Characterization of peptide-oligonucleotide heteroconjugates by mass
spectrometry
Characterization of peptide-oligonucleotide heteroconjugates by mass spectrometry
Ole N.
Jensen
+
,
Sandhya
Kulkarni1,
w
,
Jane V.
Aldrich1,
W
and
Douglas F.
Barofsky
2,
*
Department of Biochemistry and Biophysics,
1
College of Pharmacy,
2
Department of Agricultural Chemistry and Environmental Health Sciences Center,
Oregon State University, Corvallis, OR 97331-7301,
USA
Received April 8, 1996;
Revised and Accepted August 16, 1996
ABSTRACT
Two peptide-oligothymidylic acids, prepared by joining an 11 residue synthetic peptide containing one internal carboxyl group (Asp
side chain) to amino-linker-5
'
pdT
6
and amino-linker-5
'
pdT
10
oligonucleotides, were analyzed by matrix-assisted laser desorption/ionization (MALDI) on a linear time-of-flight mass spectrometer and by electrospray ionization (ESI)
on a triple-quadrupole system. These synthetic compounds model peptide -nucleic acid heteroconjugates encountered in antisense research and in
studies that use photochemical crosslinking to investigate molecular aspects of protein-nucleic acid interactions. MALDI and ESI sensitivities for the two hybrid
compounds were found to be similar respectively to their sensitivities for the
pure oligonucleotide parts. In general, MALDI proved to be less affected by
sample impurities and more sensitive than ESI, while ESI on the quadrupole
produced greater mass accuracy and resolution than MALDI on the time-of-flight instrument. A hybrid's behavior in a MALDI-matrix or an ESI-spray-solvent was found to be governed mainly by the
oligonucleotide. A single positive ESI tandem mass spectrum of the peptide-dT
6
accounted for the heteroconjugate's entire primary structure including the point of the oligonucleotide's covalent attachment to the peptide.
INTRODUCTION
Covalent protein-nucleic acid conjugates occur in nature, for example, in the genome of
certain RNA and DNA viruses (
1
), in DNA topoisomerases (
2
), or as amino acid-RNA conjugates in aminoacyl-tRNAs (
3
). Covalent nucleoprotein complexes can be produced in the laboratory by
irradiating cells or protein-nucleic acid mixtures with ultraviolet (UV) light (
4
-
6
). The bonds induced in such a photochemical reaction form at or near points of
contact between protein and nucleic acid; thus, identification of the amino
acids that are crosslinked to nucleobases can provide information about the
nucleic acid binding domain of the protein (
7
). Peptide-oligonucleotide heteroconjugates have recently been synthesized for use in antisense research; the peptidic components of these hybrid complexes are thought to help stabilize, transport or
target antisense oligonucleotides (
8
-
13
).
Recently, Jensen
et al
. described a protocol for studying protein- nucleic acid interactions (
14
,
15
) that makes extensive use of matrix-assisted laser desorption/ionization (MALDI) (
16
,
17
) and electrospray ionization (ESI) (
18
,
19
) mass spectrometric techniques to analyze photochemically crosslinked protein-nucleic acid complexes. Bioanalytical applications of MALDI and ESI
techniques have been reviewed (
20
-
22
). MALDI and ESI mass analyses of a covalent peptide-oligonucleotide hybrid pose special problems because the two polymeric components generally have conflicting
requirements for ionization. Some of these difficulties were overcome and
certain others were delineated in recent investigations of various UV-crosslinked protein-nucleic acid systems (
14
,
15
,
23
).
In order to further develop mass analysis as a tool for investigating covalent nucleopeptides, we have chemically synthesized peptide- oligothymidylic acid conjugates and examined their behavior under MALDI
and ESI mass spectrometric conditions. We report our observations and discuss
their analytical import in this paper.
MATERIALS AND METHODS
Synthesis of peptide-oligonucleotide hybrids
The peptide was assembled on a PAL
®
resin (Millipore, 0.5 g) using amino acids protected with 9-fluorenylmethoxycarbonyl and a Biosearch 9500 automated peptide synthesizer as previously described (
24
), except that
N
,
N
-dimethylacetamide (HPLC grade, Aldrich) was used in place of
N
,
N
-dimethylformamide (DMF). The side chains of Arg and Asp were protected respectively with the
groups 2,2,5,7,8-pentamethylchroman-6-sulfonyl and t-butyl. After assembly and removal of the 9-fluorenylmethoxycarbonyl group off the N-terminus, the peptide was treated with a 4-fold excess of acetyl 1-imidazole (0.4 M) for 2 h. The
resin was washed and dried; then the peptide was released by treating the resin
with trifluoroacetic acid (TFA) containing scavengers, Reagent K (
25
), for 3 h and isolated by a previously described method (
24
).
