Abstract

Total Synthesis of Illicium Sesquiterpenoids Through an Aliphatic

C–H Functionalization Approach

by

Kevin Hung

Doctor of Philosophy in Chemistry

University of California, Berkeley

Professor Thomas J. Maimone, Chair

The Illicium sesquiterpenoids are a family of natural products exclusively produced by plant species in the genus Illicium. Certain species in the genus are known for their use as spices and fragrance, while others are notorious for their potentially lethal toxicity. In Chapter I, the intricate structural features of the three main subtypes of Illicium sesquiterpenoids, namely the seco-prezizaane, allo-cedrane, and anislactone subtype, along with a brief history of past isolation efforts were reviewed. Next, biological effects such as neurotoxicity and more importantly, neurotrophic activities and their implications were briefly discussed. The neuroprotective effects displayed by select sesquiterpenoids have led to their proposed application as small-molecule drug targets for neuro-degenerative diseases. The apparent interplay between toxicity and neurotrophic activity with oxidative decorations on the carbon skeleton of seco-prezizaane natural products was also noted by us. This was followed by a survey of previous total synthesis endeavors of the seco-prezizaane- and allo-cedrane- type natural products with emphasis on the disparate and ingenious solutions past contributors have devised for the synthesis of their respective cores. Finally, the proposed biosynthetic origin of these sesquiterpenoids was examined with the intention of gaining critical insight for a de novo synthesis of these architecturally ornate natural products.

In Chapter II, the overarching retrosynthetic analysis of our approach to the seco-prezizaane sesquiterpenoids was presented. Whereas past synthetic efforts have relied on embedded oxidation states in the starting materials, our strategy incorporated various site-selective and orthogonal aliphatic C–H functionalization chemistry for the installation of the requisite oxidations on the carbon scaffold. In turn, taking inspiration from the proposed biosynthetic origin, the seco-prezizaane carbon system was envisioned to arise from the cedrane core in a biomimetic manner. Furthermore, we sought a unified strategy to access all members of the seco-prezizaane family of natural products. Our initial plan for the construction of the seco-prezizaane core through the allo-cedrane skeleton was presented. Although this path was fraught with challenging transformations, it nevertheless showed that a skeletal reorganization of the cedrane framework was feasible. Next, we were able to successfully arrive at the seco-prezizaane motif via a modified strategy involving an -ketol rearrangement. From this scaffold, the key C–H functionalization of the C-4 methine position was achieved through a non-heme iron(oxo) catalysis methodology. Subsequent oxidative transformations furnished (–)-14-dehydroxydunnianin, a precursor to the natural product (–)-debenzoyldunnianin. Although the final late-stage C–H functionalization of the C-14 methyl position was met with significant difficulties, we have garnered critical insights into the reactivity of synthetic intermediates we have encountered along the way.

Building upon the lessons we learned in Chapter II, the eventual first synthesis of pseudoanisatinoids such as (+)-pseudoanisatin and (–)-3-deoxypseudoanisatin was recounted in Chapter III. Instead of a late-stage C–H functionalization of the C-14 methyl position, we accomplished the desired transformation via an early-stage radical-mediated methodology. This material was then carried forward under optimized reaction conditions following those we have discovered in the previous Chapter to allow access to pseudoanisatinoid sesquiterpenoids. In the process, we were confronted with functional group compatibility issues for the aforementioned C–H functionalization of the C-4 methine position. This was ultimately overcome with uncovering an optimal catalyst and also reagent conditions for this daunting transformation after substantial experimentation. Hence, we have demonstrated the viability of our overarching retrosynthetic approach in utilizing two aliphatic C–H functionalizations for the synthesis of pseudoanisatinoid natural products.

