Enantioselective total synthesis of (-)-nardoaristolone B via a gold(I)-catalyzed oxidative cyclization.

The first enantioselective total synthesis of (-)-nardoaristolone B is accomplished by the implementation of an enantio- and diastereoselective copper(I)-catalyzed conjugate addition/enolate trapping sequence and a gold(I)-catalyzed oxidative cyclization (intermolecular oxidant), employed for the first time in total synthesis.

Nardoaristolone B (1) was isolated in 2013 from Nardostachys chinensis batal, a plant of the genus Nardostachys endemic of the Himalayan mountains.[1] The synthesis of the racemic mixture has been recently reported.[2] Closely related sesquiterpene (−)-aristolone (2) was isolated much earlier, in 1955, from the roots of Aristochia debilis(3) and has been synthesized in racemic form by various research groups.[4] Nardoaristolone B (1) exhibits protective activity on the injury of neonatal rat cardiomyocytes.

We were intrigued by the possibility of accessing 1 and other members of the aristolone family by combining the highly efficient Cu(I)-catalyzed asymmetric conjugate addition/α-alkylation cascade of α,β-unsaturated cyclic ketones developed by the groups of Alexakis and Cramer[5−8] with the Au(I)-catalyzed oxidative cyclization of enynes recently discovered by Liu (Scheme 1).[9] This last method, based on the gold(I)-catalyzed oxidative functionalization of alkynes pioneered by Toste[12] and Zhang[10,11] could offer direct access to this family of compounds from cyclohexanone 4 as the common intermediate. Here we report the first enantioselective total synthesis of nardoaristolone B (1) and additional studies on the scope of the Au(I)-catalyzed oxidative cyclization of 1,6-enynes.

Scheme 1

Synthetic Plan toward Enantioenriched (−)-Nardoaristolone B and (−)-Aristolone

Although the conjugate methylation of 2-methyl-2-cyclohexenone proceeded satisfactorily at −35 °C,[5b] the subsequent α-alkylation proved to be challenging using methallyl bromide. However, employing methallyl iodide under high concentration and using 1:1 mixture of HMPA/THF before addition of MeLi led to trisubstituted cyclohexanone 3 in 45–55% yield (Scheme 2). Employing the optimal chiral phosphoramidite ligand as reported by Alexakis,[5a] the reaction proceeded with 91–92% enantiomeric excess and 3:1 dr. Careful purification by standard column chromatography allowed us to isolate 3 in essentially pure form (>30:1 dr). The isomerization of exo-olefin 3 into the corresponding trisubstituted endo-alkene 4 was not trivial, and a range of conditions was screened.[13] Fortunately, the use of RhCl3·xH2O (5 mol %) in ethanol at 75 °C led to the desired endo-olefin 4 in 74% yield. The conversion of cyclohexanone 4 into enol triflate 5 was performed under standard conditions (82% yield). Sonogashira cross-coupling of 5 with trimethylsilyl acetylene employing Pd(PPh3)2Cl2 (2 mol %) and CuI (5 mol %) in a DMF/Et3N mixture, followed by methanolysis of the TMS group led to 1,5-enyne 6 in satisfactory yield (74% over two steps) (Scheme 2).

Scheme 2

Synthesis of 1,5-Enyne 6

Pleasingly, 1,5-enyne underwent the desired gold(I)-catalyzed oxidative cyclization using 8-methylquinoline N-oxide (PNO5) and IPrAuNTf2 as catalyst in 1,2-dichloroethane ((CH2Cl)2) at 80 °C,[9] albeit with low isolated yield (20%, along with 25% of the simply cycloisomerized enyne 9). Careful scrutiny of conditions revealed that the choice of the oxidant was crucial in order to favor the desired oxidative cyclization over the cycloisomerization. Thus, 3,5-dichloropyridine N-oxide (PNO3) proved to be superior to all the other N-oxides screened (Table 1, entries 3 and 11).[13,14] Interestingly, using isomeric 2,6-dichloropyridine N-oxide (PNO4) led exclusively to diene 9 (Table 1, entry 4), whereas a complex mixture was obtained under standard cycloisomerization conditions in the absence of any oxidant (Table 1, entry 6). The competitive formation of cycloisomerization product 9, along with 7, in these reactions suggests that both products result from a common cyclopropyl gold(I) intermediate.[15] However, the alternative mechanism involving an earlier oxidation of the terminal alkyne to form an α-oxo gold(I) carbene intermediate that leads to 7 by intramolecular cyclopropanation, in parallel with a simple gold(I)-catalyzed cycloisomerization, cannot be excluded.[9]

