Strategies for the Synthesis of the Halenaquinol and Xestoquinol Families of Natural Products

Abstract

The halenaquinol family of naphthoquinol natural products includes a few closely related polycyclic compounds that feature an activated, electrophilic furan ring. This motif is likely responsible for the rich biological activity attributed to these secondary metabolites. Their interesting structures—related via their electrophilic furan to the biologically important furanosteroids—and their activities prompted significant efforts by organic chemists that resulted in many strategically compelling laboratory syntheses of these targets. These different strategies are compared and contrasted in this Microreview, and the authors’ recent work on the structurally different but biogenetically related natural product exiguaquinol is put into the context of the previous studies on halenaquinol-type targets.

Graphical Abstract

The wealth of different strategies for the chemical synthesis of the halenaquinol/xestoquinol family of natural products is discussed. Similar themes throughout the different achievements are highlighted, and a discussion of the recently discovered relative exiguaquinol is included.

1. Introduction to the Halenaquinol and Xestoquinol Families of Natural Products

The furanosteroid class of fungus-derived natural products includes wortmannin (1, Figure 1) and viridin (2). These bioactive compounds have long been known to be electrophilic natural products by virtue of their doubly activated (acylated) furans, which can undergo addition/elimination reactions with biologically relevant nucleophiles.^1–4 In an interesting example of possible convergent evolution, the halenaquinol/xestoquinol family (see halenaquinol sulfate and xestoquinol sulfate, 3 and 4, respectively) also feature the activated furan incorporated into a structurally complex secondary metabolite; the structural origins of these natural products are thought to be polyketide biosynthesis.⁵ The halenaquinol and xestoquinol family members, which are the primary focus of this review, have demonstrated antibiotic,^5a cardiotonic,⁶ cytotoxic,⁷ antifungal,⁷ and protein tyrosine kinase inhibitory activity,^5b among others. Recently, this famiy has expanded, with the addition of exiguaquinol (5), a probable biosynthetic descendant of halenaquinol, and a possible antibiotic.⁸

Figure 1.

Structures of representative furanosteroid natural products (steroid nucleus highlighted in red in viridin), halenaquinol-type secondary metabolites, and exiguaquinol

Halenaquinone (7) was isolated in 1983 by the Clardy group from the marine sponge Xestospongia exigua (Figure 2).^5a Its chemical structure and relative stereochemistry were determined by X-ray diffraction, but its absolute configuration could not be confirmed at that time. In 1985, the Kitagawa group reported the isolation of two related natural products, halenaquinol (9) and halenaquinol sulfate (4), from Xestospongia sapra and confirmed their absolute stereochemistry by total synthesis.⁹ In the same year, xestoquinone (6), a structurally similar furanosteroid lacking a C3 ketone, was isolated from Xestospongia sapra by the Nakamura group.⁶ Xestoquinol (8) and xestoquinol sulfate (3) were also identified from the same organism several years later.¹⁰

Figure 2.

Additional members of the xestoquinol and halenaquinol families

To date, there have been four syntheses of halenaquinone or halenaquinol (three asymmetric) and four syntheses of xestoquinone or xestoquinol (two asymmetric). In addition, several approaches have been reported to access the core structural features of halenaquinol and xestoquinol. The following sections will highlight the synthetic efforts towards the halenaquinol and xestoquinol classes of natural products and will conclude with some work towards exiguaquinol.

2. Total Syntheses of Halenaquinol, Xestoquinol, and Thiohalenaquinol

2.1. Common Approaches to the Syntheses of Halenaquinol and Xestoquinol

Since 1988, there have been eight reported syntheses of halenaquinol or xestoquinol and one synthesis of a related analogue, thiohalenaquinol. In two instances, the successful strategy to access halenaquinol was applied to xestoquinol with little perturbation to the overall approach. While each group’s approach was unique, there appear to be four main C–C bond disconnections used in the seven distinct strategies (Figure 3). The groups of Harada,^11,12 Kanematsu,¹³ and Rodrigo^14,15 based their syntheses on a similar disconnection in which the naphthalene portion of the molecule is formed by a [4+2] cycloaddition (red). Although their tricyclic dienophiles were different in each case, all three groups utilized a stable o-quinodimethane precursor to annulate the naphthalene moiety.

