Enantioselective Synthesis of (Z)-1,2-anti-2,5-anti-Triol Monosilyl Ethers Using a Cross-Metathesis Allylboration Sequence

Abstract

The enantioselective synthesis of (Z)-1,2-anti-2,5-anti-triol monosilyl ethers via a two-step sequence involving olefin cross-metathesis of β-alkoxyallylboronate 4 and subsequent allylboration of the derived bisboryl intermediate 6 provides triol monoethers 7 with good to excellent diastereoselectivity.

Stereodefined 1,2,5-triol subunits are present in many biologically active natural products. The zooxanthellamides, stagonolides, and annonaceous acetogenins are three classes of natural products in which this moiety is found.¹

The boron-mediated allylation reaction has proven to be an important method for carbon-carbon bond formation.² This reaction proceeds with predictable control of stereochemistry via cyclic, chair-like transition states, and has been widely applied in the synthesis of natural products. In recent years, our laboratory has focused on developing higher-order applications of the allylboration reaction in which two C—C bond forming events are performed in a single operation with a bifunctionalized allylboron reagent (i.e., the double allylboration reaction).^3,4 One limitation to this methodology, which provides 1,5-diols with excellent stereochemical control, is the general difficulty of synthesizing reagents that enable placement of additional substituents on the 1,5-diol scaffold as a direct result of the double allylboration event.

In an effort to expand the utility of the double allylboration reaction, we imagined that manipulation of the terminal olefin of boronate ester 3 via cross metathesis might provide a means to introduce additional functionality (Scheme 1).⁵ Owing to our long-standing interest in the synthesis of polyhydroxylated natural products,⁶ we concentrated on metathesis partners that would enable us to access 1,2-anti-2,5-anti triols of type 7. Specifically, we studied the cross metathesis reactions of silyl ether protected allylboronates 4 and pinacol vinyl boronate (5), which is a well-established metathesis coupling partner.⁷ Accordingly, we report herein a highly diastereoselective synthesis of (Z)-1,2-anti-2,5-anti-1,2,5-triols 7 by the sequence summarized in Scheme 1.

Scheme 1.

The Allylboration-Cross-Metathesis-Allylation Sequence

Allylboronate 3a (R₁ = CH₂CH₂Ph, X = H) was synthesized via the hydroboration of allene 1a (X = H) with ^lIpc₂BH followed by treatment of allylboronate 2 with hydrocinnamaldehyde at −78 °C.^3a,8 Initial attempts to effect cross metathesis reactions of silyl ethers 4a (derived from 3a) with pinacol vinylboronate 5 using Grubbs’ 2^nd generation catalyst were unsuccessful.⁹ The lack of reactivity was attributed to the steric bulk associated with the tetraphenyl-substituted dioxaborolane unit of 4a (R₁ = CH₂CH₂Ph, X = H), along with the bulky pinacol boronate ester of the coupling partner 5. Therefore, we turned to use of the more robust Hoveyda-Grubbs’ 2^nd generation catalyst¹⁰ in order to accommodate the increase in reaction temperature necessary to effect the cross metathesis reaction. However, treating a mixture of 4a (R₁ = CH₂CH₂Ph, X = H, −SiR₃ = TBS), 5 and the second generation Grubbs-Hoveyda catalyst in toluene at 80 °C led to olefin 14–24% isomerization of 4a to the corresponding vinylboronate after 24 h, presumably due to a Ru-H species generated in situ upon decomposition of the ruthenium catalyst.¹¹ Following an extensive screening of additives to suppress the olefin isomerization, tetrafluoro-1,4-benzoquinone was selected as the most effective reagent to prevent this side reaction.¹²

A third variable that had to be optimized to maximize selectivity for the (E)-olefin geometry in the cross metathesis product 6 is the size of the alcohol protecting group. Because it is known that steric bulk at the allylic position generally enhances selectivity for the (E)-olefinic metathesis product,¹³ we studied the cross metathesis of a series of silyl ethers generated from 3a (R₁ = PhCH₂CH₂, X = H) (Table 1). These experiments led to the identification of the TIPS ether as the alcohol protecting group that gives greatest selectivity for (E)-6.

Table 1.

(E)-Selectivity in the Cross-Metathesis Reactions of Silyl Ether Derivatives of Allylboronate 3a

entrya	R3Si	E/Z ratio in 5b
1	TMS	3:1
2	TES	6:1
3	TBS	8:1
4	TBDPS	16:1
5	TIPS	≥ 20:1

^aReactions were performed by treating silyl ethers derived from allylboronate 3a (R₁ = CH₂CH₂Ph, X=H; 1.0 equiv) and pinacol vinyl boronate 5 (1.5 equiv) with the 2^nd generation Hoveyda-Grubbs catalyst in toluene (0.5 M) at 80 ° C for 24 h in the presence of tetrafluoro-1,4-benzoquinone (0.10 equiv).

^bOlefin geometry in 6 was determined by ¹H NMR analysis of the crude product.

The results of cross-metathesis reactions of a range of allylboronate substrates are summarized in Table 2. In all cases, the selectivity for the (E)-olefinic product was excellent (≥20 : 1). Most of these reactions provided the vinylboronate products 6a–k in 60–75% yield, along with some recovered 4a–k. Presumably, these reactions did not proceed to completion owing to the high steric of use demands of the cross coupling partners—a consequence of the TIPS ether to maximize selectivity for the (E)-olefin product in the metathesis reaction.