Moving forward to higher oxidized Illicium sesquiterpenoids in the majucinoid series, we investigated a different strategy in Chapter IV to obviate problematic transformations we experienced in our synthesis of pseudoanisatinoids. Both the C-14 position and the C-4 methine site was oxygenated with a radical-mediated methodology. This procedure eliminated the use of the esoteric designer iron catalyst that was previous applied. Then, two remarkable reactions involving ruthenium and selenium for the installment of the prerequisite oxidation states was accomplished. The majucinoid scaffold was forged from the resulting intermediate through a similar but no-less impressive -ketol rearrangement. A first total synthesis of four sesquiterpenoids, (–)-3,4-dehydroneomajucin, (–)-neomajucin, (–)-majucin, and (–)-jiadifenoxolane A, along with a formal synthesis of (–)-jiadifenolide, (–)-jiadifenin, and (–)-(1R,10S)-2-oxo-3,4-dehydroxyneomajucin was completed in this manner. Thus, our retrosynthetic approach was broadened to include entry to majucinoids in addition to pseudoanisatinoids.

Finally, in Chapter V, we detailed our expansion of the synthetic platform we have designed for other Illicium natural products. A formal synthesis of allo-cedrane sesquiterpenoid (–)-11-O-debenzoyltashironin was realized from an intermediate described in Chapter IV by way of non-heme iron(oxo) C–H functionalization technology. Furthermore, we disclosed our efforts to polyoxygenated natural product (–)-anisatin via a rhodium-catalyzed C–H silylation of the C-13 methyl group, providing an avenue toward the last subset of seco-prezizaane-containing Illicium sesquiterpenoids.

To conclude, we have developed a versatile synthetic platform to allow access to not only all of the seco-prezizaane but also the allo-cedrane families of Illicium sesquiterpenoids through an aliphatic C–H functionalization approach. We hope that our endeavors will stimulate further incorporation of this retrosynthetic paradigm in the broader field of total synthesis. We have also laid the foundation to support future interrogations into the fascinating biological properties of these natural products.

Natural products have long been a rich source of medicinal compounds and inspirations for drug discovery. Among these, molecules that interact with their targets through covalent bonds are highly attractive in biological applications. This dissertation describes our studies on the biosynthetic relationship and chemical synthesis of potentially covalently acting natural products of gymnastatin family, as well as the application of the cysteine-targeting limonoid natural product nimbolide and triterpene derivative bardoxolone for targeted protein degradation.Gymnastatins are a family of cytotoxic natural products derived from tyrosine, which possess a variety of electrophilic functional groups. Through bio-inspired strategies, the total syntheses of gymnastatin G, dankastatin B and dankastatin C was achieved. The previously presumed biomimetic strategy failed to provide the desired stereochemistry during formation of the bicyclo[3.3.1]nonane core of gymnastatin G, complicating our understanding of the biosynthesis of gymnastatins. In addition, the connection between gymnastatin G and a related natural product aranorosin was established through a simple chemical transformation. These results provide further insight into the biosynthesis and chemical reactivity of gymnastatins (Chapter 1). In Chapter 2, we describe the use of nimbolide as a novel E3 ligase recruiter in targeted protein degradation (TPD) applications. While TPD and proteolysis-targeting chimeras (PROTACs) have arisen as powerful therapeutic modalities, the lack of E3 ligase recruiters remains one of the major limitations of current TPD strategies. Extending our recent discovery of nimbolide as a covalent recruiter for the E3 ligase RNF114, we show the broader utility of this E3 ligase recruitment by developing a nimbolide-based PROTAC BT1, which selectively degrades oncogenic fusion protein BCR-ABL over c-ABL in leukemia cancer cells. The selectivity profile of BT1 is opposite to existing cereblon or VHL-recruiting BCR-ABL degraders of similar linker composition. This work further establishes nimbolide as an additional general E3 ligase recruiter for use in PROTACs and demonstrates the importance of expanding the arsenal of E3 ligase recruiters. In Chapter 3, an unexplored mode of E3 ligase recruitment, namely, covalent, reversible E3 ligase recruitment was explored. Efficient proteasome-mediated degradation of BRD4 was achieved by a bifunctional small molecule linking the KEAP1-Nrf2 activator bardoxolone to a BRD4 inhibitor JQ1, which further expanded the TPD toolbox to include bardoxolone as an easily prepared E3 ligase recruiter and further highlighted the power of natural products in such applications.

