Enantioselective Ruthenium Catalyzed Benzocyclobutenone-Ketol Cycloaddition: Merging C-C Bond Activation and Transfer Hydrogenative Coupling for Type II Polyketide Construction

Abstract

The first enantioselective intermolecular metal catalyzed cycloadditions of benzocyclobutenones via C-C bond oxidative addition are described. In the presence of a ruthenium(0) complex modified by (R)-DM-SEGPHOS, tetralone-derived ketols and benzocyclobutenones combine to form cycloadducts with complete regio- and diastereoselectivity and high enantioselectivity. Using this method, the “bay region” substructure of the angucycline natural products arenimycin was prepared.

Graphical Abstract

Following pioneering work by Liebeskind,^1,2 Jun³ and Murakami,⁴ the insertion of π-bonds into metallacycles obtained through C-C bond oxidative addition has matured into an active area of research.⁵ Despite substantial progress in this field, several key challenges persist. The use of directing groups is often required to overcome the kinetic and thermodynamic barriers posed by the activation of relatively strong and stable C-C bonds. Many C-C bond activation initiated C-C couplings are restricted to intramolecular processes. Finally, beyond reactions of donor-acceptor cyclopropanes,^6,7 enantioselective intermolecular metal-catalyzed C-C bond activation initiated C-C couplings are exceptionally uncommon.⁸

In connection with studies of transfer hydrogenative C-C bond formation,⁹ we recently developed a novel class of ruthenium(0)-catalyzed cycloadditions of vicinal diols, ketols or diones.^9d,10 These processes operate through a common mechanistic motif wherein oxidative formation of ruthenium(II) metallacycles, which is accompanied or triggered by dione addition, is followed by diol- or ketol-mediated transfer hydrogenolysis of the ruthenacycle to release product and regenerate the reactive dione (Figure 1).^9d Recently, we found that ruthenacycles obtained via C-C bond oxidative addition to benzocyclobutenones are catalytically competent in transformations of this type, providing a kinetic pathway for the insertion of adjacent carbon atoms of vicinal diols, ketols or diones into C-C σ-bonds.¹⁰ⁱ Based on this finding, we now report the first examples of enantioselective intermolecular metal catalyzed cycloadditions of benzocyclobutenones.^11–14

Figure 1.

General catalytic mechanism for ruthenium(0) catalyzed cycloaddition of diols, ketols or diones.

In an initial set of experiments (Table 1), a series of benzocyclobutenone derivatives la-ld were exposed to equimolar quantities of the tetralone-derived ketols 2a-2c (or diols) in the presence of the mononuclear ruthenium(0) catalyst¹⁵ derived from Ru₃(CO)₁₂ and (R)-SEGPHOS in m-xylene (1 M) at 150 °C. The reaction of the unsubstituted benzocyclobutenone la with the unsubstituted diol 2a led to the formation of cycloadduct 3a with complete regio- and diastereoselectivity, but in nearly racemic form (Table 1, entry 1). However, using the corresponding ortho-methoxy-substituted benzocyclobutenone lb, cycloadduct 3b was formed with significant levels of enantiomeric enrichment (Table 1, entry 2). Based on this result, an effort was made to define substituents that would both enforce optimal levels of enantiomeric enrichment and facilitate entry to naturally occurring type II polyketides of the angucycline class (vide infra). Toward this end, incorporation of a triisopropylsilyl (TIPS) ether at C2 of the benzocyclobutenone, as in lc, and, for the ketol partner, a benzyloxy substituent at C8, as in 2c, conspired to improve the degree of asymmetric induction. Thus, the reaction of benzocyclobutenone lc and ketol 2c delivered cycloadduct 3h with complete regio- and diastereoselectivity as a 94:6 ratio of enantiomers (Table 1, entry 8). Finally, using (R)-DM-SEGPHOS as ligand and a slight excess of benzocyclobutenone lc, cycloadduct 3h was generated in 96% yield as a single regio- and diastereomer as a 97:3 ratio of enantiomers (Table 1, entry 9).

Table 1.

Selected optimization experiments in the enantioselective intermolecular ruthenium catalyzed cycloaddition of benzocyclobutenones.^a

entry	R1	R2	1a-1d	R3	2a-2c	3a-3h	yield (%), dr, rr	er
1b	H	H	1a	H	2a	3a	87, >20:1, >20:1	53:47
2b	OMe	H	1b	H	2a	3b	83, >20:1, >20:1	81:19
3	OMe	H	1b	OMe	2b	3c	77, >20:1, >20:1	58:42
4	OMe	H	1b	OBn	2c	3d	79, >20:1, >20:1	76:24
5b	OMe	OTIPS	1c	H	2a	3e	75, >20:1, >20:1	87:13
6	OMe	OTIPS	1c	OMe	2b	3f	78, >20:1, >20:1	89:11
7	OMe	OTBS	1d	OMe	2b	3g	74, >20:1, >20:1	80:20
8	OMe	OTIPS	1c	OBn	2c	3h	79, >20:1, >20:1	94:6
➡ 9c	OMe	OTIPS	1c	OBn	2c	3h	96, >20:1, >20:1	97:3

^aYields are of material isolated by silica gel chromatography. Enantioselectivity values were determined by HPLC analysis on a chiral stationary phase. (R)-SEGPHOS (Ar = phenyl).

^bReaction was conducted from the diol oxidation level, 1a-1c (200 mol%).

