Bioinspired Total Synthesis of Erectones A and B, and the Revised Structure of Hyperelodione D.

The field of biomimetic synthesis seeks to apply biosynthetic hypotheses to the efficient construction of complex natural products. This approach can also guide the revision of incorrectly assigned structures. Herein, we describe the evolution of a concise total synthesis and structural reassignment of hyperelodione D, a tetracyclic meroterpenoid derived from a Hypericum plant, alongside some biogenetically related natural products, erectones A and B. The key step in the synthesis of hyperelodione D forms six stereocentres and three rings in a bioinspired cascade reaction that features an intermolecular Diels-Alder reaction, an intramolecular Prins reaction and a terminating cycloetherification.

Polycyclic polyprenylated acylphloroglucinols (PPAPs) are a vast family of plant derived meroterpenoid natural products with a broad spectrum of biological activity. Although most PPAPs contain a bicyclo[3.3.1]nonane ring system (e.g. hyperforin), a growing number of “non‐canonical” PPAPs with diverse and often unique polycyclic structures have recently been isolated. For example, in 2021 the structural elucidation of a complex di‐geranylated tetracycle, hyperelodione D (1), from Hypericum elodeoides Choisy was reported (Figure 1). The proposed structure of hyperelodione D has the same scaffold as two PPAP meroterpenoids previously found in Hypericum erectum, erectones A and B (2 and 3), which possess either prenyl or geranyl side chains at C6a and C8.

Figure 1

Non‐canonical PPAP natural products of interest in this work.

As suggested in the original isolation work,[ , ] the complex tetracyclic structures of 1, 2 and 3 could arise from a Diels–Alder based cascade reaction between E‐β‐ocimene and a suitable dihydroxyquinone dienophile such as 4 (Scheme 1a). Cascade reactions founded on the Diels–Alder cycloaddition are some of the most powerful methods in organic chemistry for the rapid generation of molecular complexity. Furthermore, two of the most spectacular cascade reactions applied in total synthesis employ E‐β‐ocimene, a naturally abundant monoterpene which serves as a reactive diene in intermolecular Diels–Alder reactions (Scheme 1b). In a 2011 full paper based on an earlier communication, Majetich and co‐workers showed that hydroxyquinone 5 and E‐β‐ocimene underwent a Lewis acid catalysed cascade initiated by an intermolecular Diels–Alder reaction to give perovskone (6). In 2013, Liu and co‐workers reported the bioinspired synthesis of bolivianine (8) from unsaturated aldehyde 7 and E‐β‐ocimene via consecutive intermolecular and intramolecular Diels–Alder reactions. Inspired by these classic total syntheses, we initially targeted a concise synthesis of the proposed structure of hyperelodione D (1) using a Diels–Alder cascade reaction between E‐β‐ocimene and the relatively simple geranylated dihydroxyquinone 4. This bioinspired strategy would employ a complex cascade reaction of E‐β‐ocimene after just a few steps, thus maximizing its impact.

Scheme 1

a) Proposed bioinspired synthesis of hyperelodione D. b) Some previous Diels–Alder cascades Involving E‐β‐ocimene.

A detailed biosynthetic proposal for hyperelodione D (1) is outlined in Scheme 2. As an example of a non‐canonical PPAP, 1 could derive from di‐geranylation of an acylphloroglucinol such as 9, a common polyketide biosynthetic intermediate in Hypericum plants, to give 10 (Scheme 2). Indeed, hyperelodiones E and F (which were co‐isolated with 1) both contain isobutyryl groups that could also be derived from 9. A Dakin oxidation of 10 with concomitant loss of isobutyric acid, followed by aerobic oxidation of the resultant hydroquinone, would give the dearomatized dihydroxyquinone 4. Finally, an intermolecular Diels–Alder reaction between 4 and E‐β‐ocimene would form the endo adduct 11, which is primed for an intramolecular Prins reaction and cyclotherification to give 1.

Scheme 2

Proposed biosynthesis of hyperelodione D.

The biosynthesis of the proposed structure of hyperelodione D (1) served as the blueprint for its concise total synthesis, alongside a di‐prenylated analogue 15 (Scheme 3). Commercially available 2,4,6‐trihydroxybenzaldehyde was di‐prenylated or di‐geranylated under aqueous conditions to give 12 and 13. Dakin oxidation of aldehydes 12 and 13 under acidic conditions then afforded dihydroxyquinones 14 and 4 in good yield via FeCl3‐mediated oxidation of the intermediate hydroquinones. The key Diels–Alder cascades between 14 or 4 and E‐β‐ocimene were achieved using Eu(fod)3 as a mild Lewis acid catalyst in C6H6 at 85 °C to give 15 (56 % yield) and 1 (58 % yield). The Diels–Alder cascade reactions could also be achieved less efficiently in 10 % aq. HCl at 50 °C or in PhMe at 150 °C. These highly predisposed cascade reactions generate five stereocentres, four skeletal bonds and three rings in a single step. The initial Diels–Alder reaction of the cascade is both regioselective and diastereoselective, with the less hindered methylene group of E‐β‐ocimene attacking the more hindered prenyl/geranyl‐bearing carbon atom of 14 or 4 via an endo transition state. Conducting the reaction between quinones 4 or 14 and E‐β‐ocimene “on‐water” at neutral pH and at 50 °C gave the endo Diels–Alder adducts 11 and 16 as single diastereomers, which were converted into 1 and 15 using catalytic Eu(fod)3 in C6H6. The bowl‐shaped, tetracyclic structure of 15 was confirmed by single crystal X‐ray crystallography. However, at this point we were disappointed to observe that NMR data for synthetic 1 did not quite match the data reported for natural hyperelodione D.

