|Home | About | Journals | Submit | Contact Us | Français|
Detailed in this account are our efforts towards the total synthesis of vinigrol. A highly expedient and convergent synthetic approach made possible by the use of a strategic oxidative dearomatization reaction coupled with a series of ensuing substrate controlled transformations is discussed.
In 1987 researchers at Fujisawa Pharmaceutical company reported the isolation of a novel diterpenoid, they had named vinigrol (Figure 1, ,1),1), with an unprecedented 1,5-butane tethered cis-decalin core.1 Vinigrol has since been shown to be potent as an antihypertensive and platelet-inhibiting agent, as well as having an inhibitory effect on Ca2+ movement.2 Subsequently vinigrol was shown to be a tumour necrosis factor (TNF) antagonist3 with the ability to inhibit the progression of AIDS-related complex to AIDS.4 Giving this promising biological profile and structure, it is not surprising that a number of synthetic research groups took notice.5 Twenty years since its isolation, vinigrol has yet to succumb to total synthesis. Interestingly, the first published route by Hanna remains one of the most advanced progresses towards vinigrol.
In our quest towards completing an expedient total synthesis of vinigrol we have focused our efforts on synthetic strategies utilizing an oxidative dearomatization coupled with an intramolecular Diels-Alder reaction as the key steps. Towards that end we have disclosed our efforts employing the Wessely and Adler-Becker oxidative dearomatization reactions.6 Although these disconnections were shown to be conceptually sound they both suffered from weakeness in either one of the key transformations. This communication details our current route, which overcomes the limitations encountered in the previous routes in addition to being more efficient and easily processable.
Retrosynthetic analysis for our proposed total synthesis of vinigrol (1) is detailed in Scheme 1. Late stage substrate controlled hydrogenation of both olefins followed by formation of an enol triflate, cross coupling of the triflate to form the allylic alcohol and deprotection will advance intermediate 2 to vinigrol. The natural product core will be revealed using a retro-Michael fragmentation of caged structure 3, which in turn can be rapidly accessed via 4 following substrate controlled hydrogenation of the exo olefin and samarium mediated deoxygenation of the mixed ketal moiety. Tandem 6-exo radical cyclizations will deliver the pre-fragmentation polycyclic core (4) from cycloadduct 5. This intermediate will be assembled from the oxidative dearomatization/Diels-Alder union of pyrogallol derivative 6 and acrylic acid 7. In the ideal synthetic scenario a one pot samarium cascade can be envisioned starting from 5 directly to ring expanded 3 via consecutive 6-exo ketyl mediated radical cyclizations followed by double α-keto deoxygenation and a retro Michael fragmentation of the resulting enolate.
Our studies first focused on dearomatizing commercially available symmetrical 1,3-dimethoxy pyrogallol. Unfortunately it was quickly realized that the resulting mono-quinone ketal was very unstable and rearomatized before the desired Diels-Alder cycloaddition could occur. Electronic deactivation with 4-halopyrogallols was not sufficient, thus leading us to investigate the reactivity profile of mono-tosylated derivative 9 (Scheme 2). This strategy was shown to be very effective, as first demonstrated by the hypervalent iodide mediated oxidation of 9 in the presence of allyl alcohol, which afforded cycloadduct 10. Unfortunately more relevant tri-substituted allylic alcohols were not trapped very efficiently in our early attempts. We therefore turned our attention to lead(IV) acetate oxidations. Gratifyingly, these oxidations proceeded smoothly forming mixed quinone ketal 11 and more importantly allowing incorporation of tiglic acid (12). Unlike the dialkyl ketal ortho-quinones these ketals are more stable as a result of the acyl groups inductive effects and need to be heated to access the [2.2.2] cycloadducts.7
Our initial efforts have focused on establishing the viability of the proposed oxidative dearomatization strategy (Scheme 3). A suitable carboxylate sidechain (17) was accessed in three steps. This was accomplished by alkylation of triethyl phosphonoacete (13) with bromide 14, followed by a Horner-Wadsworth-Emmons condensation with aldehyde 168 and hydrolysis to form trisubstituted acrylic acid 17. We were delighted to learn that when phenol 9 was oxidized in the presence of 17 a union of the two took place to form a stable adduct, which upon heating provided desired bicyclic cycloadduct 18 in good yield. Deprotection of the silyl ether and Swern oxidation of the primary alcohol then afforded aldehyde 19. This substrate provided us with several 6-exo/6-exo cyclization strategies by initiating cyclization either at the aldehyde or alkyne terminus. We decided to first explore samarium mediated cyclizations. Upon treatment of 19 with samarium diiodide in the presence of HMPA a rapid cyclization took place forming 20 as a single diastereomer, with the desired correct stereochemistry of the newly formed secondary hydroxyl group. For this substrate (19), the expected second 6-exo radical cyclization did not take place to form the tetracyclic core (21) of vinigrol.9
We had predicted that there might be both steric and conformational challenges with the second radical cyclization. One way to improve the cyclization chances would be to use cyclization partners that were mostly sp2 and sp3 hybridized in order to create more space for the the more challenging second cyclization to occur. We decided to evaluate such a less encumbered system in combination with a more reactive starting vinyl radical. The rapid synthesis and evaluation of this substrate (26) is shown in Scheme 4. Known bromide 2210 was utilized to afford phosphonate 23. Condensation of aldehyde 24 and ester hydrolysis furnished acrylic acid 25, which again was reliably and selectively coupled to pyrogallol derivative 9 using the the same lead(IV) oxidative dearomatization protocol. Upon heating, cycloadduct 26 was obtained in only seven steps from propargylic alcohol 22. Unfortunately all attempted radical cyclizations of this vinyl iodide substrate only afforded mono-cyclized product 27, with no sign of any products arising from the desired 6-exo/6-exo radical cascade.