The peptide was purified on a Beckman Model 100A HPLC system equipped with a 4.6
mm * 250 mm, 5 [mu]m particle, 300 Å pore, Vydac C
18
column (The Separations Group) and a Waters Model 220 UV-detector. The mobile phase comprised (A) 0.1% TFA in water and (B) 0.08%
TFA in acetonitrile (Burdick and Jackson); solvent was delivered at 1 ml/min
with a gradient of 5-35% B for the first 10 min, 35-65% B for the next 30 min, and 60-90% B for the last 5 min. Detection was at 220 nm. The
peptide eluted at ~33.5 min.
The purified peptide was coupled to the aminolinker-oligodeoxythymidylic acids NH
2
-(CH
2
)
6
-5'-p(dT)
6
and NH
2
-(CH
2
)
6
-5'- p(dT)
10
(aminolinker-dT
6
and aminolinker-dT
10
respectively; gifts from Oligos Etc., Inc., Wilsonville, Oregon) in separate
reactions carried out in DMF (Burdick and Jackson) at room temperature. The DMF stock solution was sparged with N
2
for 2 h prior to use. HPLC-purified peptide (1.57 mg, 1.16 [mu]mol) in 100 [mu]l DMF was sonicated briefly in a 1.5 ml microcentrifuge tube.
N
,
N
-Diisopropylethylamine (Millipore; 10% in DMF, 8.5 [mu]l), 1,2-benzotriazolyloxy-tris[pyrrolidino]phosphonium hexafluorophosphate (PyBOP; NovaBiochem; 100 mg/ml in DMF, 11.8 [mu]l), and 1-hydroxybenzotriazole (50 mg/ml in DMF, 7 [mu]l) respectively were added to the peptide solution, and the
mixture was allowed to react for 10 min. The peptide cocktail was then
transferred to another microcentrifuge vial containing 820 [mu]g of aminolinker-oligothymidylic acid dissolved in 73 [mu]l DMF. The reaction solution was vortexed at room temperature for
incubation periods that varied from 4 h to 3 days. Samples of the reaction
mixture were analyzed at various intervals by MALDI mass spectrometry and microbore reverse-phase HPLC. The chromatography system consisted of a Perkin-Elmer ABI 140B syringe pump, a 1 mm * 250 mm, 5 [mu]m particle, 300 Å pore, Vydac C
18
column (The Separations Group), and an ABI/Kratos Model 783 UV-detector equipped with a microbore flow cell (1 [mu]l dead volume). The solvent system was (A) 5 mM ammonium acetate and (B) acetonitrile; solvent
was delivered at 40 [mu]l/min with a gradient of 5-90% B over 45 min. The eluted species were monitored at 256 nm. The retention times of individually chromatographed synthetic peptide, aminolinker-dT
6
, aminolinker-dT
10
and PyBOP were used to locate the peaks corresponding to these primary reagents
in the chromatograms of the reaction mixtures.
The crude peptide-linker-oligothymidylic acids were purified on the same HPLC system used to purify the synthetic peptide, but the solvents and gradient employed were those used on the microbore HPLC system to monitor the coupling reaction (see preceding paragraph).
The flowrate was 1 ml/min, and the heteroconjugates were monitored at 256 nm.
Matrix-assisted laser desorption/ionization mass spectrometry
MALDI mass analyses were performed on a custom-built, linear time-of-flight mass spectrometer. Both the instrument and the
operating conditions have been described elsewhere (
15
). Each mass spectrum was produced from 30 consecutive pulses of 355 nm photons from a Nd:YAG laser (Spectra Physics GCR-11). Signals from matrix ions or ions from various standards added to the
sample were used for mass calibration.
