Enantio- and Diastereoselective Copper-Catalyzed Allylboration of Alkynes with Allylic gem-Dichlorides.

Allylic gem-dichlorides are shown to be efficient substrates for catalytic asymmetric allylboration of alkynes. The method employs a chiral NHC-Cu catalyst capable of generating in a single step chiral skipped dienes bearing a Z-alkenyl chloride, a trisubstituted E-alkenyl boronate and a bis-allylic stereocenter with excellent levels of chemo-, regio- enantio- and diastereoselectivity. This high degree of functionalization makes these products versatile building blocks as illustrated with the synthesis of several optically active compounds. DFT calculations support the key presence of a metal cation bridge ligand-substrate interaction and account for the stereoselectivity outcome.

The presence of multiple stereochemical elements is a recurrent structural feature found in many drugs and natural products. Thus, catalytic enantioselective transformations that provide in a single operation a densely functionalized building block with several stereodefined elements in its structure are highly sought after. Since pioneering studies by Hosomi and Miyaura, copper‐catalyzed borylative couplings of unsaturated hydrocarbons have become a powerful tool for the preparation of synthetically versatile organoboron compounds. In particular, asymmetric allylboration provides access to chiral multifunctional molecules featuring both a boronic ester and an olefin group. Efficient methods either based on copper/bisphosphine or synergistic Cu/Pd catalysis where the stereogenic center is installed in the LCu–Bpin alkene insertion step have been reported for the allyboration of styrene derivatives (Scheme 1a). A different approach relies on the enantioselective allylic substitution of the catalytically generated organocopper intermediate, as described by Hoveyda in the borylative coupling between allenes and allylic phosphates (Scheme 1b). We recently applied this strategy to the copper‐catalyzed asymmetric allylboration of alkynes which involves the regio‐ and enantioselective coupling of an allylic bromide with a catalytically generated β‐boryl‐substituted alkenylcopper intermediate (Scheme 1c).[ , ] Realizing that the presence of two chlorine atoms in an allylic gem‐dichloride might result in a dual role as leaving group and as a stereocontrol element, we envisaged that an asymmetric alkyne allylboration reaction based on this type of allylic substrates might lead to an unprecedented degree of complexity in which two new stereodefined olefins and a bis‐allylic stereocenter would be created in a single step (Scheme 1d). The products of this transformation would bear two orthogonal functionalities, an alkenyl boronate and an alkenyl chloride, which offer diverse possibilities for the functionalization of the important skipped diene core. Such a process represents a formidable challenge since diastereoselective formation of the alkenyl chloride must be achieved besides control over the chemo‐, regio‐ and enantioselectivity. It is important to note that copper‐catalyzed enantioselective allylic substitution (EAS) has been mainly applied to primary allylic substrates for the synthesis of terminal olefins,[ , ] and processes involving acyclic secondary 1,1‐disubstituted allylic substrates are rare. Indeed, the use of this type of substrates in copper‐catalyzed EAS has been limited to the introduction of alkyl Grignard compounds, and the control over the geometry of the resulting internal double bond represents a major issue.

Scheme 1

Copper‐catalyzed enantioselective allylboration of unsaturated hydrocarbons.

Herein we describe the successful use of allylic gem‐dichlorides in a copper‐catalyzed asymmetric borylative coupling with alkynes which efficiently provides chiral densely functionalized skipped dienes bearing a Z‐alkenyl chloride and a trisubstituted stereodefined alkenyl boronate. The fact that C−C and C−B bond formation is achieved with high chemo‐, regio‐, enantio‐ and diastereoselectivity while establishing one stereocenter and two stereodefined internal olefins makes this method significant and distinct from previous copper‐catalyzed carboboration and EAS reactions. Moreover, it represents a new synthetic approach among the few catalytic methods available for the stereoselective synthesis of acyclic 1,2‐disubstituted alkenyl chlorides.

