Relay Cross Metathesis for the Iterative Construction of Terpenoids and Synthesis of a Diterpene-Benzoate Macrolide of Biogenetic Relevance to the Bromophycolides.

We report a relay cross metathesis (ReXM) reaction for the construction of terpenoids in an iterative protocol. The protocol features the cross metathesis of a relay-actuated Δ6,7-functionalized C10-monoterpenoid alcohol with C10-monoterpenoid citral to form a C15-sesquiterpene. Subsequent functional group manipulation allows for the method to be repeated in an iterative fashion. The method is used for the synthesis of a diterpene-benzoate macrolide of biogenetic relevance to the bromophycolide family of natural products.

Terpenoids, consisting of “head-to-tail” and “head-to-head” arrangements of five-carbon isoprene units, are a diverse and very large class of linear and (poly)cyclic naturally occurring biomolecules with more than 40,000 distinct chemical structures, thereby accounting for approximately 60% of known natural products.[1] They mediate vital biological functions including light harvesting and photo-oxidative protection, lipid membrane modulation, electron transport, intercellular signaling as hormones, and interspecies defense among others.[2] Traditional herbal remedies from plants have utilized the medicinal benefits of terpenoids for centuries,[1] with the subsequent development of terpenoid derivatives (e.g., steroidal medicines) as blockbuster drugs in the 20th century through to the present day.[3] While a comprehensive account of their biogenesis is beyond the scope of this document, it is important to note that (poly)cyclic terpenoids all arise from their linear precursors.[4] Nature assembles these linear precursors by enzyme-mediated sequential addition of C5 units of isopentenyl pyrophosphate (IPP) to (C5)-terpenyl pyrophosphates in the mevalonate pathway (Figure a).[5] However, despite the long-term recognition that these linear compounds are essentially C5-repeating isoprene units, a general and iterative chemical protocol for their synthesis—using naturally occurring, terpenoid building blocks—does not exist.[6] We recently reported an olefin cross metathesis reaction between relay-actuated Δ6,7-functionalized monoterpenoid alcohols with trisubstituted alkenes as partner olefins to form new trisubstituted alkenes (Figure b).[7] We now report that the use of readily available and nonexpensive citral as the partner olefin—a monoterpenoid with two electronically distinguishable alkenes—allows for the iterative construction of terpenoids in line with the above aims (Figure c).[8] Furthermore, we report the application of this method for the synthesis of a diterpene-benzoate macrolide of biogenetic relevance to the bromophycolide natural product family.

Figure 1

(a) Terpene biosynthesis via sequential addition of C5 units of isopentenyl pyrophosphate (IPP). (b) Previously reported cross metathesis reaction between relay-actuated Δ6,7-functionalized monoterpenoid alcohols with trisubstituted alkenes to form new trisubstituted alkenes. (c) Relay cross metathesis for the iterative construction of terpenoids.

To commence our investigations enantiopure epoxide 2, as a relay-actuated Δ6,7-functionalized monoterpenoid derivative, was prepared from diol 1(7) using the method of Corey et al. (Scheme ).[9] Using our previously identified conditions (10 mol % ruthenium benzylidene precatalyst 5,[10] alkene [5 equiv], 50 °C, 1 h),[7] attempted relay cross metathesis reaction[11] between epoxide 2 and citral (3)[12,13] to give C15-sesquiterpenoid 4 using 5 was unsuccessful (Table , entry 1). Further attempts with 10 equiv of 3 (entry 2) or at room temperature (entry 3) or with 2 mol % catalyst loading (entry 4) also failed. In these attempts, truncated olefin 6 was observed in the 1H NMR spectra of the crude reaction mixtures, along with a triplet with a characteristic coupling constant of 9.6 Hz that we attributed to 2,3-dihydrofuran,[14] implicating ruthenium hydride-induced isomerization of the expected 2,5-dihydrofuran byproduct.

Scheme 1

Synthesis of Relay-Modified Δ6,7-Functionalized Monoterpenoid 2

Table 1

ReXM of Relay-Actuated Δ6,7-Functionalized Monoterpenoid 2 with Citral (3) Using GII Catalyst (5)a

entry	equiv of 3	mol % 5	T (°C)	additive(s) (mol %)	% yieldb
1	5	10	50		0c
2	10	10	50		0c
3	5	10	RT		0c
4	5	2	50		0c
5	5	10	50	pBQ (20)	0c
6	5	10	50	AcOH (20)	64
7	5	10	70	AcOH (20)	19
8	5	10	RT	AcOH (20)	0c
9	10	10	50	AcOH (20)	30
10	5	2	50	AcOH (20)	4
11	5	10	50	AcOH (20), CuI (15)	68
12	5	20	50	AcOH (20)	80
13	5	20	50	AcOH (20), CuI (30)	84
14	5	20	50	AcOH (40), CuI (30)	88
15	5	10d	50	AcOH (20)	14

Reactions conducted on a 0.25 mmol scale.

