14 Peptides By C. E. DEMPSEY Department of Biochemistry University of Oxford South Parks Road Oxford OX1 3QU 1 Introduction This review covers literature on peptides published since the last peptide review in Annual Reports for 1980’ with emphasis on recent work from 1983/4. Research on peptides is increasing faster than ever (as evidenced by the appearance of three new journals for peptides Neuropeptides Peptides and Regulatory Peptides) due in part to significant advances in the methodology of solid-phase peptide synthesis in recent years which is providing peptide biologists with an ever-increasing supply of biologically active peptides and their analogues. The impact of peptide synthesis is widespread and synthetic peptides of some complexity are used for example to generate antisera for the isolation of proteins expressed by newly sequenced genes to probe the structure and specificity of antibody molecules and as synthetic vaccines.Although many of these applications are beyond the scope of this review they illustrate the importance of solid-phase peptide synthesis in many areas of biological research. It is timely that the 1984 Nobel prize for chemistry was awarded to R. B. Merrifield for his pioneering work in developing (and continuing to improve) these methods. It is impossible to make more than a passing survey of the huge volume of literature on peptides published in 1981-1984. The reviewer has chosen to illustrate those papers he feels to be particularly significant or illustrative of trends and to follow up lines of research that were developing when the field was last reviewed.It is emphasized that a selection of this type strongly reflects the personal bias of the reviewer. As in the last review of this series’ an attempt has been made to cover all areas of recent research in the peptide field and the biological aspects have been strongly emphasized. 2 Synthesis The series ‘Peptides’ edited by Gross and Meienhofer continues to be an excellent source of methods in peptide synthesis and includes a recent comprehensive treat- ment of protecting groups.2 Only a few recent examples of new protecting groups and their applications are illustrated here. ’ C. E. Dempsey and C. A. Vernon Annu. Rep. hog. Chem. Sect. B 1980 77,323. * ‘The Peptides Analysis Synthesis Biology’ Vol.3 ed. E. Gross and J. Meienhofer Academic Press New York 1981. 353 C. E. Dempsey Protecting Groups and Solution Syntheses.-Although many peptides have been synthesized using standard conditions [based on t-butyloxycarbonyl (Boc) or benzyl- oxycarbonyl (Z) protection] some sequences remain troublesome particularly those which contain amino-acids (such as Cys Tyr Trp and Met) sensitive to acid deprotection treatments in combination with these protecting groups. The use of phosphinamides as a new class of amino protecting group has been de~cribed.~ Phosphinamides have similar acid lability to the t-butyl- or benzyl-urethane groups but are free of the carbenium ion-mediated side reactions that can occur with side chain functions of sensitive amino-acids during deprotection.The acid lability of the P-N bond of phosphinamides can be modulated by varying the substituents on phosphorus and the diphenylphosphinamide (Dpp) group was chosen to evaluate these compounds as amino protecting groups. N" -Diphenylphosphinyl amino-acids were prepared (with 75-8570 yield in crystalline form) via the benzyl or methyl esters (Scheme 1) and the use of the Dpp group was illustrated with the synthesis Cl-H,fh-CH(R)CO,Me A Dpp-NH-CH(R)-CO,Me 1ii Dpp-NH-CH( R)-C02H (Dpp-= diphenylphosphinyl) Reagents i Ph,PO(CI)-N-methylmorpholine 0 "C;ii OH-H+ Scheme 1 of the partially protected tetrapeptide amide (1) using 0.25 M hydrochloric acid (in trifluoroethanol :water (9 :1) and HC1-methanol for successive cleavages of the N-protecting group (which was monitored by 31P n.m.r.).This sequence is the C-terminal tetrapeptide amide of gastrin. The same tetrapeptide having a free carboxyl terminus (protected as the phenyl ester) was synthesized taking precautions to evaluate cleavage conditions which leave the C-terminal phenyl ester intact and could be converted into the protected tetrapeptide amide in 90% yield by treatment Trp-Met-Asp( 0Bu')-Phe-NH2 (1) with ammonia in dichloromethane. This method is suggested to be a convenient general route in the synthesis of peptide amides. N"-diphenylphosphinyl amino- acids preserve their chirality during condensation reactions and the Dpp group is selectively cleaved in combination with side-chain protection based on t-butyl and benzyl alcohols.Phosphinic chlorides (2) have been used for carboxyl activation via the mixed amino-acid carboxylic-phosphinic anhydride and both diphenylphos- phinic chloride (2a) and 1-oxo-1 -chlorophospholane (3) are useful the latter having the advantage that the cyclic phosphinic acid ammonolysis by-product is water soluble and readily removed from the product ~eptide.~ R. Ramage D. Hopton M. J. Parrott G. W. Kenner and G. A. Moore J. Chem. SOC.,Perkin Trans. 1 1984 1357. R. Ramage C. P. Ashton D. Hopton and M. J. Parrott. Tetrahedron Lett. 1984 25 4825. Peptides 355 A new protecting group for the histidine imidazole is the v-benzyloxymethyl group and the (commercially available) protected histidine (4) has been used in a solid-phase synthesis of the octapeptide angiotensin and the solution synthesis of His-His-His.’ Compound (4) is activated and coupled without significant racemiz- ation [whereas the 7-benzyloxymethyl analogue of (4) is optically labile] and the protecting group is cleaved by hydrogenolysis as other benzyl-based groups.BOC-NH-CH-COzH Many biologically important polypeptides are coupled to carbohydrate via side-chain 0-or N-glycosidic linkages at Ser or Thr residues. The function of peptide- linked carbohydrate is of considerable interest and this has stimulated attempts to synthesize 0-glycopeptides for biological studies. Because of the lability (to both acid and base) of the 0-glycosidic bond side-chain protecting groups removable under mild neutral conditions have been sought and the allyl group has been shown to be useful both as a carboxyl-6 and as an a-amino- (as allyloxycarbonyl; Aloc) protecting group.7 The glycopentapeptide (6) was prepared from (5) by ethyl-2- ethoxy-l,2-dihydroquinoline-1-carboxylate (EEDQ)-promoted coupling of dipep- tide allyl esters according to Scheme 2.The use of the allyl group is based on its Z-Ser-Ala- Ala-Gly- Ala-OAll ZNHCHCO,All ZNHCHC0,H ii iii iv OBzl OBzl (6) (Bzl = benzyl; All = Allyl) Reagents i (Ph3PI4Pd morpholine THF (tetrahydrofuran) ; ii HCI- H-Ala-Ala-OAl1 NEt3 EEDQ CH,Cl,; iii deprotection (as i) ; iv HC1.H-Gly-Ala-OAlI NEt, EEDQ DMF (dimethyl- formamide) Scheme 2 ’ T. Brown J. H. Jones and J.D. Richards J. Chem. Soc. Perkin Trans. I 1982 1553. H. Kunz and H. Waldemann Angew. Chem. Inz. Ed. Engl. 1984 23 71. ’H. Kunz and C. Unverzagt Angew. Chem. Inf. Ed. EnrI.. 1984. 23,436. -!-+ Ser( But)-OBu' +Aloc-Ala-OH 356 C. E. Dempsey mild quantitative and selective cleavage by 10 mol% of tetrakis(tripheny1phos- phane)palladium(o) with morpholine as the allyl acceptor. The same catalytic transfer method (using 5,5-dimethylcyclohexane-1,3-dione as the allyl acceptor) cleanly removes the Aloc group from the a-amino-group and the protected tetrapep- tide ester (9) was prepared in high yield (95%) according to Scheme 3. The high yield in the dipeptide condensation was attributed to the small size of the Aloc group. Aloc-Ala-Ser( But)-OBu' A H-Ala-Ser( Bu')-OBu' (7) Aloc-Met-OH + Phe-OBut -!+ Aloc-Met-Phe-OBu' -Aloc-Met-Phe-OH (8) (7) + (8) -% Aloc-Met-Phe-Ala-Ser( But)-OBu' (9) (Aloc = allyloxycarbonyl) Reagents i EEDQ CH,CI,; ii (Ph3P),Pd 5,5-dimethylcyclohexane-1,3-dione, THF; iii trifluoroacetic acid (TFA); iv EEDQ THF Scheme 3 The formyl group is seldom used for amino protection in peptide synthesis but has proved useful in the selective modification of amino groups in native peptides.& N*-amino groups of the bee venom peptides apamin (19) and peptide-401 (sequence not shown) were formylated with high selectivity using the mixed formic-acetic anhydride in formic acid.p-Nitrophenyl formate at pH 9.5 was usefully selective for the &-amino group. The N-formyl group was cleaved from apamin in MeOH-HCl (0.1 M) at room temperature with acceptably low levels of methanolysis of side chain or a-amides.Selective reductive methylation (using formaldehyde and sodium cyanoborohydride at pH 6.8) of the a-or &-amino group of apamin was therefore possible by using the formyl group as a selective reversible amino protecting group. Using this strategy it could be shown that a decrease in biological activity on reductive methylation of apamin was due to dimethylation of the a-,and not the &-amino group. PHI (Peptide having N-terminal Histidine and C-terminal Isoleucine amide) was isolated from porcine intestine using a chemical assay for C-terminally amidated peptides.' The sequence" of PHI (10) shows a marked homology to a number of gut peptides of the glucagon family including secretin and VIP and also to a recently isolated growth hormone releasing factor (13).A surprising difference is at residue-24 10 His-Ala-Asp-Gly-Val-Phe-Thr-Ser-Asp-Phe-Ser-Arg-Leu-Leu-Gly-Gln-Leu- 20 27 -Ser-Ala-Lys-Lys-Tyr-Leu-Glu-Ser- Leu-Ile-NH (10) C. E. Dempsey J. Chem. SOC.,Perkin Trans. 