Methods of peptide sequencing. Part I. Conversion of oligopeptides into cyclic dipeptides: a gas chromatographic–mass spectrometric study

摘要

1975Methods of Peptide Sequencing. Part 1. Conversion of Oligopeptidesinto Cyclic Dipeptides : a Gas Chromatographic-Mass SpectrometricStudyBy Robert A. W. Johnstone. and T. Jeffery PovaII, The Robert Robinson LaboratoriesThe University, LiverpoolThe chemical degradation of oligopeptides to cyclic dipeptides and identification of the latter by gas chromato-graphy-mass spectrometry has been investigated as a method of amino-acid sequencing. Advantages anddisadvantages of the method are described.L69 3BXTHE transformation of peptides into cyclic dipeptides(dioxopiperazines) by pyrolysis, followed by identific-ation of the latter by g.1.c. has been suggested as amethod of amino-acid sequencing.l This method yieldsall possible ' sequential ' cyclic dipeptides from anoligopeptide but it is not yet clear whether misinform-ation can be generated by application of the vigorouspyrolytic conditions to peptides larger than the actino-mycins.We have examined an alternative acid-catalysed conversion of oligopeptides into cyclic di-peptides based on earlier limited examples in whichpeptides were heated with 2-naphthol or acetic acid.4As examples the tetrapeptide Leu-Gly-Gly-Leu gaveonly the cyclic dipeptide cy~Zo(-Leu-Gly-),~ and the tri-peptide Phe-Gly-Gly gave cycZo(-Phe-Gly-) .* There areother reports of the formation of cyclic dipeptides fromtripeptides and of their formation in pyridine from2,4,5-trichlorophenyl esters of tetrapeptides61 A. B. Mauger, Chem. Comm., 1971, 39.8 N. Lichtenstein, J .Amer. Chem. SOC., 1938, 60, 560.J. D. Baty, R. A. W. Johnstone, and T. J. Povall, J.C.S.N. F. Albertson and F. C. McKay, J . Amer. Chem. SOC., 1953,75, 5323; F. C. McKay and N. F. Albertson, ibid., 1967,79, 4686.Chem. Comm., 1973, 392.As an illustration of the procedure envisaged in thiswork, a tetrapeptide, ABCD, would be convertedsequentially into two cyclic dipeptides, AB and CD.Removal of the N-terminal amino-acid by the Edmanmethod would enable this unit to be identified and givea tripeptide, BCD, which could be converted into onecyclic dipeptide, BC. This method was suggested inessence many years ago but was not pursued becauseidentification of the components of the product mixturewas impossible at the time on the small scale required.However, the identification of cyclic peptides by g.1.c.-mass spectrometry provides a sensitive means of investi-gating such mixtures on a small scale and was used inthis work.The use of g.1.c. alone to identify cyclicdipeptides is not attractive because of the large numberof possible cyclic dipeptides (210 from the 21 commonamino-acids) . Use of relative retention times requiressynthesis of all possible cyclic dipeptides, not to mentionPeptides:Chemistry and Biochemistry, ed. B. Weinstein, Marcel Dekker,New York, 1970, pp. 419-434; J. C. Sheehan and D. N. Mc-Gregor, J . Amer. Cheon. SOC., 1962, 84, 3000; G. Lucente and P.Frattesi, Tetrahedron Letters, 1972, 42, 4285.6 I<. Titlestad, Chem. Comm., 1971, 1627.6 J.Meienhofer, Y . Sanq, and R. P. Pate1 iJ.C.S. Perkin Idiastereoisomers, and still does not guarantee a positiveidentification. Use of mass spectrometry with g.1.c.removes these difficulties.Several methods of converting peptides into cyclicdipeptides were examined by using the tetrapeptideVal-Gly-Leu-Phe as test material. The peptide washeated in (i) dimethylformamide (DMF), (ii) DMFcontaining phenol, (iii) DMF containing acetic acid,(iv) acetic acid alone, and (v) ethyl phosphorodichloriditein pyridine.' Only glacial acetic acid alone or in DMFproved successful. Thus, acidic catalysis is necessaryfor the sequential production of cyclic dipeptides, perhapsexpected cyclic dipeptides were observed (Table 1).However, unequivocal identification of the cyclic di-peptides is only possible through separation of theproducts by g.1.c.followed by mass spectrometry of theseparated components, and so g.1.c.