Enantioselective Trifunctionalization of Terminal Alkynes

Abstract

The stereoselective synthesis of alkenes has been one of the central objectives in organic chemistry. Significant advances in synthetic methodology have made mono- and disubstituted alkenes widely accessible from a variety of readily available precursors. However, increased steric hindrance in more highly substituted alkenes limits the effectiveness of these methods, and as a result, efficient and selective synthesis of highly substituted alkenes remains a formidable challenge. Here, we demonstrate palladium-catalyzed trifunctionalization of terminal alkynes, using organoboranes and allylic carbonates as coupling partners. This transformation provides tetrasubstituted alkenes with excellent regio- and diastereoselectivity. Moreover, regiodivergent, diastereo- and enantioselective incorporation of the allylic fragment provides access to a wide range of complex 1,4-diene products. We present evidence that the palladium catalyst controls the selectivity of the tetrasubstituted alkene formation and the selectivity of the allylic substitution.

Graphical Abstract

INTRODUCTION

Tri- and tetra-substituted alkenes are prevalent in bioactive molecules¹ and materials.² They also serve as versatile substrates in asymmetric transformations for stereospecific construction of vicinal chiral centers.³ Relative to mono- and disubstituted alkenes, highly substituted alkenes exhibit increased steric congestion and severe eclipsing interactions. These features significantly limit the efficiency, scope, and stereoselectivity of traditional olefination methodologies,⁴ such as carbonyl olefination⁵ and alkene metathesis⁶ in synthesis of highly substituted alkenes. Currently, the most versatile strategy for accessing tetrasubstituted alkenes is based on difunctionalization of internal alkynes, which has seen rapid development in the last decade.^7,8 However, the key carbometallation step in these reactions lacks inherent regioselectivity and the reactions require substrates with an electronically,⁹ sterically,^9a,10 or coordinatively¹¹ biased alkyne limiting their scope. Overall, further development of methods for selective synthesis of highly substituted alkenes from readily available starting materials remains necessary.

An intriguing approach to the synthesis of highly substituted alkenes involves electrophile-triggered 1,2-metalate shift of alkynyl boron-ate complexes, as first demonstrated by Brown¹² and Pelter¹³ in 1970s. Although providing efficient access to tetrasubstituted alkenes, this approach initially afforded poor diastereocontrol.^13a,14 Good diastereoselectivity was limited to reactions forming cyclic products¹⁵ or reactions with a narrow range of heteroatomic electrophiles.¹⁶ In 2007, Murakami et al. addressed these limitations by introducing a Pd-catalyzed coupling of alkynyl boron-ate complexes with aryl halides.¹⁷ More recently, Aggarwal reported an exceptionally versatile metal-free alkylation of alkynyl boron-ate complexes,¹⁸ with the two approaches delivering excellent and complementary diastereoselectivity (Scheme 1c).

Scheme 1.

Synthesis of Tetrasubstituted Alkenes and 1,2-Metallate Shift of Alkynyl Boron-ate Complexes

Building on the unique reactivity of alkynyl boron-ate complexes, we recently introduced direct metal-catalyzed trifunctionalization of terminal alkynes through electrophilic functionalization of alkynyl boron-ate intermediates formed in situ from alkynes and organoboranes (Scheme 1c).¹⁹ This process provides a highly convergent strategy for the synthesis of tetrasubstituted alkenes and eliminates the need for stoichiometric organolithium reagents used in the formation of alkynyl boron-ate complexes. At the same time, the use of transition metal catalysts to initiate the 1,2-metallate allowed us to expand the scope of electrophiles in functionalization of alkynyl boron-ate complexes. We recognized that the same approach can be used to develop an enantioselective trifunctionalization of terminal alkynes by using prochiral electrophiles. Intriguingly, the same transition metal catalyst that promotes 1,2-metalate shift could then also be used to control the enantioselectivity of the overall process that would afford highly complex chiral molecules featuring tri- and tetrasubstituted alkenes.

In this work, we present a palladium-catalyzed direct trifunctionalization of terminal alkynes using organoboranes and allylic carbonates as coupling partners (Scheme 1d). The new method enables efficient synthesis of tri- and tetrasubstituted alkenes with excellent regio- and diastereoselectivity. In addition to promoting 1,2-metallate rearrangement of the alkynyl boron-ate intermediate, the palladium catalyst plays a crucial role in controlling the substitution of allylic carbonates. Notably, the choice of ligand influences the regioselectivity of the allylation, enabling regiodivergent access to either linear products with excellent E-selectivity or branched products with high enantioselectivity. Finally, we also show that the alkenyl borane products can be further functionalized into a variety of tetrasubstituted alkenes without isolation or purification.

