Enantioselective S-Alkylation of Sulfenamides by Phase-Transfer Catalysis

Abstract

A general phase-transfer catalyst (PTC) mediated enantioselective alkylation of N-acylsulfenamides is reported. Essential to achieving high selectivity was the use of the triethylacetyl sulfenamide protecting group along with aqueous KOH as the base under biphasic aqueous conditions to enable the reaction to be performed at −40°C. With these key parameters, enantiomeric ratios up to 97.5:2.5 at the newly generated chiral sulfur center were achieved with an inexpensive cinchona alkaloid derived PTC. Broad scope and excellent functional group compatibility was observed for a variety of S-(hetero)aryl and branched and unbranched S-alkyl sulfenamides. Moreover, to achieve high selectivity for the opposite enantiomer, a pseudoenantiomeric catalyst was designed and synthesized from inexpensive cinchonidine. Given that sulfoximines are a bioactive pharmacophore of ever-increasing interest, selected product sulfilimines were oxidized to the corresponding sulfoximines with subsequent reductive cleavage affording the free-NH sulfoximines in high yields. The utility of the disclosed method was further demonstrated by the efficient asymmetric synthesis of atuveciclib, a phase I clinical candidate for which only chiral HPLC separation had previously been reported for isolation of the desired (R)-sulfoximine stereoisomer.

High oxidation state sulfur pharmacophores incorporating nitrogen, and particularly sulfoximines, are an increasingly employed pharmacophore in drug discovery and development with multiple entering clinical trials (Scheme 1A).^[1] When the two carbon substituents of a sulfoximine are different, the sulfur atom is a configurationally stable asymmetric center, which has motivated the development of innovative asymmetric syntheses of sulfoximines,^[2] including recent reports utilizing sulfinate esters and sulfinamides.^[3] Recently, we reported the first example of enantioselective catalytic alkylation of N-acylsulfenamides at sulfur to provide sulfilimines (Scheme 1B),^[4,5] which are readily oxidized in high yield to sulfoximines.

Scheme 1.

Sulfoximine drug candidates and catalytic enantioselective syntheses of sulfilimines.

Phase-transfer catalyst (PTC) mediated carbanion alkylation is a powerful tool for asymmetric synthesis that has seen extensive use in industrial settings due to the applicability of inexpensive alkylating agents and organocatalysts as well as robust and reproducible reaction conditions.^[6] In our previous work on the racemic alkylation of N-acylsulfenamides with alkyl halides, we provided a single example of a chiral phase-transfer benzylation that afforded the product in a modest 83:17 e.r. (Scheme 1C).^[7,8] To the best of our knowledge, this represents one of few examples of phase-transfer alkylation on S(II) centers,^[9] and the only one involving sulfenamides. While this example served as a proof-of-concept for asymmetric phase-transfer sulfenamide alkylation, considerable improvements were clearly needed.

Herein, we report general conditions for an enantioselective phase-transfer alkylation of N-acylsulfenamides (Scheme 1D). This method utilizes readily accessible alkyl iodides as the electrophile to form a S(IV) center in a PTC mediated reaction. The use of a cheap and commercially available cinchona alkaloid catalyst further enhances the practicality of the approach. The discovery and application of a new, bulky N-triethylacetyl protecting group resulted in greatly improved enantioselectivities. Moreover, aqueous KOH as the base with its comparatively low freezing point relative to other alkali bases allowed for a reaction temperature of −40°C to further enhance enantioselectivity. The utility of the alkylation products was demonstrated by high yielding oxidation and subsequent reductive removal of the N-triethylacetyl protecting group. We further established the applicability of our approach to drug synthesis by an efficient asymmetric synthesis of the phase I clinical candidate atuveciclib^[10] in five steps and 64% overall yield from an enantioenriched sulfilimine obtained by phase-transfer asymmetric alkylation.

