Efficient Synthesis of α-Haloboronic Esters via Cu-Catalyzed Atom Transfer Radical Addition

Abstract

The synthesis of α-haloboronic esters via atom transfer radical addition (ATRA) is constrained due to its limited range of compatible substrates or the need to manipulate the olefin coupling partners. Herein, we present a novel approach for their synthesis via Cu-catalyzed ATRA to vinyl boronic esters. The catalyst is proposed to mediate a traditionally inefficient halogen atom transfer of the α-boryl radical intermediate, thus significantly expanding the range of participating substrates relative to established methods. The forty-eight examples illustrate that a wide range of radical precursors, including primary, secondary, and tertiary alkyl halides, readily add across both unsubstituted and 1-substituted vinyl pinacol boronic esters. Further, a one-pot, two-step protocol is presented for direct access to an array of α-functionalized products. Finally, the synthetic utility of this methodology is demonstrated in the synthesis of an ixazomib analogue.

Interest in α-haloboronic esters has increased recently as valuable building blocks in organic synthesis.¹ Consequently, achieving their synthesis using efficient, versatile, and straightforward methods stands as an important objective in reaction development.^1d,2 The Matteson homologation is the most common method for their synthesis; however, this approach necessitates the use of strong bases (e.g., n-BuLi) and cryogenic conditions, and is restricted in the product of α-chloroboronic esters (Scheme 1A).^1d,3 Atom transfer radical addition (ATRA) to vinyl boronates or vinyl boronic esters offers an alternative approach for their synthesis. Pre-activated coupling partners, e.g. vinyl N-methyliminodiacetic acid boronates^2a and potassium vinyl trifluoroborates,^2e generate a reactive radical 1 which undergoes efficient propagation with electron-deficient alkyl halides (Scheme 1B).⁴ Nonetheless, the existing tetracoordinate boron in 2 limits the direct functionalization of the α-halogen.⁵ Recently, the superior reactivity of lithium aryl/alkyl vinyl boronates in the ATRA reaction has been employed in the synthesis of α-aryl/alkyl boronic esters: upon formation of the ATRA intermediate, rapid 1,2-migration occurs to generate the product.⁶ The general direct ATRA to vinyl boronic esters remains a significant challenge; the α-boryl radical 3, formed upon radical addition step, exhibits relative stability leading to inefficient propagation with all but highly electron deficient radical precursors (e.g. di-tert-butyl bromomalonate; Scheme 1B).⁷ To overcome this limitation, we propose employing a judiciously designed halogen atom transfer (XAT) mediator to facilitate the halogen atom transfer step, irrespective of the electronic nature of the alkyl halides.

Scheme 1.

Strategies for the synthesis of α-haloboronic ester.

To accomplish the general ATRA reaction to vinyl boronic esters, we envisioned using a transition metal catalyst [Mⁿ] that can act as (i.) a radical initiator to reduce alkyl halides (R–X), generating the R• and [Mⁿ⁺¹]–X; (ii.) an XAT mediator that favorably undergoes halide abstraction by 3, formed upon the addition of R• to the vinyl boronic ester, to afford the desired product 4. A potential challenge for ATRA to vinyl boronic esters arises from the adverse radical addition of electron-deficient radicals to vinyl boronic ester due to the polarity mismatch.⁴ However, a favorable halide abstraction by the α-boryl radical from the XAT mediator may overcome this challenge.

We hypothesized that Cu would be a suitable catalyst to facilitate this desired transformation. Cu(I) complexes are well known to activate a wide range of alkyl halides, including electron-rich and electron-deficient variants.⁸ Further, Cu(II) halide complexes are excellent XAT mediators (Scheme 1C); extensive studies by Matyjaszewski and Kochi have shown them to undergo halogenation at nearly diffusion-controlled rates.^8,9 Notably, α-haloboronic esters have not been identified as a radical precursors that can be activated by Cu(I) complexes, suggesting that Cu(II) complexes may irreversibly halogenate 7.

