Cyclization of N-Terminal S-Carbamoylmethylcysteine Causing Loss

of 17 Da from Peptides and Extra Peaks in Peptide Maps. Kieran F. Geoghegan,* Lise R. Hoth, Douglas H. Tan, Kris A. Borzilleri, Jane M. Withka, and. J...
0 downloads 0 Views 110KB Size
Cyclization of N-Terminal S-Carbamoylmethylcysteine Causing Loss of 17 Da from Peptides and Extra Peaks in Peptide Maps Kieran F. Geoghegan,* Lise R. Hoth, Douglas H. Tan, Kris A. Borzilleri, Jane M. Withka, and James G. Boyd Pfizer Global Research and Development, Eastern Point Road, Groton, Connecticut 06340 Received January 17, 2002

Enzymatic digests of proteins S-alkylated with iodoacetamide may contain peptides with N-terminal S-carbamoylmethylcysteine. These can be partly converted to a form with 17 Da lower mass and increased HPLC retention. Proof by synthesis supported by MS/MS and NMR spectroscopy was used to show that N-terminal S-carbamoylmethyl-L-cysteine can cyclize, losing NH3 to form an N-terminal residue of (R)-5-oxoperhydro-1,4-thiazine-3-carboxylic acid. The abbreviation Otc is proposed for the (R)-5-oxoperhydro-1,4-thiazine-3-carbonyl residue. The rate of cyclization is significant in 0.1 M NH4HCO3 at 37 °C, with the half-life of the acyclic form being 10-12 h for several peptides tested. This is similar to the rate at which N-terminal pyroglutamate forms from N-terminal glutamine. Keywords: iodoacetamide • peptide mapping • cyclization • mass spectrometry • S-carbamoylmethylcysteine

Introduction The use of mass properties to match a single peptide with its theoretical equivalent in a database allows proteins to be identified with high fidelity, sensitivity, and throughput.1-4 Two important search algorithms for this process1,2 use the peptide molecular mass in their first phases. This places a match at risk of being missed when chemical modification causes the experimental mass to differ from the mass of the corresponding unmodified peptide in the database. Search programs take account of this and allow potential modifications to be considered in the mass-matching phase.2,5 These could include phosphorylation of serine, threonine, or tyrosine, oxidation of methionine, or any other modification that may be present. With enough computational power and a complete catalog of modifications, all possible variations could be considered and the information yield in proteomics maximized. This paper describes a secondary modification affecting peptides with S-carbamoylmethyl-L-cysteine (CamC) at the N-terminus. These exist in enzymatic digests of proteins that have been S-alkylated with iodoacetamide. In a reaction that parallels the formation of N-terminal 2-oxopyrrolidine-5-carboxylic acid (also called pyroglutamic acid, 5-oxoproline, or 2-pyrrolidone-5-carboxylic acid) from N-terminal glutamine,6,7 cyclization of N-terminal CamC gives an N-terminal residue of (R)-5-oxoperhydro-1,4-thiazine-3-carboxylic acid (Figure 1). Peptides in which this has occurred become more hydrophobic and lose 17 Da from the N-terminal residue. The effects on peptide mapping studies can include partial loss of signal from the original form of affected peptides and the creation of new peaks that are not correctly identified by the search program. * To whom correspondence should be addressed. Fax: (860)441-3783. E-mail: [email protected]. 10.1021/pr025503d CCC: $22.00

 2002 American Chemical Society

Figure 1. Cyclization of N-terminal CamC to an N-terminal residue of (R)-5-oxoperhydro-1,4-thiazine-3-carboxylic acid. -NH-R indicates a continuing peptide chain.

Experimental Section Biochemicals. Iodoacetamide and two synthetic peptides (CGYGPKKKRKVGG and CDPGYIGSR) were obtained from Sigma (St. Louis, MO). CTYNSLGAK was obtained by custom synthesis from AnaSpec (San Jose, CA). Each peptide was separately dissolved to 1.0 mM in a fresh 1.2 mM solution of iodoacetamide in 0.1 M Tris HCl, pH 8.0, and the reaction mixtures were incubated at 22 °C in the dark for 30-60 min before being fractionated by RP-HPLC on a Waters Symmetry C18 column (type T81051Q42). In each case, the major product was isolated and shown by mass spectrometry to be the N-terminally S-carbamoylmethylated peptide. Aliquots of the product were dried in a SpeedVac (Thermo Savant). HPLC and Mass Spectrometry. LC-MS peptide mapping was run on a Hewlett-Packard model 1090 chromatograph interfaced with a Thermo Finnigan LCQ ion-trap mass spectrometer controlled by Xcalibur (Thermo Finnigan) software. A Vydac C18 microbore column (type 218TP51; 1.0 mm × 250 mm) was run at 0.1 mL/min with a gradient of acetonitrile from 1.6 to 50% in the first 100 min after injection (nominally a change of 0.5% in the acetonitrile concentration per minute) and from 50 to 80% in the next 40 min. The solvents were as Journal of Proteome Research 2002, 1, 181-187