-mass spectrometryof the O-trimethylsilyl derivatives of the cyclic peptideswas carried out (Table 2). Thus, the pentapeptideH-Gly-Leu-Leu-Gly-Gly-OH gave only one cyclic di-peptide, cycZo(-Gly-Leu-) , on treatment with glacialacetic acid. Edman degradation with phenyl isothio-cyanate enabled identification of the terminal glycine asthe corresponding phenylthiohydantoin (identified by00SCHEME 1by the mechanism shown in Scheme 1. Although acidicconditions are necessary, more strongly acidic conditions(HC1) degrade a peptide into amino-acids and smallpeptides but prevent cyclic dipeptide formation, probablyby protonation of the terminal amino-group.Additionalmass spectrometry), and gave a tetrapeptide, Leu-Leu-Gly-Gly. With acetic acid, this tetrapeptide gave twocyclic dipeptides, cycZo(-Leu-Leu-) and cycZo(-Gly-Gly-) .These, and similar experiments with other peptides, werecarried out on about 20 pmol of material and yields ofTABLE 1Degradation of peptides to cyclic dipeptides aPeptide b Cyclic dipeptides detectedAla-Phe-Leu cycle( -Ala-Phe-)cyclo (-Trp-Leu-)cyclo (-Leu-Phe-)Trp-Met-Asp(0Me) -Phe cyclo (-Trp-Met-)cycle-Asp (OMe) -Phe-Gly-Phe-Gln-G1 y-G1 y cycZo (-Gly-Phe-)cyclo (-Gln-Gly-)Gly-Pro-Trp-Leu C ~ C Z O ( -Gly-PrO-)Trp-Gly-Leu-Phe C ~ C Z O (-Trp-Gly-)Identifying ions in mass spectrum C44, 91, 114, 120, 128, 176, 21870, 83, 98, 111.154130, 29930, 77, 103, 130, 24391, 113, 120, 141, 169, 204, 26077, 130, 31791, 114, 120, 126, 167, 186, 232, 245, 27691, 113, 20484, 126, 168, 185a Identified by mass spectrometry of the mixture without g.1.c. separation. b For each peptide, the same experiment was repeatedm/e values; those italicised are on the peptide remaining after removal of the N-terminal amino-acid by the Edman method.molecular ions.TABLE 2Degradation of peptides to cyclic dipeptides aPeptide Cyclic dipeptide detected Identifying ions in mass spectrumGl y -L ys-Gl y C ~ C ~ O (-Gly-Lys-) 174,267,468,473Ala-T yr-Leu-Phe cyclo (-Ala-Tyr-) 179, 271, 436, 450Ser-Gly-Leu-Phe cycle( -Ser-Gly-) 103, 147, 267, 346, 360Ala-Glu(0Me) -Leu-Gly cyclo -Ala-Glu (OMe)Lys-Val-Phe-GI y cyclo (-Phe-Gly) 91, 229, 267, 333, 348G1 y-Leu-Leu-G1 y-Gly cyclo (-Leu-Gly-) See above0 Identified by mass spectrometry of g.1.c.separated components (trimethylsilyl derivatives).cycZo(-Leu-Phe-)cycZo(-Leu-Phe-) See abovecyclo (-Leu-G1 y-)267, 285, 313, 347, 389, 404241, 257, 326, 327, 343, 358166, 184. 257, 258, 271, 299, 341As Table 1.evidence for the sequential formation of cyclic dipeptidesis provided by the pentapeptide Elu-Gly-Pro-Trp-Leu, which has no free terminal amino-group and did notform any cyclic dipeptides.Our study of the conversion of peptides into cyclicdipeptides followed by mass spectrometry of the productmixture was encouraging, since ions due only to thecyclic dipeptides were about 50.Since these werereadily identifiable on this scale, the method appearedsuitable for use on a much smaller scale. All except oneof the simple peptides shown in Tables 1 and 2 behavedsimilarly to the above pentapeptide and no difficultieswere encountered.by J. M. Turner, Ph.D. Thesis, Cambridge, 1965.7 Personal communication from G. W. Kenner; original w-or1975 1299The first indication that the method might not begeneral, at least under the conditions used here, wasprovided by the tetrapeptide Lys-Val-Phe-Gly, fromwhich only the cyclic dipeptide, cycZo(-Phe-Gly-) wasidentified although presumably cyclo (-Lys-Val-) musthave been formed initially (Table 2). However, in thecases of two hexapeptides, H-Thr-Ala-Ile-Gly-Val-Gly-OH and H-Ala-Phe-Ile-Gly-Leu-Val-OH, none of theexpected cyclic dipeptides could be detected by g.1.c.