Enantioselective Catalyst Systems from Copper(II) Triflate and BINOL–Silanediol

Abstract

Silanediol and copper catalysis are merged, for the first time, to create an enhanced Lewis acid catalyst system for enantioselective heterocycle functionalization. The promise of this silanediol and copper catalyst combination is demonstrated in the enantioselective addition of indoles to alkylidene malonates to give rise to the desirable adducts in excellent yield and high enantiomeric excess. From these studies, 1,1′-bi-2-naphthol (BINOL)- based silanediols emerge as one-of-a-kind cocatalysts. Their potential role in the reaction pathway is also discussed.

The interesting properties of the silanediol functional group have enticed investigators to pursue its study in a number of applications, such as materials science, medicine, molecular recognition, and catalysis.^[1–4] In the context of enantioselective catalysis, the hydrogen-bonding and anion-binding abilities of silanediols have significant value.^{[5, 6]} For instance, enantioselective silanediol catalysis has emerged as a promising platform for the functionalization of nitrogen- and oxygen-based heterocycles. These reactions plausibly benefit from a silanediol anion-binding mode of action; the silanediol may directly activate an ionic substrate through hydrogen-bonding interactions with the anionic component.^{[6, 7]} Notably, silanediols can offer reactivity patterns that are inaccessible with other types of catalysts. For instance, (R)-3,3′-diphenyl-2,2′-bi-1-naphthalol (VANOL)- and 1,1′-bi-2-naphthol (BINOL)-based silanediols are, to date, distinct in their ability to control the stereochemical outcome of reactions of 4-silyloxybenzopyrylium triflates.^[6a]

The unique abilities of silanediols in enantioselective catalysis enticed us to stop limiting their application to the direct activation of substrates (Figure 1A). We reasoned that silanediols could work together with transition-metal catalysts, such as Cu(OTf)₂ (OTf = triflate), thereby generating an enhanced catalyst system (Figure 1B) for the discovery of useful reactions.^[8–10] Herein, we report, for the first time, a silanediol and copper catalyst system for the enantiocontrolled addition of indoles to alkylidene malonates.

Figure 1.

Expanding the role of silanediols in catalysis: A) substrate activation and B) catalyst activation.

It is well known that metals catalyze the addition of indoles to arylidene and alkylidene malonates.^[11] In the case of arylidene malonates, the reactions can be rendered highly enantioselective in the presence of an appropriate chiral ligand, such as tris-oxazolines.^[11e] Attaining high levels of enantiocontrol in the addition of indole to alkylidene malonates, on the other hand, remains a significant challenge. With this in mind, we set out to explore silanediol and transition-metal catalyst systems as a solution for the enantioselective addition of indole (3a) to cyclohexylidene malonate (2a; Table 1).

Table 1.

Identification of reaction conditions for silanediol and copper cocatalysis in the reaction of 2a and 3a.

Entry[a]	1 [mol%]	Co-catalyst [mol%]	Yield[b]	ee[c]
1	30	Cu(OTf)2 (20)	92	72
2	30	Sc(OTf)3 (20)	95	9
3	30	In(OTf)3 (20)	88	10
4	20	Zn(OTf)2 (20)	0	–
5	30	CuOTf (20)	10	31
6	0	Cu(OTf)2 (20)	15	0
7	30	Cu(OTf)2 (0)	0	–
8	30	CuSO4 (20)	0	–
9	30	CuCl (20)	0	–
10	30	CuI (20)	0	–
11	30	HOTf (10)	0	–
12	20	Cu(OTf)2 (20)	93	73
13	10	Cu(OTf)2 (10)	70	70
14[d]	20	Cu(OTf)2 (20)	99	75

^[a]See the Supporting Information for details on the experimental procedure.

^[b]Yield determined by ¹H NMR spectroscopy.

^[c]The enantiomeric excess (ee) determined by HPLC.

^[d]1,1,1-Trifluoroisopropanol (TFIP; 0.5 equiv) was added.

