precipitates as the ammonium
salt (K,P (NaHzPO4) = 26, K,Ip (NI-IJ-IZPO‘) - 10]. Al-
though the calculation can be carried to very small
solution volumes, there is a point where the system can
no longer be considered as a solution—where there are
an insufficient number of solvent molecules per solute
molecule. Thus, the sample becomes a solid with par-
tially hydrated species at a volume around points B/ C.
Thus, it is probably not the Kw values of NaHzPO4 and
NI-thPO, that determine which cation is captured by
the phosphate group; this is probably reflecting the
relative availability/concentrations of the ammonium
ions vs. sodium ions.

Obviously, the calculation resulting in Figure 7 is not
rigorous—it was done to assist in considering possibil-
ities. Concentrations, not activities, were used, and ionic
strengths can become considerable during solvent
evaporation. Equilibrium (aqueous) values are of lim-
ited use in such a system when volatile solvents are
rapidly evaporating. The mixed solvent system presum-
ably changes composition during evaporation. Trap-
ping of analyte and salt molecules, in neutral and
charged forms, in precipitating matrix crystals is not
considered but obviously takes place [40-43].

In such a calculation, it is difficult to make general
statements concerning the behavior of polypeptides/
proteins. Solubilities of proteins in aqueous solutions
vary enormously [44]. Solubilities can be enhanced with
mixed solvents, acids, salts, and other solubilizing
agents. Also, proteins can “salt out” at high salt concen-
trations, notably when ammonium salts are used [44,
45]. Ammonium sulfate has been used to help precipi-
tate nonphosphorylated hydrophobic peptides prior to
MALDI analysis [33]. Peptides frequently have solubil-
ities on the order of tens of mg/mL. For example,
angiotensin I, angiotensin II], and myoglobin have
aqueous solubilities of 10, 25, and 20 mg/mL, respec-
tively [46]. Phosphorylation leads to an increase in
solubility in some cases [2, 47].

The calculation suggests a useful model. As solvent
evaporates, matrix first begins to precipitate. If sodium
ions are present in a sufficient amount, citrate ions can
form sodium citrate. This will limit sodium ion concen-
tration. The phosphopeptide can then interact with
ammonium ions in the later stages of evaporation,

79

I Am Soc Mass Spectrom 1999. 10. 35-44

leading to the ultimate "neutralization" of the phos-
phate group(s). Also of note, ammonium acetate fre-
quently was found to be as effective as ammonium
citrate as a matrix additive. The Ks? for sodium acetate
is smaller than that for ammonium acetate, analogous to
the relationship for the citrate salts, so in such systems,
acetate can limit the role of sodium ions as does citrate.
Such calculations can be particularly instructive when
ionic species are involved.

Conclusions

The use of diammonium citrate as an additive in the
MALDI matrix assists in the analysis of phosphopep—
tides as it does for oligonucleotides, presumably by
lowering the charge state of the analyte, allowing for
desorption to occur. The approach is simple, has real
utility, and can be used in a variety of ways. For
example, suppose a valuable polypeptide sample is
subjected to MALDI analysis for a molecular weight
determination, and no peak is detected. It may be
phosphorylated, and is exhibiting a poor response. In
this case, the matrix/analyte target can be removed
from the instrument, and 1-2 ILL of the diammonium
citrate solution added. It has been verified that the
target will redissolve and new crystals will be formed,
incorporating the additive. This may lead to successful
detection, and would suggest that the analyte may be
phosphorylated. If so, the additive increases the prob-
ability that sufficient numbers of ions will be available
to perform further analyses, to determine if phosphate
groups are present, and how many. The use of this
simple additive will allow investigations of more highly
phosphorylated polypeptides to establish their frag-
mentation patterns, and to establish the maximum
number of phosphate groups that could be determined
by PSD or CAD.

Acknowledgments

The authors would like to thank Dr. John R. Strahler for perform-
ing the tryptic digest of B—casein as well as many helpful discus-
sions. Mass spectral data were acquired at the MSU Mass Spec-
trometry Facility, which is partially supported by grant (RR-00480)
from the Biotechnology Research Technology Program of the
National Center for Research Resources of the NIH.

References

1. Annan. K S.; Carr, 5. A. Anal. Chem. 1996, 68, 3413-3421.
2. Weller, M. Protein Phosphorylation; Pion: London, 1979.
3. Annan, R. S.; Carr, 5. A. I. Protein Chem. 1997, 16, 391—402.
4. Yip, T.-T.; Hutchens, T. W. FEBS lett. 1992, 308, 149-153.
5. Craig, A. (3.; Hoeger, C. A.; Miller, C. L.; Goedken, T.; Rivier,
I. E.; Fischer, W. H. Biol. Mass Spectrom. 1994, 23, 519-528.
. Liao, P.-C.; Leykam, I.; Andrews, P. C.; Gage, D. A.; Allison, I.
Anal. Biochem. 1994, 219, 9-20.
7. Carr, S. A.; Huddleston, M. I.; Annan, R. 5. Anal. Biochem. 1996,
239, 180-192.
8. Busman, M.; Schey, K. L; Oatis, I. E., Jr.; Knapp, D. R. I. Am.
Soc. Mass Spectrom. 1996, 7, 243—249.

