Unlocking catalytic asymmetric transformations of sulfilimidoyl fluorides toward chiral sulfinimidate esters

Abstract

The rational design of bioactive molecules increasingly utilizes sulfur-containing groups such as sulfoximines and sulfilimines. While sulfonimidoyl fluorides have emerged as versatile chiral S(vi) electrophiles enabling efficient asymmetric synthesis of sulfoximines, their S(iv) counterparts, sulfilimidoyl fluorides, have remained inaccessible in asymmetric catalysis due to their intrinsic instability and configurational lability. Central to this strategy is a bifunctional tertiary amine–hydrogen-bond donor catalyst that enables selective conversion of a transient sulfur(iv) intermediate. Mechanistic studies support a dynamic kinetic resolution process involving a sulfilimidoyl fluoride intermediate, providing a distinct strategy for accessing S-stereogenic sulfur(iv) compounds. The method exhibits a broad substrate scope, encompassing a wide range of sulfenamides and alcohols. Notably, an idebenone-derived sulfinimidate ester (3bg) exhibits prolonged intracellular persistence, while both S-3bg and racemic 3bg exhibit improved anti-inflammatory activity. This work provides a new strategy for asymmetric transformation of sulfilimidoyl fluoride intermediates and expands access to synthetically and biologically valuable sulfur stereocenters for drug discovery.

Dynamic kinetic resolution enables the first catalytic asymmetric transformation of sulfilimidoyl fluorides to chiral sulfinimidate esters, giving idebenone-derived sulfilimine 3bg with improved metabolic stability and anti-inflammatory activity.

Introduction

The rational design of bioactive molecules increasingly relies on the strategic incorporation of bioisosteric functional groups to modulate physicochemical, pharmacokinetic, and pharmacodynamic properties.^1–4 In this context, sulfoximines have emerged as privileged aza-sulfur(vi) motifs,^5–10 with well-documented utility as surrogates for sulfones and sulfonamides (Scheme 1A, left). Their successful application in pharmaceutical development—exemplified by Ceralasertib, an ATR inhibitor currently in Phase III clinical trials—has established sulfoximines as robust design elements in both medicinal and agrochemical chemistry.^11–18

Scheme 1. (A) Bioisosteric design in drug development: established sulfoximines versus underexplored sulfilimines. (B) Sulfonimidoyl fluorides as enabling intermediates for asymmetric synthesis of sulfoximines. (C) Key challenges associated with sulfilimidoyl fluorides in asymmetric catalysis. (D) Asymmetric oxidative halogenation for the construction of chiral S(iv) centers. (E) This work: catalytic asymmetric synthesis of sulfinimidate esters via dynamic kinetic resolution of sulfilimidoyl fluorides.

By contrast, sulfilimines, the mono-aza analogues of sulfoxides featuring a stereogenic sulfur center and a tunable nitrogen substituent, remain significantly underexplored.^19–21 This disparity largely stems from the inherent synthetic challenges associated with sulfilimine construction, including the difficulty of achieving precise stereochemical control and the propensity for overoxidation during the reaction process. Consequently, sulfilimines are underrepresented in lead-compound libraries, limiting the exploration of their biological and physicochemical properties. Recent studies, however, have demonstrated that selected sulfilimines exhibit high chemical and configurational stability under physiologically relevant conditions,^22–36 along with favorable drug-like properties such as enhanced aqueous solubility, improved metabolic stability, and reduced plasma protein binding (Scheme 1A, right).^37,38 As such, sulfilimines represent a highly promising yet underexplored class of S(iv)-based scaffolds, and the development of general and efficient stereoselective synthetic methods, together with systematic investigations into their application in drug discovery, is therefore highly desirable.

In recent years, sulfonimidoyl fluorides have emerged as versatile chiral S(vi) electrophiles for stereospecific sulfur(vi) fluoride exchange (SuFEx) reactions (Scheme 1B).^39–41 Notable advances include Bull's stereospecific fluorination of chiral sulfinamides with electrophilic fluorine reagents,^42–44 Gao's bis-oxazoline-mediated enantioselective desymmetrization of symmetrical iminosulfur oxydifluorides,⁴⁵ and Lopchuk's development of bench-stable chiral S(vi) transfer reagents derived from tert-butylsulfinamides.^46,47 More recently, Li and co-workers reported a catalytic asymmetric synthesis of chiral sulfonimidoyl fluorides from prochiral sulfenamides via dynamic kinetic hydrolysis and stereospecific oxidative fluorination.⁴⁸ These chiral sulfonimidoyl fluorides can undergo diverse nucleophilic substitutions to deliver sulfoximines, sulfonimidates, and sulfonimidamides with high efficiency and excellent stereocontrol, thereby significantly expanding the synthetic utility of SuFEx chemistry.