1 1982 2625. K. Tatemoto and V. Mutt Nature (London) 1980 285 417. 10 K. Tatemoto and V. Mutt Proc. Natl. Acad. Sci. USA 1981 78 6603. Peptides 357 where PHI has glutamic acid; all other members of this family of peptides have a neutral amino-acid at this position. It was considered that an error in sequence determination or desamidation during isolation of PHI might be responsible for this difference and to test these possibilities porcine PHI and its Gln-24 analogue were synthesized." Six fragments corresponding to residues 1-6 7-1 1 12-14 15-18,19-23 and 24-27 of (10) were prepared by a stepwise active ester procedure using N-hydroxysuccinamide esters and catalytic hydrogenolysis to remove the N"-Boc group.Side-chain protection groups were derived from t-butyl alcohol and 1-adamantanol and the fragments were coupled with dicyclohexylcarbodi-imidein the presence of 1,2-dinucleophiles to suppress racemization. In assembling the Gln-24 analogue the pentapeptide fragment 23-27 (and the fragment 19-22) was used because of the tendency for the C-terminal tetrapeptide to cyclize to the pyroglutamyl peptide. The synthetic peptides were deprotected with trifluoroacetic acid (TFA) and purified to apparent homogeneity as judged by amino-acid analysis and high performance liquid chromatography (h.p.1.c.).The synthetic Glu-24 PHI was indistinguishable from native PHI by h.p.1.c. whereas the Gln-24 analogue could be resolved from PHI. This confirms the original assignment of glutamic acid to position-24 of PHI. Native PHI and the Gln-24 analogue were equally effective in a number of biological assays leaving unresolved the possibility that Gln-24 of PHI arose as an artifact of the purification procedure. Later sequence analysis of cloned DNA coding for the biosynthetic precursor to the human equivalent of PHI (PHM; having a C-terminal methionine amide) showed glutamic acid at residue-24,12 and it seems likely that the original sequence of PHI shown to be correct from the synthetic study is also the native structure of PHI in uiuo.A limiting factor in solution synthesis of peptides is often the solubility of protected fragments and this has been a particular problem with C-terminal fragments of VIP-related peptides. In the PHI synthesis solubility problems were overcome by interchanging side-chain protecting groups derived from t-butyl alcohol and 1-adamantol the latter group improving the solubility of hydrophilic peptides in organic solvent^.'^ In a synthesis of C-terminal fragments of VIP the temporary replacement of C-terminal asparagine with aspartic acid (protected as the t-butyl ester) greatly improved the solubility of the fragment^.'^ After removing side-chain protecting groups the C-terminal pentadecapeptide fragment of VIP was amidated via the mixed anhydride (with succinate) to generate the asparagine residue.An attempt has been made to predict the solubility of protected peptide fragments in solvents used for fragment coupling based on the sequence-dependent tendency of peptides to aggregate by forming intramolecular hydrogen-bonded @-sheet structure^.'^ It was suggested that the solubility of poorly soluble peptides can be im- proved by protecting the amide bond at suitable intervals in the sequence thus precluding intermolecular association by hydrogen-bond formation. Some success in these predictions and in the improvement of fragment solubility (using the I' L.Moroder W. Gohring P. Lucietto J. Musiol R. Scharf P. Thamm G. Bovermann E. Wunsch J. Lundberg K. Tatemoto and V. Mutt Hoppe-Seyler's 2. PhysioL Chem. 1983 364,1563. 12 N. Itoh K. Obata N. Yanaihara and H. Okamoto Nature (London) 1983 304,547. l3 R. Geiger and W. Konig in ref 2 p. 34. 14 W. M. M. Schaaper and D. Voskamp Red. Trau. Chim Pays-Bas 1984 103 17. l5 M. Narita K. Ishikawa J.-Y. Chen and Y. Kim fnt. J. Pepride Protein Res. 1984 24,580. 358 C. E. Dempsey 2,4-dimethyloxybenzyl or 2,4,6-trimethyloxybenzylprotecting group16) was achieved with a number of poorly soluble peptides. Many large peptides have been synthesized by fragment coupling indicating that with suitable strategies solubility problems can generally be overcome Recent examples include syntheses of human17 and mouse18 epidermal growth factor (homologous 53-residue peptides with three disulphide bonds) the first using the maximal protection strategy with a final deprotection in liquid hydrogen fluoride (HF) and the second using a milder protection strategy based on protecting groups labile to TFA or trifluoromethylsulphonic acid.Solid-phase Synthesis.-From a survey of the recent literature on the preparation of medium to large peptides for biological studies solid-phase synthesis -in many cases using commercial synthesizers -appears to be the preferred method. For many large peptides a successful strategy is the solid-phase synthesis of fragments which can be purified before coupling in solution or after re-immobilization on solid supports.Recent improvements have greatly extended the versatility of the solid- phase method and it has been stated” that ‘almost any desired sequence can be synthesized routinely’. Limitations in solid-phase synthesis remain however and these include the possibility of incomplete coupling during chain extension and side-reactions during the acid treatments required to remove protecting groups and cleave the peptide from the support. High resolution purification methods are therefore essential in the isolation of the synthetic target. In the development of solid-phase synthesis as a routine method the premium is on those with expertise in peptide chemistry to provide improvements in these areas and much progress has been made.Sheppard and his colleagues have published extensive details of the methods used in their laboratories.20 Polar polymethylacrylamide resins are used as supports and the authors believe these permit optimal solvation of the growing peptide and reagents and minimize aggregation within the matrix. The vigorous acid treatments required in the conventional Merrifield method are replaced by a single mild acid-cleavage step (in TFA) by using the base-labile p-fluorenylmethoxycarbonyl (Fmoc) protecting group for the a-amino function and acid-labile t-butyl or p-alkoxybenzyl groups for side chain and carboxy-terminal groups. These methods were illustrated by comparative syntheses of human P-endorphin (11) using a 10 Tyr-Gly-Gly-Phe-Met-Thr-Ser-Glu-Lys-Ser-Gln-Thr-Pro-Leu-Val-Thr-Leu- 20 30 -Phe-Lys-Asn-Ala-Ile-Ile-Lys-Asn-Ala-Tyr-Lys-Lys-Gly-Glu (11) ‘’ R.Geiger and W. Konig in ref. 2 p. 55. 17 D. Hagiwara M. Neya Y. Miyazaki K. Hemmi and M. Hashimoto J. Chem. SOC.,Chem. Commun. 1984 1676. H. Yajima K. Akaji N. Fujii K. Hayashi K. Mizuta M. Aono and M. Moriga J. Chem. SOC.,Chem. Commun. 1984 1103. 19 P. A. Kiberstis A. Pessi E. Atherton R. Jackson T. Hunter and D. Zimmern FEBS Lett. 1983 164 355. ** (a) E. Atherton M. Caviezel H. Fox D. Harkiss H. Over and R. C. Sheppard J. Chem. SOC.,Perkin Trans. 1 1983 65; (b) E. Brown R. C. Sheppard and B. J. Williams ibid. 1984,75; (c) E. Brown R. C. Sheppard and B. J. Williams ibid. 1983 1161. Peptides 359 polydimethylacrylamide resin first with conventional protection (N"-Boc and ben- zyl side-chain protection with HF deprotection) and then with the mild protection strategy based on the Fmoc group.The yield of purified homogeneous /3-endorphin was 41% using the mild protection strategy and 9% using Boc-based protection. The advantages of mild deprotection are clearly seen in the chromatography (on carboxymethylcellulose) of the crude products obtained on deprotection and cleavage from the resin. A large number of by-products were resolved in the run of the HF-cleaved product whereas the crude product obtained by mild acid cleavage ran as a major peak with only minor by-products. Despite the demonstrated benefits of a mild deprotection strategy the majority of recent syntheses have employed conventional protection and cleavage procedures.This is due in part to the development of routine procedures using commercial synthesizers with the attendant ease and rapidity with which peptides of considerable length can be prepared and the use of h.p.1.c. to isolate the target peptide from complex product mixtures. It is commonplace for the isolation and characterization of novel peptides to be closely followed by published syntheses and an example is the family of natriuretic peptides isolated from rat atria. It has been known for some time that endogenous factors exist which promote the excretion of sodium from the kidney (natriuresis). Because of the potential link between the regulation of extracel- lular fluid volume (by variation in the concentration of plasma sodium) and essential hypertension the characterization of these factors is of considerable interest.A number of natriuretic peptides were isolated in 1983 and 1984; all are contained within the