Ethene-1,1-disulfonyl Difluoride (EDSF) for SuFEx Click Chemistry: Synthesis of SuFExable 1,1-Bissulfonylfluoride Substituted Cyclobutene Hubs

Abstract

We present the synthesis of 1,1-bis(fluorosulfonyl)-2-(pyridin-1-ium-1-yl)ethan-1-ide, a bench-stable precursor to ethene-1,1-disulfonyl difluoride (EDSF). The novel SuFEx reagent, EDSF, is demonstrated in the preparation of 26 unique 1,1-bissulfonylfluoride substituted cyclobutenes via a cycloaddition reaction. The regioselective click cycloaddition reaction is rapid, straightforward, and highly efficient, enabling the generation of highly functionalized 4-membered ring (4MR) carbocycles. These carbocycles are both valuable structural motifs found in numerous bioactive natural products and pharmaceutically relevant small molecules. Additionally, we showcase diversification of the novel cyclobutene cores through selective Cs₂CO₃-activated SuFEx click chemistry between a single S–F group and aryl alcohol, yielding the corresponding sulfonate ester products with high efficiency. Finally, density functional theory calculations offer mechanistic insights about the reaction pathway.

Sulfur-Fluoride Exchange (SuFEx) is a highly versatile and powerful click transformation, with extensive scope, first reported by Sharpless and colleagues in 2014^[1–3]. SuFEx adheres to many of the original click chemistry principles (e.g., atom economy, high yield, simple purification, and selectivity for a single product) necessary to achieve click-status and is one of the few “true” click reactions compatible with biological systems^[4–6]. SuFEx leverages a delicate balance between the stability and reactivity of high oxidation S–F compounds, which, in contrast to their S–Cl counterparts, are resistant to reductive collapse, permitting a clear pathway for nucleophilic exchange.

Traditional SuFEx reactions typically involve aryl silyl ethers and aryl sulfonyl fluorides or fluorosulfates. These reactions are catalyzed by Lewis basic tertiary amines or phosphazenes^[1], although bifluoride salts and Cs₂CO₃ activation routes are also feasible^[7–9]. Additionally, silica-free SuFEx protocols have been developed^[10], including Accelerated SuFEx Click Chemistry (ASCC), which utilizes a synergistic hexamethyldisilazane (HMDS)-BTMG (Barton’s base) catalytic system to generate stable S–O linkages directly from aryl and alkyl alcohol nucleophiles^[11–13].

SuFEx has profoundly influenced various disciplines, encompassing drug discovery and materials science^{[2,7,8,14–23]}. This success is, in part, attributed to the development of versatile connective hubs that facilitate new sulfur-centered linkages for the swift synthesis of complex molecules. The family of S–F connective hubs can be categorized into two groups: i) inorganic high oxidation sulfur compounds, which include connective gases such as sulfuryl fluoride (SO₂F₂) and thionyl tetrafluoride (SOF₄), both possessing multiple exchangeable S–F bonds; and ii) carbon-based high oxidation state S–F bond-containing hubs, including Michael acceptors and dipolarophiles such as ethene sulfonyl fluoride (ESF) ^[1,20,24,25], 1-bromoethene-1-sulfonyl fluoride (BESF)^[26–28], and the versatile 2-substituted-alkynyl-1-sulfonyl fluorides (SASFs) ^[29–31].

The fluorosulfuryl isocyanate (FSI) and methanedisulfonyl fluoride (MDSF) reagents represent the latest additions to the SuFEx family, facilitating the synthesis of fluorosulfuryl carbamates (and ureas) and unsaturated 1,1-disulfonyl fluorides, respectively^[32,33]. Building on our previous work in developing α,β-unsaturated SuFEx hubs (BESF, SASF), we present ethene-1,1-disulfonyl difluoride (EDSF), an unprecedented SuFEx reagent for the synthesis of SuFExable 1,1-bissulfonylfluoride substituted cyclobutene hubs.

4-Membered ring (4MR) carbocycles are important, albeit underexploited motifs in bioactive natural products and pharmaceutically relevant small molecules (Fig. 1D). Examples include the natural product cyclomegistine (6) and the cyclobutene analog (7) of combretastatin A-4, both of which have potent cytotoxic activity against several cancer cell lines^[34–36]. The strained carbocycle can favorably contribute to drug properties of small molecules, such as reduced planarity, increased metabolic stability, and conformation restriction

Figure 1.

A) Examples of SuFEx hubs; B) Work by Alcaide; C) This work: in situ generation and application of ethene-1,1-disulfonyl difluoride; D) 4-Membered carbocycles in bioactive small molecules and natural products.

