Stereo-reversed E2 Unlocks Z-selective C–H Functionalization

Abstract

The stereoselective functionalization of C–H bonds represents a central challenge in modern organic synthesis. Despite decades of innovation in C–H activation chemistry, methods for Z-selective functionalization of alkenes have eluded synthetic practitioners. Terminal alkenes present the most vexing challenge for Z-selectivity, as they require selective cleavage of the more hindered of two otherwise virtually identical C–H bonds. Herein, we describe the transformation of alkenes into transient 1,2-bis-sulfonium intermediates found to undergo Z-selective elimination, overturning a textbook E2 stereoselectivity rule via stabilizing interactions. We identify paired electrolysis as a uniquely enabling strategy to both selectively generate the requisite bis-sulfonium intermediate and drive its rapid elimination in situ. The resultant Z-alkenyl sulfonium linchpins provide access to a wide array of Z-alkene targets from inexpensive feedstocks through robust cross-coupling reactions.

The transformation of C–H bonds is a longstanding goal of organic synthesis. Decades of research have delivered an array of mechanistically diverse approaches to replace these classically inert bonds with new C–C bonds (1, 2). Within this broad arena, strategies to control stereoselectivity are of paramount importance. While elegant examples of atropo- (3) and enantioselective (4) aryl and alkyl C–H functionalization reactions have recently emerged, Z-selective alkenyl C–H functionalization reactions remain elusive (5, 6). Such processes present a formidable challenge due to the inextricable introduction of unfavorable allylic-1,3 strain (Fig. 1A) (7,8). For example, a Z-selective variant of the Heck reaction has yet to arise despite over half a century of innovation on this Nobel Prize-winning reaction (5, 9–11). While other synthetic strategies to access Z-alkenes are known (12–21), all methods to selectively cleave the pro-Z C–H bond rely on directing groups to enforce stereoselectivity. This strategy intrinsically limits Z-selective C–H functionalization to a small pool of rigid alkenes bearing strongly coordinating auxiliaries (22). Overall, the realization of a complementary method for undirected, Z-selective C–H functionalization will require a new mechanistic strategy to overcome the inherent thermodynamic bias against Z-alkene formation.

Fig. 1: Overview of this work.

(A) C–H Functionalization approach to Z-olefin synthesis from terminal alkenes. (B) Envisioned strategy: vicinal dinucleofuge eliminates to diversifiable Z-linchpin.

We envisioned a fundamentally distinct approach to Z-selective C–H functionalization. Specifically, transformation of a terminal alkene into a transient 1,2-dinucleofuge followed by Z-selective elimination would deliver a Z-alkenyl linchpin (Fig. 1B) (23). This deceptively simple proposal requires stereoselective elimination to a Z-1,2-disubstituted alkene. Unfortunately, this seemingly elementary reaction requires violation of a basic rule taught in introductory organic chemistry: E2 stereoselectivity is predicted by the least sterically hindered Newman projection capable of anti-elimination (24). Indeed, while dinucleofuges, such as dihalides, can be readily prepared from terminal alkenes, they categorically eliminate to the predicted E-alkenyl halides alongside branched products (25–27). In fact, the E-selective elimination of dihalides served as the first case study in the seminal disclosure of this pedagogical tool by Newman in 1955 (28). Herein, we describe a strategy for Z-selective C–H functionalization that hinges on the discovery of a unique Z-selective elimination. Our approach transforms an alkene into a transient 1,2-bis-sulfonium dinucleofuge that engages in stabilizing interactions to override steric hindrance in the pro-Z transition structure, overturning the 70-year-old E2 stereoselectivity model.