at

PHOSPHOI’EPTIDE DETECTION BY MALDI 43

9. Kang, 5.; Liao, P.-C.; Gage, D. A.; Esselman, W. I. I. Biol. Chem.
1997, 272, 11588-11596.

10. Huddleston, M. I.; Annan, R. 5.; Bean, M. F.; Carr, 5. A. I. Am.
Soc. Mass Spectrom. 1993, 4, 710-717.

11. Ding, I.; Burkhart, W.; Kassel, D. 8. Rapid Commun. Mass
Spectrom. 1994, 8, 94-98.

12. Hunter, A. P.; Games, D. E. Rapid Commun. Mass Spectrom.
1994. 8, 559-570.

13. Tsutakawa, S. E.; Medzihradszky, K. F.; Flint, A. I.; Burlin-
game, A. L; Koshland, D. E.. Ir. I. Biol. Chem. 1995, 270,
26807-26812.

14. Iaffe, H.; Shetty, K. T.; Pant, H. C. Biochemistry 19”, 37,
3931-3940.

15. 'Yates, I. R., 111; Speicher, 5.: Griffin, P. R.; Hunkapiller, T. Anal.
Biochem. 1993, 214, 397-408.

16. Phillips, D. R.; Karas, M.; Ehring, H.; Hillenkamp, F. Presented
at the 41st ASMS Conference on Mass Spectrometry and
Allied Topics, San Francisco, CA, 1993.

17. Hiller, I.; Levine, G.; Szekely, 6.; Allison, I. Presented at the
43rd ASMS Conference on Mass Spectrometry and Allied
Topics, Atlanta, GA, 1995.

18. Wu, Z.; Biemann, K. lnt. I. Mass Spectrom. Ion Processes 1997,
165, 349-361.

19. Huang, Z.-H.; Shyong, B.-I.; Gage, D. A.; Noon, K. R.; Allison,
I. I. Am. Soc. Mass Spectrom. 1994, 5, 935-948.

20. Asara, I.; Uzelmeier, C. E.; Dunbar, K. R.; Allison, I. Inorg.
Chem. 1998, 37, 1833-1840.

21. Nordhoff, E; Ingendoh, A.; Cramer, R.; Overberg, A.; Stahl, 8.;
Karas, M.; Hillenkamp, F.; Crain, P. F. Rapid Commun. Mass
Spectrom. 1992, 6, 771-776.

22. Currie, G. I.; Yates, I. R., 111. I. Am. Soc. Mass Spectrom. 1993, 4,
955-963.

23. Pieles, U.; Ziircher, W.; Schar, M.; Moser, H. E. Nucleic Acids
Res. 1993, 21, 3191-3196.

24. Zhu, Y. F.; Taranenko, N. I.; Allman, S. L.; Martin, S. A.; Half,
L; Chen. C. H. Rapid Commun. Mass Spectrom. 1996, 10,
1591-1596.

. Murray, K. K. I. Mass Spectrom. 1996, 31, 1203-1215.

. Strahler, I. R.; Smelyanskiy, Y.; Lavine, 6.; Allison, I. lnt. I.
Mass Spectrom. Ion Processes 1997, 169, 111-126.

27. Beavis, R. C.; Chait. B. T. In Methods in Enzymology: High
Resolution Separation and Analysis of Biological Macromolecules;
Hancock, W. 5.; Karger, B. L., Eds; Academic: Boston. 1996;
Vol. 270, pp 519-551.

28. Strupat, K.; Karas, M.; Hillenkamp. F., Eckerskom, C.; Lotts-
peicl't. F. Anal. Chem. 1994, 66, 464-470.

29. Schreiner, M.; Strupat, K.; Lottspeich, F.; Eckerskom, C. Elec-
trophoresis 1996, 17, 954-961.

30. Vestling, M. M.; Fenselau, C. Mass Spectrom. Rev. 1995, 14,
169-178.

31. Zaluzec, E. I.; Gage, D. A.; Allison, I.; Watson, I. T. I. Am. Soc.
Mass Spectrom. 1994, 5, 230-237.

32. Patterson. S. D. Electrophoresis 1995, 16, 1104-1114.

33. Woods, A. 5.; Huang, A. Y. C.; Cotter, R. I.; Pasternack,
G. R.; Pardoll, D. M.; Iaffee, E. M. Anal. Biochem. 1995, 226,
15-25.

34. Griffin, P. R; MacCoss, M. I.; Eng, I. K.; Blevins, R. A.;
Aaronson, I. 5.; Yates, I. R., 111. Rapid Commun. Mass Spectrom.
1995, 9, 1546-1551.