sequence of the largest of the isolated peptide (12). Three of large biologically active fragments of (12) were reported in 1984. Atlas et al. 10 Leu- Ala-GI y-Pro-Arg-Ser-Leu- Phe-Gly-Gly-Arg-Ile-Asp- Arg- Arg-Ser-Ser-Cys- 20 33 -Arg-Ile-Gly-Ala-Gln-Ser-Gly-Leu-Gly-Cys-Asn-Ser-Phe-Arg-Tyr synthesized a 24-residue peptide corresponding to residues 9-32 of (12) using a commercial synthesizer.2* Nu-Boc amino-acids were coupled using a water soluble carbodi-imide (the coupling solvent was not reported) and repetitive removal of the Boc group was done with TFA.Conventional side-group protection was used. After cleavage from the resin with HF the peptide was purified by gel filtration oxidized to the disulphide and directly compared with the native peptide of the same sequence in its ability to relax histamine-contracted aortic rings. The peptide was equally effective as the native peptide and 10-fold more active than the unoxidized synthetic peptide. Although the authors claim the first synthesis of a mammalian natriuretic and vasoactive factor no evidence was given to establish the homogeneity amino- acid composition or sequence of the synthetic peptide. In a later synthesis of two peptides corresponding to residues 10-30 and 9-33 of (12) by similar methods on a commercial synthesizer the crude HF-cleaved peptides required purification by gel-filtration ion-exchange and repeated h.p.1.c.before homogeneity was 21 S. A. Atlas H. D. Kleinert M. J. Camargo A. Januszewicz J. E. Sealey J. H. Laragh J. W. Schilling J. A. Lewicki L. K. Johnson and T. Maack Nature (London) 1984 309,717. 360 C.E. Dempsey achieved., The pure synthetic peptides (obtained in 15% yield) were sequenced by Edman degradation. The longer peptide was indistinguishable from the native peptide by h.p.1.c. and in several biological assays and was 70-fold more effective than the 21-residue peptide in natriuretic activity. A third synthesis23 of an atrial natriuretic factor [residues 8-33 of (12)] was done by solution condensation of fragments corresponding to sequences 8-15 16-21 22-25 and 26-33 of (12) which were synthesized on solid supports.The advantages of this approach are i small peptides of suitable sequence can be prepared with minimal side-group protection ;here fragment one required only thiol protection (as carboxamidomethyl ; Acm); ii alternative cleavage methods are often possible with small peptides and fragments one two and three were cleaved by transesterification or catalytic hydro- genation; and iii fragments may be purified before coupling in solution (to 97-99% homogeneity in the present case). Side-chain protecting groups [Asp( Bzl) and Arg(NO,)] were removed by hydrogenolysis and the fragments were condensed in sequence using a modified azide procedure. The Acm protecting groups were removed with simultaneous formation of the disulphide bond by treatment with iodine.The purified peptide was characterized by amino-acid and sequence analysis h.p.l.c. and fast atom bombardment mass spectrometry of fragments and was equipotent with a native peptide [residues 8-33 of (12)] in contracting smooth muscle and more potent in natriuretic activity. The rapid syntheses of atrial natriuretic peptides were no doubt facilitated by the small number of ‘difficult’ amino-acids in these peptides [sequence (12) contains no Lys Trp or Met]. Merrifield has continued to explore alternatives to the final cleavage with HF which is recognized to be the main source of side reactions in the standard solid-phase method. Side reactions arise from acylation or alkylation of the peptide by carbenium ions generated (by an SNl mechanism) after protonation of the protecting group carbonyl in HF.It was argued that these could be avoided if the acid deprotection was diverted through an SN2-type mechanism by-passing formation of reactive carbenium ions.24 This could be achieved by diluting the HF (and so reducing its acidity function) with a weak base (dimethyl sulphide; DMS) which remains unprotonated under the acid conditions required to protonate the protecting group and so is available for nucleophilic participation in the cleavage reaction. A low concentration of HF in DMS (HF DMS :p-cresol; 25 :65 :10 v/v) was suitable and could deprotect synthetic peptides containing most standard protecting groups in high yield with a low degree of side reactions.For those protecting groups stable to low HF concentrations (such as tosyl for Arg or 4-methylbenzyl for Cys) the low HF treatment must be followed by the standard high concentration of HF. The latter treatment is now relatively free of side reactions however because most of the protecting groups generating carbenium ions are removed in the low HF treatment. Rat transforming growth factor a fifty-residue peptide containing three disulphide bonds has been synthesized using standard 22 M. Sugiyama H. Fukumi R. T. Grammer K. S. Misono Y. Yabe Y. Monsawa and T. Inagami Biochem. Biophys. Res. Commun. 1984 123 338. 23 N. G. Seidah C. Lazure M. Chretien G. Thibault R. Garcia M. Cantin J. Genest R. F. Nutt S. F. Brady T.A. Lyle W. J. Paleveda C. D. Colton T. M. Ciccarone and D. F. Veber Proc. Natl. Acad. Sci. USA 1984,81 2640. 24 J. P. Tam W. F. Heath and R. B. Memfield J. Am. Chem. Soc. 1983 105 6442. Peptides 361 Merrifield methods in combination with the 'low-high' HF deprotection and ~leavage.~' The crude peptide was oxidized (using glutathione) to avoid polymeriz- ation during purification by gel-filtration and reverse-phase h.p.1.c. The peptide was characterized by amino-acid analysis (of both enzymatic and acid hydrolysates) and co-eluted with native rat transforming growth peptide on h.p.1.c. The yield of homogeneous peptide was a remarkable 31% based on the ini?ial loading of alanine to the resin. The same strategy has also been used in an improved synthesis of crystalline glucagon in 41 YO yield.26 Improvements have been made in the use of hydrogenolysis as an alternative to acidolytic deprotection and cleavage although the method is still limited by the presence of sulphur-containing amino-acids.Thymosin (28 amino-acids with no Cys or Met) has been synthesized on a benzhydrylamine resin without acidolytic treatments2' N"-Fmoc protecting groups were repetitively cleaved with 55YO pyridine in DMF and hydrogenolysis-labile side-chain protection (benzyl and Boc) was used. The peptide was cleaved and deprotected by repeated (3-times) catalytic transfer hydrogenation with cyclohexa- 1,4-diene in the presence of palladium black in 84% yield and the overall yield of purified peptide was 22%.In another example,28 an ACTH pentadecapeptide fragment analogue (having no Cys or Met) was synthe- sized using N"-Boc amino-acids and benzyl side-chain protection on a poly(styrene- 1'10 divinylbenzene) resin. The peptide was cleaved and deprotected by catalytic transfer hydrogenation using a palladium catalyst and ammonium formate as the in situ hydrogen source with a yield after work up of 94%. 3 Sequence Determination Instrumental improvements in the automated peptide sequencer based on the Edman degradation continue to be reported and it is now possible to carry out routine sequence analysis with less than 100 picomoles of pol~peptide.~~ A recent advance is the so-called 'gas-phase' microsequencerZ9 in which the peptide is non-covalently immobilized on a Polybrene" support and the Edman reagents are delivered as vapours in a stream of argon.This modification has several advantages over the conventional spinning-cup sequencer including the minimization of loss of peptide from the reaction chamber and the reduction in volume of reagents required to effect both the Edman coupling and cleavage reactions and extraction of the resulting anilinothiazolinones. The latter are converted into phenylthiohydantoin (PTH) amino-acids and analysed by h.p.1.c. with picomole sensitivity. The increased efficiency and sensitivity afforded by technical improvements of this type can be important as the amount of peptide available for sequencing is often very small. An example is the isolation and sequencing of human hypothalamic growth hormone releasing factor (GRF).The hypothalamus has long been known to contain factors 25 J. P.Tam H. Marquardt D. F. Rosberger T. W. Wong and G. J. Todaro Nature (London) 1984,309 376. 26 S. Mojsov and R. B. Merrifield Eur. J. Biochem. 1984 145 601. 27 R. Colombo J. Chem. Soc. Chem. Commun. 1981 1012. 28 M. K. Anwer A. F. Spatola C. D. Bossinger E. Flanigan R.C. Liu D. B. Olsen and D. Stevenson J. Org. Chem. 1983 48 3503. 29 M. W. Hunkapillar J. E. Strickler and K. J. Wilson Science 1984 226 304. * Polybrene is a quaternary ammonium salt which adheres strongly to both polypeptides and glass surfaces. C. E. Dempsey with stimulatory and inhibitory effects on growth hormone secretion but previously only the inhibitory factor (somatostatin) could be characterized.It was found however that certain pancreatic tumours secrete a factor with GRF activity and using an immobilized antibody raised against a synthetic fragment of the pancreatic factor 1.3 nmol of hypothalamic GRF could be ~btained.