Discovery of a Z-selective E2 and translation to Z-selective thianthrenation

While studying the reactivity of bis and mono, electrogenerated adducts of thianthrene (TT) and an alkene (29), we made a surprising discovery. We observed an unambiguous correlation between the bis-to-mono ratio and the Z-to-E ratio of the resultant elimination products (Fig. 2A). This trend predicted that bis would eliminate selectively to Z-alkenyl thianthrenium product 1. To validate this prediction, we isolated the small amount of bis formed through electrolysis and treated it with trifluoroacetate, a weak base found to not promote alkenyl thianthrenium isomerization. This experiment confirmed that bis eliminates with nearly perfect Z-selectivity. Given that alkenyl sulfonium salts can be engaged as cross coupling electrophiles (30, 31), this discovery sets the stage for a general approach to Z-selective C–H functionalization.

Fig. 2: Reaction design.

(A) Unexpected discovery: bis- and mono-adducts undergo stereodivergent eliminations. (B) Major challenges in achieving a bis-selective thianthrenation protocol. Reaction parameters: divided cell, constant current electrolysis, i = 10.0 mA, alkene (0.4 mmol), thianthrene (varying equiv.), TFA (5.0 equiv.), KPF₆ (4.0 equiv.), 1:1 MeCN:PhCN (0.05 M), 25°C, 2.6 hour. (C) Design strategy for Z-selectivity: undivided cell paired electrolysis. (D) Validation of strategy: reactor-dependent thianthrenation selectivity. For experimental details, see SI. TFA, trifluoroacetic acid.

Having established that bis eliminates Z-selectively, we sought to translate this discovery into a useful Z-selective C–H functionalization protocol. Unfortunately, bis-thianthrenium adducts are exclusively minor products in established electrochemical processes (27, 32, 33) and are not observed under S-oxide-based thianthrenation (31). However, we reasoned that the divergent mechanisms of formation for bis and mono offer an untapped opportunity to develop a bis-selective protocol. Our current working mechanistic model for adduct formation posits that bis is formed by iterative addition of TT^•+ across the alkene, whereas mono is formed through rapid cycloaddition of TT²⁺ with the alkene. While chemical thianthrenation conditions are designed to access TT²⁺ and therefore exclusively form mono, electrochemical thianthrenation generates both TT^•+ and TT²⁺, providing access to bis as a minor product. While TT^•+ is formed at the anode throughout the reaction, formation of TT²⁺ requires an endergonic disproportionation of TT^•+. Therefore, the rate of mono formation is sharply sensitive to the concentrations of both TT^•+ and TT (Fig. S9). Consistent with this scenario, electrochemical thianthrenation is bis-selective at low conversion, when the concentration of TT is high. However, as TT is consumed anodically, disproportionation of TT^•+ to generate TT²⁺ becomes increasingly feasible. As a consequence, mono formation accelerates as the reaction proceeds, furnishing mono as the major product by the end of electrolysis (Fig. S7).

We hypothesized that increasing the concentration of TT should suppress mono by comproportionation with TT²⁺. Excitingly, we found that increased TT loading dramatically suppressed mono formation, although, as expected, the absolute yield of bis was not impacted (Fig. 2B, left). While on its own, this approach delivered low yield of bis, it suggested that simply extending electrolysis time with a large excess of TT would deliver a high-yielding, bis-selective protocol. Unexpectedly, however, even with a large excess of TT (8 equiv.), allowing electrolysis to proceed to high conversion resulted in a significant erosion in bis selectivity (10:1 decreased to 3:1). Given that high TT loading unambiguously suppresses the direct pathway to form mono, these data implicated an indirect pathway wherein bis converts to mono over time (34). Stirring isolated bis in reaction solvent revealed that bis is intrinsically unstable and converts to mono at a rate competitive with that of electrochemical adduct formation (Fig. 2B, right). In principle, this bis-to-mono conversion could be outpaced by rapid in-situ elimination; however, we found that addition of a wide range of bases inhibited adduct formation altogether (Table S3), presumably due to interactions with the electrophilic oxidized TT species.