35. Spengler, B. I. Mass Spectrom. 1997, 32, 1019-1036.

36. Gonnan, I. I.; Ferguson, 8. L; Nguyen, T. 8. Rapid Commun.
Mass Spectrom. 1996, 10, 529-536.

37. Resing, K. A.; Johnson, R. S.; Walsh, K. A. Biochemistry 1995,
34, 9477-9487.

38. Liao, P.-C.; Allison, I. I. Mass Spectrom. 1995, 30, 408-423.

Bil}

80

44

39.

41.

ASARA AND ALLISON

K,p values were calculated from aqueous solubility data from
(a) Budavari, S. The Merck Index, 12th ed.; Merck & Co.:
Whitehouse Station, NI, 1996. (b) Dean, I. A. binge's Handbook
of Chemistry, 14th ed.; McCraw-Hill: New York, 1992.

. Doktycz, S. I.; Savickas, P. I.; Krueger, D. A. Rapid Commun.

Mass Spectrom. 1991, 5, 145—148.
Strupat, K.; Karas, M.; Hillenkamp, F. Int. I. Mass Spectrom. Ion
Processes 1991, 111, 89-102.

. Dai, Y.; Whittal, R. M.; Li, L. Anal. Chem. 1996, 68, 2494-2500.

81

43.

45.

I Am Soc Mass Spectrom 1999, 10, 35—44

Beavis, R. C.; Bridson, I. N. I. Phys. D: Appl. Phys. 1993, 26,
442-447.

. Creighton, T. E. Proteins: Structures and Molecular Properties,

2nd ed.; W. H. Freeman: New York, 1993.
Lehninger, A. L. Principles of Biochemistry; Worth: New York,
1982.

. Fluka Chemical Corp.; Milwaukee, WI: 1997/1998 Catalog.
47.

Decker, R. H.; McMahon, N. I. Proc. Soc. Exp. Biol. Med. 1969,
132, 1178-1182.

 

3‘11

 

APPENDIX C

82

And. Oran. 1m. 71, 2808-2870

Enhanced Detection of Oligonucleotides in UV
MALDI MS Using the Tetraamine Spermine as a

Matrix Additive

Jot-alths's-tddolnflllson'

09:74:11:de WSmsUdmsity, EastLanshg,Mdigan48824

Mania-named lasa desorption/inflation mass spec-
u'omeuy(MALDIMS)hmbeenusedmmlyzeoligo-
nucleotides. However, successhnsbeenlimimdbycation
adduction andhighdetectionlimita. Bothoftheseprob-
lemsareduetothehighnetnegativechargedutoligo-
nucleotides carry on the phosphodiester backbone. Co-
man'ixes such as ammonium salts with UV absorbers such
as 3-hydrmrypicolinic acid, 2,4,6-trihydrmwacetophe-
mde-m—Z-thiothymiriehavebeenusedtohnpmve
the spectral quality for oligonucleotides In MALDI MS.
Ormicbaseshavealsobeenusedasco-matrixes;how-
ever, the most popular matrix, 3-hydroxypicolinic acid,
lsnotcompatiblewiththese additives. Wehavefoundthat
dietetr'aaminesperminemamstrixadditivecansuc-
cessfully eliminate cation adduction and lower the dac-
tion limits for DNA in the MALDI experiment, without
having to resort to desalting steps. 'Ihe results argued that
multiply protonated spermine molecules fundion better
than ammonium ions in neutralin'ng oligonucleotides and
displacing alkali cations. Protonated spermine is chemi-
callysimilartoammoniumionssinceitbindstothe
phosphatebackboneandreleaseaprotonstothephos-
plasma-pups. Sperminecanbeusedsuccesafullywiththe
man-ins 6---2-thiothyrnine -d 80% anthranilic acid!
20% nicotinic acid but not with 3—hydroaypicolinic acid.
'I'headditivealaowurlrswellbrtheanalysiaofmaalated
DNA.

Mach-assisted laser desorption] ionization mass spectrometry
(MALDI MS) has been used to analyse oligonucleotides.H
ApplicationsofMAlDlinthisamcanbelimitedbyalkalication
adduction and high detection limits!” In MALDI MS, spectra
for oligonucleotides are “typically acquired from 1 to 10 pmol of
a nucleic acid sample".1 In contrast. low lemtomole levels of DNA
havebeen analyzed using nanoelecu'ospraywith Fr-MS‘The high

'TowbomWahmflbeMI-‘rmaitafliaon.

m
(l)Nordbo&&lGrpéc.F4Roepsbrfl.P.Ms-$mha 1906.15,
67-138.

(2) MW. K K]. Kala Spectra-r. I998. 31, 1203-1215.

(3) Umbxh. P. A. Ala-Spectrom. 26.1996. 15. 297-336.

(4) Rnslrey. M.T.;Juhaaa. P.;Smimov.l.P.:1‘akadi.E.J.;Maflin.S.A.;Hafl.
LA hue. Natl. Acad. Sci. (1.5.4 1996. 93, 4724-4729.

