Towards The Total Synthesis Of Terpenoid Natural Products

Abstract

The construction of diverse molecular architecture conceived and created by Nature, continues to be the most exiting and challenging task to the practitioners of organic synthesis. As a result of refinement in isolation and purification techniques, recent advances in the spectroscopic methods particularly two-dimensional NMR spectroscopy and routine use of single crystal X-ray crystallography, the isolation and structural elucidation of the complex natural products has become a routine exercise. Even those natural products which are present in minute quantity, are being unraveled from the newer and exotic sources such as marine flora and fauna, microbial organisms and insect world. This has been a big boon for the synthetic organic chemists, providing them with increasing number of exciting objectives. The recent advances in the field of natural product synthesis testify to the organic chemists endeavors to meet these emerging challenges.

Nature's expertise and virtuosity in creating a phenomenal array of carbocyclic frameworks is most notably highlighted in the terpenoid group of natural products. Indeed, the number and type of carbocyclic skeleta among terpenes continues to grow unabated as more and more natural products are being routinely isolated from the various sources. Thus, various polycyclic natural products bearing new and novel fused assemblies of five, six, seven and eight membered rings and replete with dense functionalization and stereogenic centers are being regularly encountered. The present investigation represents synthetic efforts towards some novel and recently isolated terpenoid natural products. Two main themes have been pursued. The first involves the construction of a functionalized hydroazulene framework employing RCM as the key step and its further elaboration to the 5,7,6-tricyclic framework present in diterpene guanacastepene-A and 5,11-fused bicyclic system present in neodolabellane diterpenes. The second theme explores the synthetic versatility of the well-established photo-thermal metathetic approach to linear triquinanes through its application to the total synthesis of novel and recently isolated natural product cucumin E. It further explores the utility of 5-5-5 fused ring system to access 5-8 system. This strategy has led to the stereoselective total synthesis of natural product asterisca-3(15),6~diene belonging to the rare asteriscane family.

The present thesis entitled "Towards the Total Synthesis of Terpenoid Natural Products" describes our endeavors towards the synthesis of 5-7-6, 5-11, 5-5-5 and 5-8 fused natural products and has been organized under four chapters. Chapter I. Studies toward the total synthesis of novel diterpene antibiotic guanacastepene A. Chapter EL Synthesis of the novel 5,11-fused bicyclic framework of neodollabellane diterpenoids. Chapter HI. A Stereoselective total synthesis of the novel triquinane sesquiterpene cucumin E. Chapter IV. Total synthesis of 5-8 ring fused sesquiterpene hydrocarbon asterisca-3(15),6-diene.

The Chapter I describes a stereoselective approach towards the construction of the novel 5,7,6-rig fused framework present in the diterpene antibiotic guanacastepene A 1, recently isolated from an unidentified fungus growing on the tree Daphnopsis americana by Clardy et al. Besides its structural novelty, guanacastepene A exhibits impressive activity towards methicilline-resistant Staphylococcus aureus and vancomycine-resistant Entereococcusfaecium. Thus, 1 has evoked an unprecedented attention from the synthetic community and we too were enticed to enter this arena. Scheme 1 (structural formula)

The synthetic approach towards guanacastepene A 1, envisage in this study, was revealed through a retrosynthetic analysis which identified hydroazulenic core 2 (AB rings) with requisite level of functionalities as an advanced precursor to which a six membered ring could be annulated through appropriate protocols. The hydroazulene core 2 was to be accessed from the substituted cyclopentenone 4 through the intermediacy of 3 and the former in turn could be prepared from the readily available endo tricyclo[5.2.1.026]deca-3,8-diene-5-one 6, Scheme 1. In this approach to the AB ring hydroazulenic core 2 of 1, some essential requirements were recognized at the outset. These were the setting up the key cis relationship of the angular methyl group at C11 and the neighboring bulky-isopropyl group at C12, installation of a desirable level of functionalization in the five membered ring and a functional group handle in seven-membered ring to Scheme 2 (structural formula)

append the six membered ring with requisite functionality. Keeping these considerations in mind, readily available endo-tricyclo[5.2.1.02-6]deca-3,8-diene-5-one 6 with well-established propensity toward reactivity on exro-face was identified to be starting point, Scheme 1.