Figure 3.

Four unique approaches to xestoquinol/halenaquinol

The second most common strategic bond disconnection involved scission of the C–C bond between the C6 quaternary center and the naphthalene group (blue, Figure 3). The Keay¹⁶ and Wipf¹⁷ groups noticed that this bond could be constructed through an intramolecular Heck reaction of an aryl halide and a pendant alkene. Using a chiral ligand, the Keay group was able to perform their transformation enantioselectively, highlighting the power of this tactical disconnection.

The Shibasaki group completed their synthesis of halenaquinol using a late-stage assembly of the furan ring (grey, Figure 3).^18,19 This strategy was unlike previous approaches in that it first formed the CDE-ring tricycle and postponed construction of the A and B rings. This approach resulted in the longest linear synthesis of halenaquinol and perhaps suggests the greater step economy of the alternative routes.

The most recent synthesis of halenaquinol was disclosed by the Trauner group and took advantage of a previously unexplored [4+2] cycloaddition to assemble the naphthalene portion of the molecule (green, Figure 3).²⁰ Their strategy centered on the early functionalization of the furan ring and a late-stage intramolecular Diels–Alder reaction of a vinyl quinone diene. This interesting strategy enabled a concise synthesis of halenaquinol in a high overall yield.

2.2 Harada’s Syntheses of Halenaquinol, Xestoquinol, and Adociaquinones A & B

The first total syntheses of (+)-halenaquinone (7) and (+)-halenaquinol (9) were performed by the Harada group in 1988 (Scheme 1).¹¹ Using their π-electron SCF-CI-dipole velocity MO method to calculate the CD spectrum,²¹ they were able to predict the natural configuration of the C6 methyl quaternary stereogenic center to be (S), and then complete the first enantioselective syntheses of halenaquinone and halenaquinol (Scheme 1).²² This achievement required 15 steps in the longest linear sequence for halenaquinone and one additional transformation to access halenaquinol. Their approach features a benzocyclobutene ring-opening/o-quinodimethane Diels–Alder cycloaddition to stitch together the molecule.

Scheme 1.

Synthesis of halenaquinol by the Harada group

Starting from (−)-Wieland–Miescher ketone (10), the Harada group accessed enone 11 in nine steps (Scheme 1). Dienophile 11 underwent an o-quinodimethane Diels–Alder cycloaddition with benzocyclobutene 12 upon heating to temperatures above 200 °C to generate 13 in 33% yield. DDQ was used to aromatize 13 to the corresponding naphthalene, which was then treated with O₂ and t-BuOK in tBuOH to provide diosphenol 14 in 90% yield. To forge the furan ring, the Harada group developed a modified Pfitzner–Moffatt oxidation protocol that utilized DCC, DMSO and trifluoroacetic acid.^11,23 After ketal deprotection, the 1,3-diol was oxidized to the β-ketoaldehyde and underwent spontaneous cyclization to reveal furan 15. Finally, pentacycle 15 can be converted to halenaquinone (7) by oxidative deprotection with CAN; subsequent reduction with sodium dithionite affords halenaquinol (9). Owing to the high instability of 9 to light, air, and heat, the final reduction to halenaquinol was performed in the dark. With synthetic samples of enantiopure halenaquinone and halenaquinol, the Harada group was able to verify that the CD spectrum of the synthetic sample matched the spectrum calculated by their SCF-CI-DV MO method, indicating that the enantiomer synthesized was identical to the one isolated by Kitagawa.^11,21

Using a nearly identical route, Harada and coworkers were able to complete a synthesis of xestoquinone (6)¹² and adociaquinones A and B (17–18)²⁴ starting from enone 16 (Scheme 2). Overall, Harada’s syntheses provided halenaquinone in 15 steps and 2% overall yield and xestoquinone in 17 steps and 1% overall yield from 10.

Scheme 2.