^c1c (130 mol%), (R)-DM-SEGPEIOS (Ar = 3,5-xylyl). See the Supporting Information for further experimental details.

The scope of the enantioselective ruthenium catalyzed benzocyclobutenone-ketol cycloaddition was evaluated (Table 2). Retaining key structural features required to enforce high levels of asymmetric induction, an otherwise diverse set of benzocyclobutenones 1c, 1e-1l, were reacted with ketol 2c to furnish cycloadducts 3h-3p. In each case, complete regio- and diastereoselectivity were accompanied by excellent levels of enantiomeric enrichment. The formation of 3o and 3p illustrates tolerance of activated aryl chlorides. Variation of the ketol partner also was explored. As demonstrated by the formation of cycloadducts 3q-3s, electron releasing substituents that reside in para-relationship with respect to the ketone improve enantioselectivity, precluding the need for an ortho-benzyloxy-group. For example, in the formation of cycloadduct 3r, a 99:1 enantiomeric ratio is observed. The formation of 3r also demonstrates compatibility with amine functional groups.

Table 2.

Enantioselective intermolecular ruthenium(0) catalyzed cycloaddition of benzocyclobutenones 1c, 1e-1l with ketols 2c-2f to form cycloadducts 3h-3s.^a

^aAll reactions were performed on a 0.20 mmol scale. Yield of material isolated by silica gel chromatography. Enantioselectivities were determined by chiral stationary phase HPLC analysis. See Supporting Information for further experimental details.

^bX-ray structure obtained after removal of silyl ether (3h′).

^c36 h.

Underscoring the utility of this method vis-à-vis construction of type II polyketides, in particular, the angucycline antibiotics arenimycins A-D, collinone and SF22446A1-A3/SF22446B1-B3,^16–19 benzocyclobutenone lc was exposed to ketol 2g under standard conditions for ruthenium-catalyzed cycloaddition (Scheme 1). The cycloadduct 3t was formed in good yield and incorporates the highly congested “bay region” motif found in arenimycin. To further demonstrate relevance to type II polyketide construction, cycloadduct 3h was subjected to silyl-deprotection followed by oxidation of the resulting alcohol to form the dione 4h (eq 1). Additionally, cycloadduct 3o was subjected to an SNAr reaction to form compound 4o, which incorporates a dimethylamino motif. Dimethylamino groups frequently occur in type II polyketide antibiotics(e.g. the tetracyclines) and, as recently demonstrated by Hergenrother, introduction of unhindered amines endows certain rigid “Gram-(+) only” antibacterials with Gram-(−) activity by conferring membrane permeability.²⁰

(eq. 1)

(eq. 2)

Scheme 1.

Arenimycins A-D and collinone, antibacterial angucyclines that incorporate a bridgehead diol motif, and reaction of benzocyclobutenone 1c with ketol 2g to form cycloadduct 3t^a.

^aFor structurally related angucycline natural products, SF22446A1-A3 and SF22446B1-B3, see Ref. 19. Reaction was performed on 0.20 mmol scale. Yield of material isolated by silica gel chromatography. Enantioselectivities were determined by chiral stationary phase E1PEC analysis. See Supporting Information for further experimental details.

Finally, the ability to achieve high levels of enantioselectivity in cycloadditions of chiral racemic benzocyclobutenones 1c, 1e-1l has implications regarding the nature of the C-C bond oxidative addition event. Fligh levels of enantiomeric enrichment persist at equimolar loadings of benzocyclobutenone and ketol (Table 1, entry 8) and do not depend on stoichiometry, which suggests kinetic resolution of benzocyclobutenones lc, le-11 is not operative. Rather, these data corroborate a catalytic mechanism wherein stereoablative cycloreversion to form a discrete α-oxo-ortho-quinodimethane or “ortho-quinoid ketene methide”²¹ is followed by enantiodetermining C-C bond oxidative addition to form a ruthenaindanone complex (Figure 2).²² The indicated stereochemical model for ruthenaindanone formation accounts for structural features of the benzocyclobutenone that are required to enforce high levels of enantioselectivity, specifically, non-bonded interactions between the methoxy and triisopropylsiloxy groups of the α-oxo-ortho-quinodimethane with the xylyl moieties of DM-SEGPHOS.

Figure 2.

Stereoablative cycloreversion precedes oxidative addition and stereochemical model for ruthenaindanone formation.

In summary, we report the first enantioselective intermolecular metal catalyzed cycloadditions of benzocyclobutenones via C-C bond oxidative addition. Using a mononuclear¹⁵ ruthenium(0) complex derived from Ru₃(CO)₁₂ and (R)-DM-SEGPHOS, tetralone-derived ketols and benzocyclobutenones react to form cycloadducts with complete regio- and diastereoselectivity and high enantioselectivity. Whereas the majority of methods for convergent type II polyketide construction require use of stoichiometric carbanions under cryogenic conditions (e.g. the Hauser-Kraus annulation,^23,24 aryne-mediated cycloadditions²⁵), the present method assembles type II polyketide motifs in the absence of stoichiometric byproducts under non-cryogenic conditions. In an initial synthetic application, this method was used to prepare the congested “bay region” substructure characteristic of the angucycline antibiotics arenimycin, collinone and SF22446A1-A3/SF22446B1-B3.^16–19 Future studies are aimed at defining ketol partners for the synthesis of linear tetracycline antibiotics.²⁶