Table 2.

Cross-Metathesis Reactions of Allylboronates 4a–k.

entry	R1	product	catalyst (mol %)	additivec (mol %)	yieldd (%) / brsme (%)
1		6aa	10	10	56 / (90)
			10	none	51 / (−)
		6bb	10	10	61 / (82)
			12	none	74 / (−)
2		6cb	10	10	50 / (57)
3		6db	10	10	69 / (87)
4		6eb	10	10	67 / (80)
5		6fb	10	none	52 / (−)
6		6gb	10	8	55 / (71)
			10	none	67 / (−)
7		6hb	10	10	46 / (61)
			10	none	88 / (−)
8		6ib	10	10	74 / (83)
			15	none	75 / (−)
9		6jb	10	10	45 / (54)
			10	none	43 / (−)
10		6kb	10	none	67 / (−)

^aX = H.

^bX = F.

^cTetrafluorobenzoquinone was added to prevent olefin isomerization.

^dIsolated yields of 6a–k.

^eYield based on recovered starting material.

^fOlefin geometry determined by ¹H NMR analysis of the crude reaction products.

At the outset, we anticipated that the allylboration reactions of 6 would proceed via the chair-like transition state TS-A (see Scheme 1) with the α-silyloxy carbon chain residing in an axial position to avoid non-bonded steric interactions with the tetraphenyl-1,3,2-dioxaborolane unit of the reagent, which become quite significant in the alternative (disfavored) transition state TS-B.^3a Accordingly, it was anticipated that this allylboration sequence would provide (Z)-1,2-anti-2,5-anti-1,2,5-triol monoethers 7 with excellent selectivity following oxidative workup. However, we recognized that a third allylboronate, 9, would be produced in the allylboration of reactions of 6, and that 9 could, in principle, react with a second equivalent of the aldehyde partner to provide double allylboration products of type 10 or 11 (Scheme 2). Indeed, in experiments performed using 6a (with the para substituent X = H), mixtures of 7 (major) and 10/11 were obtained when 6a was treated with 1.0 equiv. of aldehydes at various reaction temperatures and concentrations (data not shown). These results indicated that the reactivities of 6a and 9a (X = H) towards aldehydes are comparable.

Scheme 2.

Allylboronate Intermediate 9 May Undergo a Second Allylboration.

Consequently, we directed our attention to adjusting the reactivity of the allylboronate functional group in 6 relative to that of the intermediate allylboronate species 9. Based on earlier studies in which we observed that a p-nitrophenyl substituent on a 1,3,2-dioxaborolane increased the Lewis acidity of the allylboronate and correspondingly increased the rate of its reaction with aldehydes, compared to the parent allylboronate with a phenyl-substituted dioxaborolane,¹⁴ we synthesized the tetra-p-fluorophenyl substituted allylboronate 6b (X = F) by a sequence paralleling that used for the synthesis of 6a (X = H). Treatment of 6b with hydrocinnamaldehyde in CH₂Cl₂ (0.1 M) for 24 h, and subsequent workup with H₂O₂ and NaOH afforded triol 7a in 67% yield with ≥20 : 1 d.s. Increasing the allylboration reaction period to 36 h provided 7a in 71% yield, and while after 48 h the yield of 7a was 77% (see Scheme 3). In all cases, the double allylboration products 10/11 were not observed. Thus, this proof of principle experiment established that the inductive effect of the p-fluoro substituents increased the reactivity of 6b such that allylboration reactions of intermediate 9 were no longer competitive.

Scheme 3.

Allylboration Reactions of 6b—Scope of Aldehyde Substrates.a

^aUnless indicated otherwise, all allylboration reactions were performed for 36 h at ambient temperature. ^bDetermined by advanced Mosher ester analysis. ^cDiastereomer ratio determined by ¹H NMR analysis of the crude product.

The scope of allylboration reactions of 6b with additional aldehydes is presented in Scheme 3, and results of allylboration reactions of 6c, d, f, g and j with hydrocinnamaldehyde are presented in Scheme 4. In the vast majority of cases (the one exception being the reaction of 6b with hydrocinnamaldehyde in Scheme 3), the reactions were performed at ambient temperature for 36 h. The triol monoethers 7 were obtained in 52–89% yield, 82–93% e.e., and ≥14: 1 d.s. for all examples except 7g (10 : 1 d.s., Scheme 3) and 7n (9 : 1 d.s., Scheme 4).

Scheme 4.

Allylboration Reactions of 6c, 6d, 6f, 6g and 6j with Hydrocinnamaldehyde.

^aDetermined by advanced Mosher ester analysis. ^bDiastereomer ratio determined by ¹H NMR analysis of the crude reaction product.

The relative and absolute stereochemistry of the triol monoethers presented in Schemes 3 and 4 were assigned via advanced Mosher ester analysis of the corresponding MPTA esters.^15,16

In summary, the allylboration-cross metathesis-allylboration sequence presented in Scheme 1 constitutes a convergent and highly enantio- and diastereoslective method for synthesis of (Z)-1,2-anti-2,5-anti-triol monosilyl ethers 7 with synthetically useful efficiency. This new method represents an expansion of our double allylboration methodology,³ and defines a strategy for introduction of substituents at positions internal to the 1,5-diol motif. Applications of this new method in the synthesis of natural products, along with further extensions of the double allylboration methodology, are in progress and will be reported in due course.