Having not succeeded in forming the vinigrol core using a radical cyclization cascade we decided to evaluate a stepwise approach, that would build upon our existing route and the lessons learned but employ instead a substituted pyrogallol precursor. We postulated that commercially available pyrogallol 30 would offer us an excellent entry point (Scheme 5). All eight of its core carbon atoms and an oxygen atom would end up in the final structure. Moreover, it is symmetrical and has a deactivating group (acyl group instead of a phenolic tosylate). Towards that end, acrylic acid 29 was accessed using classic chemical transformations from phosphonoacetate 28. The oxidative dearomatization union of acid 29 and phenol 30 afforded bicyclic ketal 31 in very good yield. X-ray crystallography of the vinyl bromide analog (32) confirmed chemical connectivity.11
Radical cyclization of 31 yielded a single diastereomer (Scheme 6), which was shown to be the desired ketone 33 having trapped the radical from the bottom face. Olefination attempts to access RCM precursor 34 proved extremely challenging and were met with no success. We were able to olefinate the less hindered radical cyclization precursor (32). When the resulting tetraene (36) was used for the radical cyclization en route to metathesis precursor 34, triene 37 was instead isolated as the only product. Following cyclization the intermediate allylic radical had migrated to form a tetrasubstituted alkene.
Inspired by the convergent and rapid union of pyrogallol ketone 30 and carboxylate moiety 29 we decided to make two critical modifications. We felt that the oxidative dearomatization union could be further improved and we were eager to suppress the acyl migration often observed during the intramolecular Diels-Alder reaction. Additionally, although the lactone was clearly attractive for eventual oxidation economy in the short term it was complicating our search for the optimal cyclization sequence. By using an alcohol instead of a carboxylate in the oxidative dearomatization step we would access a more stable ketal, which would be helpful for later steps. We also expected that the intermediate ketal would more likely undergo the intramolecular Diels-Alder reaction in situ while not suffering from any competing acyl migrations. The main question we faced was whether we could use the allylic alcohol as a nucleophile, which challenged us earlier (Scheme 2), or alternatively to pre-tether it to the aromatic core. Our efforts are detailed in Scheme 7.
We used the sequence detailed in Scheme 5 to obtain allylic alcohol 38. This was done by reducing the ester obtained from the Horner-Wadsworth-Emmons reaction with Dibal-H. We were delighted to learn that when alcohol 38 was coupled with phenol 30 in the presence of iodobenzene bis(trifluoroacetate) the mixed ortho-quinone mono-ketal was formed, and as predicted underwent an in situ Diels-Alder reaction to cycloadduct 39. As for the previous substrates the radical cyclization proceeded smoothly, exclusively forming exo-ketone 40.12 This hindered ketone was successfully bis-olefinated (41) using the Peterson olefination protocol. Gratifyingly, and despite significant steric hindrance, tetraene 41 could be cyclized using the Grubbs-Hoveyda catalyst (42)13 in the presence of benzoquinone14 to the desired vinigrol pre-fragmentation cage structure 43. This seemingly simple core contains all except three of the carbon atoms needed for the total synthesis.
In summary, we have detailed a new and efficient synthetic strategy for vinigrol. It relies on a strategic oxidative dearomatization/Diels Alder reaction followed by a short sequence of critical substrate controlled transformations. This latest approach constructs vinigrol’s tetracyclic core (43) in only four steps from commercially available ketone 30. Our efforts towards the completion of the total synthesis of vinigrol (1) will be reported in due course.
We would like to thank NIH-NIGMS (RO1 GM086584) and Cornell University for support of this synthetic program. This material is also based upon work supported under a National Science Foundation Graduate Research Fellowship (JGMM). We would like to gratefully thank Emil Lobkovsky (Cornell University) for X-ray structure analysis.