Revisiting Oxytocin Through the Medium of Isonitriles

Abstract

The reaction of thioamino acids and N-terminal peptides, mediated by hindered isonitriles and HOBt, gives rise to peptide bonds. In one pathway, oxytocin was synthesized by eight such reiterative amidations. In another stereospecific track, oxytocin was constructed by native chemical ligation, wherein the two building blocks were assembled by thioacid amine amidation. The NMR spectra of oxytocin and dihydro-oxytoxin suggest a high level of pre-organization in the latter, perhaps favoring oxidative folding.

Oxytocin is a long-known, well-appreciated 9-mer peptide hormone which enables parturition and lactation in nursing mothers.^1,2 More recently, there has been a growing perception that the receptor–binding characteristics of nonapeptides, including oxytocin, are of considerable moment in other neuropathological phenomena, including autism. Indeed, some encouraging effects, including indications of memory enhancement and improved social associations, have been demonstrated in autistic subjects following inhalation of oxytocin, in the context of an organized clinical trial.^3,4 It has been argued that a breakdown in the bioprocessing of polypeptidic precursors en route to oxytocin may be implicated in provoking and maintaining the autistic state.⁵ Accordingly, it seemed to us worthwhile to revisit the oxytocin manifold – first from a chemical perspective, and then for the purpose of promoting improved biological function via wisely selected oxytocin-inspired congeners. Though oxytocin (1) itself has been synthesized several times,⁶ the corresponding “dihydro” bis cysteinyl compound, 2, had apparently not been characterized. We also hoped to reach characterizable 2 as part of our total synthesis effort. We could then be in a better position to understand the structural changes which must occur as a consequence of oxidative disulfide formation in the parent system, and particularly in analogs we would be synthesizing and evaluating.^7,8

In addition to a clear basis for interest in oxytocin at various biological levels, we also viewed it as a compelling object molecule for testing the usefulness and scope of methodology which we have recently been developing, involving the use of isonitriles in the chemical synthesis of amide bonds, including those in peptidic contexts.⁹ The reaction of particular interest for us, drawn from our menu of new isonitrile chemistry, involves coupling of a thioacid (3) with an amine (5) in the presence of a hindered isonitrile (4). As we have shown,^10,11 this three-component reaction gives rise to an amide linkage under remarkably mild conditions. Evidence has been provided that the reaction commences with an addition event, which gives rise to a thio-FCMA (6). The latter is apparently a very powerful acyl donor and, upon reaction with a suitable amine, provides an amide (7).¹² While the thioacid–amine coupling can occur with no further additives, it was recently shown to be enhanced by the presence of HOBt.¹³ It seems not unlikely that under these HOBt conditions, the reaction progresses from the thio-FCMA through an HOBt ester (8), en route to amide 7. (Figure 1) Thus, we viewed oxytocin as an attractive target to test the scope of the hindered isonitrile/HOBt–mediated amidation process, hopeful that information gleaned from the parent target system could be applied to synthesizing our proposed analogs.

Figure 1.

Oxytocin (1)and dihydro-oxytocin (2).

The oxytocin directed synthesis program began by coupling of glycine methyl ester 9 and the leucine-derived thio-acid 10. Indeed, under mediation by t-BuNC and HOBt, dipeptide 11 was obtained in 92% yield. The latter was converted to tripeptide 13 in 66% yield over two steps, through t-Boc cleavage and subsequent isonitrile-mediated amidation with thioacid 12. Cleavage of the t-Boc protecting group of 13, followed by coupling the resulting N-terminus with 14 afforded tetrapeptide 15 in 64% yield over two steps. The isonitrile–mediated “peptide homologation” strategy was then demonstrated to be compatible with a suitable asparagine derivative (16), providing pentapeptide 17 in 84% yield. Similarly, following cleavage of the carbamate linkage of 17, the resulting N-terminus served to amidate glutamine-derived thioacid 18 to afford hexapeptide 19 in 82% yield over the two steps (Scheme 1).

Scheme 1.

Synthesis of intermediate 19.

Our journey then progressed through the same two step coupling–deprotection elongation process, using, in sequence, isoleucine-derived thioacid 20, tyrosine-derived thioacid 22 and cysteine-derived thioacid 14, in 77%, 78%, 84% yields, respectively, thereby affording 24. Conversion of the C-terminal methyl ester to the corresponding primary amide followed by global deprotection provided the linear nonapeptide 2. Happily, the chemistry described above allowed us to characterize “dihydro-oxytocin” 2 as a homogeneous entity. Finally, aeration of 2 in pH=7 aqueous solution furnished oxytocin 1 in 63% isolated yield (Scheme 2). The 600 MH_z H¹ spectra of homogeneous 2 and 1 are provided for comparison (Figure 2). Their striking similarity, both in terms of levels of definition, and in appearance, suggests significant preorganization of 2 en route to 1.¹⁴

Scheme 2.

Synthesis of oxytocin (1) and dihydro-oxytocin (2).

Figure 2.

¹H NMR Spectra of of oxytocin (1), in red, and dihydro-oxytocin (2), in blue.