(5) McLafierty. F. W.; Kelleher, N. L; Beg'lcy. T. P; Fridriksson. E. X.; Zubarev.
RMMD.M.CM1. Opin. Oat. Biol 1998.5. 571-578.

ass Anslyflclehamlsn'y, Vol. 71. No. 14, July 15, 1999

 

83

netnegativechargethatoligonucleotideacanaccumulate. dueto
the -l chargeoneschphomhodiestergroupalongtheDNA
badrbonecomplimtestheMAIDIanalysia'I'herearetwoaspects
tothechargestateoftheanalytethatarerelevantinMALDl
analysis.cheanalyteistr'appedinthemau'ixinionicIorm.
desorption may or may not be possible depending on the charge.
Suficientenergyappearstobeunavailableforthedesorption of
multiply charged species in MALDI. which predominantly forms
singlycharged specissflhismltsineitherhighdetectionlhnits
ornosignal atalL lithe multiplychargedanionicanalytelowers
its charge in either the precipitation process or the desorption
process. it frequently does so by extracting protons from the
man'ixand/orfonningadductswithalkalimetal ionssuchasNa+
andK‘tdv’Inthiswathechargestateofthemoleculeislowered,
allovn’ng for desorption; however the signal may be distributed
overmanym/umlues, sinceavarietyotionicspeciesoithetype
[(W‘) 01*).(Na"),(K").l' are formed. For hisb-nm species
thattannotbeisotopically resolved thepreaenceotthesepeaks
results in poor resolution and peak asymmetry. making accurate
rut/a assignments dificult.l his problem also raises detection
limits and frequsatly requires cleanup steps such as RP-l-IPIL,
microdialysis, or cation exchange in the analysis to remove salts
prior to MALDI target formation.l These complications have
limited the ability of MALDI MS ‘to routinely sequence oligo-
nucleotides 15-20 bases long”! It has been shown in fast atom
bombardmentmassspectmmetry (FAB MS), anotherdclorption/
ionization (D/I) technique that produces mainly singly charged
ions, that matrix additives can lower the charge on polyionic
analytes and allow for more eficient D/l of highly charged
analytes. Additives developed at Michigan State University for
specific applications include the use of triflic acid for the analysis
of positively charged organometallic complexes' and the use of
octylnicotinium bromide for the enhanced detection of negatively
charged phosphate-containing biomolecules.’

Highly charged oligonucleotides and phomhopeptides have
beenanalyeedwith improvedD/Ieficiencythroughtheuseof

(6) Wu.2.;Biemann. KM]. Mum loam 1997, 165-166,
349—361.

(7) Henry. C. Anal. Gal. 1997. “SSA-246A.

(8) Asara]. M.; Uzelmeier, C. E:Dunbar.l(.R.;Alliaon.J.Ir-az Gut. 1998.
37. 1833-1840.

(9) Huang. LIL; 91m, 8:]; Gun. D. A.; Noon. K. R.; AliaanJJ. Ant. See.
More W 1994. 5, 935-948.

 

10.1021/ac981406l CCC: $18.00 0 1999 American CW Society
mm on Wfl 05/27/19”

nntrixadditiveam° Matrix additivescanreduceadduct formation
with alkali cations and provide protons to lower the net negative
charge on these complexes. Ammonium acetate was first used to
displace alkali metal ions with NH." successfully in the MALDI
experiment by Currie et al.“ Since then, many applications using
ammonium salts have appeared in the literature“~13 with the most
successful being the use of diammonium hydrogen citrate
(DAHQbyPielesetall‘Armnoniumsaltsareveryuseml
additives since complexed ammonium ions lose ammonia upon
desorption/ionization, leaving wotons behind. DAHC as a matrix
additive in MALDI has been used in combination with my
matrhtes such as 2.4,6—trihydroxyacetophenone (IHAP), Gan-Z
thiothymine (GATT). 80% anthranilic acid/20% nicotin'x: acid (80/
20), and 3hydroxypicolinic acid (El-HPA).l 3-HPAhas become the
most popular matrix for the analysis of DNA.‘1 However. am-
monium ions do not completely displace alkali metal cations, and
theyhavelessofaneffectforhigherniassanalytes.1 Basicneutnl
amines such as triethylamine and imidasole have also been shown
to reduce the formation of alkali metal cation adducts and improve
spectral quality for DNAin MAIDIMSaswell.‘-"1he development
of new additives for DNA analysis by MALDI MS is important.
offeringadvantagessuchasinnnovedresolutionandenhanced
sensitivity.
Herewedemonsu'atetheuseofthetetraamineapermineasa
useful matrix additive for the direct MAIDI analysis of oligo-
nucleotides. without incorporating a desalting step. Spermine.
NH2(CH2)3NH(CH;)4NH(CH2)3NH2. is found in all otganisrnslw
Polyamines such as putrescine, spermidine. and spermine are
essential to all living cells; their depletion has been shown to
inhibit cell growth and affect gene expression." Spermine. in its
tetraprotonated form at cell pH,“ stabilizes DNA by shielding the
negative charges from the phosphate backbone. thus reducing
negative charge repulsion within and between DNA madam”
With completed spermine. DNA condenses into a more compact
structuremHl Divalent cations such as Mg“ have also been
shown to serve this function.""n Polyamines are therefore
regulators of secondary and tertiary structures of nucleic acids”
The X-ray structure of spermine crystallized from a phosphate
buffered solution shows that its amino groups form direct
hydrogen bonds to phosphate oxygens.’o In DNA. spermine
molecules are almost always found in areas of high phosphate
clensity.‘7 However, X-ray crystallographic data reported in recent

 

(10) Aura. I. M.; Allison. 1.]. Ant. Soc. Moss Spectrom. 1999. 10. 35—44.