We questioned whether we could redesign the electrochemical system to both suppress mono and also prevent bis-to-mono conversion. While divided cell electrolysis was initially used to maintain high faradaic efficiency, we envisioned that the use of an undivided cell, which introduces a cathode into the reaction chamber, could offer two strategic advantages (Fig. 2C) (35). First, the cathode would maintain an elevated steady-state concentration of TT through reduction of anodically generated TT^•+. This would disfavor disproportionation and therefore suppress mono generation without requiring a large excess of TT. Second, cathodic hydrogen evolution reaction (HER) of trifluoroacetic acid would generate trifluoroacetate base at a rate intrinsically coupled to TT oxidation. This paired electrolysis manifold would ensure that base is only generated as adduct is formed, mitigating base-induced reaction inhibition. To our delight, undivided cell electrolysis inverted the typical stereoselectivity of thianthrenation, delivering model alkenyl thianthrenium salt 1 in both high yield and Z-selectivity with minimal additional optimization (Fig. 2D).

Scope and synthetic utility of Z-alkenyl thianthrenium salts

We next probed the scope of alkenes amenable to this Z-selective thianthrenation protocol (Fig. 3A). Throughout these studies, we found that the intrinsic Z-selectivity of the process was consistently high. Furthermore, the exceptional crystallinity of these Z-alkenyl thianthrenium salts enables recrystallization to afford nearly diastereopure products. This represents a practical advantage over traditional alkene syntheses, wherein separation of geometric isomers is notoriously challenging. Under these undivided cell conditions, Z-alkenyl thianthrenium salts are readily accessed from terminal alkenes bearing a diverse array of common functional groups (3–25). Notably, while amines are typically oxidized at lower potentials (36) than those required to oxidize TT (Fig. S11), protonation under standard reaction conditions allows selective thianthrenation of alkenes bearing unprotected amines (11 and 12). While both alcohols and pyridines are known to attack oxidized TT species (37–39), Z-alkenyl thianthrenium salts can still be accessed from alkenes bearing both of these important functional groups (13 and 14). Pyridines are protonated analogously to aliphatic amines, whereas alcohols are protected in situ with a labile trifluoroacetyl group by trifluoroacetic anhydride. Alkene substrates bearing other protic functional groups, such as amides (15 and 16), carboxylic acids (17), and sulfonamides (18), are each thianthrenated without in-situ transformation. The oxidizing conditions do render some electron-rich groups, such as sulfides and enolizable ketones, incompatible. Nonetheless, ketone-containing Z-alkene products remain accessible since alkenes bearing Weinreb amides (19) undergo efficient Z-selective thianthrenation. Common electrophiles such as aryl and alkyl bromides are fully preserved in this Z-selective thianthrenation process (20 and 21). Notably, for substrates containing both terminal alkenes and acrylates, only the terminal alkene undergoes thianthrenation (22). This selectivity pattern complements transition metal-catalyzed alkenyl C–H activation reactions that commonly employ acrylate derivatives as substrates. Selective functionalization of a single alkene is also observed for substrates bearing two terminal alkenes (23). In this case, we posit that distal electronic communication from the electron-deficient alkenyl thianthrenium moiety in the product inhibits iterative thianthrenation. Our approach also successfully engages gaseous feedstocks using TT as the limiting reagent to access simple Z-alkene building blocks (24) with high diastereoselectivity. While monosubstituted alkenes are readily engaged, more substituted alkenes are not amenable to this electrochemical process: 1,1-disubstituted alkene substrates provide intractable mixtures of products, and internal alkenes fundamentally shift the stereodetermining step as they no longer possess both pro-Z and pro-E C–H bonds in the bis-thianthrenium intermediates. However, alkenes bearing allylic substituents still undergo Z-selective thianthrenation despite challenging the system with a significant increase in 1,3-allylic strain (25). Finally, underscoring its practical synthetic utility, this undivided cell thianthrenation scales efficiently in batch using inexpensive graphite and stainless-steel electrodes to deliver multigram quantities of pure Z-alkenyl thianthrenium 1 after recrystallization (Fig. 3B).

Fig. 3: Z-selective C–H thianthrenation.

