A Picomole-Scale Method for Charge Derivatization of Peptides for

Localization of the O-Glycosylated Sites in Peptides by Fixed-Charge ... Berkel , J. Martin E. Quirke, Rudolfo A. Tigani, and Anja S. Dilley , Thomas ...
0 downloads 0 Views 312KB Size
Download PDF
Anal. Chem. 1997, 69, 137-144

A Picomole-Scale Method for Charge Derivatization of Peptides for Sequence Analysis by Mass Spectrometry Zhi-Heng Huang,† Jiang Wu,‡ Kenneth D. W. Roth,‡ Ying Yang,‡ Douglas A. Gage,† and J. Throck Watson*,†,‡

Departments of Biochemistry and Chemistry, MSU-NIH Mass Spectrometry Facility, Michigan State University, East Lansing, Michigan 48824

A highly activated ester containing a fixed positive charge, S-pentafluorophenyl [tris(2,4,6-trimethoxyphenyl)phosphonium]acetate bromide (TMPP-AcSC6F5 bromide), has been synthesized as a reagent for N-terminal modification of peptides. Stable in aqueous acetonitrile solution during extended storage, TMPP-AcSC6F5 bromide reacts with unprotected peptides through p-(dimethylamino)pyridine (DMAP)-promoted amidation in aqueous acetonitrile (15 min, ambient temperature) to form N-TMPP-Ac derivatives of peptides. These peptide derivatives are readily amenable to analysis by fast atom bombardment (FAB) and matrix-assisted laser desorption/ionization (MALDI) mass spectrometry. Greater than 90% conversion has been observed in transforming low-nanomole quantities of analyte using molar ratios of 1:5:10 (peptide/reagent/ DMAP). For reactions at the picomole level a slightly modified stoichiometry, with molar ratios of 1:10:500, is employed. Owing to the high reaction efficiency and the tolerance to moderate excess reagent and base during analysis by FAB- and MALDI-MS, the reaction mixture containing the modified peptides can be analyzed directly in most cases, without sample cleanup. Examples of the preparation and analysis of a variety of N-TMPP-acetylpeptides (TMPP-Ac-peptides) ranging from hexamers to 15-mers are given. Collisionally activated dissociation tandem mass spectrometry of TMPP-Ac-derivatives showed dominant a-type ions, accompanied by d- and c-type ions in some cases, allowing sequence determination to be made in a straightforward manner. In pursuing the analytical advantages of charged derivatives to simplify the spectra during subsequent collisionally activated dissociation tandem mass spectrometry (CAD MS/MS),1-3 several research groups have incorporated different head groups into the N-terminus of peptides to facilitate analysis by FAB-MS and other desorption techniques.3-14 One of the first examples was based †

Department of Biochemistry. Department of Chemistry. (1) Papayannopoulos, I. A. Mass Spectrom. Rev. 1995, 14, 49. (2) Biemann, K. In Methods in Enzymology; McCloskey, J. A., Ed.; Academic Press: San Diego, CA, 1990; Vol. 193, pp 455-518, 886. (3) Zaia, J. In Methods in Molecular Biology, Vol. 61: Protein and Peptide Analysis by Mass Spectrometry; Chapman, J. R., Ed.; Humana Press Inc.: Totowa, NJ, 1996; pp 29-41. (4) Zaia, J.; Fenselau, C. Proceedings of the 43rd ASMS Conference on Mass Spectrommetry and Allied Topics; Atlanta, GA, May 21-26, 1995; p 513. (5) Zaia, J.; Biemann, K. J. Am. Soc. Mass Spectrom. 1995, 6, 428. ‡

