Synthesis of chiral sulfilimines by organocatalytic enantioselective sulfur alkylation of sulfenamides

Abstract

Sulfilimines are versatile synthetic intermediates and important moieties in bioactive molecules. However, their applications in drug discovery are underexplored, and efficient asymmetric synthetic methods are highly desirable. Here, we report a transition metal–free pentanidium-catalyzed sulfur alkylation of sulfenamides with exclusive chemoselectivity over nitrogen and high enantioselectivity. The reaction conditions were mild, and a wide range of enantioenriched aryl and alkyl sulfilimines were obtained. The synthetic utility and practicability of this robust protocol were further demonstrated through gram-scale reactions and late-stage functionalization of drugs.

An organocatalytic platform for the enantioselective synthesis of chiral sulfilimines is established.

INTRODUCTION

Recent years have witnessed the rapid development of asymmetric synthetic methods for sulfur stereogenic centers (1–3), laying the foundation for wide applications in drug discovery (4–7). Replacement of carbon atoms with nitrogen atoms has become a popular strategy for drug optimization, and the introduction of nitrogen atoms can greatly improve the pharmacological properties for some drug candidates (8, 9), which is also applicable for sulfur stereogenic centers. This is proven by the increasing importance of chiral sulfoximines (10, 11), the aza-analogs of sulfones. The N-substitution at the sulfur center can generate intrinsic chirality and provide additional modification sites. The lone pair of the nitrogen atom also makes it Lewis basic for potentially favorable hydrogen bonding and coordination. Because of these properties, the incorporation of nitrogen can increase structural diversity, modulate conformational preferences, and improve pharmacological parameters and physicochemical properties (Fig. 1, A and B) (12). These beneficial effects are significant during drug development. Therefore, sulfoximines (13–15) and other sulfur stereogenic centers containing S─N bonds [e.g., sulfonimidoyl halides (16–20), sulfonimidates (21, 22), sulfonimidamides (23–25), sulfinamides (26–30), and thio-oxazolidinone (31, 32)] have drawn lots of attention from both academia and industry, and several candidates have entered clinical trials (33–35).

Fig. 1. Importance of sulfur stereogenic centers containing S─N bonds and enantioselective synthesis of sulfilimines.

(A) Sulfur stereogenic centers containing S─N bonds. (B) Bioactive molecules containing S─N bonds. (C) Synthesis of chiral sulfilimines by enantioselective sulfur alkylation of sulfenamides (Ellman group). (D) Pentanidium-catalyzed enantioselective sulfur alkylation of sulfenamides (this work). Me, methyl; tBu, tert-butyl; Bn, benzyl; e.r., enantiomeric ratio.

Sulfur can exist in multiple oxidation states and form diverse S(IV) and S(VI) stereogenic centers. Sulfoxide is a long and well-studied S(IV) stereogenic center (36–38), which is an important moiety in some marketed drugs such as esomeprazole and armodafinil. In the meanwhile, chiral sulfilimines, the aza-analogs of sulfoxides, have been greatly underestimated in drug discovery likely due to the limited synthetic methods and the resulting unclear physiochemical properties (39–43). Sulfilimines are underdeveloped S(IV)-derived scaffolds, which can serve as versatile intermediates for other sulfur stereogenic centers (44). Recent studies also proved the importance of sulfilimines in bioactive molecules such as agrochemicals (45, 46) and collagen IV networks (Fig. 1B) (47). To obtain enantioenriched sulfilimines, commonly used methods are stereospecific transformations of stoichiometric amounts of chiral reagents (48–52) and catalytic enantioselective imidation of thioethers (53–56). However, the imidation usually requires transition-metal catalysts and highly active electrophilic imidating reagents. Furthermore, the two thioether carbon substituents are required to be sterically or electronically differentiated to achieve high enantiocontrol. Thus, new strategies for the enantioselective synthesis of sulfilimines are highly desirable.