181

Published on Web 02/19/2002

research articles follows: A, 0.1% TFA; B, 0.085% TFA in acetonitrile. The experiments on synthetic peptides with N-terminal CamC were analyzed by HPLC on a model 1090 chromatograph running at 0.2 mL/min with a Vydac C18 narrowbore column (type 218TP52; 2.1 mm × 250 mm) and a gradient from 0 to 48% acetonitrile between 1 and 14 min after injection (nominally a change of 3.7% in the acetonitrile concentration per minute). The solvents were as follows: A, 0.1% TFA; B, 0.085% TFA in acetonitrile:water (80:20) (v/v). p25 Protein. His-tagged cdk5 and its regulatory partner p25 were coexpressed in Sf-9 cells, and their heterodimer was isolated by metal-affinity chromatography. Briefly, a DNA fragment encoding p25 was generated by PCR using human p35 cDNA as template and subcloned into a pFastBac vector. After confirmation of the sequence, the pFastBac-p25 construct was transformed into DH10Bac E. coli for bacmid preparation. Bacmids containing the p25 insert were isolated, confirmed, and transfected into Sf-9 insect cells to obtain P1 p25 virus. The p25 sequence was composed of amino acid residues 99307 of the precursor p35 protein (SWISS-PROT accession Q15078) preceded by an initiator methionine. The subunits of the purified cdk5-p25 heterodimer were separated by RP-HPLC on a Vydac C18 column (type 218TP510) running at 4 mL/min with a gradient of 0-64% acetonitrile in 0.1% TFA over 29 min. p25 was eluted several minutes after cdk5, and ES-MS indicated that the major forms of the protein (23 289 Da and 23 370 Da) lacked the initiator Met but included an R-N-acetyl group (confirmed in peptide mapping by LC-MS), while about onehalf of the protein was monophosphorylated (also confirmed). The protein was dried in a SpeedVac concentrator. Chemistry. (R)-5-Oxoperhydro-1,4-thiazine-3-carboxylic acid was prepared by a modified form of the method of Cundari et al.8 L-Cysteine (4.85 g, 40.0 mmol) was stirred until dissolved in a mixture made from 50 mL of saturated sodium bicarbonate, 20 mL of water, and 10 mL of methanol. Ethyl bromoacetate (6.7 g, 40.2 mmol) was added dropwise over several minutes, after which time the reaction mixture was stirred at room temperature for 2 h and then refluxed overnight. After cooling, the reaction mixture was carefully acidified with 6 N HCl and then evaporated to an orange solid. The solid was dissolved in 70 mL of ethyl acetate, and the solution was heated under reflux and filtered hot before being evaporated to a solid. Crude yield: 0.80 g (12%). The crude product was dissolved in 15 mL of ethyl acetate, and the organic solution was extracted twice with water (15 mL, 8 mL). The combined water layers were evaporated to a clear oil that crystallized upon standing. Final yield: 0.50 g (7.7%). 400 MHz 1H NMR (CD3OD) δ: 4.39 (m, 1H), 3.39 (d, J ) 17 Hz, 1H), 3.20 (d, J ) 17 Hz, 1H), 3.16 (dd, J ) 3.4, 13 Hz, 1H) 3.05 (dd, J ) 1.3, 5.8 Hz). APCI MS: MH+ ) 162.0 (theoretical: 162.1). Peptide Synthesis. Peptide synthesis was performed on an Applied Biosystems model 433A solid-phase synthesizer using Synthassist software version 2.0. Syntheses were run on a 0.25 mmol scale using preloaded Fmoc amino acid (Wang) resins and standard Fmoc amino acids. Activation with O-benzotriazol-1-yl-N,N,N′,N′-tetramethyluroniumhexafluorophosphate (HBTU) and single amino acid coupling cycles were employed. Following peptide synthesis, a portion of the resin was manually coupled to (R)-5-oxoperhydro-1,4-thiazine-3-carboxylic acid using HBTU activation. The resin was filtered, thoroughly washed, and then dried under vacuum. The peptide product was cleaved from the resin and deprotected by treatment with a 5 mL solution of 83% TFA, 5% phenol, 4.5% water, 5% 182

