Highly Stereoselective Synthesis of Tetrasubstituted Acyclic All-carbon Olefins via Enol Tosylation and Suzuki–Miyaura Coupling

Abstract

A highly stereocontrolled synthesis of tetrasubstituted acyclic all-carbon olefins has been developed via a stereoselective enolization and tosylate formation, followed by a palladium-catalyzed Suzuki–Miyaura cross-coupling of the tosylates and pinacol boronic esters in the presence of Pd(OAc)₂/RuPhos catalytic system. Both the enol tosylation and Suzuki–Miyaura coupling reactions tolerate an array of electronically and sterically diverse substituents and generate high yield and stereoselectivity of the olefin products. Judicious choice of substrate and coupling partner provides access to either E- or Z-olefin with excellent yield and stereochemical fidelity. Olefin isomerization was observed during the Suzuki–Miyaura coupling. However, under the optimized cross-coupling reaction conditions, the isomerization was suppressed to <5% in most cases. Mechanistic probes indicate that the olefin isomerization occurs via an intermediate, possibly a zwitterionic palladium carbenoid species.

Graphical Abstract

INTRODUCTION

Tetrasubstituted acyclic all-carbon olefins have attracted tremendous attention owing to their unique structural, physical, and electronic properties.¹ For example, they have been explored extensively for their potential use in molecular devices and liquid crystals² and are widely present in biologically active compounds such as anticancer agents tamoxifen,³ idoxifene,⁴ and etacstil,⁵ as well as the selective estrogen receptor degrader GDC-0810 (Figure 1).⁶ Moreover, stereodefined tetrasubstituted olefins are the foundation for numerous asymmetric reactions such as dihydroxylation,⁷ epoxidation,⁸ and hydrogenation⁹ to stereospecifically establish vicinal sp³-hybridized carbon centers.

Figure 1.

Biologically active compounds containing stereodefined tetrasubstituted olefins

Consequently, a wide variety of synthetic strategies to construct tetrasubstituted acyclic all-carbon olefins have been developed.^10,11 Classical synthetic methods including Wittig,¹² Horner–Wadsworth–Emmons (HWE),¹² Julia,¹³ Peterson,¹⁴ and McMurry¹⁵ reactions generally afford poor stereoselectivities. Eliminations of highly functionalized tertiary alcohols produce tetrasubstituted olefins;¹⁶ however, the synthesis of the alcohol precursors is often challenging, and multiple elimination pathways frequently result in erosion of stereocontrol.¹⁶ Carbometallation of internal alkynes is commonly employed but suffers from chronic lack of regiocontrol and functional group tolerance.¹⁷ Recent alternative approaches using Lewis-acid-based electrophiles to activate internal alkynes followed by further functionalization are limited by combinations of harsh reaction conditions, expensive reagents, and poor regioselectivities.¹⁸ Therefore, the development of a general, operationally simple, scalable, highly regio- and stereoselective preparation of tetrasubstituted acyclic all-carbon olefins, while a daunting challenge, will provide a useful new tool for synthetic organic chemists.

In support of our clinical and commercial development of GDC-0810, we were tasked to develop a highly efficient and stereoselective synthesis of tetrasubstituted acyclic all-carbon olefins. Herein we wish to report a highly viable method based on stereoselective enol tosylate formation and subsequent Suzuki–Miyaura coupling (Scheme 2).

Scheme 2.

