Synthesis and applications of a new class of phosphorus donorligands for asymmetric catalysis

摘要

HN Me HN Me H H N P N Ar Me H H 1a Ar = Ph (100) b Ar = p-MeOC6H4 (96) c Ar = o-MeOC6H4 (100) d Ar = o-MeSC6H4 (98) e Ar = o-EtC6H4 (95) f Ar = o-PriC6H4 (100) g Ar = o-BrC6H4 (97) h Ar = o-MeOCH2C6H4 (97) i Ar = O Ar P(NMe2)2 + i Me N P N H H H3B Me Me ndash; 2 OMe + OAc Ph Ph MeO2C CO2Me Ph Ph MeO2C CO2Me Ph Ph or i R S 3 Synthesis and applications of a new class of phosphorus donor ligands for asymmetric catalysis Heather Tye,a Donald Smyth,a Colin Eldredb and Martin Wills*a a Department of Chemistry, University of Warwick, Coventry, UK CV4 7AL b Glaxo-Wellcome Research and Development Ltd, Glaxo-Wellcome Medicines Research Centre, Gunnels Wood Road, Stevenage, Hertfordshire, UK SG1 2NY The synthesis and applications to allylic substitutions of a range of novel ligands based on diazaphospholidines is described; enantiomeric excesses of up to 89 were achieved in an allylic substitution reaction.Recent years have seen dramatic developments in asymmetric catalysis.1 Allylic substitution reactions in particular have been the subject of attention from a number of research groups and several excellent ligands have been developed for this particular application.2 Of these, diphosphines and combined phosphinendash; amine donors have proven to be especially well suited to the required role.2 As part of an ongoing programme3 directed at the development of a range of phosphorus donor ligands with a wide range of applications to asymmetric synthesis, we wished to investigate diazaphospholidines 1 in this capacity.We anticipated that such ligands would benefit from a high level of conformational rigidity and would project a suitable chiral environment within organometallic complexes of which they may form a part. Furthermore we were aware of the facile synthesis of such materials from the corresponding C2 symmetric chiral diamines.4 Although similar compounds have been reported to act as effective chiral shift reagents, we were unaware of any catalytic asymmetric transformations to which they had been applied.5 The preparation of a series of ligands was achieved following literature precedents (Scheme 1).4 The bis(dimethylamino) phosphine precursors to 1a, 1b, 1c, 1e, 1h and 1i were formed by the reaction of the corresponding lithiated aromatic compound with preformed (Me2N)2PCl.The corresponding precursors for 1d and 1f could not be prepared by this method. However lithiation and transmetallation with zinc dichloride, followed by reaction with phosphorus tirchloride and subsequently treatment with an excess of dimethylamine furnished the required intermediates.The precursor to compound 1g was synthesised following the published procedure of Drewelies and Latscha.6 Each intermediate bis(dimethylamino)phosphine was heated under reflux with (R,R)-N,NA-dimethyl-1,2-diaminocyclohexane until no further dimethylamine could be detected in the nitrogen stream from the reaction.At this point solvent removal provided the pure ligands 1. Although always handled under an inert atmosphere when in solution, each of the ligands proved to be resistant to atmospheric oxidation when in solid form. Treatment of 1c with BH3middot;SMe2 complex resulted in formation of a highly stable borane complex 2 which could be purified by chromatography. We have in previous work demonstrated that related borane complexes could be used as a source of protected ligand3 (borane may be removed by an amine), however in the case of 2 all attempts to remove the borane resulted in decomposition.We have applied the series of ligands thus produced to a number of asymmetric transformations. In particular the ligands appear to be particularly well disposed to the nucleophilic allylic substitution reaction of 1,3-diphenyl-3-acetoxyprop- 1-ene 3. In most cases the substitution reaction with dimethyl malonate (Scheme 2) gave good yields and, in the best cases, excellent asymmetric inductions (Table 1).A number of interesting trends become apparent upon analysis of the results. In particular a general improvement and reversal of selectivity is observed upon introduction of an ortho substituent. The level of electron