The Ellman group has recently reported the Rh-catalyzed enantioselective sulfur alkylation of sulfenamides (Fig. 1C) (57). They used diazo compounds as alkylation reagents and obtained chiral sulfilimines in high yields and with high enantioselectivities via attack of sulfenamide on rhodium carbenoid. Later, the transition metal–free sulfur alkylation of sulfenamides to produce racemic sulfilimines was presented by several groups (58–60). In particular, the Ellman group provided a proof-of-concept asymmetric alkylation using cinchonidine-derived catalysts (60).

Our group is working on pentanidium-based (27, 61, 62) organocatalysis and asymmetric synthesis of chiral sulfur structures. We and Tan group have developed an efficient method for the enantioselective synthesis of chiral sulfinate esters with pentanidiums as chiral cationic catalysts (27). The obtained enantioenriched sulfinate esters are easily transformed to diverse S(IV) and S(VI) stereogenic centers except sulfilimines, which has led us to exploit previously unknown methodologies. Here, we report an efficient pentanidium-catalyzed sulfur alkylation of sulfenamides with high enantiocontrol and wide substrate scope (Fig. 1D). Furtherly, we demonstrated the synthetic utility and practicability through gram-scale reactions (with complete recovery of chiral catalysts) and late-stage functionalization of drugs.

RESULTS

Sulfenamides are versatile and known for many useful transformations. After deprotonation under basic conditions, sulfenamides form prochiral sulfenamide anions, which are ambident nucleophiles and result in the challenging control of selective sulfur or nitrogen alkylation. On the basis of our and Tan group’s previous work (27, 62), we proposed to modulate the reaction sites and directions to control the chemoselectivity and enantioselectivity using pentanidiums as chiral cationic catalysts to form ion pairs with the sulfenamide anions. We started our investigations using tetrabutylammonium bromide as the achiral cationic catalyst to study the influence of N-substitutions of sulfenamides and alkylation reagents (Fig. 2). The reaction between the sulfenamide N-protected by ethyloxycarbonyl group and benzyl bromide gave a mixture of S/N-alkylation products in 55 and 44% yields, respectively (Fig. 2, entry 1). Then, alkylation reagents with different leaving groups were tried. Benzyl chloride, benzyl 4-toluenesulfonate, and benzyl 4-nitrobenzenesulfonate all gave a mixture of S/N-alkylation products (Fig. 2, entries 2 to 4). Following investigations with sulfenamides protected by other oxycarbonyl groups including methyloxycarbonyl, benzyloxycarbonyl, and 2-(trimethylsilyl)ethoxycarbonyl groups gave similar S/N-alkylation ratios without obvious improvements of chemoselectivity (Fig. 2, entries 5 to 7). 9-Fluorenylmethoxycarbonyl–protecting group made the sulfenamide less active, and only S-alkylation product was isolated in 34% yield (Fig. 2, entry 8). Sulfenamides N-protected by acyl protecting groups including benzoyl, phenylacetyl, pivaloyl, and trifluoroacetyl groups were synthesized (Fig. 2, entries 9 to 12). These sulfenamides showed high chemoselectivity, and only S-alkylation products were obtained.

Fig. 2. Optimization of substrates and reaction conditions.

^aReaction conditions: Sulfenamides 1a-1i (0.1 mmol), Bn-X (1.5 equiv), catalyst (3 mol %), 50% aqueous (aq.) KOH (3.0 equiv), toluene (1.0 mL), −20°C, and 24 hours. ^bReactions were conducted at 25°C for Bn-Cl. ^cReaction conditions: Sulfenamide 1h (R = COtBu) (0.1 mmol), Bn-Br (1.5 equiv), catalyst (3 mol %), base (3.0 equiv), solvent (1.0 ml), temperature (T; °C), and 24 hours. Isolated yields were reported, and e.e. values were determined by chiral high-performance liquid chromatography (HPLC) analysis. Teoc, 2-(trimethylsilyl)ethoxycarbonyl; Fmoc, 9-fluorenylmethoxycarbonyl; Ts, p-toluenesulfonyl; Ns, 4-nitrobenzenesulfonyl; EA, ethyl acetate; e.e., enantiomeric excess.