Synthesis of xestoquinone and adociaquinones A and B by the Harada group

2.3 Kanematsu’s Synthesis of Xestoquinone

In 1991, the Kanematsu group published the second synthesis of xestoquinone based on a furan ring transfer strategy developed previously in their lab (Scheme 3).^25,13 This reaction rearranges a propargyl furfuryl ether (19) into a dihydroisobenzofuran (22) by a base-mediated isomerization to the allene (20) followed by an intramolecular Diels–Alder/ring-opening cascade (Scheme 3). Other furfuryl ethers, bearing alkyl substituents on the furan or linker, were found to participate in this transformation, affording substituted dihydroisobenzofurans in excellent yields.²⁵

Scheme 3.

Kanematsu’s furan ring transfer method

Starting with furfuryl alcohol (23), a four-step sequence involving furan ring transfer produced 24,²⁶ which can engage methyl acrylate in a conjugate addition followed by regioselective methylation to afford 25 (Scheme 4). The ketone in 25 was reduced via the tosylhydrazone and the methyl ester was saponified and cyclized to 27 through a Friedel–Crafts acylation. Oxidation of the cyclohexanone provided enone 28, which served as the dienophile in the subsequent Cr-mediated o-quinodimethane Diels–Alder reaction between 28 and 29. To complete their formal synthesis of xestoquinone, the Diels–Alder adduct was aromatized to generate 30, which can be converted to xestoquinone (6) by oxidation with CAN.¹²

Scheme 4.

Formal synthesis of (±)-xestoquinone by the Kanematsu group

The Kanematsu synthesis completed xestoquinone in 14 steps from furfuryl alcohol with an overall yield of 0.8%. The key transformations included a furan ring transfer reaction and an o-quinodimethane Diels–Alder cycloaddition originating from the Cr-mediated reductive metallation of an o-bis(bromomethyl)arene.

2.4. Rodrigo’s Syntheses of Halenaquinone and Xestoquinone

In 2001, the Rodrigo group disclosed a short formal synthesis of racemic halenaquinone in which they rapidly assemble the ABC tricycle using an intramolecular Diels–Alder cycloaddition that was previously discovered in their lab (Scheme 5).^14,27,28 Afterwards, the naphthalene portion was appended through the cycloaddition of an electron-rich isobenzofuran.

Scheme 5.

Synthesis of halenaquinone by the Rodrigo group

Diene 32 was prepared from propargyl alcohol in four steps^29–31 and subjected to the Rodrigo group’s optimal conditions for o-benzoquinone monoketalization/IMDA with methylguaiacol 31 and PIFA (Scheme 5). Adducts 33 and 34 were obtained as an inseparable mixture; however, heating this mixture to reflux in 1,2,4-trimethylbenzene converted 33 into 34 by means of a Cope rearrangement, and 34 was obtained in 36% overall yield. Treatment of enone 34 with dimethoxyisobenzofuran 35 afforded pentacycle 36 in high yield. Naphthalene 37 was formed after base-mediated aromatization and acid-mediated elimination of the ketal. To complete their formal synthesis, 37 was aromatized to the furan with p-chloranil and the vinyl sulfide was converted into ketone 15 by heating with TiCl₄ in wet acetic acid. 15 can be transformed into halenaquinone using the previously disclosed methods.¹¹ This sequence provided 7 in 12 steps (longest linear) from commercially available materials and with roughly 3% overall yield.

Prior to their synthesis of halenaquinol, the Rodrigo group completed a formal synthesis of xestoquinone using a similar strategy (Scheme 6).²⁷ In this case, their dienol fragment (38) was lacking the thiophenyl ether group and required a late-stage hydrogenation of the cyclohexene to afford xestoquinone (6). This route efficiently provided 6 in eight steps and 7% overall yield from 31. Following their synthesis, several methoxy-substituted congeners of xestoquinone were synthesized to validate their proposed structures and substitution patterns.¹⁵

Scheme 6.

