Enantioselective Synthesis of (E)-δ-Stannyl Homoallylic Alcohols Via Aldehyde Allylboration Using α-Stannylallylboranes Generated by Allene Hydroboration Followed By a Highly Diastereoselective 1,3-Boratropic Shift

The asymmetric carbonyl allylation reaction is an important transformation in organic synthesis.¹ Although extensive efforts in the past three decades have provided many useful allyl metal reagents, new developments in this area continue to emerge.^2–6 One disadvantage of the vast majority of prior allylation methods, however, is that homoallylic alcohols with terminal vinyl groups are generated. Several step manipulations are often required for further carbon–carbon bond formation at this position.⁷ Several procedures exist for synthesis of non-racemic α-substituted allylboron reagents, which undergo aldehyde allylboration reactions to give products with substituted olefins,¹ but such reagents typically require multistep syntheses (e.g., involving the Matteson homologation⁸ as in Hall's recent work^5e) and are not widely adopted in the literature. An important advance in this area involving the lithiathion-borylation of allylic carbamates has been recently achieved by Aggarwal.⁹ Olefin cross metathesis is one method that permits a terminal olefin to be modified directly. However, the efficiency as well as the E/Z selectivity of cross metathesis depends on the properties of the olefin metathesis partners.¹⁰ Consequently, olefin metathesis does not provide a general solution to the problem of synthesis of homoallylic alcohols with substituted alkenes. Therefore, a simple method that provides direct access to homoallylic alcohols with functionalized olefins is highly desirable. Toward this end, we have discovered and report herein a direct, one step, highly enantioselective synthesis of (E)-δ-stannyl-homoallylic alcohols via an allene hydroboration-aldehyde allylboration sequence.

In connection with an ongoing problem in natural products synthesis, we explored the hydroboration of allenylstannane 1 with diisopinocampheylborane [(^dIpc)₂BH]. In principle, hydroboration of 1 could provide allylboranes 2 or 3, depending on the hydroboration regioselectivity and the rate and equilibrium constant for 1,3-boratropic isomerization in this system.¹¹ Based on the ability of −SnR₃ groups to stabilize a β carbocation,¹² we anticipated that the formation of allylborane intermediate 2 might be thermodynamically favored.¹³ If so, (E)-δ-stannyl-homoallylic alcohols 4 would be available following the reaction of 2 with an aldehyde (Figure 1).

Figure 1.

Treatment of allenylstannane 1 at −40 °C with (^dIpc)₂BH in diethyl ether with the solution being allowed to warm to −20 °C over 5 h to complete the hydroboration, followed by treatment of the resulting allylborane with hydrocinnamaldehyde at −78 °C provided (E)-δ-stannyl-homoallylic alcohol 4a in 64% yield and, remarkably, with >95% e.e. (entry 1, Table 1). Application of this procedure to a variety of other representative aldehydes (entries 1–9, Table 1) provided homoallylic alcohols 4a–h in 51–78% yield and 92% to >95% e.e. (absolute configurations were assigned by Mosher ester analysis¹⁴). The olefin geometry of homoallylic alcohols 4a–h was E (J^E = 18.8–19.2 Hz); the corresponding (Z)-olefin isomers as well as the regioisomeric homoallylic alcohols 5 (or dienes accessible by elimination of 5) were not detected in any of the experiments performed under these conditions.

Table 1.

Synthesis of δ-Stannyl Homoallylic Alcohols 4 via Hydroboration of 1 at −40°C to −20°C (via Kinetically Controlled Allylborane Isomerization)^a vs. at 0°C (Thermodynamically Controlled Allylborane Isomerization)

entry	RCHO	product	yield	%e.e.b
1	Ph(CH2)2CHO	4a	64%	>95
2	PhCH2CHO	4b	67%	>95
3	PhCHO	4c	78%	93
4	BnO(CH2)2CHO	4d	68%	>95
5	BnOCH2CHO	4e	71%	>95
6	PhCH=CHCHO	4f	73%	>95
7	CyCHO	4g	55%	92
8	t-BuCHO	4h	51%	94
9c	Ph(CH2)2CHO	ent-4a	71%	94
10d	Ph(CH2)2CHO	ent-4a	73%	30

^(a)Reactions were performed by treating 1 with (^dIpc)₂BH (0.7 equiv.) in Et₂O at −40°C and warmed to −20°C over 5 h, followed by the addition of RCHO (0.5 equiv.) at −78°C. The mixture was then allowed to stir at −78°C for 8 h. The reactions were subjected to a standard workup (NaHCO₃, H₂O₂) at 0°C prior to product isolation.

