Albert Einstein College of Medicine
Multiple Antigen Peptide System (MAPS) peptides were originally designed to produce peptide antigen carriers with a defined number of conjugated peptides (1, 2). The resulting haptens usually contain four or eight identical peptides in a dense structure than can simply be dialyzed before immunization. Antisera produced to MAPS peptides will not contain antibodies to the commonly used carrier proteins, such as keyhole limpet hemocyanin, although they may contain antibodies specific to the MAPS core structure. MAPS peptides have been designed that contain both B-cell and T-cell epitopes (1, 2). The ease in use for the investigator requiring sequence-specific antibodies places all responsibility for antigen preparation on the synthetic chemist.
MAPS peptides are usually prepared or purchased as either four-branch or eight-branch forms (Figure 1). A Lys residue with both the a- and e-amino groups blocked with the same labile protecting group is coupled to the b-Ala carboxyl-terminal residue. When deblocked, this Lys coupling is repeated once or twice so that either four or eight amino groups are available for further coupling reactions. Three residues of Gly per chain are sometimes added to provide a flexible link between the core of the MAPS matrix and the peptides to be synthesized. For amino acid analysis, the MAPS core adds 0.875 residues of Lys per chain and 0.125 residues of b-Ala per chain for the eight-chain form, or 0.75 Lys and 0.25 b-Ala for the four-chain form. For analysis by mass spectrometry (MS), the contribution of the MAPS core to the mass is 986 amu for the eight-chain form or 473 amu for the four-chain form.
Amino acid and Edman sequence analysis have been the primary analytical tools available for quality control of these peptides. Since MAPS peptides were first introduced, improvements in analytical technology, particularly MS, have raised the standards for preparing synthetic peptides in general. Studies conducted by the ABRF Peptide Synthesis Research Committee suggest the quality of peptides produced by member resource laboratories has improved concomitant with the increased availability of modern mass spectrometers (3, 4). However, analysis of MAPS peptides has remained a challenge because the product is rarely a homogeneous species.
The data shown in the examples below do not represent a systematic study of MAPS synthesis, an effort not possible within the context of our laboratory's operations. They represent our efforts to provide MAPS peptides as well characterized structurally as the other peptides that our laboratory synthesizes, and they may provide a frame of reference for other laboratories with the same goals. Over the past two years, our laboratory has synthesized 25 MAPS peptides. These were prepared by FastMoc procedures on an ABI 433A peptide synthesizer, with a ratio of activated amino acid to resin of 4:1 and with MAPS core resins purchased from Perkin Elmer/Applied Biosystems Division. The following sidechain protecting groups were used: t-butyl (tBu) for Asp, Glu, Ser, Thr, and Tyr; t-butyloxycarbonyl (Boc) for Lys; 2,2,5,7,8-pentamethylchroman-6-sulfonyl (Pmc) for Arg; and trityl (Trt) for Asn, Gln, and His. Amino acid analysis, reversed-phase high pressure liquid chromatography (HPLC), and electrospray ionization or matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (ESI-MS or MALDI-MS) were used to analyze the peptides. Amino acid analysis was performed on a Hewlett-Packard HP1090M AminoQuant system with fluorescence detection, and reversed-phase HPLC was carried out on a Hewlett-Packard HP1090M HPLC equipped with a diode array detector, using either a Vydac C18 (218TP54, 4.6 x 250 mm) or C4 (214TP5215, 2.1 x 150 mm) column. The columns were eluted with a gradient from 0.1% aqueous trifluoroacetic acid (TFA) to 100% acetonitrile/0.1% TFA at a flow rate of 0.75 ml/min (C18) or 0.2 ml/min (C4), increasing acetonitrile concentration at 1%/min, and monitored at both 214 and 280 nm. A PE-Sciex API-III mass spectrometer was used for ESI-MS, and a Perseptive Voyager instrument for MALDI-MS.
Figure 1: Schematic representation of MAPS peptides. Structure A: 4-MAPS peptide, structure B: 8-MAPS peptide, structures C and D: "positional isomers" of the 4-MAPS peptide. The cross-hatched circle represents the b-Ala residue, closed circles represent Lys residues, open circles represent the antigenic peptides, and triangles denote protecting groups.