Rodrigo’s synthesis of xestoquinone

2.5. Keay’s Synthesis of Xestoquinone

The Keay group disclosed an asymmetric formal synthesis of xestoquinone in 1996 that took advantage of a Pd-catalyzed polyene cyclization to assemble the pentacyclic scaffold (Scheme 7).^16,32 This work constituted the third synthesis of xestoquinone and the second asymmetric synthesis. Their convergent strategy relied on joining two key fragments: one containing the furan and another bearing the naphthalene.

Scheme 7.

Synthesis of xestoquinone by the Keay group

The naphthalene segment was prepared in 11 steps from 2,5-dimethoxybenzoic acid (41, Scheme 7). According to the procedure developed by Wallace, 41 can be transformed into 1,3,6-trimethoxybenzocyclobutene (42) in seven steps and 49% yield utilizing a [2+2] cycloaddition of an in situ generated benzyne.³³ With heating, 42 can rearrange to the o-quinodimethane and undergo a Diels–Alder reaction with ethyl 3-bromopropiolate (43) to generate 44.³⁴ With three additional steps, 44 was converted into acid chloride 45.

Trisubstituted furan building block 46 was assembled in a straightforward manner, and it was united with naphthoyl chloride 45 by lithiation of the unsubsituted furan position and nucleophilic addition to afford 47 in 54% yield. Formation of aryl triflate 48 followed by their pivotal asymmetric Heck bicyclization with catalytic Pd₂(dba)₃ and (S)-BINAP produced pentacycle 49 in 82% yield over two steps and up to 68% ee. Finally, 49 was hydrogenated and demethylated to provide xestoquinone (6).¹²

In their synthesis of xestoquinone, the Keay group demonstrated the importance of Pd catalysis in asymmetric total synthesis. Although the syntheses of their two key fragments (45 and 46) were fairly lengthy, the steps following their combination rapidly produced xestoquinone (6). Overall, the Keay group’s enantioselective synthesis afforded xestoquinone in 17 steps (longest linear) and 4% overall yield.

2.6. Wipf’s Synthesis of Thiohalenaquinone

The activated furan is believed to play a significant role in the biological activity of the furanosteroids,^1–4 as well as the xestoquinone/halenaquinone family. However, as a potent electrophile, this group is also likely responsible for its poor target specificity and in vivo toxicity.^17,35 In order to probe the biological significance of the furan present in halenaquinone (7), the Wipf group synthesized thiohalenaquinone (56, Scheme 8), an analogue containing a thiophene ring in place of the typical furan. They reasoned that the reduced strain energy due to the longer C–S bonds in the thiophene and the increased resonance stabilization energy should make thiohalenaquinone less prone to nucleophilic addition compared to halenaquinone, thus making the sulfur variant a more selective inhibitor for its target enzymes.¹⁷ Their strategy initially revolved around functionalization of the thiophene ring. The naphthalene portion was installed via an o-quinodimethane Diels–Alder reaction and the carbon framework was assembled with a Heck reaction followed by a late-stage RCM.

Scheme 8.

Wipf’s synthesis of thiohalenaquinone (56)

Conversion of 3,4-dibromothiophene (50) into highly substituted thiophene 51, bearing a bromo-ynone dienophile, was accomplished through a 12-step protocol. Treatment of 51 with benzocyclobutene 46 at 220 °C afforded aryl bromide 52, which was then subjected to a Pd-catalyzed 6-exo Heck reaction under microwave irradiation to forge the cyclohexanone ring of 53 in 10% yield over four steps. To form the final 6-membered ring, the protected alcohol was converted to the aldehyde and allylated with allyltributyltin and BF₃·OEt₂ to give 54. Ru-catalyzed isomerization to the internal olefin and RCM with Hoveyda–Grubbs catalyst produced pentacycle 55 in 56% yield over the two steps. Interestingly, RCM of the vinylated compound (not shown) proceeded in poor yield; fortunately, their allylation/isomerization sequence was shown to work significantly better. Finally, the synthesis of thiohalenaquinone was completed in three steps by Dess–Martin oxidation, conjugate reduction, and CAN oxidation to unveil the naphthoquinone in 56. In the end, the Wipf synthesis of thiohalenaquinone was accomplished in 22 linear steps and 0.4% overall yield and utilized a convergent Diels–Alder/Heck sequence to assemble the majority of the carbon framework of 56. Biological evaluations of these potentially interesting halenaquinone analogues have yet to be reported.