Prior to being mass analyzed, aliquots from the reaction mixtures were diluted 100-fold with Millipore-filtered water, and HPLC purified samples were lyophilized and redissolved in 50-200 [mu]l water/acetonitrile (1:1, v/v). A saturated solution of
4-hydroxy- [alpha]-cyanocinnamic acid (Aldrich) in 0.1% TFA/33% acetonitrile (2:1, v/v) (
26
) was used as the laser desorption/ionization matrix for the synthetic peptide, and a saturated solution of 2,4,6-trihydroxyacetophenone (Aldrich) in 50 mM diammonium hydrogen citrate/50% acetonitrile (
27
) was used as the matrix for the aminolinker-oligothymidylic acids, the crude reaction mixtures, and the HPLC-purified peptide-linker-oligothymidylic acids. Before depositing a sample onto the
sample introduction probe, the sample stage was seeded with a layer of matrix
crystallites by depositing 0.5 [mu]l of matrix solution onto it and then, at the first sign of crystal
formation (10-15 s after deposition when viewed under a stereo microscope), gently
wiping the droplet off with a Kimwipe tissue. Subsequently, a 0.5 [mu]l droplet of analyte/matrix solution (1:4 or 1:9, v/v) was deposited on top
of the seed layer and allowed to dry. In some cases, the analyte/matrix
deposits were rinsed by gently dipping the sample stage into water before the
sample had dried.
Electrospray ionization mass spectrometry
Pneumatically-assisted ESI (ionspray) mass analyses were performed on a Perkin-Elmer Sciex API III Plus triple quadrupole mass spectrometer system.
The potential of the electrospray needle was placed at 5000 V or -4100 V to produce positive or negative ions respectively, and the
potential of the orifice leading into the mass analyzer was set at 60-110 V or -110 V to extract positive or negative ions, respectively. The flow
rates for the nebulizer gas (oxygen) and the curtain gas (nitrogen) were both
set at 0.6 l/min for either positive or negative ions. The instrument was mass-calibrated in the positive ionization mode with a mixture of polypropylene
glycols.
Samples were injected at 3.3 [mu]l/min with a Harvard syringe pump connected to the electrospray interface
via a Rheodyne 8125 injector (5 [mu]l sample injection loop). The synthetic peptide was analyzed in an aqueous
solution of 0.1% formic acid/50% methanol; the aminolinker-oligothymidylic
acids, the crude reaction mixtures, and the peptide-linker-oligothymidylic acid heteroconjugates were analyzed in a solution of 5 mM ammonium acetate/50% acetonitrile.
The analyzing section of the Sciex API III consists of two quadrupole mass filters separated by an open quadrupole structure that serves as a gas collision chamber. Single stage mass analysis was conducted by scanning the first quadrupole mass filter over the mass-to-charge range of interest (typically
m/z
400-2200) in ~4 s. Double stage or tandem mass spectrometry was carried out by setting
the first quadrupole mass filter to select a doubly charged analyte ion and
scanning the second quadrupole mass filter over the mass-to-charge range of interest (
m/z
50-1300 for peptide,
m/z
50-2000 for aminolinker-dT
6
, and
m/z
50-2400 for peptide-linker-dT
6
) in ~4 s. Fragmentation was induced in the middle quadrupole chamber by introducing argon as a collision gas (mean collision gas density/collision energy: 125/24 eV for peptide; 110/41 eV for aminolinker-dT
6
; 185/22 eV for peptide-linker-dT
6
).
RESULTS
Synthesis of heteroconjugates
The synthetic peptide
1
was patterned after the sequence of the 11-residue N-terminal segment of Ung-T19
2
, one of the tryptic peptides of
Escherichia coli
Uracil-DNA glycosylase (Ung) that has been shown to form part of the protein's
DNA binding domain (
23
). The peptide was designed to possess one free carboxyl group by replacing the
Asn residue in T19 with an Asp residue. To simplify the synthesis of the
peptide-nucleic acid conjugates, the Cys and His amino acids in
2
, whose side chains contain potentially reactive functionalities, were replaced
with the less reactive residues Ala and Arg respectively. To prevent side
reactions during conjugation of the peptide to the nucleic acid, the N-terminal was blocked by acetylation, and the C-terminal carboxyl group was amidated. MALDI and ESI mass analyses confirmed that the synthesized peptide had the predicted molecular mass (Table
1
), and ESI tandem mass spectrometry confirmed that it had the expected sequence
(data not shown).