When subjected to amino acid analysis, all the MAPS peptides had compositions with no more than 10% error for each amino acid. However, this type of error could mask a deletion or truncation present in only one of four or eight chains and would not reveal residues with protecting groups still attached. Our laboratory chooses not to analyze synthetic peptides by automated Edman sequence analysis to avoid any possible risk of contamination interfering with our high sensitivity sequencing work. However, sequence analysis would be similarly insensitive to detection and quantitation of small errors in these populations of molecules. Because our laboratory has had access to various types of mass spectrometers for eight years, we have generally synthesized to achieve the correct mass as determined by the spectrometer, or to "please" the mass spectrometer. Our first instrument was a Finnigan MAT-90 fast atom bombardment mass spectrometer, which was not suitable for analyzing peptides as large as most MAPS peptides. Approximately four years ago we acquired an ESI-MS instrument, and in the following year a MALDI-MS instrument. At first, it was generally thought that MAPS peptides could not be analyzed with these newer instruments. Several years ago, a brief discussion on the ABRF electronic mailing list elicited some comments from a few groups who were beginning to obtain ESI-MS data on their MAPS peptides. Since then, our laboratory has been more aggressive in trying to solve the synthetic and analytical issues surrounding MAPS peptides. In the discussion below, we first present a detailed analysis of the products obtained from a reasonable MAPS synthesis, followed by similar analytical data from an excellent MAPS synthesis and by data from a poor synthesis that we were unable to analyze by HPLC or mass spectrometry.
Figure 2 illustrates the analytical data for a representative four-chain MAPS peptide with the sequence SLERE GKLPY SWR. The ESI-MS analysis (Figure 2, upper panel) shows the desired species, plus molecules with one or two tBu or Pmc groups, TFA and EDT adducts, and combinations of these. The TFA and EDT adducts disappear with lyophilization. The MALDI-MS spectrum (Figure 2, middle panel) is simpler, showing predominantly the desired product and the +tBu derivative, with smaller amounts of the +2tBu and +Pmc species. Only a few species at lower m/z were detected. Therefore, this synthesis shows no significant problems in assembly but some problems with incomplete deprotection. It is also of interest that the MALDI-MS profile more closely resembles the HPLC profile (Figure 2, lower panel) than does the ESI-MS data. It is important to remember that MALDI-MS analysis also can overestimate contaminants if the amount of sample analyzed is too high (5); several dilutions of each sample should always be analyzed by MALDI-MS.
Separation by reversed-phase HPLC usually gives a better representation of the quantity of each component in a mixture. In the HPLC profile (Figure 2, lower panel), there are several peaks eluted with sufficient resolution to permit their isolation. Only a few truncated products were produced in this synthesis (peak 1 in the chromatogram). When these are major components, it can be difficult to identify the exact nature of each byproduct, because within each chain of the MAPS molecule a distinct problem may exist with assembly or deprotection. Peak 2, the major peak, was identified as the correct MAPS peptide and represents about 45% of the total product. Peaks 3, 4, and 5 contain molecular species with one additional tBu group, and peaks 4 and 6 contain species with two additional tBu groups, as identified by ESI-MS. Peptides with a single Pmc group are present in peaks 7, 8, and 9. The fact that the HPLC peaks are well resolved suggests that adducts with the same mass but different elution times represent distinct positional isomers; that is, in each of the HPLC peaks the same protecting group may be on different residues (Figure 1C-D). This peptide shows one of the inherent problems in providing very pure material for MAPS syntheses. Peaks with a single tBu group are peptides that contain three correct chains of the four. Peaks with two tBu groups have either two or three correct chains, depending upon whether the tBu groups are on the same or different chains, which cannot be easily determined. Thus, if one considers the peptide motifs as individual species instead of tetrameric assemblies of peptides, the synthesis and deprotection of this synthesis was greater than 90% successful. In this example, the desired molecular complex is easily purified. In many cases, the contaminating byproducts are eluted as a series of small, unresolved shoulders on a main peak.
Figure 2: Analysis of a 4-MAPS peptide (SLERE GKLPY SWR). The expected mass is 6,885. Upper panel: ESI-MS analysis. A 5 µl aliquot of crude peptide was analyzed in positive-ion mode by infusion at a rate of 1.5 µl/min. Reconstructed data are shown. Middle panel: MALDI-MS analysis. The sample was analyzed in linear, positive-ion mode after dilution in a-hydroxy 4-cinnamic acid. Lower panel: HPLC elution profile. The crude peptide was separated on an HP1090M HPLC with a Vydac 218TP54 column, at a flow rate of 0.75 ml/min, eluted with 0.1% TFA and increasing acetonitrile concentration at 1%/min. The absorbance at 214 nm is shown. All fractions were analyzed by ESI-MS and MALDI-MS: (1) truncation peptides; (2) desired product; (3) +tBu; (4) +tBu and +2tBu species; (5) +tBu and +TFA; (6) +2tBu and +2tBu+TFA; (7) +Pmc; (8) +Pmc; (9) +Pmc and +Pmc+tBu. The TFA and EDT adducts are not present after lyophilization. The added masses for each group are (reference 6): +56 (tBu); +93 (ethanedithiol, EDT); +96 (TFA); +242 (Trt); +266 (Pmc).