In the first chapter of this dissertation, a three-step total synthesis of the antimalarial (+)-cardamom peroxide, an endoperoxide-containing monoterpene dimer, is reported. This concise synthesis utilized (–)-myrtenal as an inexpensive terpene building block, and applied a novel oxygen stitching strategy to install all the oxygen atoms in the natural product from molecular oxygen. This strategy has enabled the preparation of large quantities of (+)-cardamom peroxide, and subsequent studies have determined its absolute configuration, and this molecule’s reactivity in iron(II)-induced reductive cleavage. Additionally, the antimalarial activities of cardamom peroxide against the clinical isolates of Plasmodium falciparum from several Cambodia provinces have been evaluated.

In the second chapter, a double allylation strategy for the construction of the [5,7,5]-fused carbocyclic system found in guaianolide sesquiterpenoids is disclosed. This strategy features a robust intramolecular allylation mediated by tin(II) chloride to assemble the seven-membered core. An efficient and scalable total synthesis of the trans-fused 8,12-guaianolide (+)-mikanokryptin via this strategy is reported in nine to ten steps from the abundant monoterpene (+)-carvone, which constitutes the first example of gram-scale total synthesis of any guaianolide natural product.

Subsequently, the oxygen stitching strategy has been expanded from the synthesis of endoperoxides to the strategic installation of multiple oxygen atoms on complex guaianolides in a stereocontrolled manner. A concise total synthesis of (–)-nortrilobolide and a formal synthesis of the anticancer agent (–)-thapsigargin are achieved by merging the oxygen stitching strategy and double allylation strategy.

Preliminary studies on a modular synthesis of guaianolides from the Apiaceae family of plants are also disclosed via an intramolecular allylation strategy. These studies include concise total syntheses of sinodielide A and grilactone from chiral-pool building block linalool, and studies towards the total syntheses of prutenin, ammolactone A, and montanolide. Notably, this strategy has the potential to access the Apiaceae guaianolides encompassing all the oxidation levels in this family.

In this two-part dissertation, several different strategies are developed to attempt concise synthesis of complex natural products. In the first part, a concise total synthesis of garsubellin A, is reported. Starting from commercially available 2-methyl cyclopentenone, this synthesis of garsubellin A was enabled by a series of powerful C-C bond forming reactions, including a tandem oxy-Cope/methylation reaction, a diketene annulation, a ring expansion reaction and a novel regioselective prenyl coupling reaction. Meanwhile, a novel palladium catalyzed Wacker-type oxidation was developed to construct the characteristic tetrahydrofuran ring of garsubellin A. As we hoped this synthetic strategy could serve as a modular synthetic solution of diverse PPAP meroterpenes, an enhanced substrate scope of the diketene reaction was reported.

In the second part of this dissertation, our synthetic studies of acutumine is described. In this chapter, a background introduction is first given, including the isolation and structure determination of acutumine alkaloids, biosynthetic studies, bioactivates of acutumine family, and the prior synthetic studies of (–)-acutumine. Meanwhile, several different strategies for achieving a concise total synthesis of (–)-acutumine were developed and discussed in this dissertation. A novel [3+2] cyclization protocol was developed, which, in three steps, converts commercially available starting materials to a [3.3.0] aza-bicyclic intermediate. From this advanced bicyclic intermediate, a series of efforts to transfer this advanced bicyclic intermediate to (–)-acutumine was explored and reported.

The ophiobolin sesterterpenes are notable plant pathogens which have recently elicited significant chemical and biological attention because of their intriguing carbogenic frameworks, reactive functionalities, and emerging anticancer profiles. In Chapter 1, a 14-step total synthesis of (+)-6-epi-ophiobolin A is disclosed. This work demonstrates a streamlined synthetic platform to complex ophiobolins from (–)-linalool by leveraging the disparate termination modes of a radical polycyclization cascade for divergent elaboration and functionalization.