Ac-Gly-Phe-Phe-Gly-Ala-Asp-Arg-Phe-Val-Leu-Ala-NH
2
(
M
r
1240.4)
1
Gly-Phe-Phe-Gly-Cys-Asn-His-Phe-Val-Leu-Ala
2
Oligothymidylic acids were used in the synthesis of the heteroconjugates partly
because they were available as gifts and partly because they are generally more
stable and better behaved in mass spectrometric analyses. The experimentally
determined masses of the two oligonucleotide reactants (Table
1
) were consistent with the structures NH
2
-(CH
2
)
6
-5'-p(dT)
6
and NH
2
-(CH
2
)
6
-5'- p(dT)
10
specified by the supplier, Oligos Etc., Inc.
PyBOP was used to catalyze formation of the amide linkage between the carboxyl
group on the side chain of the aspartic acid in
1
and the aminolinker-oligothymidylic acid because of its known efficiency in promoting peptide amide bonds during peptide synthesis (
28
).
.
Average molecular masses of the synthetic peptide, aminolinker- oligothymidylic
acids and peptide-linker-oligothymidylic acids
Molecular mass (Da)
Compound
Theoretical
MALDI
ESI
Peptide
1240.4
1240.9
1240.4
Aminolinker-dT
6
1942.4
1942.4
1942.0
Aminolinker-dT
10
3159.2
3156.6
3158.0
Peptide-linker-dT
6
3164.8
3164.4
3164.3
Peptide-linker-dT
10
4381.6
4379.6
4382.0
The theoretical values are calculated from molecular formulas. The experimental values listed under MALDI and ESI are averages over two or more measurements; accuracies (relative to the theoretical values) are within 0.08% for MALDI and
0.02% for ESI.
MALDI mass spectrometry of heteroconjugates
A MALDI mass spectrum of a coupling reaction mixture containing the synthetic
peptide and the aminolinker-dT
10
is shown in Figure
1
; a similar spectrum (not shown) was obtained from the reaction mixture
containing the aminolinker-dT
6
. Formation of the heteroconjugate is indicated in the spectrum by the
appearance of a signal corresponding to a compound with a mass equal to the sum
of the masses of the peptide and the nucleotide (both of whose signals are
present in the spectrum) less the mass of a water molecule, which is lost
during the coupling reaction. There is no significant signal in the mass
spectrum that might conceivably be construed as a non-covalent complex.
ESI mass spectrometry of heteroconjugates
Negative ESI mass analysis of the two crude reaction mixtures (containing primarily synthetic peptide, aminolinker-oligothymidylic acid, PyBOP
and peptide-linker-oligothymidylic acid) generated signals corresponding only to the peptide and the
aminolinker-oligothymidylic acid (data not shown). Only peptide ions were
detected in the positive ESI mode.
Whereas we could not find ion signals corresponding to the heteroconjugates in
the ESI mass spectra of the reaction mixtures, we could easily do so in both
the negative and positive ESI mass spectra of the HPLC-purified forms. The negative ion spectrum of the purified peptide-linker-dT
6
conjugate is similar in appearance to the positive ion one except sodium adduction is more pronounced in the
latter case (Fig.
2
). In this instance, the total-ion-current, i.e. the sum of all the ion signals detected in the m/z-range scanned, is more than three times greater in the
positive ion mode than the negative. The negative and positive ESI mass spectra
of the peptide-linker-dT
10
are also similar in appearance (Fig.
3
). In this case, sodium adduction is quite extensive in the positive ion mode,
and the total-ion-current in the positive mode is slightly smaller than in the
negative.
ESI tandem mass spectrometry of peptide-linker-dT
6
In the positive ESI mode, tandem mass spectrometry of the doubly charged synthetic peptide (
m/z
621) indicated fragmentation at every peptide bond in the oligomer (data not shown). In the negative mode,
the doubly charged parent of the aminolinker-dT
6
(
m/z
969.5) fragmented primarily into [dT-H
2
O-H]
-
at
m/z
303.2 and [dT]
-
at
m/z
321.2 (data not shown). Signals corresponding to PO
3
-
(
m/z
79), [thymine base]
-
(
m/z
125), [dT-base-H
2
O-H]
-
(
m/z
177.1), and [dT-base-H]
-
(
m/z
195.1) were also observed. Two series of peaks seen in the spectrum, [dT
n
-H]
-
and [dT
n
-H
2
O-H]
-
(
n
= 2, 3), are specific to the oligothymidylic acid's sequence. Only very weak
signals due to larger oligomeric fragments (
n
= 4 and 5) and no peaks corresponding to ion fragments originating from the aminolinker moiety were observed.
