Microscale Esterification of Peptides and Analysis by MALDI-MS

W. H. Fischer and A. G. Craig

The Clayton Foundation Laboratories for Peptide Biology, The Salk Institute, La Jolla, CA, USA

C-terminal amidation is a common post-translational modification of bioactive peptides. During biosynthesis C-terminal amides are formed from glycine extended precursors by the action of the enzyme peptidylglycine a-amidating monooxygenase (1) . In many cases the biological activity of a hormone depends on the presence of the amidated C-terminus; the corresponding peptide with a C-terminal free acid being inactive. It is therefore essential when determining the primary structure of a novel peptide to establish whether or not the C-terminus is amidated. Time-of-flight instruments when coupled with matrix assisted laser desorption ionization (MALDI) result in extremely high sensitivity (sub pmol) which is a desirable asset for determining the primary structure of peptides. We have developed a procedure that allows the determination of amidation of a peptide by carrying out an esterification reaction. A number of procedures and conditions using hydrochloric acid and thionyl chloride reagents have been compared. We investigate the question of whether the C-terminus can be characterized using sub picomole amounts of either the free carboxyl or the amidated forms of peptides as test substrates and what limitations oxidation places on the microscale preparations proposed. Using the procedures established we have determined that the C-terminus of a novel gonadotropin releasing hormone from a cichlid (Haplochromis burtoni) is amidated (2) .
In order to use esterification followed by mass analysis for determination of C-terminal amidation the following prerequisites have to be met: (i) the sequence of the peptide or at least the presence and number of carboxyl side chain amino acids should be known, (ii) side reactions such as oxidation should be minimized, since they may complicate the interpretation of results, (iii) deamidation of either the C-terminus or side chain amides should not interfere or the esterification reaction should go to completion within the shortest possible time so as to exclude deamidation side reactions.
We chose two model peptides for this investigation for which both the amidated and free carboxyl forms were available: the gonadotropin-releasing hormone (GnRH) agonist antide (monoisotopic [M+H]+ = 1590.8 Da) and the amphibian peptide bombesin (1619.8 Da). These peptides do not contain any carboxyl side chains. Bombesin contains several glutamines and one asparagine as well as the oxidation sensitive residues tryptophan and methionine.

Esterification reactions were carried out with methanol, ethanol, hexanol and benzyl alcohol; ester formation was detected by the appropriate mass shift. Of these alcohols, only methanol and ethanol were found to lead to a conversion of >80% in less than 1 hour. Reactions carried out for longer periods with all alcohols led to significant formation of side products and peptide degradation. For peptides containing tryptophan or methionine, oxidized species were observed in the mass spectra. This represents a particular problem in the interpretation of spectra of methyl esters as the mono- or di-oxidized forms of the peptide will contribute to a shoulder on the peak corresponding to the methyl or ethyl esters. To overcome this complication b-mercaptoethanol (2-10 mM) was included in the reaction mixture. Initial experiments were carried out comparing the alcohol/HCl procedure to alcohol/thionylchloride. We found significant side reactions occured with 1 M thionyl chloride which were significantly reduced when the concentration was lowered. However, an additional problem observed was that reactions with alcohol/thionylchloride resulted in considerable ester formation from peptides that were C-terminally amidated. Therefore, all the results presented below use the acidic alcohol procedure.

Figure 1.