S0003-2700(96)00857-8 CCC: $14.00

© 1997 American Chemical Society

on the two-step procedures developed by Biemann5-7 and Stults8,10 involving the condensation of a peptide with iodoacetic anhydride (IAA, in solution or in the gas phase) followed by reaction with trimethylamine or other higher amines to form a N-trimethylammonium acetyl (TMAA) derivative or its higher homologs. Direct N-terminal coupling of peptides with quaternized amino acids (betaines) with a water-soluble carbodiimide such as 1-ethyl-3-[3(dimethylamino)propyl]carbodiimide (EDC) in aqueous tetrahydrofuran (THF) solution has been used by Biemann’s group to prepare several types of derivatives with a fixed charge.5 Nucleophilic substitution at the N-terminus of peptides with (triphenylphosphonium)ethyl bromide (TPP-ethyl bromide) involving somewhat different chemistry also has been employed by Watson and co-workers11-14 for modification of peptides. All these methodologies, however, have common technical pitfalls due to insufficient reactivity of the derivatizing agents, resulting in the need for large quantities of peptide, long reaction times (2-16 h), and large excess of reagents and buffers that necessitate sample cleanup. Newer techniques based on the use of active esters like N-hydroxysuccinimide esters (OSu esters) containing a charged group [Q+(CH2)nCOOSu, where Q ) Me3N, Ph3P, and others, n ) 4 or 5]15-19 have the advantages of higher yields and purity of (6) Vath, J. E.; Biemann, K. Int. J. Mass Spectrom. Ion Processes 1990, 100, 287. (7) Vath, J. A.; Zollinger, M.; Biemann, K. Fresenius Z. Anal. Chem. 1988, 331, 248. (8) Wetzel, R.; Halualani, R.; Stults, J. T.; Quan, C. Bioconjugate Chem. 1990, 1, 114. (9) Stults, J. T.; Lai, J.; McCune, S.; Wetzel, R. Anal. Chem. 1993, 65, 1703. (10) Stults, J. T. Proceedings of the 40th ASMS Conference on Mass Spectrometry and Allied Topics; Washington, DC, May 31-June 5, 1992; p 1815. (11) Chang, Y.-S.; Gage, D. A.; Watson, J. T. Biol. Mass Spectrom. 1993, 22, 176. (12) Wagner, D. S.; Nieuwenhuis, T. J.; Chang, Y.-S.; Gage, D. A.; Watson, J. T. Proceedings of the 40th ASMS Conference on Mass Spectrometry and Allied Topics; Washington, DC, May 31-June 5, 1992; p 1837. (13) Wagner, D. S.; Salari, A.; Gage, D. A.; Leykam, J.; Fetter, J.; Hollingsworth, R.; Watson, J. T. Biol. Mass Spectrom. 1991, 20, 419. (14) Watson, J. T.; Wagner, D. S.; Chang, Y.-S.; Strahler, J. R.; Hanash, S. M.; Gage, D. A. Int. J. Mass Spectrom. Ion Processes 1991, 111, 191. (15) Mamer, O. A.; Just, G.; Li, C.-S.; Preville, P.; Watson, S.; Young, R.; Yergey, J. A. J. Am. Soc. Mass Spectrom. 1994, 5, 292. (16) Bartlet-Jones, M.; Jeffery, W. A.; Hansen, H. F.; Pappin, D. J. C. Rapid Commun. Mass Spectrom. 1994, 8, 737. (17) Jeffery, W. A.; Bartlet-Jones, M.; Pappin, D. J.; Keane, A.; Harrison, M.; Cottrell, J. S. Proceeings of the 43rd ASMS Conference on Mass Spectrometry and Allied Topics, Atlanta, GA, May 21-26, 1995; p 621. (18) Hines, W.; Peltier, J.; Hsieh, F.; Martin, S. A. Proceedings of the 43rd ASMS Conference on Mass Spectrometry and Allied Topics; Atlanta, GA, May 2126, 1995; p 387.

Analytical Chemistry, Vol. 69, No. 2, January 15, 1997 137

Scheme 1. General Strategy for TMPP Acetylation of Peptides

products. Bartlet-Jones et al.16 published an example of a subpicomole scale reaction. More recently, Spengler et al.19 reported the MALDI postsource decay (PSD) analysis of charged derivatives of protected peptides in which the basic amino acid residues (Arg, Lys) had been modified. However, the general utility of active esters for preparing charged derivatives remains to be tested. Here, we report a simple and efficient chemical procedure for the preparation of a new type of charged derivative of peptides. We have developed a new reagent, S-pentafluorophenyl [N-tris(2,4,6-trimethoxyphenyl)phosphonium]acetate bromide (TMPPAcSC6F5 bromide), which can be used to convert peptides (up to 15-mers) at the low-picomole to nanomole level to their corresponding [tris(2,4,6-trimethoxyphenyl)phosphonium]acetyl (TMPPAc) derivatives. The overall strategy is outlined in Scheme 1. Attractive features of the new method include high conversion (greater than 90%) of the analyte under mild conditions within 15 min. Only a 5-10 molar excess of reagent is used so that cleanup is minimal, and usually unnecessary, prior to analysis by desorption/ionization mass spectrometry. This procedure, after a slight, but important modification, was successfully applied to lowpicomole quantities of analyte for analysis by MALDI-MS. The TMPP-Ac-mediated derivatizaion compares favorably to currently available methodologies used for the same purpose (see Table 1). The resulting N-terminal TMPP-Ac derivative has the desirable attribute of directing fragmentation during CAD to produce a series of N-terminal ions (predominantly a-type ions; for the possible structures of fragment ions produced by the N-terminally charged derivatives, see refs 9 and 14). The advantages are illustrated herein by comparison of FAB and MALDI spectra of peptides obtained before and after derivatization. EXPERIMENTAL SECTION Chemicals. All peptides and chemicals were obtained from Aldrich (Milwaukee, WI) or Sigma (St. Louis, MO) Chemical Co. Preparation of Reagents. Experiments should be carried out under a hood because of the use of benzene and the evolution of HBr gas during step i. The new active ester reagent, S(19) Spengler, B.; Luetzenkirchen, F.; Metzgers, S.; Kaufmann, R.; Jeffery, W.; Pappin, D. J. C. Proceedings of the 44th ASMS Conference on Mass Spectrometry and Allied Topics, Portland, OR, May 12-16, 1996; p 352.