Isolated yields of sesquiterpene 4 after chromatography; E/Z ratio determined as ca. 3:1 at the newly formed olefin (Δ6,7) and as ca. 2–3:1 at the α,β-unsaturated aldehyde by 1H NMR and assigned on the basis of characteristic 13C NMR shielded methyl resonances for E-isomers (see the Supporting Information).

Purification not attempted due to complex mixtures of products.

Hoveyda–Grubbs II catalyst employed.

Known hydride scavengers 1,4-benzoquinone (pBQ, entry 5) and AcOH (entry 6) were therefore explored as possible additives for the reaction.[15] Pleasingly, the use of AcOH was beneficial, and C10-monoterpene epoxide 2 now underwent smooth ReXM with C10-monoterpene citral (3) to provide C15-sesquiterpene 4 in good yield (entry 6). The effect of temperature (entries 7 and 8), equivalents of citral (3) (entry 9), and catalyst loading (entry 10) were also explored, with lower yields obtained. Further addition of CuI[16] (entry 11) was found to be beneficial, as was increasing the catalyst loading (20 mol %, entry 12). Increasing quantities of added CuI and AcOH (entries 13–14) resulted in a higher yield, providing a final optimized yield of 88% for this challenging transformation. We note that the use of Grubbs II catalyst (5) is important in this process: the use of the Hoveyda–Grubbs II catalyst[17] (entry 15) under conditions that worked well (cf. entry 6) for catalyst 5 surprisingly gave only a low yield of product from a complex product mixture.

We then turned our attention to demonstrating that the protocol is suitable for iteration (Scheme ). Accordingly, aldehyde 4 was reduced[18] to alcohol 7 and O-allylated[19] to provide C15-relay metathesis substrate 8. A second ReXM with citral (3), using the optimized conditions developed above, now produced C20-diterpene 9,[20] which could be readily reduced to the C20-alcohol 10 as the first step of a further iteration.

Scheme 2

Iterative Reduction-Allylation-ReXM

With the ability to synthesize enantiopure, Δ14,15 regioselectively functionalized geranylgeraniol 10, we now targeted diterpene-benzoate macrolide 19 (P = Et), pertinent as a putative biogenetic precursor of the bioactive bromophycolide halogenated natural product family (Scheme ).[21] Accordingly, aryl iodide 13 was produced in a two-step sequence from ethylparaben 11. Alcohol 10 was converted to its bromide 15 and coupled to aryl iodide 13 to give diterpene-benzoate 16. Alternatively, taking advantage of our previous observation[7] that prenylbenzene was unreactive to the ReXM conditions, geranyl benzoate 14 was combined in excess (5 equiv) with relay sesquiterpenoid 8 to also provide diterpene-benzoate 16 in good yield. Subsequent ester hydrolysis gave acid 17 and regioselective epoxide ring-opening with bromide gave bromohydrin 18, which was readily separated away from its minor bromohydrin regioisomer 18a. With the scene now set for macrocyclization, we anticipated that the inseparable E/Z alkene isomers that had built up in the ReXM iteration sequence[22] would become chromatographically distinguishable upon conversion to conformationally constrained rings. Much to our delight, Shiina macrolactonization[23] of seco acid 18 proceeded with high conversion of substrate (91%) and provided (E,E,E)-macrocycle 19 (P = Et) as the major macrocyclic component (29%) which was readily separable from the other more polar Z-olefin containing macrocycles.[24]

Scheme 3

Synthesis of Diterpene-Benzoate Macrolide 19 (P = Et) Pertinent to the Bromophycolide Family of Natural Products (cf. Structure of Bromophycolide A, Boxed, Bottom Left)

MNBA = 2-methyl-6-nitrobenzoic anhydride; DMAP = 4-dimethylaminopyridine.

In conclusion, we have demonstrated the use of a relay cross metathesis reaction between a relay-actuated Δ6,7-functionalized monoterpenoid and citral as readily available, inexpensive and naturally occurring building blocks. This methodology allows the unprecedented construction of terpenoids in a C5 to C5( fashion (from a monoterpene, to a sesquiterpene, to a diterpene) via an iterative ReXM-reduction-relay installation sequence. Although the iterative protocol necessarily gives rise to geometrical mixtures of products because of the current limitations of olefin metathesis catalysts for the formation of geometrically pure trisubstituted olefins, we have used the method to construct an enantiomerically and geometrically pure diterpene-benzoate macrolide of relevance to bioactive substances from marine organisms. The method reported should allow for the synthesis of myriad bespoke terpenes.[25,26]