Nickel-Catalyzed Alkyl-Alkyl Cross-Electrophile Coupling Reaction of 1,3-Dimesylates for the Synthesis of Alkylcyclopropanes

Abstract

Cross-electrophile coupling reactions of two Csp³–X bonds remain challenging. Herein we report an intramolecular nickel-catalyzed cross-electrophile coupling reaction of 1,3-diol derivatives. Notably, this transformation is utilized to synthesize a range of mono- and 1,2-disubstituted alkylcyclopropanes, including those derived from terpenes, steroids, and aldol products. Additionally, enantioenriched cyclopropanes are synthesized from the products of proline-catalyzed and Evans aldol reactions. A procedure for direct transformation of 1,3-diols to cyclopropanes is also described. Calculations and experimental data are consistent with a nickel-catalyzed mechanism that begins with stereoablative oxidative addition at the secondary center.

Graphical Abstract

Cross-electrophile coupling (XEC) reactions have the potential to construct carbon-carbon bonds in an efficient manner. To favor cross-reactivity, reactions often pair two substrates of different reactivity, in part to differentiate oxidative addition events.¹ For example, development of aryl-alkyl XEC reactions have been fruitful.^1d,e,2 Reactions that combine two substrates with similar reactivity can be challenging and result in homocoupled products.³ As such, many known examples of nickel-catalyzed XEC reactions that forge Csp³–Csp³ bonds employ one activated substrate as a coupling partner, e.g., allylic or benzylic electrophiles.^1d,e,4,5,6,7 There are few examples of nickel-catalyzed XEC reactions that engage two unactivated alkyl electrophiles.^8,9 Development of an XEC reaction that employs readily-available, unactivated alkyl electrophiles as both reactive partners would significantly expand the scope of these transformations.

A 1,3-diol is a compelling functional group motif for use in XEC reactions. Largely due to their prevalence in polyketides, 1,3-diols have robust, well-established synthetic routes for their preparation.¹⁰ For example, substituted 1,3-diols are easily accessed through the reduction of aldol products. We hypothesized that sulfonates derived from 1,3-diols would be engaged by a nickel catalyst and undergo an intramolecular XEC reaction to form cyclopropanes. Furthermore, since aldol reactions can provide outstanding levels of enantioselectivity,¹¹ this strategy could provide straightforward and predictable access to enantioenriched cyclopropanes, leveraging a well-established and powerful field. Additionally, these reactions would complement traditional asymmetric cyclopropanation routes that typically engage alkene starting materials.¹²

Alkyl sulfonates have a history of use in nickel-catalyzed cross-coupling (XC) and XEC reactions.^13,14,15 Notably, sulfonates utilized in these reactions are frequently generated in situ from the corresponding alcohols. In the context of cross-coupling reactions, nickel-catalyzed Negishi couplings of benzylic mesylates and Kumada reactions of cyclic sulfates have been reported (Scheme 1a).^16,17 The use of alkyl sulfonates as electrophiles in nickel-catalyzed XEC reactions developed contemporaneously, beginning with homocoupling reactions.¹⁸ Chemoselective XEC reactions have utilized alkyl sulfonates with aryl and vinyl halides and pseudohalides (Scheme 1b).^19,20,21 Cross-selective pairing of two alkyl sulfonates was demonstrated using methyl tosylate (Scheme 1c).^7a,22 However, to the best of our knowledge, no cross-selective XEC reaction of two primary or secondary alkyl sulfonates has been reported. Based on the accessibility of 1,3-diols and reactivity of sulfonates, we sought to develop a nickel-catalyzed XEC reaction of two alkyl mesylates for cyclopropane synthesis (Scheme 1c).¹²

Scheme 1.

