Enantioselective Inter- and Intramolecular Sulfenofunctionalization of Unactivated Cyclic and (Z)-Alkenes

Abstract

A method for the enantioselective, Lewis base-catalyzed sulfenofunctionalization of cyclic and (Z)-alkenes is reported. The intermediate thiiranium ion generated in the presence of a selenophosphoramide catalyst is intercepted by a variety of nucleophiles. A diverse array of inter- and intramolecular functionalizations proceed in high yield and good to high enantioselectivity (86:14–98:2 er). Prior experimental and computational studies indicated such enantiotopic face discrimination to be poor; however, the results disclosed herein remediate the previous findings. Control experiments were performed to investigate the different behavior of (Z)-alkenes and their more established (E)-counterparts.

Graphical Abstract

INTRODUCTION

Functionalization of unactivated olefins via enantioselective group transfer is a powerful strategy for generating molecular complexity from simple building blocks. Among them, a variety of epoxidations,¹ aziridinations,² halo-,³ sulfeno-,⁴ and selenofunctionalizations⁵ have been reported. The methods of greatest utility have demonstrated excellent substrate recognition of various double-bond substitution patterns to functionalize alkenes with high levels of diastereo- and enantiocontrol. However, such processes are rare, and the catalysts often require engineered substrates bearing coordinating groups to achieve high enantioselectivity.

Sulfenium ion transfer to unactivated alkenes by enantioselective thiiranium ion formation, pioneered in these laboratories and expanded upon by other groups, has been successfully applied for the functionalization of unactivated olefins.⁴ Enantioenriched thiiranium intermediates generated from mono-, di-, and trisubstituted alkenes can be captured by a variety of nucleophiles including oxygen,⁶ nitrogen,⁷ and carbon⁸ moieties, acting in both an intra- and intermolecular fashion (Scheme 1).

Scheme 1.

Generalized Asymmetric Sulfenofunctionalizations of Alkenes

Despite great advances in recent years, reports of enantioselective sulfenylations of both (Z)- and (E)-configured 1,2-disubstituted alkenes using one catalyst are extremely rare. For example, Shi’s sulfenoamination protocol catalyzed by chiral phosphoric acids is well suited for alkyl-substituted (Z)-amino-alkenes to form corresponding pyrrolidines and piperidines with high enantiomeric ratios (89:11–93:7 er), whereas moderate enantioselectivities were obtained when (E)-substituted substrates were employed (72:28–85:15 er).^7b In 2014, Zhao reported the enantioselective construction of trifluoromethylthiolated azaheterocycles in the presence of an indane-based, bifunctional, chiral selenide catalyst.^7d High levels of enantiocontrol were achieved using alkyl- and aryl-substituted (E)-alkenes; however, no product was observed when a (Z)-alkene was employed, which was attributed to inherent reactivity and steric hindrance in the alkene.⁹ It is clearly shown in Figure 1 that regardless of the trajectory for sulfenium ion transfer to the alkene, cis-alkenes necessarily have unavoidable steric repulsions rendering a higher energy transition state.

Figure 1.

Steric interactions of thiiranium ions.

Indeed, our own efforts to engage (Z)-configured 1,2-disubstituted substrates have been plagued by reduced reactivity and moderate or poor enantioselectivity. For example, in the seminal report from these laboratories on the enantioselective sulfenylation of simple alkenes, the enantiocontrol with substrate (E)-3 (96:4 er) was far superior to the corresponding Z-analogue (54:46 er) (Scheme 2A).^6a Later in 2014, another attempt to employ aryl-substituted (Z)-6 in an intramolecular sulfenoamination resulted in reduced reactivity (68% yield), poor endo/exo-selectivity, and weak enantioinduction (63:37 er) (Scheme 2B).^7a Several additional reports from these laboratories have shown a similar trend. These failures were explained owing to the topology of the BINAM-derived selenophosphoramide Lewis base catalyst and the corresponding active species using a diagonally symmetric quadrant model of occupied and unoccupied space.¹⁰ In addition, the presence of the bulky Lewis base makes both thiiranium ion intermediates stemming from the (Z)-alkene (syn and anti) more sterically hindered than its (E)-counterpart, disfavoring reactions with (Z)-alkenes.^10b,11

Scheme 2.

