Anal. Chem. 1998, 70, 5322-5325
Technical Notes
Matrix-Assisted Laser Desorption/Ionization Mass Spectrometry of Hydrophobic Peptides Kari B. Green-Church and Patrick A. Limbach*
Department of Chemistry, 232 Choppin Hall, Louisiana State University, Baton Rouge, Louisiana 70803
Hydrophobic peptides, especially those with acid-labile protecting groups, are difficult to characterize using mass spectrometric methods. We have developed a new procedure for matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) mass spectrometric analysis of such samples. Hydrophobic peptides, which are insoluble in aqueous solutions, are dissolved in chloroform and combined with matrixes prepared in chloroform or chloroform/methanol solutions. The use of a common solvent for the matrix and the analyte improves the analyte isolation step in MALDI mass spectrometry. The lack of acidic solutions previously used for electrospray ionization or MALDI mass spectrometry of hydrophobic peptides extends this methodology to cyclic or protected hydrophobic peptides. Conventional peptide matrixes, such as 2,5-dihydroxybenzoic acid and sinapinic acid, as well as 3-indoleacrylic acid are shown to be suitable for hydrophobic peptides. Cyclic hydrophobic peptides and linear hydrophobic peptides with blocked termini are detected as the cation-adducted pseudomolecular ion due to the lack of suitable sites of protonation on the analyte. In this work, we demonstrate a new procedure for the analysis of hydrophobic peptides using matrix-assisted laser desorption/ ionization mass spectrometry (MALDI-MS). In this method, matrixes soluble in chloroform or chloroform/methanol solutions are used, permitting the sample and matrix solutions to be mixed prior to spotting on the MALDI sample plate. The lack of acidic solutions for solubilizing the peptide allows this method to be used with hydrophobic peptides containing acid-labile protecting groups. Ionization of hydrophobic peptides in which both the N-terminus and C-terminus positions are blocked, e.g., cyclic peptides, is achieved via cationization with residual salt from the matrix. Hydrophobic peptides and proteins are important in many biological systems. Examples include proteins associated with the regulation of cell membranes1-3 and antimicrobial R-helical4,5 and * To whom correspondence should be addressed. Phone (504) 388-3417. Fax (504) 388-3458. E-mail
[email protected]. (1) Schaller, J.; Pellascio, B. C.; Schlunegger, U. P. Rapid Commun. Mass Spectrom. 1997, 11, 418-426. (2) Schey, K. L.; Papac, D. I.; Knapp, D. R.; Crouch, R. K. Biophysics 1992, 63, 1240-1243. (3) Schindler, P. A.; Van Doresselaer, A.; Falick, A. M. Anal. Biochem. 1993, 213, 256-263.
5322 Analytical Chemistry, Vol. 70, No. 24, December 15, 1998
cyclic peptides.6,7 MALDI-MS has become a routine technique for the characterization of peptides and proteins.8,9 The vast majority of peptides analyzed via MALDI-MS, however, are soluble in aqueous media. Hydrophobic peptides pose a unique problem for MALDI-MS due to their limited solubility in aqueous solutions, which limits the effectiveness of conventional peptide matrix/ sample preparation procedures. Hydrophobic peptides have been analyzed using electrospray ionization mass spectrometry (ESI-MS)1,3 and MALDI-MS.2,10 To facilitate the ESI mass spectrometric analysis of these peptides, the sample must be dissolved in formic acid or analyzed using a chloroform/methanol/water mixture containing glacial acetic acid or trifluoroacetic acid. For MALDI mass spectrometric analysis, the peptides are dissolved in aqueous formic acid, which permits their analysis with sinapinic acid. The methods to analyze hydrophobic peptides described above all use acidic conditions to solubilize the peptides in aqueous media. Acidic conditions can cause the loss of protecting groups from synthetic peptides and can potentially result in further decomposition of the sample. Such extreme sample preparation conditions can significantly impair the ability to confirm sample purity and identity of synthetic hydrophobic peptides using MALDI-MS. Protected hydrophilic peptides have been characterized using fast atom bombardment11 and MALDI-MS.12 Schmidt et al.12 found that matrix acidity in MALDI correlated to molecular ion stability. 