In sharp contrast, sulfilimidoyl fluorides—the sulfur(iv) analogues—have yet to be successfully applied in the asymmetric synthesis of sulfilimines (Scheme 1C).^49–51 Their limited utility may arise from several formidable challenges. First, rapid racemization at the stereogenic sulfur center precludes the isolation of configurationally stable, enantioenriched intermediates.⁵² Second, their inherent instability necessitates precise synchronization between in situ generation and nucleophilic trapping. Third, the catalytic systems capable of achieving both high reactivity and enantioselectivity under these conditions remain elusive. Given the emerging significance of chiral sulfilimines in medicinal chemistry and their unique S–N connectivity, establishing general and effective asymmetric strategies based on sulfilimidoyl fluoride intermediates constitutes a highly valuable yet unresolved objective in modern catalysis.

Although elegant approaches to sulfinamides/sulfinimidate esters via halogenative oxidation have been developed very recently (Scheme 1D),^52–54 these methods rely on stereocontrol through the generation of stereochemically defined sulfinimidoyl iodide/chloride intermediates. In these approaches, stereochemical information is introduced during formation of a configurationally defined sulfinimidoyl halide and then transferred through stereospecific substitution. By contrast, asymmetric strategies that utilize configurationally labile sulfilimidoyl fluoride intermediates and achieve stereocontrol during their subsequent transformation remain largely unexplored.

Herein, we report the first catalytic asymmetric transformation of sulfilimidoyl fluorides, providing direct access to enantioenriched sulfinimidate esters through a dynamic kinetic resolution (DKR) strategy (Scheme 1E). Mechanistic studies support the involvement of a transient sulfilimidoyl fluoride intermediate, whose configurational lability⁵⁵ enables the DKR process during the subsequent transformation. Enantioselectivity is achieved through a bifunctional chiral catalyst^56–59 that promotes selective conversion of this sulfur(iv) intermediate. The reaction exhibits broad substrate scope and good stereocontrol. Notably, an idebenone-derived sulfinimidate ester displays prolonged intracellular persistence compared to the parent compound, while both the racemic derivative and the (S)-enantiomer exhibit anti-inflammatory effects. These results demonstrate that incorporation of the sulfilimine motif can modulate biological properties and underscore the importance of further exploration of this emerging scaffold.

Results and discussion

Reaction development

To develop a mild protocol for the in situ generation of sulfilimidoyl fluorides and a catalytic system capable of synchronizing their consumption, we initiated our studies with the reaction between sulfenamide 1a and benzyl alcohol 2a (Table 1). Selectfluor was chosen as the electrophilic fluorinating agent for its safety, balanced reactivity and functional group compatibility, with Na₂CO₃ as the base. Initial chiral catalyst screening showed that the sodium salt of the anionic Λ-(S,S)-C1 complex gave sulfinimidate ester 3a in 64% yield with moderate enantioselectivity (38 : 62 er, entry 1). Chiral phosphoric acids and N-triflylphosphoramides (C2–C5) delivered the product in good yields but with only modest enantioselectivities (up to 33 : 67 er, entries 2–5), indicating that conventional Brønsted acid catalysis is insufficient for effective stereocontrol in this transformation.

Table 1. Optimization of reaction conditions^a.

Entrya	Cat.	Oxidant	Yield (%)	Er	Entry	Cat.	Oxidant	Yield (%)	Er
1	C1	Selectfluor	64	38 : 62	10	C10	Selectfluor	59	55 : 45
2	C2	Selectfluor	75	33 : 67	11	C11	Selectfluor	50	93 : 7
3	C3	Selectfluor	76	46 : 54	12	C12	Selectfluor	52	18 : 82
4	C4	Selectfluor	84	60 : 40	13	C11	NFSI	63	46 : 54
5	C5	Selectfluor	92	58 : 42	14	C11	NIS	80	62 : 38
6	C6	Selectfluor	61	77 : 23	15	C11	TCCA	60	56 : 44
7	C7	Selectfluor	48	41 : 59	16b	C11	Selectfluor	65	94 : 6
8	C8	Selectfluor	76	55 : 45	17b,c	C11	Selectfluor	90	82 : 18
9	C9	Selectfluor	72	57 : 43	18b,d	C11	Selectfluor	96	97 : 3

^aReaction conditions: 1a (0.1 mmol), 2a (0.15 mmol), catalyst (10 mol%), oxidant (0.15 mmol), Na₂CO₃ (0.15 mmol), MgSO₄ (30 mg) in toluene (1.0 mL) at −20 °C. Isolated yields.