~' Two samples of 0.5 nmol were sequenced using the gas-phase sequencer and the first 43 residues were directly established. The C-terminal leucine was deduced from the quantitative amino-acid composition. It was not possible to determine by sequencing whether the C-terminal carboxyl was amidated or free and this was done by synthesizing the a-amidated 44 residue peptide and its free a-carboxyl analogue. Comparison with the native peptide in two h.p.1.c. systems established the C-terminal amide and indirectly confirmed the sequence (13).10 Tyr-Ala-Asp-Ala-Ile-Phe-Thr-Asn-Ser-Tyr-Arg-Lys-Va~-Leu-Gly-Gln-Leu-Ser- 20 30 -Ala-Arg-Lys-Leu-Leu-Gln-Asp-Ile-Met-Ser-Arg-Gln-Gln-Gly-Glu-Ser-Asn-Gln- 40 44 -Glu-Arg-Gl y-Ala-Arg-Ala-Arg-Leu-NHZ (13) The sensitivity in sequencing by direct Edman degradation is ultimately limited by that achievable in identifying the cleaved N-terminal amino-acid derivative. Dimethylaminoazobenzene isothiocyanate [DABITC ;(14)] has been introduced as a modified Edman reagent for manual ~equencing.~' DABTH*-amino-acids absorb strongly at visible wavelengths and can be analysed by t.1.c. or h.p.l.c. in the latter case with subpicomole ~ensitivity.~~ Fluorescein isothiocyanate is also used in manual sequencing and an improved h.p.1.c.separation of fluorescein thiohydantoin amino acids with an analytical sensitivity of 50 femtomoles has been reported.33 These modified Edman reagents are limited by low solubility and coupling efficiency and must be used with an additional PITCT coupling to ensure maximal cleavage at each cycle. A modified reagent with similar sequencing efficiency as PITC is 4-(Nu-Boc-aminomethyl) phenylisothiocyanate (15).34 Compound (15) contains a cryptic amino-group which is revealed by hydrolysis of the Boc protecting group during acid cleavage of the coupled N-terminal amino acid. This amino group can be 30 N. Ling F. Esch P. Bohlen P. Brazeau W. B. Wehrenberg and R. Guillemin Proc. Natl. Acad. Sci.USA 1984,81 4302. 31 J.-Y. Chang Methods Enzymol. 1983 91 455. 32 C.-Y. Yang and S. J. Wakil Anal. Biochem. 1984 137 54. 33 K. Muramoto H. Kamiya and H. Kawauchi Anal Biochem 1984 141 446. 34 J. J. L'Italien and S. B. H. Kent J. Chromafop-.,1984. 283 149. * DABTH = dimethylaminoazobenzene thiohydantoin. t PITC = phenyl isothiocyanate. Peptides 363 derivatized to yield a fluorescent (or otherwise modified) thiohydantoin to enhance detection sensitivity. It is still possible in practice to approach the sensitivity of automated sequencers by manual sequencing using the dansyl-Edman method. Porcine neuropeptide Y a 36 residue C-terminally amidated peptide from mammalian brain was sequenced using 17 nanomoles of ~eptide.~’ The required sensitivity was achieved by using h.p.1.c.to isolate tryptic peptides and a subtractive dansyl-Edman technique in which an aliquot of the acid hydrolysate obtained from the dansylated peptide after each cycle of Edman degradation was redansylated and analysed by t.1.c. This allows the amino-acid lost in each cycle to be identified and provides an independent check on the course of the degradation. The sequence of secapin (from bee venom) has been revised by dansyl-Edman ~equencing.~~ The new sequence (16) corrects two previously proposed sequences one of which was obtained by mass spectrometry. 10 Tyr-Ile-Ile-Asp-Val-Pro-Pro-Arg-Cys-Pro-Pro-Gly-Ser-Lys-Phe-Ile-Lys- 20 25 -Asn- Arg-C& Arg-Val-Ile-Val- Pro (16) Fast atom bombardment (FAB) mass spectrometry has been used to study pep- tides.37 Underivatized polar molecules can be ionized by bombarding with a stream of fast neutral atoms (typically argon or xenon) and pseudomolecular ions have been observed from polypeptides of molecular weights above 9000.Fragment ions are considerably less abundant than the molecular ion but can be used to obtain sequence information if sufficient peptide is used (2-50 nmol). The technique is most useful for peptide sequencing in combination with more traditional methods of peptide chemistry and this approach was used in sequencing human Calcitonin Gene-Related Peptide [CGRP; (17)].38The molecular weight of intact CGRP and 10 Ala-C$s- As p-Thr-Ala-Thr-C$s-Val-Thr-His-Arg- Leu- Ala-Gly-Leu-Leu-Ser- Arg- 20 30 -Ser-GI y-Gly-Val-Val-Lys- Asn- Asn- Phe-Val-Pro-Thr-Asn-Val-Gly-Ser-Lys-37 -Ala- Phe- NH of tryptic peptides (in an unfractionated digest) were obtained by FAB mass spectrometry.Several of the tryptic peptides were sequenced by a combination of mass spectrometry further proteolytic digestion and amino-acid analysis (which was necessary to distinguish leucine from isoleucine). The low-mass region of the FAB spectrum of intact CGRP gave C-terminal sequence ions which established 35 K. Tatemoto Proc. Natl. Acad. Sci. USA 1982 79 5485. 36 L. K. Liu and C. A. Vernon J. Chem. Res. (S) 1984 10. 37 (a) M. Barber R. S. Bordoli R. D. Sedgwick and A. N. Tyler J. Chem. Soc. Chem. Commun. 1981 325; (6) D. H. Williams C. Bradley G. Bojesen S.Santikarn and L. C. E. Taylor J. Am. Chem. SOC. 1981 103 5700. 38 H. R. Moms M. Panico T. Etienne J. Tippins S. I. Girgis and I. MacIntyre Nature (London),1984 308 746. 364 C. E. Dempsey the sequence 20-3 1. Other sequence information was obtained by N-terminal analysis (by the dansyl method) of purified tryptic fragments and mass spectrometry of the reduced and carboxymethylated peptide. It is noteworthy that the disulphide bond and the presence of a C-terminal amide could be unambiguously established from the fragmentation information. An alternative method for peptide sequencing39 makes use of the greatly increased detection limit for charged compounds relative to uncharged ones in FAB or secondary ion mass spectrometry (SIMS).If the N-terminal amino group of a peptide is selectively derivatized with a charged group non-specifically cleaved with acid and the resulting mixture esterified and acylated (to eliminate residual charged sites) then the mass spectrum of the resulting mixture will preferentially reveal ions containing the charged group. These ions originate from the N-terminal end of the peptide and the sequence can thus be determined from the mass differences between the ions. The method has so far only been applied to simple tripeptides. Mass spectrometry is uniquely suited to structural analysis of small amounts of peptide containing structural units other than the standard amino-acids. The struc- tures of urinary sleep-promoting factors have been solved by FAB mass spectrometry in combination with structural studies on muramyl peptides of known structure.40 These factors appear to be a family of related muramyl peptides and a representative structure is (18) (where Dap is diaminopimelic acid).The necessary information CH,OH I NHCOMe ,J NHCOM~ Me CH(CO)-Ala-y-D-Glu-Dap-D-Ala for structure determination was achieved by FAB molecular weight and fragmenta- tion analysis of components purified by h.p.l.c. then acetylation and methylation to determine (by mass spectrometry) amino and carboxyl groups respectively and enzymatic cleavage of the disaccharide followed by FAB analysis of the purified free sugar and peptide. Endogenous muramyl peptides probably arise from bacterial cell wall fragments that are absorbed in the gut and incorporated into neuromodu- lators in the brain.The isolation and characterization of these factors is a result of some fifteen years effort an interesting historical account of which is available.41 Some of the most difficult structures are those peptides from micro-organisms many of which are cyclic and may contain D-amino-acids novel blocking groups and new or highly conjugated amino-acids. The structural studies of D. H. Williams and his collaborators are notable in this area and have been reviewed?2 39 D. A. Kidwell M. M. Ross and R.J. Colton J. Am. Chem. Soc. 1984 106 2219. 40 S. A. Martin M. L. Karnovsky J. M. Krueger J. R. Pappenheimer and K. Biemann J. Biol Chem. 1984 259 12652. 41 J. R. Pappenheimer J. Physiol.(London) 1983 336 1. 42 D. H.Williams Chem. SOC.Rev. 1984 13. 131. Peptides 365 Peptide Sequences from DNA Sequencing.