Stereoselective Olefin Synthesis via Enol Tosylate Formation and Suzuki–Miyaura Coupling

RESULTS AND DISCUSSION

We initiated our studies by investigating the enolization and tosylate formation of commercially available 1,2-diphenylbutan-1-one (1a) by employing 2.0 equiv of MHMDS or MOt-Bu (M = Li, Na, K) in THF and quenching with 2.0 equiv of Ts₂O in DCM. Unfortunately, the reactions generated less than satisfactory conversions (<5–97%) and E/Z selectivities (23:77–82:18). It has been reported previously that a highly E-selective enol silylation of acyclic ketones mediated by LiHMDS/R₃N in toluene was readily achieved in which ≥2.0 equiv of LiHMDS was required for maximal E/Z selectivity via an 8-membered transition state (eq 1).¹⁹

(1)

We thus focused on the enolization and tosylate formation of ketone 1a using 2.0 equiv of LiHMDS²⁰ with a tertiary alkyl amine (5.0 equiv) in toluene and quenching with 2.0 equiv of Ts₂O in DCM. As shown in Table 1, despite reasonable conversions (87–97%), reactions with bulkier amines tend to form more side products, which lowers the yield (Table 1, entries 1–3). Also, the chelating diamine base TMEDA only affords 44:56 ratio of the E/Z stereoisomers (Table 1, entry 4). With Me₂NEt being the most promising amine, the stoichiometry of Me₂NEt was then examined (Table 1, entries 5–9). Two equiv of Me₂NEt were optimal, affording quantitative conversion and 99:1 E/Z ratio. At <2.0 equiv of Me₂NEt the E/Z selectivity remained high, but the yield suffered (Table 1, entries 8–9). A preparative scale reaction using 10.0 g (44.6 mmol) of ketone 1a under optimized conditions employing 2.0 equiv of LiHMDS and 2.0 equiv of Me₂NEt at 23 °C afforded pure E-isomer in 88% yield (Table 1, entry 7). The structure of E-isomer 2a was determined by standard analytical methods including X-ray crystallography (Supporting Information).

Table 1.

Optimization of LiHMDS-mediated Enolization and Tosylation^a

entry	R3N	equiv of R3N	convb (%)	yieldc (%)	2a/2a′b
1	Cy2NEt	5	97	90	99:1
2	i-Pr2NEt	5	87	33	96:4
3	Et3N	5	97	71	98:2
4	TMEDA	5	97	74	44:56
5	Me2NEt	5	100	96	99:1
6	Me2NEt	3	100	95	99:1
7	Me2NEt	2	100	96(88)d	99:1
8	Me2NEt	1	99	84	98:2
9	Me2NEt	0	75	56	97:3

^aStandard conditions: ketone 1a (1.0 mmol), LiHMDS in PhMe (0.90 M, 2.0 equiv, 2.2 mL), R₃N (0–5 equiv), PhMe (1.0 mL), 23 °C; then Ts₂O (2.0 equiv) in DCM (5.0 mL), 23 °C.

^bConversion and E/Z ratio were determined by HPLC analysis of the reaction mixture.

^cAssay yield of desired 2a was determined by quantitative HPLC analysis. The number in parentheses is the isolated yield at both 1.0 mmol and 10.0 g (44.6 mmol) scale.

We next investigated the scope and limitations of the enolization and tosylation reactions using the optimized conditions (Table 2). The aryl group directly connected to the carbonyl moiety tolerated electron-donating (4-Me, 4-MeO, 4-Cl) and electron-withdrawing (4-F, 4-CF₃, 4-CO₂Et) substituents. All reactions afforded >95:5 E/Z selectivities and 66%–87% isolated yields of the E-tosylate product (2b–g). Substitution of a 3-MeO group did not significantly impact the E/Z selectivity and yield (2h, 98:2, 87%). However, a 2-Me group substantially eroded the E/Z selectivity to 82:18, affording E-isomer 2i and Z-isomer 2i′ in 61% and 12% yield, respectively, after chromatographic purification. The 2-pyridyl substituted ketone afforded 98:2 E/Z selectivity and 82% isolated yield of the desired tosylate 2j.

Table 2.

Reaction Scope of Enol Tosylation^a,^b,^c

^aStandard conditions: ketone 1 (2.0 mmol), LiHMDS in PhMe (0.90 M, 2.0 equiv, 4.4 mL), Me₂NEt (0.43 mL, 2.0 equiv), PhMe (2.0 mL), 23 °C; then Ts₂O (2.0 equiv) in DCM (10.0 mL), 23 °C.