(11) Carrie. 6.1.; Yam-.1. Rm]. All. Soc. Mos SW 1993. 4955-98.

(12)&u.Y.F.;Taranenko.N.I.;Allman.S.L;Mu1in.S.A.;Ilsl.L:Chut.C.
H. Rapid Coma. Moss SW 1996. 10, 1591-1596.

(13) Ii. Y. C. L; Cheng. S-W; Chan. T.-W. D. Rapid Coma. Moss Spectrom.
1998. 12, sea-soc

(14) Pieles, U.; Ztlrcher. W.; Schlr. M.; Moss. H. E. Nadir“ h. 1993.
21, 3191-3196.

(15) Smmona. T. A; limbach. P. A. Rapid Coats-a. Moss Spectrom. 1997. II.
567—572.

(16) Eoosito. 0.; Del Vecchio. P; Hanna. 6.]. All. Our. Soc. 1997. 119.
2606-2613.

(17) m D. 8.; Stmdlalingam. MJ. Mol Biol 1997. 167, 1171-1185.

(18) Nakai. C.; Glinamaa. W. Bioehasiatry 1977. 16. 566-5641.

(19) Indiah.li.; BaltimD;Berk.A;Zipmaky.S.L;Mataudaira.P.:DarnelL
J. Molecular Cell Biology, Scientific American: New York. 1995; pp 344-
346.

(20) Feuerstein.B.G.:WilEams.LD.:Basu.I-I.S.;MmLI./. cam
1991. 46. 37-47.

(21) Bloomfield. V. A BM 1997. 44, 269-287.

(22) Cohen. S. 5. Nature 1978. 274. 209-210.

yearssuggestthatsperminemayalsobindtoDNAbasestoa
smallextentm°Whilemanysperminemoleculescouldcomplet
with a single DNA strand. crystallized forms of DNAwith only a
small number of spermine molecules attached have been clar-
acterised‘mm This raises the question of whether there are
amnsiveorlimitedinteractionsofspenmnewithDNAlnsohnioa.
Therole ofspermineinreducingthenetchargeofDNAin
sohrfionsuggestsfltathshouldbeevaluatedasmaddifiveto
participateinasimihrprocessintheMAIDIuperimentltshould
alsobenotedthatspennineisfrequentlyaddedtoaohnionshom
whicholigonucleotidesarecr-ystallisedJnsomecaseaitsples»
case is required foraystal growthPM although thepolyamine
isaotincorporatedintotliea'ystaltltatisfornledSinceu'ystal
formafionispmtoftheMAIDlprocesstheroleofsperminein
this experiment may provide information on the extent of spec
mine-DNA interactions in the solution from which the MAIDI
targetisformed. Sperminehasbeencitedinthepastasa
contaminantintheMAIDlanalysisofoligonucleotides.“although
thisisnotwrprisingsinceB-HPAwasthematrixusedinthat
smdy:ithasbeenfoundthat3-IIPAasamatrbtisincompatible
withorganicbases.“Whensperminewasconsideredasan
imptn-ity,itwasreportedthat.foraparticular21-mer.theintendty
ofaapennineadductoftheanalytematchedthatofthetwotonated
analyteintheMAIDlspectrumataspermineconcentr-ationof
10mM.Inthespecu-ashownhae.usingothermanixes.spa1nine
adductsarenotobserved.ltwillbeshownthatspermine.when
usedasamatrbtadditiveinMAIleortheanalysisofoligomrcle
otides. an completely eliminate cation adduction and lower
detection limits. The spermine molecule functions much like
ammonium ions since it appears .to bind to the phosphate
backbone and release protons to the phosphate groups. either in
the cysts] growth procsss or in the subsequent desorption]
ionisationstepofMAIDl.’lhisadditivecanbeusedwiththose
most commonly used MAIDI matrixss for oligonucleotide analy-
sisnotably6A'I'I‘and80/20thillalsobeshownthattheMAIDl
spectrumofmetalatedDNAcanbeenhancedusingspermineas
well.