The analytical data for another four-branch MAPS peptide (VLFLL PLRLG HNLWR T), which showed few problems with either assembly or deprotection, are shown in Figure 3. The expected mass of 8,243 was observed. The ESI-MS, MALDI-MS, and HPLC data all indicate the presence of small amounts of byproducts containing (in order of abundance) either one Pmc group, one Trt group, or two tBu groups. Once again, the ESI-MS selectively ionizes certain species, overemphasizing the abundance of the contaminating byproducts when compared to HPLC.
Figure 4 shows our attempts to analyze the 4-branch MAPS peptide DNEYG YSNRV VDLMA YMA, which has an expected mass of 8,841. The raw ESI-MS data (Figure 4, upper panel) could not be reconstructed. The MALDI-MS analysis, shown in Figure 4, middle panel, reveals no significant signal. The HPLC profile from a C18 column showed a number of sharp, well-resolved peaks, all of low mass. Because it was possible that this peptide might be too hydrophobic to be readily eluted from a C18 resin, a C4 column was then used (Figure 4, lower panel). Several small, sharp peaks were followed by two broad, asymmetric peptide-containing peaks. Attempts to analyze these latter pools by either ESI-MS or MALDI-MS were also unsuccessful. Because the amino acid analysis of this peptide was within acceptable limits, several reasons for this HPLC elution pattern might be considered. The peptide might fold or aggregate in such a way as to hinder analysis by HPLC, ESI-MS, and MALDI-MS. Alternatively, the peptide could be polydisperse, with multiple protecting groups located at various positions, enhancing the peptide's hydrophobicity. This MAPS peptide could be correct or incorrect. At this time, methods are not available to make the distinction.
Figure 3: Analysis of a 4-MAPS peptide (VLFLL PLRLG HNLWR T). The expected mass is 8,243. Upper panel: ESI-MS analysis. A 10 µl aliquot of crude peptide was analyzed in positive-ion mode by infusion at 30 µl/min in 30% acetonitrile/0.1% TFA. The reconstructed spectrum is shown. Middle panel: MALDI-MS analysis. Conditions were the same as for Figure 2, middle panel. The species at 4,122 is the doubly protonated peptide. Lower panel: HPLC elution profile. The peak with the desired mass is indicated by *. Absorbance was monitored at 214 and 280 nm.
Our laboratory also provides traditional, non-MAPS peptide antigens of a single molecular species that can be used with no purification on the part of the investigator. This is because the investigator rarely has any interest or expertise in purification procedures. On a pragmatic basis, we release MAPS peptides if the percentage of correct peptide chains is very high (80-90%). However, we do not guarantee the presence of a single molecular species, and we do not purify MAPS peptides for the requesting laboratory. Purification of MAPS antigens is expensive and not always as feasible as shown in the examples in Figures 2 and 3. Although it has been claimed that gel filtration and ion-exchange chromatography can be used effectively to purify MAPS peptides, these do not give high-resolution separations and have not usually been used in conjunction with analysis by mass spectrometry. Our solution to this dilemma has been to focus on optimization of the synthesis and deprotection, so that our crude, lyophilized product is as homogeneous as possible. This is complicated by the fact that the density of the coupling sites on the MAPS scaffold is comparable to that of higher substitution resins. The density of the protecting groups on the 8-MAPS peptide resins may be the major obstacle in obtaining a product that can be correctly analyzed. As in the examples above, there seems to be less problem with peptide assembly than with deprotection, more so if the large protecting groups are located toward the carboxyl-terminus of the sequence. The density of protecting groups may be sufficiently high to promote their migration and reattachment during the cleavage process. MAPS peptides containing protecting groups could be more potent stimulators of the immune response, thereby eliciting greater production of antibodies than the unprotected MAPS, which might also be more biodegradable. We now synthesize most MAPS peptides as four-chain species with antigenic motifs no longer than 15 residues. Eight-chain MAPS are discouraged and will only be prepared if the peptide sequence is hydrophilic and short (no more than 14 residues). Both 4-branch and 8-branch MAPS peptides made according to this strategy have generally been more amenable to analysis and verification of structure and are still large enough to be dialyzed.