We began this study by evaluating various parameters including solvents, catalysts, bases, temperatures, and concentrations (see Tables S1-6 in Section III of SI). Of all the catalysts screened, the relatively inexpensive PTC-1 was found to provide the greatest selectivity with a 30% aqueous KOH and toluene biphasic solvent mixture at a −40°C reaction temperature. With this data in hand, we next evaluated several protecting groups on sulfenamide 1a when alkylated with benzyl iodide 2a (Tablet). While the N-pivaloyl protecting group in P1 provided a decent e.r. of 88:12, increasing the steric bulk of the N-protecting group as seen in P2, P3, and P4 increased the selectivity of the reaction up to a 95:5 e.r., while maintaining high yields up to 92% (entries 1–4). The use of bulky P5 incorporating an adamantyl group provided a selectivity of only 91:9 (entry 5). We hypothesize that P4 provides a higher selectivity because two of the ethyl groups point toward the nucleophilic sulfur center to avoid unfavorable steric interactions with the third ethyl substituent. Use of the planar N-benzoyl protecting group in P6 provided a nearly racemic product (entry 6), although incorporating ortho methyl groups on the N-benzoyl group in P7 forces the benzamide out of plane to afford a greater selectivity of 82:18 (entry 7). Continuing to add to the bulk of the benzamide as seen in the N-triisopropylbenzoyl derivative P8 was detrimental to the selectivity of the reaction, providing the product in only a 70:30 e.r. (entry 8).

Using P4 as the standard protecting group, we next evaluated several parameters to establish the standard reaction conditions. Performing the reaction at 0°C instead of −40°C still provided the product in high yield; however, the selectivity dropped from 95:5 to 91:9 (entry 9). Decreasing the catalyst loading from 10 to 5 mol % maintained the 95:5 e.r.; however, the yield dropped to 85% (entry 10). Changing the solvent mixture to include CH₂C1₂ in a 4:1 ratio reduced the selectivity slightly to 94:6 (entry 11). Decreasing the concentration of the reaction to 0.025 M resulted in a reduced yield (entry 12), while increasing the concentration to 0.1 M decreased the selectivity to 93:7 (entry 13). The use of benzyl bromide or tosylate in place of benzyl iodide had little effect on the yield but the selectivity dropped to 89:11 and 93:7, respectively (entries 14–15). Changing the counterion of the catalyst from chloride to iodide had almost no effect on the reaction yield or selectivity (entry 16). When the reaction was quenched after only one hour, 7% of 3aa had formed with 95:5 e.r., identical to the selectivity observed at full conversion confirming that no deterioration of selectivity occurs as the reaction progresses (entry 17). Additionally, after reaction completion, 90% of the phase-transfer catalyst was recovered and hydroxyl or quinoline alkylation could not be detected (see Section IX, SI).

Having determined the optimal reaction conditions, the scope with respect to both alkyl iodides and sulfenamides was evaluated (Scheme 2). As listed in Table 1, benzyl iodide provided the product 3aa in a near quantitative yield and 95:5 e.r., a result that was reproducible at 1.0 mmol scale. The use of electron-rich benzyl iodides provided sulfilimines 3ab and 3ac in excellent yields and enantiose-lectivities, while an electron-deficient benzyl iodide gave 3ad in high yield and 92:8 e.r. Benzyl iodides with halides on the aryl ring were effective electrophiles for this reaction, with the p-chloro benzyl iodide providing 3ae with a quantitative yield and 95:5 e.r. Meta substituted benzyl iodides, including m-bromo and m-nitro, gave the products 3af and 3ag in a 95:5 e.r. and in high yields. Ortho substituents did not negatively affect the transformation as the o-bromo benzyl iodide gave product 3ah in an 87% yield with a 96:4 e.r. The fully substituted pentafluoro benzyl iodide provided sulfilimine 3ai in high yield and a 96.5:3.5 e.r., and the bulkier 1-naphthyl iodide afforded 3aj in a quantitative yield with an impressive 97.5:2.5 e.r., a result that was reproduced at 1.0 mmol scale.

Scheme 2.

Substrate Scope. Reactions performed on 0.1 mmol scale of 1 with 2.0 equiv. of 2. Isolated yields reported. Enantiomeric ratios determined by chiral HPLC analysis. [a] 48 h, 3.0 equiv. of 2. [b] 48 h, 10 equiv. of 2 n.

Table 1:

Reaction optimization.^[a]

Entry	P	Variation from standard conditions	Yield 3 aa[b]	e.r. 3 aa[c]
1	P1	none	80%	88:12
2	P2	none	87%	90:10
3	P3	none	79%	93:7
4	P4	none	92%	95:5[d]
5	P5	none	65%	91:9
6	P6	none	14%	51:49
7	P7	none	73%	82:18
8	P8	none	99%	70:30
9	P4	0 °C	95%[e]	91:9
10	P4	PTC-1 (5 mol %)	85%[e]	95:5
11	P4	toluene/CH2Cl2 (4:1)	92%[e]	94:6
12	P4	0.025 M	69%[e]	96:4
13	P4	0.1 M	92%[e]	93:7
14	P4	Br instead of I in 2 a	95%[e]	89:11
15	P4	OTs instead of I in 2 a	99%[e]	93:7
16	P4	I− instead of Cl− in PTC-1	94%[e]	95:5
17	P4	1 h	7%	95:5

^[a]Reactions performed on 0.1 mmol scale of 1 a with 2 equiv. of 2 a.