”MINT“ DETAILS

The 12-mer oligonucleotide d(CCI‘CICGI‘CICC) was pur-
chased from the Yale University Oligonucleotide Synthesis lihcility
(New Haven, C'D. 'Ihe 17-mer sequencing primer d(GTAAAAC-
GACGGCCAG’D was purchased from Boehringer Mannheim
(Indianapolis, IN). The platinated 12-mer d(CCTCT{GGPt‘*}-
’I‘C'ICC) wasagiflfromDrJGmRDunbaratMlchiganState
University. All oligonucleotides were desalted via size exclusion
chromatography using a Sephadex 025 Fine (Phat-triads Biotech.
Piscataway. NJ) column. The oligonucleotides were [separed in
wamr at a stock concentration of 5 pmol/pl. and then further
diluted to 1 pmol/pl. and 75 fmol/uL An aliquot of d(CCTCTG-
GTCTCC) was also prepared at a concentration of 1 pmol/pl. lnlO
mM NaCl. An aliquot of d(CCI‘Cl‘{GGPt”}‘ICI‘CC) was prepared
in 20 mM sodium acetate at a concentration of l pmol/al. 'Ihe
d(CCTCTGGICI‘CC) strandwasdigastedinwaterwith ~50units
ofexonuclease III. a3’-5’exonuclease isolatedfromE. coliThe

 

(23)MM4WLD;G&QJMLW199LM.II3BS-
11402.

(24) Staler.T.A;Wickham.J.N.;Sannes.K.MWu.K.J.;Becker.C.II.Ansl.
Oat. 1996. 68. 576-579.

”snowman. Vol. 71, No. 14, July 15. 1999 ass?

84

 

 

 

 

 

a) I'M-Hr ZHT Wit
r ma
'” W
i . um
s) M
‘ M11
& 17 g I ,Yu ‘ ; f g
u:

Flume 1. MALDI mass spectra of d(GTMAACGACGGCCAG‘I‘)
intheprsssncsot 10mMNaCIh(s)3—hydrnrrypiooliricecldwim
diammonium hydrogen citrate. (b) Mama-Wis with diammo-
nium hydrogen citrate. and (c) O-aza-Z-thlothymtns with spsnnirie

digestionwasmonitoredevery30minfor6handthenst12h
and24h1henratr'ixes3-hydroaypicolinicscidfi-an-2-thiothym-
irremtlrranflicaddandnicotinicacidwerepm'chasedfmmSigrna
Chemical (St. Louis, MO). When spermine (Fluka. Milwaukee.
WI) was used as an additive. it was prepared at a concentration
of12.5 mM in saturated matrix solution (1/1 acetonitrile/water).
When diammonium hydrogen citrate O. T. Baker. Phillipsburg,
NDwasused.itwaspreparedataconcenu-ationof25mMin
1/ l scetonitrile/ water in saturated matrix solution. Samples were
prepared by mixing 1 [al. of analyte solution and lpl. of co-matrix
solution and allowed to air-dry on a gold sample plate.
linearMAlDImassspectr-awererecordedonal’erseptive
Biosystems (Framingham, MA) Voyager Elite delayed extraction
timeof-tlight reflectron mass spectrometer equipped with a
nitrogen laser (37 run, 3 as pulse). The accelerating voltage was
23 kV, the delay time selected was 50 ns, the grid voltage was set
to 93.0% of the accelerating voltage, and the guide wire voltage
was set to a magnitude of 0.25% of the accelerating voltage.
Typically, 64-128 laser shots were averaged for each spectrum
Allspectrashownweretakeninthenegative-ionmode.

RESULTS AND 03608810"

We show here that the addition of spermine to the MALDI
matrix solution results in improved spectra for singlestranded
oligonucleotides and that it eliminates the need for a desalting
step when such analytes are under study. Figure 1a shows the
MALDI spectrum of I pmol of a 17-mer deoxyn'bonucleotide using
3-HPA/DAHC as the co-matrix. This is a commonly used matrix
for such analytes. To demonstrate the complications that arise
when salts are present, the sample was introduced in a solution
thatwasalso lOmM in NaCl. Notethatthesmmoniumsaltcannot

sass W W, Vol. 71. No. 14. M 15, 1999

85

succemfully displace all of the sodium ions for the high concentra-
tionofNaClusedandthelargenumberofpotentialsimsatwhich
Na+ could form a complex. For thisdiscussion. M will be used to
indiuteflieneuualfonnofflreanalyteinwhichallofthenegative
charges have been neutralized by protons (oligo" + all“). ‘Ihus.
iftheanalyteformsanegativelychargedspeciesintheMAlDI
experiment which contains one Na‘, it would lose two Vetoes.
hence the designation of [M + Na — Zill‘. Figure lb shows the
MALDI spectrum obtained for the same l7-mer, with another
commonly used matrix combination. GATT/DAHC. While the
rangeofsodium ion adductsmaybe smallerthanthatshownin
Figure 1a. the resolution is lower. This results in a very broad
signalthatspansnearly lOODaduetoeationadductionJigrn'e
lcshowsthespecn'umwhenGA'IT/spermineisusedasthe
matrix. Even at the large salt concentration. spermine is efiective.
much more so than ammonium ions. in displacing Na+ ions and
providing protons to negatively charged phomhate groups for
neutralization-yielding a single peak for the deprotonated form
of the neutral molecule at rut/r 5223. Without alkali metal ion
adducts. the highest resolution of the instrument can be realised.
andmostnreaningfrdar/zvalueseanbedeterrninedThecapsbfl-
itytosimplyeliminateeationintionisveryimportantespedally
when mbrtures of oligonucleotides with similar molecular weights
sreundersmdynnextensiveseriesofalkalimetalionadducts
can mask the presence of a second component. The result of
spermine addition is also an increase in the signal-to-noise (,S/N)
ratio of the experiment. Spermine is not as eflective with 3-l-IPA;
therefore, other rnatrixes such as GATT must be used. For 6ATI‘.
spermine clearly is more effective than ammonium salts.