2,5-Dihydroxybenzoic acid (DHB), the most acidic matrix studied, caused significant cleavage of the acid-labile side-chain protecting groups of synthetic peptides, while more neutral matrixes such (4) Blondelle, S. E.; Houghten, R. A. Biochemistry 1992, 31, 12688-12694. (5) Javapour, M. M.; Juban, M. M.; Lo, W.-C. J.; Bishop, S. M.; Alberty, B.; Cowell, S. M.; Becker, C. L.; McLaughlin, M. L. J. Med. Chem. 1996, 39, 3107-3113. (6) Kondejewski, L. H.; Farmer, S. W.; Wishart, D. S.; Hancock, R. E. W.; Hodges, R. S. Int. J. Pept. Protein Res. 1996, 47, 460-466. (7) Gibbs, A. C.; Kondejewski, L. H.; Gronwald, W.; Nip, A. M.; Hodges, R. S.; Skyes, B. D.; Wishart, D. S. Nature Struct. Biol. 1998, 5, 284-288. (8) Hillenkamp, F.; Karas, M.; Beavis, R. C.; Chait, B. T. Anal. Chem. 1991, 63, 1193A-1202A. (9) Roepstorff, P. Trends Anal. Chem. 1993, 12, 413-421. (10) Schey, K. L. Protein and Peptide Analysis by Mass Spectrometry; Methods in Molecular Biology 61; Humana Press: Totowa, NJ, 1996; pp 227-230. (11) Grandas, A.; Pedroso, E.; Figueras, A.; Rivera, J.; Giralt, E. Biomed. Environ. Mass Spectrom. 1988, 15, 681-684. (12) Schmidt, M.; Krause, E.; Beyermann, M.; Bienert, M. Pept. Res. 1995, 8, 238-242. 10.1021/ac980667s CCC: $15.00
© 1998 American Chemical Society Published on Web 11/12/1998
as 2,4,6-trihydroxyacetophenone or 2-amino-5-nitropyridine caused little fragmentation of the protected hydrophilic peptide. The general approach to the analysis of hydrophobic compounds by MALDI-MS has been to utilize nonaqueous solvents for preparation of both the matrix and analyte solutions. Juhasz and Costello prepared gangliosides in a chloroform/methanol mixture and several standard peptide matrixes in acetonitrile/ water mixtures for subsequent MALDI-MS experiments.13 Hydrophobic synthetic polymers have been detected with MALDI using typical peptide matrixes (DHB and sinapinic acid) dissolved in acetone.14 Presently, there is no suitable approach for the characterization of protected hydrophobic peptides or for the characterization of cyclic hydrophobic peptides that do not contain sites amenable to protonation. We have investigated the use of nonaqueous solvents for preparation of the matrix and protected hydrophobic peptide solution. The method described here extends further the analytical utility of MALDI-MS for peptide analysis. EXPERIMENTAL SECTION Materials. Dithranol, 3,5-dimethoxy-4-hydroxycinnamic acid (sinapinic acid, SA), 2,5-dihydroxybenzoic acid (DHB), R-cyano4-hydroxycinnamic acid (HCCA), and 3-indoleacrylic acid (IAA) were obtained from Aldrich (Milwaukee, WI) and were used without further purification. All solvents used were HPLC grade. Methanol was desalted using AG501-X8(d) resin beads (Bio-Rad, Hercules, CA). Approximately 50 mg of beads was added per 100 mL of solvent and allowed to stir for 1 h. Chloroform was desalted using ammonium carbonate. Activated AG50W-X8 cation-exchange resin beads (Bio-Rad) were added to the solvents or prepared samples prior to spotting on the MALDI sample plate where noted. All peptides used were provided by Dr. Maria Ngu-Schwemlein at Southern University. Abbreviations used in this paper for peptide protecting groups are as follow: O-phenacyl ester, O-Pa (C-terminus protecting group); tert-butoxycarbonyl, t-Boc (Nterminus protecting group); γ-O-benzyl, γ-O-Bzl (Glutamic acid residue protecting group); and 