Stereospecific Nickel-Catalyzed Reductive Cross-Coupling of Alkyl Tosylate and Allyl Alcohol Electrophiles

Abstract

The stereospecific cross-coupling of easily-accessed electrophiles holds significant promise in the construction of C–C bonds. Herein, we report a nickel-catalyzed reductive coupling of allyl alcohols with chiral, non-racemic alkyl tosylates. This cross-coupling delivers valuable allylation products with high levels of stereospecificity across a range of substrates. The catalytic system consists of a simple nickel salt in conjunction with a commercially available reductant and importantly represents a rare example of a cross-coupling involving the C–O bonds of two electrophiles.

Graphical Abstract

Transformations which enable the catalytic, stereospecific construction of C–C bonds using simple chiral, non-racemic electrophiles are unique, enabling tools for asymmetric synthesis.¹ Recent notable advances in this nascent area have capitalized on the availability of enantioenriched benzylic or aliphatic secondary alcohols in diverse metal-catalyzed bond constructions.² For example, we have reported several stereospecific, carbonylative transformations of alkyl tosylates using cobaltate catalysis for the concise asymmetric preparation of diverse carbonyl compounds.³ We recently became intrigued with the possibility of expanding the scope of stereospecific C–C bond construction using secondary alkyl tosylates in cross-electrophile-type couplings.^2e

The stereoselective allylation of organic electrophiles is among the most useful C–C bond-forming reactions in chemical synthesis,⁴ and we became interested in pursuing a stereospecific allylation of unactivated secondary tosylates. In particular, we sought a cross-coupling variant that used a simple C–O allyl electrophile as a coupling partner and avoided the preparation of highly reactive organometallic reagents.⁵ In notable related work involving a stereospecific alkyl-allyl coupling, Jarvo and co-workers have reported an intramolecular cyclization of 2-vinyl-4-halotetrahydropyrans to access vinyl cyclopropanes.⁶ Recent studies have also reported cobalt- and nickel-catalyzed reductive allylations of alkyl bromides with allylic acetates or carbonates, although these radical-mediated couplings are stereoablative.⁷

We became intrigued by the possibility of a stereospecific nickel-catalyzed alkyl-allyl cross-coupling of two C–O electrophiles. Interestingly, in studies of a nickel-catalyzed silylation of allyl ethers with chlorosilane electrophiles, Kambe and co-workers provided preliminary evidence for the possibility of using alkyl tosylates as electrophiles in couplings with allyl ethers, although the stereoselectivity of the reaction was not addressed (Figure 1).⁸ A stereospecific coupling of chiral, non-racemic alkyl tosylates and simple allyl C–O electrophiles (especially allyl alcohols) would leverage these two readily available coupling partners to significantly extend the capabilities of stereospecific cross-coupling. Herein, we report the successful development of such a reaction forging a C–C bond between two common C–O electrophiles in stereospecific fashion.

Figure 1.

Intermolecular cross-coupling of secondary alkyl tosylates and allyl electrophiles.

Our investigations commenced with secondary tosylate 1 and allyl alcohol 2 as our two C–O electrophiles (Table 1). Following simple in situ silylation of the alcohol by hexamethyldisilazane (HMDS), a catalytic system comprised of 2 mol % of NiCl₂ and 2.2 equivalents of vinyl magnesium chloride as reductant delivered the allylation product 3 in good yield with excellent stereocontrol (80% ¹H NMR yield, 98% es, entry 1) in a 1:1 mixture of Et₂O:THF at room temperature. Substituting NiCl₂ with either Ni(acac)₂ or Ni(cod)₂ significantly lowered the yield (entries 2 and 3). Increasing the amount of catalyst to 5 mol % NiCl₂ had little impact on the reaction (entry 4). Substituting vinyl MgCl for MeMgCl led to no observed product (entry 5), whereas the use of PhMgCl led to conventional Kumada-type cross-coupling of the allyl electrophile (entry 6). Decreasing the amount of reductant or performing the reaction at 0 °C also reduced reaction efficiency (entries 7 and 8). Removing Et₂O from the solvent mixture led to a slightly lower yield (entry 9). Little reaction occurred in the absence of either the reductant or nickel precatalyst (entries 10 and 11). The low but measurable yield in the experiment of entry 11 is likely the result of a small amount of allyl MgCl forming (vide infra).

Table 1.

Stereospecific allylation of secondary alkyl tosylate 1 with allyl alcohol.

entry	variation from standard conditions	% yielda	% esb
1	none	80	98
2	2 mol % Ni(acac)2 instead of NiCl2	65	98
3	2 mol % Ni(cod)2 instead of NiCl2	68	97
4	5 mol % NiCl2	81	98
5	2.2 equiv MeMgCl instead of vinyl MgCl	0	–
6	2.2 equiv PhMgCl instead of vinyl MgCl	0	–
7	1.0 equiv vinyl MgCl	35	98
8	0 °C instead of rt	66	99
9	No Et2O	73	95
10	No vinyl MgCl	0	–
11	No NiCl2	9	–

Reactions were performed with [2]₀= 0.5 M.

^aYields determined by ¹H NMR spectroscopy of crude reaction mixture using an internal standard.

^bEnantiospecificity (es) = (ee_product/ee_substrate) × 100% determined by chiral SFC analysis.