138

Analytical Chemistry, Vol. 69, No. 2, January 15, 1997

pentafluorophenyl [tris(2,4,6-trimethoxyphenyl)phosphonium]acetate bromide (TMPP-AcSC6F5 Br), was prepared using an approach previously reported in the literature20,21 with modifications as described below: (i) Bromoacetic Acid S-Pentafluorophenyl Ester (Br-Ac-SC6F5). A mixture of 500 mg (2.5 mmol) of pentafluorothiophenol (C6F5SH) and 500 mg (2.5 mmol) of bromoacetyl bromide in 3 mL of benzene was refluxed overnight in a sand bath at 120 °C (with a calcium chloride drying tube attached to the condenser to exclude moisture). After cooling, the reaction mixture was diluted with 5 mL of benzene, washed successively with aqueous NaHCO3 and water, and dried over anhydrous MgSO4. The dried organic layer was allowed to pass through a Pasteur pipet filled with ∼4 cm of silica gel. The solvent was removed using a rotary evaporator below 50 °C to leave the intermediate as a colorless liquid; yield 800 mg (95.6%). (FAB-MS in glycerol, MH+ at m/z 321). This product was used as such for the next reaction step. (ii) TMPP-Ac-SC6F5 Bromide. To a solution of 122 mg (0.38 mmol) of the above described intermediate in 0.5 mL of benzene was added 195 mg (0.37 mmol) of tris(2,4,6-trimethoxyphenyl)phosphine (TMPP) in 2 mL of benzene; a voluminuous white precipitate was formed immediately. After standing at room temperature for 0.5 h, the mixture was diluted with an equal volume of hexane. The solvent was decanted after a brief vortexing. This washing process was repeated three more times to remove any unreacted material. After drying overnight in a vacuum desiccator over phosphorus pentoxide and NaOH pellets, the pure product (198 mg, 62%) was obtained as a white crystalline powder: FAB-MS [matrix, m-nitrobenzyl alcohol (NBA) or glycerol] m/z 773 (C+, formula weight of the [TMPP-AcSC6F5]+ cation, intense), 573 [TMPP-CHdCdO]+, 533 [TMPP + H]+, 365 [(MeO)3Ph]2P+, m/z 181 (MeO)3PhCH2+ (intense). (For FAB fragmentation of alkyl triphenylphosphonium salts, see ref 22.) Nanomole-Scale Derivatization. Working solutions of TMPPAc-SC6F5 bromide and DMAP were prepared by dissolving the reagents in 1:1 (v/v) acetonitrile/water at a concentration of 10 nmol/mL. A typical procedure for derivatization at the lownanomole level follows: To a solution of 1 nmol (1 µL of 1 nmol/ µL) of peptide in acetonitrile/water (1:1, v/v) in an Eppendorf tube was added TMPP-Ac-SC6F5 bromide (0.5 µL of 10 nmol/µL) followed by DMAP (1 µL of 10 nmol/µL). The mixture was vortexed for 1 min. After standing at room temperature for 15 min, the mixture was analyzed by HPLC. Alternatively, the solution was diluted with acetonitrile/water to a final concentration of 50 pmol/µL (or 2 pmol/µL) for analysis by FAB- (or MALDI-) MS. Picomole-scale Derivatization. To a solution of the peptide (1 µL of 10 pmol/µL in aqueous acetonitrile, 1:1, v/v) was added TMPP-Ac-SC6F5 (1 µL of 100 pmol/µL) followed by DMAP (1 mL of 5 nmol/µL). After standing at room temperaure for 15 min, the mixture was diluted with 7 µL of a saturated solution of R-cyano-4-hydroxycinnamic acid (R-CHCA) in 1:1 (v/v) acetonitrile/0.1% trifluoroacetic acid (TFA). An aliquot of 2 µL (2 pmol equiv) was taken to record the MALDI spectra. HPLC. HPLC was used to assess the optimization and efficiency of the coupling reaction, as well as for purification of (20) Stewart, F. D.; Mathes, M. A. J. Org. Chem. 1949, 14, 1111. (21) Dalgliesh, C. E.; Mann, F. G. J. Chem. Soc. 1947, 559. (22) Claereboudt, J.; Baeten, W.; Geise, H.; Claeys, M. Org. Mass Spectrom. 1993, 28, 71.

Table 1. Comparison of the N-Terminal Charged Derivatization through Amidation Reactions method

stoichiometry (molar ratio)

scale

purification

carbodiimide-mediated couplinga 5 two-step protocolb 8 N-hydroxysuccinimide ester couplingc,d 15,16

[pepide]:[betaine]:[EDC] ) 1:103:103 [pepide]:[IAA]:[Me3N] ) 1:50:104 [pepide]:[TPP+(CH2)4CO2Su]:[NaOH] ) 1:2:1 [peptide]:[Me3N+(CH2)5CO2Su]:[Me3NH2CO3] ) 1:105:2 × 107 [peptide]:[TMPP+-AcSC6F5]:[DMAP] ) 1:5:10 or ) 1:10:500

50 nmol 0.1-2 nmol 500 nmol 50 fmol 0.1-2 nmol 10 pmol

RP-HPLC RP-HPLC RP-HPLC no workup no workup no workup

pentafluorothiophenol ester couplinge

a Peptide in THF was treated with (i) betaine, (ii) 1% HCl, and (iii) EDC to final pH 4.5, room temperature, 2 h. b Peptide in 0.3 M MES buffer, pH 6, was treated with (i) iodoacetic anhydride (IAA), 0 °C, 5 min, and (ii) aqueous Me3N, 37 °C, 2 h. c Peptide in water was treated with (i) equimolar amount of 1 M NaOH and (ii) TPP+(CH2)4CO2Su, room temperature, 2 h. To date, this protocol has been used mainly for the modification of peptidoleukotrienes. d The only example of subpicomole reaction reported by the authors employed Me3N+(CH)5CO2Su in aqueous (Me3NH)HCO3 (pH 8.5) for direct modification of [Glu1]fibrinopeptide B at 0 °C, 10 min. e For details, see Experimental Section.