For most laboratories, traditional peptide antigens provide the most flexible experimental material. A pure peptide can be conjugated either to a carrier protein for antibody production or to a resin matrix for later antibody purification. However, MAPS antigens are easier to prepare for immunization and do have a defined number of peptides per molecule. In practice, our laboratory has found that coupling cysteine-containing peptides to commercially available maleimide-activated carrier proteins or amine-containing peptides to carrier proteins by glutaraldehyde crosslinking always produces a useful antigen. However, quantitating the coupling of linear peptides to carrier proteins can require techniques such as amino acid analysis or spectro-photometric assays, which can be daunting to the user laboratory.
MAPS antigens can provide a powerful alternative approach when classical procedures fail to produce a useful antiserum. For example, one investigator in our research community failed to produce an antibody with a traditionally conjugated peptide antigen but was able to produce a highly specific antibody with a MAPS antigen. This antibody worked well not only in immunoblots but in fluorescence confocal microscopy. However, it should be noted that even using conventional linear peptides conjugated to keyhole limpet hemocyanin, the rabbit may be the most variable component in antiserum production. One investigator associated with our laboratory has made antisera to several regions of an organic anion transport protein, injecting each antigen conjugate into several rabbits. In most cases, only one rabbit of a set has produced useful antibodies for both immunoblots and more demanding applications such as fluorescence confocal microscopy.
Figure 4: Analysis of a 4-MAPS peptide (DNEYG YSNRV VDLMA YMA). The expected mass is 8,841. Upper panel: ESI-MS analysis. The sample was analyzed in positive-ion mode by infusion at 2 µl/min, after diluting in an equal volume of 50% methanol/0.5% ammonium hydroxide (final pH about 8.5). Raw, unreconstructed data are shown. Other solvents and conditions were tried, but none gave better data. Middle panel: MALDI-MS analysis. The analysis conditions were the same as for Figure 2, middle panel, but a higher laser power setting was used; this spectrum is typical of the ones obtained using other laser power settings or other matrix conditions. Extending the mass range to more than 20,000 m/z did not reveal additional signals. Lower panel: HPLC elution profile. Absorbance from a C4 column was monitored at 214 nm and 280 nm.
MAPS antigens do provide an excellent alternative approach for making antibodies. Now that analytical tools have raised our capability to detect and define synthetic products, standards of acceptability have been raised. Our laboratory takes a conservative approach, even though most MAPS antigens are probably usable. In essence, we "synthesize for the mass spectrometer." It could be claimed by some that mass spectrometry simply is not suited for analysis of certain samples. However, our laboratory learned a critical lesson when working to improve our techniques for producing very long peptides (49, 65, and 76 residues) for structural studies. Once we had solved certain strategic synthetic issues for these peptides, the mass spectra obtained were of very high quality. Therefore, in our view, inability to obtain a mass spectrum of a synthetic peptide after having tried a variety of methods for sample preparation strongly points toward a problem with the synthetic product. Better strategies are needed to achieve the goal of producing a single molecular species from a MAPS synthesis, in particular, optimization of deprotection procedures.
Acknowledgments
The authors thank Silvie Bourassa for stimulating this discussion on the ABRF electronic mailing list and Lynda Bonewald for critical comments.
References
1. D.N. Posnett, H. McGrath and J.P. Tam (1988) J. Biol. Chem. 263: 1719-1725.
2. J.P. Tam (1989) Methods in Enzymology 168: 7-15.
3. G.B. Fields, R.H. Angeletti, S.A. Carr, A.J. Smith, J.T. Stults, L.C. Williams and J.D. Young (1994) Techniques in Protein Chemistry V (ed. J.W. Crabb): 501-507.
4. G.B. Fields, R.H. Angeletti, L.F. Bonewald, W.T. Moore, A.J. Smith, J.T. Stults and L.C. Williams (1995) Techniques in Protein Chemistry VI (ed. J.W. Crabb): 539-546.
5. R. Hogue Angeletti, L. Bibbs, L.F. Bonewald, G. Fields, J. Kelly, J.S. McMurray, W.T. Moore, and S. Weintraub. Analysis of Racemization During "Standard" Solid Phase Peptide Synthesis: A Multicenter Study. Techniques in Protein Chemistry VII, in press.
6. URL address: http://www.medstv.unimelb.edu.au/WWWDOCS/SVIMRDocs/MassSpec/deltamassV2.html
Ruth Angeletti may be contacted at Albert Einstein College of
Medicine, Laboratory for Macromolecular Analysis, 1300 Morris Park
Avenue, Bronx, NY 10461, Tel: (718) 430-3475, Fax: (718) 430-8939,
E-mail:
angelett@aecom.yu.edu.