^[b]Isolated yields reported.

^[c]Enantiomeric ratios determined by chiral HPLC analysis.

^[d]Absolute configuration determined by X-ray crystallographic analysis of resulting sulfoximine (see Scheme 3).

^[e]Yields determined by ¹H NMR using trimethyl(phenyl)silane as a standard.

Substituted pyridines could also be incorporated. An alkylating reagent relevant to the clinical candidate VIP152 (Scheme 1A),^[11] 4-(iodomethyl)-2-chloropyridine, gave 3ak in 89% yield and 92:8 e.r. Switching the locations of the nitrogen and chloro substituents gave 3al in a 99% yield and 93:7 e.r. Alkylation with 1-iodo-2-butyne in place of a benzyl iodide gave 3am in a 93% yield and 94:6 e.r. Finally, the use of methyl iodide provided 3an though with a lower 87:13 e.r. and a large excess of this alkylating agent was necessary to achieve the 75% yield.

Naphthyl iodide 2j was then used to evaluate sulfenamide scope. Electron-rich and electron-deficient sulfenamides are suitable reactants as the p-methoxy and p-trifluoromethyl sulfenamides delivered products 3bj and 3cj in near quantitative yields and with selectivities of 97.5:2.5 and 96:4 e.r., respectively. Sulfenamides containing halides at the para, meta, and ortho positions were efficient coupling partners, providing the corresponding products in a 96.5:3.5, 97.5:2.5, and 97:3 e.r., although the ortho substituted 3fj sustained a slight reduction in yield to 81% when compared to the 99% yields for 3dj and 3ej. Despite the aqueous basic conditions, more reactive functional groups were well tolerated, with an aldehyde successfully incorporated in sulfilimine 3gj and an ester in 3hj, both in 99% yields and with selectivities of 97:3 and 96:4 e.r., respectively. A bulkier naphthyl sulfenamide provided sulfilimine 3ij in a 99% yield with a 94:6 e.r. Heteroaryl groups could also be utilized as demonstrated with the synthesis of 2-thiophenyl sulfilimine 3jj in an 83% yield and 96.5:3.5 e.r. and 3-thiophenyl sulfilimine 3kj in a 94% yield and 97.5:2.5 e.r. Sulfenamides with S-alkenyl substituents were also effective substrates with sulfilimine 311 obtained in an 80% yield and 97:3 e.r.

Alkyl sulfenamides were also useful inputs, although alkylation proceeded more slowly and required more forcing conditions with 3 equivalents of electrophile 2 and a 48 hour reaction time. We suspect this decrease in the rate was due to the lower acidity of S-alkyl sulfenamides and thus less efficient deprotonation. A cyclohexyl group was incorporated into sulfilimine 3mj in a 73% yield with a 93:7 e.r., and an S-benzyl sulfenamide provided sulfilimine 3nj in a lower 55% yield but maintaining a high 96.5:3.5 e.r. Given the close similarity of the two benzyl groups in 3nj, this sulfilimine would clearly be extremely challenging to prepare by an asymmetric thioether imination approach. Significantly, an S-methyl sulfenamide was an effective coupling partner, providing 3oj in a 97% yield and 96:4 e.r. Reacting an S-methyl sulfenamide with the 3-nitrobenzyl iodide provided sulfilimine 3og in an 84% yield and 90:10 e.r. at 1.0 mmol scale. This material was subsequently utilized in a synthesis of the clinical candidate atuveciclib (see below).

The synthesis of the opposite enantiomer of the product is not necessarily straightforward because the true enantiomer of the catalyst cannot be accessed from the chiral pool. While the standard commercially available cinchonine-derived catalyst PTC-1 provided 3aj in a 97.5:2.5 e.r. favoring the R-enantiomer (Table 2, entry 1), the commercially available cinchonidine-derived catalyst PTC-2 favors the S-enantiomer, although with a drop to 93:7 e.r. (entry 2). It is a common trend for pseudoenantiomeric catalysts, where only two of the five chiral centers have been inverted, to afford lower selectivity. Deng and co-workers have made significant steps towards understanding this uneven efficiency in pseudoenantiomeric catalysts and have discovered that inverting the chiral center of the vinyl group on the catalyst tends to largely restore the selectivity of the pseudoenantiomer.^[12] Inspired by Deng’s work, we prepared the epimer of the pseudoenantiomeric catalyst (PTC-3) by a six-step sequence starting from inexpensive cinchonidine (see Section X, SI). Using this catalyst, 3aj can be synthesized in a 96.5:3.5 e.r. favoring the S-enantiomer (entry 3), a significant improvement over PTC-2. Satisfyingly, all three catalysts gave the product in near quantitative yields. Understanding the sense of induction provided by these phase-transfer catalysts is challenging due to the biphasic reaction conditions and ionic substrate catalyst interactions and will be the subject of a future report.^[13]