Table 1

Screening of Conditions for the Gold(I)-Catalyzed Oxidative Cyclization of 6

entry	[Au]	oxidant	yield of 7/9a,b (%)
1	IPrAuNTf2	PNO1	31/5
2	IPrAuNTf2	PNO2	20/36
3	IPrAuNTf2	PNO3	74/15
4	IPrAuNTf2	PNO4	0/55
5	IPrAuNTf2	PNO5	20/25
6	IPrAuNTf2	none	complex mixture
7	(JohnPhos)AuCl/AgNTf2	PNO3	43/15
8c	tBuXPhosNTf2	PNO3	18/15
9	[(ArO)3P]AuCl/AgNTf2	PNO3	55/2
10	IMesAuNTf2	PNO3	55/2
11	IPrAuNTf2d	PNO3	74 (74)e/15

Yields determined by 1H NMR analysis of the crude mixture using diphenylmethane as internal standard.

Full conversion of starting material was observed unless otherwise stated.

13% unreacted starting material were also visible.

5 mol % of catalyst; Ar = 2,4-(tBu)2C6H3.

Isolated yield.

With the optimal conditions in hand, we performed an oxidative cyclization in the presence of only 5 mol % of IPrAuNTf2 and 3,5-dichloropyridine N-oxide. The desired product 7 was isolated in good yield (74%) along with 15% of cycloisomerized product 9. The last step consisted in the allylic oxidation, which was accomplished in high yield (93%) using a Pd-catalyzed radical oxidation in the presence of Pearlman’s catalyst (Pd(OH)2/C) and t-BuOOH[16] (Scheme 3). The spectroscopic data are in excellent agreement with the ones reported for the isolated compound and further support for the structure was obtained by X-ray diffraction analysis.[17]

Scheme 3

Last Step in the Synthesis of (−)-Nardoaristolone B

Having accomplished the first total synthesis of nardoaristolone B, we were very keen on applying our strategy to a higher enyne in order to gain access to the core of the aristolone family of natural products. Our synthetic effort first involved the conversion of key intermediate enol triflate 5 into the corresponding 1,6-enyne 8. Although the direct Kumada cross-coupling of 5 with propargylmagnesium bromide in the presence of various Pd- or Ni-based catalysts did not take place, we uncovered an unprecedented Kumada cross-coupling of TMS-protected propargylmagnesium bromide with enol triflates. This coupling proceeded smoothly on our model system (4-tert-butylcyclohexanone-derived enol triflate) employing only 2 mol % of Pd(PPh3)4 and 2 equiv of freshly prepared Grignard reagent.[13] However, the cross-coupling was significantly slower on 5 and 20 mol % of Pd complex as well as 4 equiv of Grignard reagent were necessary to obtain full conversion at 23 °C. Under these conditions, the cross-coupling of 5 proceeded smoothly to afford 8 in 77% yield after methanolysis of the TMS group (Scheme 4). This substrate was then treated under a variety of conditions in order to prepare aristolone; however all attempts resulted in the formation of 6a-formyl-6-deoxonardoaristolone 10 in 65% yield when employing IPrAuNTf2 as catalyst.[15] Although (−)-aristolone (2) was not prepared via our original strategy, the preparation of enantioenriched cyclohexanone 4 constitutes a formal synthesis of (−)-2, since racemic 2 has already been prepared from (±)-4 in five steps.[4b]

Scheme 4

Synthesis and Fate of 1,6-Enyne 8

Under all the conditions examined, the gold(I)-catalyzed reaction of 1,6-enyne 8 proceeded exclusively by the 6-exo-dig mode. It is interesting that this result is in contrast to that observed in the reaction 1-ethynyl-2-allylbenzene, which yielded a 6-membered ring ketone by a 6-endo-dig oxidative cyclization.[9] This different behavior can be ascribed to the different substitution pattern of the alkene, which usually controls the outcome in gold(I)-catalyzed cycloisomerizations of 1,6-enynes.[18]

In conclusion, we have developed the first enantioselective synthesis of (−)-nardoaristolone B (1) in seven steps and 14−17% overall yield. Our expedient strategy, by implementation of an enantio- and diastereoselective conjugate addition/enolate alkylation and the first example of a gold(I)-catalyzed oxidative cyclization of enynes in total synthesis, is perfectly suited for the rapid preparation of analogues of this natural product.