Previous work

The 1,4-addition/aldolisation reaction is classified as a domino type reaction. This means that these two reactions are occurring successively and in our case catalyzed by copper(I) complexes. Indeed 1,4-addition reactions using in situ formed Cu-B or Cu-Si (see general introduction) can be further exploited when ran in the presence of an electrophile, inducing an inter- or intramolecular trapping of the addition product. In our case, the aldolisation process involves trapping the copper intermediates with aldehydes or ketones to furnish aldol products in one step (Scheme 4). This type of domino process was already developed in our group with Cu-H species as described earlier in the general introduction. The idea here is to use other nucleophiles such as silylboranes or bisboronates.

I.2.1. Borylation reaction

During his PhD thesis, Alexandre Welle developed a diastereoselective domino borylative aldol reaction using B2pin2 13 as pronucleophile.5 A copper(I) catalyst generated from [(Ph₃P)CuF].2MeOH 25 and (rac)- BINAP 26 promoted the conjugate boryl transfer onto various acyclic Michael acceptors and the resulting enolate was trapped with various aldehydes (Scheme 5). Aldols adducts were obtained with good yields and moderate diastereoselectivities (up to 3 to 1 in favour of syn isomer). It is important to mention that because boronate compounds are unstable on silica gel, products were isolated after oxidation of the boronic moiety under mild conditions to the corresponding alcohol.6

Scheme 5. Copper(I)-catalyzed domino borylative aldol reaction.

Relative configuration was determined after treatment of compound 27 with DMP in the presence of a catalytic amount of PTSA (Scheme 6). Dioxolane 28 was obtained in quantative yield and NMR analysis

5 _{Welle, A.; Cirriez, V.; Riant, O., Tetrahedron 2012, 68, 3435-3443.} 6 _{Mun, S.; Lee, J.-E.; Yun, J., Org. Lett. 2006, 8, 4887-4889.}

showed unambiguously a trans relationship between the amide and the phenyl for the major product.

Scheme 6. Determination of the relative configuration of 27.

While modest diastereoselectivities were observed with acyclic Michael acceptors, they decided to check the reactivity of simple cyclic enones as models. Geometry of the enolate is blocked in cyclic compounds thus the diastereocontrol would be better. As the true nature of the enolate intermediate that will react with the aldehyde has still not been established as well as the transition state involved in the aldol process between the metal enolate and the aldehyde, previsions on the stereochemical outcome of cyclic substrates are not straightforward. However, a recent study by Hoveyda group gave one example of a highly diastereoselective tandem conjugated addition/aldol reaction on cyclohexenone using NHC complex (Scheme 7).7

7 _{Lee, K.-S.; Zhugralin, A. R.; Hoveyda, A. H., J. Am. Chem. Soc. 2009, 131,}

7253-7255. For the first metal-free enantioselective conjugate borylation of α,β- unsaturated compounds with phosphorus ligands, see: Bonet, A.; Gulyás, H.; Fernández, E., Angew. Chem. Int. Ed. 2010, 49, 5130-5134.

Scheme 7. Tandem conjugated addition/aldol reaction using NHC

complex by Hoveyda’s group.

Likewise, we tested cyclohexenone and cyclopentenone as model substrates and found different behaviour depending on the ring size of the Michael acceptor. In all cases, the boron ester arising from the tandem process was identified in the crude reaction mixture but was not isolated and was directly oxidized to the corresponding diol, which was then protected as a dioxolane in a one pot sequence for relative configuration determination (Scheme 8). Although the unoptimised yield with cyclohexenone is modest, only one diastereoisomer was detected in the crude reaction mixture prior to oxidation. In contrast, the spontaneous dehydration after boronic ester oxidation when using cyclopentenone was observed, although diastereocontrol was very good. Unfortunately, every attempt to prevent water elimination failed.

Scheme 8. Domino reaction using cyclic enones.

The use of a functionalised chiral Michael acceptor was also investigated and focussed on the enantiopure enone 30, which was easily prepared from D-ribose using a protocol from the literature.8 _{After reacting 30 in} the standard catalytic conditions, the desired adduct was obtained as the major product and only one diastereoisomer was detected in the crude reaction mixture, meaning that the absolute configuration of three new centres was controlled (Scheme 9). As the oxidation of the boronate ester was not possible for cyclopentenone, the pure adduct was not isolated and stereochemistry of the newly formed chiral carbon centres was not confirmed.

8 _{Choi, W. J.; Park, J. G.; Yoo, S. J.; Kim, H. O.; Moon, H. R.; Chun, M. W.;}

Scheme 9. Domino reaction using enantiopure cyclic enone.

However, organoboranes derivatives are quite unstable in some cases although they are versatile intermediates in organic synthesis. Indeed their main application is found in the Suzuki cross-coupling reaction. Thus, valorization of these adducts with palladium catalysis might be interesting. After condensation of benzaldehyde, methyl acrylate and 13 using the standard catalytic conditions, the solvent was evaporated from the crude reaction mixture without any work-up and replaced by DMF. The Suzuki reaction was then directly carried out with iodobenzene using Pd(PPh₃)₄ as a catalyst and cesium carbonate as a base using an optimised protocol from the literature.9_{The corresponding adduct could} be isolated as a separable mixture of syn/anti diastereoisomers in an excellent 89% yield after two steps (Scheme 10).

9 _{Lawrence, J. D.; Takahashi, M.; Bae, C.; Hartwig, J. F., J. Am. Chem. Soc. 2004,} 126, 15334-15335.