In Chapter 2, the synthetic studies toward the polycyclic norcembranoid diterpene natural product ineleganolide is described. The synthetic strategy employs a Liebeskind-Srogl coupling to unite (–)-linalool- and (–)-perillaldehyde-derived fragments followed by sequential Ueno-Stork radical cyclization and intramolecular etherification to forge the 5,5,5,6 tetracyclic core of ineleganolide. Subsequent efforts to construct the pentacyclic core of ineleganolide are disclosed. Efforts to advance the late-stage intermediate to the natural product are currently ongoing.

Resiniferatoxin (RTX, 1.1.1, Figure 1.1.1) belongs to the family of daphnane diterpene orthoesters (DDO’s), a structurally and biologically very diverse subfamily of the daphnane diterpene natural products of with over 200 members have been isolated to date

Podophyllotoxin-based glycosidic derivatives have seen numerous uses in cancer chemotherapy. However, these analogs are primarily derived from semisynthetic modifications of the natural product, and the inherent restriction of semisynthesis has prevented further development of new analogs. In the first chapter of this dissertation, a concise and modular total synthesis of the prototypical aryltetralin lignan, podophyllotoxin, is disclosed. Central to the overall strategy is a palladium-catalyzed C-H arylation reaction as the point of diversification. From an advanced intermediate, a two-step sequence furnished not only the natural product but also fully synthetic podophyllotoxin analogs. Moreover, this work uncovered subtle previously overlooked conformational effects governing reductive elimination from high-valent palladium centers.

In the second chapter, a general strategy for the synthesis of complex meroterpene natural products is reported. First, a modular 10-step synthesis of the flagship PPAP, hyperforin, is disclosed. The synthetic approach includes two key transformations 1.) a novel annulation reaction between lithium enolates and diketene, and 2.) an oxidative ring expansion reaction mediated by hypervalent iodine. Second, the substrate scope of the diketene annulation reaction is reported. Finally, a synthesis of berkeleyone A, a complex meroterpene derived from 3,5-dimethylorsellinic acid, is reported in thirteen steps.

Pyrrole-containing alkaloids have long captured the attention of the synthetic chemistrycommunity due to their intriguing and reactive chemical architectures; such is the case of curvulamine and associated polypyrrole natural products. Curvulamine, curindolizine, and the related bipolamines share a unique compact pentacyclic skeleton flanked by two electron-rich pyrroles and up to seven stereocenters. Synthetic access to these complex alkaloids has previously been limited to curvulamine and bipolamine I. In this work, we aim to chronicle the results of our synthetic investigations into this family of unique alkaloids. In the first chapter, we provide an overview into their isolation and antibiotic properties, and discuss our initial synthesis of curvulamine which frames our studies in subsequent chapters. The second chapter describes the total synthesis of curindolizine and the substrate scope of a microwave-assisted method to access pyrroloazepinones, bicyclic heterocycles used in the synthesis of curvulamine and curindolizine. Finally, in the last chapter we describe the synthetic strategies used to access bipolamine C, D, G, and I and describe the unexpected reactivity of some pyrrole-containing intermediates along this journey.

Part I: Altemicidin and related Streptomyces-derived monoterpene alkaloids possess, highly-polar, dense azaindane cores as well as potent cytotoxic and tRNA synthetase inhibitory properties. The congested alpha-amino acid motif decorating their presumed ten-carbon iridoid-like core structure has proven both a synthetic challenge and biosynthetic mystery. This work relates the synthetic process which culminated in a distinct and concise synthesis of altemicidin from simple chemical feedstocks. Key chemical findings include the exploitation of dearomative pyridinium addition and dipolar cycloaddition sequence to stereospecifically install the quaternary amine moiety, and a chemoselective molybdenum-mediated double reduction to establish the fully functionalized azaindane nucleus with minimal redox manipulations.Part II: Synthetic efforts to the tropone-containing cephalotaxane natural product, harringtonolide, are disclosed. The synthetic approach involves construction of the 7-6-5-6 ring system through a Diels-Alder-Barbier coupling sequence followed by ring-closing metathesis. Subsequent advancement efforts via oxidation and stereocenter epimerization are reported.