Journal of Proteome Research • Vol. 1, No. 2, 2002

Geoghegan et al. thioanisole, and 2.5% ethanedithiol (Reagent K) for 30 min at 23 °C. The mixture was filtered, the TFA filtrate was diluted into 40 mL of diethyl ether, and the precipitated crude peptide salt was collected by centrifugation. The peptide was purified by preparative RP-HPLC (20 mm × 250 mm Waters C18 column) using a water/acetonitrile (0.1% TFA) gradient. Fractions were assayed by analytical HPLC (4.5 mm × 250 mm Vydac C18 column), appropriate fractions were pooled, and the concentration was estimated by measuring absorbance at 276 nm. Product peptide was then lyophilized to a white powder. It was >97% pure by analytical LC-MS. Monoisotopic MH+ for (R)-5-oxoperhydro-1,4-thiazine-3-carbonyl-Thr-Tyr-AsnSer-Leu-Gly-Ala-Lys ) 996.8 (theor 996.4). NMR Spectroscopy. One-dimensional proton and twodimensional 1H-1H COSY were performed separately on des17 Da CamC-TYNSLGAK formed by incubation of the peptide at pH 8.0, and R-N-(5-oxoperhydro-1,4-thiazine-3-carbonyl)TYNSLGAK was made by coupling (R)-5-oxoperhydro-1,4thiazine-3-carboxylic acid to TYNSLGAK. Both samples contained 3.1 mM peptide in 99.6% DMSO-d6. Experiments were run on a Bruker Avance 600 MHz spectrometer at 25 °C. Thirtytwo scans were run in each one-dimensional proton experiment and sixteen in each COSY experiment. In addition, chemical shift assignments for the peptide were determined using standard homonuclear two-dimensional NMR methods including NOESY and TOCSY experiments. TOCSY spectra were acquired at a mixing time of 60 ms, and NOESY spectra were acquired at a mixing time of 500 ms at 25 °C. Data were processed and analyzed on a Silicon Graphics Octane workstation using XWINNMR, FELIX, and XWINPLOT software.

Results Following a published method,9 dried recombinant p25 was dissolved in 0.4 M NH4HCO3 containing 8 M urea, reduced with DTT, and S-alkylated with iodoacetamide. The sample was then diluted with 3 volumes of water to make the buffer 0.1 M NH4HCO3 containing 2 M urea, and the reduced and S-alkylated p25 was digested with trypsin by incubation at 37 °C for 20 h. As recommended,9 the trypsin/substrate ratio was 1:25 (w/w). The digest was predicted to include two peptides with Nterminal CamC, both of which were detected in a peptide map obtained by LC-MS (Figure 2). Mass range plots corresponding to MH+ for the predicted peptides CamC-LSVINLMSSK (1251.6 Da) and CamC-LGEFL-CamC-R (1054.5 Da) are shown in Figure 2, parts C and E, respectively. The identities of both peptides were confirmed by signal-dependent MS/MS in the ion-trap10 and postrun analysis using SEQUEST (data not shown). Two other peaks in the map were unidentified by SEQUEST, but close examination during a search for phosphorylation sites showed they were related to the peptides with N-terminal CamC. Each of these products had mass 17 Da less than its parent (MH+ of 1234.6 and 1037.6 Da, respectively; Figure 2D,F). Each was eluted several minutes later than its parent, indicating greater hydrophobicity. For MH+, base peak intensity for the derivative peak was 3-5 times greater than for the parent (Figure 2C-F), but this ratio was nearly inverted for the corresponding MH22+ ions (not shown). Where well-separated absorbance peaks were available for comparison, they indicated a ratio ranging from about even to 2:1 in favor of the des-17 Da form. MS/MS of these products placed the mass change in the N-terminal CamC residue (Figure 3). For the des-17 Da