Copper(I)mediated stereoselective 1,4-addition of isopropylmagnesium iodide on 6, followed by sequential a-alkylation with allyl bromide and methyl iodide led to 7 as a single diastereomer and correctly installed the methyl and isopropyl groups in the required cis-relationship, Scheme 2* Retro-Diels-^Ider reaction in 7 under flash vacuum pyrolysis (FVP) liberated the cyclopentenone 8. For the annulation of a seven-membered ring to cyclopentenone 8, recourse was taken to a ring closing metathesis-(RCM) based protocol. Barbier-type addition of 4-bromo-1-butene to 8 in the presence of lithium metal and oxidative transposition of the resulting allylic alcohol with PCC furnished enone 9 in good yield. On exposure to Grubbs' catalyst, enone 9 underwent smooth RCM reaction to deliver the desired hydroazulenic framework 10, Scheme 2.

The bicyclic hydroazulenic enone 10 was now poised for the elaboration of functionalities in the context of evolution to the natural product 1. Thus, 10 was elaborated to epoxy alcohol 11 in a three step sequence, Scheme 2. TMSOTf mediated opening of epoxide ring to yield cis-enediol, protection of the resultant diol as an acetonide and allylic oxidation furnished the key enone 12, Scheme 2. Attempts to alkylate the enone 12 to install the C16 methyl group and the precursor side chain for six membered ring annulation failed consistently. Recourse was then taken to a-carboethoxylation in 12 using Mander's reagent proved to be quite effective and further alkylation with methyl iodide furnished 13 as a single stereoisomer with the correct stereochemical positioning of the quaternary methyl group at C8.

Intermediate 13 was elaborated to tricyclic framework 14 of guanacastepane A in five steps, by setting up NaOEt mediated intramolecular aldol reaction as key step, for the construction of six membered ring, Scheme 2. In tricyclic cross-conjugated dienone 14, complete carbon framework of the natural product guanacastepene A 1, with a copious disposition of functionalities was realized. Further efforts to transform 14 to 1 were not very encouraging. However a variant of ring C annulation on 12 is being investigated by a colleague in the group to achieve the total synthesis of the natural product.

In travails towards 1 and 14, several deviations from the expected course, leading to the new tricyclic structural variants of the biologically promising guanacastepene A 1 were encountered and these findings will also be detailed in this chapter of the thesis.

In the Chapter n of the thesis, synthetic studies directed towards the bicyclic framework present in novel neodolabellane type diterpenes like a-and (3-neodolabellenol 17a and 17b isolated from an unknown species of Australian soft coral by Coll et al will be delineated. The readily accessible bicyclic hydroazulenic enone 13, also served as an advanced intermediate for the construction of the 5-11 fused bicyclic skeleton 16 of neodollabellane diterpenoids via an oxy-Cope rearrangement. Elaboration of 13 to the divinyl carbinol 15 and its [3.3] sigmatropic rearrangement (oxy-Cope rearrangement) to 16 and related reactions will be described, Scheme 3. Scheme 3(Structral formula)

Chapter III describes the first total synthesis of the sesquiterpenoid natural product cucumin E 26 bearing a novel triquinane framework, and reported recently from the mycelial cultures of agaric Macrocystidia cucumis (Pers ex Fr.) by the groups of Steglich and Anke. Synthesis of 26 was accomplished following an interesting variant of the photo-thermal metathetic approach to linear triquinanes delineated by us sometime ago, Scheme 4. Cucumin E 26 attracted our attention as this sesquiterpene [Scheme 4 (Structural formula)] bears an Interesting biogenetic relationship to the related hirsutane group of linear triquinanes from which it can be derived through the migration of a methyl group. Towards the synthesis of 26, the readily available pentacyclic dione 18 was identified as the key starting material and was elaborated to 19 using FVP (flash vacuum pyrolysis) conditions under which 18 underwent [2+2]-cycloreversion of the cyclobutane ring to furnish the cis, syn, cis-triquinane, Scheme 5. On exposure to base, 19 could be equilibrated through back and forth double bond isomerization to furnish the cis,antecis-isorner 20 in reasonable yield. Attention was now turned towards the installation of the network of four methyl groups present in 26 and relevant functional group adjustments. Catalytic hydrogenation of 20, selective mono-Wittig olefination and subsequent gem-dimethylation furnished olefinic ketone 21. At this stage, the carbonyl group in 21 was sought to be removed and recourse was taken to the Barton deoxygenation protocol to furnish 22, Scheme 5. The next task en-route to the cucumin skeleton was the introduction of the angular methyl group at C7 to generate the complete Cis carbon framework. For this purpose, the ketal group in 22 was deprotected to furnish the ketone 23. Angular methylation in 23 exhibited fair regioselectivity to yield 24 as the major product. The enone moiety in 24 was established through allylic oxidation following the Sharpless catalytic selenium dioxide oxidation followed by PDC oxidation to afford 25. Rh(III)-mediated isomerization of the exocyclic double bond in 25 delivered cucumin E 26, whose spectral characteristics were exactly identical to the natural product as established through direct comparison, Scheme 5.