Table 2:

Asymmetric synthesis of opposite enantiomer.^[a]

Entry	Catalyst	Yield 3aj[b]	e.r. 3aj (S:R)[c]
1	PTC-1	99%	2.5:97.5
2	PTC-2	99%	93:7
3	PTC-3	99%	96.5:3.5

^[a]Reactions performed on 0.1 mmol scale of 1 a with 2.0 equiv. of 2 j.

^[b]Isolated yields reported.

^[c]Enantiomeric ratios determined by chiral HPLC analysis.

The enantioenriched sulfilimines 3 obtained by PTC-mediated sulfur alkylation are useful intermediates to enantioenriched sulfoximines, an important component in many bioactive molecules. Oxidation of sulfilimine 3aa furnished sulfoximine 4 in a stereospecific manner with a 99% yield (Scheme 2A). The configuration of this product was rigorously determined using X-ray crystallographic analysis to be the R-enantiomer. Next, deprotection of sulfoximine 4 was conducted to obtain the free-NH sulfoximine 5 that can be further utilized to explore vector space or contribute a hydrogen bond donor. Adapting conditions developed by Chiba, Takita and co-workers,^[14] reductive cleavage with ZnH₂ provided 5 in a quantitative 98% yield without racemization. Use of LiAlH₄, a common reducing reagent, also afforded 5 albeit in a 73% yield.

Atuveciclib, which has an (R)-configuration at the sulfur center, is a PTEFb/CDK9 inhibitor that advanced to Phase I clinical trials for the treatment of cancer.^[10] Chiral HPLC separation was used in the drug discovery and development studies,^[10,15] and in the only published asymmetric synthesis, enantioselective oxidation of the corresponding thioether proceeded in only 60:40 e.r. when the archetypal asymmetric sulfoxidation catalyst developed by Kagan was used.^[16] We therefore developed an asymmetric synthesis of atuveciclib as a means to demonstrate the applicability of our approach to drug structures (Scheme 2B). First, sulfilimine 3og was oxidized using ruthenium tetroxide to deliver the sulfoximine that, without purification, was treated with Fe/HCl to reduce the nitro group, thereby providing aniline 6 in an excellent 98% yield over the two steps. Reductive cleavage of the triethylacetyl protecting group using ZnH₂ provided the free sulfoximine 7 in a 73% yield. Next, an S_NAr reaction between 7 and dichlorotriazine was selective for the aniline over the sulfoximine nitrogen to afford 8 in a 98% yield. Finally, a Suzuki coupling^[17] with the 4-fluoro-2-methoxyphenyl boronic acid with Pd(dppf)Cl₂ as the catalyst provided the desired R-enantiomer of atuveciclib in a 91% yield while maintaining a 90:10 e.r. (Scheme 3).

Scheme 3.

Synthetic applications of sulfilimines.

In summary, we report an enantioselective S-alkylation of sulfenamides employing cinchona alkaloid derived catalysts to afford sulfilimines in up to 99% yields and 97.5:2.5 e.r. The reaction proceeds with broad scope and high functional group compatibility. An inexpensive cinchona alkaloid derived PTC was used to access one enantiomer of the sulfilimines, while a readily prepared catalyst derived from cheap and commercially available cinchonidine was developed to provide the opposite enantiomer with minimal loss in enantioselectivity. The sulfilimine products from this transformation were oxidized and cleaved to the corresponding free-NH sulfoximines in a straightforward two step reaction sequence. Finally, the method was applied to an efficient asymmetric synthesis of the phase I clinical candidate atuveciclib. Given the increasing importance of sulfoximines and other high oxidation state pharmacophores in drug discovery, we foresee that the reported enantioselective phase-transfer catalyzed approach will be of significant value for the discovery and development of bioactive compounds.

Data Availability Statement

The data that support the findings of this study are available in the supplementary material of this article.