Cyclization of N-Terminal S-Carbamoylmethylcysteine

research articles

Figure 2. LC-MS analysis of a tryptic digest of recombinant p25: (A) Absorbance chromatogram at 220 nm; (B) base peak chromatogram from the mass spectrometer; (C) mass chromatogram for range 1251.0-1252.0 Da, corresponding to MH+ (theor) of 1251.6 for predicted peptide CamC-LSVINLMSSK (peak and signal intensity indicated by arrow); (D) mass chromatogram for des-17 Da form of CamCLSVINLMSSK (plot for 1234.0-1234.8 Da); (E) mass chromatogram for the range 1054.0-1055.0 Da, corresponding to MH+ (theor) of 1054.5 for predicted peptide CamC-LGEFL-CamC-R; (F) mass chromatogram for des-17 Da form of CamC-LGEFL-CamC-R (plot for 1037.01038.0 Da).

Figure 3. MS/MS spectra for des-17 Da forms of two tryptic peptides from p25: (A) des-17 Da form of CamC-LSVINLMSSK; (B) des-17 Da form of CamC-LGEFL-CamC-R. Fragment ions are assigned on the basis that the loss of 17 Da occurred in the N-terminal residue.

derivative of CamC-LSVINLMSSK (Figure 3A), b ions corresponding to the sequence with 17 Da lost from CamC were

detected beginning at b4, and detection of y9 (978.5 Da obsd; 978.5 Da theor) ruled out a mass change in Ser-3. For the desJournal of Proteome Research • Vol. 1, No. 2, 2002 183

research articles

Figure 4. HPLC analysis of CamC-GYGPKKKRKVGG incubated in tryptic digest conditions of 0.1 M NH4HCO3 at 37 °C. Mass spectrometry (not shown) confirmed the identity of the starting material and showed that the peak increasing with time was the des-17 Da derivative.

17 Da derivative of CamC-LGEFL-CamC-R (Figure 3B), b ions beginning at b4 (both peptides coincidentally had b4 at 443 Da) and the y6 ion at 781.2 Da (781.4 Da theor.) also placed the loss of 17 Da in the N-terminal CamC. The internal CamC was not changed, and the peak at m/z ) 946.4 (Figure 3B), which is 91 Da lower than MH+, was interpreted as due to cleavage of the βC-γS bond in its side chain in the mass spectrometer.11 This process may be favored by the proximity of the internal CamC to C-terminal Arg.11 Analyses of other proteins giving tryptic peptides with N-terminal CamC showed similar partitioning of these peptides between standard and des-17 Da forms (data not shown). As the result appeared general, the chemical basis of the mass and mobility shift was investigated. The structural change seemed likely to occur after digestion, when peptides with N-terminal CamC were already formed. To confirm this and measure the rate, two synthetic peptides with N-terminal CamC (CamC-GYGPKKKRKVGG and CamC-DPGYIGSR) were incubated in 0.1 M NH4HCO3 at 37 °C, and formation of the des-17 Da form was followed by HPLC. In both cases, one shown in Figure 4, conversion to the des-17 Da form had a half-life of about 10 h. This was consistent with the rate observed for two tryptic peptides in the p25 digest (Figure 2). LC-MS (not shown) confirmed the identities of all starting and product peaks, and automated MS/MS showed that the loss of 17 Da occurred in the N-terminal residue for both peptides (not shown).