Pentanidiums were later used as chiral catalysts to make the reaction enantioselective. When PN1 [Cambridge Crystallographic Data Center (CCDC), 2346344] was used, only S-alkylation products were obtained for all sulfenamides. For sulfenamides protected by oxycarbonyl groups (Fig. 2, entries 1 and 5 to 8), the products were isolated in high yields but with moderate enantiomeric excess (e.e.) values in reaction with benzyl bromide. The leaving groups of alkylation reagents were also important for the enantiocontrol. The enantioselectivities using benzyl chloride, benzyl 4-toluenesulfonate, and benzyl 4-nitrobenzenesulfonate were not satisfying (Fig. 2, entries 2 to 4). For sulfenamides protected by acyl groups (Fig. 2, entries 9 to 12), the reactions proceeded with high enantiocontrol, and up to 93% e.e. value was obtained for sulfenamides 1f (Fig. 2, entry 9) and 1h (Fig. 2, entry 11) protected by benzoyl and pivaloyl groups. Pivaloyl-protected sulfenamide 1h and benzyl bromide were then used as models to investigate the influence of reaction conditions on the enantioselectivity. Parameters including catalysts, bases, solvents, and temperatures were carefully examined (Fig. 2, entries 13 to 24, and see the Supplementary Materials for details). The reaction conditions comprising PN1 and 50% aqueous KOH solution in toluene at −20°C were found to be optimal and gave sulfilimine 2h in 99% yield and 93% e.e. value (Fig. 2, entry 13). The reactions catalyzed by PN2 and PN3 gave products with 91 and 88% e.e. values (Fig. 2, entries 14 and 15), and PN4 resulted in unsatisfactory enantiocontrol (Fig. 2, entry 16). The influence of different bases was investigated. Cesium carbonate was not efficient, and only trace products were detected (Fig. 2, entry 17). When sodium hydroxide was used, the product 2h was isolated in low yield and with 84% e.e. value (Fig. 2, entry 18). To our delight, potassium hydroxide well promoted the reaction, and 2h was isolated in 98% yield and 93% e.e. value (entry 19). Following screening of solvents showed the reactions in dichloromethane, and diethyl ether and ethyl acetate gave low to moderate enantioselectivities (Fig. 2, entries 20 to 22). Reactions at 0°C offered 2h in 99% yield and with 85% e.e. value (Fig. 2, entry 23). When the reaction temperature was decreased to −40°C, the e.e. value was increased to 99%, but the product was only isolated in 43% yield (Fig. 2, entry 24).

With the optimized N-protecting group and reaction conditions in hand, the substrate scope was examined (Fig. 3). A series of aryl sulfenamides with different substitutions were investigated (Fig. 3A). Phenyl sulfenamide reacted efficiently to give the sulfilimine 2j in 97% yield and with 91% e.e. value. Several aryl sulfenamides substituted by different alkyl groups including sterically hindered 2-methyl, 3,5-dimethyl, and 4-tertbutyl groups all proceeded smoothly to give products 2k-2m in high yields and with high enantiocontrol. Aryl sulfenamides N-protected by pivaloyl or trifluoroacetyl groups with strong electron-donating methoxy group at para-position were also synthesized. Their reactions with benzyl bromides both gave products 2n and 2o with 95% e.e. value. This reaction was also efficient to yield a variety of sulfilimines 2p-2u substituted with halogen atoms at different positions. Electron-withdrawing trifluoromethyl, ester, aldehyde, and nitro groups were also well tolerated, and the expected products 2v-2y were obtained with high enantioselectivities. Naphthyl group was suitable for our methodology, and product 2z was obtained with 97% e.e. value. Electron-rich and electron-poor heteroaryl sulfenamides afforded the corresponding sulfilimines 2aa-2ac in high yields and with high enantioselectivities. Then, the reactions with alkyl sulfenamides were tired. To our delight, primary, secondary, and tertiary alkyl sulfenamides all proceeded well to give the expected sulfilimines 2ad-2aj with high enantiocontrol.

Fig. 3. Reaction scope.