We began our study by exploring the reaction between phenylacetylene 1, (E)‐(3,3‐dichloroprop‐1‐en‐1‐yl)benzene 2 and B2pin2 (Table 1). In our previous work on the copper‐catalyzed enantioselective alkyne allylboration with allylic bromides, we found that the use of a chiral sulfonate‐bearing NHC ligand was key to control the selectivity of the process by establishing a metal cation bridge interaction between the copper complex and the allylic bromide. Thus, we envisaged that this type of supramolecular interaction might be also crucial to control both the enantioselectivity and the diastereoselective formation of the resulting alkenyl chloride in the reaction with an allylic gem‐dichloride. Indeed, the use of other types of chiral bisphosphine or NHC ligands led to nearly equimolar mixtures of dienes 3‐ and 3‐ in low yield with negligible enantioselectivity (entries 1–3).According to our hypothesis, a major increase both in efficiency and selectivity was observed when sulfonate‐bearing NHC ligands were employed. By using NaO Bu as base and toluene as solvent, the chiral Cu catalyst derived from ligand L4 which features a NMes unit provided diene 3‐ as the major isomer (E/Z : E/E=84 : 16) with a very good enantioselectivity (97 : 3 er) (entry 4). Interestingly, the use of ligand L5 bearing a slightly less bulky 2,6‐dimethyl substituted aryl ring led to a similar result with an improved 99 : 1 enantiomeric ratio (entry 5). In contrast, the use of ligands with bulkier aryl groups resulted in diminished conversion and selectivity (entry 6). Evaluation of different bases and solvents using L5 as ligand highlighted the important role of the base metal cation. A significant drop in selectivity was observed when KO Bu was used (entry 7). In contrast, product 3‐ was almost exclusively obtained with excellent enantiomeric ratio by using LiO Bu (entry 8). Low conversion and a slight decrease in selectivity were observed when a less polar solvent such as hexane was used (entry 9). Full conversion was recovered with the more polar solvent THF, although it caused a change in selectivity affording product 3‐ as the major isomer, although as a nearly racemic mixture (entry 10). Gratifyingly, the use of ligand L4 under the best set of conditions (LiO Bu as base and toluene as solvent) led to the exclusive formation of product 3‐ with perfect enantioselectivity (entry 11). X‐ray diffraction analysis of 3‐ revealed the relative and absolute stereochemistry of the product.

Table 1

Optimization studies.

[a] Reaction conditions: 1 (0.4 mmol), 2 (0.2 mmol), B2pin2 (0.4 mmol), CuCl (10 mol %), ligand (12 mol %), base (0.4 mmol), solvent (1.5 mL) at 30 °C. [b] Conversion (2 consumption) was determined by 1H NMR analysis using trimethyl benzene 1,3,5‐tricarboxylate as internal standard. [c] Determined by GC analysis of reaction crude. [d] Yield of isolated product. [e] Enantioselectivity determined by SFC analysis. [f] n.d.=not determined. Minor diastereomer could not be detected by GC analysis. [g] Hexane used instead of toluene. [h] THF used instead of toluene.

Entry[a]

L*

M

Conv. [%][b]

, / , ratio[c]

3 (, / , ) yield [%][d]

3 (, / , ) er[e]

1

L1

Na

21

n.d.[f]

4/12

57 : 43/53 : 47

2

L2

Na

40

41 : 59

12/24

54 : 46/50 : 50

3

L3

Na

23

n.d.[f]

4/13

56 : 44/49 : 51

4

L4

Na

84

84 : 16

51/13

97 : 3/49 : 51

5

L5

Na

88

85 : 15

64/12

99 : 1/57 : 43

6

L6

Na

44

78 : 22

20/8

7 : 93/47 : 53

7

L5

K

76

82 : 18

37/12

92 : 8/49 : 51

8

L5

Li

Full

92 : 8

67/6

99 : 1/50 : 50

9[g]

L5

Li

16

84 : 16

10/2

97 : 3/49 : 51

10[h]