The esterification of bombesin-OH (pGlu-Gln-Arg-Leu-Gly-Asn-Gln-Trp-Ala-Val-Gly-His-Leu-Met-OH) using HCl with methanol or ethanol was followed at various time points over 1 hour (see Figure 1.). A conversion of >80% was observed when using methanol after 30 minutes, whereas the ethyl esterification required 60 minutes. A low level of ester fromation was observed for bombesin-NH2 which appears to result from deamidation followed by esterification of both the C-terminus and side chain amides (based on comparison with the non-acidic residue containing antide GnRH agonist). As this reaction was considerably slower than the conversion of the free carboxyl group, it did not interfere with the interpretation of results. No oxidation or other side reactions were observed for samples incubated in the presence of b-mercaptoethanol. In particular, asparagine and glutamine as well as C-terminal amide groups are stable under the reaction conditions employed. Based on these results the above criteria (ii) and (iii), are fulfilled with either the HCl/methanol protocol for short time periods (up to 15 minutes) or the HCl/ethanol procedure for 60 minutes.
The esterification protocol was applied to the characterization of a novel gonadotropin releasing hormone (GnRH) isolated from cichlid pituitaries (2) , since less than 1 pmol of material remained to determine the C-terminus of the peptide. Based on the sequence analysis, we were confident that the peptide contained no acidic residues (pGlu-His-Trp-Ser-Tyr-Gly-Leu-Ser-Pro-Gly). Esterification with ethanol was carried out in the presence of bombesin free acid as an internal control. The mass of native cichlid GnRH (m/z 1114) could be determined on the Bruker Reflex (Figure 2a) using considerably less than 1 pmol of the precious GnRH material, as shown by the comparison of 1 pmol of bombesin-OH (m/z 1621) (Figure 2 b). Using an equivalent amount of the native GnRH peptide mixed with 1 pmol of bombesin-OH we were able to determine that the peptide could not be esterified (Figure 2c). After 1 hr of reaction the bombesin free acid was completely converted to its ethyl ester (and observed to be shifted by 28.5 Da). It could thus be concluded that the cichlid GnRH contained no free carboxyl function and therefore that the C-terminus was amidated.

The HCl/ethanol esterification protocol was found to be a generally useful method for modifying free acid groups in small peptides. In particular, the procedure was found to be suitable for analysis of sub pmol amounts of peptides. This is particularly advantegous when analyzing native materials which are available in limited quantities. Methanol and ethanol were found to give the best results in the esterification reaction; reaction with the former requiring the least amount of time to go to completion, the latter affording a greater mass shift. We conclude that micro scale esterification will have considerable application not only in determining the C-terminus of peptides but also in distinguishing glutamic acid and lysine residues when sequencing peptides or when determining the presence of an acidic residue in a sequence.

Experimental Procedures
Synthetic peptide (1-2 pmol) or native peptide (approx. 0.1 pmol) in 0.1% aq. trifluoroacetic acid and acetonitrile was evaporated under vacuum in 0.5 ml micro centrifuge tubes (polypropylene). To the dried peptide 2µl of a 10 mM solution of b-mercaptoethanol was added followed by 10 µl of a 2M solution of HCl in the appropriate alcohol (this solution was generated in situ by addition of 32 µl acetylchloride to 200 µl of alcohol generating HCl and the corresponding acetyl ester (3) ; the reaction was allowed to proceed for at least 1 hour on ice before addition to the dried peptide). When carrying out the modified Fischer esterification reaction, 1 % thionyl chloride was added to methanol on ice, and an aliquot (10 µl) was added to the dried peptide. The esterification was carried out at room temperature for the specified time after which 10 µl of 10 mM b-mercaptoethanol was added and the solvents were removed in vacuo. An aliquot (2 µl) of matrix solution (saturated solution of a-cyano-4-hydroxycinnamic acid in water/acetonitrile (2:1) containing 0.1 % trifluoroacetic acid) was added to the micro centrifuge tube. An aliquot of this solution (1µl) was spotted onto a stainless steel target and analyzed after drying in a Bruker Reflex time-of-flight mass spectrometer in the linear mode at 30 kV accelerating voltage.
We have found that adding matrix directly to polypropylene tubes to extract the samples consistently gave a more intense molecule ion signal than other methods (including the addition of 0.1 % aq. trifluoroacetic acid and varying concentrations of acetonitrile, with subsequent application of the sample onto a thin layer of the matrix) of extracting sub pmol amounts of peptides.

1. A. S. N. Murthy, R. E. Mains and B. A. Eipper (1985). Purification and Characterization of Peptidylglycine a-Amidating Monooygenase from Bovine Neurointermediate Pituitary. J Biol Chem 261, 1815-1822

2. J. F. F. Powell, W. H. Fischer, M. Park, A. G. Craig, J. E. Rivier, S. A. White, R. C. Francis, R. D. Fernald, P. Licht, C. Warby and N. M. Sherwood (1995). Primary structure of solitary form of gonadotropin-releasing hormone (GnRH) in cichlid pituitary; three forms of GnRH in brain of cichlid and pumpkinseed fish. Reg. Pep. 57, 43-53

3. D. Hunt, J. R. Yates III, J. Shabanowitz, S. Winston and C. R. Hauer (1986). Protein sequencing by tandem mass spectrometry. Proc Natl Acad Sci USA 83, 6233-6237

Back to Home Page