(A) Scope of Z-alkenyl thianthrenium electrophiles. Standard conditions: undivided cell, constant current electrolysis, i = 93 mA, alkene (3.1 mmol), thianthrene (2.0 equiv.), TFA (10.0 equiv.), TFAA (1.4 equiv.), KPF₆ (3.0 equiv.), 1:1 MeCN:PhCN (0.125 M), 25°C, 26.8 hours. Additional experimental details are available in supplementary information (Page S17). ^§Reaction conducted with 11 equiv of TFA. ‡Reaction conducted with 2.5 equiv. of TFAA. ^†Yield was determined by ¹H-NMR analysis using CH₂Br₂ as an internal standard. (B) Decagram-scale Z-selective C–H functionalization in batch. TFA, trifluoroacetic acid; TFAA, trifluoroacetic anhydride.

In principle, Z-selective C–H thianthrenation provides an electrophilic linchpin for rapid diversification via stereospecific cross-coupling. However, alkenyl thianthrenium salts can undergo isomerization under basic conditions (40, 41). In all reported C–C coupling reactions of alkenyl thianthrenium salts derived from terminal alkenes, the more stable stereoisomer was used to generate an E-alkene. This has obscured whether stereochemistry is maintained through a stereospecific process or if different stereoisomers could converge to the same product (31). We were pleased to find that the stereochemistry of Z-alkenyl thianthrenium 7 was translated to a variety of C–C bonds with minimal deviations from typical cross-coupling conditions (Fig. 4A) (31, 42). Pd-catalyzed Sonogashira, Suzuki, and Negishi coupling protocols each afforded the corresponding C(sp), C(sp²), and C(sp³) cross-coupled products 26–28 with excellent retention of alkene geometry. A Heck reaction delivered E,Z-diene 29 with minimal erosion of diastereoselectivity despite proceeding through a Pd–H intermediate, which can promote Z- to E-alkene isomerization. Borylation transposed the polarity of the Z-alkenyl thianthrenium handle to furnish the nucleophilic Z-alkenyl building block 30 with an organoboron coupling handle (43). Finally, a Pd-catalyzed carbonylation furnished Z-acrylate product 31. Additionally, substrates bearing pendant nucleophiles known to react with alkenyl thianthrenium salts under basic conditions (27, 44) did not impede productive coupling (see Page S40). Overall, successful implementation of these reactions suggests that Z-alkenyl thianthrenium salts will serve as Z-alkenyl pseudohalides across a wide range of stereospecific coupling reactions.

Fig. 4: Synthetic applications.

(A) Cross-coupling diversifications of a Z-alkenyl thianthrenium linchpin. (B) Streamlined approach to bioactive natural product synthesis via Z-selective C–H functionalization.

We envisioned that this Z-selective C–H functionalization protocol would enable approaches to a wide array of valuable synthetic targets from inexpensive petrochemical feedstocks through intuitive cross coupling disconnections (Fig. 4B). Through our C–H functionalization approach, 1-hexene underwent Z-selective thianthrenation (32) and cross-coupling with commercially available 10-undecyn-1-ol to provide key intermediate 33 on path to Clathculin B (45). Similarly, Z-selective thianthrenation of 1-pentene provided Z-linchpin 34, setting the stage for a Suzuki coupling to furnish the Silk Moth sex pheromone Bombykol (35) in excellent yield in three steps (46). We next transformed 1-dodecene into the corresponding Z-thianthrenium salt 36. Subsequent C(sp²)–C(sp³) Negishi coupling furnished 37, a key intermediate in the synthesis of Spongy Moth sex pheromone (±)-cis-Disparlure (47). Finally, we synthesized pear fragrance ethyl decadienoate (39) in two steps through a Z-selective thianthrenation-Heck sequence from 1-heptene and ethyl acrylate (48). Beyond offering an efficient route to each of these compounds, our strategy presents several practical advantages. Recrystallization of each Z-alkenyl thianthrenium intermediate consistently provides stereopure material, contrasting traditional approaches to Z-alkene synthesis (e.g., Wittig olefination) that have variable stereoselectivity and lack any straightforward process to separate stereoisomers. Furthermore, the key Z-selective C–H functionalization step is robust, requiring no precautions to exclude air or moisture. This offers additional operational simplicity relative to emerging Z-selective metathesis strategies that rely on sensitive metal catalysts. Given the success of these proof-of-concept experiments, we anticipate the broad implementation of this synthetic tactic in the preparation of a wide range of Z-alkene targets.