Geoghegan et al. We hypothesized that N-terminal CamC can cyclize (Figure 1) to lose 17 Da (NH3). The hypothesis was tested by synthesizing a peptide with an (R)-5-oxoperhydro-1,4-thiazine-3-carbonyl residue at the N-terminus, and comparing it with the des17 Da product formed from a peptide with the same sequence except for unaltered CamC at its N-terminus (Figure 5). (R)-5-Oxoperhydro-1,4-thiazine-3-carboxylic acid was synthesized and then coupled to TYNSLGAK assembled by solidphase synthesis (see the Experimental Section). Des-17 Da CamC-TYNSLGAK was obtained by incubating CamCTYNSLGAK in 0.01 M sodium phosphate, pH 8.0, for 92 h and then isolating the des-17 Da product by HPLC (similar to Figure 4). Both products were analyzed by LC-MS. They had identical retention times, and their MS/MS spectra were identical with respect to all major peaks (Figure 6). Standard homonuclear 2D COSY, TOCSY, and NOESY NMR experiments were used to compare the two products and determine the chemical shift assignments. The two products gave essentially identical spectra in all modes of spectroscopy, except for the exchangeable hydroxyl protons for Thr and Ser residues still present in the peptide prepared by coupling (R)5-oxoperhydro-1,4-thiazine-3-carboxylic acid to TYNSLGAK. A comparison of the COSY spectra for both products and the chemical shift assignments for the 5-oxoperhydro-1,4-thiazine3-carbonyl ring are shown in Figures 7 and 8, respectively. These results supported a conclusion that N-terminal CamC can cyclize to an N-terminal residue of (R)-5-oxoperhydro-1,4thiazine-3-carboxylic acid. Khandke et al.7 showed that the ionic content of the solution can strongly affect the rate of cyclization of N-terminal Gln, with the rate in 0.01 M sodium phosphate, pH 8.0, being about twice that in 0.2 M NH4HCO3, pH 7.9. A similar effect was found in the present work (not shown), but no detailed study was undertaken. A brief study was also made of the behavior of S-carboxymethyl-GYGPKKKRKVGG incubated in 0.01 M sodium phosphate, pH 8.0, in which only trace levels of Nterminal cyclization occurred during a 24 h incubation at 22 °C. This appeared consistent with the results of Ozawa,12 who used temperatures well in excess of 100 °C to obtain good yields of cyclization of S-carboxymethylcysteine, but there is also at least one report of considerable cyclization of N-terminal S-carboxymethylcysteine in an immunoglobulin digested with papain.13

Discussion The complexity of LC-MS peptide maps can make it necessary to accept computer-based assignments of identity to a fraction of the peaks while leaving others unexplained. In principle, it would be preferable to account for every peak, as this would ensure that all the biological information captured by an experiment is actually recovered. With proteomic experiments moving to industrial scale,14-17 the fullest possible recovery of information ensures the maximum gain to biology.

Figure 5. Proof by synthesis that loss of 17 Da from N-terminal CamC gives an N-terminal residue of (R)-5-oxoperhydro-1,4-thiazine3-carboxylic acid. 184

Journal of Proteome Research • Vol. 1, No. 2, 2002

Cyclization of N-Terminal S-Carbamoylmethylcysteine

research articles

Figure 6. (A) Base peak chromatogram of the des-17 Da product formed from Cam-TYNSLGAK; (B) MS of the 27.43 min peak; (C) MS/MS of the MH+ ) 996.4 Da ion; (D, E, F) corresponding data for synthetic R-N-5-oxoperhydro-1,4-thiazine-3-carbonyl-TYNSLGAK.

Figure 7. Correlated spectroscopy (COSY) experiment at 25 °C comparing (A) des-17 Da product formed from Cam-TYNSLGAK (1.7 mg in 0.5 mL of 99.6% d6-DMSO), and (B) synthetic (R)-5-oxoperhydro-1,4-thiazine-3-carbonyl-TYNSLGAK (2.0 mg in 0.65 mL of 99.6% d6-DMSO).

It is not unusual for different peptides to be derived from the same fragment of amino acid sequence. Incomplete modification, which may or may not have a biological origin, leads to this distribution between multiple products. Phosphorylations, oxidations, glycosylations, and modification by acrylamide in gels are examples of agents that can lead to this. Another cause is cyclization of N-terminal Gln to N-terminal pyroglutamate, which causes peptides beginning with Gln to

be split between two species.7 Maximizing the sensitivity and completeness of the analysis requires that as many components as possible of the product mixture be identified and accounted for. This paper shows that peptides with N-terminal CamC also undergo division between two forms based on cyclization. Formation of the six-membered ring of (R)-5-oxoperhydro-1,4thiazine-3-carboxylic acid occurred at a similar rate to formaJournal of Proteome Research • Vol. 1, No. 2, 2002 185

research articles

Geoghegan et al.