charged derivatives. The HPLC system consisted of two Waters Model 6000A pumps controlled by a Waters Millennium data system. All experiments were performed on a Vydac C18 reversephase column (5-µm particle size, 4.6 × 250 mm) with the eluant monitored at 215 nm. The mobile phase contained the following: (A) 0.1% TFA/acetonitrile (90:10, v/v); (B) 0.1% TFA/ acetonitrile (10:90, v/v). The gradient elution was carried out at a flow rate of 1 mL/min with gradient 0-70% B in 30 min; peak areas were calculated by the data system. Mass Spectrometry. FAB-MS spectra were acquired on a JEOL HX-110 double-focusing mass spectrometer. Ionization was effected by bombardment of the sample/matrix solution with a 6-keV beam of xenon atoms. Sample solutions (0.5-2 µL, or 25100 pmol equiv) were mixed on the probe tip, which had been preloaded with 1 µL of the matrix [thioglycerol (TG) + 2-hydroxyethyl disulfide (HEDS), 1:1, v/v], which has the capacity to produce stronger signals (2-5-fold) for quaternary ammonium and phosphonium salts than the commonly used matrices glycerol and m-nitrobenzyl alcohol. For CAD-MS/MS, helium was used as the collision gas in a cell located in the first field-free region. The helium pressure was adjusted to reduce the abundance of the presursor ion by 50%. Linked scans were generated by the JEOL MS-MP8020D data system. MALDI mass spectra were recorded on a Voyager Elite timeof-flight (TOF) mass spectrometer (Persepive Biosystems, Boston, MA) equipped with a N2 laser (337 nm). Data were acquired in the linear mode of operation. Time-to-mass conversion was achieved by either external or internal calibration using bradykinin, [M + H]+ at m/z 1061.2, and insulin, [M + H]+ at m/z 5734.5, as standards. To prepare a sample for MALDI measurement, 1 µL of the analyte containing 2 pmol equiv of the derivative and 1 µL of a saturated solution of R-CHCA in 1:1 acetonitrile/0.1% TFA were mixed and briefly vortexed. Two microliters of this solution was applied to the sample plate and allowed to dry in air before being introduced into the mass spectrometer. RESULTS AND DISCUSSION Reagents. Three desirable attributes of a charged derivative of a peptide are as follows: (i) efficient preparation, including high yield and simplicity of the derivatization procedure, (ii) specific fragmentation to afford easy recognition of sequence ions, and (iii) selective modification of the N-terminus (or alternatively, the C-terminus). In our search for a more potent active ester than those commonly employed for N-terminal modification, such as hydroxy-

succinamide esters, we noted that the reactivity of thioesters (acyl thiols) far exceeds that of the corresponding oxygen esters in aminolysis23,24 (for example, acyl-S-coenzyme A shows up to a 100fold higher reactivity toward amines relative to that of oxygen esters25) and in condensation with peptides.26,27 Furthermore, pentafluorothiophenol active esters have recently been reported to be superior to many active esters (e.g., (EtO)2POCN, (PhO)2PON3, mixed carbonic anhydrides, pentafluorophenol esters, and the like) for amidation reactions.28 Therefore, we synthesized the intemediate BrCH2COSC6F5 by direct esterifiation of pentafluorothiophenol with bromoacetyl bomide. Attempted quaternization of the above-mentioned intermediate with common triarylphosphines (e.g., Ar ) p-MeOC6H4, R-furyl, C6H5, FC6H4, and C6F5) resulted in very poor yields of products because of their inherent low basicities. This observation prompted us to turn to TMPP. The latter is a nucleophile with high basicity (the highest for all known phosphines)29 apparently owing to the multiple MeO substituents (positive mesomeric effect), even though the bulky Ar3P ring system is sterically hindered (a pseudopropeller arrangement).

Unlike its phosphine congeners, TMPP reacted instantaneously with BrCH2COSC6F5 to form the expected TMPP+-CH2CO-SC6F5 bromide (TMPP-Ac-SC6F5 bromide) in good yield (Scheme 1). In addition to the simplicity of reagent preparation, the improved solubility of TMPP-Ac-SC6F5 in aqueous environments, relative to other triarylphosphonium reagents, is particularly beneficial in the derivatization of peptides. The amphithetic character of the TMPP head group also maintains the solubility of the peptide derivative in aqueous and aqueous organic solvents. We noted also that modified peptides having a Me3N+(CH2)5CO moiety at the N-terminus have been reported to contribute additional fragment ions [b - 59]+ in PSD spectra,18 presumably (23) Douglas, K. T. Acc. Chem. Res. 1986, 19, 186. (24) Bruce, T. C. In Organic Sulfur Compounds; Kharasch, N., Ed.; Pergamon Press: New York, 1961; Vol. 1, pp 421. (25) Tamvakopoulos, C. S.; Anderson, V. E. Anal. Biochem. 1992, 200, 381. (26) Schmidt, U.; B. Potzolli, B. Liebigs Ann. Chem. 1987, 935. (27) Lloyd, K.; Young, G. T. J. Chem.Soc. C 1971, 2890. (28) Davis, A. P.; Walsh, J. J. Tetrahedron Lett. 1994, 35, 4865. (29) Wada, M.; Higashizaki, S. J. Chem. Soc., Chem. Commun. 1984, 482.

