Asymmetric Catalytic Hydroboration of Enol Carbamates Enables an Enantioselective Matteson Homologation

Abstract

Rh-catalyzed asymmetric hydroboration of enol carbamates yields α-boryl carbamates in good enantioselectivity. The enol carbamate starting materials can be prepared with moderate Z selectivity using a modified Juila olefination and used as an E/Z mixture, taking advantage of the faster reactivity of the major Z isomer in the directed hydroboration. Optically active α-boryl carbamates participate in a Matteson-type homologation with Grignard reagents in which the O-carbamate is substituted with high conservation of optical activity to provide enantioenriched secondary boronic esters.

Graphical Abstract

The Matteson homologation remains a valuable synthetic transformation and the subject of continued development 60 years after its initial disclosure.1 The homologation involves a 1,2-metalate rearrangement of a boronate that features a leaving group alpha to boron (A, Scheme 1A). Noteworthy characteristics include 1) the ability to access the key intermediate A via addition of an organometallic to an α-functionalized boronic ester or via addition of a carbene equivalent (e.g. LiCH₂Cl) to a boronic ester, 2) the stereospecificity of the reaction, both in terms of the migrating group (retention) and the carbon undergoing substitution (inversion), and 3) the synthetic utility of the resulting boronic ester, which can be converted to a wide range of useful functional groups.^2,3

Scheme 1.

Approaches to stereoselective Matteson homologations.

Several strategies have been described to use the Matteson homologation to generate optically active boronic esters. Boronic acids derived from chiral non-racemic diols can lead to α-chloroboronic ester with high diastereoselectivity (Scheme 1B).² These products can participate in a second, stereospecific homologation reaction (not shown). Alternatively, optically active α-sulfinyl benzoates can undergo metalation with retention of stereochemistry and subsequent trapping with boronic esters (Scheme 1C).⁴ Substitution of the benzoate provides enantioenriched boronic esters. Similarly, Hoppe demonstrated that directed, asymmetric lithiation of terminal carbamates provides a route to α-boryl carbamates (B), which can participate in Matteson homologation with Grignard reagents (Scheme 1D).⁵ Finally, Lewis acids complexed with chiral non-racemic ligands have been shown to promote the enantioselective Matteson rearrangement with lithium dichloromethane. A Yb(III)(bisoxazoline) complex was shown to work in super-stoichiometric quantities,⁶ while an in-situ generated Li salt was found to catalyze a highly enantioselective reaction (Scheme 1E).^7,8

The Hoppe asymmetric lithiation/borylation strategy is attractive because it relies on readily available alcohol starting materials and gives rise to a reagent that can be used in asymmetric Matteson homologations.⁵ However, it requires stoichiometric quantities of sparteine, which has become challenging to access as the natural (−)-enantiomer⁹ while the unnatural (+)-enantiomer requires de novo synthesis.¹⁰ We reasoned that asymmetric hydroboration of enol carbamates could produce α-boryl carbamates with equal access to either enantiomer and the requirement for only catalytic quantities of the chiral controller (Scheme 1F). Here we describe the realization of this objective.

Asymmetric hydroboration has emerged as an effective strategy to synthetize optically active boronic esters.¹¹ Directing groups near the olefin can impart high regioselectivity, as shown by Takacs with oximes and phosphonates.¹² Takacs¹³ and Li¹⁴ also demonstrated that amides are also effective directing groups (Scheme 2A), while both Li¹⁵ and Tang¹⁶ have shown that enamides can direct asymmetric hydroboration (Scheme 2B). Since carbamates are effective directing groups for lithiation,¹⁷ we hypothesized that they might direct asymmetric hydroboration as well.^{18, 19}

Scheme 2.

Carbonyl-directed catalytic asymmetric hydroborations.