Sulfenofunctionalization of 1,2-Disubstituted (E)- vs (Z)-Alkenes

Hence, engaging both (Z)- and (E)-alkenes for sulfeno-functionalization reactions in a highly enantioselective fashion using a single catalyst remains a challenge and is addressed in this report.

RESULTS AND DISCUSSION

To investigate the reactivity of (Z)-alkenes, a general substrate that rules out the influence of coordinating groups (e.g., −OH or −NHR) on enantioselectivity was selected for the model reaction. On the basis of our recent report on the intermolecular sulfenoamination of alkenes,¹² the combination of (Z)-β-methylstyrene 9 and 4-bromoaniline was well suited for the catalyst survey (Table 1). All reactions were performed in 1,1,1,3,3,3-hexafluoroisopropyl alcohol (HFIP) using 2,6-diisopropylphenyl sulfenylating agent 10, which affords higher enantioselectivities compared to its less bulky analogues.¹⁰

Table 1.

Sulfenoamination of (Z)-β-Methylstyrene–Catalyst Screen^a,b,c

^aAll reactions performed on a 0.05 mmol scale.

^bYields determined by ¹H NMR spectroscopic analysis using 1,1,1,2-tetrachloroethane as an internal standard.

^cEnantiomeric ratio determined by CSP-HPLC analysis.

ND = not determined.

Orienting experiments to identify (Z)-alkene-selective Lewis bases began with bisimidazoline-derived selenophosphoramides, which previously showed promising enantioselectivity in bromocycloetherification.¹³ However, in this case, very low reactivity was observed using catalyst 12, with most of the starting materials returned unreacted. Because tetrahydrothiophene is a privileged Lewis base for sulfenium ion transfer, a chiral analogue was examined. Inspired by the desymmetrization of alkenoic diols with chiral tetrahydrothiophene 13¹⁴ on 2,2-disubstituted alkenes, it was synthesized and tested in its efficiency on the model reaction. Unfortunately, Lewis base 13 afforded no enantioinduction and product formation was very slow. Aggarwal’s isothiocineole 14¹⁵ applied in sulfur-ylide-mediated epoxidations and aziridinations resulted in good reactivity (76% yield); however, minimal levels of enantiocontrol were obtained (54:46 er). Dithiin 15 is reported by Pasquato and co-workers to generate sulfonium salts from MeSCl and SbCl₅ followed by stoichiometric delivery of the sulfenium ion to (E)-3-hexene with good selectivity.¹⁶ However, catalyst 15 was insoluble in HFIP and resulted in no reactivity. Good yield but moderate levels of enantioselectivity were observed using BINAS-based disulfide 16¹⁷ (66:34 er). BINOL-derived bifunctional Lewis base 17 bearing a free hydroxy group inspired by Shirakawa for applications in enantioselective bromolactonizations¹⁸ afforded high conversion to the desired product, albeit in a racemic fashion.

Owing to the lack of significant enantioinduction, it was decided to benchmark these results against catalysts previously developed in this group. Selenophosphoramidate (R)-18 and selenophosphoramide (R)-2b were subjected to the model reaction conditions. Whereas (R)-18 afforded promising enantioselectivity (89:11 er), much to our surprise, using BINAM-derived (R)-2b, sulfenoamination product 11a was isolated in 90% yield and 97:3 er. Importantly, comparable yield (80%) and selectivity (98:2 er) were obtained for the sulfenoamination of (E)-β-methylstyrene with 4-bromoaniline under similar reaction conditions.¹² The high enantioselectivity observed contradicted what has been established over the last 10 years, namely, (Z)-alkenes were not well recognized by these catalyst scaffolds.^5a,6a With this unexpected result in hand, the generality of catalyst (R)-2b using a variety of (Z)-alkenes in both inter- and intramolecular sulfenofunctionalization reactions was investigated using 4-chloroaniline as the model nucleophile (Table 2).

Table 2.

Intermolecular Sulfenofunctionalization of Z-Alkenes^a,b,c,d,ef

^aAll reactions performed on a 1.00 mmol scale.

^bYield of isolated, analytically pure product.

^cDiastereomeric ratio determined by ¹H NMR spectroscopic analysis of the crude reaction mixture. Unless otherwise noted, the dr of the products is >99:1.

