Coupling in the absence of tertiary amines. V. Removal of the 2-nitrobenzenesulfenyl (NPS) group with 1-hydroxybenzotriazole in the presence of aniline or N-methylaniline.

In a search for conditions of acidolytic removal of amine protecting groups leading to salts of the deblocked amine that can be acylated without addition of a tertiary amine, cleavage of the 2-nitrobenzenesulfenyl (Nps) group with hydroxybenzotriazole (HOBt) in 2,2,2-trifluoroethanol was attempted. The Nps group was smoothly removed, but the resulting salt of the amine component could not be acylated unless deprotonated with a tertiary base. A rationale is now proposed for this unsatisfactory outcome of the cleavage reaction and for the concomitant surprising reduction of HOBt to benzotriazole. Based on the proposed mechanism, a new approach was designed for the removal of the Nps group. It was cleaved with HOBt in the presence of weakly basic nucleophiles such as aniline, N-methylaniline or 8-aminoquinoline. The protecting group was transferred smoothly to the amino group of the nucleophilic acceptor leaving the deblocked amine component in the form of its HOBt salt. This was then readily acylated without addition of a tertiary amine.

Reactions of C-terminal omega-amino acid residues in liquid hydrogen fluoride.

The benzyl-ester bond linking the C-terminal delta-aminovaleric acid residue of a peptide to a polymeric support was cleaved with liquid hydrogen fluoride in the presence of anisole, added as scavenger. Instead of the expected peptide with a free carboxyl group at the C-terminus, a peptide terminating in a ketone derivative was obtained. The unusual extent of this known side-reaction was attributed to the effect of the distance between the amino group and the carboxyl group in the C-terminal residue. The results of model experiments corroborated this view.

Contractile activity of vasotocin, oxytocin, and vasopressin on mammalian prostate.

The neurohypophyseal peptides arginine vasotocin, oxytocin and arginine vasopressin contracted guinea pig, rat, canine and human prostates with potencies and efficacies that were comparable to those of noradrenaline and methacholine. All three neuropeptides raised prostatic tone and elicited contractions at 10(-9) or 10(-8) M, with an order of efficacy: arginine vasotocin greater than oxytocin greater than arginine vasopressin. The findings suggest a physiological role for these peptides in prostatic smooth muscle contraction and possibly also in other aspects of male reproductive function.

The myth of coupling reagents.

Coupling reagents have been proposed for peptide synthesis ever since the introduction of the method because of the convenience of the procedure, which consists of peptide bond formation by addition of a specific condensing agent to the mixture of carboxyl and amine components. However, truly efficient and yet innocuous coupling reagents are hard to find. Most coupling reagents give rise to unwanted reactions, which can be traced to certain characteristic structural features in their molecules. Such features are illustrated with the intermediates generated by dicyclohexylcarbodiimide and N-ethyl-5-phenyl-isoxazolium-3'-sulfonate (Woodward's reagent K). Some side reactions can be avoided if peptide bond formation is carried out in two separate steps: activation of the carboxyl group of the carboxyl component in the form of an active ester, followed by the catalyzed aminolysis of the active ester.

Active esters in solid-phase peptide synthesis.

Incorporation of single amino acid residues into peptide chains built on insoluble polymeric supports a priori appeared promising: the use of isolated, well defined (and potentially commercially available) reactive intermediates were expected to reduce the extent of undesired side reactions. In spite of these expectations active esters were only infrequently used in solid-phase peptide synthesis, mainly because the reaction rates achieved with them were insufficient for rapid chain-lengthening that became possible with automated instruments. In recent years, however, a certain revival of the active ester principle can be noted. This is the consequence of two factors: the application of highly reactive esters and the discovery of efficient catalysts of the ester-aminolysis reaction. The mechanism of catalysis and its explantation for further improvements are also discussed.

Conformation of peptides of the secretin-VIP-glucagon family in solution.