To test this hypothesis, we investigated a variety of Cu catalysts’ abilities to promote the ATRA reaction between electron-rich radical precursor, ethyl α-bromoisobutyrate 9, and vinylboronic acid pinacol ester (vinyl Bpin) 10 (Table 1). Our initial efforts focused on Cu(I) salts (4.0 mol %) with highly soluble dtbbpy ligands (8.0 mol %) in DCE. We observed a significant formation of 11’ (X = Cl) instead of the anticipated ATRA product 11 (X = Br) (entries 1 and 2). However, when employing Cu(OTf)₂, in the presence of N-methylaniline and K₃PO₄ as reductants,¹⁰ we observed 66% yield of 11 along with trace 11’ (entry 3). We then explored several ligands and found those resulted in low reactivity and the formation of the double insertion products with electron-rich and electron-deficient ligands, entries 4 and 5.¹¹ 2,2’-bipyridine (bpy) ligand shows similar reactivity to dtbbpy (entry 6). Utilizing the precomplexed [Cu(dtbbpy)₂](OTf)₂ catalyst gives a moderate increase to 72% yield, entry 7. With the optimized copper complex, we explored different solvents and found that using solvents lacking C_sp3-Cl bonds, such as PhCl, prevents the generation of 11’ and results in the formation of 11 in excellent yield (89% in-situ; 87% isolated). Additionally, the reaction is scalable; 1.27 g of 11 (73% yield) is isolated on a 5.0 mmol scale (entry 9). The reactions are also feasible to perform on the benchtop and under ambient air with a moderate decrease in the in-situ yield (entries 10 and 11). Control experiments demonstrate that copper is essential for promoting the reaction; in its absence, no product is observed (entry 12). Likewise, radical initiators, known to promote the ATRA with alkyl halides, yield less than 5% of the desired product (entries 13 and 14). Purification of the α-haloboronic esters via silica chromatography is difficult. In response to this challenge, we either use B-doped silica¹² or replace the Bpin with the tetraethyl (3,4-diethylhexane-3,4-diol) analogue (BEpin) that has comparable reactivity and can be readily purified by standard silica chromatography.¹³

Table 1.

Optimization of ATRA^a

entry	catalyst	ligand	solvent	yield (%)
1	CuBr	dtbbpy	DCE	47b
2	CuCl	dtbbpy	DCE	51b
3 c	Cu(OTf)2	dtbbpy	DCE	66
4 c	Cu(OTf)2	dOMebpy	DCE	56
5 c	Cu(OTf)2	dCF3bpy	DCE	7
6 c	Cu(OTf)2	bpy	DCE	66
7 c	[Cu(dtbbpy)2](OTf)2	-	DCE	72
8 c	[Cu(dtbbpy)2](OTf)2	-	PhCl	89, 87d
9 c,e	[Cu(dtbbpy)2](OTf)2	-	PhCl	73d
10 c,f	[Cu(dtbbpy)2](OTf)2	-	PhCl	57
11 c,f,g	[Cu(dtbbpy)2](OTf)2	-	PhCl	63
12	-	-	PhCl	N.R.
13	AIBN (10 mol %)	-	PhCl	<5
14 e	BEt3 (10 mol %)	-	PhCl	N.R

^aReaction condition: 9 (0.20 mmol), 10 (1.5 equiv), Cu (4.0 mol %), and ligand (8.0 mol %) in degassed solvent (0.50 M), 80 °C, 24 h; in situ yields determined by GC and comparison to an internal standard.

^bYield of 11’.

^cN-methylaniline (8.0 mol %) and K₃PO₄ (8.0 mol %) were added as reductants.

^dIsolated yield.

^e5.0 mmol scale reaction.

^fReaction was set up on the benchtop and open to air.

^gReaction was sparged with N₂. dtbbpy = 4,4’-di-tert-butyl-2,2’-dipyridyl.