2-chlorocarboxybenzyl, 2-Cl-Cbz (lysine residue protecting group). Caution: Protective measures such as wearing goggles and a lab coat and performing all work in a well-ventilated fume hood should be practiced when handling chloroform. Mass Spectrometry. Hydrophobic peptides analyzed by MALDI-MS using the method described previously10 were prepared by dissolving the peptide in a 7:3 (v:v) solution of formic acid/hexafluoro-2-propanol at a 10 mM concentration. Then, 50 mM sinapinic acid, 100 mM DHB, and 100 mM HCCA matrixes were prepared in 70% formic acid. Next, 1-2 µL of a 500:1 matrix/ analyte solution was spotted onto the probe tip and allowed to dry. Peptides prepared using the method discussed here were prepared as 100 µM solutions in chloroform. The solution was sonicated for 1-2 min to ensure complete solubilization of the peptide. Sinapinic acid, DHB, and HCCA were prepared as 10 mM solutions in a 2:1 chloroform/methanol mixture. IAA and dithranol were prepared as 10 mM solutions in chloroform. Typically, 2-10 µL of the matrix solution and 1-2 µL of the sample solution were combined (resulting in a 500:1 matrix/ (13) Juhasz, P.; Costello, C. E. J. Am. Soc. Mass Spectrom. 1992, 3, 785-796. (14) Danis, P. O.; Karr, D. E. Org. Mass Spectrom. 1993, 28, 925-925.
Figure 1. (a) MALDI-TOF mass spectrum of the hydrophobic peptide t-Boc-Glu-(γ-O-Bzl)-Ala-Leu-Ala-COOCH2COPh dissolved in a 7:3 (v:v) solution of formic acid/hexafluoro-2-propanol and analyzed with 10 mM DHB in 70% formic acid. (b) The same peptide dissolved in chloroform and analyzed with 10 mM solution of DHB in chloroform. No molecular or pseudomolecular ions can be detected in (a) as the N-terminus protecting group is acid-labile. Abundant pseudomolecular ions corresponding to sodium and potassium adducts are observed in (b) along with some minor loss of the N-terminus protecting group. Peaks indicated by an asterisk are fragment ions.
analyte mole ratio), and 1-2 µL of this mixture was spotted onto the MALDI plate and allowed to air-dry. The MALDI-TOF experiments were performed on a PerSeptive Biosystems Inc. (Framingham, MA) Voyager linear MALDI-TOF instrument with an N2 laser. Laser power was set at the threshold level required to generate signal. Accelerating voltage was set to 28 kV. Each spectrum is an average of 32 scans, and each experiment was conducted three separate times at different locations on the sample spot to average results and ensure reproducibility. RESULTS AND DISCUSSION Investigations of Conventional Peptide Matrixes. Hydrophobic peptides pose a general problem for routine MALDI-MS analysis due to the difficulty in preparing such samples with aqueous solutions of matrixes. The procedure reported previously utilizes acidic solutions to facilitate matrix/analyte mixing.10 Unfortunately, such an approach is unsuitable for peptides with acid-labile protecting groups. Figure 1a is the mass spectral results obtained on a model hydrophobic peptide, t-Boc-Glu-(γ-O-Bzl)Ala-Leu-Ala-COOCH2COPh, prepared by dissolving the peptide in a formic acid/hexafluoro-2-propanol mixture and analyzed with DHB as the matrix. No molecular or pseudomolecular ions are observed, and the major analyte peak arises from the loss of the t-Boc group. Presumably, the t-Boc protecting group is lost from the peptide in the formic acid/hexafluoro-2-propanol solution used to dissolve the peptide in aqueous medium, although the protecting group may be lost during the desorption/ionization step.12 Analytical Chemistry, Vol. 70, No. 24, December 15, 1998
5323
Figure 2. MALDI-TOF mass spectra of the hydrophobic peptide t-Boc-Glu-(γ-O-Bzl)-Ala-Leu-Ala-COOCH2COPh dissolved in chloroform and analyzed with (a) sinapinic acid and (b) HCCA each in a 2:1 chloroform/methanol mixture. Peaks indicated by an asterisk are fragment ions. Unlabeled peaks are matrix ions.