-mass spectrometry.Similarly, the €3-chain of insulinyielded no identifiable cyclic dipeptides with aceticacid. Thus, the yields of cyclic dipeptides varied fromca. 50 down to 0. The reaction of a peptide withacetic acid afforded, in addition to cyclic dipeptides,brown materials which gave no peaks on g.1.c.-massspectrometry and are probably products of extensivedegradation and/or polymerisation. Mass spectrometryof the crude products at high temperature indicated thepresence of considerable amounts of high molecularweight material.For reference purposes, a variety of cyclic dipeptideswas synthesised, mostly by heating the methyl ester ofthe corresponding dipeptide.For g.l.c., the cyclicdipeptides were made less polar by conversion intotrimethylsilyl derivatives,g thereby allowing the morecomplex ones containing amino-acid residues such asserine, tyrosine, and lysine to be chromatographed aswell as those containing simpler amino-acids. Thepolarity of the cyclic peptides was also reduced by perme-t h y l a t i ~ n , ~ but the mass spectra of these permethylatedcompounds were inferior to those of the trimethylsilylderivatives for identification purposes.The mass spectra of several cyclic dipeptides have beenreported ; linear dipeptides gave mass spectradue to a combination of the spectra of the linear andcyclic dipeptides,1° the latter being formed in themass spectrometer.The mass spectra of the cyclicpeptides investigated here generally gave abundantmolecular ions; fragment ions at M - 28, Af - 29, andAf - 43 were sometimes but not invariably present andwere often of low abundance, in contrast with earliersuggestions.l0. l1 Similarly, fragment ions of the typeH,N=CHR (R = amino-acid side-chain), were often butnot always present (Scheme 2). With the exception ofthose containing proline, the cyclic dipeptides showedabundant fragment ions due to loss of the side-chain ofone of the amino-acid residues with or without hydrogentransfer. The loss of the side-chain of leucine, valine,or serine occurred so readily that no molecular ion, oronly one of very low abundance, was present (Scheme 2).For cyclic dipeptides containing aromatic side-chains,the major ions were due to the side-chains themselves,e.g. m/e 81 and 82 for histidine.In cyclic dipeptidescontaining tryptophan, the mass spectra showed littleelse than an ion at m/e 130 together with a molecularion. Lysine-containing cyclic dipeptides gave an ionA. B. Mauger, J . Chromatog., 1968,87, 315.G. Marino, L. Valente, R. A. W. Johnstone, F. Moham-madi-Tabrizi, and L. Sodini, J.C.S. Chem. Comm., 1972, 357.a t m/e 84, characteristic of lysine, as well as ions due toloss of NH, ; for glutamine-containing cyclic dipeptides,by far the most abundant fragment ion was produced byloss of NH, from the molecular ion. The mass spectra ofpermethylated cyclic dipeptides showed little or noimprovement over those of unmethylated ones.The mass spectra of trimethylsilylated cyclic dipep-tides were much more satisfactory.Apart from ions atm/e 73 and 75 from the trimethylsilyl group itself, thederivatives gave abundant molecular and M - 15 ions.Trimethylsilylated cyclic dipeptides containing aromaticside-chains gave fragment ions resulting from loss of theside-chain; for phenylalanine the charge was retained onthe cyclic dipeptide ring but for tyrosine and tryptophanthe charge was retained by the side-chain (m/e 179 and202, respectively). Lysine-containing cyclic dipeptidesR'CH NH1 -+ HLKcHR 0 'n(n = 1 or 2)SCHEME 2gave a tristrimethylsilyl derivative under mild conditionsbut longer reaction times gave the tetrakistrimethylsilylderivative.Conclusion.