We started our investigation by examining the hydroboration of enol carbamate 3a, which was prepared with modest E/Z selectivity using a modified Julia-Kocienski olefination (Table 1).^{20, 21} A survey of ligands identified bisphosphines as a promising class to pursue, selected examples of which are shown in Table 1. Additional optimization studies shown in the supporting information. While Duanphos, BPE and Duphos gave poor reactivity and selectivity (Entries 1–3), ferrocene-containing ligands showed encouraging results. Ferrocelane gave modest yield and selectivity (Entry 4) while Josiphos-type ligands showed good reactivity but poor selectivity (Entry 5). JoSPOphos ligand L6b, featuring an alkyl phosphine and an aryl secondary phosphine oxide, combined high enantioselectivity with promising reactivity. By contrast, the isomeric ligand L6a did not support catalysis (Entries 6, 7). An evaluation of solvents revealed trifluorotoluene as optimal (see supporting information for more details) for reactivity. Whereas reactions in THF consumed only the Z isomer, the hydroboration in trifluorotoluene converted both olefin isomers. Since enantioselectivity dropped in this experiment, we speculated that E and Z olefin isomers could react with different rates and enantioselectivities (Entry 8). A control reaction with a ~1:1 E/Z mixture further depressed enantioselectivity and suggested that the E isomer might be converted with lower selectivity than the Z isomer (Entry 9 and see below). Finally, shorter reaction times or lower reaction temperatures selectively consumed the Z isomer and provided the hydroboration product in good yield and excellent enantioselectivity (Entries 10, 11).

Table 1.

Synthesis of Z-Enriched Enol Carbamate^a and Optimization of Hydroboration Conditions^b

Entry	L	Solvent	Temp (°C)	Time (h)	Yield (%)c	Ee (%)d
1	L1	THF	50	12	<5	7
2	L2	THF	50	12	13	34
3	L3	THF	50	12	21	n.d.
4	L4	THF	50	12	52	65
5	L5	THF	50	12	69	10
6	L6a	THF	50	12	0	n.d.
7	L6b	THF	50	12	50	91
8e	L6b	PhCF3	50	12	80f	78
9g	L6b	PhCF3	50	12	77	70
10e	L6b	PhCF3	35	18	65f	97
11e	L6b	PhCF3	rt	48	67f	97

^aReaction carried out on a 4.0 mmol scale in THF (0.1 M) using Julia reagent 1 (1.5 equiv), LiHMDS (1.7 equiv) and aldehyde 2a (1.0 equiv).

^bReactions were run on a 0.05 mmol scale using enol carbamate 3a (1.0 equiv), [Rh(cod)₂]OTf (5 mol%), ligand (6 mol%) and HBpin (2.0 equiv) unless noted otherwise.

^cYields determined by ¹H NMR.

^dEe values determined by HPLC (n.d. = not determined).

^e0.35 mmol scale.

^fIsolated yield.

^g3a with Z : E = 55 : 45 was used.

Two optimized conditions emerged from these studies. Conditions A were generally best for aryl-substituted olefins (35 °C, 18h, 5 mol% Rh), while Conditions B (rt, 48h, 10 mol% Rh) were usually preferred for alkyl-substituted olefins (Scheme 3).²²

Scheme 3.

Substrate scope for the asymmetric hydroboration. ^aAll reactions were run in PhCF₃ (0.10 M) using enol carbamate 3 (1.0 equiv), [Rh(cod)₂]OTf (5–10 mol%), ligand L6b (6–12 mol%) and HB(pin) (2.0 equiv). Starting material, reagents and solvent were mixed in a N₂-filled glove box. Yields of isolated products. Enantiomeric excess (ee) values were determined by HPLC. ^bWith [Rh(cod)₂]OTf (5 mol%) and ligand L6b (6 mol%). ^cWith [Rh(cod)₂]OTf (10 mol%) and ligand L6b (12 mol%). ^dDiastereomeric ratio (dr) was determined by ¹³C NMR. ^eReaction time = 14 h. ^fDiastereomeric ratio (dr) was determined by HPLC.

For aryl substituted olefins, the electronic properties of the aryl ring had minimal impact, with strongly electron donating substituents like alkoxy (4c, 4g, 4h) and amino (4i) groups performing similarly as electron withdrawing -CF3 (4e) and esters (4f). Substitution was tolerated in the meta and para positions, and an ortho-Br (4l) potentially gives rise to various ortho substitution patterns. Non-basic heteroaryl rings were accommodated by the reaction, including indole (4p, 4w), phenothiazine (4q), benzothiophene (4r) and furan (4s). Alkyl-substituted enol carbamates also reacted selectively, with both linear (4t) and cyclic (4u) substrates providing the hydroboration product in high ee. Single crystals derived from thioether 4d were suitable for X-ray diffraction analysis (CCDC 226136). The crystal structure showed coordination of the carbamate carbonyl to the boron and established the absolute stereochemistry of 4d. Other products were assigned by analogy.

