Catalytic conjunctive cross-coupling enabled by metal-induced metallate rearrangement

See allHide authors and affiliations

Science  01 Jan 2016:
Vol. 351, Issue 6268, pp. 70-74
DOI: 10.1126/science.aad6080

A two-for-one twist on Suzuki coupling

The Suzuki-Miyaura coupling reaction is one of the most widely used ways of making carbon-carbon bonds. Essentially a palladium catalyst activates one carbon fragment and then links it to a second fragment pulled from boron. Zhang et al. now demonstrate a twist on the conventional pathway (see the Perspective by Fyfe and Watson). In their system, the palladium initially coaxes together two carbon fragments on one boron center. Then the catalyst stitches a second C-C bond to a third, external fragment. A chiral ligand renders the reaction highly enantio-selective.

Science, this issue p. 70; see also p. 26


Transition metal catalysis plays a central role in contemporary organic synthesis. Considering the tremendously broad array of transition metal–catalyzed transformations, it is remarkable that the underlying elementary reaction steps are relatively few in number. Here, we describe an alternative to the organometallic transmetallation step that is common in many metal-catalyzed reactions, such as Suzuki-Miyaura coupling. Specifically, we demonstrate that vinyl boronic ester ate complexes, prepared by combining organoboronates and organolithium reagents, engage in palladium-induced metallate rearrangement wherein 1,2-migration of an alkyl or aryl group from boron to the vinyl α-carbon occurs concomitantly with C–Pd σ-bond formation. This elementary reaction enables a powerful cross-coupling reaction in which a chiral Pd catalyst merges three simple starting materials—an organolithium, an organoboronic ester, and an organotriflate—into chiral organoboronic esters with high enantioselectivity.

Organoboronic acids and their derivatives are widely available and broadly useful starting materials for organic synthesis (1). In addition to being environmentally benign and generally inexpensive, these reagents exhibit a near-ideal balance of stability and reactivity. Although chemically and configurationally stable, organoboronic esters engage in a broad array of carbon-carbon and carbon-heteroatom bond-forming processes upon activation. The most commonly practiced such reaction is the transition metal–catalyzed Suzuki-Miyaura cross-coupling reaction between organic electrophiles and organoboron compounds (2). In broad strokes, the mechanism of the Suzuki-Miyaura reaction involves a sequence of (i) oxidative addition between a metal catalyst and the electrophile, (ii) transmetallation with the organoboron reagent, and (iii) reductive elimination of the C-C bonded product (3). Here, we used an alternative pathway to the organoboron transmetallation step. The overall putative catalytic cycle enables a class of organoboron cross-coupling that we term “conjunctive cross-coupling” (Fig. 1A) because it merges two nucleophilic reagents into one during the course of the reaction. Overall, the conjunctive cross-coupling constructs chiral products by merging three simple starting materials: an organolithium reagent, an organoboronic ester, and an organic electrophile. The reaction, which adds to recent catalytic alkene carboboration reactions (4, 5), establishes two new carbon-carbon bonds and forms a stereogenic center bearing a useful organoboronic ester with high enantioselectivity.

Fig. 1 Metal-induced metallate shift as a strategy for catalytic reaction design.

(A) The catalytic conjunctive cross-coupling process. (B) Common metallate shift to saturated carbon centers requires a leaving group. L, ligand. (C) Metallate shift to unsaturated carbons is often activated by addition of an external electrophile. M, metal; E, electrophile. (D) A metallate shift promoted by a metal complex can serve as an alternative to a transmetallation process. (E) Proposed catalytic cycle for conjunctive cross-coupling. Ar, aryl group.

Organoboronic esters are known to participate in a wide range of reactions that occur by stereospecific 1,2-metallate rearrangements (6, 7). The preponderance of these reactions occur with a four-coordinate anionic boron-centered “ate” complex (Fig. 1B) that rearranges by a 1,2-carbon shift from boron to an adjacent sp3-hybridized electrophilic center (designated “A”) bearing an attached leaving group (X). The stereoelectronic requirements of the metallate shift dictate an anti-periplanar arrangement of the migrating carbon atom (R) and the leaving group such that the overall process is stereoretentive at R but stereoinvertive at A (8). These processes are common (9) for transformations where the “A” group is a carbon, nitrogen, or oxygen atom and less common but still established for sulfur and phosphorus (10, 11). Because of the stereospecific nature of the above-described processes, stereoselectivity in metallate shifts is generally subject to substrate control (12), although Jadhav and Man reported an example with selectivity dictated by a chiral catalyst (13). Metallate rearrangement from boron to sp (14) and sp2 (15) hybridized carbons often require the addition of an external electrophilic activator (Fig. 1C), and generally the 1,2-metallate shift is followed by elimination to reestablish unsaturation (16). In this context, we considered that in place of stoichiometric electrophilic activating agents, π-acidic late transition metals might similarly promote the 1,2-metallate shift of alkenyl boronates (Fig. 1D) and that the resulting chiral organometallic intermediate might be used in subsequent bond-forming processes.