^bCaSO₄ (30 mg) was used instead of MgSO₄.

^cAt 0 °C.

^dC11 (20 mol%) and 2a (0.2 mmol) were used.

To exploit potential cooperative effects of Brønsted bases and hydrogen-bond donors, we screened various bifunctional tertiary amine–hydrogen-bond donor catalysts. Cinchona alkaloid-derived urea, thiourea, and squaramide catalysts (C6–C8) provided the desired product with improved but still moderate enantioselectivities (entries 6–8). Alternative diamine-based scaffolds (C9 and C10) proved less effective (entries 9 and 10), highlighting the importance of a well-defined bifunctional framework. Further optimization of the hydrogen-bond donor motif proved crucial. Replacement of the 3,5-bis(trifluoromethyl)aniline unit in C8 with the corresponding benzylamine-derived squaramide in C11 led to a significant improvement in enantioselectivity (93 : 7 er, entry 11). This result suggests that increased conformational flexibility of the aryl substituent may facilitate more effective noncovalent interactions in the stereodetermining transition state. The corresponding diastereomer C12 delivered inferior selectivity (entry 12), underscoring the critical role of catalyst geometry.

Evaluation of alternative halogenating reagents (NFSI, NIS, and TCCA) resulted in markedly diminished enantioselectivities (entries 13–15), highlighting the unique effectiveness of Selectfluor in enabling both efficient fluorination and subsequent asymmetric induction. Further optimization of drying agents, bases, solvents, temperature, and catalyst loading established the optimal conditions as 20 mol% of C11 with CaSO₄ as the drying agent at −20 °C, affording 3a in 96% yield and 97 : 3 er (entry 18).

Substrate scope

To examine the generality and functional group tolerance of our method, a diverse array of sulfenamide substrates 1 was subjected to the optimized reaction conditions with benzyl alcohol 2a as the nucleophile (Scheme 2). Sulfenamides bearing both electron-rich and electron-deficient substituents at the different positions of the aryl group were well tolerated, delivering the desired sulfinimidate esters 3b–3g in 52–93% yields and good to excellent enantioselectivities (up to 98 : 2 er). Sterically demanding ortho-substituted substrates bearing methyl (3h) and halogen (3i) groups were also efficiently transformed, highlighting the excellent compatibility of the protocol with substituents having diverse electronic and steric effects. The methodology accommodates extended aromatic systems, such as naphthyl (3j), as well as heterocycles like pyridine (3k) and thiophene (3l).

Scheme 2. Substrate scope of sulfenamide substrates. Reaction conditions: 1 (0.1 mmol), 2a (0.2 mmol), C11 (20 mol%), Selectfluor (0.15 mmol), Na₂CO₃ (0.15 mmol), CaSO₄ (30.0 mg) in toluene (1.0 mL) at −20 °C for 60 h. Isolated yields. Er values were determined by HPLC analysis.

In addition to aryl substrates, a range of aliphatic sulfenamides, including linear (3m, 3n) and cyclic (3o, 3p), afforded the corresponding sulfinimidate esters with good yields (up to 96 : 4 er). Variation of the N-acyl group was also well tolerated (3q–3ad), with most substrates affording the desired products in good to excellent yields and enantioselectivities. A broad range of N-acyl substrates bearing diverse substituents on the aromatic ring—including electron-withdrawing groups (Br, F, CF₃, CN, NO₂, Cl) and electron-donating groups (alkoxy groups)—are well tolerated (3q–3aa). The methodology further accommodates extended aromatic systems such as 2-naphthyl (3ab), 1-naphthyl (3ac), and thiophene (3ad). Good enantioselectivity is consistently observed, with enantiomeric ratios up to 97 : 3 er and the majority of substrates exceeding 95 : 5 er. These results collectively demonstrate the generality of the method across a wide range of electronically and structurally diverse sulfenamides. The reaction was also tested with the substrate N-(p-tolylthio)pivalamide under the standard conditions. Although the desired product was obtained, the reaction proceeded with only low ee, indicating that an alkyl substituent adversely affects catalyst–substrate chiral recognition and asymmetric induction.