-The sequences of some endogenous peptides have been determined by DNA sequencing even before their existence in vivo has been realized. This peculiar situation arises from the fact that most if not all endogenous peptides are biosynthesized as part of large precursors which may contain the sequences of several peptides. It is sometimes possible to isolate messenger RNA coding for peptide precursors and this can be used as a template for in vitro biosynthesis of the complementary DNA (cDNA). The cDNA is isolated hybridized (to form the base-paired double helical dimer) and inserted into the genetic material of a bacterium (via a bacterial plasmid).With suitable manipulations (see for example ref 43) a bacterial clone containing the desired precursor cDNA is obtained from which the latter can be isolated and sequenced by chemical methods. Potential endogenous peptides are sequences bounded by pairs of basic amino-acids from which the peptide is liberated by proteolytic excision (although C-terminal proteolytic cleavage at single arginine residues is known). A C-terminally amidated peptide is predicted if the amino-acid preceding the C-terminal proteolytic excision point is glycine. The amino-acid sequence of an enkephalin precursor (bovine proenkephalin A) was obtained in this way4 and a novel C-terminally amidated peptide of 26 amino-acids was predicted (residues 104-129 of the precursor sequence).Extracts of bovine adrenal medulla were fractionated using antibodies raised against synthetic fragments of the predicted peptide and a peptide having the requisite immunoreactivity was isolated.45 The peptide (named amidorphin) was sequenced and had the structure predicted from the precursor sequence. This confirms that amidorphin a peptide predicted from analysis of a DNA sequence is a true endogenous peptide (the physiological function of which remains to be determined). It is of course not possible to determine from the DNA sequence whether post-translational modification of the polypeptide will occur; this can only be done by isolation and structural characterization of the expressed polypeptide. The sequence of human T-cell growth factor (also called interleukin 2) was predicted by cDNA sequencing but when the purified peptide was subjected to Edman degradation an unassignable PTH amino-acid was found at position-3.It was subsequently found by analysis of the N-terminal tryptic peptide using FAB mass spectrometry that the third amino-acid residue of the growth factor was serine with an 0-linked N-acetylgalactosamine residue.46 The function of polypeptide-linked carbohydrate in this case and in general remains obscure. 4 Physical Studies Many studies of the conformational properties of peptides in solution continue to be reported. The information available from high resolution n.m.r. has been enhan- ced by the application of two-dimensional (2-D) n.m.r. methods to peptide 43 J.E. Davies and H. G. Gassen Angew. Chem. Inf. Ed. EngL 1983 22 13. 44 (a) M. Noda Y. Furutani H. Takahashi M. Toyosato T. Hirose S. Inayarna S. Nakanishi and S. Nurna Nume (London) 1982 295 202; (b) U. Gubler P. Seeburg B. J. Hoffrnann L. P. Gage and S. Udenfriend ibid. 1982 295 206. 45 B. R. Seizinger D. C. Liebisch C. Grarnsch A. Herr E. Weber C. J. Evans F. S. Esch and P. Bohlen Nature (London) 1985 313 57. 46 R. J. Robb R. M. Kutny M. Panico H. Morris W. F. DeGrado and V. Chowdhry Biochem. Biophys. Res. Comrnun. 1983 116 1049. 366 C. E. Dempsey conformations and the use of versatile techniques for following the exchange of magnetization between nuclei or among chemical species (reviewed in re$ 47).The most useful of the latter techniques for peptide conformations are the nuclear Overhauser effect (n.O.e.) which allows the separation of nuclei to be estimated over short (24 A) distances (assuming a reasonable degree of conformational stability) and saturation transfer which can (in favourable cases) yield exchange rates between interconverting conformational states and more generally be used to measure the rates of exchange of peptide amide protons with water. The potential of high resolution 2-D n.m.r. for studying the conformation (and conformational fluctuations) even of small proteins is best illustrated in the work of Wuthrich and his collaborators on the bovine pancreatic trypsin inhibitor.48 A conformational study49 of apamin (19) using one dimensional n.m.r.was illustrated in Annual Reports for 19801 and the conformation in water has been re-investigated using 2-D n.m.r. methods5' An essentially complete assignment of the proton n.m.r. spectrum of apamin was achieved by 2-D correlated spectroscopy (which establishes a complete set of J-coupling relationships in the molecule) and 2-D n.0.e. spectroscopy. Based on revised resonance assignments and the distance constraints between protons from n.0.e. experiments a new model for the conforma- tion of apamin in water has been proposed. The most interesting feature of the model is the presence of a helix comprising residues 9-18 of sequence (19). This helix contains the residues (Arg-13 Arg-14) whose side-chain guanidino groups are important for the biological activities of apamin.The p-turn enclosing residues 2-5 proposed by Bystrov et is retained and hydrogen bonds between the amide proton of Thr-8 and the Ala-5 carbonyl and between the amide of Lys-4 and the side-chain carbonyl of Asn-2 are added. The new conformation is consistent with previously determined49 coupling constants and amide exchange rates. lo 18 I I Cys-Asn-Cys-Lys-Ala-Pro-Glu-Thr-Ala-Leu-Cys-~a-~g-Arg-Cys-Gln-Gln-His-NH~ I I It continues to be true that the most useful and unambiguous information on the conformations of peptides in water is obtained from large polypeptides or those containing multiple disulphide bonds i.e. peptides having a high degree of conforma- tional stability. Small linear peptides of which many of the known hormones and neuropeptides are examples are highly flexible in physiological solutions ; n.m.r.studies on ~omatostatin~~ for example have been concerned with and brad~kinin,~~ their conformational diversity. An extreme example of the difficulty in defining unambiguous conformational information for small peptides is provided by work 47 R. Benn and H. Giinther Angew. Chem. Int. Ed. EngL 1983 22 350. 48 (a) K. Wuthrich G. Wider G. Wagner and W. Braun J. Mol. Biol. 1982 155 311; (6)G. Wagner C. I. Stassinopoulou and K. Wuthrich Eur. J. Biochem. 1984 145 431. V. F. Bystrov V. V. Okhanov A. I. Miroshnikov and Yu. A. Ovchinnikov FEBS Lett. 1980 119 113. D. Wemmer and N. R. Kallenbach Biochemistry 1983,22 1901. M.Knappenberg A. Michel A. Scarso J. Brison J. Zanen K. Hallenga P. Deschrijver and G. Van 49 50 51 Binst Biochem Biophys. Acta 1982 700 229. 52 L. Denys A. A. Bothner-By G. H. Fisher and J. W. Ryan Biochemistry 1982 21,6531. Peptides 367 on leucine-enkephalin (20). N.m.r. and model building studies have indicated that a conformation with a P-turn involving a hydrogen bond between the amide of Phe-4 and the Tyr-1 carbonyl is a dominant low energy conformer although it has also been shown that the conformations of enkephalins are sensitive to changes in the properties of the solvent.53 The enkephalins have been studied by proton and 13C n.m.r. in the presence of phospholipid vesicles and the lipid environment is proposed to promote a folded conformation with a &turn between Leu (or Met) and Gly-2 and the side chains of residues 1,3 and 5 buried within the lipid matrix.54 Both leucine-enkephalin5’ and methionine-enkephalins6 have been crystallized (from water-dimethylformamide mixtures and from water respectively) and in each case an extended anti-parallel P-sheet conformation has been found.Moreover even in the crystal the conformations are heterogeneous; the unit cell of the leucine- enkephalin crystal contains four independent pentapeptide conformers and that of methionine-enkephalin contains two. These studies indicate the variety of conforma- tions accessible to the enkephalins but say little about the important receptor-bound conformations. More direct information has come from the use of conformationally constrained analogues.A semi-rigid analogue (21)(where A,bu is a,y-diaminobutyric acid) has been synthesized and compared with leucine-enkephalin for binding ability in a guinea pig ileum (GPI) assay (which is relatively specific for the psubclass of opiate receptors) and the mouse vas deferens (where opiate effects are mediated Tyr-Gl y-Gl y-Phe- Leu Tyr-cyclo[-NY-~-A,bu-Gly-Phe-Leu-] (20) (21) mainly by S-receptors).” The cyclic analogue (21) was 17-times more potent than leucine-enkephalin in the preceptor assay and seven times less potent as a &receptor agonist. A linear analogue of (21) was equipotent in both assays. Compound (21) was therefore assumed to have a conformation highly favourable for interaction with the opiate preceptor but less able to adopt the conformation required to bind optimally to the &receptor.These results are of interest because they indicate that p-and S-opiate receptor subtypes have different conformational requirements for ligand binding. Because the analogue (21) is semi-rigid it was possible to explore the preferred conformation for the preceptor and notably it could be shown that (21) was unable to adopt the P-turn conformation suggested by the n.m.r. and model building studies.58 The use of classical structure-activity studies to identify functional groups together with conformational analysis to direct the synthesis of conformationally constrained analogues having predicted agonist or antagonistic properties has been a fruitful approach to peptide drug design.With this strategy the minimal functional structure of somatostatin (14 amino-acids) has been reduced to a cyclic hexapeptide having 50-100-times the potency of the native peptide in the inhibition of release 53 L. Zetta and F. Cabassi Eur. J. Biochem. 1982 122 215. 54 B. A. Behnam and C. M. Deber J. Biol. Chem. 1984,259 14935. 55 A. Camerman D. Mastropaolo I. Karle J. Karle and N. Camerman Nature (London) 1983,306,447. 56 T. Ishida M. Kenmotsu Y. Mino M. Inoue T. Fujiwara K. Tornita T. Kirnura and S. Sakakibara Biochem. J. 1984 218 677. 