Origins of Z-selectivity

Having demonstrated the synthetic value of this Z-selective C–H functionalization protocol, we sought to interrogate the mechanistic origins of the highly Z-selective elimination of bis-adducts. In introductory organic chemistry, students are taught to predict E2 stereoselectivity by identifying the least sterically encumbered Newman projection with an anti-periplanar arrangement of leaving group and hydrogen. Given that steric clashes from the ground state conformers reliably translate into the elimination transition structures, this approach has proven virtually infallible. In the present context, however, it incorrectly predicts E-selective elimination of bis because the pro-E conformer (gauche-E) has one fewer gauche interaction relative to its pro-Z counterpart (gauche-Z). To reconcile this disconnect, we computationally modeled these gauche conformers and their corresponding anti-elimination transition structures using 1-butene as a simple model substrate. These calculations establish that the relative stabilities of these conformers invert during elimination. While the rudimentary conformational analysis holds and gauche-Z is slightly higher in energy than gauche-E (ΔG° = 0.9 kcal/mol), the Z-forming anti-elimination transition structure (TS-anti-Z) is significantly lower in energy (ΔΔG^‡ = 3.2 kcal/mol) than its E-forming analog (TS-anti-E). Further investigation using distortion-interaction analysis (49) indicated that distortion energies drive the energetic difference between these two anti-elimination transition states (Fig. S20). To rationalize the difference in distortion energies, we analyzed the geometric changes that each gauche conformer undergoes to reach its elimination transition structure. This revealed that gauche-E requires extensive geometric reorganization to access TS-anti-E, whereas the geometry of gauche-Z is largely preserved in TS-anti-Z. This difference is clearly reflected in the change in dihedral angle between the two thianthrenium groups (Δ∠S–C–C–S), which is much larger for the E-forming pathway (contracts by 15.5°) than it is for the Z-forming pathway (expands by 1.0°).

To elucidate the specific interactions underlying the pronounced difference in geometric distortion, we performed non-covalent interaction (NCI) (50) and natural bond orbital (NBO) (51) analysis on the two gauche conformers and their respective anti-elimination transition structures (Figs. S21–24). We found that both conformers engage in a host of stabilizing interactions: (i) π-π contacts between the aromatic rings of the two thianthrenium units (52) and (ii) hydrogen bonding contacts between the neutral sulfur of each thianthrenium unit and its α-alkyl C–H bond (53). Each of these interactions was found to play a critical role in influencing Z-selectivity. First, the π-π contacts are stronger for gauche-Z than they are for gauche-E, offsetting the steric penalties of this more hindered conformer. Crucially, this differential carries into the corresponding elimination transition structures, where TS-anti-Z engages in stronger π-π contacts than TS-anti-E. Second, the terminal thianthrenium unit hydrogen bonds to the pro-E C–H bond (H_E) in both gauche conformers. Since the pro-E C–H bond must be deprotonated in the E-forming pathway, TS-anti-E sacrifices this C–H···S hydrogen bonding interaction. Moreover, in TS-anti-E, the terminal thianthrenium unit must rotate away from the site of deprotonation, which forces the internal thianthrenium unit to rotate accordingly. This structural accommodation preserves π-π contacts and prevents destabilizing steric clashes but comes at the cost of the internal C–H···S hydrogen bonding interaction. In contrast, TS-anti-Z retains its entire network of stabilizing interactions, including both of its C–H···S hydrogen bonding contacts, rationalizing the minimal geometric reorganization required for the Z-forming pathway.