Spectra obtained by tandem mass analysis of the double-negatively charged peptide-linker-dT
6
ion (
m/z
1581) and the double-positively charged peptide-linker-dT
6
ion (
m/z
1583) are shown and described in Figures
4
and
5
. Using the mass spectra of aminolinker-dT
6
(not shown) as a guide, it is relatively easy to discern several series of
negative and positive ion peaks separated by 304 Da, the mass of one thymidylic acid residue, that result from fragmentation of the oligonucleotide. Close inspection of the positive ion spectrum (Fig.
5
) in the
m/z
-region below the parent ion's signal reveals peaks corresponding to
fragments cleaved from both the N- and C-termini of the peptide.
DISCUSSION
As pointed out in the introduction, the optimum matrix-assisted laser desorption conditions for ionizing the oligopeptide and
oligonucleotide components of a covalent hybrid generally clash. In Table
2
, six matrices commonly used for MALDI mass analyses of proteins, peptides and
oligonucleotides are ranked from top to bottom in the peptide-protein column in
order of decreasing suitability. The suitability of these matrices for
oligonucleotides is clearly seen to be inverse to that for peptides and
proteins. Based on prior work (
14
,
15
,
23
), it seems to us that the ionization of a covalent complex roughly reflects the
relative masses of its oligopeptide and oligonucleotide components. Thus on the
one hand, sinapinic acid turns out to be the best primary matrix for
nucleoprotein heteroconjugates that are 80% or more polypeptide by mass, e.g. phage T4 gene 32 protein UV-crosslinked to dT
20
(
15
),
Escherichia coli
transcription termination factor rho photoaffinity labeled with the ATP-analog 4-thio-uridine diphosphate (
15
), or Ung UV-crosslinked to dT
20
(
23
). On the other hand, sample preparation conditions favoring MALDI analysis of the oligonucleotidic
component, e.g. 2,4,6-trihydroxyacetophenone in diammonium-hydrogen citrate and acetonitrile (
27
), are necessary in order to produce the sharpest, most intense ion signals from
nucleopeptide conjugates whose peptide component accounts for only 10-40% of the complex's mass, e.g. the nucleopeptidic conjugate from tryptic
digests of UV-crosslinked Ung-dT
20
(
23
) or the synthetic peptide-linker-oligothymidylic acids used in the present study.
.
Some compounds commonly used as primary matrices for MALDI mass analyses of
proteins, peptides and oligonucleotides
Matrix
Peptide
Oligo-
Reference
and protein
nucleotide
Sinapinic acid
-----
-
(37)
4-hydroxy-[alpha]-cyano-cinnamic acid
-----
-
(26)
2,5-dihydroxybenzoic acid
----
--
(38)
Ferulic acid
---
---
(37)
2,4,6-trihydroxyacetophenone
--
-----
(27)
3-Picolinic acid
-
-----
(39)
The general suitability of a given matrix for a given compound class is
subjectively indicated on a qualitative scale that ranges from - (poor) to ----- (excellent).
The interaction between the oligopeptidic and oligonucleotidic parts of a
heteroconjugate also seems to manifest itself in the MALDI mass analysis of
mixtures. In general, excellent MALDI mass spectra can be produced from
mixtures containing large numbers of peptides as is done in peptide mass-mapping of proteins. In our hands, mixtures of peptides and nucleopeptidic
conjugates have not exhibited this desirable analytical trait. So far, we have
only been able to produce usable mass spectra from mixtures containing two or
three heteroconjugates (
23
), and the strength and definition of the ion signals in the best of these
spectra are inferior to those observed in spectra generated from individual, HPLC-purified heteroconjugates. Competition among the various components in the sample for incorporation into the
crystalline matrix may be one of the primary factors contributing to the refractory behavior of mixtures of nucleopeptide conjugates.
Conflicting conditions for ionization of oligopeptides and oligonucleotides also
exist in electrospray. The heteroconjugates investigated in this study show
preference for a minor variation of a solvent system reported for ESI mass
analysis of oligonucleotides (
29
). This seems sensible given that the ratios of peptide-mass to oligonucleotide-mass are ~7:10 for peptide-linker- dT
6
and 4:10 for peptide-linker-dT
10
. The fact that a 40% decrease in the size of the oligonucleotide component increases the total-ion-current in the positive mode >3-fold over that in the negative mode (compare Fig.
2
with Fig.