(A) The scope of sulfenamides. (B) The scope of alkyl halides. (C) Late-stage modification of drugs. Reaction conditions: Sulfenamides (0.1 mmol), alkyl halides (1.5 equiv), PN1 (3 mol %), 50% aqueous KOH (3.0 equiv), toluene (1.0 ml), −20°C, and 24 hours. Isolated yields were reported, and e.e. values were determined by chiral HPLC analysis. ^aReaction was conducted at 0°C. ^bReaction was conducted at −40°C. Piv, pivaloyl.

Different alkyl halides were also investigated for the reaction (Fig. 3B). Reactions with benzyl bromides substituted by different halogen atoms and cyano groups afforded products 2ak-2ao with high enantioselectivities. 2-(Bromomethyl)naphthalene and 3-bromomethylthiophene were also excellent alkylation reagents to give products 2ap and 2aq with 97 and 92% e.e. values, respectively. Benzyl bromide substituted by two trifluoromethyl groups was also well tolerated to give product 2ar in 99% yield and with 96% e.e. value. Allylic bromides were used under our conditions, but we failed to obtain the expected sulfur allylation product 2at which was rapidly isomerized to achiral nitrogen allylation product 2at′ through [2,3]-sigmatropic rearrangement (39). When allylic bromide was substituted by phenyl group at the terminal carbon, the expected products 2au and 2av were obtained with high enantiocontrol. Reactions with 1-bromo-2-butyne afforded products 2aw and 2ax with 88 and 90% e.e. values, respectively. Other unactivated alkyl halides gave unsatisfactory results (see the Supplementary Materials for details). The absolute configurations of 2p, 2s, and 2ah were confirmed by x-ray crystallographic analysis (CCDC, 2346341, 2346342, and 2346343). We also used our methodology for the late-stage modification of marked drugs valdecoxib and celecoxib (Fig. 3C). The achiral sulfonamide groups of these drugs were successfully modified with S(IV) stereogenic centers. Drug derivatives sulfilimines 2ay and 2az were obtained with high enantiopurities.

To further demonstrate the synthetic utility and practicability of our methodology, we conducted gram-scale experiments and further transformations of sulfilimines (Fig. 4). Both sulfenamides N-protected by pivaloyl and trifluoroacetyl groups proceeded well on the gram scale without loss of yields and enantioselectiveities, giving 2.72 and 2.41 g of sulfilimines, respectively (Fig. 4A). It is remarkable that our chiral catalysts can be completely recovered (>99%) by column chromatography and can catalyze enantioselective reactions without loss of reactivity. Sulfilimine 2as was subjected to stereospecific oxidation and amination to afford the enantioenriched sulfoximine 4 and sulfondiimine 5 (Fig. 4B). The pivaloyl group of sulfoximine 4 was removed under basic conditions to give N-unprotected sulfoximine 6, which was the intermediate for the synthesis of BAY 1143572 (34). The trifluoroacetyl protection group of the sulfilimine 2i was also removed, and the resulting N-unprotected sulfilimine 7 was directly subjected to next steps without further purification (Fig. 4C). The N atom of the free sulfilimine 7 is an ideal reaction site for late-stage modification of drugs which was successfully coupled with several drug molecules. The sulfonyl chloride intermediates of celecoxib and sildenafil reacted with the N atom of 7 to afford the drug derivatives 8 and 9 with 92 and 94% e.e. values, respectively. The N─H-free sulfilimine 7 was also able to condense with carboxylic acid–containing drugs under coupling conditions using 1-(3-dimethylaminopropyl)-3-ethylcarbodiimide hydrochloride. Racemic lesinurad with axial chirality, (S)-ibuprofen, penicillin G potassium salt, and lumacaftor reacted smoothly with 7 to afford the respective condensation products 10-13. These drug derivatives incorporated with the chiral sulfiliminyl moieties are previously underdeveloped and can expand the chemical space for drug discovery.

Fig. 4. Gram-scale synthesis and product derivatizations.

(A) Gram-scale synthesis. (B) Stereospecific oxidation/amination of 2as and synthesis of BAY 1143572. (C) Synthesis of N─H-free sulfilimines and incorporation into drugs. Isolated yields were reported, e.e. values were determined by chiral HPLC analysis, and d.r. values were determined by chiral HPLC or NMR analysis. rt, room temperature; EDCI, 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide hydrochloride; HOBt, 1-hydroxybenzotriazole; d.r., diastereomeric ratio.