-A general theoretical approach to ob-taining the sequence of amino-acids in an oligopeptide byconverting it into cyclic dipeptides has been shown towork practically in several instances.Analysis andidentification of cyclic dipeptides in mixtures by g.1.c.-mass spectrometry of their trimethylsilyl derivativeshas been demonstrated. The actual sequential conver-sion of oligopeptides into cyclic dipeptides was found togive variable yields of products under the conditions usedmaking the method at present unacceptable as a generalprocedure where only limited amounts of peptide areavailable.EXPERIMENTALThe following analytical conditions were satisfactory forcyclic dipeptides. T.1.c.was carried out on silica gel withCHC1,-MeOH (9 : 1) or CHC1,-MeOH-AcOH (70 : 10 : 5 ) .lo H. J. Svec and G. A. Junk, J . Amer. Chem. Soc., 1964, 86,2278.l1 K. Jankowski and L. Varfalvy, Bull. Acad. polon. Sci., Sbr.Sci. chim., 1971, 19, 651; 1972, 20, 423; R. Nagarajan, J. L.Occolowitz, N. Neuss, and S. M. Nash, Chem. Comm. 1969, 3591300 J.C.S. Perkin IG.1.c. was performed on 5 OV1 columns (2.1 m x 6mm)a t 190-320 "C with a nitrogen flow rate of 45 ml min-1.Mass spectra were obtained with an A.E.I. MS12 spectro-meter at 70 eV. G.1.c.-mass spectrometry was carried outon a column (2-1 m x 6 mm) of either 5 OV1 on GasChromQ or 3 OV 17 on GasChrom Q initially a t 166-170 "C andthen temperature-programmed at 4' min-l to 250 "C withthe carrier gas flowing a t 30 ml min-1.Synthesis of Cyclic Dipeptides.-In a typical experiment,the dipeptide H-Leu-Phe-OMe (20 mg) was heated inrefluxing dimethylformamide (3 ml) for 12 h.The solution(ninhydrin-negative) was evaporated in vacuo and theresidue sublimed a t 140-170 "C and 0.5 mmHg to give thecyclic dipeptide. The purity was checked by analysis,mass spectrometry, t.l.c., and in some instances, lH n.m.r.spectroscopy. Similarly, the following cyclic dipeptideswere prepared cycZo(-A-B-) ; A = Gly, B = Gly, Ala, Val,Leu, Pro, Lys, Ser, Phe, or Glu; A = Pro, B = Val, Pro, Leu,Ile, or Phe; A = Phe, B = Leu or Asp(0Me); A = Tyr,B = Ala; A = Trp, B = Leu or Met; A = Lys, B = Val;A = Leu, B = Leu; A = Ile, B = His.The cyclic dipeptide (< 1 mg) was Trimethylsilylation.left in a stoppered vial with NO-bistrimethylsilylacetamide(0.1-0.2 ml) a t 80 "C for 19 h. A sample of this solutionwas injected straight onto the g.1.c. column.This was carried out in the usualway with phenyl isothiocyanate to give a phenylthio-hydantoin (identified by mass spectrometry) and a residualpeptide which was then degraded to cyclic dipeptides.Degradation of Peptides to Cyclic Dipe$tides.-(a) Thepeptide (1.5 mg) was dissolved in dry pyridine (1 ml) andethyl phosphorodichloridite (0.01 ml) was added a t 0 "C.The solution was allowed to reach room temperature andwas then heated on a steam-bath for 48 h. The solution wasevaporated in vacuo to leave a brown gum in which nocyclic dipeptides were detected.(b) The peptide (4 mg) was heated a t 120 "C in dimethyl-formamide (6 ml) containing phenol (50 mg) for 5 h.Evaporation left a brown gum in which no cyclic dipeptideswere detected.(c) The peptide (10 mg) was heated in glacial acetic acid(4 ml) at 100-110 "C for 10 h. Evaporation in vacuoleft a brown residue containing cyclic dipeptides.5/100 Received, 16th January, 19761Edman degradation