In Chapter IV, the total synthesis of the bicyclo[6.3.0]undecane-based sesquiterpene hydrocarbon asterisca-3(15),6-diene 38, isolated from Lippia integrifolia (Griseb) by Konig et al. and representing the simplest member of the asteriscane family, is described. Our approach to the bicyclo[6.3.0]undecane system was based on the 'carbocyclic ring equivalency' concept. Thus, bicyclo[3.3.0]octane ring system is an eight-membered ring equivalent and tricyclo[6.3.0.02'6]undecane (linear triquinane system) is the latent form of the bicyclo[6.3-0]undecane system through the scission of the central bond as shown in Scheme 6.

Following this concept a synthesis of 38 was envisaged from the cfe,syn, cis-triquinane bis-enone 28, readily and quantitatively available from the pentacyclic-caged dione 27, through flash-vacuum pyrolysis (FVP), as described earlier. More stable bis-enone 29 was obtained from 28 by relocation of one of the enone moieties in 28 through thermal activation under static conditions. The two double bonds in 29 could be now easily differentiated and hence it served as an appropriate substrate for further elaboration. Thus, bis-enone 29 on selective catalytic hydrogenation and regioselective gem-dimethylation afforded 30, Scheme 7. At this stage, the two-carbonyl functionalities in 30 were sought to be removed and this was achieved in a stepwise manner. The sequence involved chemoselective thioketalisation of the enone carbonyl followed by reductive desulfurization in metal-ammonia milieu and led to a diastereomeric mixture of alcohols (resulting from the concurrent reduction of the saturated ketone under metal-ammonia conditions). The diastereomeric mixture of alcohols was deoxygenated following the Barton protocol to yield tricyclic hydrocarbon 31, Scheme 7. Catalytic ruthenium mediated oxidative fragmentation of the tetrasubstituted olefinic bond in 31 afforded the 5,8-fused os-bicyclic dione 32. Wittig olefination of cis-bicyclic dione 32 proceeded regioselectively at the carbonyl group distant from the ring junction and furnished keto-olefin 33. However, the isomerization of exocyclic double bond in 33 to the desired endo position (corresponding to C6-C7 in the natural product) to yield 34 proved to be difficult due to unwanted transannular cyclization. Consequently, the transformation of 33 to the desired 34 was carried out through a five-step sequence. The sequence involved the reduction of the carbonyl group in 33 to yield alcohol, protection of the resultant alcohol as IMS-ether and RhCb mediated isomerization of the exo-double bond to the desired endo position. Further deprotection of the TMS ether and oxidation led to the acquisition of the expected enone 34, Scheme 7. Finally, the exo- methylene unit present in the natural product was installed by Wittig olefination in 34 to furnish 35, corresponding to the 'assigned structure' of the natural product. However the spectral data of synthetic 35 was distinctly different from that reported for the natural product and a revision of the natural product structure was warranted. A careful analysis of the spectral data led us to the surmise that the natural product could be the trans-isomer and we embarked on its synthesis.

Consequently, cis-bicyclic diketone 32 on exposure to base could be readily equilibrated to the more stable trans-isomer 36 in which the later was the major product (1:4). Bicyclic trans-dione 36, like its cis sibling 32 underwent a facile regioselective Wittig olefination to yield keto-olefin 37, Scheme 8. RhCk-mediated double-bond isomerization in 37 proceeded without any complications and gave a readily separable mixture of regiomeric olefinic ketones 38 and 39 in the ratio 2:3, respectively. Wittig olefination on the required keto olefin 39 proceeded smoothly to furnish the bicyclic hydrocarbon 40 whose spectral characteristics [lH NMR, 13C NMR) exactly matched those reported for the natural product, Scheme 8.

A total synthesis of the natural product asterisca-3(15),6-diene has been accomplished. These synthetic efforts necessitate the revision of the earlier assigned structure of the natural product from cis-35 to trans-38. (For structural formula pl see the original document)