Analytical Chemistry, Vol. 69, No. 2, January 15, 1997

139

Table 2. HPLC Conversion of Selected Peptides to TMPP-Ac Derivativesa peptide VGVAPG sleep-inducing peptide (WAGGDASGE) bradykinin (RPPGFSPFR) small cardioactive peptide B (MNYLAFPRM-NH2) neuromedin B (GNLWATGHFM-NH2) buccalin (GMDSLAFSGGL-NH2) granuliberin R (FGFLPIYRRPAS-NH2) allatostatin I (APSGAQRLYGFGL-NH2) icaria chemotactic peptide (IVPFLGPLLGLLT-NH2) a

no. of amino % conversion acid residues (by HPLC) 6 9 9 9

95 95 64 76

10 11 12 13 13

95 93 95 95 96

[Peptide]:[reagent]:[DMAP] ) 1:5:10 (molar ratio).

due to cyclization/elimination (repeated loss of Me3N neutrals), which can complicate spectral interpretation. In view of the potential complications caused by similar fragmentation, the preparation of charged derivatization agents with longer spacers between the charged group and the linking group was not pursued. The successful preparation of the unusual derivatizing agent, TMPP-Ac-SC6F5 bromide, opens the possibility for studying a new type of modified peptides that are linked to a charged, amphithetic head group. Coupling Reaction. TMPP-AcSC6F5 bromide is a white, crystalline powder; it is soluble in aqueous alcohols, alcohols, and aprotic solvents, including tetrahydrofuran and acetonitrile. Working solutions are prepared in aqueous acetonitrile or aqueous THF, which show a pH around 6.5-7.0. These solutions can be stored in a freezer for up to six months without appreciable decomposition. Nevertheless, TMPP-AcSC6F5 readily reacts with NH3 and primary and secondary amines, giving rise to the corresponding product, TMPP+CH2CONRR′ (R, R′ ) H or alkyl). In the presence of a basic catalyst like Et3N, it reacts also, though more slowly, with water and alcohols to form TMPP+CH2CO2R, where R ) H or alkyl. Practically no coupling occurs in the absence of a base. A number of bases and buffer solutions have been tested to facilitate the condensation reaction under conditions that maintain the compatible solubility of both the peptide and coupling agent. Included in the trials were (p-N,N-dimethylamino)pyridine (DMAP, a hypernucleophilic acylation promoter), Et3N, iPr2NEt, pyridine, 0.5 M NaHCO3 (pH 8.1), 0.2 M phosphate (pH 7.5-8.2), and 0.1 M Na2B4O7 (pH 9.5). Condensation can be readily achieved by using DMAP, or alternately, by aqueous NaHCO3, or phosphate. Aqueous NH4HCO3 (pH 7.5-8) is effective, but it consumes a large proportion of the reagent owing to the fast competitive reaction (amidation). The course of the reaction was followed by HPLC. The completeness of reaction was measured on the basis of the diminution of the peak corresponding to the underivatized peptide under conditions where different quantities of coupling reagent and base (or buffer) were used. Based on the data obtained from an evaluation study (Table 2), it was established that the coupling at the nanomole level is complete (generally over 90% conversion) at room temperature for 15 min under conditions where peptide: reagent:DMAP ) 1:5:10 (molar ratio), while the resulting pH remains in the range of 6.5-9. In our experience, failure to control the pH may result in incomplete derivatization. 140 Analytical Chemistry, Vol. 69, No. 2, January 15, 1997

Figure 1. Mass spectra of TMPP-acetyl neuromedin B recorded by MALDI-MS (A) [C+ 1705.9 (average); matrix, R-CHCA, 2 pmol loaded from a mixture of 10-pmol reaction] and FAB-MS (B) [C+ 1704.7 (monoisotopic); matrix, TG + HEDS, 50 pmol equiv loaded]. Monoisotopic masses for FAB mass spectra are used in all FAB spectra throughout.

Direct Analysis by FAB and MALDI-MS. Under FAB conditions TMPP-Ac derivatives provide an intense molecular cation peak (C+) 572 u above that representing the MH+ ion of the underivatized peptide. As nearly quantitative modification can be achieved with a 5-10-fold molar ratio of the reagent and base over the analytes, this relatively minor excess allows direct analysis of the reaction mixtures. In general, an aliquot containing 2550 pmol of the derivative is sufficient for producing a good-quality FAB spectrum, or 50-100 pmol to obtain a MS/MS spectrum. Similarly, an aliquot containing 1-2 pmol equiv of the peptide derivative was used to generate a spectrum by MALDI-MS, without additional workup. We have found that peptides composed mainly of amino acids that lack a polar side chain respond poorly to FAB and MALDI. However, these peptides (e.g., penta-Gly, hexa-Ala, VGVAPG, and icaria chemotactic peptide IVPFLGPLLGLLT amide) showed significant enhancement (up to 10-20-fold) upon TMPP-Ac derivatization. In agreement with the results of Vath and Biemann6 and Stults10, signal enhancement is not always observed. A comparison of the abundance of the protonated molecules (MH+) of 12 randomly selected peptides (Mr 500-1500 Da) to that of the molecular cations (C+) of their TMPP-Ac derivatives under FAB conditions indicated only three exhibited significant enhancement (factor of 2-15-fold based upon peak intensities of the MH+ and C+); the remaining 9 peptides showed an insignificant increase in ion abundance (enhancement factor 0.7-1.5). The procedure described in the Experimental Section works well in most cases for reactions at the 0.1-2-nmol level. However, this stoichiometry is not effective when applied to picomole-scale reactions because the concentration of base is too low to initiate the coupling. A modified procedure for derivatization using only 10 pmol of the starting material has been developed that is based on a slightly different stoichiometry of peptide:reagent:DMAP ) 1:10:500 (mole ratio). An example is given for the TMPP-Ac derivative of neuromedin B in Figure 1, obtained either by the picomole-scale procedure and analyzed by MALDI-MS (A) or by the nanomole-scale procedure and analyzed by FAB-MS (B), respectively, without further purification.