With suitable conditions in hand, we surveyed a diverse range of alkyl tosylates and allyl alcohols (Figure 2). Reactions of secondary tosylates with varying alkyl chain lengths all proceeded efficiently (3–5). The reaction tolerated the presence of both aryl ethers and aryl chlorides (6–7), which have the potential to react with low valent nickel.¹⁰ Substrates containing common pyrrole and thiophene heterocycles afforded cross-coupling products 8 and 9, respectively, in good yield with excellent stereospecificity. Chemoselective tosylate allylation proceeded in the presence of an alkyne as well, delivering product 10. The catalytic cross-coupling is also successful with cyclic substrates, as demonstrated by the reactions leading to products 11–14.

Figure 2.

Nickel-catalyzed allylation of diverse substrates. Yields are of isolated product unless otherwise noted. Enantiospecificity determined by chiral SFC and HPLC analysis. ^a(S)-enantiomer of alkyl tosylate used as substrate. ^bReactions performed with 5 mol % NiCl₂. ^cYield determined by ¹H NMR spectroscopy of crude reaction mixture using an internal standard. ^dTriethylsilyl ether used as substrate. ^eReaction performed with 1 equiv allyl alcohol, 1.2 equiv alkyl tosylate, and 0.55 equiv HMDS.

We next examined the cross-coupling of a range of allyl alcohols (Figure 2, right). Transformations of 2-substituted allyl alcohols delivered products 15–17 in good yield with excellent stereospecificity. Cross-coupling of penta-1,4-dien-3-ol proceeded with high regioselectivity, yielding skipped diene 18. Reactions involving 1,2-disubstituted alkenes afforded products in good yield and enantiospecificity, albeit with poor diastereoselectivity owing to the production of the branched product as demonstrated by the reaction of cinnamyl alcohol to provide 19. With the goal of identifying a 1,2-disubstituted alkene favoring the linear product, we attempted the cross-coupling of (E)-3-(PhMe₂Si)-prop-2-en-1-ol. We were pleased to find that this reaction produced the linear vinyl silane 20 exclusively, with excellent stereocontrol. Such linear-selectivity has been previously observed in other metal-catalyzed allylations and provides opportunity for further synthetic elaboration via cross-coupling.^4a,9 Finally, we demonstrated that other allyl ethers are viable coupling partners in the reaction, using both methyl and phenyl allyl ethers. As expected, allyl acetate–the most common electrophile in transition metal catalyzed allylations–is not applicable likely owing to the reductant involved.¹¹

The stereoselective introduction of the alkene functionality herein provides for numerous attractive avenues in synthesis. As a representative example, we sought to apply our transformation to an asymmetric synthesis of an alpha-2β receptor agonist (24, Scheme 1).¹² The relative absence of functionality in proximity to the stereogenic center of 24 poses a challenge in designing an asymmetric synthesis. Starting with readily prepared chiral, non-racemic tosylate 21, stereospecific allylation with allyl alcohol as the coupling partner delivered 22 with excellent stereoselectivity on gram scale. Next, ring-closing metathesis successfully delivered the cyclohexene 23 in good yield. Silyl deprotection, oxidation, and Van Leusen imidazole synthesis then delivered 24.

Scheme 1.

Asymmetric synthesis of alpha-2β receptor agonist 24 featuring the stereospecific cross-coupling.

We considered two likely mechanistic scenarios for the cross-coupling, which differed in the identity of the nucleophile: 1) catalytic formation of an allyl Grignard reagent, or 2) catalytic formation of a nucleophilic allyl nickelate species. We first performed a control experiment where the allyl silyl ether formed in situ was allowed to react with the catalytic system, followed by quenching with benzaldehyde (eq 1). We hypothesized that significant amounts of the allylation product should be observed if allyl Grignard had formed. The allylation product was indeed produced in good yield as a 5.0:1 mixture with the vinylation product, consistent with allyl Grignard formation. Furthermore, we performed a simple addition of allylmagnesium chloride to tosylate (S)-1 (eq 2). Previous studies have demonstrated the ability of allyl Grignard reagents to add to tosylate electrophiles, but the stereochemical outcome of these additions was not studied.¹³ The addition produced 3 in virtually the same yield and stereospecificity as the cross-coupling with allyl alcohol (Figure 2), consistent with the allyl Grignard as the active nucleophile in the cross-coupling.

A mechanistic hypothesis consistent with our results is shown in Scheme 2. Reduction of the nickel precatalyst delivers a nickel(0) butadiene complex, which undergoes oxidative addition to the allyl silyl ether previously generated in situ from the allyl alcohol. The allylnickel then transmetallates with another equivalent of the reductant to generate an allylnickel(II) vinyl complex, which subsequently undergoes nucleophilic attack to generate an allylnickelate species. Transmetallation of the allylnickelate then produces the allyl Grignard nucleophile which adds to the alkyl tosylate with high stereospecificity.

Scheme 2.

Plausible mechanism for the catalytic, stereospecific cross-coupling.

In conclusion, we have developed a stereospecific cross-coupling of alkyl tosylates and allyl alcohols providing a range of allylation products with high levels of stereospecificity. This reaction forges a C–C bond from two widely available C–O electrophiles using a simple catalytic system. Our studies support the intermediacy of an allyl Grignard reagent as the active nucleophile. We anticipate that the use of reductive nickel catalysis will enable other fundamental, stereospecific C–C bond constructions using simple substrates.