Because the previous protocol results in the loss of acid-labile protecting groups from the peptide, a different approach to solubilizing the sample and matrix was sought. We initially focused our efforts on dissolving the standard peptide matrixes in chloroform or mixtures of chloroform and alcohols. Figure 1b is the mass spectrum obtained on the same peptide/matrix combination as in Figure 1a, but prepared in a chloroform/methanol solution. The acid-labile protecting group is still lost to a minor extent, but an abundant pseudomolecular ion is now detected. Two other common peptide matrixes, sinapinic acid and HCCA, also were prepared in a chloroform/methanol solution and used to analyze our model hydrophobic peptide (Figure 2). Sinapinic acid resulted in a more abundant (M - t-Boc + H)+ ion than DHB, and several fragment ions corresponding to cleavages along the peptide backbone are detected. The spectrum obtained using HCCA is characterized by a number of abundant peptide fragment ions. Thus, among the three common peptide matrixes investigated, DHB afforded the most abundant pseudomolecular ions of the analyte and the least amount of fragmentation. Because the standard peptide matrixes are not soluble in chloroform, the matrixes had to be prepared in a solvent mixture to ensure that they were at least partially dissolved in the solution. It was found that if the matrix was dissolved in a solution of chloroform and methanol, then abundant pseudomolecular ion signal from the hydrophobic peptide could be obtained. When the DHB matrix is dissolved in 100% methanol, a weak signal is obtained for the (M + Na)+ and (M - t-Boc + H)+ ions. This is likely due to the limited solubility of the peptide in methanol. When the volume of methanol is greater than that of chloroform, the (M - t-Boc + H)+ ion abundances are greater than the (M + Na)+ ion abundances. When the volume of chloroform exceeds 5324 Analytical Chemistry, Vol. 70, No. 24, December 15, 1998
Figure 3. MALDI-TOF mass spectra of the hydrophobic peptide t-Boc-Glu-(γ-O-Bzl)-Ala-Leu-Ala-COOCH2COPh dissolved in chloroform and analyzed with 10 mM solutions of (a) dithranol and (b) IAA in chloroform. Abundant pseudomolecular ions corresponding to sodium and potassium adducts are observed along with some loss of the N-terminus protecting group. Peaks indicated by an asterisk are fragment ions. Unlabeled peaks are matrix ions.
the volume of methanol in the solvent mixture, the (M + Na)+ ion is the most abundant peak in the mass spectrum. A 2:1 chloroform/methanol ratio was found to be the optimal solvent mixture because the (M + Na)+ ion is the most abundant peak in the mass spectrum, the (M - t-Boc + H)+ ion abundance is low, and the overall quality of the mass spectrum is sufficient for analytical use. The next series of experiments were performed to determine if fragmentation was dramatically influenced by the matrix-toanalyte ratios useds100:1, 500:1, 1000:1, 5000:1, and 10000:1 matrix/analyte mole ratios were investigated for DHB, sinapinic acid, and HCCA. No appreciable differences in the amount of fragmentation or quality of the spectra within each matrix investigated were seen during these studies (data not shown). Optimal results (i.e., minimal fragmentation and abundant pseudomolecular ion) were found with DHB in 500:1 and 1000:1 matrix/ analyte ratios. It was assumed that the sodium and potassium contributing to the production of the pseudomolecular ions arise from the solvents and matrixes used in these experiments. A series of experiments were performed to determine if desalting the matrix solution would adversely affect the pseudomolecular ion abundance. Desalting the matrix and solvents with ion-exchange resin beads prior to MALDI-MS resulted in a >50% reduction in the (M + Na)+ and (M + K)+ ion abundances (data not shown). Investigations of Alternative Matrixes. Because the cationized molecular ions (M + Na)+ and (M + K)+ are the preferential forms of the pseudomolecular ions of the hydrophobic peptides, it seemed reasonable to explore other matrixes which have a greater solubility in chloroform. For example, dithranol and IAA
Cyclic hydrophobic peptides, like the linear hydrophobic peptides having protecting groups at each terminus, do not have suitable sites of protonation for analysis via standard MALDI-MS techniques. Thus, such samples should be amenable to the protocol developed here. To demonstrate the applicability of this method to cyclic hydrophobic peptides, cyclo[-Ala-Leu-]2 (Figure 4a) and a cyclic peptide containing (2-Cl-Cbz-protected) Leu and Lys residues (Figure 4b) were prepared and analyzed using the same procedure as for the model peptide in Figure 1b. In Figure 4a, the most abundant ion detected in the mass spectrum is the (M + Na)+ ion, along with a small (M + H)+ ion and peaks corresponding to the dimer of this cyclic peptide. In Figure 4b, only the (M + Na)+ ion is detected. There was some question whether the sample analyzed in Figure 4b was a cyclic or linear peptide, as the difference between the sodiated cyclic peptide and the linear peptide is 5 u. Angiotensin (m/z 1297.5) was used to internally calibrate the mass spectrum, from which it was determined that the peak was the sodiated cyclic peptide rather than the protonated linear peptide.