RESULTS AND DISCUSSION

Initial Results and Reaction Development.

Seeking to incorporate complex prochiral electrophiles into direct trifunctionalization of alkynes, we were drawn to allylic electrophiles. While the enantioselective allylations of various boron-ate complexes are known,²⁰ reactions of preformed alkynyl boron-ate complexes with allylic electrophiles have been mired by low diastereoselectivity and only the products of the linear allylic substitution have be obtained.²¹ Hoping to address these issues and control regio-, diastereo-, and enantioselectivity of the allylic substitution, we explored the reaction of terminal alkyne 1, organoborane 2, and cinnamyl bromide 3 in the presence of transition metal catalysts (Scheme 2). Various Ir, Rh, Co and Cu complexes proved ineffective, while Pd catalysts bearing diverse phosphine ligands furnished the desired product 4 in moderate yields and excellent diastereoselectivity. To our delight, phosphoramidite ligand L₁ derived from TADDOL (α,α,α′,α′-tetraaryl-2,2-disubstituted-1,3-dioxolane-4,5-dimethanol) enabled branched-selective allylation, delivering branched product 5 in 17% yield, along with 15% of the linear isomer 4. These findings prompted a systematic ligand evaluation to develop a ligand-controlled, regiodivergent allylation for accessing structurally diverse, highly substituted alkenes.

Scheme 2. Preliminary Experiments^a.

^aYields were determined by gas chromatography analysis of crude reaction mixtures, using 1,3,5-trimethoxybenzene as the internal standard. Bn, benzyl; Ac, acetyl; PPh₃, triphenylphosphine; N.D., not detected.

We evaluated a range of ligands in a reaction shown in Scheme 3, using allylic carbonate 6 as the electrophile. Ligand identity proved critical for both reaction efficiency and regioselectivity. Most phosphine ligands favored formation of the linear product as commonly observed in palladium catalyzed allylation reactions (see SI, Table S1). The best results were obtained with phosphinooxazoline ligand L₄, which furnished 4 in 85% yield with >99:1 l:b regioselectivity, while other phosphinooxazoline ligands provided significantly lower yields (see SI, Table S14). Phosphoramidite ligands derived from scaffolds such as L₂ (entry 2) and L₃ (entry 3) also favored the formation of linear products, while TADDOL-derived phosphoramidite ligands promoted branched selectivity.

Scheme 3. Ligand Effects^a.

^aYields and regioselectivity were determined by gas chromatography analysis of reaction crudes, using 1,3,5-trimethoxy benzene as the internal standard. Enantiomeric ratio (er) of branched products was determined by HPLC analysis after hydroboration-oxidation (see SI). ^bModified conditions: Pd(OAc)₂ (5 mol%), L₄ (7.5 mol%), KHCO₃ (1.5 equiv), in cyclohexane (0.1 M), at 60 °C for 16 h. Boc, tertbutoxycarbonyl; dba, dibenzylideneacetone; Ac, acetyl; NA, not applicable.

We improved branched selectivity of the TADDOL-derived ligands (entries 1 and 6) by replacing the dimethylamino moiety in the phosphoramidite backbone with piperidine (see SI, Tables S4 and S5). Further investigation revealed that aryl substituents on the TADDOL backbone significantly influenced enantioselectivity. Electron-neutral (L₆, entry 7) and electron-donating (L₇, entry 8) substituents had minimal impact, whereas electron-withdrawing substituents (L₈ and L₉, entries 9 and 10) led to a significant enhancement in enantioselectivity. The branched product 5 was obtained in 85% yield, 95:5 er, and 99:1 b:l regioselectivity using ligand L₉ under the optimized conditions. To the best of our knowledge, this is the first TADDOL-derived phosphoramidite ligand that favors branched selectivity in Pd-catalyzed allylic substitution.

We also examined the structure of the tetrasubstituted alkenyl boranes formed in reactions shown in entries 5 and 10. In both reactions we observed a single regioisomer of the product. NMR analysis allowed us to establish the alkene configurations in both linear and branched allylation products (see SI), revealing a strong preference for anti-addition, which places allylic fragment and the group migrating in 1,2-metallate shift trans to each other. This selectivity complements the syn-selectivity observed in Aggarwal’s recent work.¹⁸

Reaction Scope and Synthetic Applications.