Additionally, 4MR systems offer versatile cores for accessing more complex structures. However, achieving regioselective synthesis of functional cyclobutenes presents a significant challenge. Therefore, we sought to develop a clickable cyclobutene that would serve as a connective hub, allowing for derivatization of the 4MR core. Alcaide et al. demonstrated that 2-(pyridinium-1-yl)-1,1-bis(triflyl)ethanides (1) are efficient 1,2-dipole precursors in [2+2] cycloaddition reactions with alkynes, resulting in the formation of 1,1-bis(trifluoromethylsulfonyl)-substituted cyclobutenes (2)^[37,38]. Our vision was to use EDSF (4), a reagent that replaces the trifyl unit with sulfonyl fluorides, to produce sulfonyl fluoride functionalized cyclobutenes that can be further diversified through late-stage SuFEx click chemistry.

In light of the expected instability of EDSF, we aimed to synthesize a bench-stable zwitterionic precursor. Our approach involved heating methanedisulfonyl fluoride (MDSF) with paraformaldehyde and a range of 2-substituted pyridines, such as 2,4-dimethyl-, 2-acetyl-, 2-fluoro-, 2-chloro-, and 2-bromopyridine, along with pyridine in anhydrous DCE^[39]. While the reactions involving 2-substituted pyridines resulted in complex product mixtures^[40], the reaction with pyridine successfully yielded 1,1-bis(fluorosulfonyl)-2-(pyridin-1-ium-1-yl)ethan-1-ide (3) as a bench-stable precursor. This precursor was obtained in gram-scale quantities (2.4 g) within a short reaction time of 1 h (Scheme 1A)^[41,42].

Scheme 1.

A) Synthesis of 1,1-bis(fluorosulfonyl)-2-(pyridin-1-ium-1-yl)ethan-1-ide (3) and its reaction with terminal alkynes. A) Synthetic procedure. B) Substrate scope. C) Representative NOESY spectrum for 11e. Reactions were conducted on a 100 μmol scale unless stated otherwise in the Supplementary Information. Isolated yields are reported.

We subsequently investigated the in situ generation of EDSF and its ensuing 1,2-dipolar cycloaddition with alkynes, employing the readily accessible electron-rich aryl acetylene, 4- methoxyphenylacetylene, as a model substrate. Our findings revealed that the addition of 2.0 equivalents of H₂SO₄ to a 1:2 mixture of 4-methoxyphenylacetylene and 3 at −10 °C in MeCN/H₂O, followed by warming to room temperature and stirring for 1 hour, resulted in the formation of cyclobutene 11a as a single regioisomer with an 84% yield^[43]. The reaction exhibited favorable performance with a range of aryl acetylenes and enynes 10a–10q (Scheme 1A), yielding 17 novel 1,2-disubstituted cyclobutene products 11a–11q in good to excellent yields (56–97%). Notably, the aryl enynes 10m and 10n reacted with 3 via the triple bond, while non-conjugated alkyne substituents remained unaffected, showcasing remarkable chemoselectivity.

The syn-regiochemistry of the cyclobutene products was elucidated by NOESY NMR. A strong NOE correlation between protons H_A and H_B of cyclobutene 11e supports aryl substitution in the 2-position of the cyclobutene carbocycle relative to the 1,1-disulfonyl fluoride substituted carbon (Scheme 1B).

Our assignment was substantiated through the use of HMBC NMR analyses, which demonstrated a robust correlation between the methine proton and the methylene carbon within the carbocyclic framework (refer to Supporting Information, Fig. S2). Moreover, the ¹H chemical shifts exhibited a notable correspondence with the analogous trifluoromethyl derivative, as synthesized by Alcaide et al^[37].

The reaction involving compound 3 and a variety of non-terminal aryl acetylenes and enynes (12a–12i) was subsequently investigated. By employing slightly modified conditions [1.0 eq. of alkyne, 2.0 eq. of compound 3, and 2.0 eq. of H₂SO₄ stirred in MeCN at −10 °C, followed by stirring at room temperature for 1 h], a series of highly substituted cyclobutenes (13a–13i) were obtained with yields up to 90% (Scheme 2A). The regiochemistry of the products was verified through NOESY NMR analysis, exemplified by compound 13e in Scheme 2B. The observed NOE correlation between proton H_A in the carbocyclic structure and proton H_B of the cyclohexyl ring implies aryl substitution at the 2-position of the cyclobutene carbocycle relative to the 1,1-disulfonyl fluoride substituted carbon.

Scheme 2.