Figure 8. Two-dimensional TOCSY data showing cross-peaks for the correlated protons of the 5-oxoperhydro-1,4-thiazine-3-carbonyl ring at 25 °C. Chemical shifts determined for ring protons, referenced relative to the water resonance in DMSO-d6 at 3.43 ppm, are shown.

tion of a five-membered ring by cyclization of N-terminal Gln. Clearly, when iodoacetamide is used to S-alkylate Cys in proteins before digestion, peptides with N-terminal CamC will be generated when the protein sequence permits, and these will undergo some conversion to their N-terminally cyclized forms. Being more hydrophobic than their parent peptides, the new products will appear as extra peaks and increase the complexity of the peptide map. The term pyroglutamyl is less systematic than other names for the product of cyclization of N-terminal Gln, but it fits better than others with standard peptide nomenclature. This extends to an abbreviation (Glp) for the pyroglutamyl residue recommended by the IUPAC-IUBMB Joint Commission on Biochemical Nomenclature18 (see also http://www.chem. qmul.ac.uk/iupac/AminoAcid/). By analogy, we propose Otc as an abbreviation for the N-terminal (R)-5-oxoperhydro-1,4thiazine-3-carbonyl residue formed by cyclization of N-terminal CamC. The extent to which a modification changes the HPLC retention of a peptide varies with the length and hydrophobicity of the peptide, and a change will typically have a more pronounced effect in a shorter peptide. For the peptides referred to in Figures 2 and 4, N-terminal cyclization delayed elution by the following apparent percentages of acetonitrile: for CamC-LSVINLMSSK, 4%; for CamC-LGEFL-CamC-R, 6%; for CamC-CDPGYIGSR, 3% (not shown); and for CamCGYGPKKKRKVGG, 2%. The observed half-life for formation of the N-terminal 5-oxoperhydro-1,4-thiazine-3-carbonyl group at 37 °C was about 10 h, corresponding to an apparent first-order rate constant of 2 × 10-5 s-1. This degree of efficiency under mild conditions prompted curiosity as to the mechanism of the reaction (Scheme 1). Oxothiomorpholine synthesis by O f N acyl transfer (I) is well precedented, but we are not aware of other examples of this synthesis proceeding via intramolecular N f N acyl transfer (II). However, formation of the analogous, sulfur-free piperidin-2-one from 5-aminopentanoic acid amide (III) was studied in detail by Perrin et al.19 These investigators reported a first-order rate constant of 2 × 10-5 s-1 at 25 °C in dilute NaOH, quite close to the rate observed in the present case. The N-terminal 5-oxoperhydro-1,4-thiazine-3-carbonyl residue (Otc) has a monoisotopic mass of 144.0 Da, including the 186

Journal of Proteome Research • Vol. 1, No. 2, 2002

Scheme 1

mass of hydrogen on its lactam nitrogen. This is not isobaric with any standard amino acid residue, although it might in a rare case be mistaken for an R-N-acetylthreonyl or O-acetylthreonyl residue at the N-terminus. The absence of mass equivalence with a common residue reduces the chances that MS/MS spectra of peptides with the modification will be misinterpreted as good matches to predicted fragmentation patterns. It is more likely that peaks corresponding to these species will give no high-scoring match to a known sequence fragment. If peptide mass alone is considered, the loss of 17 Da associated with cyclization of CamC could be mistaken for the effect of formation of a succinimide ring consequent to deamidation at Asn, which causes the loss of 17 Da.20 Cyclization of N-terminal Gln to pyroglutamate also causes loss of 17 Da. NMR provided the final evidence that the same N-terminal residue was formed both by loss of 17 Da from N-terminal CamC and by coupling (R)-5-oxoperhydro-1,4-thiazine-3-carboxylic acid to the N-terminus of a peptide. The chemical shift

research articles

Cyclization of N-Terminal S-Carbamoylmethylcysteine assignments for the new N-terminus (Figure 8) agreed well with those determined for the same residue in another context.21 Finally, a reviewer of this paper noted that the cyclization of N-terminal CamC to give the Otc residue has been known anecdotally among some protein chemists, and we wish to acknowledge this. Although the result was new to us when we encountered it experimentally, we were surprised to find no previous account of it in the literature. As we are not aware of strong reasons to prefer iodoacetamide as an S-alkylating agent over iodoacetic acid or 4-vinylpyridine, the possibility of occasional complications due to the cyclization may be a point that weighs against the selection of iodoacetamide.