The interaction of metal ions such as platinum with DNA is
important to be able to study. since the anticancer activity ofcis-
platin has been related to strong DNA-Pt interactions.” 0f
cornse,forMAIDItobeusefulinstudyingsuchinteractions.the
complexmustremainintactduringalloftheprocessesassodated
with the analysis That is, the metal should not be displaced from
the oligonucleotide by either matrix molecules or additives. ane
2a shows the spectrum ofa platinated DNA strand using 3-IIPA/
DAHC as the co-rnatrixes The spectrum shows thatmany sodium
adductsuefonnedlnthiscasethesamplewaswepared/
supplied for analysis in a sodium acetatebuflered solution. While
thespacingofthepeaksclearlyindicatesthatsodiumionadducts
arebeingformed,eveninthepresenceofammoniumions,one
cannotassurne that the lowest rut/speak in the cluster represents
the [M - H1” ions. Again, the alkali metal ion adduct forrmtion
makes the detection of additional species of similar mass dificult.
Figure 2b shows the MALDI spectrum of the same sample using
the 80/20/DAHC matrix. Apparently. with this matrix. ammonium
ionsaremoreefiectivein displacing sodiumions. Ithasbeen
shown previously that the 80/20 matrix is a good choice when
one attempts to detect platinated DNA." Figure 2c shows the
MALDI spectrum of the same platinated DNA compound using
80/23 and spermine. The sodium adducts have completely been
eliminatedbyspermine.givingasharppeakrepresenting [M-l-
Pt“(NH3): - 3!!!" at m/z 3773. A peak for the free DNA is also
seenasthe [M — Hl‘ muster/23546. Webelievethatthisis

(25) Suntan. S. 8.: Lin-d. S. J. Gas Res. 1987. 87. 118—1181.

(26) Reediik. J. In. as... Add 1992. 1913—200. 873-881.

(27) Gonnet, F4 Kocher, F; BlalsJ. C.; Bolbach, GgstelJ. C; M]. C.
I. MW 1996. 31. 802-809.

 

 

c) WW

 

 

 

I I I I I T I

seen sass

Figure 2. MALDI mass spectra of d(CCTCT{GGPt’*}TCTCC) in
(a)3-Wypicoinicaddwithdlammirmhydmgenctfiats.(b)80%
mnmucodricscldwlmdlammuumhydrogsncim.
and (c) 80% anthranilic «moss nicotinic acid with spennlns.

indicative ofmincompleterescfionratherthananindication of
an equilibrium that exists between the platinated and platinum-
free forms. since other spectra have been obtained for the same
metal complex. more highly purified by HPLC, in which the peak
at rut/a 3546 is not present. In working with this Pt-DNA complex.
wehavefound thatthe 80/20matrixissuperiort06A'IT, interms
of allowing the metalated complex to remain intact. This is
consistentwitltdatareportedbyGonnetetalnAcommentshmld
be made that. using other matrixes and excitation wavelengths.
PtOV'IiQ't-DNA complexes have been shown to exhibit mono-
and diammine losses.” This is not observed for the systems
studied here. While the data shown are typical for platinum
complexes, using the two matrixes and two additives shown,
complexes involving other metals are much more sensitive to the
chemicalnamreofthematrixP'IhedatashowninFIgmeZshow
thatspermineasaco-matrixeliminatestheneedfordesalting
and purification procedures.

Spermine. particularly with the GATT matrix, also allows for
detection of oligonucleotides at lower levels than with DAHC.
Without spermine, even with purified samples that are Na+ free.
detection limits are higher. Figure 3a shows a portion of the
MALDI spectrum of 75 final of a desalted lZ-mer deoxyribonucle-
otide using the most common matrix combination of 3-HPA and
diammonium hydrogen citrate. Using our MALDI-10F instru-
ment. no signal is obtained at this low femtomole concentration.
Figure 3b shows the MALDI spectrum of the same 12-mer using

 

(20MCENN6ME4HMEIIL1MMWIOI
m1994.132.239—249.

(29)Aaua.].M.:Dtmbar.K.R.;Allison.J.Unpubliahedmlts.

(30)Guo.L-H;WtaR.lnMMiu£amColowi&S.P,Kaplan.N.
0.ch New York. 133; Vol. 1m.pfio.

