cis-Selective Direct Halocyclopropanation of Alkenes Mediated by Nucleophilic Cobalt Photocatalysis

Abstract

The cyclopropyl group is an invaluable motif in both medicinal and synthetic chemistry, with 1,2-disubstituted halocyclopropanes serving as important building blocks for the preparation of more complex cyclopropane-containing structures. Herein, we report a Vitamin B₁₂-photocatalyzed approach for the direct synthesis of 1,2-disubstituted halocyclopropanes, starting from alkenes and simple haloforms under visible-light irradiation. Our method yields a diverse range of 1,2-disubstituted halocyclopropanes with high cis-selectivity, providing efficient access to these traditionally difficult-to-prepare stereoisomers. Mechanistic studies highlight the crucial role of Vitamin B₁₂ in influencing the stereochemical outcome of the reaction.

Introduction

Cyclopropane rings are invaluable to both the medicinal and synthetic chemistry communities, owing to their ability to modulate key pharmacological properties − and their unique chemical properties that can be exploited in chemical synthesis. − Given their importance, reliable and efficient strategies for the synthesis and derivatization of these scaffolds are of great value. In this context, 1,2-disubstituted halocyclopropanes have been shown to be key building blocks for the construction of more complex cyclopropane-containing structures. − Furthermore, monochlorinated cyclopropanes can also be found in natural products with promising bioactivities, − further emphasizing the importance of these motifs in potential drug discovery and development. Among the synthetic strategies for the synthesis of halocyclopropanes, common approaches are the Simmons–Smith cyclopropanation reaction − and a two-step process involving the preparation of a dihalocyclopropane followed by a second monodehalogenation step (see Scheme A). The latter, and most widely used, approach involves the [2 + 1] cycloaddition of a dihalocarbene, generated from a haloform under basic conditions, with an alkene, which generally proceeds in good yields for a broad range of alkene substrates. − However, current methods for the selective monodehalogenation step, which often relies on organometallic reagents or strongly reducing metals, suffer from a myriad of drawbacks, including harsh reaction conditions, low yields, low functional group tolerance, and poor diastereoselectivity of the resulting 1,2-disubstituted halocyclopropane. ,− More recently, a range of one-step methods that address many of these shortcomings have been developed, which include leveraging techniques such as transition-metal catalysis, − electrochemistry, and photoredox catalysis. , Of note, Tambar and coworkers have reported the photomediated isomerization of cinnamyl chlorides to 1,2-substituted chlorocyclopropanes, which proceeds with remarkable selectivity for the trans-isomer (>95:5). More recently, Fasan and coworkers disclosed an enzyme-catalyzed carbene transfer protocol to generate bromo- and chlorocyclopropanes from β-halostyrenes in high diastereo- and enantioselectivity. In other recent developments, Nagib and coworkers reported an iron catalyzed direct halocyclopropanation of 1,1-diphenylethylene, which was proposed to proceed via iron carbene intermediates. Despite these elegant advances, reports of direct approaches for the synthesis of cis-halocyclopropanes are limited to iodocyclopropanes. For cis-bromo- and chlorocyclopropyl analogs, their preparation and isolation often relies on the separation of cis- and trans-isomers from unselective reactions. Given prior observations from both our group and from Chen and Zhang that Vitamin B₁₂-catalyzed cyclopropanation reactions showed slight preference for the formation of the cis-isomer, we hypothesized that we could leverage this catalytic platform and our prior expertise to develop a one-step, cis-selective halocyclopropanation of alkenes directly from haloforms (CHX₃).

1. (A) Preparation of cis-1,2-Disubstituted Halocyclopropanes; (B) Proposed Mechanistic Design for the Direct Halocyclopropanation of Alkenes Mediated by Nucleophilic Cobalt Photocatalysis.