^bE/Z ratio was determined by HPLC analysis of the reaction mixture.

^cIsolated yield.

^dThe reaction was also performed employing 10.0 g (44.6 mmol) of ketone 1a.

^e12% of the Z-isomer 2i′ was isolated.

^f10% of the Z-isomer 2t′ was isolated.

The aryl group distal to the carbonyl also tolerated electronically diverse substituents and afforded the tosylates 2k–q in 95:5–99:1 E/Z selectivity and 75–95% isolated yields. Interestingly, the 2-Me substituent had a less profound effect on the E/Z selectivity (2r, 95:5) than in the case of 2i, albeit with a lower isolated yield (50%). Surprisingly, the 2-pyridyl substituted ketone generated a 2:98 E/Z selectivity favoring the unexpected Z-isomer 2s′ in 61% isolated yield. We attribute this reversal in selectivity to lithium chelation by the 2-pyridyl nitrogen forcing Z-enolate formation (eq 2).

(2)

The enol tosylation reaction is sensitive to steric changes on the alkyl moiety proximal to the enolization site. Replacing the ethyl group with a smaller methyl group surprisingly afforded lower conversion and E/Z selectivity (83:17). After purification, only 47% of the E-tosylate 2t was isolated, along with 10% of the Z-tosylate 2t′. The hindered isopropyl moiety in 1u afforded no product. Fortunately, the allyl, cyclopropylmethyl, and n-butyl substituted ketones readily afforded the desired tosylates in high yields and selectivities (2v–x). X-ray crystallographic analysis of tosylates 2 unambiguously established the E-olefin configuration (Supporting Information). A few representative X-ray structures are shown in Figure 2.

Figure 2.

X-ray structures of E-tosylates 2a, 2e (top), 2m and 2v (bottom)

Next, we examined the Suzuki–Miyaura coupling²¹ of E-tosylate 2a and 4-fluorophenyl pinacol boronic ester (3a) by high-throughput screening of phosphine ligands using Pd(OAc)₂ as the precatalyst and K₃PO₄•H₂O as the base in PhMe/H₂O (3:1) at 70 °C. CMPhos,²² Xantphos,²³ and RuPhos²⁴ generated the highest conversions and lowest E→Z olefin isomerization (Figure 3).²⁵ Further validation and optimization led to the optimal conditions using 1.0 equiv of E-tosylate 2a, 1.1 equiv of boronic ester 3a, 1 mol % of Pd(OAc)₂, 2 mol % of RuPhos, and 1.5 equiv of K₃PO₄•H₂O in PhMe/H₂O (0.5 M, 3:1) at 70 °C (eq 3). Under this set of conditions, the reaction afforded complete conversion with only 2% of isomeric Z-olefin observed in the crude product. The E- and Z-isomers were inseparable by silica gel flash column chromatography. Thus, the desired tetrasubstituted olefin 4a containing 2% of the Z-stereoisomer was isolated in 98% yield on a 5.0 g (12.7 mmol) scale (Table 3, entry 1). 4-Fluorophenylboronic acid and potassium 4-fluorophenyl trifluoroboronate could be used in place of boronic ester 3a, affording the same 98% yield and 98:2 E/Z ratio.

Figure 3.

High–throughput screening of Suzuki–Miyaura coupling of E-tosylate 2a and 4-FC₆H₄Bpin (3a)

Table 3.

Scope of Boronic Esters in the Suzuki–Miyaura Reaction^a,^b,^c

^aStandard conditions: E-tosylate 2a (1.0 mmol), boronic ester (1.1 equiv), Pd(OAc)₂ (1 mol %), RuPhos (2 mol %), K₃PO₄•H₂O (1.5 equiv), PhMe/H₂O (1.5/0.5 mL), 70 °C.

^bE/Z ratio were determined by HPLC analysis of the isolated inseparable products.

^cIsolated yield.

