Enantioselective cross-coupling synthesis of allylic silanes

keywords: enantioselective catalysis, cross-coupling, stereoselective allylation of carbonyls

The first installment of this blog features a recent communication from Sarah Reisman's laboratory, "Synthesis of Enantioenriched Allylic Silanes via Nickel-Catalyzed Reductive Cross-Coupling" (Hofstra et al., 2017 ASAP).  This publication builds on her laboratory's earlier work on enantioselective nickel-catalyzed cross-couplings (Cherney and Reisman, 2014).  The recent extension to the synthesis of allylic silanes is important because the functionalized products are valuable synthons in their own right.  This presentation will be divided into two parts, corresponding to different chapters in volume 2: 

1) enantioselective cross-coupling, and 
2) stereoselective addition of chiral allylic silanes to organic electrophiles.  
__________________

1) Enantioselective reductive cross-coupling with nickel catalysis: 

Vinylic bromides undergo enantioselective reductive cross-coupling with sp3-hybridized alpha-trialkylsilyl benzylic chlorides, promoted by a chiral nickel bis(oxazoline) catalyst in the presence of cobalt phthalocyanine (CoPc) co-catalyst, and manganese powder as the stoichiometric reductant (Figure 1).  The vinylic bromides are prepared by regioselective and stereospecific hydrozirconations of terminal alkynes (Li et al., 2006), and the chiral racemic silanes arise from LDA deprotonation of benzylic chlorides followed by silylation (Manzano et al., 2013).  

Figure 1.  Enantioselective synthesis of chiral allylic silanes by
Ni-catalyzed reductive cross-coupling

The likely mechanism begins with two-electron reduction of the Ni(II) pre-catalyst to Ni(0) (Figure 2).  

Figure 2.  Two-electron reduction of Ni(II) bisoxazoline pre-catalyst to Ni(0)

The corresponding reduction of cobalt phthalocyanine (CoPc) is a single-electron reduction  from Co(II) to Co(I) (Figure 3).  The chemoselectivity favoring activation of the sp3-hybridized benzylic chloride arises from nucleophilic substitution of reduced CoPc, Co(I), to give a PcCo-alkyl intermediate, Co(III) (Ackerman et al., 2015).   

Figure 3.  Single-electron reduction of Co(II) phthalocyanine to Co(I), and
nucleophilic substitution of the benzylic chloride to form a Co(III)-alkyl

Although the catalytic cycle appears to follow the typical order of steps in cross-coupling, namely oxidative addition, transmetallation, and reductive elimination, this mechanism likely involves a Ni(0) ⇄ Ni(II) ⇄ Ni(III) ⇄ Ni(I) catalytic cycle (Cherney and Reisman, 2014, Figure 4).  Oxidative addition of the Ni(0) catalyst with the vinylic bromide gives a Ni(II) intermediate, which accepts the benzylic group from the Co(III)-alkyl complex in a transmetallation step, to form a pentacoordinate Ni(III) intermediate.  Presumably it is this step in which the chiral center is set.  This transmetallation step represents single electron changes in the formal oxidation states of cobalt and nickel.  Alkyl transfer via a benzylic radical intermediate would allow either enantiomer of the chiral Co(III)-alkyl to afford a single enantiomer of the chiral Ni(III)-alkyl (Ackerman et al., 2015).  However, the corresponding couplings with simpler benzylic chlorides under similar conditions are inconsistent with a radical chain mechanism, suggesting that the alkyl transfer step may occur without involving a free radical intermediate (Cherney and Reisman, 2014).  Reductive elimination from the organonickel (III) intermediate produces the allylic silane product and a Ni(I) catalytic intermediate, which then undergoes single-electron reduction with Mn(0) to regenerate the Ni(0) catalyst, with MnBr2 as the stoichiometric byproduct.  In contrast to classical cross-coupling methods, an organomanganese reagent generated in situ is unlikely (Cherney and Reisman, 2014); instead the organocobalt catalytic intermediate is the alkyl group donor.  

Figure 4.  Possible catalytic cycle for enantioselective reductive cross-coupling

The authors do not offer a model for stereoinduction from the chiral catalyst.  From examining a three-dimensional model of a square planar-pyramidal Ni(III) complex, one may speculate on this intermediate with the trimethylsilyl projecting away from all steric bulk (Figure 5).  The hydrogen at the chiral center might face toward one of the flaps, perhaps even favored by an aromatic pi-H interaction.  Such an aromatic pi-H interaction might favor the orientation of the phenyl substituent from the benzylic chloride.  At least this is consistent with the apparent requirement of an aromatic group in the benzylic chloride reactant (see also Cherney & Reisman, 2014).  This model presumes stereospecific retention of configuration in the reductive elimination step, and also does not account for the possibility of equilibration in the alkyl transfer step from the Co(III)-alkyl to the Ni(III)-alkyl intermediates.   