Our studies began with an examination of the influence of BINOL-based silanediol 1 (30 mol%) and a wide selection of transition metals (20 mol%), a few of which are listed in Table 1, in the addition of 3a to 2a. Only the combination of 1 and Cu(OTf)₂ gave rise to a high yield of 4a with high levels of enantiocontrol (92% yield, 72% ee; Table 1, entry 1). Other transition metals tested could generate high yields of desired product, but low levels of enantiocontrol (e.g., Sc(OTf)₃, In(OTf)₃; Table 1, entries 2 and 3). CuOTf also proved inferior to its Cu(OTf)₂ counterpart (Table 1, entry 5). Control experiments revealed the important roles of both the silanediol and Cu(OTf)₂ in achieving excellent yields and stereocontrol: Cu(OTf)₂ (20 mol%) alone, with no silanediol added, resulted in a dramatic drop in the yield of 4a (15 %; Table 1, entry 6). No product was observed if the silanediol (30 mol%) was tested as the sole catalyst of the reaction (Table 1, entry 7). Copper sources other than Cu(OTf)₂, such as CuSO₄, CuCl, and CuI, were also evaluated, but they did not yield product; this suggests that the OTf counterion is essential to the success of the reaction (Table 1, entries 8–10). Triflic acid was ruled out as the cocatalyst because control experiments resulted in no reaction (Table 1, entry 11). Optimization of the copper to silanediol ratio resulted in an improvement in the enantiomeric ratio (e.r.): silanediol (20 mol%) and Cu(OTf)₂ (20 mol%) was identified as the best combination in terms of yield and stereocontrol (93% yield, 73% ee; Table 1, entry 12). The catalyst loadings could be dropped to 10 mol% with no significant drop in ee, but the yield was slightly lower (70% yield, 70% ee; Table 1, entry 13). Finally, a screening of additives enabled us to pinpoint that TFIP (0.5 equiv) gave a small boost in ee (75% ee; Table 1, entry 14).

Having identified a workable copper and silanediol catalyst system for the introduction of 3a to 2a, we further probed the influence of the structures of 2 and 3 on the outcome of the reaction (Table 2).^[12] The substituent on the ester had a rather dramatic effect on enantiocontrol; the trend was that smaller substituents led to higher e.r. values.^[12] Methyl esters (2) resulted in the best stereocontrol of the reaction. Alkylidene malonates with R¹ = Cy afforded 4a in excellent yield with high levels of enantiocontrol (98 %, 75% ee; Table 2, entry 1). Various substitution patterns about the 3 scaffold enabled the preparation of 4b–4g. Alkyl and aryl substituents in the 5-position of 3 were well tolerated to produce 4b–4 f in excellent yields and high ee (Table 2, entries 2–6). Substituents at the 7-position of 3 were also well accepted. For example, 7- methylindole afforded 4 g in 96% yield and 75% ee (Table 2, entry 7).

Table 2.

Effects of substrate structures on the reaction.

Entry[a]	R1	R2	Product	Yield [%][b]	ee [%][c]
1	Cy[d]	H	4a	98	75
2	Cy[d]	5-Me	4b	96	77
3	Cy[d]	5-tBu	4c	92	68
4	Cy[d]	5-Ph	4d	92	71
5	Cy[d]	5-(2-naphthyl)	4e	66	72
6	Cy[d]	5-(3,5)-CF3Ph	4 f	58	86
7	Cy[d]	7-Me	4g	96	75
8	Ph	H	4h	99[e]	49
9	p-MeO−Ph	H	4i	16[e],[f]	52
10	p-NO2−Ph	H	4j	99[e]	30
11	o-tolyl	H	4k	50[e]	49
12	2-naphthyl	H	4l	99[e],[f]	58
13	n-propyl	H	4m	92[e]	30

^[a]See the Supporting Information for details on the experimental procedure.

^[b]Yields of products isolated after flash column chromatography on silica gel.

^[c]Determined by HPLC.

^[d]Cy = cyclohexyl.

^[e]Yield determined by ¹H NMR spectroscopy; 30 mol% 1.

^[f]Addition of 40 mol% 1.