Table 3. TMPP Acetylation of Peptides for Direct Analysis by FAB- and MALDI-MS peptide Val-Gly-Val-Ala-Pro-Gly Leu enkephalinamide Leu enkephalin-Arg sleep-inducing peptide des-Arg1-bradykinin des-Arg9-bradykinin Val-His-Leu-Thr-Pro-Val-Glu-Lys angiotensin III allatostatin IV angiotensin II buccalin bradykinin allatostatin II des-pGlu1-leuteinizing hormone releasing hormone neuromedin B small cardioactive peptide B des-Asp1-angiotensin I Ile-Ser-bradykinin Ala-Pro-Gly-[Ile3,Val5]-angiotensin II allatostatin I icaria chemotactic peptide granuliberin R bovine adrenal medula dodecapeptide mastoparan ves. somatostatin Tyr-somatostatin

Mr nativea

C+ deriva

A:B ratiob

nanomole (0.1-2-nmol reaction)c

picomole (10-pmol reaction)d

VGVAPG YGGFL-NH2 YGGFLR WAGGDASGE PPGFSPFR RPPGSFPF VHLTPVEK RVYIHPF DRLYSFGL-NH2 DRVYIHPF GMDSLAFSGGL-NH2 RPPGFSPFR GDGRLYAFGL-NH2 HWSYGLRPG-NH2

498.6 554.6 711.8 848.8 904.0 904.0 922.1 931.1 969.1 1046.2 1053.2 1060.2 1067.2 1071.2

1072.2 1128.2 1285.4 1422.4 1477.6 1477.6 1495.7 1504.7 1542.7 1619.8 1626.8 1633.8 1640.8 1644.8

0:0 0:0 0:1 2:0 0:1 0:1 1:2 0:2 1:1 1:2 1:0 0:2 1:1 0:2

yes yes yes yes yes nd yes yes nd nd yes nd yes yes

+++ +++ +++ +++ +++ nt +++ ++e nt nt +++ nt ++ ++e

GNLWATGHFM-NH2 MNYLAFPRM-NH2 RVYIHPFHL ISRPPGFSPFR APGDRIYVHPF APSGAQRLYGFGL-NH2 IVPFLGPLLGLLT-NH2 FGFLPIYRRPAS-NH2 YGGFMRRVRPE INLKALAALAKKIL-NH2 AGCKNFFWKTFTSC (Disulfide bridge 3-14) YAGCKNFFWKTFTSC (Disulfide bridge 3-14)

1132.2 1141.4 1181.4 1260.5 1271.4 1335.5 1351.7 1422.7 1424.6 1478.9 1637.9

1705.9 1715.0 1755.0 1834.1 1845.0 1909.1 1925.3 1996.3 1998.2 2052.5 2211.5

0:1 0:1 0:3 0:2 1:2 0:1 0:0 0:2 1:3 0:3 0:2

yes yes yes yes yes yes yes yes yes yes yes

+++ ++ ++ ++ ++ +++ +++ +++ + +++ +++

1801.1

2374.7

0:2

nt

+++

sequence

a Average molecular (and cationic) weights. b Ratio of [no. of acidic amino acids]/[no. of basic amino acids]. c Products analyzed by FAB without cleanup. Generally, 50-100 pmol of the analyte is needed to obtain a MS/MS spectrum. Key: yes, good reaction; nd, not detected; nt, not tested. d Products analyzed by MALDI (2 pmol) without cleanup: +++ strong, ++ moderate, + weak; nt, not tested. e Moderate to strong signals can be obtained by an alternative two-step procedure (in aqueous acetonitrile 1:1, v/v): To a peptide (2 µL × 1 nmol/µL) was added in order (i) DMAP (2 µL × 10 nmol/µL), (ii) iodoacetic acid N-hydroxysuccinimide ester (IAAOSu, 1 µL × 10 nmol/µL), and (iii) TMPP (1 µL × 10 nmol/µL). After 15 min at room temperature, an aliquot containing the derivatized product was analyzed directly by FAB (0.1 nmol) or MALDI (2 pmol).

As summarized in Table 3, a wide range of peptides has been successfully modified by the protocols described above. We have observed that the reactivity of this novel reagent is not noticeably reduced for N-terminal amino acids with bulky side chains (e.g., Val, Ile). Furthermore, N-TMPP-Ac derivatives are obtained as the only products from peptides involving aliphatic and/or aromatic HO groups (Ser, Thr, Tyr). An important aspect of the present methodology is the apparent selective derivatization at the N-terminal for lysine-containing peptides. Using the TMAA technique, Vath and Biemann6 observed that a mixture of monoand diderivatized species formed when VHLTPVEK and INLKALAALAKKIL amide (mastoparan, a tetradecapeptide from Vespula lewisii) were derivatized (3:1 and 5:1 ratios by FAB ion abundance, respectively). In contrast, we observed only mono-TMPP-Ac adducts of these two peptides by the DMAP-promoted protocol (detected by FAB- or MALDI-MS). Similar directed modifications were obtained in a phosphate buffer medium at pH 7.6 or 8.2. We also found that these two peptides provided the mono-TMPPAc products only at the N-terminus of the chain as indicated by FAB-CAD-MS/MS (see Figure 5B) and MALDI-PSD analyses.30 The TMPP-Ac derivatization reaction can be adversely influenced by the presence of certain amino acids located at the N-terminus. For example, bradykinin and des-Arg9-bradykinin, both bearing an N-terminal Arg in the chain, reacted poorly with (30) Watson, J. T.; Huang, Z. H.; Liao, P. C.; Wu, J.; Roth, K. D. W.; Lavine, G.; Gage, D. A.; Allison, J. Proceedings of the 44th ASMS Conference on Mass Spectrometry and Allied Topics; Portland, OR, May 12-16, 1996; p 376.