Our proposed mechanistic design for our cis-selective direct halocyclopropanation of alkenes mediated by nucleophilic cobalt photocatalysis is outlined in Scheme B. Based on prior art using Vitamin B₁₂ and other macrocyclic tetradentate cobalt complexes, − including our recent contribution for the activation of CH₂Cl₂ as a radical precursor and methylene source in cyclopropanation reactions, we anticipate that generation of the highly nucleophilic Co­(I) oxidation state of Vitamin B₁₂ would enable an S_N2-type nucleophilic substitution with CHX₃, generating Co­(III)–CHX₂ intermediate I. Intermediate I could then undergo photolysis upon exposure to green (525 nm) LED irradiation to form a ^•CHX₂ radical and a persistent Co­(II) radical. Addition of the electrophilic ^•CHX₂ radical to an activated alkene, such as styrene, followed by trapping of the carbon radical intermediate by Co­(II) would yield intermediate II, which we tentatively propose gets reduced to liberate the carbanion that would ultimately yield the halocyclopropane after a polar 3-exo-tet cyclization.

Results and Discussion

We began our investigation of the direct halocyclopropanation of alkenes using styrene (1) as the model substrate (Table ). After optimization of each of the key reaction parameters, we found that chlorocyclopropanation of 1 proceeded efficiently using 4 equiv CHCl₃, 5 mol % Vitamin B₁₂, 5 equiv of Zn as the stoichiometric reductant and 1 equiv NH₄Cl in DMA under 525 nm LED irradiation for 8 h, generating cyclopropyl adduct 2 in 64% NMR yield (entry 1). Importantly, high selectivity for the cis-2 adduct (cis:trans 86:14) was observed under these optimized reaction conditions. The reaction was also compatible with CHBr₃, generating 3 in 58% yield, albeit with a slightly decreased cis:trans ratio of 78:22 (entry 2). Switching to CHI₃ did provide the corresponding cycloadduct 4 with good cis-selectivity (84:16); however, a low yield was observed, with the remaining mass balance being recovered starting material (entry 3). We hypothesize the low yield stems from side reactivity between Zn and CH₃I. DMF was also effective as a solvent, albeit with a decreased yield (entry 4). The use of Mn as a stoichiometric reductant led to a low yield of 2 and a sharp decrease in cis:trans selectivity, likely owing to off-cycle reactivity between Mn and CHCl₃ (entry 5). Switching to a 456 nm LED also led to a decrease in reaction efficiency (entry 6), likely owing to the weaker absorption of the Co­(III)–CHX₂ ligand-to-metal charge-transfer (LMCT) band under 500 nm (see SI). Control reactions indicated degassed conditions were optimal and Zn was crucial for product formation (entries 7–8). Removing Vitamin B₁₂ resulted in only 8% yield for CHCl₃ (entry 9) and no observable product for CHBr₃ (entry 10). Furthermore, the observed product formed in the absence of Vitamin B₁₂ was found to have minimal cis:trans selectivity (55:45). These results highlight that Vitamin B₁₂ plays a critical role in generating the corresponding halocyclopropanes with high cis-selectivity, with minimal competing Simmons–Smith reactivity being observed under our standard conditions. These findings are consistent with a recent report from Nagib and coworkers, who observed that Zn was unable to efficiently activate CHCl₃ in the absence of LiI. While removal of NH₄Cl only resulted in a slight reduction in yield (entry 11), we have observed greater reaction reproducibility using this acid additive, removing any variation that could be observed related to inconsistent Zn activation. Consistent with our prior findings, 30% yield of 2 could be obtained in the absence of LED irradiation, likely resulting from inefficient thermolysis of the weak Co­(III)–C bond at room temperature (entry 12). Increasing the LED intensity, while resulting in the highest cis-selectivity, was found to negatively impact the yield (entry 13). Finally, adding TEMPO to the reaction resulted in a decreased yield of 2, highlighting the potential involvement of radical intermediates (entry 14).