For a comparison to alkylidene malonate 2a, several arylidene malonates were tested as substrates in the reaction. Interestingly, the arylidene malonates did not afford ee values as high as 2a. For instance, arylidene malonates containing either electron-donating or -withdrawing substituents yielded 4h–4l in low to excellent yield (16–99 %), but with modest levels of ee (30–58% ee; Table 2, entries 8–12). The importance of the alkylidene malonate structure on enantioselectivity was further highlighted with malonate 2 (R¹ = n-propyl), which gave rise to 4m in 92% yield and 30% ee (Table 2, entry 13).

The reaction is amenable to scale up (Scheme 1). If 2a was subjected to reaction with 3a under the influence of silanediol and copper cocatalysis, compound 4a was produced in 85% yield and 74% ee. The process is synthetically useful because the enantioenriched product is easily crystallized from isopropanol and hexanes to give rise to 4a in an excellent enantioselectivity of 97% ee.

Scheme 1.

Scale-up and recrystallization for the preparation of 4a.

Interesting observations were made while probing the possible role of the silanediol. First, it was found that the silanediol was a uniquely effective cocatalyst: all of our attempts to catalyze the reaction with a combination of Cu(OTf)₂ and other popular dual hydrogen-bond donors (HBDs) were entirely fruitless (Scheme 2). For instance, the combination of Cu(OTf)₂ and popular chiral thiourea catalyst 5 yielded no reaction. The addition of enantiopure BINOL (7) and VANOL (8) resulted in low yields of racemic product. Evidence of the importance of the silanediol functional group was found because bis-dimethoxy silanediol 6 gave rise to racemic 4a in 25% yield.

Scheme 2.

Unique combination of silanediol and Cu(OTf)₂.

Given the unique ability of silanediol 1 to cooperate with catalytic copper(II) triflate to affect the enantioselective addition of 3 and 2, we became curious to learn more about the role of the silanediol in the reaction. Experimental evidence collected during the optimization of the reaction (Table 1) suggested that neither the silanediol nor Cu(OTf)₂ alone were efficient catalysts: there is something unique about the combination of the two. The observation had been made (Table 1) that the triflate counterion was required for a successful reaction. Furthermore, it has been previously demonstrated that silanediols can recognize triflate anions;^[6a] thus there is the possibility that the silanediol is activating Cu(OTf)₂ through anion binding to the triflate to create copper-silanediol catalyst complex I (see Scheme 4, below). Alternatively, a reaction pathway in which the silanediol may be operating as a ligand on copper was also considered.

Scheme 4.

Possible catalytic pathway.

Efforts to probe the role of the silanediol and plausible catalytic species began with UV/Vis studies of the Cu(OTf)₂ in the presence of silanediol 1 and a competing ligand: achiral bis-oxazoline 9 (Scheme 3). In the first set of experiments, it was noted that the combination of Cu(OTf)₂ and 1 did not produce a band in the spectrum indicative of a d–d transition, whereas the combination of Cu(OTf)₂ and 9 produced a band at λ_max ≈754 nm (Scheme 3a). In a second experiment, the titration of 1 (1–5 equiv) into a solution of Cu(OTf)₂ and 9 did not significantly alter the UV/Vis spectrum (Scheme 3b). These data suggest that the silanediol may not be operating as a ligand, and is not likely to displace a bis-oxazoline ligand on copper.

Scheme 3.

Probing of the plausible role of silanediol.

To better understand the silanediol mode of action, an experiment was set up to in which achiral bis-oxazoline 9 (20 mol%) was added to an otherwise standard reaction (Scheme 3c). Based upon the UV/Vis data, it was reasoned that the bis-oxazoline would be a superior ligand on copper to that of the silanediol. If 4a was prepared in enantiomerically enriched form from this reaction, it might suggest that the silanediol is operating as an anion-binding catalyst and not as a ligand. Indeed, compound 4a was isolated in 90% yield with 27% ee; this possibly provides evidence for the reaction pathway depicted in Scheme 4. More specifically, the catalytic cyclic may begin with the silanediol interacting with Cu(OTf)₂ to generate enhanced Lewis acid species I.^[13] The complexation of I with alkylidene malonate 2 may give rise to II. The subsequent enantioselective addition of 3a to II yields intermediate III, which, after proton transfers, affords desired product 4 and frees complex I for continued participation in the catalytic cycle.