3
) suggests that the peptide's effect on ionization is disproportionate to its contribution to the size of the heteroconjugate. It should be noted that, for the peptide-linker dT
10
, most of the signal in the positive ion mode is due to sodium adduction (Fig.
3
) and that this effectively reduces sensitivity in this case despite the large total-ion-current. More efficient desalting may remedy this problem.
MALDI mass analysis of a crude reaction mixture proves to be a quick, sensitive
and straightforward method for unequivocally confirming formation of a covalent peptide-linker-oligothymidylic acid product (Fig.
1
), although it does not provide quantitative information. This capability was
used extensively in a study of the DNA binding site of Ung (
23
), an enzyme that initiates the uracil-excision DNA-repair pathway by removing a uracil-base introduced into DNA (
30
-
32
). Results from MALDI mass analyses were combined in this study with concurrent
results from Edman sequencing and biological testing to demonstrate that amino
acids 58-80 (tryptic peptide 6) and amino acids 185-213 (tryptic peptides 18 and 19) reside at the interface of an Ung-DNA complex. These two domains in the primary sequence of
Ung, which are heavily conserved in general, are especially similar to the
corresponding amino acid regions in human uracil-DNA glycosylase (UNG 15) (
23
). Viewing their findings for Ung in light of this fact, Bennett
et al
. ventured that the two UNG 15 amino acid residues Asp145 (corresponding to
Asp64 in Ung) and His268 (corresponding to His187 in Ung), whose substitution
was recently shown by site-directed mutagenesis studies on the human uracil-DNA glycosylase gene to inactivate UNG 15 (
33
), reside in the vicinity of the DNA-binding pocket (
23
). This speculation has since been substantiated by the 3-dimensional structure of UNG 15 obtained from X-ray analysis (
34
). Another recent X-ray diffraction structure, this one of herpes simplex virus uracil-DNA glycosylase (
35
), further corroborates the general validity of the Ung-findings by Bennett
et al
. and the utility of the mass spectrometric approach they used to obtain them.
ACKNOWLEDGEMENTS
This work was supported by grants from the National Institutes of Health
Biomedical Research Support (RR07079), from the Medical Research Foundation of
Oregon (9146), and from the National Institutes of Health (ES00210 and
GM32823). The Sciex API III Tandem Electrospray Mass Spectrometer system was purchased in part through grants from the National Science Foundation (BIR-9214371) and from the Anheuser-Busch Companies. Oligos Etc. of Wilsonville, Oregon generously donated the aminolinker-oligothymidylic acids that made this study possible. O.N.J. was supported by predoctoral fellowships from the
Danish Research Academy and the Carlsberg Foundation, Denmark. Dr Seksiri
Arttamangkul, College of Pharmacy, provided invaluable advice on the synthesis of the peptide-linker-oligothymidylic acids.
7 Williams,K.R. and Konigsberg,W.H. (1991) In Sauer,R.T. (ed.), Methods in Enzymology-Protein-DNA Interactions. Academic Press, San Diego, Vol. 208, pp. 516-539.
8 Bayard,B., Bisbal,C. and Lebleu,B. (1986) Biochemistry, 25, 3730-3736.MEDLINE Abstract
9 Bashkin,J.K., McBeath,R.J., Modak,A.S., Sample,K.R. and Wise,W.B. (1991) J. Org. Chem., 56, 3168-3176.MEDLINE Abstract
14 Jensen,O.N., Barofsky,D.F., Young,M.C., von Hippel,P.H., Swenson,S. and Seifried,S.E. (1994) In Crabb,J.W. (ed.), Techniques in Protein Chemistry V. Academic Press, San Diego, pp. 27-37.
15 Jensen,O.N., Barofsky,D.F., Young,M.C., von Hippel,P.H., Swenson,S. and Seifried,S.E. (1993) Rapid Commun. Mass Spectrom., 7, 496-501.MEDLINE Abstract
16 Karas,M., Bachmann,D., Bahr,U. and Hillenkamp,F. (1987) Int. J. Mass Spectrom. Ion Processes, 78, 53-68.MEDLINE Abstract
Present addresses:
+
European Molecular Biology Laboratory, Meyerhofstrasse 1, D-69012 Heidelberg, Germany,
[sect]
Molecular Probes, 4849 Pitchford Avenue, Eugene, OR 97402, USA and
[para]
University of Maryland at Baltimore, Department of Pharmaceutical Sciences, 20
North Pine Street, Baltimore, MD 21201-1180, USA