To prove the proposed sulfenamide anionic intermediate, deuteration experiments of 1h with different bases were conducted (Fig. 5A). When KOH was used, we got the product with 99% D-incorporation. This supported the deprotonation process of starting material. On the basis of the absolute configuration of products and previous reports about pentanidium catalysts (63), the model for stereoinduction was proposed (Fig. 5B). The cationic catalyst formed ion pair with the sulfenamide anion. The benzyl hydrogen atoms of catalysts formed multiple hydrogen bonds with the sulfenamide anion and benzyl bromide, which directed the spatial orientation of nucleophilic substitution to account for the observed enantioselectivity.

Fig. 5. Mechanistic studies.

(A) Deuteration with different bases. (B) Proposed model for stereoinduction.

DISCUSSION

In summary, we have developed an efficient enantioselective S-alkylation of sulfenamides for the synthesis of chiral sulfilimines with pentanidiums as the chiral cationic catalysts. This methodology shows exclusive chemoselectivity for sulfur over nitrogen and high enantioselectivity. A variety of enantioenriched sulfilimines were obtained in excellent yields with high enantiocontrol, and the gram-scale synthesis further demonstrated the practicability. The mild conditions and wide substrate tolerance render the reaction suitable for late-stage modification of drugs. The achiral sulfonamide groups of celecoxib and valdecoxib were efficiently replaced by enantioenriched sulfiliminyl moieties using the established method. The utility of the enantioenriched sulfilimine products was also demonstrated by the condensation with sulfonyl chlorides or carboxylic acid–containing drugs. We anticipate that this methodology and these previously unknown drug derivatives incorporated with chiral sulfiliminyl moieties can unearth the potential of sulfur stereogenic centers in the discovery and development of pharmaceuticals.

MATERIALS AND METHODS

Commercially available materials and other solvents purchased from commercial suppliers were used as received. ¹H, ¹³C, and ¹⁹F nuclear magnetic resonance (NMR) spectra were recorded on JEOL JNMECZ400S (400 MHz for ¹H, 100 MHz for ¹³C, and 376 MHz for ¹⁹F) with CDCl₃ as the solvent and tetramethylsilane as the internal standard. Chemical shifts were recorded as δ in units of parts per million (ppm). The residual solvent peak was used as an internal reference (CDCl₃: δ 7.26 ppm for ¹H NMR and δ 77.0 ppm for ¹³C NMR). Multiplicity was indicated as follows: s (singlet), d (doublet), t (triplet), q (quartet), m (multiplet), and br (broad). High-resolution mass spectrometry was recorded on a quadrupole time-of-flight mass spectrometry (AB SCIEX X500R with ESI source and Agilent 7250 with EI source), which combines quadrupole precursor ion selection and a high-resolution accurate-mass time-of-flight mass analyzer to deliver mass accuracy. e.e. values were determined by high-performance liquid chromatography (HPLC) analysis on Shimadzu LC-2050 HPLC workstations. Optical rotations were measured using a 2-ml cell with a 0.5-dm path length on TH-P300 polarimeter with a sodium lamp of wavelength of 589 nm and reported as follows: [α]^T_D (c g/100 ml, solvent). Flash chromatography separations were performed on silica gel (200 to 300 mesh, Huanghai). Visualization was performed using an ultraviolet lamp or potassium permanganate stain. Procedures involving air- or moisture-sensitive materials were conducted with anhydrous solvents under an inert atmosphere of nitrogen or argon using standard Schlenk techniques.

General reaction procedure

To a glass via (4 ml), sulfenamide 1 (0.1 mmol), PN1 (3.9 mg, 3 mol %), toluene (1 ml), 50% aqueous KOH (0.3 mmol) were sequentially added. The mixture was cooled to −20°C and stirred for 3 min. Then, alkyl halides (0.15 mmol) were added. The mixture was stirred at −20°C for 24 hours and then purified by flash chromatography column to give the desired chiral sulfilimine 2.

Supplementary Materials

This PDF file includes:

Supplementary Text

Tables S1 to S5

References