57 P. W. Schiller and J. DiMaio Nature (London) 1982 297 74. 58 J. DiMaio and P. W. Schiller Roc. Nafl. Acad. Sci. USA 1980 77 7162. 368 C.E.Dempsey of insulin glucagon and growth hormone in rats.59 A bicyclic analogue [(22) where Mpa is P-mercaptopropionic acid and the side chains of Asp-5 and Lys-8 are linked by an amide bond] of vasopressin (30) (see p. 000) was predicted and found to be an antagonist of the antidiuretic activity of the native peptide and in addition was selective in that the antipressor activity of the latter peptide was only partially antagonized.60 Mpa-Tyr-Phe-Gln-cyclo[ -P-Asp-Cys-Pro-Lys-N'-] -Gly-NH (22) Much evidence has arisen for the notion that some linear peptide hormones can adopt ordered conformations at membrane surfaces and that these conformations are important for biological activity. A number of peptides have been shown by circular dichroism studies to adopt helical conformations in the presence of deter- gents or phospholipids,61 and in one case that of adrenocorticotropin (ACTH) a 39-residue peptide which induces the synthesis of steroid hormones in the adrenal cortex the biological activities of fragments correlate with their ability to penetrate phospholipid bilayers6* ACTH( 1-21) is a fully potent agonist of native ACTH and it is known from structure-activity studies that the N-terminal decapeptide is responsible for triggering the biological responses of ACTH while the C-terminal sequence modulates the activity with respect to different receptors.The conformation of ACTH(1-21) in lipid bilayers was studied by a combination of photoaffinity labelling using a hydrophobic photoactivable compound 3-trifluoromethyl-3-( m-['''1 ]iodophenyl)diazirine which upon photolysis labels regions of the peptide that enter the hydrophobic core of the membrane and infrared attenuated total reflection spectroscopy which can be used to estimate the orientation of ordered peptide segments in the membrane.63 It was suggested from these experiments that the N-terminal decapeptide enters the membrane as a perpendicular helix while the C-terminal region lies along the membrane surface stabilizing the association of peptide and membrane.Interestingly neither the N-terminal decapeptide nor the C-terminal fragment of ACTH( 1-21) was able in isolation to associate with membranes. Kaiser and his colleagues have proposed that the potential for some peptides to adopt so-called amphiphilic helices in which hydrophobic and hydrophilic amino- acids segregate on different faces of the helix (Figure 1) may be an important property but one which is relatively non-specific with respect to the amino-acid sequence.64 Their approach to this idea is illustrated for @-endorphin (11).p-Endorphin has the potential to form an amphiphilic n-helix (see Figure 1) or a-helix and is known to adopt a helical conformation on interaction with phospholipids. An analogue of @-endorphin was synthesized in which the C-terminal residues 59 D. F. Veber R. Saperstein R. F. Nutt R. M. Freidinger S. F. Brady P. Curley D. S. Perlow W. J. Paleveda C. D. Colton A. G. Zacchei D. J. Tocco D. R. Hoff R. L. Vandlen J. E. Gerich L. Hall L. Mandarino E.H. Cordes P. S. Anderson and R. Hirschmann LifeSci. 1984 34 1371. 60 G. Skala C. W. Smith C. J. Taylor and J. H. Ludens Science 1984 226 443. 61 C.-S. C. Wu A. Hachimori and J. T. Yang Biochemistry 1982 21 4556. B. Gysin and R. Schwyzer FEBS Lett. 1983 158 12. 63 H.-U. Gremlich U.-P. Fringeli and R. W. Schwyzer Biochemistry 1984 23 1808. 64 E. T. Kaiser and F. Z. Kezdy Science 1984 223 249. Pept ides 369 Tyrl -Gly-Gly-Phc-Met ---1tyl2 Figure 1 Amphiphilic distribution of amino-acids in the C-terminal sequence of P-endorphin [see (1l)] on formation of a .rr-helix (viewed down the helix axis). Hydrophobic amino-acids are shaded (residues 14-31) were replaced by a non-homologous sequence but which was predicted to generate a similar amphiphilic structure on helix formation.Only leucine glutamine and lysine were used for this region and the N-terminal sequence (residues 1-13) was left unchanged. The analogue was more potent than P-endor- phin in the GPI assay (which is of little significance because the preceptor has little sensitivity to C-terminal structure of enkephalin-containing sequences) and was about one-fifth as active in the rat vas deferens assay in which the opiate effects are mediated by a third class of opiate-receptors the &-receptor. The latter activity is remarkable because the &-opiate receptor is highly specific for P-endorphin and is very sensitive to changes in the C-terminal part of the peptide. An analogue having a similar potential for C-terminal amphiphilic helix formation but consisting only of D-amino-acids in residues 13-31 again had high activity in the &-receptor assay.65 A third analogue was designed to minimize the potential for the amphiphilic helix but retained a similar amino-acid composition to the active analogues.This (negative) model showed little activity in the &-receptor assay. These experiments support the idea that the potential for amphiphilic helix formation is an important function for the C-terminal sequence of P-endorphin and similar conclusions have been reached for several other biologically active pep tide^.^^ 5 Endogenous Peptides An increasing number of peptides are isolated using antibodies raised against (synthetic) peptides originally purified from non-mammalian species or predicted from the sequences of cloned DNA coding for peptide precursors.Systematic screening of tissue extracts using chemical or immunochemical assays as well as bioassay has produced many novel peptides. Much research continues to be aimed at defining the distribution of these peptides in vivo and their physiological functions. J. P. Blanc and E. T. Kaiser J. Bid. Chem. 1984. 259 9549. 370 C.E. Dempsey CGRP (171 for example was originally predicted from a cDNA sequence coding for a precursor generated by alternative processing of the calcitonin gene.66 In the thyroid the calcitonin gene is processed to form the calcitonin precursor whereas in non-thyroid tissue (particularly the central and peripheral nervous system) the precursor for CGRP is formed.CGRP has been localized in the nervous system (using a fluorescently labelled antibody),66 was released from stimulated cultured neurons and when injected into the brains of rats produced a marked increase in blood pressure and heart These experiments indicate that CGRP is an endogenous neuropeptide. CGRP also produced peripheral eff ecd7 and it has been shown that injection of as little as 15 femtomoles (15 x mol) into rabbit or human skin induces a marked vasodilatory response.68 This extraordinary potency suggests that the vasodilatory action of CGRP has a physiological function. If a peptide can be shown to have a biological activity at concentrations similar to those occurring in vivo then this supports the idea that the activity is physiologically relevant.An interesting example is Nerve Growth Factor (NGF). NGF is character- ized by its activity on the development of sympathetic and sensory neurons. The main sources of NGF are diverse (mouse salivary gland snake venoms and placenta) but the concentrations in mammalian sera are very low -a few pg per litre. Indeed the low concentrations of NGF complicate its quantification and there is some dispute over serum levels.69970 Nevertheless it has been shown that NGF is a power- ful anti-inflammatory agent in the rat (it suppresses the inflammatory response to an irritant injected into the rat hind paw) and is effective at concentrations of 5 pg kg-'.69 These levels are similar to the concentration of NGF in serum that can be measured by bioassay.NGF has also been shown to induce directed migration in vitro of human cells that are involved in the inflammatory response (although at higher concentrations than are required for anti-inflammatory a~tivity).~' Whether these new activities of NGF are related is not known. Many sequences of biosynthetic precursors for peptide hormones and neuropep- tides have now been obtained by analysis of cloned DNA sequences and this has been of great value in understanding the biosynthetic origins and interrelationships of endogenous peptides. Immunohistochemical co-localization of structurally related peptides has in some cases indicated a common biosynthetic origin and PHI (10) and VIP for example have been identified within the same cell and subsequently found to share a common precursor.12 The evidence that the neurohypophysial hormones oxytocin and vasopressin are processed from precursors which also encode their respective neurophysins has been confirmed by cDNA sequence analysi~.'~ The precursor relationships between families of endogenous peptides are best characterized for the opiate peptides.All the known endogenous opiate peptides 66 M. G. Rosenfeld J.-J. Mermod S. G. Amara L. W. Swanson P. E. Sawchenko J. Rivier W. W. Vale and R. M. Evans Nature (London) 1983 304,129. 67 L. A. Fisher D. 0. Kikkawa J. E. Rivier S. G. Amara R. M. Evans M. G. Rosenfeld W. W. Vale and M. R. Brown Nature (London) 1984 305 535. 