Scheme 10. Domino borylation/aldol/Suzuki coupling.

This preliminary result shows that it is possible to have access to structural diversity on this family of adducts by simple variation of the aldehyde and the aryl iodide. In order to go further in the synthetic versatility of the three component adducts, they finally checked the transformation of a model adduct bearing a Weinreb amide group. The borate ester intermediate 31 was not isolated and directly reacted with a large excess of methyl lithium to carry out the transformation of the amide group into the corresponding methyl ketone 32 (Scheme 11).

Final oxidation of the boronic ester by sodium perborate delivered the desired keto-diol 33 in a 48% overall yield for the three steps and a 3/1 syn/anti ratio.

In summary, an efficient procedure for the tandem borylation/aldol reaction based on copper catalysis was developed. A wide variety of aldehydes were used and the desired adducts were obtained in good to excellent yields. Regarding the Michael acceptor, the catalytic system tolerates α substitution. Different types of electronwithdrawing groups on the Michael acceptor may be employed and low level of diastereoselectivity is observed with linear substrates. However, the use of cyclic substrates increases dramatically the diastereoselectivity and enantiopure substrates lead to the control of the absolute configuration of three new contiguous centres. It is also possible to combine this protocol with a Suzuki coupling.

I.2.2. Silylation reaction

Our group wanted also to develop a tandem process using silylboranes as pronucleophiles. Although the use of silyl-metal species is not suitable for silylative aldol because of potential direct addition of the silyl group to the aldehyde (this problem of chemoselectivity is well known in copper hydride chemistry),10 _{they described a diastereoselective domino}

10 _{(a) Lipshutz, B. H.; Amorelli, B.; Unger, J. B., J. Am. Chem. Soc. 2008, 130,}

14378-14379. (b) Chuzel, O.; Deschamp, J.; Chausteur, C.; Riant, O., Org. Lett.

silylative aldol reaction using silylborane 19.11 _{At the same time, an} asymmetric conjugated addition of silylborane on electrophilic double bonds with high enantioselectivities was reported by Hoveyda using chiral copper-diaminocarbene complexes with an example of a tandem aldol reaction.12

Initial experiments performed by our group showed that reaction between various carbonyl compounds (aldehydes and ketones), various Michael acceptors and a stoichiometric amount of silylborane 19 in toluene at room temperature in the presence of [CuF(PPh3)3]·2MeOH 25, and (rac)-BINAP 26 readily afforded the desired aldol adduct in good

yield (Scheme 12).

Scheme 12. Copper(I)-catalyzed domino silylative aldol reaction.

They observed an excellent reactivity of their catalytic system and excellent yields of the three components adducts were obtained in each case with low catalyst loading and short reaction time. However, no or very low diastereoselectivity was achieved except for the intramolecular

11 _{Welle, A.; Petrignet, J.; Tinant, B.; Wouters, J.; Riant, O., Chem. Eur. J. 2010,} 16, 10980-10983.

version, cyclic enones or Weinreb amides. They therefore decided to use other substrates. Their hypothesis was that the stereochemistry could be controlled by using chiral auxiliaries such as oxazolidinones on the Michael acceptors. After optimization of the catalytic system and a survey of various achiral diphosphine ligands, they found that a better reproducibility could be achieved by replacing (rac)-BINAP 26 by DPPF (1,1’-bis(diphenylphosphino) ferrocene) and toluene by THF. Under these optimised catalytic conditions, the reaction between various aromatic aldehydes and acryloyloxazolidinones 34 gave the corresponding syn-aldol adducts in good yields (59-80%) and with almost complete control of the stereochemistry in most cases (Scheme 13). The absolute configuration of these compounds was determined by X-ray diffraction.

Scheme 13. Copper-catalyzed silylative–aldol reaction between

They next investigated the use of methacryloyloxazolidinones 35 with various aromatic and heteroaromatic aldehydes. Moderate to good isolated yields were obtained (59-77%) and diastereoselectivities in the range of 78/22 to 92/8 were achieved (Scheme 13). This means that the absolute stereochemistry of the chiral quaternary carbon chiral centres was controlled. The structure of the aldol adduct was first attributed to a simple open structure but a rearranged structure was later attributed when an X-ray crystal-structure analysis of the 2-furyl carboxaldehyde adduct was carried out to analyze the absolute configuration of the newly formed asymmetric carbon centres. The rearranged adduct 36 is assumed to arise from an intramolecular ring opening of the oxazolidinone moiety by the hydroxyl group of the aldol adduct (Scheme 14).

Scheme 14. Formation of rearranged adduct 36.

They finally investigated the effect of the structure of the chiral auxiliary (Scheme 15). The optimization was carried out with thiophene-2- carboxaldehyde because it gave the lowest diastereomeric ratio. The selectivity is dependent on the steric hindrance because the benzyl- substituted substrate led to a lower selectivity. The isopropyl-substituted oxazolidinone led to an improved dr and the tert-butyl derivative restored the full control of the selectivity.

Scheme 15. Effect of the structure of the chiral auxiliary.

In summary, they have developed a domino silylation–aldol reaction between enoyloxazolidinones and various aldehydes. The process catalyzed by a copper–diphosphine complex in the presence of silylborane led to high diastereoselectivities. Yields ranging from 59 to 90% and diastereoisomeric ratios up to 95/5 were obtained. When α methyl-substituted acryloyloxazolidinones were used, a controlled chiral quaternary carbon was generated.