Substrate Plasticity Enables Group-Selective Transmetalation: Catalytic Stereospecific Cross-Couplings of Tertiary Boronic Esters

Abstract

Activation of enantiomerically enriched tertiary alkylboronic esters with adamantyllithium generated in situ enables stereoretentive boron-to-copper transmetalation. The resulting alkylcopper species can undergo cross-coupling reactions with an array of electrophiles to furnish synthetically useful compounds bearing quaternary stereocenters. DFT calculations on the transmetalation process provide insights for reactivity and selectivity.

Graphical Abstract

Catalytic construction of enantiomerically-enriched quaternary stereocenters is a challenging and yet important task in synthetic organic chemistry.¹ In spite of impressive advances in stereospecific cross-coupling of configurationally stable secondary nucleophiles,² transmetalation of a configurationally stable tertiary alkyl group to a transition metal is non-trivial and has hampered extension of metal-catalyzed cross-coupling to the installation of quaternary centers. Most of the existing examples of transmetalation of tertiary organometallic reagents employ nucleophilic Grignard reagents,³ organozinc reagents⁴ and organocopper reagents (Scheme 1a).^3e,5 These methods are limited by the availability of nucleophiles, which can be difficult to prepare and are often configurationally unstable at room temperature. Compared to the above-mentioned reagents, tertiary organoboronic esters exhibit higher configurational stability and may be prepared by a variety of efficient enantioselective synthesis methods.⁶ However, these reagents are less nucleophilic and thus challenging for transmetalation. Harris and coworkers utilized tertiary cyclopropyl organoboron reagents for Suzuki-Miyaura and Chan-Lam type coupling reactions (Scheme 1b),⁷ and Uchiyama reported Suzuki-Miyaura cross-couplings of tert-butyllithium activated bicyclo[1.1.1]pentylboronic esters (Scheme 1c).^3e Despite this progress, these examples are limited to strained or benzylic substrates and do not address the issue of stereospecificity. Recognizing the general challenge of coupling a tertiary nucleophile, Molander established a single-electron transmetalation process,⁸ (Scheme 1d) in which a tertiary organoboron was first converted to a free radical, which was then trapped with a nickel complex; this sequence enabled catalytic cross coupling with aryl halides, but the radical nature of the reaction makes it challenging to control stereochemistry.⁹ In this report, we describe a practical catalytic method for generating adamantyllithium in situ, which then reacts with tertiary alkylboronic esters at room temperature. The resulting borate species can engage in stereospecific copper-catalyzed coupling reactions with an array of electrophiles. Insofar as we are aware, the examples presented in this manuscript represent the first that involve stereospecific transmetalation from a tertiary boronate to another metal. In addition to providing a useful synthesis tool, these studies shed light on features that impact transmetalation efficiency.

Scheme 1.

Transmetalation of Tertiary Nucleophiles in Cross-Coupling Reactions.

Recently, we demonstrated that tert-butyllithium-activated primary and secondary organoboronic esters can undergo stereospecific transmetalation to both copper and zinc salts, thereby enabling a broad array of cross-coupling reactions.¹⁰ We considered that alkyllithium-activated tertiary alkylboronic esters might participate in similar processes. An obvious challenge would be the selectivity between transferring the activator versus the substrate alkyl group. While previous activation of secondary alkylboronic esters relied on the faster transmetalation of a secondary alkyl group relative to a tert-butyl group, effective activation of a tertiary group would require an activator that transmetalates slower than the transferring tertiary alkyl substrate. We were inspired by the observation of relatively low reactivity of cyclohexyl boronic esters under tert-butyllithium activation and wondered if the similarly rigid and cyclic nature of an adamantyl group would preclude its transmetallation in the presence of other tertiary groups.