68 S. D. Brain T. J. Williams J. R. Tippens H. R. Morris and I.MacIntyre Nature (London) 1985,313 54. 69 B. E. C. Banks C. A. Vernon and J. A. Warner Neurosci. Lett. 1984,47 41. 'O S. Korsching and H. Thoenen Roc. Natl. Acad. Sci USA 1983 80 3513. 71 A. P. Gee M. D. P. Boyle K. L. Munger M. J. P. Lawman and M. Young Proc. Natl. Acad. Sci. USA 1983 80 7215. 72 H. Land G. Schutz H. Schmale. and D. Richter Nature (London) 1982 295 299. Peptides 37 1 are encoded within the sequences of three precursors which have been sequenced from their cDNA. The precursor named pro-opiomelanocortin contains &endorphin (plus several non-opiate peptides like ACTH) proenkephalin A contains the enkephalins (plus a number of C-terminally extended enkephalins such as amidor- phin) and proenkephalin B contains the dynorphin pep tide^.^^ Each of these biosynthetic precursors is processed within anatomically distinct neuronal systems and may undergo differential processing to generate distinct sets of bioactive peptides in different areas of the nervous Two (bovine) precursors containing the sequence of substance P (23) have been sequenced from their cDNA~~ and one of these additionally has the sequence of a peptide homologous to the frog skin peptide kassinin (24).The new peptide [(25) named alternatively substance K,75neurokinin or neuromedin L77] has been independently purified from porcine spinal cord by two Japanese groups76v77 in addition to a second homologous peptide (26) neurokinin p76or neuromedin K.77 Arg-Pro-Lys- Pro-Gln-Gln- Phe- Phe- Gly-Leu-Met- NH (23) Asp-Val- Pro-Lys-Ser- Asp-Gln- Phe-Val- Gly-Leu-Met- NH (24) His-Lys-Thr-Asp-Ser-Phe-Val- Gly-Leu-Met- NH (25) Asp-Met-His-Asp-Phe- Phe-Val- Gly-Leu-Met- NH (26) Peptide (26) is presumably generated from a novel precursor.These peptides are members of a larger class of peptides named tachykinins having the common (italicized) C-terminal sequence and originally discovered in frog skin and octopus salivary gland. As with the opioids tachykinin receptor sub-types exist and it is proposed that peptides (25) and (26) are the endogenous agonists of one subtype (SP-E receptors) while substance P acts at ‘so-called’ SP-P receptors.78 Other peptides homologous to frog skin peptides have been isolated from mammalian tissues. These include neuromedin B and C from porcine spinal which with one amino-acid substitution have the same sequence as the C-terminal decapeptide of bombesin and ovine corticotropin-releasing factor (CRF),80a 41 residue peptide homologous to sauvagine.The enzymatic mechanism for generating C-terminally amidated peptides from their precursors has been studied using model peptides having a CJterminal glycine 73 R. J. Miller J. Med. Chem. 1984 27 1239. 74 N. Zamir E. Weber M. Palkovits and M. Brownstein Roc Nutl. Acud. Sci. USA 1984 81 6886. 75 H. Nawa H. Kotani and S. Nakanishi Nuture (London) 1984 312 729. 76 S. Kimura M. Okada Y. Sugita I. Kanazawa and E. Munekata Roc. Jpn. Acud. Ser. B 1983,59 101. 77 K. Kangawa N. Minamino A. Fukuda and H.Matsuo Biochem. Biophys. Res. Commun. 1983,114,533. 78 J. C. Hunter and J. E. Maggio Eur. J. PharmucoL 1984 97 159. 79 N. Minamino K. Kangawa and H. Matsuo Biochem. Biophys. Res. Commun. 1984 124 925. 80 W. Vale J. Spiess C. Rivier and J. Rivier Science 1981 213 1394. C.E. Dempsey which is known from precursor sequences to precede the basic C-terminal cleavage point of amidated peptides.81 When peptide (27) was synthesized using "N-labelled glycine and incubated with extracts of porcine pituitary (a source of the amidating enzyme) the label was shown by mass spectrometry to be retained in the dipeptide amide (28). When a I4C a-carboxyl analogue of (27) was used the label was found in glyoxylate. These experiments support an oxidative mechanism involving removal of hydrogen from the C-terminal glycine and spontaneous hydrolysis of the resulting imino group (Scheme 4).Results from later experiments have been consistent with this mechanism.82 D-Tyr-Val- NH -CH2C0,H + D-Tyr-Val- N=CHC02H -* D-Tyr-Val- NH2 + CHOC0,H (27) (28) Scheme 4 A large number of peptides have been identified in neurons and nerve terminals of the mammalian central nervous system and these include almost all the known brain and gut peptides many of the classical hormones and a number of peptides first identified in non-mammalian species. Much effort has been spent on determining the distribution of these neuropeptides and their receptors with the expectation that some indication of function might be inferred.Immunohistochemical methods can be used to provide precise localization of immunoreactive species within individual neurons although cross-reaction between antisera and antigens structurally related to the 'target' peptide remains a source of error. This problem is particularly acute with the existence of multiple forms and homologous families of peptidesg3 Studies on neuropeptide Y are illustrative. Immunoreactive peptides recognized using anti- bodies raised against avian pancreatic polypeptide (APP) were localized in peripheral non-adrenergic neurons and in neurons of the human cerebral cortex although attempts to isolate and characterize the immunoreactive peptide have been unsu~cessful.~~ Neuropeptide Y has considerable homology with APP and when antibodies were raised against the former peptide immunoreactive material was found to co-exist in neurons with APP-like irnmunoreacti~ity.~~ When these neurons were re-examined using an antiserum more specific for APP the latter peptide was now undetectable.It was concluded that neuropeptide Y is the endogenous pan- creatic-polypeptide-like species in the mammalian nervous system. In peripheral tissues neuropeptide Y is localized in nerve fibres around blood vessels suggesting an effect on the vasculature and accordingly the peptide has been found to induce vasoconstriction although the concentrations required were rather high. At much lower concentrations neuropeptide Y enhanced the response of blood vessels to other vasoconstrictor agents.86 It is interesting that neuropeptide Y is co-localized 81 A.F. Bradbury M. D. A. Finnie and D. G. Smyth Nature (London) 1982 298 686. 82 C. C. Glembotski B. A. Eipper and R. E. Mains J. Biol. Chem. 1984 259 6385. 83 T. Hokfelt 0.Johansson and M. Goldstein Science 1984 225 1326. 84 J. M. Lundberg T. Hokfelt A. Angglrd L. Terenius R. Elde K. Markey M. Goldstein and J. Kimmel Proc. Natl. Acad. Sci. USA 1982 79 1303. 85 T. E. Adrian J. M. Allen S. R. Bloom M. A. Ghatei M. N. Rossor G. W. Roberts T. J. Crow K. Tatemoto and J. M. Polak Nature (London) 1983. 306 584. 86 L. Edvinsson E. Ekblad R. HHkanson and C. Wahlestedt Br. J. Pharmacol. 1984 83 519. Peptides 373 in sympathetic noradrenergic neurons with the transmitter norepinephrine which itself produces vasoc~nstriction.~~ It is suggested that in this system86 (and perhaps generally83) the peptide co-localized with a classical transmitter supports (or other- wise modulates) the action of the transmitter.Examples of the co-existence of multiple peptides within the same neurons have also been reported. In single rat hypothalamic neurons for example three immunoreactive peptides corresponding to PHI enkephalin and CRF have been identified.87 From the known actions of these peptides their co-release would be predicted to result in the parallel secretion of prolactin ACTH and growth hormone from the anterior pituitary. Such an effect is known to occur in the physiological response to stress and it is suggested that the neuronal co-existence of the former set of peptides may be involved in this response.The study of the function of neuropeptides and particularly their status as neurotransmitters is limited by the lack of truly selective antagonists for their effects in the brain. Recent work on cholecystokinin (CCK) provides an example. CCK is one of the most abundant of neuropeptides. It has been confirmed that a CCK-%like component is the major form in rat brain and that gastrins (which share the C-terminal pentapeptide amide) do not contribute significantly to CCK-like immunoreactivity in the brain.88 In addition CCK-like immunoreactivity has been purified from post-mortem human brain characterized by Edman degradation and negative ion FAB-MS and shown to be largely the sulphated CCK octapeptide (29).89CCK-like immunoreactivity is co-localized with the transmitter dopamine Asp-Tyr(OS03H)-Met-Gly-Trp-Met- Asp-Phe-NH (29) and because there is evidence that a hyperactive dopamine system underlies some of the symptoms of schizophrenia a role for CCK in the disease has been sought.Accordingly it has been shown that CCK increases the neuronal activity of dopamine-containing neurons suppresses the release of dopamine from neurons and is deficient in the brain and cerebrospinal fluid of untreated schizophrenics (see re$ 90). Confirmation of the central effects of CCK requires that they be blocked by a specific antagonist. An antagonist of the peripheral effects of CCK is proglumide (a glutaramic acid derivative). When proglumide was tested on the effects of CCK on dopamine-containing neurons the drug was effective in blocking the increase in neuronal activity of dopaminergic neurons by CCK but not the attenuation of dopamine release.It was suggested that the two actions of CCK are mediated by different receptor subtypes one of which is unresponsive to the effects of progl~mide.