^dEnantiomeric ratio determined by CSP-HPLC analysis.

^eEnantiomeric ratio determined on the dinitrobenzoyl derivative.

^fObtained as a mixture of constitutional isomers.

The influence of steric bulk on the sulfenium ion transfer and the electronic effects were explored by changing the substituents near the alkene fragment. No profound steric effect was observed in the presence of an ortho-methyl group such that the corresponding sulfenyl amine 11b was formed in good yield and enantiocontrol. In the case of an electron-withdrawing CF₃-group, the reactivity was maintained, and desired product 11c was isolated in synthetically useful yield (78%) and enantioselectivity (98:2 er). Having a methoxy group in the 4-position resulted in the formation of 11d as a diastereomeric mixture (64:36 dr) favoring the formation of the anti-product. Both diastereoisomers were obtained with high enantiocontrol (96:4 and 98:2 er). This reaction likely proceeded through an open carbocation intermediate stabilized by the strongly electron-donating nature of the methoxy group. Preferential formation of anti-11d is attributed to relieving A_1,3-strain of the methyl group with the aromatic ring prior to interception by the nucleophile. In addition, (Z)-β-ethylstyrene reacted with exceptional yield and enantioselectivity (11e), while (Z)-β-isopropylstyrene had a precipitous drop in conversion but retained satisfactorily high enantiocontrol (11f).

Cyclic alkenes were also explored since the challenge of enantioselective sulfenium ion transfer to an internal double bond is well known.^7b Gratifyingly, the reaction of indene with 4-chloroaniline gave 11g in excellent yield and enantioselectivity (Table 2). It is worth noting that the indane-based chiral bifunctional sulfide catalyst, developed by Zhao and co-workers for asymmetric sulfenofunctionalization,^{6c–f,7c,d,8a,b} can be accessed through this process in a single synthetic step. 1,2-Dihydronaphthalene behaved similarly to indene to afford near-quantitative conversion to highly enantioenriched product 11h. 2H-Chromene analogues represent a privileged class of substrates for medicinal chemistry¹⁹ and chromene derivative 11i was formed as a single diastereomer albeit with reduced enantioselectivity (86:14 er). The exact reason for the stereochemical erosion using this substrate is yet to be determined.

The scope of the reaction was extended by surveying the nucleophilic partners, for which (Z)-β-methylstyrene was chosen as the model alkene (Table 2). It is important to note that a judicious selection of nucleophiles allowed the suppression of a competitive opening of the thiiranium ion by HFIP.²⁰ In general, to avoid the oxysulfenylated product arising from solvent capture, the amine nucleophile must not be basic enough to deprotonate HFIP, thus rendering it inactive, but nucleophilic enough to perform the desired transformation. In addition to anilines, the reactivity of secondary and primary amines was examined. Morpholine and piperazine are abundantly used as substituents or scaffolds in drugs and medicinal chemistry owing to their positive impact on the ADME properties.²¹ Accordingly, products 11k and 11l were obtained in good yield (~80%) and with consistently high enantioselectivity (95:5 and 94:6 er, respectively). Although tosylamide could not outcompete HFIP, the more acidic triflamide was a competent nucleophile to furnish 11m in 69% yield and 90:10 er. To establish the absolute stereochemical course for the reactions of (Z)-alkenes, the full stereostructure of 11m was determined by X-ray crystallographic analysis and was shown to be R,R.²² This outcome comports with the observed stereochemical pathway for (E)-alkenes using the same catalyst. The scope was further extended to primary amines and other anilines, with benzyl-amine providing 11n in good yield and enantioselectivity, whereas p-anisidine (11o) and 2-fluoroaniline (11p) afford the sulfenoamination products in near-quantitative yields. Next, oxygen nucleophiles were explored. 4-Nitrophenol afforded highly enantioenriched product 11q in 96% yield. Incorporation of oxygen using TMSOAc resulted in the desired product formation (11r) with good enantiocontrol (90:10 er). Finally, benzoic acid was also competent leading to the formation of benzoate 11s in 96% yield and 93:7 er. It is notable that both carboxylate nucleophiles suffered from poor selectivity between the α and β-positions of the styrene. It is posited that the initial attack of the nucleophile proceeded kinetically at the α-position but reformation of the thiiranium-ion with concomitant ejection of the nucleophile followed by recapture allowed an equilibration event to happen during the course of the reaction.

