|Home | About | Journals | Submit | Contact Us | Français|
A total synthesis of the putative structure of nagelamide D from imidazole is described. A Stille cross-coupling is used to construct the bis imidazole skeleton, and the pyrrolecarboxamides are introduced via a double Mitsunobu reaction using a pyrrolehydantoin derivative. Discrepancies between the published spectroscopic data and that reported in the literature cast doubts either on the assigned structure or the reported data.
The oriodin alkaloids are a growing family of pyrrole-imidazole alkaloids produced by members of the Agelisida, Axinellida, and Halichondrida orders of marine sponges (Figure 1).1 These natural products range in complexity from the simple monomeric parent system oroidin (1)2 to the most complex member reported to date, the tetrameric styllisadine A.3 Formally, these natural products arise from a variety of oxidation and oligomerization pathways of the basic building block oroidin (1) and congeners, but little is known about the details of their biosynthesis.4, 5 In large part due to the challenging structural problems presented by the dimeric members of this group, we and others have become interested in developing general synthetic approaches to the pyrrole-imidazole alkaloids including ageliferin (2),6 axinellamine A,7 palau'amine,8 and the nagelamides (3-6).9-11 The original group of nagelamides, is a collection of eight oroidin dimers, which were isolated from an Okinawan marine sponge, Agelas sp. by Kobayashi and coworkers.9 Subsequently, additional members of this subfamily have emerged which reinforce the structural diversity found in the oroidin natural products. Of the initial collection of natural products, two broad groups of compounds were described reflecting differences in one C-C bond (between C9 and C9′) leading to a polysubstituted tetrahydrobenzimidazole type of system (e.g., ageliferin (2) and nagelamide G (4)) and the seco systems (e.g., nagelamides A (5), C (3) and D (6)). The examples lacking the additional ring exhibited different levels of oxidation and may in fact serve as biosynthetic precursors to other members of this family. To date, only one total synthesis of a nagelamide, nagelamide E (10-epi-2) has appeared in the literature, and this was as a byproduct in an approach to ageliferin (2).12 Generally our approaches to the various members of this family of natural products have relied on the elaboration of simple imidazole derivatives as a means to construct highly functionalized imidazole-containing synthons.11, 13 In this manuscript, we describe an application of this strategy in a total synthesis of the reported structure of nagelamide D (6).14
Retrosynthetically it was envisioned that the core bis imidazole skeleton 7 found in nagelamide D (6) would be constructed through an appropriate cross-coupling reaction between an imidazolyl fragment 10 and vinyl fragment 11 (Figure 2). Experience gained in earlier endeavors suggested that Z-11 (X = SnBu3) and 10 (Y = I) would serve these roles.15 Unfortunately, initial attempts to access the Z-vinylstannane 11 were not successful, and therefore we investigated the preparation of the corresponding and more accessible E-vinylstannane which upon cross-coupling and reduction would provide the same intermediate 7 suitable for elaboration to 6. Both cross-coupling fragments can be traced back to the protected diiodoimidazole 12.15 Elaboration of the diimidazole core 7 through double C2-amination (Figure 2, 7→8)16 and double amidation of the diol would provide the complete skeleton of nagelamide D. Reduction of the azides and deprotection would then furnish the natural product 6.
The imidazolyl iodide component 17 (Scheme 1) was assembled in rapid order from the DMAS-protected 4,5-diiodoimidazole 12 (DMAS = dimethylaminosulfonyl), which can be formylated by metallation at C5 with EtMgBr and treatment with N-(2-pyridyl)-N-methylformamide forming 1317, 18 A Horner-Wadsworth-Emmons reaction provides the acrylate derivative 15, which can be reduced to the allylic alcohol 16 on treatment with DIBAL. Reaction of allylic alcohol 16 with TBSCl affords the silyl ether 17 and the first fragment needed for cross-coupling. The E-vinylstannane 20 was prepared via the hydrostannylation of TBS-protected propargyl alcohol 19 (Scheme 2). Despite extensive experimentation with conditions and protecting groups, mixtures of the α- and β–isomers (~3:2) were always obtained, and unfortunately the regioselectivity of the hydrostannylation reaction could not be improved. The regiochemistry and the stereochemistry of this reaction were determined by the magnitudes of the vicinal coupling constants and Sn/H coupling constants. The imidazole-substituted propargyl alcohol 19 in turn was prepared via a Sonogashira reaction between 4-iodoimidazole 18 and TBS-protected propargyl alcohol.16