Acknowledgment. We thank L. Contillo for arranging peptide synthesis and Dr. Henry B. F. Dixon of Cambridge University for helpful discussions and advice. References (1) Eng, J. K.; McCormack, A., L.; Yates, J. R., III. J. Am. Soc. Mass Spectrom. 1994, 5, 976-989. (2) Mann, M.; Wilm, M. Anal. Chem. 1994, 66, 4390-4399. (3) Shevchenko, A.; Jensen, O. N.; Podtelejnikov, A. V.; Sagliocco, F.; Wilm, M.; Vorm, O.; Mortensen, P.; Boucherie, H.; Mann, M. Proc. Natl. Acad. Sci. U.S.A. 1996, 93, 14440-14445. (4) Yates, J. R., III. Trends Genet. 2000, 16, 5-8. (5) Yates, J. R., III; Eng, J. K.; McCormack, A. L.; Schieltz, D. Anal. Chem. 1995, 67, 1426-1436. (6) Sanger, F.; Thompson, E. O. P.; Kitai, R. Biochem. J. 1955, 59, 509-518. (7) Khandke, K. M.; Fairwell, T.; Chait, B. T.; Manjula, B. N. Int. J. Pept. Prot. Res. 1989, 34, 118-123. (8) Cundari, S.; Dalpozzo, R.; De Nino, A.; Procopio, A.; Sindona, G.;

Athanassopulos, K. Tetrahedron 1999, 55, 10155-10162. (9) Stone, K. L.; Williams, K. R. In A Practical Guide to Protein and Peptide Purification for Microsequencing; Matsudaira, P., Ed.; Academic Press: San Diego, 1993; pp 43-69. (10) Jonscher, K. R.; Yates, J. R. Anal. Biochem. 1997, 244, 1-15. (11) Wolf, S. M.; Biemann, K. Int. J. Mass Spectrom. Ion Process. 1997, 160, 317-329. (12) Ozawa, H. Bull. Soc. Chem. Jpn. 1963, 36, 920-922. (13) Smyth, D. S.; Utsumi, S. Nature 1967, 216, 332-335. (14) Lopez, M. F. Electrophoresis 2000, 21, 1082-1093. (15) Patterson, S. D. Curr. Opin. Biotechnol. 2000, 11, 413-418. (16) Gavin, A.-C.; Bo¨sche, M.; Krause, R.; Grandi, P.; Marzioch, M.; Bauer, A.; Schultz, J.; Rick, J. M.; Michon, A.-M.; Cruciat, C.-M.; Remor, M.; Ho¨fert, C.; Schelder, M.; Brajenovic, M.; Ruffner, H.; Merino, A.; Klein, K.; Hudak, M.; Dickson, D.; Rudi, T.; Gnau, V.; Bauch, A.; Bastuck, S.; Huhse, B.; Leutwein, C.; Heurtier, M.-A.; Copley, R. R.; Edelmann, A.; Querfurth, E.; Rybin, V.; Drewes, G.; Raida, M.; Bouwmeester, T.; Bork, P.; Seraphin, B.; Kuster, B.; Neubauer, G.; Superti-Furga, G. Nature 2002, 415, 141-147. (17) Ho, Y.; Gruhler, A.; Heilbut, A.; Bader, G. D.; Moore, L.; Adams, S.-L.; Millar, A.; Taylor, P.; Bennett, K.; Boutilier, K.; Yang, L.; Wolting, C.; Donaldson, I.; Schandorff, S.; Shewnarane, J.; M Vo; Taggart, J.; Goudreault, M.; Muskat, B.; Alfarano, C.; Dewar, D.; Lin, Z.; Michalickova, K.; Willems, A. R.; Sassi, H.; Nielsen, P. A.; Rasmussen, K. J.; Andersen, J. R.; Johansen, L. E.; Hansen, L. H.; Jespersen, H.; Podtelejnikov, A.; Nielsen, E.; Crawford, J.; Poulsen, V.; Sørensen, B. D.; Matthiesen, J.; Hendrickson, R. C.; Gleeson, F.; Pawson, T.; Moran, M. F.; Durocher, D.; Mann, M.; Hogue, C. W. V.; Figeys, D.; Tyers, M. Nature 2002, 415, 180-183. (18) Anonymous. Biochem. J. 1984, 219, 345-373. (19) Perrin, C. L.; Arrhenius, G. M. L. J. Am. Chem. Soc. 1982, 104, 2839-2842. (20) Hekman, C. M.; DeMond, W. S.; Kelley, P. J.; Mauch, S. F.; Williams, J. D. J. Pharm. Biomed. Anal. 1999, 20, 763-772. (21) Zanotti, G.; Pinnen, F.; Lucente, G.; Cerrini, S.; Gavuzzo, E. J. Chem. Soc., Perkin Trans. 1 1988, 2647-2652.

PR025503D

Journal of Proteome Research • Vol. 1, No. 2, 2002 187