2.7. Shibasaki’s Synthesis of Halenaquinone

In 1996, Shibasaki and coworkers disclosed the second total synthesis and the first catalytic asymmetric synthesis of halenaquinone (Scheme 9).^18,19 Their strategy revolved around the naphthalene portion of the molecule while the remaining three rings were annulated sequentially. Their key enantiodetermining Suzuki cross-coupling/Heck cyclization was the first example of this type of cascade. Although their Pd-cascade forges the same C–C bond as the Keay and Wipf strategies, the Shibasaki group performed this step much earlier in their synthesis and relied on additional key steps to complete their synthesis.

Scheme 9.

Shibasaki’s synthesis of halenaquinone featuring a cascade of Suzuki–Miyaura coupling and asymmetric Heck cyclization

The synthesis of the tricyclic motif of 60 began with a five-step protocol to access naphthalene 58 from commercially available tetralone 57. Double triflation of 58 followed by the remarkable one-pot Suzuki cross-coupling/asymmetric Heck cascade using Pd(OAc)₂ and (S)-BINAP directly provided 60 in 85% ee, although in low yield. Alternative stepwise pathways afforded 60 with improved yields (not shown).

Tetrahydroanthracene 60 was converted over ten steps to ynone/vinyl iodide 61. Finally, Pd-catalyzed furan formation forged the fused pentacyclic scaffold of halenaquinol (15) after desilylation with TBAF. Naphthoquinol demethylation/oxidation following Harada’s protocol¹¹ yielded halenaquinone (7).

Although the Shibasaki synthesis accessed halenaquinone in the greatest number of steps (18 longest linear steps using their most direct route, 1% overall yield), it constituted the first catalytic asymmetric synthesis of halenaquinone. Key features included their Pd-catalyzed Suzuki cross-coupling/Heck reaction cascade and the metal-mediated furan formation.

2.8. Trauner’s Synthesis of Halenaquinone

To showcase their methodology involving the inverse-electron-demand Diels–Alder reaction of vinyl quinones,^36,37 the Trauner group developed a concise asymmetric total synthesis of (−)-halenaquinone (Scheme 10).²⁰ First, 3,4-diiodofuran (62) was converted in a five step sequence to enantioenriched Heck precursor 63. The 6-exo Heck reaction using Pd(OAc)₂ and TBAB closed the cyclohexane ring of 64 with good diastereoselectivity (dr = 7:1). Nucleophilic addition of the stannate of 65 to aldehyde 64 produced an inconsequential mixture of carbinol diastereomers (66) in high yield. Next, deprotection and Ley–Griffith oxidation of both secondary alcohols yielded a chalcone-type intermediate, which was oxidized to vinyl quinone 67 with AgO and HNO₃. The desired inverse-electron-demand Diels–Alder cycloaddition was performed under high pressure, and the immediate product was aromatized with MnO₂ to afford ent-halenaquinone (ent-7).²⁰

Scheme 10.

Trauner’s synthesis of halenaquinone featuring an intramolecular vinylquinone Diels–Alder cycloaddition

The Trauner group accomplished their asymmetric synthesis of the unnatural enantiomer of halenaquinone in 12 linear steps and 8% overall yield from diiodofuran 62. The route features a diastereoselective intramolecular Heck cyclization and a Diels–Alder cycloaddition of a vinyl-substituted benzoquinone to assemble the furanosteroid pentacyclic scaffold.

3. Syntheses of Halenaquinol and Xestoquinol Analogues and Model Systems

A number of studies toward the xestoquinol/halenaquinol natural products culminated in the syntheses of the complex tetracyclic core structure of these targets. In most cases, the o-quinodimethane Diels–Alder cycloaddition plays a key role, as it did in some of the total syntheses described above.