Nickel-Catalyzed XC and XEC Reactions of Alkyl Sulfates and Sulfonates

We began our investigation by employing 1,3-dimesylate 1. This substrate was designed to provide an alkylcyclopropane with low volatility to facilitate isolation and analysis. We evaluated a series of ligands in the presence of Ni(cod)₂ and methylmagnesium iodide (MeMgI) in DCM/PhMe (Table 1). The diphosphine ligands rac-BINAP and dppm, in addition to the pyridyl ligand Bphen, produced the highest yields of cyclopropane 2 (entries 1–3). Additionally, a 70% yield of the desired cyclopropane was achieved utilizing bench-stable ((R)-BINAP)NiCl₂ as the catalyst (entry 9). In general, across a range of substrates, rac-BINAP and dppm provided robust reaction yields, and so we selected these ligands for further experiments.²³

Table 1.

Ligand Evaluation of XEC Reaction of Unactivated 1,3-Dimesylates

Entry	Deviation from reaction conditions	Yielda (%)
1	none	75
2	BPhen	78
3	dppm	71
4	PCy3	56
5	DPEPhos	44
6	SiMes*HBF4	33
7	Xantphos	16
8	no ligand	13
9	((R)-BINAP)NiCI2	70
10	PhMgBr	5
11	PhMgBr + Mgl2	17
12	no Ni, no ligand	5
13	no MeMgl	0

^aYield determined by ¹H NMR based on comparison to PhTMS as internal standard.

Following the ligand evaluation, we next investigated the importance of the Grignard reagent and nickel catalyst. Modifying the Grignard reagent to phenylmagnesium bromide almost completely shut down the reaction, with 5% of the desired cyclopropane observed (entry 10). Adding MgI₂ to PhMgBr reaction conditions provided a similar result (entry 11). A control reaction without nickel and ligand (MeMgI only) produced a 5% yield of the desired cyclopropane (entry 12), while a control reaction in the absence of MeMgI provided no conversion to the desired cyclopropane and only recovered starting material (entry 13).

With optimized conditions in hand, we investigated the tolerance of various substituted aromatic and heterocyclic groups (Scheme 2a). Isolated yields are reported, however for certain substrates volatility or polarity complicated isolation. Therefore, yield determined by ¹H NMR by comparison to an internal standard is also reported. Heterocycles such as thiophene, furan, and benzofuran were well tolerated under the reaction conditions (3–5), as was the substituted heterocycle 2-methoxypyridine (6). Electron-donating groups were well tolerated in the synthesis of 2 and 7, as well as electron-withdrawing groups such as aryl CF₃ and aryl fluoride (8,9).

Scheme 2.

Isolated yields of reaction products. ^aYield determined by ¹H NMR based on comparison to PhTMS as internal standard. ^b3.0 equiv of MeMgI

Our next efforts focused on testing the impact of steric bulk near the forming cyclopropane (Scheme 2b). For compounds such as 11, standard reaction conditions employing rac-BINAP as the ligand provided modest yields.²³ Fortunately, we found that for these more hindered substrates, using dppm as the ligand and performing the reaction at 0 °C provided good yields. A series of alkyl and aryl groups were well tolerated in the β-position relative to the cyclopropane (10–12). For example, the diol precursor for cyclopropane 12 was prepared by a stereospecific Kumada ring-opening reaction.²⁴ We were also interested in the functional group compatibility of the reaction. Compounds containing benzylic ethers (PMB), aryl chlorides, alkyl ethers, and silyl protecting groups (TBS) underwent smooth cyclopropane formation (13–15).