Reaction of non-terminal alkynes with 1,1-bis(fluorosulfonyl)-2-(pyridin-1-ium-1-yl)ethan-1-ide. A) Substrate scope. B) Representative NOESY spectrum for 13e. Reactions were conducted on a 100 μmol scale unless stated otherwise in the Supplementary Information. Isolated yields are reported.

Next, we delved into the late-stage SuFEx diversification of the newly developed cyclobutene hubs, a direction unexplored in previous studies. Upon conducting the reaction between the representative substrate 11a and an aryl silyl ether (4-OMePhOTBS) using 20 mol% DBU in MeCN, no discernible reaction occurred. This could potentially be attributed to the diminished reactivity of alkyl sulfonyl fluorides compared to their analogous aryl substrates.

Upon screening reaction conditions, we found that Cs₂CO₃ could facilitate the SuFEx reaction between the S–F and the corresponding aryl alcohol, presumably through the phenolate formation^[9]. Consequently, under optimized conditions [1.0 eq. of cyclobutene 11a, 1.1 eq. of phenol 14a, and 1.1 eq. of Cs₂CO₃ stirred at room temperature for 3 h in MeCN], compound 15a was exclusively produced with an 80% yield. Employing the modified protocol, we successfully synthesized 10 diverse sulfonate products (15a-15j), including alkynyl- (15b), nitro- (15c), ester- (15d), dimethylamino- (15f), trifluoromethyl(thio)- (15g), and nitrile- (15h), achieving good to excellent yields ranging from 72–88% (Scheme 3)^[44].

Scheme 3.

Reaction of phenols with cyclobutenes and substrate scope. Reactions were conducted on a 250 μmol scale unless stated otherwise in the Supplementary Information. Isolated yields are reported.

Mechanistically, we hypothesize a stepwise [2+2] cycloaddition for ethene-1,1-disulfonyl difluoride, as corroborated by density functional theory (DFT) calculations at the ωB97XD/Def2-TZVP level (Scheme 4). This computational approach facilitates an in-depth examination of the reaction mechanism, elucidating key factors that contribute to the observed outcomes. Consequently, the addition of H₂SO₄ to 1,1-bis(fluorosulfonyl)-2-(pyridin-1-ium-1-yl)ethan-1-ide facilitates the dissociation, yielding the highly reactive EDSF, which undergoes stepwise cycloaddition with the alkyne through two competing pathways. Pathway 1 is the distal addition of the alkyne (TS1distal), which has a ΔG value of +68 kJ/mol and leads to the zwitterionic intermediate (INTdistal, ΔG = −4 kJ/mol), pathway 2 is the proximal addition (TS1prox) with a ΔG of +71 kJ/mol leading to the zwitterionic intermediate (INTprox, ΔG = −10 kJ/mol). Although only pathway 2 is proposed in the analogous calculations by Alcaide et al.^[37], we note close competition between these two addition pathways, and both zwitterionic intermediates ultimately converge to reach the very low energy TS2 (ΔG = 16 kJ/mol) after INtdistal undergoes conformational reorganization. Finally, the favorable cyclobutene product is formed (ΔG = −98 kJ/mol) in a concerted fashion. The outstanding regioselectivity observed in our experimental findings is further supported by our analysis of the alternative regioisomeric pathway. This pathway, which involves the addition of the alkyne via the internal carbon, was determined to have a significantly higher activation barrier (TSB, ΔG = 151 kJ/mol). The substantial energy barrier renders this pathway unfeasible, providing a strong rationale for the exceptional regioselectivity observed experimentally.

Scheme 4.

DFT calculations at: ωB97XD/Def2TZVP//ωB97XD/Def2SVP (SMD=MeCN), relative ΔG in kJ/mol at 298K

In summary, we have developed a practical approach for the regioselective synthesis of 26 unprecedented sulfonyl fluoride functionalized cyclobutenes with up to 97% yield from ethene-1,1-disulfonyl difluoride (EDSF). EDSF is conveniently generated in situ from the novel bench-stable precursor, 1,1-bis(fluorosulfonyl)-2-(pyridin-1-ium-1-yl)ethan-1-ide. Moreover, we demonstrate the capacity of the cyclobutene products to be rapidly diversified through late-stage SuFEx modification under specific conditions, selectively exchanging just one S–F group with a library of phenols to isolate 10 mono-sulfonate products. Lastly, a plausible reaction mechanism supported by molecular modeling studies is provided, in which we propose a stepwise [2+2] cycloaddition between ethene-1,1-disulfonyl difluoride and the alkyne starting materials.