In one embodiment of catalysis based on metal-induced 1,2-metallate rearrangements, we considered that an electrophilic palladium complex (II, Fig. 1E), generated by oxidative addition of I with an organic electrophile, might induce 1,2-migration in a vinyl boronate-derived ate complex (IIIIV) and establish a new C–C bond and a boron-substituted stereogenic center. Subsequent reductive elimination (IVI) would serve to establish a second C–C bond, release the product, and concomitantly furnish a reduced Pd complex that might continue a catalytic cycle. The net reaction, in this cycle, is aligned with important work from Murakami, who studied reactions of alkynyl boron ate complexes (17). Although π-acidic late transition metal complexes are well known to activate alkenes for nucleophilic attack, nucleopalladation with Pd(II) complexes generated by oxidative addition reactions are less common (18, 19). For the envisioned process to be successful, several key questions emerged: Would the Pd(II) aryl complexes be sufficiently π-acidic to facilitate metallate rearrangement (IIIIV)? Would common direct transmetallation (IIIV) dominate the reaction and furnish Suzuki-Miyaura products? Could facial selectivity in olefin binding render the migration (IIIIV) enantioselective? Last, would β-hydrogen elimination in intermediates such as IV compete with reductive elimination?

To begin our studies of the conjunctive cross-coupling reaction, we selected phenyl triflate as the electrophile, anticipating that the outer-sphere triflate anion would leave open a coordination site on Pd for alkene binding (Fig. 1E, III). In an effort to minimize steric penalties that might inhibit metallate shift, we opted for an unsubstituted vinyl group in construction of the boron ate complex. Similarly, on the basis of recent studies from Mayr and Aggarwal (20, 21), we selected a neo-pentylglycolato ligand for boron because the ate complexes derived from this ligand (versus others readily available) were determined to be the most nucleophilic. Last, in an effort to favor reductive elimination relative to β-hydrogen elimination in putative intermediate IV, we selected ligands with large bite angles for the palladium complex. Initial experiments with achiral bidentate ferrocenyl 1,1′-diphosphines were promising and suggested that the conjunctive cross-coupling could indeed operate. A survey of a number of chiral diphosphine structures revealed that Josiphos (22) and MandyPhos (Solvias AG, Kaiseraugst, Switzerland) (23) are particularly effective ligand classes, with the MandyPhos ligand L1 providing an outstanding level of enantioselectivity and very good catalyst efficiency in the reaction.

In optimizing the reaction conditions, we found tetrahydrofuran (THF) to be the most effective solvent. Therefore, when the migratory R group was appended to boron by addition of hydrocarbon solutions of organolithium reagents to the neopentyl glycol–derived vinyl boronic ester vinylB(neo) (Fig. 2), the resulting ate complexes were evaporated to dryness and redissolved in THF before reaction. The efficiency of the reaction was also greatly diminished by chloride, bromide, or iodide ions, an obstacle easily overcome by the continued use of aryl and alkenyl triflates as the electrophile. For coupling of alkenyl triflate electrophiles, the reaction selectivity was markedly enhanced with boron ate complexes derived from pinacolato ligands in place of the neo-pentylglycol derivative [e.g., product 15 is formed in 53% yield and 82:18 enantiomeric ratio (er) when using the neo-pentylglycol ligand], whereas for aryl triflates, the neo-pentylglycol ligand furnished higher selectivity [1 formed in 94% yield, 93:7 er using B(pin) group]. With these features optimized, the scope of the conjunctive coupling was explored with an array of ate complexes and electrophiles. During these experiments, it was most convenient to oxidize the product organoboronic ester to the derived alcohol by treatment with NaOH and H2O2; however, isolation of the organoboron product itself is also possible (the organoboron precursor to alcohol 1 was isolated in 76% yield by silica gel chromatography). As shown in Fig. 2, both aryl and alkyl groups proved competent migrating elements in conjunctive coupling reactions. In cases where the yield of product is low, analysis of the reaction mixture before oxidation showed that by-products generally consist of recovered organoboronic ester (likely generated during workup by protonolysis of one carbon ligand from the ate complex) and direct Suzuki-Miyaura products. Primary and secondary alkyl groups migrate. as do functionalized alkyl appendages [e.g., (trimethylsilyl)methyl]. Conjunctive couplings were also effective for both electron-rich and electron-poor electrophiles. Moderately encumbered electrophiles, such as ortho-substituted arenes, can engage in the reaction, although the more highly substituted 2,6-dimethylphenyl derivative suffered from lower selectivity. With respect to preparation of chiral hydrocarbon frameworks, a range of substituted alkenyl triflates participate, and the configuration of the product alkene directly reflects that of the precursor electrophile.

Fig. 2 Scope of the catalytic conjunctive cross-coupling reaction starting from vinyl boronic esters.