With the optimized conditions in hand, we next explore the scope of the reaction. The Cu-catalyzed ATRA can be utilized with both nucleophilic and electrophilic radical precursors, including tri- and di-substituted alkyl halides (Table 2, A and B). Notably, nucleophilic radical precursors afford higher yields in PhCl while electrophilic radical precursors perform better in DCE. The functional group compatibility of this method complements that of the Matteson homologation, as nucleophilic radical precursors bearing easily oxidizable groups, acid- and base-sensitive moieties participate in this transformation giving ATRA products (17‒20) in good yields. Chiral auxiliaries on either alkyl halide or vinyl boronic ester do not efficiently control diastereoselectivities of the resulting products 27 and 28. Likewise, disubstituted alkyl halides are tolerated forming the ATRA products (29‒34) as a 1:1 mixture of diastereoisomers. Moreover, we found that tertiary electrophilic radical precursor, diethyl 2-bromo-2-methylmalonate, has low reactivity under the optimized conditions. We hypothesized that utilizing a more active ligand may facilitate this process by increasing the activation of the alkyl halide while slowing the deactivation‒halogenation, which increases the concentration of R• and therefore facilitates its addition to 10. With that, we employed TPMA in the reaction and successfully rescued the reactivity to afford the desired product 22 in 57% yield.

Table 2.

Scope of alkyl halides and substituted vinyl boronic esters.^a

Reaction conditions:

^aAlkyl halides (1.0 equiv), vinyl B(pin) (1.5 equiv), [Cu(dtbbpy)₂](OTf)₂ (4.0 mol %), N-methylaniline (8.0 mol %), K₃PO₄ (8.0 mol %), PhCl (0.5 M), 80 °C.

^bDCE instead.

^cAlkyl halide (2.0 equiv), vinyl boronic ester (1.0 equiv), DCE.

^dAlkyl halide (3.0 equiv), vinyl B(pin) (1.0 equiv), DCE.

^eAlkyl halide (2.0 equiv), vinyl B(pin) (1.0 equiv), Cu(OTf)₂ (4.0 mol %), TPMA (4.0 mol %), N-methylaniline (8.0 mol %), K₃PO₄ (8.0 mol %), PhCl/MeCN (3:1).

^fAlkyl halide (2.0 equiv), vinyl B(Epin) (1.0 equiv), Cu(OTf)₂ (4.0 mol %), TPMA (4.0 mol %), K₃PO₄ (8.0 mol %), DCE (1.0 M), 90 °C.

Subsequently, we aimed to expand our scope by exploring primary halides which have posed a challenge in ATRA reaction. A significant challenge lies in the selective activation of primary halides over inherently more activated secondary halides. Following that, we examined the reactivity of 11 with 10 under the optimized condition. Product 35 was detected in trace amounts; hence, we hypothesized that a broad range of primary halides could be compatible due to the stability of the α-haloboronic ester against the Cu catalyst. However, a trace amount of ATRA product and S_N2 side product with N-methylaniline are observed when ethyl α-bromoacetate is subjected to the optimized conditions. As seen in Table S6, control reaction between of 9 and 10, in the absence of N-methylaniline, affords 11 in 33% yield.¹¹ This result suggests K₃PO₄ can promote the formation of a Cu(I) active catalyst, as supported by UV-Vis spectroscopy.^{11, 14} Based on this observation, we sought conditions for the ATRA to 10 with primary halides in the absence of N-methylaniline. Moreover, similar to the reaction involving diethyl 2-bromo-2-methylmalonate, TPMA exhibits the ability to facilitate this transformation.¹¹ With that, we successfully accessed ATRA product in 70% in-situ yield under the following conditions: ethyl α-bromoacetate (2.0 equiv), 10 (0.20 mmol) in the presence of Cu(OTf)₂ (4.0 mol %), TPMA (4.0 mol %), K₃PO₄ (8.0 mol %) and DCE (1.0 M), at 90 °C for 24 h.¹¹ Subjecting vinyl B(Epin) in the condition gives a comparable result, as 36 is isolated in 66% yield. Gratifyingly, the optimized conditions present excellent compatibility with a wide variety of primary halides (Table 2C).