To assess the generality of the optimized conditions, we explored the reactivity of a broad array of substrates (Table 1). After in situ protodeboronation of the initially formed alkenyl boranes with acetic acid, the products were isolated and characterized as trisubstituted alkenes. Reactions proceed with excellent regio- and diastereoselectivity under both branched- and linear-selective conditions. For branched products, the high enantiomeric ratios initially observed with model substrates remained consistent across the expanded substrate scope. In the formation of linear products, the cinnamyl moiety was constructed with high E-selectivity.

Table 1.

Reaction Scope^a

^aYields of isolated products are reported. Reactions performed on 0.50 mmol scale. Single isomer (> 20:1 d.r. and > 20:1 r.r.) was detected in ¹H NMR spectra of isolated products unless otherwise noted. Enantiomeric ratios determined by chiral HPLC analysis (see SI).

^bThe linear isomer of allylic carbonate was used as a substrate

^c80 °C;

^dPd(OAc)₂ (5 mol%) and K₃PO₄ (2.0 equiv) were used instead of Pd(dba)₂;

^eL₇ instead of L₉;

^f2.5 equiv organoborane, 45 °C;

^g40 h. Cy, cyclohexyl; pin, pinacolato; Ts, p-toluenesulfonyl; PMP, p-methoxylphenyl; Phth, phthaloyl.

Initial experiments conducted with a range of different allylic carbonates revealed broad and consistent reactivity. Various aryl- and heteroaryl-substituted allylic carbonates gave satisfactory yields, including those with electron-donating (8, 15) or electron-withdrawing substituents (7, 10, 14, 18). Cross-coupling handles, such as halides (12, 13) and boronic esters (16), were well tolerated. The use of sterically demanding electrophiles (9, 11) did not hinder construction of linear isomers. Alkyl-substituted allylic carbonates gave high yields when used in the synthesis of branched isomers, albeit with a decrease in enantioselectivity (20). Linear isomer of allylic carbonate 6 provided comparable results in the formation of linear product 4, and a significantly lower yield in the formation of branched products 5, albeit with high regio- and enantioselectivity.

We next evaluated the scope of alkyne and organoborane coupling partners. The reaction could be successfully performed in the presence of ketones (21), amides (28), phenols (26), quinoxalines (22), pyridines (24, 37), pyrazoles (33), indoles (38), alkyl (pseudo)halides (25, 30), N-Boc piperidines (32), phthalimides (34), and epoxides (35). Alkyl alkynes, arylboranes (31, 36) and primary alkylboranes proved reliable in the synthesis of both branched and linear isomers. Additionally, aryl alkynes (23) and secondary alkylboranes (29) were competent in delivering linear products. We also noted that sterically hindered alkynes and alkylboranes provided low yields of desired products (see SI, Table S17)

With access to a variety of tetrasubstituted alkenyl boranes, we explored their further functionalization to acyclic tetrasubstituted alkenes (Scheme 4).

Scheme 4. Transformations of Alkenyl Boranes^a.

^aYields of isolated products are reported. Reaction performed on 0.50 mmol scale. Single isomer (> 20:1 d.r. and > 20:1 r.r.) was detected in ¹H NMR spectra of isolated products. The alkenylborane intermediates were prepared via slightly modified conditions and transformed in situ (see SI). Conditions I: aryl iodide (2.5 equiv), NaOt-Bu (3.0 equiv), Pd(OAc)₂ (5 mol%), RuPhos (7.5 mol%), 60 °C. Conditions II: alkyl halide (3.0 equiv, bromide for 41, iodide for 45), LiOi-Pr (2.5 equiv), CuI (50 mol%), 45 °C (for 41) or 80 °C (for 45) in DMAc (0.1 M). Conditions III: trimethylamine N-oxide (3.0 equiv), 25 °C.

Oxidation of the alkenylboranes with trimethylamine N-oxide furnishes the corresponding tetrasubstituted alkenylborinic esters (42, 46), which could be isolated and used in further transformations. Alternatively, the alkenylboranes can be directly cross coupled with aryl iodides without any workup or purification (40, 44). Coupling with C(sp³) electrophiles proceeds smoothly after a solvent swap followed by introduction of a Cu(I) catalyst and the corresponding alkyl halides (41, 45). Notably, all transformations proceed without erosion of stereochemical fidelity. Together, these transformations underscore the capability of the reaction to streamline access to previously difficult-to-access chemical space.