While this comparative analysis of anti-elimination pathways clarifies the mechanistic origin of Z-selectivity, further scrutiny revealed a surprising alternative pathway to E-product that proceeds via a sterically congested syn-elimination. This syn-elimination (TS-syn-E) is lower in energy than TS-anti-E because it retains the key terminal C–H···S hydrogen bonding contact by inverting which C–H bond is deprotonated to form E-product (Fig. 5B). This refinement of the model predicts an energy difference in excellent agreement with the experimentally observed bis-adduct elimination stereoselectivity (ΔΔG^‡ = 1.7 kcal/mol, both calculated and experimental). Taken together, these observations provide a complete mechanistic model for the reaction that explains the observed Z-selectivity: the least sterically hindered transition state that maintains key stabilizing interactions results in the Z-stereoisomer (Fig. 5C) (see Page S60 for additional commentary).

Fig. 5: Origins of Z-selectivity.

(A) Computational analysis of anti-elimination pathways. (B) Lowest-energy pathway to E-product. (C) Complete energetic landscape. (D) Structural analog probing origin of stereocontrol. DFT calculations were performed at the UB3LYP-D3/def2TZVPP-CPCM(acetonitrile)//UB3LYP-D3/6–311G(d,p)-CPCM(acetonitrile) level of theory. NCI and Distortion Interaction Analysis calculated at the UB3LYP-D3/6–311G(d,p)-CPCM(acetonitrile) level of theory.

Finally, we designed an experiment to selectively probe the unusual hydrogen bonding interaction at the heart of our stereochemical model. We envisioned that swapping the neutral sulfur atom of TT with a smaller oxygen atom (PT) would disrupt this hydrogen bonding interaction (Fig. 5D). While oxygen is typically a better hydrogen bond acceptor than sulfur, the geometric constraints of the rigid bis-adduct system rely on sulfur’s diffuse orbitals to engage in these long-range interactions (54). Indeed, NCI and NBO analyses predict that swapping the neutral sulfur atom with a smaller oxygen atom (bis-O) precludes the critical hydrogen bonding interactions while maintaining a similar steric profile and π-π interactions. As a result, bis-O is predicted to eliminate with poor stereoselectivity (ΔΔG^‡ = 0.1 kcal/mol). Experimentally, electrogenerated bis-O eliminates to a 1.6:1 Z:E mixture of alkenyl sulfonium products (Figs. S27–28). This dramatic decrease in selectivity upon such a seemingly distal structural modification supports the hypothesis that the hydrogen-bonding interaction plays a critical role in enforcing high Z-selectivity. Moreover, the fact that the bis-O system still does not revert to high E-selectivity underscores the contribution of π-π contacts that offset the steric penalties associated with the Z-forming pathway. Overall, these data establish that strategically positioned non-covalent interactions can override steric control in elimination reactions.

Conclusion

We have provided a solution to a longstanding challenge: Z-selective C–H functionalization of unactivated alkenes. This transformation was unlocked by an unexpected Z-selective elimination reaction of a transient bis-thianthrenium intermediate that overturns textbook E2 stereoselectivity predictions through stabilizing interactions. We illustrate how electrochemical cell design can be leveraged as a critical parameter to both selectively generate the requisite bis-thianthrenium dinucleofuge and drive its rapid elimination in situ to circumvent its inherent instability. Beyond providing a practical approach to Z-alkene synthesis, this work illustrates how non-covalent interactions can override intrinsic steric preferences in elimination reactions. We anticipate that both the synthetic methods and mechanistic principles uncovered herein will find broad application in accessing stereodefined alkenes.

Supplementary Material

Supplementary Materials:

Materials and Methods

Supplementary data

Figures S1 to S29

Tables S1 to S18

References (55–82)

Data and materials availability:

X-ray data are available free of charge from the Cambridge Crystallographic Data Centre under accession number CCDC 2379265. All other data and methodological details are in supplementary materials, including substrate preparation, cross-coupling procedures, product isolation and characterization, computational data, and NMR spectra.