The significance of a well defined molecular architecture in hormone receptor interaction and the methods available for the study of preferred conformations are discussed. The conformational freedom in glucagon is a major obstacle in the determination of its biologically relevant geometry. In the secretin molecule intramolecular forces generate a folded, partially helical conformation. In respect of long range cooperative interactions resulting in a compact molecule with secondary-tertiary structure secretin is similar to globular proteins. In VIP some characteristics of secretin and also of glucagon can be recognized. Further progress in conformation analysis can be expected from the study of rigid, cyclic analogs in which the biological activities of the parent hormones are retained or even enhanced. Such analogs have well defined conformations without external stabilization from membrane mimicking lipids. Therefore, they provide information on the biologically relevant geometry of the hormones and contribute also to our knowledge of receptor sites.

In search of new methods in peptide synthesis. A review of the last three decades.

An attempt was made to discern ideas and trends in the development of peptide synthesis and to recognize general principles of the discipline. Introduction of efficient methods of activation and coupling during the early years of the reviewed period was followed by only moderate further improvements. Major advances were achieved by the discovery of novel methods of protection and by techniques of facilitation. Improvements in the methods of deblocking hold considerable promise and might bring significantly closer the goal of peptide synthesis: the direct preparation of homogeneous products.

Self-association and solubility of peptides. Solvent-titration study of protected C-terminal segments of porcine secretin.

Self-association of peptides (related to the C-terminal sequence of porcine secretin) in methylene chloride was disrupted by adding dimethylsulfoxide in increasing amounts. This structural transition was monitored by the disappearance of the amide-I C = O stretching band of strongly intermolecularly hydrogen-bonded molecules (1625-1630 cm-1) in the infrared absorption spectra. The effects induced by main-chain length and sequence, type of N alpha-protection, and concentration were assessed. Hexamethylphosphortriamide was compared for its structure-disrupting properties to dimethylsulfoxide. The increasing propensity to aggregate displayed by these peptides is paralleled by a decrease in their solubility. The impact of these results on the planning of peptide syntheses is briefly discussed.

Cyclization studies with a model pentapeptide.

A pentapeptide with the sequence Gly-Ala-D-Val-Leu-Ile was designed for a study of cyclization. Isoleucine was selected as the C-terminal residue in order to determine, from the amount of alloisoleucine in the cyclic product, the extent of racemization during activation and ring closure. The insolubility of cyclo (glycyl-alanyl-D-valyl-leucyl-isoleucyl) in the commonly used solvents facilitated its isolation and thus the evaluation of comparative experiments. Because of its thermal stability the cyclopentapeptide could be purified by sublimation in vacuo. The results of cyclization experiments carried out with this model suggest that separation of the steps of activation and coupling is preferable to cyclization with the help of coupling reagents, that is to the execution of activation and ring closure in a single operation.

9-Fluorenylmethyl esters.

N alpha-Protected amino acid 9-fluorenylmethyl esters (Fm esters) were prepared by imidazole-catalyzed transesterification of active esters with 9-fluorenylmethanol (9-hydroxymethylfluorene). The new carboxyl protection is unaffected by acids, but is efficiently removed by beta-elimination under the influence of secondary and tertiary amines. Primary amines and ammonia can cause slight amide formation. Deblocking was achieved also by catalytic hydrogenation.

Experiments on the protection of the thioether in methionine.

The decrease in the solubility of peptides when their methionine residues are replaced by methionine sulfoxides prompted the exploration of an alternative approach to the protection of the thioether in methionine side chains. Alkylation of tert.-butyloxycarbonyl methionine p-nitrophenyl ester with methyl p-toluenesulfonate yielded the crystalline derivative of methionine S-methyl p-toluenesulfonate which could be incorporated into peptide chains. Alternatively, methionine S-methyl p-toluenesulfonate (Mmt) residues could be generated by the action of methyl p-toulensulfonate on methionine containing peptides. The protecting group remained intact under the conditions of aminolysis and ammonolysis commonly used in peptide synthesis, and it was unchanged after the removal of other blocking groups with trifluoroacetic acid or diethylamine. On treatment with hydrobromic acid in acetic acid the toluenesulfonate anion was replaced by bromide ion, while hydrogenation resulted in the decomposition of the modified methionine side chain. In the process of deprotection Mmt residues could be smoothly converted to methionine residues by thiolysis. Thus, the protecting group functioned well in several respects but an increase in solubility (in dimethylformamide) on alkylation was observed only in a part of the peptide derivatives tested. Therefore, the value of the new approach for the protection of the methionine side chain in peptide synthesis remains to be established.