1. Optimization and Control Reactions for the Halocyclopropanation of Styrene .

Entry	X	Deviations from standard conditions	Product: yield (%)	cis:trans
1	Cl	none	2: 64	86:14
2	Br	none	3: 58	78:22
3	I	none	4: 14	84:16
4	Cl	DMF instead of DMA	2: 50	85:15
5	Cl	Mn instead of Zn	2: 24	56:44
6	Cl	456 nm LED	2: 45	86:14
7	Cl	not degassed	2: 50	87:13
8	Cl	no Zn	2: N.D.	-
9	Cl	no Vitamin B12	2: 8	55:45
10	Br	no Vitamin B12	3: N.D.	-
11	Cl	no NH4Cl	2: 53	87:13
12	Cl	no hν, rt	2: 30	79:21
13	Cl	100% LED Intensity	2: 47	91:9
14	Cl	2 equiv TEMPO	2: 33	85:15

^aStandard conditions: 1 (0.25 mmol, 1 equiv), CHX₃ (4 equiv), Vitamin B₁₂ (5 mol %), Zn (5 equiv), NH₄X (1 equiv) in DMA (2 mL) were irradiated under Ar with a Kessil PR-160L 525 nm LED (75% intensity) at 30 °C for 8 h. Yields and dr ratios were determined by ¹H NMR analysis of the crude reaction mixture using 1,3,5-trimethoxybenzene as an external standard. N.D.: not detected.

With the optimized reaction conditions in hand, we then examined the scope of the halocyclopropanation reaction (Scheme ). Using either CHCl₃ or CHBr₃, a series of cis-selective halocyclopropyl adducts could be generated in moderate to good yields and good cis:trans selectivity (>80:20) from a range of styrenes (2–15). In most of these cases, the minor trans-isomer could be removed during purification, leading to highly pure cis-halocyclopropanes. The reaction proceeded equally efficiently regardless of whether the styrene contained electron donating (6–10) or electron withdrawing (11–14) functional groups. The reaction could tolerate aryl halides (14) and aryl boronic esters (15), which can serve as functional handles for further cross-couplings and derivatizations. Styrenyl analogues of benzoic acid and aniline were not compatible with our method. While 1,1-disubstituted alkenes were well tolerated (16–19), albeit with decreased cis-selectivity, 1,2-disubstituted styrenes were found to be unreactive. Indene (20) also underwent chlorocyclopropanation in 46% yield, and trans-1-phenyl-1,3-butadiene (21) was also compatible, with chlorocyclopropanation taking place exclusively at the terminal alkene of the diene. Styrenes derived from eugenol (22), naproxen (23), L-tyrosine (24), and probenecid (25) all provided the desired cis-selective halocyclopropyl adducts in moderate to good yields, highlighting the potential of our method for the late-stage functionalization of drugs and natural products. Vinyl heterocycles derived from pyridine (26), indole (27), benzofuran (28) and benzothiophene (29) all reacted smoothly under our optimized conditions to give the corresponding chlorocyclopropyl adducts in moderate to good yields. Electron-deficient alkenes, while still compatible owing to the mild nucleophilicity of the ^•CHX₂ radical, gave lower yields of the desired cis-halocyclopropane (see SI). On the other hand, unactivated alkenes did not yield any of the desired halocyclopropyl adducts. Reactions could also be directly scaled up to 1.0 mmol scale with no loss in reactivity (12, 19). Given the recent interest in deuterium incorporation in small molecule pharmaceutics, − we also investigated the direct halodeuterocyclopropanation of alkenes using readily available CDCl₃. In all cases, good to excellent yields and remarkable diastereoselectivity of the desired chlorodeuterocyclopropyl adducts were observed (30–34). Finally, a series of silyl enol ethers were also examined as suitable alkene substrates for our halocyclopropanation protocol. In all cases, good to excellent yields of the desired trisubstituted cyclopropanes were observed (35–40). Furthermore, subsequent TMS-deprotection following the photomediated halocyclopropanation reaction led to successful isolation of the trisubstituted cyclopropanols 41 and 42, which have been shown to be useful building blocks in organic synthesis. −