Figure 5.  Hypothetical model for stereoinduction from the chiral ligand

____________________

2) Stereoselective addition of chiral allylic silanes to organic electrophiles:

These transformations proceed with excellent stereospecificity from the chiral allylic silane with a variety of organic electrophiles activated by Lewis and Brønsted acids, including aldehydes, acetals, and aminals.  Some of these applications are analogous with earlier work from the Panek laboratory, using chiral allylic silanes prepared by other methods (Arefolov & Panek, 2005; Wu et al., 2010). 

a) Building on an open transition state model for additions of allyltrialkylsilanes to aldehydes, triggered by Lewis acids, the intermolecular reaction may proceed via the antiperiplanar arrangement of pi-reactants, with hydrogens anti- in the developing carbon-carbon bond (Figure 6).  The chirality of the allylic silane favors a reactive conformation in which the hydrogen at the chiral center is s-trans to the proximal alkene hydrogen, so that the Lewis acid-activated aldehyde reacts onto the opposite face of the allylic silane, leading to the trans-alkene product.  Diastereoselectivity is presumably a function of the trans- stereochemistry of the allylic silane.  However, this stereospecificity has not been rigorously tested, as the corresponding cis-allylic silanes cannot be prepared by the methodology described in part 1.  The enantio- and diastereoselectivities were evaluated not on the homoallylic alcohol, but on the disubstituted tetrahydrofuran arising from intramolecular substitution of the pendant chloride with the secondary chiral alkoxide, under basic conditions. 

Figure 6.  Intermolecular addition of chiral allylic silane with an aldehyde

b) Lewis-acid promoted transacetalation of the primary alcohol linked to the chiral allylic silane results in intramolecular cyclization, giving the diastereomer of the disubstituted tetrahydrofuran with the opposite configuration of the chiral center arising from the acetal intermediate (Figure 7).  This corresponds to a synclinal orientation of the pi-reactants, which controls the diastereoselectivity of this transformation.  The enantioselectivity arises from addition of electrophile onto the face of the allylic silane opposite to the trimethylsilyl group, once again arising from a reactive conformation in which the hydrogen at the chiral center is s-trans to the proximal alkene hydrogen. 

Figure 7.  Intramolecular addition of chiral allylic silane with acetal-derived electrophile

c) To demonstrate the versatility of this method, the phthalimido-substituted allylic silane described above in part 1 underwent chemoselective NaBH4 reduction to provide the aminal, tethered to the chiral allylic silane.  Brønsted acid-activation forms an N-acyliminium ion intermediate, which reacts with the tethered allylic silane with measurable diastereoselectivity and high enantioselectivity for the major diastereomer (Figure 8).  The major stereoisomer corresponds to a reactive conformation with an antiperiplanar orientation of the pi-reactants, although the most significant of the minor diastereomers may have arisen from a synclinal orientation.  

Figure 8.  Intramolecular addition of chiral allylic silane via N-acyliminium ion intermediate

References: 
- Ackerman, L. K. G.; Anka-Lufford, L. L.; Naodovic, M.; Weix, D. J. Chem. Sci. 2015, 6, 1115. 
- Arefolov, A.; Panek, J. S. J. Am. Chem. Soc. 2005, 127, 5596. 
- Cherney, A. H.; Reisman, S. E. J. Am. Chem. Soc. 2014, 136, 14365.
- Hofstra, J. L.; Cherney, A. H.; Ordner, C. M.; Reisman, S. E. J. Am. Chem. Soc. 2017 ASAP; DOI: 10.1021/jacs.7b11707

- Li, D.-R.; Zhang, D.-H.; Sun, C.-Y.; Zhang, J.-W.; Yang, L.; Chen, J.; Liu, B.; Su, C.; Zhou, W.-S.; Lin, G.-Q. Chem. Eur. J. 2006, 12, 1185.

- Manzano, R.; Rominger, F.; Hashmi, A. S. K. Organometallics 2013, 32, 2199.

- Wu, J.; Chen, Y.; Panek, J. S. Org. Lett. 2010, 12, 2112.