Derivatives of S-9-fluorenylmethyl-L-cysteine.

The 9-fluorenylmethyl (Fm) group was examined with respect to its potential for blocking the sulfhydryl function. The S-Fm group is resistant to acids and to catalytic hydrogenation but is cleaved by ammonia in methanol or by organic bases, such as a 20% solution of piperidine in dimethylformamide. Synthesis of N-tert.-butyloxycarbonyl-S-9-fluorenylmethyl-L-cysteine p-nitrophenyl ester and of cysteinyl peptides protected with the S-Fm group are described.

Coupling in the absence of tertiary amines.

In order to avoid base catalyzed side reactions during coupling, attempts were made to render superfluous the addition of tertiary amines to the reaction mixture. Weak acids were applied for the removal of acid labile protecting groups. Acetic acid and other carboxylic acids were considered unsuitable for this purpose coupling step. Pentachlorophenol and 2,4-dinitrophenol cleaved the Bpoc, Nps and Trt groups but more practical rates were reached with solutions of 1-hydroxybenzotriazole (HOBt) in trifluoroethanol, in acetic acid, or in a mixture of phenol and p-cresol. In addition to acidolysis, HOBt salts of amino components could also be obtained through hydrogenolysis of the Z group or thiolysis of the Nps group in the presence of HOBt, or by the displacement of acetic acid from acetate salts with HOBt. Acylation of HOBt salts of amino components with symmetrical or mixed anhydrides or with active esters did not require the addition of tertiary amine.

Selective modification of the ability of cholecystokinin to cause residual stimulation of pancreatic enzyme secretion.

In the C-terminal heptapeptide of cholecystokinin (-Tyr(SO3H)-Met-Gly-Trp-Met-Asp-Phe-NH2), replacing the aspartic acid residue by beta-aspartic acid did not alter the ability of the peptide to cause stimulation, desensitization, or residual stimulation of enzyme secretion from dispersed pancreatic acini. Replacing the tyrosine sulfate residue by hydroxynorleucine sulfate did not alter the ability of the heptapeptide to cause stimulation or desensitization, but caused a 50-fold decrease in the potency with which the peptide caused residual stimulation of enzyme secretion. These findings suggest that a modification of the N-terminal region of cholecystokinin heptapeptide, which does not alter the ability of the peptide to bind to its receptor on pancreatic acini and by so doing cause stimulation and desensitization of enzyme secretion, can increase the rate at which the bound peptide dissociates when the acini are washed and reincubated. This increased dissociation is reflected by a reduction in the potency with which the peptide causes residual stimulation of enzyme secretion.

Allomalformin.

In an attempt to find explanation for the initial erroneous sequence assignment for malformin, a sequence-isomer of the natural product, 3-isoleucine-5-valine malformin or briefly "allomalformin" that on partial acid hydrolysis could have given rise to misleading fragments, was synthesized and compared with both natural and synthetic preparations of malformin. Allomalformin is identical to the parent microbial peptide (malformin A, or briefly malformin) with respect to biological activity and conformation (ORD and CD spectra) and is indistinguishable from it by high pressure liquid chromatography. Yet, the two isomers have slightly different Rf values on thin-layer chromatograms and by this method no allomalformin could be detected in samples of the natural product. On the other hand both high pressure liquid chromatography and thin-layer chromatography demonstrated the presence of the lower homolog, 5-valine malformin, in the samples examined. On partial acid hydrolysis this natural analog should liberate Val-Cys, while Cys-Val forms from malformin itself. Similarly, the corresponding desthio cyclopentapeptides should give rise to Val-Ala and Ala-Val respectively; the former being more resistant to further hydrolysis persists in the partial hydrolysates. The presence of Val-Cys in partial hydrolysates of malformin and of Val-Ala in the partial hydrolysates of desthiomalformin, both originating from the accompanying lower homolog rather than from malformin itself, is likely to have led to the postulation of the erroneous Cys-Val-Cys partial sequence.