2. Substrate Scope for the Direct cis-Selective Halocyclopropanation of Alkenes .

^a Yields of isolated, purified products after 8 h of irradiation using the optimized conditions (see General Procedures C1 and C2 in the Supporting Information). Cis:trans ratios were determined by ¹H NMR analysis of the crude reaction mixture. ^bRepresents cis:trans ratio of the isolated, purified products. ^cReaction performed at 1.0 mmol scale (see General Procedure C3 in the Supporting Information). ^dDiastereomeric ratios for silyl enol ether examples were determined by ¹H NMR analysis of the crude reaction mixture. ^eDeprotection conditions: cat. TMSCl, MeOH, 0°C, 10 min. Yields reported as isolated yields over two steps.

Next, we sought to determine the synthetic potential of this cis-selective halocyclopropanation of alkenes (Scheme ). First, the reaction was optimized for a more synthetically relevant scale. Using 5 mmol of styrene (1) with CHBr₃, 66% yield of 3 was obtained in high diastereoselectivity (cis:trans 83:17, Scheme A). Notably, the loading of both Vitamin B₁₂ and Zn could be lowered without sacrificing reaction efficiency, and the amount of CHBr₃ could be decreased to only 2 equiv by adding it dropwise over the course of the reaction. Next, we explored a series of synthetic manipulations of the cis-bromocyclopropane products. Using in situ generated lithium morpholide, bromocycloadduct 12 underwent efficient nucleophilic amination, giving the trans-2-arylcyclopropylamine 43, a class of compounds known for their rich biological activity, in 62% yield and 92:8 dr (Scheme B). In the presence of excess potassium tert-butoxide, bromocyclopropane 19 could be converted into the corresponding cyclopropene (44), an important class of building blocks in organic synthesis (Scheme C). − cis-Bromocyclopropane 3 readily underwent magnesium-bromide exchange with retention of configuration, which upon trapping with DMF yielded aldehyde 45 (Scheme D). Subsequently, the aldehyde was further elaborated to the alkene using a Wittig reaction to yield a common radical clock (46) in 42% yield over two steps. − Finally, 3 could also be employed as a radical precursor under nucleophilic Vitamin B₁₂-photocatalyzed conditions, and upon reaction with methyl acrylate was converted to Giese adduct 47 in 49% yield with good selectivity for the trans-isomer (dr 94:6).

3. Synthetic Elaborations of Halocyclopropanes .

^a (A) 5 mmol scale halocyclopropantion. (B) Nucleophilic amination. (C) Cyclopropene synthesis. (D) Synthesis of radical clock. (E) Giese reaction. Conditions: (i) 1.1 equiv morpholine, 1.2 equiv n-BuLi, THF, – 78 °C to −20 °C. (ii) 11, THF, −20 °C. (iii) 4 equiv tert-BuOK, DMSO, rt, 24 h. (iv) 1.1 equiv iPrMgCl·LiCl, 1,4-dioxane, rt, 8 h. (v) DMF, rt, 4 h. (vi) 1.1 equiv Ph₃PMeBr, 1.1 equiv tert-BuOK, THF, rt, 16 h. (vii) 5 mol % Vitamin B₁₂, 3 equiv Zn, 1 equiv NH₄Br, DMF, Ar, hν (525 nm), 30 °C, 16 h.