We next sought out to confirm our hypothesis that aromatic groups were not required for desired reactivity. To confirm this, we successfully synthesized cyclopropanes containing only sp³ centers such as borneol derivative 16 and β-sitosterol derivative 18. Acetonide 17 was formed smoothly from the corresponding tetraol derivative, demonstrating tolerance to a typical protecting group employed in polyketide synthesis. These results confirm that – in contrast to our laboratory’s previously published XC and XEC reactions – this XEC reaction does not require benzylic or allylic electrophiles to engage the nickel catalyst.^5a,5b,25

Based on these results, we concluded that arene coordination would no longer be critical for reactivity, but the reaction could nonetheless provide a strategy for synthesis of arylcyclopropanes. In our prior work, synthesis was restricted to cyclopropanes bearing extended aromatic rings, due to requirements to activate substrates for oxidative addition.^5a,5b,25 To thoroughly investigate the generality of this XEC reaction, we evaluated simple benzylic substrates. Notably, under standard conditions for preparation of the 1,3-dimesylates, benzylic mesylates undergo substitution to afford the benzylic chlorides.²⁶ Subjection of 1,3-chloromesylates to our standard reaction conditions provided cyclopropanes bearing a range of substituted aromatic substituents (Scheme 2c; 19–22).

The potential impact of this transformation would be expanded if 1,3-diols could be employed as starting materials for the reaction. We were encouraged that other XC and XEC reactions that employ sulfonates generated in situ have been reported.^16,20a A procedure was developed such that diol 23 was treated with MsCl and base, followed by addition to catalyst and Grignard reagent. Cyclopropane 11 was formed in good yield, similar to that observed when employing the corresponding 1,3-dimesylate. Therefore, this method allows direct conversion of a 1,3-diol to the corresponding cyclopropane (eq 1).

(1)

In order to elucidate a potential reaction mechanism, our first question was whether the mechanism involves oxidative addition (OA) of an alkyl mesylate, or if alkyl iodides are formed as intermediates in the reaction.²⁷ A control experiment demonstrated that under the reaction conditions, in the absence of nickel catalyst, MeMgI converts 1,3-dimesylate 24 to 1,3-diiodide 25 (Table 2, entry 1). Monitoring the XEC reaction at early conversion produced 1,3-diiodide 25 in 56% yield and cyclopropane 7 in 18% yield (entry 2). Therefore, we hypothesized that 1,3-diiodide 25 is formed in situ and is a reactive intermediate in the nickel-catalyzed XEC reaction. Indeed, subjecting 25 to XEC reaction conditions provided clean conversion to cyclopropane 7 (eq 2).

(2)

Table 2.

Formation of 1,3-diiodide 25 in situ

^aYield determined by ¹H NMR based on comparison to PhTMS as internal standard. See SI for details.

Our second question was whether oxidative addition initiates at the primary or secondary center. To begin to address this question, we calculated the relative barrier heights for oxidative addition of 1° and 2° alkyl iodides, via both radical and S_N2 mechanisms (Figure 1).²⁸ For both the secondary and primary center, the barrier heights for reaction via the radical pathway are significantly lower. The calculated lowest energy transition state (TS3) engages the secondary center by halogen atom abstraction to generate a 2° alkyl radical intermediate (INT7).^29,30,31 This barrier height is lower for reaction at the secondary center, because the more substituted center is better able to stabilize the incipient alkyl radical.^29,32,33,34 Therefore, we propose that the XEC reaction begins with iodine atom abstraction at the 2° alkyl iodide moiety, and proceeds through a 2° alkyl radical (INT7). This alkyl radical recombines with the nickel catalyst to form organonickel species INT8.

Figure 1.

Comparison of DFT-computed barriers for Ni(0)-catalyzed C–I oxidative addition, employing dppm as the ligand. Free energies in toluene are in kcal/mol and were calculated under the M06/def2-TZVPP-SMD(toluene)//B3LYP-D3(BJ)/def2-SVP-SMD(toluene) level of theory.