The reaction scope was further evaluated by probing (Z)-alkenes of different chain lengths bearing various nucleophilic moieties that can engage in intramolecular sulfenofunctionalizations and deliver useful heterocyclic motifs (Table 3). Phenyl-substituted alcohol 19a afforded corresponding ether 20a in 63% yield and 95:5 er. In addition, pyrrolidine 20b was obtained in a highly enantioenriched form (entry 2). As expected, lactonization of 4-methoxyphenyl-substituted carboxylic acid 19c proceeded through an open carbocation, which resulted in a mixture of diastereomers (20c), albeit in an improved ratio (80:20 dr) compared to the 4-methoxy-substituted substrate in the intermolecular sulfenofunctionalization (64:36 dr, Table 2). Both isomers were obtained in excellent yields and enantioselectivities (entry 3). Naphthyl-substituted lactone 20d derived from acid 19d was furnished with high efficiency (entry 4). Pyran 20e was formed as the exclusive product with no loss of enantioinduction. Finally, amide 19f affords oxazoline 20f with exclusive exo-selectivity albeit with extended reaction times owing to the reduced nucleophilicity of the double bond and perhaps solvation of the amide oxygen nucleophile. Starting material 19f remained after 48 h.

Table 3.

Intramolecular Sulfenofunctionalization of Z-Alkenes^a,b,c,def

^aAll reactions performed on a 1.00 mmol scale.

^bYield of isolated, analytically pure product.

^cDiastereomeric ratio determined by ¹H NMR spectroscopic analysis of the crude reaction mixture. Unless otherwise noted, the dr of the products is >99:1.

^dEnantiomeric ratio determined by CSP-HPLC analysis.

^eEnantiomeric ratio determined for major diastereomer.

^fEnantiomeric ratio determined on the corresponding sulfone.

Notably, all cyclizations of Z-alkenes led to the predominant formation of the kinetic exo-product as expected unless that product would be a four-membered ring.²³ The rationale for this phenomenon is depicted in Scheme 3A whereby the endo transition state for (Z)-alkenes necessarily has a destabilizing 1,3-diaxial interaction that is not observed in the exo-transition state. A head-to-head comparison is provided to demonstrate this effect on two nearly identical substrates, which vary only by alkene geometry (Scheme 3B). The trans-alkene affords the endo-cyclization product (pyran) exclusively both as the thermodynamic and the kinetic product. This sharply contrasts the cis-alkene that affords roughly a 4:1 mixture of the kinetic furan product as the major component. In light of the putative 1,3-diaxial interaction that favors exo-cyclization of cis-alkenes, a tertiary alcohol bearing a geminal dimethyl group was synthesized and subjected to standard reaction conditions (Scheme 3C) to examine how an increase in diaxial strain may improve endo/exo-selectivity. Exclusive formation of the furan provides further evidence for the role of nonbonding interactions governing the transition state, which leads to either five- or six-membered ring formation.

Scheme 3. Rationale for exo-Selective Cyclization of (Z)-Alkenes^a.

^aAll reactions performed on a 0.1 mmol scale. Yields in parentheses determined by ¹H NMR analysis of the crude reaction mixture using 1,1,2,2-tetrachloroethane as the internal standard. *Enantiomeric ratio determined on the corresponding sulfone.

The unexpected effectiveness of the BINAM-derived selenophosphoramide catalyst in enantiotopic face discrimination of (Z)-alkenes motivated a study to better understand the basis for its selectivity. Two key changes have been implemented since the inception of these research programs: the first is changing to a bulkier sulfenylating agent, which affords better selectivity, and the second is changing the solvent from CH₂Cl₂ to HFIP. To further examine how these factors may have conspired to afford such unexpected results, a series of studies were performed modulating the solvent and substrate. Accordingly, the (E)-isomer of 19a was prepared and subjected to cyclization conditions in both HFIP and CH₂Cl₂. As expected, trans-20a was produced with excellent enantioselectivity irrespective of solvent (Table 4). Furthermore, the same cyclization with (Z)-19a in dichloromethane afforded similar results as with HFIP. These preliminary findings suggest that solvent plays little role in the enantiodetermining step nor offers viable racemization pathways.^10b Furthermore, the significant increase in enantioinduction between the smaller –SPh sulfenylating agent 1 versus the bulkier agent 10 is noteworthy. For styryl alkenes, sulfenylating agent 1 afforded a 63:37 er with an amine nucleophile (Scheme 2b), whereas sulfenylating agent 10 afforded greater than 95:5 er for all cyclization products (Tables 3 and 4).