We next examined the scope of the alcohol component (Scheme 3). A broad range of primary and secondary alcohols, including methanol (3ae), ethanol (3af), isopropanol (3ah), cycloheptanol (3ai), and cyclohexanol (3ak), reacted smoothly to afford the corresponding sulfinimidate esters with good to excellent enantioselectivities (43–86% yields, 92 : 8 to 99 : 1 er). Both electron-rich and electron-deficient benzyl alcohol derivatives (3al–3as) furnished the desired products in consistently high yields and with moderate to high enantioselectivities (up to 94% yield, 97 : 3 er). Furthermore, functionalized alcohols bearing heterocycles and bulky frameworks such as thiophene (3at), phenoxy (3au), dihydro-1H-indene (3av), adamantyl (3aw), and 1,4-dioxaspiro[4.5]decane (3ax) also participated smoothly in the transformation, underscoring the versatility of the protocol. The effect of different substituents on the sulfenamide moiety was next evaluated through reactions of sulfenamides bearing different aryl electronic properties with cyclohexanol. The results indicated that all the sulfenamides reacted smoothly, delivering the desired products (3ay–3bb). In addition, the 3-chloro-substituted sulfenamide also performed well with cyclobutanemethanol (3bc) and indanol (3bd).

Scheme 3. Substrate scope of sulfenamide and alcohol substrates. Reaction conditions: 1 (0.1 mmol), 2 (0.2 mmol), C11 (20 mol%), Selectfluor (0.15 mmol), Na₂CO₃ (0.15 mmol), CaSO₄ (30.0 mg) in toluene (1.0 mL) at −20 °C for 60 h. Isolated yields. Er and dr values were determined by HPLC analysis.

The utility of this method was further demonstrated through late-stage functionalization (LSF) of complex molecules. (+)-Menthol (3be), ethenzamide (3bf), idebenone (3bg), metronidazole (3bi), pregnenolone (3bk), and d-galactopyranose (3bl) were successfully converted to the corresponding sulfinimidate esters with good levels of stereocontrol. Additional pharmaceutical derivatives, such as oxaprozin and indometacin analogues, were also compatible substrates (3bh, 3bj). These results highlight the potential of this strategy for the diversification of structurally complex molecules.

Synthetic applications

The chiral sulfinimidate ester bearing an S(iv) stereocenter is configurationally stable and serves as a versatile platform for constructing diverse S-stereogenic scaffolds via straightforward stereospecific transformations. For example, treatment of 3a with a Grignard reagent smoothly gives sulfinimine 4, and subsequent oxidation with RuCl₃/NaIO₄ yields sulfoximine 5 in a stereospecific manner with a high overall yield (77% over two steps, Scheme 4A). The absolute configuration of the S(iv) stereocenter in product 3a was established as R by comparison with product 5 of known configuration.⁴⁸ In addition to the transformations described above, the obtained chiral sulfinimidate esters proved to be versatile intermediates that could be further transformed into a range of structurally diverse chiral derivatives. For instance, under the treatment of n-butyllithium, 3al underwent nucleophilic substitution with tryptamine to generate the sulfinimidamide product 6. Similarly, it underwent nucleophilic substitution with cyclopropylmagnesium bromide to afford the cyclopropyl-substituted product 7 in 89% yield with complete retention of enantioselectivity (Scheme 4B). Furthermore, subsequent nucleophilic addition of methylmagnesium bromide to 3al followed by oxidation successfully delivered compound 9 (Scheme 4C). Notably, the sulfoximine framework embedded in 9 could serve as a recyclable directing group for Ru-catalyzed ortho-C–H alkenylation of arenes (10), thus providing a convenient platform for further functionalization.

Scheme 4. Products derivation and identification of absolute configuration. (A) Stereospecific transformation of 3a. (B) Nucleophilic substitution of 3al. (C) Recyclable sulfoximine-directed C–H alkenylation.