^dThe reaction was also performed employing 5.0 g (12.7 mmol) of E-tosylate 2a.

^ePd(OAc)₂ (2 mol %) and RuPhos (4 mol %) were employed.

^f2.1 equiv of boronic ester was employed.

^gNegishi conditions: E-tosylate 2a (1.0 mmol), cyclopropylzinc bromide (1.1 equiv, 0.50 M in THF, 2 mL), LiCl (0.50 M in THF, 2.2 mL), Pd(OAc)₂ (1 mol %), RuPhos (2 mol %), 60 °C.

^h1.1 equiv of n-butylboronic acid was used.

(3)

E-Tosylate 2a was coupled to a variety of electronically and sterically diverse boronic esters (3) under the optimized reaction conditions (Table 3). Phenylboronic esters substituted at the 4-position with either an electron-donating group (Me, MeO) or electron-withdrawing group (CO₂Et, CF₃) generated the cross-coupling products (4b–e) in excellent yields and E/Z selectivities. Reaction of 4-chlorophenyl pinacol boronic ester with E-tosylate 2a afforded 4f in 99:1 E/Z ratio but in a disappointing 30% yield (46% conversion) of the isolated product. We suspect that 4-chlorophenyl pinacol boronic ester oligomerizes under our standard reaction conditions. The sterically demanding 2-tolyl pinacol boronic ester required 2 mol % of Pd(OAc)₂ and 4 mol % of RuPhos to reach full conversion but still afforded 86%yield and 98:2 E/Z ratio of product 4g. Highly hindered mesityl-derived pinacol boronic ester provided essentially no product 4h. Similarly, 2-pyridyl pinacol boronic ester did not afford much of the coupling product 4i, presumably due to the propensity of protodeboronation of the boronic ester under the reaction conditions.²⁶ By contrast, 3-thienyl and 5-indazolyl pinacol boronic esters coupled to give 4j–k in moderate to good yields and excellent isomeric ratios. Vinyl, alkynyl, cyclopropyl and alkyl boronic esters were also examined. As expected, styryl pinacol boronic ester produced 80% yield and 98:2 ratio of the coupling product 4l. The reaction with phenylethynyl pinacol boronic ester, however, required 2.1 equiv to reach completion, affording 61% yield and 99:1 selectivity of 4m. Cyclopropyl pinacol boronic ester failed to generate the desired olefin product possibly due to decomposition during the reaction. Gratifyingly, when E-tosylate 2a was treated with cyclopropylzinc bromide under Negishi coupling²⁷ conditions, the reaction produced 95% yield and 99:1 selectivity of olefin 4n. Although n-butyl pinacol boronic ester did not afford significant yield (<20%) of the coupling product, n-butylboronic acid underwent the Suzuki–Miyaura coupling smoothly, affording 98% yield and 99:1 selectivity of olefin 4o.

We examined the reactivity of a variety of E-tosylates (2) with phenyl pinacol boronic ester as a coupling partner (Table 4). The aryl group proximal to the tosylate can tolerate both electron-donating (Me, MeO, Cl) and electron-withdrawing (F, CO₂Et, CF₃) substituents at the 4-position as well as some steric congestion at the 2-position (5a–g), affording excellent yields and stereochemical integrity of the desired olefin products. E-tosylate with a 2-pyridyl moiety afforded 5h in 65% yield and 96:4 selectivity. Olefins 5a–g are the complementary stereoisomers to olefins 4a–g in Table 3. It is noteworthy that the modular nature of this methodology provides both isomers with excellent stereocontrol. Similarly, couplings with 4-fluorophenyl pinacol boronic ester showed that the aryl moiety distal to the tosylate also tolerates electron-donating and electron-withdrawing substituents at the 4-position (5i–m) as well as steric congestion at the 2-position (5n). Replacing the ethyl group with methyl, allyl, cyclopropylmethyl, or n-butyl moieties (5o–r) caused no erosion in yields or selectivities.

Table 4.