3.1. Crews’s Syntheses of Halenaquinone and Xestoquinone Analogues

During their studies on the protein tyrosine kinase activity associated with various furanosteroids and naphthalene-derived structures, the Crews group synthesized several complex halenaquinone and xestoquinone analogues.³⁸ The structures most relevant to halenaquinone and xestoquinone include those in which the furan is replaced with a benzene ring (73–74, Scheme 11) or those lacking the fused aromatic ring altogether (75). While no yields are reported throughout their syntheses, enough of each analogue was isolated for biological evaluation. Their strategy appears to take inspiration from the work of both the Harada and the Kanematsu groups.^11,13 Readily available tricyclic intermediate 68 was oxidized with PCC to obtain diketone 69. Heating either 68 or 69 in the presence of sulfone 70 effected a cheletropic extrusion/o-quinodimethane Diels–Alder cycloaddition with the enones to provide pentacyclic compounds 71 and 72. Finally, aromatization to the naphthalene was accomplished with DDQ and oxidative demethylation with CAN led to xestoquinone analogue 73 and halenaquinone analogue 74, each containing a fused benzene ring in place of the furan.

Scheme 11.

Halenaquinone and xestoquinone analogues made by the Crews group

The synthesis of the tetracyclic analogue 75, lacking the furan ring altogether, started from Wieland–Miescher ketone (synthesis not shown). Protein tyrosine kinase inhibition assays indicated that each analogue was less active (IC₅₀ = 27 μM for 73, 9 μM for 74, and 10 μM for 75) compared to halenaquinone (7) and halenaquinol (9) (IC₅₀ = 1.5 and 0.6 μM, respectively).³⁸ However, the fact that any activity remained likely points to a non-covalent mechanism of action in these cases

3.2. Nemoto’s Synthesis of the Halenaquinone Core

In 2001, the Nemoto group described their strategy for accessing the tetracyclic core of halenaquinol (79, Scheme 12).^39,40 They planned to access a fused tetracycle resembling halenaquinol through an intramolecular o-quinodimethane Diels–Alder cycloaddition. Benzocyclobutene 76 was heated to reflux in o-dichlorobenzene to unveil the o-quinodimethane, which underwent facile intramolecular cycloaddition with the furan to afford endo product 77 exclusively. Formation of an intermediate phenylselenide by alkoxyselenenylation, followed by oxidation and elimination provided dihydrofuran 78. Finally, treatment with acid generated furan 79 with loss of methanol.

Scheme 12.

Nemoto’s IMDA approach to the halenaquinone tetracycle

The halenaquinone core synthesis by the Nemoto group provided tetracycle 79 in nine steps and 8% overall yield from 3-furoyl chloride. Although their target still lacks a ketone and an additional aryl ring, it remains a unique and concise strategy in the halenaquinone synthesis literature.

3.3. Ahn’s Synthesis of the Xestoquinone Core

In 2003, the Ahn group published a racemic synthesis of the tetracyclic core of xestoquinone (85, Scheme 13).⁴¹ Their approach focused on the early formation of the quaternary center and postponed construction of the fused ring scaffold until a later stage. Using a [4+2] cycloaddition of a propargyl aldehyde, the 6,5-ring system was assembled in a single step, although with poor conversion.

Scheme 13.

Ahn’s unusual synthesis of tetracyclic furanyl ketone 85

Reactive ynal/ynoate 81 was obtained by oxidation of 80 with Dess–Martin periodinane. Upon heating in toluene, 81 underwent the intended proposed cascade, involving a tetradehydro-hetero-Diels–Alder reaction followed by a ring rearrangement and hydride shift, and provided 84.⁴¹ Although this reaction proceeded in only 5% yield, enough of 84 was obtained to evaluate the final step. Methyl ester 84 was treated with BBr₃ to effect a one-pot demethylation/Friedel–Crafts acylation to afford the tetracyclic core of xestoquinone (85) in 42% yield.

The Ahn synthesis provided the tetracyclic core of xestoquinone (85) in 15 linear steps and 0.3% overall yield. While they demonstrated that their thermal cascade can yield furan 84, significant optimization of this step is needed increase the efficiency of the Ahn group’s overall approach.