Figure 4. MALDI-TOF mass spectra of cyclic hydrophobic peptides (a) cyclo[-Ala-Leu-]2 and (b) a cyclic peptide containing (2-Cl-Cbzprotected) Leu and Lys residues. Each cyclic peptide was prepared and analyzed using the same procedure as for the linear peptide in Figure 1b. In both cases, the most abundant ion is the (M + Na)+ peak.
have been used as matrixes, in conjunction with sodium, silver, or other salts, for MALDI-MS analysis of synthetic polymers.15-19 The model hydrophobic peptide, t-Boc-Glu-(γ-O-Bzl)-Ala-Leu-AlaCOOCH2COPh, was dissolved in a 2:1 chloroform/methanol solution and analyzed with IAA and dithranol, each prepared as 10 mM solutions in chloroform. As shown in Figure 3, the pseudomolecular ions, (M + Na)+ and (M + K)+, are the most abundant ions with these matrixes. Less fragmentation is seen when using IAA, as compared to dithranol, as a matrix, but both matrixes yield more fragment ions than DHB (cf. Figure 1b). Cyclic Hydrophobic Peptides. To date, there have been no published reports on the analysis of cyclic hydrophobic peptides. (15) Bahr, U.; Deppe, A.; Karas, M.; Hillenkamp, F.; Giessmann, U. Anal. Chem. 1992, 6, 2866-2869. (16) Belu, A. M.; DeSimone, J. M.; Linton, R. W.; Lange, G. W.; Frieman, R. M. J. Am. Soc. Mass Spectrom. 1996, 7, 11-24. (17) Scrivens, J. H.; Jackson, A. T.; Yates, H. T.; Green, M. R.; Critchley, G.; Brown, J.; Bateman, R. H.; Bowers, M. T.; Gidden, J. Int. J. Mass Spectrom. Ion Processes 1997, 165/166, 363-375. (18) Mowat, I. A.; Donovan, R. J.; Maier, R. R. J. Rapid Commun. Mass Spectrom. 1997, 11, 89-98. (19) Wong, C. K. L.; Chan, T. W. D. Rapid Commun. Mass Spectrom. 1997, 11, 513-519.
CONCLUSION A new protocol for the MALDI-MS characterization of hydrophobic peptides has been presented. Hydrophobic peptides with acid-labile protecting groups cannot be characterized using aqueous acid solutions. Several standard MALDI matrixes have been found to be sufficiently soluble in chloroform or chloroform/ methanol solutions, which permits the use of chloroform as the solvent for both the matrix and the peptide. The lack of suitable sites for protonation of such peptides requires the formation of cation-adducted pseudomolecular ions. Generation of an abundant (M + Na)+ or (M + K)+ ion from these peptides with minimal fragmentation allows for the confirmation of the synthesis procedure used in their preparation. The procedure described herein is suitable for the characterization of both linear and cyclic hydrophobic peptides. ACKNOWLEDGMENT The authors thank Dr. Maria Ngu-Schwemlein at Southern University, Baton Rouge, for providing the hydrophobic peptides used in this study and for her helpful discussions throughout this project. Financial support for this work was provided by the National Science Foundation Environmental Chemistry Initiative to LSU and Southern University (CHE-9634060) and Louisiana State University. Received for review June 18, 1998. Accepted October 14, 1998. AC980667S
Analytical Chemistry, Vol. 70, No. 24, December 15, 1998
5325