L5

Li

Full

23 : 77

18/69

82 : 18/46 : 54

11

L4

Li

Full

99 : 1

74/–

>99 : 1/–

Having established optimized conditions, we set out to explore the scope of the reaction (Scheme 2). The Cu/L4 catalyst proved to be remarkably effective with a range of alkynes and allylic gem‐dichlorides, with L5 giving slightly superior results in some cases. Difunctional skipped E,Z‐dienes were obtained in good yields and with excellent regio‐, enantio‐ and diastereoselectivity in nearly all cases. Aryl alkynes with electron‐donating groups were well tolerated and provided dienes 4–6 as pure homochiral products. Similarly, alkynes bearing electron‐withdrawing groups such as halogen (7, 8) and ester (9) substituents also performed well, although they proved to be less reactive and required higher temperatures. The reaction also proceeded efficiently with alkynes featuring heterocyclic (10–11) and cyclopropyl (12) rings or a conjugated double bond (13). The use of acyclic alkyl‐substituted alkynes led to diminished yields, although corresponding products 14 and 15 were still obtained with high regio‐, enantio‐ and diastereocontrol. Importantly, as illustrated by the synthesis of diene 3, the reaction could be scaled up to 2 mmol using lower catalyst loading (5 mol %) without any loss in yield or selectivity.

Scheme 2

Scope of the reaction. [a] Reaction conditions: see Table 1, entry 11. Yield values refer to isolated products. [b] Reaction run on a 2 mmol scale using 5 mol % of catalyst. [c] L5 was used instead L4. [d] At 40 °C. [e] At 50 °C. [f] At 60 °C. [g] Obtained from a 0.8 : 1 mixture of the allylic 1,1‐ and 1,3‐dichlorides. Yield is referred to the 1,1‐isomer.

Allylic gem‐dichlorides with aryl groups bearing different substituents at the ortho‐, meta‐ and para‐positions were also well tolerated and provided the corresponding chiral dienes 16–20 in good yield with excellent selectivities. Only a slight decrease in the diastereoselectivity was observed when an ortho‐bromo substituted substrate was used. A substrate featuring a sterically demanding naphthalene group could also be used with very high selectivity (21). Notably, allylic gem‐dichlorides bearing aliphatic substituents were also efficient partners in this transformation. Despite being used as a mixture of 1,1‐ and 1,3‐isomers, these substrates provided the corresponding skipped dienes 22 and 23 in good yield and with excellent selectivity. Interestingly, an allylic gem‐dichloride derived from (−)‐citronellal which bears an extra double bond and a stereogenic center could be used with complete levels of chemo‐ and diastereoselectivity (24).

The presence of two orthogonal functionalities at both termini of the products offers a versatile synthetic handle for further chemoselective functionalization (Scheme 3). Different stereodefined chiral skipped dienes could be accessed by Suzuki cross‐coupling (25), Matteson homologation/oxidation (26) or a sequential combination of both (27). Notably, subsequent treatment of 23 with β‐bromostyrene and 1‐octenyl boronic acid under palladium catalysis afforded skipped tetraene 29. Few methods for the synthesis of these compounds have been reported and, to the best of our knowledge, examples of chiral enantioenriched skipped tetraenes are elusive.

Scheme 3

Synthetic modifications of products. Conditions: i) 1‐bromo‐2‐iodobenzene (1.5 equiv), Pd(PPh)3 (10 mol %), NaOH 2 M, dioxane, 100 °C; ii) CH2Br2 (2 equiv), n‐BuLi (0.5 equiv), THF, −78 °C to rt; iii) NaBO3 ⋅ 4H2O (5 equiv), H2O, rt; iv) TBSCl (1.2 equiv), imidazole (2 equiv), CH2Cl2, rt; v) 4‐fluorophenylboronic acid (1.5 equiv), Pd2(dba)3 (5 mol %), XPhos (10 mol %), CsF (3 equiv), dioxane, 100 °C; vi) β‐bromostyrene (1.5 equiv), NaOH 2 M, dioxane, 100 °C; vii) 1‐octenyl boronic acid (1.5 equiv), Pd2(dba)3 (5 mol %), XPhos (10 mol %), dioxane, 100 °C.