Collectively, the examples in Scheme 3 demonstrate the functional group compatibility of the reaction. Tolerated groups include esters (4f, 4h, 4x), silyl ethers (4g, 4t), imides, amides and sulfonamides (4k, 4u, 4w), and halides (4e, 4l, 4w). Basic amines and strongly coordinating heterocycles (e.g., pyridines) are not tolerated under the current reaction conditions. With regard to other unsaturated C-C bonds, alkyne 3j, conjugated diene 3o, and the trisubstituted olefin derived from citronellal (3v) showed group-selective reduction of the enol carbamate. By contrast, a terminal olefin reacted at similar rates as the enol carbamate, giving diborated product 4m exclusively. Disubstituted olefins appear to react slightly slower than enol carbamates, as 4n was isolated with the unactivated olefin intact, albeit in reduced yield due to partial double hydroboration. Finally, hydroboration of the enol carbamates derived from citronellal demonstrated a matched/mismatched effect wherein the S-methyl isomer was formed with substantially higher ee than the R-methyl isomer.

During optimization studies with a 20:80 E:Z mixture of 3a, we noticed that ee was dependent on conversion. In particular, high ee was obtained when only the Z isomer was consumed, but it decreased as the E isomer reacted. Many of the enol carbamates could not be isolated as pure olefin isomers, but ester 3f could be resolved flash chromatography, allowing us to investigate the issue more carefully (Scheme 4). Purified E-3f reacted slowly under the reaction conditions, providing <20% of the hydroboration product. The main product in this reaction results from hydrogenation, which proved impossible to separate from the hydroboration product. By contrast, Z-3f reacted cleanly and provided the hydroboration product in 96% ee. For comparison, a 20:80 E:Z mixture generated the 4f in 88% ee. Estimates based on ee and conversion indicate that the E isomer generates the same major enantiomer, albeit in substantially reduced selectivity and rate. Thus, high ee can generally be obtained by monitoring the reaction for conversion of the Z isomer.

Scheme 4.

Impact of olefin stereochemistry on the asymmetric hydroboration. See Scheme 3 for reaction conditions.

Hoppe originally reported that α-boryl carbamates undergo stereospecific substitution with Grignard reagents.⁶ We confirmed this reactivity with products derived from asymmetric hydroboration, but also revealed some limitations (Scheme 5). Boronic ester 4a underwent substitution with a secondary (5a), aryl (5b) and alkenyl (5c) Grignard reagent in reasonable yields and high conservation of optical purity. However, reactions with linear or methyl Grignards were not successful, and PhLi and MeLi did not provide the desired product. Boronic ester 4c likewise reacted cleanly with PhMgCl, but alkyl and alkenyl Grignard reagents gave complex reaction mixtures. In general, the reactions were rapid, and performed best when the boronate was formed at −78 °C and then quickly placed into an ice bath. Prolonged stirring at −78 °C appeared to lead to racemization whereas starting the reaction at 4 °C resulted in messier reaction mixtures. Freshly prepared Grignard reagents performed most effectively, although the identity of the halogen (RMgBr vs RMgCl) did not appear to have a significant impact. Under the optimal conditions, several different boronate/Grignard combinations proved successful.²³

Scheme 5.

Reactions of hydroboration products with Grignard reagents. Reaction conditions: Et₂O (0.1 M), Grignard reagent (1.4 equiv.), −78 to 4 °C, 2 h. Yields are given for products isolated after oxidation. Enantiomeric excess (ee) values were determined by HPLC. Enantioselectivity (es) = ee of product / ee of starting material.

In summary, Julia olefination of aryl and alkyl aldehydes with carbamate 1 offers a general route to enol carbamates with predominantly Z stereochemistry. Asymmetric hydroboration using a Rh(I)(JoSPOphos) complex selectively converted the Z-olefin to α-boryl carbamates with high enantioselectivity. Finally, the resulting α-boryl carbamates participate in Matteson-type homologation reactions with high stereospecificity, although opportunities exist to further improve the generality of this transformation.