X-ray analysis of a crystal of (tert-butyl)(adamantyl)B(pin)●Li(THF)₂ (Figure 1a), obtained by slow diffusion of hexane into a THF solution of the borate, hinted at selective group transfer in transmetallation: the B–(adamantyl) bond is shorter than the B–(tert-butyl) bond (1.673 Å vs. 1.692 Å), which suggests that the latter may be more reactive.^10,11 Aligned with these observations, DFT calculations (Figure 1b) indicated that the barrier for transmetalation of the adamantyl group from adamantyl(tert-butyl)pinacol boronate to CuCN is 2.3 kcal/mol higher in energy than for transmetalation of the tert-butyl group. Distortion-interaction analysis¹² revealed that the difference in energy between these two transition states arises from different energies required to distort the transmetalating group as the transition state is approached (see Supporting Information). Analysis of the calculated transition state structures (Figure 1c) shows increased pyramidalization in the transferring group, presumably to accommodate the increased steric encumbrance associated with a transient five-coordinated carbon center. The extent of pyramidalization as the reaction progresses from the ground state to the transition state, assessed by comparing the sum of the three C–C–C bond angles of the transferring alkyl group (Figure 1d, ideal trigonal atom = 360°; ideal tetrahedron = 328.5°), was found to be greater for the tert-butyl group in TS2 than the adamantyl group in TS1 (for relevant computational experiments that calculate ΔE as POAV¹³ is varied across a series of tertiary alkanes, see the Supporting Information). We postulate that in the absence of other factors (hybridization, steric effects) the plasticity of the transmetalating group can be a factor that facilitates transmetalation: the greater flexibility of the tert-butyl group allows it to undergo more pyramidalization than the adamantyl ligand, and yet it requires less energy to achieve this distortion. This favors the tert-butyl transmetalation pathway.

Figure 1.

(a) X-ray structure of (tert-butyl)(adamantyl)B(pin) ate complex. CCDC: 2267241. (b) DFT-calculated energy surface for boron-to-copper transmetalation of either the tert-butyl group or the adamantyl group. (c) Three-dimensional structure of transition state complexes. (d) Analysis of the pyramidalization changes at carbon during transmetalation.

Encouraged by the results of DFT calculations, we sought a practical method for the synthesis of the corresponding adamantyl borate complexes. Adamantyllithium has been prepared by lithiation of adamantyl halides,¹⁴ however, this reaction requires low temperature and specialized lithium alloys. Additionally, the resulting adamantyllithium is a pyrophoric reagent that is unstable in many solvents and thus ill-suited for long term storage. As an alternative, we examined a catalytic modification of a stoichiometric in situ lithiation described by Sumida, Hosoya and Ohmiya.¹⁵ Thus, treatment of adamantyl chloride and the tertiary boronic ester 1 (Scheme 2) with 5 mol% 4,4’-di-tert-butylbiphenyl (DBB) and excess lithium metal resulted in formation of the borate complex within one hour at room temperature as indicated by ¹¹B NMR analysis.¹⁶ Of note, DBB functions as both a catalyst and as an indicator for the reaction process: upon complete consumption of adamantyl chloride, the reaction mixture takes on the dark blue color of the residual Li–DBB adduct. At this stage, the boronate complex can be separated from unreacted Li, the solvent removed, and the solid material stored for >one month with little change to the ¹¹B NMR spectrum.

Scheme 2.

Activation of Tertiary Alkylboronic Esters with Adamantyllithium Generated In Situ.

With ready access to adamantyl-activated tertiary boron ate complexes, their use in copper-catalyzed coupling reactions was investigated. In this process, a THF solution of the borate was added to 20 mol% copper cyanide followed by allyl electrophiles. For a variety of leaving groups, the desired allylation product 2 (Figure 2) was obtained in 25–60% yield (see Supporting Information) with only trace amounts of adamantyl coupling product being observed. Depending on the nature of the leaving group, moderate to high levels of stereospecificity (54–>98% es) were also observed in these reactions, however, the use of 2.0 equivalents of allyldiethyl phosphate resulted in optimal yield (64%) and consistently high levels of stereospecificity.

Figure 2.

Scope of Stereospecific Cu-Catalyzed Couplings of Tertiary Alkylboronic Esters with Electrophiles. Reactions were carried out with 0.20 mmol of alkylboronic esters. Yields are of isolated materials. Enantiomeric ratio was measured by chiral SFC analysis and have an error of ±1%. pdt:Ad–E, where measured, refers to the selectivity in coupling to the tertiary boronate versus coupling to the adamantyl group.