To evaluate the potential of the sulfilimine motif as a bioisosteric modification in medicinal chemistry, we examined the impact of structural derivatization on the biological profile of idebenone (Scheme 5A).⁶⁰ We first assessed the intracellular persistence of idebenone and its derivatives in 293T cells over a time course following treatment (0, 1, 2, 4, 8, 12, 24, 36, 48, and 72 h). Idebenone was detected only at 1 h and 2 h post-administration, whereas its derivatives—rac-3bg, R-3bg, and S-3bg—were present at multiple later time points. Among these, S-3bg exhibited the most prolonged intracellular retention and was detectable up to 24 h. These findings suggest that derivatization extends the metabolic stability of idebenone in cells, with R-3bg demonstrating the weakest persistence, while rac-3bg showed slightly better persistence than R-3bg but significantly lower than S-3bg (Scheme 5B).

Scheme 5. Rac-3bg, R-3bg, and S-3bg display distinct metabolic stability and anti-inflammatory profiles. (A) Chemical structures of idebenone, rac-3bg, R-3bg, and S-3bg. (B) Time-dependent intracellular concentrations of idebenone and its derivatives. Data are presented as mean ± s.e.m.; n = 3. (C) Effects of idebenone and its derivatives on LPS-induced nitric oxide (NO) production in RAW264.7 macrophages. Data are shown as mean ± s.e.m.; n = 3, one-way ANOVA. (D and E) Immunoblots showing the levels of phosphorylated p65 (p-p65^Ser536) and phosphorylated p38 (p-p38^T180/Y182) in RAW264.7 cells under indicated treatments. Densitometric quantification of p-p65 and p-p38 protein levels normalized to β-actin. Data are shown as mean ± s.e.m.; n = 3, one-way ANOVA.

We next evaluated the anti-inflammatory potential of these compounds using an LPS-induced RAW264.7 macrophage model. Following LPS challenge, a marked increase in intracellular NO was observed. Treatment with idebenone or R-3bg did not significantly attenuate NO production, whereas both rac-3bg and S-3bg substantially suppressed LPS-induced NO generation (Scheme 5C), indicative of anti-inflammatory activity. To further delineate the anti-inflammatory mechanisms, we assessed the phosphorylation of key inflammatory mediators. LPS stimulation significantly up-regulated p-p65^Ser536 and p-p38^T180/Y182 levels. While idebenone showed no significant effect relative to the LPS-treated group, R-3bg selectively reduced p-p38 phosphorylation without affecting p-p65^Ser536. In contrast, both rac-3bg and S-3bg markedly suppressed the activation of p-p65^Ser536 and p-p38^T180/Y182 (Scheme 5D and E), further confirming their anti-inflammatory efficacy. Collectively, these results demonstrate that rac-3bg and S-3bg maintain potent anti-inflammatory activity at 24 h post-treatment. Structural modification of idebenone thus not only prolongs its intracellular presence but also enhances its anti-inflammatory function.

Mechanistic investigations

To gain further insight into the reaction pathway and verify the involvement of the proposed sulfilimidoyl fluoride intermediate, we performed time-resolved NMR monitoring experiments. Initial mass spectrometric analysis of the reaction mixture containing sulfenamide 1a, benzyl alcohol 2a, and Selectfluor revealed a species with a molecular weight consistent with the proposed sulfilimidoyl fluoride intermediate. However, owing to its intrinsic instability, this transient species could neither be isolated nor directly visualized by thin-layer chromatography (TLC). Because Brønsted acid catalysts were found to accelerate the reaction during preliminary catalyst screening, p-toluenesulfonic acid was employed as a catalyst for intermediate monitoring experiments to facilitate its accumulation (Scheme 6A). Time-resolved ¹⁹FNMR analysis revealed the appearance of a new resonance at −11.0 ppm after 0.5 h. The intensity of this signal gradually increased, reached a maximum between 1.5 and 3.5 h, and subsequently decreased, indicating the transient generation and consumption of a fluorinated intermediate during the reaction.

Scheme 6. Mechanistic investigations and proposed origin of stereocontrol. (A) Control experiments for monitoring reaction intermediate. (B) Control experiments for mechanistic analysis and identification of the enantiodetermining step.