With the requisite two fragments in hand we began to investigate the cross-coupling reaction, and ultimately determined that the fluoride-mediated Stille cross-coupling conditions reported quite recently by Baldwin and co-workers provided the bis vinylimidazole 22 in good yield after treatment with TBAF to complete the partial desilylation.19 At this point, catalytic hydrogenation led to saturation of both double bonds to give the diol 23 (Scheme 3),20 which was then protected as the bis silyl ether 24 (Scheme 4). Double deprotonation of the imidazole C2 positions with n-BuLi and trapping with TsN3 provides the bis azide 25 in 71% yield (Scheme 4).16 Deprotection of the silyl ethers occurred on treatment with TBAF leading to the formation of the expected diol 26. At this stage the introduction of the remaining nitrogen and the two dibromopyrrole moieties was all that remained to complete the synthesis. Strategically, one might envision accomplishing this task by conversion of the primary alcohols to the primary amines via the azide or phthalimide, followed by acylation. However, we had previously demonstrated that the pyrrolecarboxamide moiety could be introduced directly as a pyrrolehydantoin via a Pd-catalyzed allylic substitution reaction.21, 22 Therefore, it was not difficult envision using a similar type of substrate as a nucleophile in a Mitsunobu reaction. Diol 26 was subjected to a double Mitsunobu reaction with the dibromopyrrolehydantoin derivative 27 leading to 28 (Scheme 4).23 In order to prevent competitive formation of the iminophosphorane by reaction of the triphenylphosphine with the azide moieties, triphenylphosphine was reacted with DIAD prior to introduction of the diol.24 In this way the full carbon skeleton of nagelamide D was accessed, all that remained to complete the synthesis was deprotection and azide reduction. Exposure of 28 to aqueous sodium hydroxide led to hydrolysis of the ureas, providing the pyrrolecarboxamide 29. Initial attempts were made to complete the sequence by reduction of the azides with NaBH4, which proceeded uneventfully; however removal of the DMAS-protecting groups at this stage was unsuccessful. Ultimately it was determined that deprotection could be accomplished with azide 29 upon treatment with methanolic HCl providing 30.25 Exposure of the bis azide 30 to hydrogen in the presence of the Lindlar catalyst resulted in reduction to the amines and provided nagelamide D (6),16 which was converted to the TFA salt by treatment with TFA.
Somewhat to our surprise, we found on comparison of the spectroscopic data for the synthetic and naturally-occurring material that they were not completely identical. In particular in the 1H NMR spectra, there are several notable discrepancies in the signals due to the aliphatic protons, and to a lesser extent in the 13C NMR spectrum for the corresponding carbon atoms. Notably, two of the reported shifts δH 2.14/2.72 (nat.) vs 1.68 (syn.) for H9′ and δH 3.20-3.30 (nat.) vs 2.40 for H10′ (syn.) are substantially different, whereas our data are very similar to that obtained for dihydrooroidin. A full comparison of the spectroscopic data for nagelamide D, natural and synthetic, in addition to the corresponding data for nagelamide A (5), nagelamide J and dihydrooroidin are provided in the Supplementary Information.9, 26
An obvious possible cause of the spectrosopic discrepancy may be a consequence of the concentration differences between the samples used, as Kobayashi and coworkers report isolation of ~ 2 mg of nagelamide D (6). However, we found essentially no differences in the 1H NMR spectra on serial dilutions down to approximately the same estimated concentration. Addition of a couple of drops of TFA to the final dilution sample exhibited no substantial differences in the 1H NMR spectrum. Prior to conversion of synthetic 6 to the TFA salt, we checked the 1H NMR spectrum of the free base in MeOH-d4, which exhibited essentially identical chemical shifts to the salt in DMSO-d6. It appears that neither concentration nor pH make a major difference in the appearance of the 1H NMR spectra.
Similarly, there are discrepancies in the 13C NMR spectra between synthetic and natural nagelamide D, in particular a CH2 signal is reported at δc 16.6 (t), whereas for synthetic material the highest field signal is observed at δc 21.1. Based on the same set of comparative molecules, our data appears to be completely consistent with the assigned structure of nagelamide D (see Supporting Information). Unfortunately, we have been unable to obtain a sample or copies of the original NMR data of the genuine natural product from the Kobayashi lab to compare the chromatographic and spectroscopic properties of the synthetic and natural material, and therefore we cannot establish the source of this inconsistency at this time. While our work was in progress, we became aware of a biomimetic total synthesis of nagelamide D by the Horne lab. It is significant that both the 1H and 13C NMR data reported matches very well with that obtained by us, in spite of there being no common intermediates in either of the two routes until the final compound, indicating that the structural identity of the synthetic material is secure.14 Therefore, pending additional information we must conclude that either the assigned structure or the reported NMR data is in error, unraveling this may require re-isolation and characterization of the natural material.
This work was supported by the Robert A. Welch Foundation (Y-1362) and the NIH (GM065503). The NSF (CHE-9601771 and CHE-0234811) is thanked for partial funding of the purchase of the NMR spectrometers used in this study.
Supporting Information Available Detailed experimental procedures and copies of 1H and 13C NMR spectra for all new compounds.