^(b)Determined by Mosher ester analysis.

^(c)(^lIpc)₂BH was used.

^(d)This reaction was performed by treating 1 with (^dIpc)₂BH (0.7 equiv.) in Et₂O at 0°C followed by addition of RCHO (0.5 equiv.) at −78°C.

The enantioselectivity of this sequence proved to be highly dependent on experimental conditions. As shown in entry 10 of Table 1, when the hydroboration of 1 was performed at 0 °C followed by addition of aldehydes at −78 °C, the enantiomeric homoallylic alcohols, ent-4a was obtained in 73% yield and 30% e.e. (measured by Mosher ester analysis¹⁴). Similarly, when the hydroboration step was carried out at –40 °C and the solution was then allowed to warm to 0 °C, addition of hydrocinnamaldehyde at −78 °C provided ent-4a, in 30% e.e. and 77% yield. Here again, products with a (Z)-vinylstannane unit (e.g. 6 in Scheme 1a) were not detected.

Scheme 1.

Hydroboration of 1 and Allylborane Isomerization Pathways

Based on recent work on allenylboronate hydroboration,¹⁵ the reaction of allenylstannane 1 with (^dIpc)₂BH is expected to provide (Z)-γ-stannylallylborane 3a as the kinetic product (Scheme 1a). If so, the results in Table 1 (entries 1–9) require that 3a undergoes a kinetically controlled and highly diastereoselective 1,3-boratropic shift at temperatures below −20 °C to give α-stannylallylborane 2a, which then reacts with aldehydes at −78 °C to give 4 via the usual chair-like transition state, TS-2. The (E)-olefin geometry of 4 dictates that the α-stannyl unit occupies a pseudoequatorial position in TS-2. The absolute configuration of the hydroxyl group of 4 (which we have assigned by conversion of 4 to known compounds or by application of the modified Mosher method¹⁴; see SI) is fully consistent with the normal sense of asymmetric induction by the −B(Ipc)₂ unit.¹⁶

Support for this analysis is provided by low temperature ¹H-NMR studies of hydroboration of 1 (see SI) which indicates that (Z)-3a is present at short reaction time, as well as by data from an experiment in which the hydroboration of allenylstannane 1 with (^dIpc)₂BH was performed at −40 °C for 2 h before addition of hydrocinnamaldehyde at −78 °C. Small amounts (<5%) of syn-β-hydroxyallylstannane 5a were isolated along with recovered allenylstannane 1 and homoallylic alcohol 4a as the major product (the allene hydroboration was not complete under these conditions). Because neither β-hydroxyallylstannanes 5 nor corresponding dienes are observed in any experiments in which the hydroboration solution was allowed to warm to −20 °C prior to addition of the aldehyde, the results of the −40 °C hydroboration experiment and the ¹H-NMR study of the hydroboration reaction (SI) are consistent with the theses that (i) the kinetic mode of hydroboration of 1 is that depicted in Scheme 1a leading to (Z)-γ-stannylallylboronate 3a (the precursor to 5), and (ii) that equilibrium constant for the 1,3-boratropic shift is highly displaced in favor of 2a (and/or 2b, depending on reaction temperature, vide supra).

The data for the experiment summarized in Table 1, entry 10, indicates that when the hydroboration of 1 is performed at 0 °C, (or if the product of the hydroboration at −40 °C to −20 °C is allowed to warm to 0 °C) a rapid and reversible 1,3-boratropic shift occurs that gives a thermodynamic mixture of allylboranes 2a and 2b, slightly favoring 2b. Indeed, when the hydroboration reaction of 1 was monitored at 0 °C, a ca. 1:2 mixture of two α-stannyl allylborane species—2a and 2b—was observed via ¹H-NMR (see SI). Allylborane 2a reacts with aldehydes via TS-2 to give 4, and 2b reacts via TS-3 to give ent-4. These data indicate that pseudoequatorial placement of the α-stannyl unit in TS-3 overrides the enantiofacial selectivity of the (^dIpc)₂B— group which, if dominant, would have dictated the formation of (Z)-δ-stannyl-homoallylic alcohols 6 via TS-4. That is, the stereodirecting influences of the α-stannylboryl stereocenter and that of the (^dIpc)₂B— group are mismatched in 2b, with the α-stannylboryl stereocenter being the more dominant of the two.