Having demonstrated the reach of isonitrile mediated thioacid amidation by the sequential, step-by-step C3N elongation synthesis of oxytocin, we next investigated the applicability of isonitrile chemistry to the enablement of a more convergent strategy. With the goal of avoiding epimerization, we first took recourse to the native chemical ligation (NCL) logic of Kent and associates.¹⁵ One of the highlights of the Kent chemistry is its extraordinary level of stereointegrity (epimerization avoidance) in the key amide-forming ligation step. We hoped to demonstrate that the thioacid/isonitrile/HOBt–mediated amidation can be used to provide coupling substrates for an NCL ligation. We particularly hoped to exploit our orthomercaptoaryl ester rearrangement (OMER) variation of the classical NCL to generate the O-hydroxy activated phenyl thioester.^16,17

Accordingly, protected glutamine 26 was prepared from 25 by esterification in 87% yield (Scheme 3). Removal of the t-Boc group followed by coupling with isoleucine-derived thioacid (20) in the presence of t-BuNC and HOBt afforded dipeptide 27 in 83% yield. After elongation with thioacid 22 by isonitrile chemistry, the resulting adduct was subjected to 20% piperidine/CHCl₃ solution to furnish tripeptide 28. Elongation of 28 via isonitrile-mediated coupling of cysteine-derived thioacid (29) and tripeptide 28, followed by deprotection of the trimethylsilylethyl ester, afforded tetrapeptide 30. Coupling of 30 with asparagine derivative 31 bearing the required C-terminal OMER^16,17 machinery, followed by global deprotection, using cocktail B(trifluoroacetic acid, phenol, triisopropylsilane, H₂O) provided NCL ligation partner 32. The complementary ligation element, 33, was prepared in two steps from 15 via aminolysis followed by global deprotection to liberate the N-terminal cysteine.

Scheme 3.

Synthesis of oxytocin (1) through NCL.

Merger of 32 and 33 occurred smoothly in a pH 7.2 buffer solution, providing a ligation product, which, when treated with O-methylhydroxylamine,¹⁸ furnished (the previously synthesized) 2 with no detected epimerization. As before, aeration of 2 resulted in oxytocin 1. Since this remarkably smooth chemistry gives rise to readily purifiable compounds, it was not surprising that the discerning high field (600 MH_z) spectra of 1 and 2 were in each case identical with those obtained by the linear route described above.

We next evaluated the feasibility of a more risky proposition, i.e. the possibility of using the thioacid/isonitrile/HOBt-mediated amidation reaction (see Figure 1) for the ligation step, itself. The question was whether “stereointegrity” at the C-terminus of the acyl donor could be maintained even when the thioacid is placed at the terminus of a peptide. It is well perceived that activated acyl donors corresponding to C-terminal peptides can be vulnerable to apparent oxazolone formation, accompanied by epimerization of the α-C-terminal stereocenter (cf 39→40) prior to ligation.^19–21 We hoped to probe this question in the context of C-terminal thioacid 36.

Toward this end, tripeptide 28 was coupled with Boc-Cys(Trt)-SH (14) in the presence of t-BuNC and HOBt (Scheme 4). The ester linkage of the resulting adduct was converted to its corresponding carboxylic acid 34 through use of TBAF (61% over the two steps from 28). Coupling of 34 with aspartate derived Fm thioester 35 followed by liberation of the thioacid afforded the required 36 in 70% yield.

Scheme 4.

Attempted synthesis of 1 through isonitrile/HOBt-mediated couplings.

The N-terminal cysteine coupling partner 37 was readily retrieved from its previously described protected form 15, as shown. Isonitrile/HOBt mediated ligation of 36 and 37 occurred reasonably well, leading to the previously characterized 24 in 61% yield. However, the desired 24 was accompanied by significant amounts of its D-aspartate–containing epimer, 38. Even after significant attempts at optimization of stereointegrity in the coupling step, the best ratio of 24:38 we obtained was only 6:1. While the feasibility of suppression of epimerization in such isonitrile–mediated thioacid ligations may well be sequence dependent, stereointegrity in this type of ligation is clearly not yet a solved problem.

In summary, the amidation of thiol acids derived from a wide range of potential single amino acids, with amines (including those from a variety of amino acids and N-terminal peptides) has been demonstrated. This central reaction has been iterated eight times to produce homogeneous dihydro (bis cysteinyl) oxytocin (2). This intermediate was oxidatively cyclized to afford oxytocin (1). The same type of isonitrile/HOBT mediated chemistry was used to prepare subunits to achieve an NCL type ligation using the OMER^16,17 method to generate the required C-terminal activated thioester. Throughout these reactions, using Boc protected single amino thioacids, there was no indication of racemization in the acyl donor. However, epimerization was noted with a C-terminal thioacid of pentapeptide 36.

Comparison of the high field (600 Mz) H′ spectra of 2 and 1 suggests that oxidative folding could well be facilitated by considerable pre-organization. Studies involving the extension of isonitrile chemistry to more automatable settings in polypeptide synthesis, as well as its application to the facilitation of SAR studies of oxytocin and related behavior modifying congeners are in progress.

Scheme 5.

Mechanism of epimerization of activated thio-FCMA or HOBt ester.