Scope of E-tosylates in the Suzuki–Miyaura Reaction^a,^b,^c

^aStandard conditions: E-tosylate 2a (1.0 mmol), boronic ester (1.1 equiv), Pd(OAc)₂ (1 mol %), RuPhos (2 mol %), K₃PO₄•H₂O (1.5 equiv), PhMe/H₂O (1.5/0.5 mL), 70 °C.

^bE/Z ratios were determined by HPLC analysis of the isolated inseparable products.

^cIsolated yield.

The pyridyl-containing Z-tosylate 2s′ unexpectedly generated in the enol tosylate reaction underwent the Suzuki–Miyaura reaction with 4-fluorophenyl pinacol boronic ester, forming the tetrasubstituted olefin 5s in a moderate 57% yield with <1% olefin isomerization (eq 4).

(4)

X-ray crystal structures of several tetrasubstituted olefins unambiguously established the geometry of the double bonds. Representative X-ray structures are shown in Figure 4.

Figure 4.

X-ray structures of olefins 4a, 5d (top), 5j and 5o (bottom)

Olefin isomerizations have been previously noted in Suzuki–Miyaura couplings, but little is known about the cause of this side-reaction.²⁵ We investigated the isomerization mechanism by examining the cross-coupling reaction of E-tosylate 2a and 4-fluorophenyl pinacol boronic ester (3a). Subjecting product 4a (E/Z = 98:2) to the standard coupling conditions caused no change in the E/Z ratio after 20 h, which indicates that the olefin product does not isomerize under the reaction conditions. Exclusion of light or addition of 2,6-di-tert-butyl-4-methylphenol (BHT, 5 mol %) to the coupling reaction had no impact on the E/Z selectivity, arguing against a free radical mechanism. Treating either E-tosylate 2a or its Z-isomer 2a′ with Pd(OAc)₂ omitting either RuPhos or the boronic ester or both showed no isomerization. Thus, we deem that a Pd(II)-catalyzed isomerization is an unlikely mechanism. Treating E-tosylate 2a or its Z-isomer 2a′ with low concentration (10 mol %) of boronic ester 3a using either Pd(0) generated from Pd(OAc)₂ and RuPhos or RuPhos-Pd-G₃²⁸ showed that the tosylates do not isomerize.²⁹

Based on the above observations, we propose a plausible mechanism for olefin isomerization shown in Scheme 3. Similar to a mechanistic proposal published by Lipshutz,³⁰ the E-tosylate 2a undergoes a stereospecific Pd(0) insertion to give Pd(II) intermediate Pd-1, which transmetalates with the aryl boronic ester to form intermediate Pd-2. A subsequent tautomerization of Pd-2 to zwitterionic palladium carbenoid species³¹ Pd-3 or Pd-3′ erodes the stereointegrity to generate intermediate Pd-4. Intermediates Pd-2 and Pd-4 reductively eliminate to regenerate Pd(0) and produce the desired olefin 4 and its stereoisomer 4′, respectively.

Scheme 3.

Proposed Mechanism of Olefin Isomerization during Suzuki–Miyaura Coupling

CONCLUSION

We have developed a highly stereoselective synthesis of tetrasubstituted acyclic all-carbon olefins via a stereoselective enolization and tosylate formation, followed by a palladium-catalyzed Suzuki–Miyaura cross-coupling with pinacol boronic esters in the presence of Pd(OAc)₂/RuPhos catalytic system. Both the enol tosylation and Suzuki–Miyaura coupling reactions tolerate an array of electronically and sterically diverse substituents. Judicious choice of substrate and coupling partner provides access to either E- or Z-olefin with excellent yield and stereochemical fidelity. Mechanistic probes show erosion of stereointegrity occurs subsequent to transmetalation through a zwitterionic palladium carbenoid species. Isomerization was minimal in most cases under optimized cross-coupling conditions.

Scheme 1.

Selected Strategies to Prepare Tetrasubstituted Olefins