2-Bromoacetonitrile and benzyl 2-bromoacetate undergo ATRA to afford the 37 and 38 in very good yield. More importantly, electronically distinct benzylic halides are also compatible and give the XAT products (39–45) in moderate yields. Again, the functional group tolerance is excellent, as aryl bromides, alkyl fluorides, thioethers, acidic C–H bonds, esters, and nitriles remain intact (36–45). Finally, bioactive moieties such thiophene and cyclic carbonates are tolerated (46 and 47).

Next, we proceeded to explore the reaction’s applicability with α-substituted vinyl Bpin derivatives. A variety of groups are tolerated in the reaction, including methyl, nonyl, ethyl TBS ether, cyclohexyl, and phenyl (48–51). Further, a diverse array of radical precursors proved suitable coupling partners under these conditions (52–57). Although the Cu(I) complex is unlikely to activate the ATRA product 11, more nucleophilic α-alkyl-α-haloboronic esters may be active enough to participate in the ATRA process as the radical precursors. Indeed, both 48 and 54 participate in the reaction with 10, yielding 59 and 60, respectively. Finally, β-substituted vinyl boronic esters are not suitable alkene coupling partners, likely due to steric hindrance slowing the radical addition step.

α-Haloboronic esters are known as versatile building blocks for further transformations through a 1,2-migration process with appropriate nucleophiles. Indeed, in the presence of a stoichiometric amount of N-methylaniline the α-aminoboronic ester 61 is formed in 50% in-situ and 22% isolated yield (Scheme 2A). Moreover, we illustrate a wide range of nucleophiles that can be incorporated to afford α-boronic ester derivatives through a telescoped, one-pot two-step strategy (Scheme 2B). Grignard reagents, including vinyl, phenyl, and allyl magnesium bromide, generate 62–64 with new C_sp3–C_sp2 or C_sp3–C_sp3 α to the boronic ester in good yields. Likewise, alkali salts, NaOMe and LiHDMS, can be used to provide the ether and amine products 65 and 66. Additionally, in the presence of iPr₂EtN, both phenol and benzenethiol undergo the substitution reaction, as 67 and 68 are isolated in 71% and 74% yield, respectively. Functionalization of the quaternary α-haloboronic esters is also demonstrated (69–71). Finally, we sought to demonstrate the synthetic utility of our transformation. Ixazomib, a drug molecule used to treat multiple myeloma, derivative 73 is synthesized in four steps: α-bromo boronic ester 11 undergoes substitution with LiHMDS and subsequent deprotection to obtain α-aminoboronic ester 72. Ultimately, peptide coupling between the free amine and N-(2,5-dichlorobenzoyl) glycine yields 73 (Scheme 2C).

Scheme 2.

Representative one-spot two step derivatives and synthesis of ixazomib derivative.^a

^aSee SI for reaction conditions.

In conclusion, we have successfully developed the Cu-catalyzed ATRA for the direct synthesis of α-haloboronic esters. This method obviates the necessity for strong bases and cryogenic temperatures. The copper acts as an exceptional XAT mediator, which bypasses the limitations associated with traditional radical chain approaches for ATRA reactions with vinyl boronic esters. This methodology has an excellent scope with respect to the radical precursors, as tri-, di-, and monosubstituted alkyl halides participate. Further, both unsubstituted and 1-substituted vinyl boronic acids are readily functionalized. Finally, a telescoped one-pot, two-step protocol is developed which allows direct access to a variety of functional groups at α-position. The operational simplicity, broad functional group tolerance, and scalability of this reaction render it suitable for adoption to the synthesis of relevant α-haloboronic ester derivatives, as highlighted in the synthesis of an ixazomib derivative.