To further characterize this species, we monitored the corresponding ¹HNMR spectra. Oxidation of the sulfur center from S(ii) in sulfenamide 1a to S(iv) in the proposed sulfilimidoyl fluoride intermediate is expected to increase the electron-withdrawing ability of sulfur, resulting in downfield shifts of the methyl protons (H_b) of the p-tolyl group. Consistent with this expectation, the new resonance appeared at δ 2.51. Importantly, the temporal evolution of these proton signals closely matched that of the ¹⁹FNMR resonance at −11.0 ppm, exhibiting simultaneous formation, accumulation, and disappearance. Together with the molecular-weight information from mass spectrometry and the characteristic chemical shift changes associated with sulfur oxidation, these observations provide strong evidence supporting the formation of the proposed sulfilimidoyl fluoride intermediate during the reaction.

To gain insight into the origin of stereocontrol, a series of control experiments were conducted (Scheme 6B). As mentioned before, modification of the catalyst structure revealed a pronounced effect on enantioselectivity. In particular, catalyst C11, featuring a benzyl-substituted squaramide motif, provided a significant enhancement in enantioselectivity relative to related analogues. This observation suggests that increased conformational flexibility of the hydrogen-bond donor unit may facilitate more effective substrate–catalyst organization in the stereodetermining transition state. In contrast, a cinchonidine-derived catalyst lacking a squaramide hydrogen-bond donor (C13) afforded product 3a with nearly racemic outcome, highlighting the importance of hydrogen-bonding interactions in achieving stereocontrol. A more nucleophilic catalyst (β-ICD, C14) delivered 3a with low enantioselectivity, suggesting that nucleophilic Lewis base activation is unlikely to be the dominant pathway in this transformation.⁵⁵ To further identify the enantiodetermining step, a stepwise experiment was performed in which the putative sulfilimidoyl fluoride intermediate was first generated under thermal conditions, followed by the addition of catalyst C11 and benzyl alcohol at −20 °C. Under these conditions, the desired product 3a was obtained with high enantioenrichment (92.5 : 7.5 er) but low yield, likely due to partial decomposition of the intermediate at elevated temperature. This result supports that stereocontrol is established during the subsequent substitution step rather than during the formation of the sulfilimidoyl fluoride.

Based on these observations, a plausible catalytic cycle is proposed (Scheme 7). The reaction is initiated by in situ formation of sulfilimidoyl fluoride I from sulfenamide 1a in the presence of Selectfluor. Intermediate I is proposed to undergo configurational interconversion at the sulfur center, thereby enabling a DKR process. In the presence of catalyst C11, the tertiary amine moiety is proposed to act as a Brønsted base to activate the alcohol nucleophile, while the squaramide unit engages the sulfilimidoyl fluoride through hydrogen-bonding interactions. Such bifunctional activation is expected to organize the reacting partners in a chiral environment, thereby favoring reaction of one enantiomer of intermediate I over the other. The resulting enantioselective substitution at the sulfur center leads to formation of the sulfinimidate ester product, with concomitant cleavage of the S–F bond and regeneration of the catalyst. Overall, the experimental observations are consistent with a DKR process in which stereocontrol is exerted during the nucleophilic substitution step.

Scheme 7. Proposed catalytic cycle and stereochemical model.

Conclusions

In summary, we have developed an enantioselective strategy for the synthesis of sulfinimidate esters via a DKR process. This work represents a rare example of catalytic asymmetric transformation of sulfilimidoyl fluorides, enabled by a bifunctional tertiary amine–hydrogen-bond donor catalyst that addresses challenges associated with stereocontrol at sulfur(iv) centers. Mechanistic studies suggest that the reaction proceeds through in situ formation of a configurationally labile sulfilimidoyl fluoride intermediate, with enantioselection occurring during the subsequent substitution step under catalyst control. Beyond its synthetic utility, incorporation of the sulfilimine motif into an idebenone scaffold was found to modulate intracellular persistence and inflammation-related cellular responses, highlighting the potential of chiral sulfur(iv) functionalities in medicinal chemistry.

Author contributions

H.-P. Li conducted the main experiments and prepared the SI. J. Zhang performed the biological activity experiments. Z. Zhao, M. Li and X. Jin prepared the substrates. X.-H. He, W. Huang, and C. Peng provided valuable discussions. G. Zhan and B. Han conceived and directed the project and wrote the paper.

Conflicts of interest

The authors declare no conflict of interest.

Data availability

CCDC 2474373 contains the supplementary crystallographic data for this paper.⁶¹

The data that support the findings of this study are available in the supplementary information (SI). Supplementary information: detailed experimental procedures, characterization data (NMR spectra, mass spectrometry data), and synthetic methods for relevant compounds. See DOI: https://doi.org/10.1039/d6sc06421c.