In conclusion, silanediols work with Cu(OTf)₂ to generate an enhanced catalyst system with applications in enantioselective synthesis. Herein, silanediol and copper catalysis has been demonstrated in the reaction of indoles and alkylidene malonates to afford desirable adducts with high levels of enantiocontrol. Our preliminary studies suggest that the silanediol may be operating as an anion-binding catalyst to activate copper triflate, but further investigations are needed to better understand the operation of the catalyst system. Ongoing studies in our laboratory are dedicated toward uncovering the full potential of silanediol and transition-metal cocatalysis in the context of asymmetric synthesis.

Experimental Section

Dimethyl cyclohexylidene malonate (113 mg, 0.5 mmol, 1.0 equiv), Cu(OTf)₂ (36 mg, 0.1 mmol, 0.2 equiv), trifluoroisopropanol (22.6 μL, 0.25 mmol, 0.5 equiv), and toluene (5 mL) were added to a 20 mL screw-top reaction vial fitted with a Teflon-coated septum. The flask was purged with dry N₂ and the reaction mixture stirred for 15 min or until a homogenous slurry was obtained. The reaction vial was then cooled to −78 °C in a dry ice/acetone bath. A 0.05 M stock solution of silanediol (2.4 mL, 82 mg, 0.24 mmol, 0.2 equiv) in toluene and a solution of 3a (2.6 mL, 88 mg, 0.75 mmol, 1.5 equiv) in toluene were added to the reaction vial dropwise. The reaction vial was transferred to a laboratory freezer (−28°C) and stirred overnight. The reaction was quenched with deionized water (2 mL), stirred for 10 min, extracted with EtOAc (3V10 mL), and dried over Na₂SO₄. Solvent was removed from the combined organic layer under vacuum to obtain the crude product. The crude product was purified by column chromatography on silica gel (eluent = 4:1, hexanes/EtOAc). The resulting material was purified further by column chromatography on silica gel (eluent = 100% dichloromethane). After removal of the solvent under vacuum, dimethyl 2-[cyclohexyl(1H-indol-3-yl)methyl]malonate was obtained as an off-white solid (159 mg, 0.46 mmol, 93 %). R_f = 0.25 (4:1, hexanes/EtOAc); ${[\alpha ]}_{\mathrm{D}}^{23}=-9.1$ (c = 4.0, CH₂Cl₂); ¹H NMR (500 MHz, CDCl₃): δ = 8.03 (s, 1H), 7.66 (ddt, J = 8.0, 1.5, 0.8 Hz, 1H), 7.32 (dt, J = 8.0, 0.9 Hz, 1H), 7.16 (ddd, J = 8.2, 7.0, 1.2 Hz, 1H), 7.10 (ddd, J= 8.0, 7.0, 1.1 Hz, 1H), 7.03 (d, J = 2.4 Hz, 1H), 4.03 (d, J = 10.8 Hz, 1H), 3.78 (dd, J = 10.8, 4.9 Hz, 1H), 3.73 (s, 3H), 3.35 (s, 3H), 1.78–1.55 (m, 6H), 1.31–1.08 (m, 2H), 1.02–0.81 ppm (m, 3H); ¹³C NMR (126 MHz, CDCl₃): δ = 169.6, 169.0, 135.8, 128.4, 122.9, 121.9, 119.7, 119.4, 113.9, 111.0, 55.7, 52.6, 52.3, 42.0, 41.2, 32.3, 28.8, 26.7, 26.5, 26.3 ppm; IR (neat): ν̃ = 3413, 2926, 2853, 1755, 1726, 1457, 1431 cm⁻¹; HPLC: 86.0:14.0 e.r., 72% ee, Chiralpak AS-H column (10% iPrOH/hexanes, 1 mLmin⁻¹, 225 nm); t_R (minor) = 8.55 min, t_R (major) = 23.80 min.