The scope of the coupling reaction was examined with a series of different allyl and propargyl electrophiles (Figure 2), which provided moderate to good yields of alkenes and allenes containing quaternary stereocenters. For product 2 and 12, the enantiospecificity was determined to be >98%. Of note, the Boc protecting group of product 9 survived the coupling sequence intact. Also of note, the borate species can engage in copper-catalyzed stereospecific amination reactions.¹⁷ Thus, enantiomerically-enriched amines can be prepared in moderate to good yield and good stereospecificity. The absolute configuration of 18 was assigned based on X-ray crystallography, which indicated that the process occurs with retention of configuration at carbon. For less reactive electrophiles such as acyl chlorides (products 21–24), β-haloenones (products 26–29), 2-iodopyridine (25), sulfonylcyanamide reagent 19 (product 20) and bromoalkynes (product 30), moderate yields can be obtained. For these electrophiles, high yielding coupling reactions occurred when more reactive cyclopropylboronic esters were employed. For several cases with more polar electrophiles (16, 26–28) it was possible to isolate the minor product of coupling to the adamantyl activator, and it was found that the group-selectivity ranged from 3:1 – >20:1. Of particular interest is product 27 where the [2.2.1]-bicycloheptyl group transfers more readily than the adamantyl group. DFT calculations on the substrate (see Supporting Information) show the B–[2.2.1]bicycloheptyl bond is shorter and presumably stronger than the B–adamantyl bond. In the ate complex, the bicycloheptyl carbon is more pyramidalized that the adamantyl carbon, however, in the transmetalation transition state structures the bicycloheptyl carbon undergoes greater pyramidalization suggesting substrate plasticity as more important element than bond strength in the reaction of this substrate.

A last point worth mention is that secondary boronic esters are suitable substrates for stereospecific adamantyl-activated couplings providing 31–33 via Cu-catalyzed coupling reactions, and 34–35 by transmetalation to Zn(II)^10b prior to coupling with electrophiles).

To examine aspects pertaining to synthesis utility, a gram-scale coupling reaction was conducted (Scheme 3). In this process, the desired product 2 could be isolated in 51% yield and >98% enantiospecificity. With larger quantities of coupling product, we examined transformations of the allylation adduct. It was found that the coupling product can engage in platinum-catalyzed diboration (→36),¹⁸ palladium-catalyzed Wacker oxidation (→37), ruthenium-catalyzed olefin metathesis (→38) and hydroboration reactions (→39), which furnished synthetically-useful building blocking containing a quaternary stereocenter.

Scheme 3.

Gram-Scale Coupling and Product Derivatizations.

To further examine the synthetic application of the above-described coupling reactions, we considered a general route for the synthesis of bicyclo[2.2.1]heptane derivatives 41. This framework has been suggested as a saturated bioisostere of meta disubstituted benzenes (40) due to the similar angle and distance of the two attachment points at the bridging carbon (Scheme 4a).¹⁹ An obvious need for the development of this class of compound is the selective and broad-scoped installation of functional groups at the tertiary carbons. Starting from inexpensive ketone 43 (Scheme 4b), Wittig olefination followed by diboration provided 44 in good yield. Ketal deprotection followed by homologation as inspired by Qin²⁰ selectively occurs at the primary boron to provide critical diboron scaffold 42 that was characterized by x-ray crystallography (Scheme 4c). Under standard conditions for copper-catalyzed amination of tertiary boronates described above, 45 was isolated in 48% yield. Further oxidation with sodium perborate generated 46, which should serve as a useful starting point for the construction of COX-2 inhibitor analogs (i.e. 47).²¹

Scheme 4.

Synthesis of Disubstituted Bicyclo[2.2.1]heptane as meta Benzene Bioisosteres.

In conclusion, a practical and mild method for activating tertiary alkylboronic esters was developed. The adamantyllithium-activated borate complex can engage in copper-catalyzed cross-coupling reactions with an array of electrophiles. Importantly, this reaction can be used for construction of challenging quaternary stereocenters. Further studies on the utility of these processes are in progress.