TMPP-AcSC6F5, possibly due to an enhanced rate of hydrolysis of the reagent as indicated by the rapid disappearance of the peak for the reagent at m/z 773 during analysis by FAB- or MALDIMS. At present, we cannot explain this result, but it is unlikely due to charge repulsion between the reagent and N-terminal Arg alone because some N-Arg peptides (e.g., des-Asp1-angiotensin I and angiotensin III) react efficiently under the same conditions. On the other hand, failure of TMPP-Ac deriviatization in such cases is not serious, as peptides with Arg at or near the N-terminus produce reasonably interpretable CAD spectra from the underivatized peptide. Angiotensin I, angiotensin II (but not their desAsp1-analogs), and allatostatin IV, all possessing contiguous AspArg residues at their N-termini, are some other examples for which the derivatization fails. In these cases, derivatization can be effected by a two-step alternative procedure by iodoacetylation [iodoacetic acid ester with N-hydroxysuccinimide (ICH2COOSu) in the presence of DMAP] followed by quaternization with the phosphine TMPP (see footnote e under Table 3) to give the expected C+ in moderate to high abundance. CAD-MS/MS Studies. It has been generally accepted that the fragmentation of a peptide backbone under CAD is strongly influenced by the position of basic amino acid residues within the chain. For example, a basic residue at the N-terminus preferentially yields a-, b-, and d-type ions. Similarly, a basic residue at the C-terminus preferentially yields v-, w-, y-, and z-type ions. Unfortunately, these ion series are often incomplete in the spectra of an underivatized peptide, and ambiguities may arise in the Analytical Chemistry, Vol. 69, No. 2, January 15, 1997

141

Figure 3. FAB-CAD-MS/MS spectrum of des-Asp1-angiotensin I, [M + H]+ m/z 1181.7 (A) and its derivative, C+ m/z 1753.8 (B).

Figure 2. FAB-CAD-MS/MS spectra of neuromedin B, [M + H]+ m/z 1132.5 (A) and its TMPP-acetyl-derivative C+ m/z 1704.7 (B and C) recorded by linked scanning with constant B/E ratio. (C) is the lower part of the MS/MS spectum (m/z 0-600) in which the major peaks at m/z 573 and 181 represent characteristic ions of the TMPPacetyl moiety and are produced according to the following pathway (Ar ) 2,4,6-trimethoxyphenyl):22

[(i) skeletal rearrangement, (ii) loss of (Ar2P - H); (iii) elimination of CO)].

interpretation of CAD mass spectra with regard to the sequence deduced. Furthermore, a basic residue in the middle of a peptide may yield a more complex pattern of both N- and C-terminal ions, making the sequence determination more difficult.31 To explore the influence of TMPP-Ac derivatization on sequencespecific fragmentation, we chose to examine several representative peptides that have no basic amino acids or have a basic amino acid residue (mainly Arg) in different positions along the chain. Figure 2 presents the CAD-B/E linked scan spectra of neuromedin B (A) and its N-TMPP-Ac derivative (B, C). Because the fixed charge has a stronger directing influence on the fragmentation than any basic residues that may be present, the appearance of the spectrum (Figure 2A) is dramatically altered. Three basic ion species were observed in the spectrum of the resulting charged derivative (Figure 2B): (i) Formation of a-type ions dominate as displayed in Figure 2C and subsequent figures. Ions formed from cleavage on the (31) Johnson, R. S.; Martin, S. A.; Biemann, K.; Stults, J. T.; Watson, J. T. Anal. Chem. 1987, 59, 2621.

142 Analytical Chemistry, Vol. 69, No. 2, January 15, 1997

C-terminal side of Gly and Ala may generally be weak or in some cases not detectable. (ii) Prominent d-type ions occur due to the elimination of the side chain through β-cleavage from certain amino acid residues. Characteristic peaks 43 (d2), 42 (d3), 14 (d6), and 16 u (d6′) can be observed in Figure 2B below the peaks for a-type ions due to Asn, Leu, and Thr, respectively. Similarly, other amino acids exhibit peaks for specific side-chain losses to form d-type ions as shown in subsequent figures; for example, the peak 85 u below the peak for the a-type ion for Arg (Figures 3B and 6B), 14 u for Val (Figures 3B and 5B), 14/28 u for Ile (Figure 3B), and 16 u for Ser (Figure 4B). For certain amino acids like Asn and Arg, additional loss of NH3 may be observed from a-type ions. These side-chain fragments can be of significant importance for the identification of specific amino acid residues. (iii) Formation of c-type ions, or [a + 45]+, is found to accompany certain amino acids, for example, Thr and Ser (Figures 2B, 4B, and 6B). This is in agreement with the observation by Downard and Biemann,32 that, in high-energy CAD, Ser, Thr, and Asp (and others) appeared to promote the production of c-type ions from the amino acid preceding it. Figure 3 presents the CAD-MS/MS spectra of des-Asp1angiotensin I and its TMPP-Ac-derivative. It has been described previously that peptides involving an N-terminal Arg (for example, des-Asp1-angiotensin I, angiotensin III, des-Arg9-bradykinin, substance P, ACTH fragment 18-39, and peptides from chymotryptic digests)33-36 were able to provide a-type ions because of the preferential charge localization at the guanidine residue which strongly drives the charge-remote fragmentation in a well-defined pattern. However, such N-terminal ion series were often incomplete, thus leaving uncertainties in spectral interpretation (Figure 3A). With the incorporation of an N-terminal TMPP-Ac group, a (32) Downard, K. M.; Biemann, K. J. Am. Soc. Mass Spectrom. 1993, 4, 874. (33) Biemann, K.; Papayannopoulos, I. A. Acc. Chem. Res. 1994, 27, 370-378. (34) Bean, M. F.; Carr, S.; Thorne, G. C.; Reilly, M. H.; Gaskell, S. J. Anal. Chem. 1991, 63, 1473. (35) Johnson, R. S.; Martin, S. A.; Biemann, K. Int. J. Mass Spectrom. Ion Processes 1988, 86, 137. (36) Rouse, J. C.; Yu, W.; Martin, S. A. J. Amer. Soc. Mass Spectrom. 1995, 6, 822.