Having explored the reaction scope and synthetic utility of halocyclopropanes, we investigated the key steps of the reaction mechanism. Using UV–vis techniques, we were able to confirm the feasibility of both the generation of the catalytically active Co­(I) oxidation state and the subsequent S_N2 reaction with both CHCl₃ and CHBr₃ (see SI). Next, we sought to probe the involvement of ^•CHX₂ radical intermediates. Upon subjecting radical clock 48 to our standard reaction conditions with CHCl₃, the anticipated ring-opened/cyclized product 49 was observed in 7% yield and confirmed by HRMS analysis, supporting the intermediacy of ^•CHCl₂ radicals (Scheme A). Interestingly, the non-ring-opened chlorocyclopropyl adduct 50 was observed in 58% yield, which was also confirmed by HRMS analysis. It is also important to note that the corresponding dichlorocyclopropane adduct was not observed in the reaction, and a control study for the dehalogenation of 51 under standard conditions yielded only 12% yield of 2 with a low cis:trans selectivity (58:42) that was inconsistent our model reaction (Scheme B). Considering these data, dihalocyclopropanes are unlikely to be key reaction intermediates, ruling out the formation of carbenes under our standard conditions. To explain the formation of 50, an analysis of the relevant rate constants is required, which are summarized in Scheme C. Upon ^•CHCl₂ radical generation, addition to radical clock 48 results in the formation of stabilized benzyl radical intermediate III. Given that ring-opening leads to the formation of unstabilized primary radical IV, the unimolecular rate constant for this reaction is estimated to be quite slow, on the order of 10⁵ s^–1 (k ₁). Conversely, the Co­(II) oxidation state of Vitamin B₁₂ and other related model cobalt complexes, which can be described as a persistent free radical, is known to react with other carbon-centered radicals to form Co­(III)–R species at bimolecular rate constants in the range of 10⁷ to 10⁹ M^–1 s^–1 (k ₂). , Therefore, trapping of benzyl radical intermediate III by Co­(II) is proposed to lead to the formation of Co­(III)-bound intermediate V, which upon single-electron reduction leads to dicyclopropyladduct 50 through a polar 3-exo-tet. In good agreement, the involvement of intermediates of type II or V have long been invoked in Vitamin B₁₂-photocatalyzed Giese reactions. , Taken together, these data highlight both the radical-polar crossover nature of the halocyclopropanation, and the intimate involvement of Vitamin B₁₂ leading up to the key 3-exo-tet step. We hypothesize that the formation and reduction of intermediate II plays a crucial role in setting the observed cis-selectivity of our Vitamin B₁₂-photocatalyzed cyclopropanation reaction (Scheme D), as 3-exo-tet cyclizations occurring in solution (in the absence of Vitamin B₁₂) are expected to be unselective based on previously reported radical-polar crossover approaches. − Our group is actively working on developing a stereochemical model to rationalize the selectivity observed in Vitamin B₁₂-catalyzed cyclopropanation reactions, and these results will be reported in due course.

4. Mechanistic Studies and Proposed Halocyclopropanation Mechanism .

^a (A) Radical clock experiment. (B) Monodehalogenation control. (C) Proposed mechanistic pathway. (D) Proposed origin of stereoselectivity.

Conclusions

In summary, we have developed a Vitamin B₁₂-photocatalyzed approach for the direct halocyclopropanation of electron-rich alkenes using haloforms. The reaction proceeds with high cis-selectivity, providing efficient access to these traditionally difficult to prepare stereoisomers. Our protocol was compatible with a broad range of styrenes, vinyl heterocycles, and silyl enol ethers, and the method could also be applied for the chlorodeuterocyclopropanation of alkenes using readily available CDCl₃. The reaction could be performed on 5 mmol scale, and the bromocyclopropyl adducts could be further derivatized into a range of high value products. Mechanistic studies suggested that Vitamin B₁₂ plays a key role in the radical-polar crossover step, ultimately influencing the stereoselectivity. We anticipate our Vitamin B₁₂-photocatalysis platform will find immediate applications in the synthesis of 1,2-disubstituted halocyclopropanes, as it provides a direct approach for the preparation of these valuable motifs that proceeds with complementary selectivity compared to known methods.

The data underlying this study are available in the published article and its Supporting Information.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/jacsau.6c00381.

• Experimental procedures, reaction optimization, compound characterization, and NMR spectra (PDF)

The authors declare no competing financial interest.