Scheme 3

Total synthesis of the proposed structure of hyperelodione D.

Although 1H and 13C NMR spectra for synthetic 1 are similar to those reported for natural hyperelodione D, there are slight differences in the chemical shifts at C2, C2a, C29 and C30. In addition, the HMBC spectrum of hyperelodione D shows a correlation between C2 and only one methyl substituent. Furthermore, 2D NMR spectra of natural hyperelodione D indicate a prenyl side chain at C8, rather than a geranyl group. We therefore proposed that 17 might be the true structure of hyperelodione D (Scheme 4a). The revised structure 17 is biosynthetically plausible, arising from an intermolecular Diels–Alder reaction between erectquione A (18) and E,E‐α‐farnesene to give 19, followed by Prins cyclization and cycloetherification. Erectquinone A, which is the proposed biosynthetic precursor of erectones A and B (2 and 3), was also isolated from Hypericum erectum, and E,E‐α‐farnesene is found in the essential oils of several Hypericum species. Next, we validated E,E,‐α‐farnesene (which was synthesized as a single stereoisomer according to a known procedure ) as an effective participant in the Lewis acid catalysed Diels–Alder cascade with quinone 14 to give 20 in 43 % yield (Scheme 4b). NMR spectra of 20 showed a much closer match to the data reported for natural hyperelodione D than our synthetic 1, which gave confidence in the suggested reassignment to 17. The relative configuration of the sixth stereocentre formed in the cascade at C2 of 20 was assigned by NOE correlations.

Scheme 4

a) Structural reassignment and revised biosynthesis of hyperelodione D. b) Validation of E,E‐α‐farnesene as a suitable diene in the Diels–Alder cascade.

To confirm the proposed structure revision of hyperelodione D through synthesis, we needed access to erectquione A (18) as the quinone dienophile for a bioinspired Diels–Alder cascade with E,E‐α‐farnesene. In addition, reaction between 18 and E‐β‐ocimene should form the constitutionally isomeric erectones A and B (2 and 3). Synthesis of 18 was achieved via a stepwise sequence of alkylations that began with a challenging mono‐geranylation of 2,4,6‐trihydroxybenzaldehyde in K2CO3/acetone to give 21, in low yield due to competing O‐alkylation and di‐alkylations (Scheme 5). Mono‐prenylation of 21 under aqueous conditions then gave 22 in 60 % yield, and finally Dakin oxidation of 22 gave erectquione A (18) in 70 % yield.

Scheme 5

Total synthesis of erectquinone A.

With 18 in hand, we investigated its Diels–Alder reactions with E‐β‐ocimene and E,E‐α‐farnesene (Scheme 6). First, an on‐water catalysed, endo Diels–Alder reaction between 18 and E‐β‐ocimene gave 23 and 24 in 70 % yield as an inseparable 1 : 1 mixture of regioisomers. We propose that these Diels–Alder adducts could be undiscovered natural products in Hypericum erectum. Treating this mixture of 23 and 24 with catalytic Eu(fod)3 in C6H6 at 85 °C then formed a 1 : 1 mixture of erectones A and B (2 and 3) in 64 % combined yield via Prins cyclization and cycloetherification. Alternatively, a one‐pot cascade reaction, also catalysed by Eu(fod)3, allowed the direct formation of 2 and 3 in 55 % combined yield, which were separable by flash chromatography on silica gel. Similarly, an on‐water, intermolecular Diels–Alder reaction between 18 and E,E‐α‐farnesene gave regioisomers 19 and 25, which potentially exist as natural products in Hypericum elodeoides Choisy. Treating 19 and 25 with catalytic Eu(fod)3 in C6H6 at 85 °C then gave a separable 1 : 1 mixture of tetracycles 17 and 26, which could also be formed directly from 18 and E,E‐α‐farnesene under identical conditions. Pleasingly, NMR data for 17 fully matched that for natural hyperelodione D, thus proving our structural reassignment of this complex meroterpenoid. The prenyl/geranyl substitution pattern of 17 and 26 was established by HMBC correlations and comparison to NMR data for erectones A and B. Alongside the Diels–Alder adducts 19, 23, 24 and 25, it is also probable that tetracycle 26 occurs in nature, and our characterization of these compounds could aid their future discovery in Hypericum plants. Finally, the Diels–Alder cascade between erectquione A and E‐β‐ocimene was also conducted in 10 % aq. HCl at 50 °C to give erectones A and B in 40 % yield, alongside the Diels–Alder adducts. (The corresponding aqueous Diels–Alder cascade with E,E‐α‐farnesene gave an inseparable mixture of products.)

Scheme 6

Bioinspired divergent total synthesis of erectones A and B, and the revised structure of hyperelodione D.

In summary, we used biosynthetic speculation to guide the total synthesis and structure revision of hyperelodione D via a series of cascade reactions of gradually increasing complexity. The final Diels–Alder cascade between erectquione A and E,E‐α‐farnesene to give hyperelodione D constructs six stereocentres and three rings in a single step, thus showcasing the power of biomimetic synthesis. Furthermore, the divergent nature of this strategy was exemplified by the synthesis of erectones A and B and five possible natural products, alongside hyperelodione D, from a common intermediate.

Conflict of interest

The authors declare no conflict of interest.

As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re‐organized for online delivery, but are not copy‐edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.

Supporting Information

Click here for additional data file.

Supporting Information

Click here for additional data file.