We have utilized B3LYP DFT to explore the transition states of the hydroboration, 1,3-boratropic rearrangement and aldehyde allylation steps leading to 4 and ent-4 (SnBu₃ groups were modeled as SnMe₃).¹⁷ As shown in Scheme 1a, hydroboration of 1 proceeds through the lowest energy pathway¹⁷ TS-1 (ΔH^‡ = 9.4 kcal/mol) and gives the (Z)-δ-stannylallylborane 3a (ΔH = −24.0 kcal/mol) as the kinetic product. These DFT calculations indicate that hydroboration transition states that lead directly to 2a and 2b are substantially higher in energy due to severe congestion between the Bu₃Sn and Ipc groups.¹⁸ Two possible concerted diastereomeric 1,3-boratropic shift transition states lead from 3a to either 2a or 2b. TS-5 (ΔH^‡ = −5.8 kcal/mol) is favored over TS-6 (ΔH^‡ = −3.6 kcal/mol) by 2.2 kcal/mol because the right-hand isopinocampheyl group is oriented so that the hydrogen and methylene positions are closest to the Bu₃Sn group, while in TS-6 the Bu₃Sn group is next to the hydrogen and large tertiary carbon center (Scheme 1b). The repulsion between the isopinocampheyl and Bu₃Sn groups in TS-6 results in an asynchronous TS with partial bond lengths of 1.76 and 2.01 Å, in contrast to the nearly synchronous partial bond lengths in TS-5 (1.88 and 1.82 Å). Finally, allylborations of 2a with aldehyde substrates provides homoallylic alcohols 4 via TS-2 with equatorial placement of the α-Bu₃Sn group.

Double asymmetric allylboration reactions of 2a with chiral aldehydes 7 and 8 are shown in Table 2. Hydroboration of allenylstannane 1 with either (^dIpc)₂BH or (^lIpc)₂BH at −40 °C (with warming to −20 °C) and then addition of aldehyde 7 at −78 °C provided 4i and 4j in 70–75% yield, each with >50:1 diastereoselectivity (¹H-NMR analysis—the alternative minor diastereomers could not be detected in either experiment). Similarly, use of the more elaborated aldehyde 8 as substrate provided 4k and 4l in 71–78% yield, both again with >50:1 diastereoselectivity (Table 3). The very high diastereoselectivity of these pairs of double asymmetric reactions attests to the enantioselectivity of reagent 2a and the remarkable, highly diastereoselective 1,3-boratropic shift that is involved in the generation of 2a.

Table 2.

Double Asymmetric Stannyl Allylboration Reactions with Aldehydes 7 and 8

entry	borane	products	yield	d.s.
1	(dIpc)2BH	4i : 4j	70%	>50 : 1
2	(lIpc)2BH	4i : 4j	75%	1 : >50

entry	borane	products	yield	d.s.
1	(dIpc)2BH	4k : 4l	71%	>50 : 1
2	(lIpc)2BH	4k : 4l	78%	1 : >50

In conclusion, we have developed a highly enantioselective synthesis of (E)-δ-stannyl-homoallylic alcohols via aldehyde allylboration reactions of the double chiral allylborane reagent 2a. Allylborane 2a is easily generated from the hydroboration of commercially available allenylstannane 1 with (dIpc)₂BH at −40 to −20 °C followed by a kinetically controlled and highly diastereoselective 1,3-boratropic shift of intermediate 3a. While the 1,3-boratropic shift, including examples that occur with 1,3-stereochemical transfer are well known,^3a,11 the discovery of the highly diastereoselective 1,3-boratropic shift of 3a to 2a was totally unexpected. To the best of our knowledge, the asymmetric induction due to the asymmetry of the −B(Ipc)₂ group (or any other chiral dialkylboryl unit) in the conversion of 3a to 2a has not been previously documented in the literature.^3a,11 Subsequent allylboration reactions of reagent 2a with aldehydes provide homoallylic alcohols 4 in good yields and with excellent enantioselectivities. Compared to conventional allylmetal chemistry, which typically provides homoallylic alcohols with a terminal olefin unit, this stannylallylboration reaction is exceptionally valuable in that it provides homoallylic alcohols with a functionalized olefin unit, suitable for use in subsequent C–C bond formations (numerous examples of which are documented in the literature).¹⁹ Applications of this methodology in the synthesis of natural products will be reported in due course.