To challenge these computational findings, we designed a series of competition experiments that probe the relative rates of the nickel-catalyzed XEC reaction at 1° and 2° centers.³⁴ In a competition experiment, equimolar amounts of 1,3-dimesylate 24 and 1° mesylate 26 were subjected to XEC reaction conditions.³⁵ Cyclopropane 7 was favored with a 7:1 ratio, relative to total products generated from reaction of 1° mesylate 26 (Scheme 3a). Similarly, in a competition experiment between 1,3-dimesylate 24 and 2° mesylate 27, cyclopropane 7 was favored, although to a lesser extent, with a 3:1 ratio (Scheme 3b). Therefore, in both cases, the reaction of the 1,3-dimesylate is favored over reaction of the monomesylate. Furthermore, comparison of the product distributions demonstrates that the secondary center engages faster than the primary center under the XEC conditions. These results are consistent with the calculated transition state barriers. They are also consistent with prior studies comparing the rates of reactions of substituted alkyl halides with nickel catalysts.^29,34 Nickel-catalyzed halogen atom abstraction provides the basis for the observed preference for reaction of the 2° over 1° alkyl iodide.

Scheme 3.

Competition experiments.

To further corroborate our proposed mechanism, we sought to determine the stereochemical outcome of the XEC reaction. Since oxidative addition of the 2° alkyl iodide is predicted to proceed via the alkyl radical, reactions are expected to be stereoablative.⁴ Consistent with this hypothesis, we observe a stereoconvergent XEC reaction to form 1,2-disubstituted cyclopropane 29 (Scheme 4). Either diastereomer of 1,3-dimesylate 28 provides trans-cyclopropane 29, consistent with epimerization via an alkyl radical intermediate.

Scheme 4.

Stereoconvergent XEC reactions

Given the yield and diastereoselectivity observed in Scheme 4, we set out to apply our transformation to a range of 1,2-disubstituted alkylcyclopropanes (Scheme 5). The majority of corresponding diols were prepared by aldol or Claisen transformations.³⁶ Compounds 30–36 were formed in moderate to good yield with preference for the trans diastereomer. Functional groups including dibenzofuran (31), methoxy phenyls (32), aryl ethers (33 and 35), and acetals (34 and 36) were tolerated in this alkyl series. Paralleling our strategy in Scheme 2, we also demonstrated arylcyclopropanes can be successfully synthesized from the corresponding benzylic chlorides (37–38).

Scheme 5.

1,2-Disubstituted cyclopropanes

^aYield determined by ¹H NMR based on comparison to PhTMS as internal standard. ^bReaction stirred at 0 °C

Based on our understanding of the reaction mechanism and observed levels of diastereoselectivity, we set out to synthesize enantioenriched 1,2-disubstituted cyclopropanes (Scheme 6). Proline-catalyzed and Evans aldol reactions were employed to prepare the corresponding substituted 1,3-diols with high enantioselectivity.^37,38 Utilizing our transformation, alkylcyclopropane 40 and arylcyclopropane 42 were both formed with high enantiomeric excess. Configuration at C2 is conserved through the XEC reaction. The reaction favors formation of the trans-cyclopropane, setting the configuration of C1. Notably, compounds 40 and 42 do not bear the signature directing groups or acyl substitution required for direct synthesis by other asymmetric cyclopropanation methods.¹²

Scheme 6.

Enantioenriched 1,2-disubstituted cyclopropanes

^aYield determined by ¹H NMR based on comparison to PhTMS as internal standard

In summary, we report a nickel-catalyzed cross-electrophile coupling reaction of 1,3-dimesylates for the synthesis of alkylcyclopropanes. This transformation does not require activation of either electrophilic partner, and engages two alkyl mesylates. This method is tolerant of various functional groups and also allows access to arylcyclopropanes without extended aromatics. Furthermore, direct transformation of a 1,3-diol to the corresponding cyclopropane was established. Initial experimental and computational investigations are consistent with formation of 1,3-alkyl iodides that react with the nickel catalyst by iodine atom abstraction to provide 2° alkyl radicals. Synthesis of 1,2-disubstituted cyclopropanes is a stereoconvergent process consistent with the proposed mechanism of oxidative addition, and favors the trans diastereomer. The products of enantioselective aldol reactions were transformed to the corresponding enantioenriched cyclopropanes, therefore capitalizing on outstanding strategies of 1,3-diol synthesis.