86

 

 

 

 

 

 

Elgar-e3. MALDlmassspsctrsotntmolotdtcm'CTGGTCTOC)
In(a)3-hyaoaypicolirucacldwithdammonlunhydrogsncltratsand
(b)6-aza-2othiothyminswithspennine.

the GATT/spermine combination. There is a substantial increase
in desorptionl' ionization eficiency obsa'ved. It is also noteworthy
thatforthisparticularsample, the 75fmolofanalyteisdeposited
with 12.5mlofspermine. Evenwhenthesperrnine/analyteratio
is ~105. this additive interacts with multiple sites and provides
protons to lower the charge state. but the interactions are
sufficientlyweakthatallinteractingspermine moleculesarelost
atsomepointintheMAlDlprocess'Ihatis,thereisnopeak
representingtheadductsofoneormoresperminemoleurleswith
theanalyte (eachsperminewouldsddamassofMDa).

We have found the 6-ATT/spermine combination to be useful
in the analysis of mixtures of oligonucleotides as well. We had
been doing an experiment to study the utility of exonuclease III.
This enzyme is typically used to digest double-stranded DNA
startingatthe3’end.”l"rgure4showstheresultsofanexo-
nuclease Ill digestion of the same lz-mer used in Figure 3. from
onepointthintoadigestionInthisexpcrimenclpmolof
desalted DNA was being digested and onefourth of the sample
wasmixed withthe common 3-HPA/DHAC matrixmixhneThus.
withspprmrimatelySOfmolofeach digestion productpresenttwo
or three components could be detected. as shown in Figure is. If
the same sample amount is analyzed using the GATT/spermine
combination. more components of the mixture are detected. and
at higher S/N ratios. Much of the sequence can be determined
usingthepealrsshownltisalsonoteworthythattherelafive
response apparently is different for the two matrix mixtures as
well.Also,noticethepeakstsr/11840is98Dahigherthanthe
peaklabeled G,representingli;l’0t.1hispeakmaybeduetoan
incomplete digestion product.

Finally, a comment should be made on the utility of spermine
as an additive. Since spermine is not compatible with 3-l-IPA. the
most effective matrix for oligonucleotide applications, we suggest
thattherealstrengthofusingspermineisintheanalysisof

AW W, Val. 71. NO. 14. Jtly 75. 1999 use

1%!“

 

m-
”A

 

 

 

 

Figs-.4. MALDImassspsctraoi~50tmoioieachcomponentot
aa'emrcleasedigestionmirdusmtalsmminitiafly-mhnol)
ddtOCTCTGGTCTCQIna-hydmxypicoiinlcscidwithdiammoruum
hydrogen citrate and in B-aaa-z-wothymine with spennine.

mplesatlpmolorbelow.withoutadmltingstep.Whenhrger
quantitiesofDNA(>lpmol) arecharacter'ued.both3-HPAINH.+
and GATT/spermine may be reasonable matrix choices. depend-
ing on the saltcontentand truss ofthe oligonucleotide.

CONCLUSION
lnthepasttwodeeades.therolethatpolyaminesplayinliving
cellsatthemolecularievelhasbeenanimportanttopicof
research. Since spermineassisted crystallization leads to the
neutral (protonated) form of DNA. it follows that proton transfer
fromprotonatedsperminenitrogenstonegativelychargedphos-

2am mammary, Vol. 71, No. 14, July 15, 1999

phategroupstakesplweintheaystallintionstepoftheMAlDl
experiment. nottlredesorptioneventlfsperminemolemhs
remainattachedafteru'ystallintioniscompleteJheyarelostin
the desorption step.
ltisnotsurprisingthatspermineismoreefiectivethan
ammonium ions in neutralizing oligonucleotides. In the dynamics
of complexation of positively charged protonated nitrogen centers
with negatively charged phosphates. multidentate ligands are
expectedtoyieldthemorestahleproductsandtohavehigher
fornntion constants-analogous to simpler systems involving
chelatingagenfl'lhefactthatthemassspectralresultashowsuch
complete formation of the deprotonated form of the DNAanalyte
suggests that spermine-oligonucleotide intaactionsareextenaive
and complete. Spermine appears to sample all phosphate groups
sndprovideprotonstotheaesitesSuchdatamaybeusefulin
modelingpolyamine-nucleicacidcomplexes,certainlysuggesting
morethanapasfivecountaionroleforpmtonatedpolyamines.

WHEN?
'IheauthorsthankDrJarneleGeiger-forsuggestingthe

useofpolyaminesin MALDIanalysisand Dr. KimR. Dunbarand
Elinbeth Lozada for desalting the oligonucleotides and for
providing the cisplatin oligonucleotide complex Mass spectral
datawereacquired attheMSUMassSpectrometry Facility.which
is partially supported by Grant 111200480 from the Biotechnology
Research'l‘echnologyProgramoftheNationalCenter-forlle
search Resources of the NIH.

Received for review December 18. 1998. Amepted April 9.
1999.

A0981 406L

 

MICHIGaN STATE UNIV. LIsRnRIEs
WWW”llllWI“llllllllllllm“IIHIIHHHI
31293020509695