Table 4.

Solvent Effect on Cyclization of Styryl-Alkene 19a^a,b

Entry	Substrate	Conditions	Product	e.r.
1		HFIP	trans-20a	97:3
2		CH2Cl2 MsOH (0.4 equiv)	trans-20a	97:3
3		HFIP	cis-20a	98:2
4		CH2Cl2 MsOH (0.4 equiv)	cis-20a	97:3

^aAll reactions performed on a 0.10 mmol scale.

^bEnantiomeric ratio determined by CSP-HPLC analysis.

The focus was then directed toward the nature of the alkene and differentiation between similar methylene substituents. Substrate (E)-3 was synthesized and cyclized to give a mixture of furan and pyran constitutional isomers but with consistently high enantioselectivity (Table 5). Curiously, a different product distribution was obtained with the more strongly ionizing MsOH component through a thermodynamic equilibration process to afford greater amounts of trans-21b than with HFIP. However, when (Z)-3 was cyclized, only weak enantioinduction was observed in either solvent albeit with slightly higher levels in HFIP. Taken together, these results indicate that (R)-2b is capable of discriminating both (Z)- and (E)-styrenes as well as (E)-dialkyl-alkenes but not (Z)-dialkyl-alkenes. The bulkier sulfenylating agent was instrumental in achieving high selectivities compared to previously employed N-phenylthiophthalimide 1. In view of the result shown in Scheme 2A for the cyclization of 3, it is clear that the combination of selenophosphoramide catalyst 2b and sulfenylating agent 10 confers a moderate increase in enantiotopic face discrimination. Similar reactivity trends for constitutional selectivity remain constant, however, namely, exo-cyclization of (Z)-alkenes is preferred in the absence of any activating groups.

Table 5.

Solvent Effect on the Cyclization of Dialkyl-Substituted Alkene 3^a,bcd

Entry	Substrate	Conditions	Selectivity (a:b)	Product	e.r.
1		HFIP	1.3:1	anti-21a + trans-21b	(6)c = 98:2 (5) = 98:2
2		CH2Cl2 MsOH (0.4 equiv)	1:2	anti-21a + trans-21b	(6)c = 98:2 (5) = 98:2
3		HFIP	>20:1	syn-21a + cis-21b	71:29d
4		CH2Cl2 MsOH (0.4 equiv)	>20:1	syn-21a + cis-21b	67:33d

^aAll reactions performed on a 0.10 mmol scale.

^bEnantiomeric ratio determined by CSP-HPLC analysis.

^cEnantiomeric ratio determined on corresponding sulfone.

^dEnantiomeric ratio determined for syn-21a.

Control studies were carried out next to demonstrate the difference in reactivity between (Z)- and (E)-alkenes for sulfenium ion capture (Scheme 4). A competition experiment between (Z)- and (E)-β-methylstyrene using one equivalent of sulfenylating agent 10 and 4-bromoaniline was performed. After only 30 min, full consumption of 4-bromoaniline was observed.¹H NMR analysis of the crude reaction mixture revealed a 72:28 ratio of the anti/syn products, demonstrating that (E)-β-methylstyrene reacted about 2.5 times faster. This result is consistent with the analysis of steric hindrance in the corresponding (Z)-alkene-derived thiiranium ion intermediate discussed above.

Scheme 4.

Competition Experiment between (E)- and (Z)-β-Methylstyrene

CONCLUSIONS

In conclusion, enantioselective sulfenofunctionalization of (Z)- and cyclic alkenes has been described. Despite the challenge of the reduced reactivity and steric congestion of these substrates, the BINAM-derived selenophosphoramide catalyst was able to engage a variety of substrates in inter- and intramolecular sulfenofunctionalization. Comparable levels of reactivity and selectivity were achieved as previously demonstrated for more established (E)-analogues. The development of a diverse scope as well as a better understanding of the difference between (E)- and (Z)-configured double bonds has been accomplished.