4. Exiguaquinol. Proposed Origin from Halenaquinol Sulfate and Synthetic Studies

4.1. Likely Biogenesis of Exiguaquinol

Exiguaquinol (5) was isolated by the Quinn group from the methanolic extracts of the Australian sponge Neopetrosia exigua.⁸ It was identified in a high-throughput screen for H. pylori MurI inhibition (IC₅₀ = 361 μM) and postulated to bind the same cryptic allosteric site as AstraZeneca’s pyrazolopyrimidinedione inhibitors.⁴² In their report, the Quinn group proposed a biogenesis of exiguaquinol (5) starting from halenaquinol sulfate (4) (Scheme 14). First, an oxidative ring opening of the furan group in 4 is proposed, generating tetracarbonyl intermediate 87. A molecule of taurine (88) can then condense with the aldehyde and cyclize onto the nearby ketone, forming hemiaminal 90. Finally, the tertiary alcohol can undergo a semipinacol rearrangement to afford the fused 5-6-5 ring system present in 5.⁸ This plausible biogenetic connection converts a largely planar polycyclic natural product to one with a dense array of four stereogenic centers and more significant challenges for synthesis.

Scheme 14.

Plausible biogenesis of exiguaquinol from halenaquinol sulfate as proposed by Quinn and co-workers

Besides exiguaquinol, other halenaquinol-derived natural products have been isolated that contain the addition of a taurine moiety (Figure 4). In 1988, a regioisomeric mixture of hypotaurine adducts of both xestoquinone and halenaquinone, named adociaquinone A–B (17–18) and ketoadociaquinone A–B (91–92) respectively, were isolated from a South Pacific marine sponge, Adocia sp.⁴³ Semi-syntheses from xestoquinone (6) and halenaquinone (7) were performed in order to confirm their structural assignment, while also supporting the likely biosyntheses of adociaquinones and ketoadociaquinones from their respective natural product precursors.^24,43

Figure 4.

Other taurine-containing members of the xestoquinone/halenaquinone family

According to reports by Harada and Kitagawa, the unprotected dihydroxynaphthalene groups of xestoquinol (8) and halenaquinol (9) are unstable to air, moisture, and light, and they readily oxidize back to the corresponding naphthoquinones (Figure 2).^9,11 Therefore, we hypothesized that the sulfated alcohols in xestoquinol sulfate (3), halenaquinol sulfate (4), and exiguaquinol (5) could serve as protecting groups for the hydroquinone forms of the natural products. With the proper sulfatase enzyme, 3, 4, and 5 would likely be converted to the unprotected dihydroxynaphthalene analogue, which will spontaneously oxidize to the naphthoquinone under aerobic conditions; this form might contribute to the observed bioactivity of this family of natural products.

4.2. A Synthesis of the Tetracyclic Core of Exiguaquinol and Exiguaquinol Dessulfate by the Vanderwal Group

Besides the incorporation of exogenous taurine, exiguaquinol (5) contains several features that distinguish it from halenaquinol sulfate (4, Figure 1). Most notably, exiguaquinol contains a rearranged pentacyclic carbon skeleton that lacks the furan moiety present in the furanosteroid framework.

In 2014, the Vanderwal group disclosed a strategy to access the tetracyclic framework of exiguaquinol (93, Figure 5),⁴⁴ which was subsequently followed up by a synthesis of exiguaquinol dessulfate.⁴⁵ This work relied on an aldol reaction to append the aryl fragment and a late-stage reductive Heck reaction to forge the C-ring cyclopentanone. This latter step is inspired by the work of the Keay and Wipf groups and can be classified in the same category of C–C bond disconnections.

Figure 5.

Exiguaquinol and the tetracyclic model system first targeted by Vanderwal and co-workers.