Exploration of the Reaction Mechanism.

Following our investigation into the scope and synthesis of tetrasubstituted alkenes, we turned to mechanistic studies. Reactions employing enantioenriched allylic carbonates (R)-6 and (S)-6 under the standard branched-selective conditions proceeded with comparable efficiency, each furnishing the product 5 in > 70% yield and high er (Scheme 5a). Moreover, conducting the reaction with racemic allylic carbonate 6 as the limiting reagent gave the product 5 in 74% yield and high er (Scheme 5b). Taken together, these experiments indicate that, while minor match/mismatch effects between the catalyst and substrates exist, a kinetic resolution is not a significant contributor to the overall selectivity of the reaction.²² Instead, these results support a dynamic kinetic resolution process during the enantioselective allylic substitution.²³

Scheme 5. Mechanistic Studies^a.

^aAll reactions are conducted on 0.05 mmol scale. Standard Conditions: Pd(dba)₂ (5 mol%), L₉ or (rac)-L₉ (7.5 mol%), PhCF₃ (0.1 M), at 25 °C for 40 h. xs., excess; R.E., reductive elimination.

The formation of a new C–C bond between allylic carbonates and alkynes may proceed via either an inner-sphere reductive elimination²⁴ or an outer-sphere pathway²⁵ that would involve a reaction of π-allyl palladium complex with alkynyl boron-ate intermediate as a nucleophile. To distinguish between these mechanistic possibilities, a deuterium-labelled allylic carbonate, (S)-Z-47, was prepared and subjected to the branched-selective reaction conditions (Scheme 5c).²⁶ Based on the established stereochemistry of Pd-catalyzed allylic substitution (Scheme 5d), oxidative addition proceeds with inversion of configuration²⁷ to generate π-allyl complex Int A, which can equilibrate with complex Int B via π–σ–π isomerization.²⁸ In an outer-sphere pathway, C–C bond formation occurs with a second inversion, furnishing (R)-Z-48 from Int A or (S)-E-48 from Int B. By contrast, an inner-sphere pathway would proceed with retention of configuration, affording (S)-Z-48 or (R)-E-48 from the respective π-allyl complexes. Under the optimized reaction conditions for the branched-selective allylation we observed the formation of (S)-Z-48, consistent with an inner-sphere pathway. This result suggests that the 1,2-metallate shift is promoted by the palladium catalyst and precedes the allylation.

Based on our experimental findings and established mechanisms of 1,2-metallate shift²⁹ and palladium-catalyzed allylation,^23b we propose the reaction mechanism presented in Scheme 5e. Initial oxidative addition of Pd(0) into the allylic carbonate provides Pd(II) complex Int II with concomitant CO₂ extrusion. Subsequent deprotonation of the terminal alkyne with Int II yields Pd acetylide Int III, which reacts with the Lewis acidic organoborane to provide complex Int IV. A diastereodetermining 1,2-metallate shift of the boron-ate complex is then triggered by the interaction with electrophilic π-allyl Pd complex, delivering the tetrasubstituted alkenyl Pd species Int V. Ligand-controlled, regiodivergent reductive elimination delivers tetrasubstituted alkene product and regenerates Pd(0).

CONCLUSION

We have developed palladium-catalyzed trifunctionalization of terminal alkynes using organoboranes and allylic carbonates as coupling partners. The new transformation provides access to highly substituted tri- and tetrasubstituted alkenes with excellent regio- and diastereoselectivity. In addition, the incorporation of the allylic electrophile can be accomplished with both branched and linear selectivity, and with excellent diastereo- and enantioselectivity, allowing access to a range of highly complex 1,4-dienes. We demonstrate the broad scope of the reaction and the ability to transform the initially formed tetrasubstituted alkenyl boranes into a range of tetrasubstituted alkenes. Finally, our mechanistic study supports the mechanism that involves alkynyl boron-ate formation followed by palladium-promoted 1,2-metalate shift that controls the regio- and diastereoselectivity of the alkene formation. The subsequent allylation of the alkenyl palladium intermediate is also controlled by the palladium catalyst, with a proper choice of the ligand enabling high regi-, diastereo-, and enantioselectivity of the transformation. Overall, we demonstrate that metal-catalyzed transformations of the alkynyl boron-ate intermediates can be used with complex prochiral electrophiles to access highly complex chiral molecules from simple starting materials.

Supplementary Material

This material is available free of charge via the Internet at http://pubs.acs.org.

Experimental procedures and product characterization (pdf).