To rationalize the high levels of enantio‐ and diastereoselectivity, DFT calculations were performed (see the Supporting Information for details). Optimized structures for the stereochemistry‐determining SN2′‐type oxidative addition step[ , ] show that the allylic gem‐dichloride approaches the alkenylcopper intermediate opposite to the sizable N‐aryl group while establishing a lithium cation bridge interaction between the NHC's sulfonate group and one (or both) of the chloride units (Figure 1a). Comparison between transition states TS‐ , and TS‐ , shows that the former is favored by 9.0 kcal mol−1 and suggests that the origin of enantioselectivity arises from the repulsive steric interaction that is engendered between the Bpin unit and the phenyl substituent of the allylic substrate in transition state TS‐ , leading to the minor S enantiomer. A similar model may apply for the transition states associated to the formation of the minor E,E‐isomer (Figure 1b). In accordance with the experimental results, both TS‐ , and TS‐ , resulted higher in energy than the preferred TS‐ , .

Figure 1

Optimized structures and energies for the stereochemistry‐determining oxidative addition transition states obtained from DFT calculations performed at the ωB97XD/Def2‐TZVPPtol(SMD)//ωB97XD/6‐31G(d) level.

Another important aspect to elucidate was the effect of the metal cation in the diastereoselective formation of the Z‐alkenyl chloride. Analysis of structures for TS‐ , and TS‐ , (Figure 2a) shows that the lithium bridge imposes a bigger CNHC−Cu−Cα angle in TS‐ , (99.9° versus 93.0° in TS‐ , ). This angle opening pushes the allyl substrate closer to the Bpin unit leading to a larger repulsive interaction. Calculations using sodium instead of lithium (Figure 2b) showed that this cation causes a similar enlargement of the CNHC−Cu−Cα angle in TS‐ , ‐Na with respect to TS‐ , ‐Na (101.5° versus 92.0°). Distortion‐interaction analysis[ , ] of the stereodeterminig transition states helped to gather insight about the different results obtained with these two metal cations. These studies revealed that TS‐ , distortion energy (ΔΔE ≠ dist) is 13.7 kcal mol−1 higher than TS‐ , , while interaction energy (ΔΔE ≠ int) only favors TS‐ , by 5.3 kcal mol−1 (Figure 2a). Distortion energy of TS‐ , ‐Na related to TS‐ , ‐Na (ΔΔE ≠ dist=13.9 kcal mol−1) is comparable with the value observed for the transition states featuring a lithium cation. However, stabilizing interaction energy is significantly higher for the sodium‐based system (ΔΔE ≠ int=−9.1 kcal mol−1), thus resulting in a lower energy difference between both sodium‐based transition states (Figure 2b). The increase in the stabilizing interaction energy may be due to the larger size of the Na cation which might facilitate a more effective interaction with both chlorine atoms. This might explain the slightly lower Z/E selectivity observed when NaO Bu is used instead of LiO Bu.

Figure 2

Optimized structures (back view) for the stereochemistry‐determining oxidative addition transition states associated with the pathways leading to R,E,Z and R,E,E isomers using a) LiO Bu and b) NaO Bu. DFT calculations were performed at the ωB97XD/Def2‐TZVPPtol(SMD)//ωB97XD/6‐31G(d) level. Energies from distortion–interaction analysis (DIA) are referenced to the corresponding TS‐ , in each pair.

A different mechanistic scenario may apply when THF is used as solvent. Its more polar and coordinating nature may weaken (or disrupt) the supramolecular cation bridging ligand‐substrate interaction leading to a looser transition state, which would result in an enantioselectivity drop and the preferential formation of the E‐alkenyl chloride (Table 1, entry 10), as it happens when other NHC ligands lacking the sulfonate group are used (Table 1, entries 2 and 3).

In conclusion, we have developed a highly chemo‐, regio‐, enantio‐ and diastereoselective coupling of alkynes, diborons and allylic gem‐dichlorides. The use of this new class of allylic substrates for asymmetric copper‐catalyzed carboboration provides chiral orthogonally functionalized skipped dienes which serve as highly versatile optically active building blocks for asymmetric synthesis. Intrinsic features of the mechanism could be unveiled by DFT calculations, showcasing the crucial role of a lithium cation bridge ligand–substrate interaction as a key stereocontrol element.

Conflict of interest

The authors declare no conflict of interest.

As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re‐organized for online delivery, but are not copy‐edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.

Supporting Information

Click here for additional data file.