Figure 4. FAB-CAD-MS/MS spectra of des-Arg1-bradykinin, [M + H]+ m/z 904.5 (A) and its derivative, C+ m/z 1476.7 (B). The peak m/z 1321 represents the specific fragment due to elimination of (Arg - H2O) or 156 Da from the C+.

complete a-type ion series is clearly discernible (Figure 3B), allowing an uninterrupted series of sequence ions to be clearly recognized. Figure 4 presents the CAD spectra of des-Arg1-bradykinin (Figure 4A) and its TMPP-Ac-derivative (Figure 4B), an example with an Arg residue at the C-terminal position. It has been established that underivatized des-Arg1-bradykinin produced predominantly C-terminal ions under high-energy CAD;36 similar trends also have been observed for a number of peptides involving a C-terminal Arg under CAD-MALDI conditions.37 The strength of the TMPP-Ac moiety in directing fragmentation during CAD is underscored by the data presented in Figure 4B, which indicate that the TMPP-Ac charged group at the N-terminus overcomes the usual control of fragmentation that would otherwise be directed by the protonated Arg at the C-terminus in the underivatized peptide. As a result, all product ions in the CAD spectrum of the TMPP-Ac derivative retain the N-terminus. Moreover, a loss of the C-terminal Arg residue at m/z 1321 in Figure 4B leads to the formation of abundant [C+ - 156]+ or [a7 + 46]+ ) [b7 + 18]+. Rearrangement product ions of this type were reported to take place under metastable transition or low-energy CAD for peptides with an Arg at the C-terminus, provided the fragmentation is N-terminally directed.38 Figure 5 depicts the CAD spectra of the amino-terminal octapeptide of the β-chain of HbS (sickle cell hemoglobin), ValHis-Leu-Thr-Pro-Val-Glu-Lys (A), and its TMPP-Ac derivative (B). Similar to the example described above (Figure 4), the prominent b and y ions caused by the C-terminal Lys of the underivatized peptide (Figure 5A) are eliminated by the incorporation of a TMPP-Ac at the N-terminus to produce exclusively a (and d) ions. Figure 6 presents CAD spectra of allatostatin I (A) and its TMPP-Ac derivative (B). The scattered distribution of fragment ions in the spectrum of the underivatized counterpart caused by the presence of an internal Arg is significantly changed when an (37) Qin, J.; Chait, B. T. J. Am. Chem. Soc. 1995, 117, 5411. (38) Thorne, G. C.; Ballard, K. D.; Gaskell, S. J. J. Am. Soc. Mass Spectrom. 1990, 1, 249.

Figure 5. FAB-CAD-MS/MS spectra of Val-His-Leu-Thr-Pro-ValGlu-Lys, [M + H]+ m/z 922.5 (A) and its derivative, C+ m/z 1494.7 (B).

Figure 6. FAB-CAD-MS/MS spectra of allatostatin I, [M + H]+ m/z 1335.7 (A) and its derivative, C+ m/z 1907.9 (B).

N-terminal TMPP-Ac is incorporated into the molecule, giving rise to a well-controlled pattern of N-terminal sequence fragments. In general, the improved quality of structural information available in the CAD spectra of the TMPP-Ac derivatives more than offsets any diminution in the CAD fragmentation process that might be expected from having added 573 Da to the mass of the peptide CONCLUSION In summary, the results of this study demonstrate the reactivity of the TMPP-AcSC6F5 reagent and analytical advantages of the resulting N-terminal derivative in structure elucidation of peptides. The involvement of a amphithetic TMPP moiety as the head group and a reactive thioester as the anchor site in the newly developed reagent has greatly facilitated the derivatization and directed subsequent fragmentation. Two major advantages accrue in the use of TMPP-AcSC6F5 bromide in the analysis of peptides by desorption/ionization-CAD-MS/MS. First, the chemistry is efficient and easy to handle as apparent from the comparisons Analytical Chemistry, Vol. 69, No. 2, January 15, 1997

143

presented in Table 1. Second, the CAD spectra are sequencespecific and are independent of the presence and/or position of basic amino acids (mainly Arg) in the peptide chain. In addition, the sequence ions are not obscured by ions formed by the loss of the charged head group. Further studies on the scope and limitations of TMPP-Ac derivatization methodology, in particular selective N-terminal derivatization for use in the analysis of Lys-containing peptide mixtures, are in progress. There is some hope that this derivatization approach may be general for Lys-containing peptides, but further modification of the reaction conditions (e.g., rigorous pH control) may be required.

144

Analytical Chemistry, Vol. 69, No. 2, January 15, 1997

ACKNOWLEDGMENT This work was supported by NIH Grant RR00480 from the National Center for Research Resources (to J.T.W.), which supports, in part, the MSU-NIH Mass Spectrometry Facility.

Received for review August 21, 1996. Accepted October 29, 1996.X AC9608578

X

Abstract published in Advance ACS Abstracts, December 15, 1996.