Through a synthesis of the C-2 hemiaminal epimer of 93, the Vanderwal group demonstrated the feasibility of their C–C bond forming strategy; however, the C2 hemiaminal did not equilibrate as expected to the natural configuration as found in exiguaquinol.⁴⁴ Ground state computational analysis of 93 and its unnatural epimer, as performed by the Houk group, suggested the presence of favorable internal hydrogen bonding in the observed (undesired) tetracycle epimer. Conversely, in the ground state analysis of fully-functionalized exiguaquinol, there exists a thermodynamic preference for the natural hemiaminal epimer (5), owing to a different arrangement of intramolecular hydrogen bonding, in this case involving the sulfonate.⁴⁴

With this computational support, the Vanderwal group set out to apply their strategy to a synthesis of 5.⁴⁵ Diene 95 and maleimide (96) were converted via a Diels–Alder reaction, subsequent N-alkylation with TBS-protected 2-bromoethanol, and alkene hydrogenation to generate meso-imide 97. BF₃·OEt₂-assisted aldolization of the lithium enolate of 97 with naphthaldehyde 94 (itself made by an interesting sequence from 3,4-dibromothiophene) gave adduct 98 in high yield with complete diastereoselectivity. Alcohol protection, oxidation to the bis(sulfoxide), and thermal elimination delivered diene 99. Regioselective reduction of the imide provided the unnatural C2 hemiaminal epimer, which was carried through the reductive Heck cyclization to give 100. Subsequent deprotection and oxidation yielded the naphthyl ketone, and the exocyclic alkene was oxidatively cleaved to afford 101. Sulfonate 102 was then accessed via a high-yielding Mitsunobu reaction followed by direct thioester oxidation with m-CPBA. With the alkyl sulfonate installed, the stereochemical inversion of the hemiaminal was carried out in two steps to form the naturally-occurring (R)-epimer (103), which could be demethylated using CAN and sodium dithionite to afford exiguaquinol dessulfate (104). Unfortunately, 104 was unstable and readily decomposed under ambient conditions, thereby preventing the regioselective sulfation to exiguaquinol (5) under the myriad conditions attempted.

While exiguaquinol dessulfate has not been isolated as a natural product, it is possible that 104 or its quinone form might be secondary metabolites that remain undiscovered. 104 bears clear familial resemblance to halenaquinol (9) and xestoquinol (8), and is plagued by the same instability issues reported for these two known natural products. Through synthesis, exiguaquinol dessulfate (104) was accessed by the Vanderwal group in a longest linear sequence of 19 steps and 1% overall yield.

5. Conclusions and Outlook

The synthetic work discussed in this review takes advantage of several key disconnections to access the natural products halenaquinol and xestoquinol, as well as several of their derivatives and analogues. Recurring approaches include Diels–Alder cycloadditions of an in situ generated o-quinodimethanes to assemble the naphthalene ring structures, highlighting the power of this transformation; and intramolecular Heck reactions, which illustrated the utility of palladium catalysis in accessing the natural product scaffolds in both asymmetric and diastereoselective fashions. While it remains unclear what future directions await in terms of biological applications of these natural products, it is clear that synthesis should be able to address the needs of both material supply and analogue production, when needed. The exception, of course, is exiguaquinol, which has yet to succumb to a complete synthesis.

Scheme 15.

The Vanderwal group’s synthesis of exiguaquinol dessulfate using desymmetrization of meso-imide 97 as a key strategic ideal

Biographies

Gregg Schwarzwalder grew up in Glen Rock, New Jersey, USA. He received his B.A. degree in Chemistry from Cornell University where he worked in the group of Professor Jon T. Njardarson. In 2010, he moved to the University of California, Irvine to pursue his doctoral studies under the direction of Professor Christopher D. Vanderwal. As an NIH Predoctoral Fellow, his research focused on the synthesis of bioactive natural products. After obtaining his Ph.D. in 2015, Gregg moved to the California Institute of Technology, where he is currently an NIH Postdoctoral Fellow in the laboratory of Professor Gregory C. Fu.

Christopher Vanderwal received a B.Sc. degree in Biochemistry (1995) and an M.Sc. degree in Chemistry (1998) from the University of Ottawa. He then moved to the Scripps Research Institute for doctoral studies in the group of Professor Erik Sorensen. After obtaining his Ph.D. in 2003, Chris joined the group of Professor Eric Jacobsen at Harvard University as a Jane Coffin Childs postdoctoral associate. In 2005, Chris began his independent academic career at the University of California, Irvine, where he is currently Professor of Chemistry.