~' Receptors for vasopressin (30)are also heterogeneous in their response to antagon- ists and this has led to controversy over the role of the peptide in stress-induced secretion of ACTH from the pituitary. The hypothalamic factor responsible for the 87 T. Hokfelt J. Fahrenkrug K. Tatemoto V. Mutt S. Werner A.-L. Hulting L. Terenius and K. J. Chang Roc. Natl. Acad. Sci. USA 1983 80 895. 88 P.D. Marley J. F. Rehfeld and P. C. Emson J. Neurochem. 1984 42 1523. 89 L. J. Miller I. Jardine E. Weissman V. L. W. Go and D. Speicher J. Neurochem. 1984 43 835. 90 R.Y. Wang F. J. White and M. M. Voigt Trends Pharmacol. Sci. 1984 5 436. 374 C.E. Dempsey stress-induced release of ACTH is known to be a multifactorial complex one component of which appears to be CRF." Arginine vasopressin (30) has been proposed as another component and the peptide is known to potentiate the ACTH- releasing activity of CRF in vitro. A potent antagonist of the physiological effects of vasopressin is the synthetic analogue (31) and this was found to inhibit the increase in ACTH levels induced by vasopre~sin.~~ Levels of ACTH induced by stress were not inhibited by the drug (31) however and it was concluded from this that vasopressin is not involved in stress-induced release of ACTH.This conclusion was later disputed when it was shown that the pituitary vasopressin receptor is unresponsive to the vasopressin antagonist (3 1).92 Recent experiments have suppor- ted the idea that there are two distinct effects of vasopressin on the release of ACTH and that the indirect effect characterized by the potentiation of CRF-induced ACTH release (and apparently unresponsive to the standard vasopressin antagonists) is of physiological relevan~e.~~?~~ There is some evidence to suggest that metabolites of several brain peptides (rather than or in addition to the native peptide) may mediate particular effects in the brain.The idea (of De Wied) that vasopressin is involved in the consolidation of long term memory was described in Annual Reports for 1980.' Synthetic fragments of vasopressin are fully effective in mediating the behavioural effects of the peptide and for this reason an active metabolite has been proposed to occur in the brain. To test this idea vasopressin was incubated with isolated rat brain synaptic mem- branes and the metabolites were recovered and ~haracterized.~~ Arginine vasopressin was found to be degraded by an aminopeptidase which cleaved the peptide from the N-terminus but left the disulphide bond intact. Proteolysis was inhibited on formation of an N-terminal pyroglutamyl residue from Gln-4 and the stable meta- bolite (32) therefore accumulated.This peptide was 1000-times more active than Cys-Tyr-Phe-Gln- Asn-Cis-Pro- Arg-GI y-NH (30) CMe,-CH,-CO-Tyr( 0Me)-Phe-Gln- Asn-Cys-Pro-Arg-Gly-NH S-1 (31) pGlu-Am-&( Cis)-Pro- Arg-Gly-NH (32) arginine vasopressin in potentiating passive avoidance behaviour in rats a behavioural test used to assess the consolidation of long term memory. The authors report that preliminary studies using radioimmunoassay and h.p.1.c. indicate that a peptide with the properties of the metabolite (32) is present in rat brain. 91 P. Mormbde Nature (London) 1983,302 345. 92 A. J. Baertschi B. Gahwiler F. A. Antoni M. C. Holmes and G. B. Makara Nature (London) 1984 308 85. 93 J. C. Buckingham Br. J. PharmacoL 1985,84 213. 94 J. P. H. Burbach G.L. Kovics D. De Wied J. W. van Nispen and H. M. Greven Science 1984,221 1310. Peptides 375 Enzymatic processing of endogenous peptides may also result in peptides having antagonist properties. @Endorphin (1-27) is present in considerable amounts in the brain and has been assumed to be an inactivated form of &endorphin (11). Synthetic /3-endorphin (1-27) was found to be a potent antagonist of P-endorphin- induced analgesia and it is suggested that the fragment may have a physiological function perhaps of modulating the activity of opiates in ~ivo.~' 6 Exogenous Peptides A number of toxic peptides have been purified from the venom of the marine mollusc Conus geogruphus. Conotoxin GI (33)96 has an amidated C-terminus and two disulphide bonds paired97 in a manner similar to the bee venom peptides apamin and peptide-401.Two homologues of conotoxin GI have in one case a C-terminal Gly-Lys-NH extension and in the other amino-acid substitutions at positions 4 Glu-C~s-Cys-Asn-Pro-Ala-Cls-Gly-Arg-His-Tyr-Ser-~y~-~~* (33) 9 and 11 of the sequence.96 The conotoxins are postsynaptic toxins and appear to bind competitively to the acetylcholine receptor. Three longer toxins have been isolated from the same source and are characterized by an unusual number of hydroxylated amino-acids including hydroxyproline (Hyp).98*99 Each has two disul- phide bonds (one has consecutive cysteines as in the conotoxins) but the disulphide pairings have not been determined. One of these peptides o-conotoxin (34),99 acts presynaptically at the frog skeletal neuromuscular junction to block transmitter release apparently by blocking the stimulated influx of calcium.loO 10 Cys-Lys-Ser-Hyp-Gly-Ser-Ser-Cys-Ser-Hyp-Thr-Ser-Tyr-Asn-Cys-Cys-~g-Ser- 20 27 -Cys-Asn-Hyp-Tyr-Thr-Lys-Arg-Cys-Tyr-NH2 (34) Precursors encoding neuropeptides of the marine mollusc ApZysia have been sequenced from their cDNA and this has led to interesting speculation about the role of neuropeptides in the mediation of behaviour in this species."' The egg-laying behaviour of ApZysiu is accompanied by a stereotyped behavioural pattern which is known to be controlled by 'so-called' bag cells in the abdominal ganglion.Peptides seem to be neurotransmitters in the regulation of egg-laying behaviour and an egg-laying hormone (ELH) has previously been isolated and characterized.The 95 R. G. Hammonds P. Nicolas and C. H. Li Roc. Natl. Acad. Sci. USA 1984 81 1389. 96 W. R. Gray A. Luque B. M. Olivera J. Barrett and L. J. Cruz J. Biol. Chem. 1981 256 4734. 97 W. R. Gray F. A. Luque R. Galyean E. Atherton R. C. Sheppard B. L. Stone A. Reyes J. Alford M. McIntosh B. M. Olivera L. J. Cruz and J. Rivier Biochemistry 1984 23 2796. 98 S. Sato H. Nakamura Y. Ohizumi J. Kobayashi and Y. Hirata FEBS Let?. 1983 155 277. 99 B. M. Olivera J. M. McIntosh L .J. Cruz F. A. Luque and W. R. Gray Biochemistry 1984 23 5087. 100 L. M. Kerr and D. Yoshikami Nature (London) 1984,308 282. 101 R. H. Scheller J. F. Jackson L. B. McAllister B. S.Rothman E. Mayeri and R. Axel Cell 1983 32 7. C. E. Dempsey precursor for ELH also contains the sequence of a second egg-laying hormone which has been isolated from bag cells and named a-bag cell peptide (35). The precursor additionally contains the sequence of eight further potential peptides and it is proposed that the co-ordinated release of these peptides from bag cells mediates the stereotyped behaviour associated with egg laying. It remains to be seen whether the predicted peptides are released and contribute to the behavioural pattern as suggested. Ala-Pro- Arg-Leu- Arg-Phe-Tyr-Ser-Leu (35) The characterization of peptides from non-mammalian sources has been stimu- lated by the finding that many of the peptides are highly conserved phylogenetically and are found as neuropeptides in mammals.An extraordinary example is the head activator peptide from the freshwater coelenterate hydra. The peptide is characterized by its ability to activate the regeneration of a head in hydra which have been surgically decapitated and has the structure (36). A peptide having an identical structure has been isolated from bovine and human hypothalami.lo2 The function of the peptide in mammalian brain is not known. pGlu-Pro-Pro-Gly-Gly-Ser-Lys-Val-Ile-Leu-Phe (36) Equally surprising is a report that pig brain contains a peptide equivalent to apamin (19).'03Apamin is a potent neurotoxin and its activity arises from its ability to block the Ca2+-activated increase in permeability to K+ in neurons (and hepatocytes) possibly by binding to a K+-selective channel.The apamin-like pig brain peptide has not been fully characterized but has the following properties i it strongly cross-reacts with an antiserum raised against apamin; ii it binds with high affinity to Ca2+-activated K+ channels in synaptosomal membranes and dis- places bound apamin; iii it contracts intestinal smooth muscle preparations in a manner similar to apamin; and iv like apamin [see sequence (19)] it is inactivated by trypsin but is resistant to chymotrypsin. This is the first indication that endogenous equivalents of natural toxins exist and the structure of the apamin-like factor is of great interest. It is likely that this finding will stimulate the search for other endogenous peptides related to channel-specific venom toxins.102 H. Bodenmuller and H. C. Schaller Nature (London) 1981 293 579. 103 M. Fosset H. Schmid-Antomarchi M. Hugues G. Romey and M. Lazdunski Roc. Natl. Acad. Sci. USA 1984,81 7228.