Yields represent isolated yield of purified material and are an average of two experiments. The absolute configuration was determined by anomalous dispersion x-ray and by chromatographic comparison to known compounds. *VinylB(neo) was replaced with vinylB(pin); †reaction conducted at 80°C. er was determined by chiral supercritical fluid chromatography. OAc, acetate; Me, methyl group; Ph, phenyl group; OTf, trifluoromethanesulfonate; OTBS, tert-butyldimethylsiloxy.

The ate complex could be generated either by addition of an organolithium reagent to vinylB(neo) (Fig. 2) or by addition of vinyllithium to organoboronic esters (Fig. 3). Considering the broad array of organoboronic esters that are already available for use in common cross-coupling processes, the latter strategy is particularly enabling. In this context, we found that the presence of lithium halide salts, even at 1 to 2 mole percent (mol %) loading, markedly erodes conjunctive coupling efficiency. Thus, to perform the reaction with highest efficiency with 2 mol % catalyst loading, it is critical that halide-free vinyllithium be used. To prepare this reagent, we developed a procedure that involved lithium-halogen exchange with n-BuLi (Bu, butyl group) in hexane, followed by low-temperature recrystallization of pure vinyllithium. When these precautions are taken, conducting the reaction as in Fig. 3 allowed the scope of the migrating group to be surveyed more completely and revealed that, although reactions requiring migration of very electron-deficient groups are still a challenge, electron-neutral and electron-rich arenes can engage in the migration, as can those that are more highly substituted (i.e., 2,6-disubstituted arenes). We also found that, with 5 mol % catalyst loading, vinyllithium prepared as above but without recrystallization was effective (e.g., 3 formed in 69% yield, 98:2 er) and that vinyllithium prepared by lithium-tin exchange (24) could be used directly (2 mol % catalyst, 1 formed in 83% yield, 97:3 er).

Fig. 3 Scope of the catalytic conjunctive cross-coupling reaction starting from alkyl or aryl boronic esters.

Yields represent isolated yield of purified material and are an average of two experiments. Cy, cyclohexyl group.

To probe the utility of conjunctive cross-coupling to practical organic synthesis, we were attracted to the natural product (–)-combretastatin (25), a member of a family of cytotoxic stilbene-derived natural products that bind β-tubulin (Fig. 4A). Although many synthetic methods have facilitated the construction of combretastatins (26), alternative processes may engage different starting materials and thereby provide access to distinct analogs. In the case of conjunctive cross-coupling, the requisite boronic ester and electrophile are readily available and, as depicted in Fig. 4A, are readily converted to the coupled product in an efficient and highly selective fashion. Removal of the silicon protecting group furnished the target structure, spectra of which were fully consistent with the natural isolate.

Fig. 4 Synthesis and mechanistic studies.

(A) Synthesis of combretastatin by conjunctive cross-coupling. (B) The stereochemical course of metal-induced metallate rearrangement. dr, diastereomeric ratio.

The mechanism of the conjunctive coupling reaction is the subject of ongoing investigations, although the substrate scope (Figs. 2 and 3) gives clues about the process. The observation that electron-deficient arenes are less prone to migration is consistent with the mechanistic hypothesis put forward in Fig. 1E. According to this hypothesis, formation of IV would likely be stereochemistry determining, and, in line with this prediction, the selectivity of the reaction depends not only on the ligand framework but also on the organoboronic ester ligand (pinacol versus neo-pentylglycol), the migrating group, and the electrophile. In addition to these observations, one preliminary experiment sheds important light on the nature of the metal-induced metallate rearrangement that appears to underlie the conjunctive coupling process. As depicted in Fig. 4B, when the reacting ate complex was constructed from stereochemically defined deuterium-labeled vinyllithium (27) and phenylB(pin), the (1R,2R) stereoisomer of the conjunctive coupling product was formed in >20:1 diastereoselection (82:18 er). Although other interpretations are possible, should the mechanism be in line with that proposed in Fig. 1E and reductive elimination occur with retention of configuration at carbon (a reasonable assumption), the observed stereochemical outcome in Fig. 4B is consistent with anti-migration of the arene group to a Pd-olefin complex (Fig. 1D). Such an outcome is reminiscent of nucleometallation reactions that do not involve preassociation of the migrating group and the metal center (28).

We anticipate that many other transition metal–catalyzed reactions might also be reengineered to incorporate metal-induced metallate rearrangements, thereby providing distinct strategies for catalytic enantioselective synthesis.

Supplementary Materials

Materials and Methods

Tables S1 to S7

References (3057)

References and Notes

  1. For lead references, see (29).
Acknowledgments: Experimental data are available in the associated Supplementary Materials. This research was supported in part by the NIH, National Institute of General Medical Sciences (GM 64451), and by Boston College. Metrical parameters for the crystal structure of compound 1 are available free of charge from the Cambridge Crystallographic Data Centre under accession number CCDC 1437520.
View Abstract

Stay Connected to Science

Navigate This Article