density in the aromatic ring does not appear to influence the selectivity to any great extent (compare ligands 1a and 1b). Although the best results, 89 and 78 ee in favour of the R enantiomer, were achieved using the ligands 1c and 1d, the same product enantiomer was obtained using ligands which do not contain coordinating groups (i.e. 1e and 1f). This suggests that all may share a common mode for directing the asymmetric transformation. The importance of the size of the ortho substituent is also difficult to gauge since the ethyl and isopropyl substituted ligands 1e and 1f gave similar selectivities. We were unable to prepare the corresponding tert-butylsubstituted reagent.At present there is insufficient evidence to determine the exact means by which asymmetric induction by the ligand is achieved. It would appear that whilst the introduction of an ortho substituent in the ligand is crucial to their success, the observation of similar results from coordinating and noncoordinating groups suggests that the ligand is probably only appended to the palladium through the phosphorus atom. If this is the case then their mechanism of action may have a parallel with the phosphine ligand MOP, which has been demonstrated to be effective for this application through the formation of a 1 : 1 phosphinendash;metal complex.7 Of significance is the pseudo-C2 symmetry of the ligand and potential participation of a conformational rigidity conferred to Scheme 1 Reagents and conditions: i, PhMe, heat, 8ndash;24 h Scheme 2 Reagents and conditions: i, 1andash;i, Pd(h-C3H5)Cl2, dimethyl malonate, (Me3Si)N:CMe(OSiMe3), NaOAc, CH2Cl2 Chem.Commun., 1997 1053N P N X H H Me Me N N P Me Me MeO bull; bull; H H Ph Ph OAc 3 Ph Ph N R1 R2 ( S)-4 a R1 = H, R2 = Bn b R1 = H, R2 = ( R)-CHMePh c R1 = H, R2 = ( S)-CHMePh d R1 = H, R2 = p-MeC6H4SO2 e R1ndash;R2 = (CH2)4 f R1ndash;R2 = (CH2)2N(Bn)(CH2)2 i the nitrogen methyl groups by the adjacent chiral centre.Studies by others,4,8 and our own X-ray crystal structure of the borane complex 29 indicate that the nitrogen atoms are partially tetrahedral, the result of the methyl groups avoiding eclipsing interactions with the adjacent Cndash;C bonds (Fig. 1). The X-ray crystal structure also reveals that the ortho substituent favours a position away from the heterocyclic ring (Fig. 1). This suggests that steric interactions may prevent a 180deg; rotation about the Pndash;C(aryl) bond. This proposition is supported by our observation that 2,6-disubstituted aromatic ring diazaphospholidines failed to be formed using our general method (Scheme 1). If this conformational preference is maintained in the palladium complex then clearly this ortho group may be projected into the region of the appended allylic group, and thus provide a means to influence the reaction selectivity.Precisely how this effect is transferred to the asymmetric reaction, however, is at present unclear. In an attempt to increase the observed selectivity we attempted some studies of diazaphospholidines bearing larger groups on the nitrogen atoms. The dibenzyl derivative of 1c gave a product of 64 ee in 78 yield for the reaction depicted in Scheme 2.The diisopropyl ligand failed to form using our standard procedure, possibly due to steric hindrance. Having identified 1c as the best of the series of ligands, we have examined the use of nitrogen nucleophiles in the substitution reaction (Scheme 3, Table 2). The results obtained using benzylamine and sodium toluene-p-sulfonamide compared favourably to those previously reported using other ligands,2,10 but were significantly lower (although in the same sense) than the selectivity achieved using dimethyl malonate as nucleophile.The presence of a chiral centre in the nucleophile (a-methylbenzylamine of either configuration) appeared to have no effect on the enantioselectivity of the reaction or the sense of induction. Use of the cyclic amines pyrrolidine and N-benzylpiperazine as nucleophiles resulted in the formation of novel allylic amines 4e and 4f respectively, in excellent yields but reduced enantioselectivities.In conclusion a novel class of diazaphospholidine ligands has been synthesised and their application to palladium-catalysed allylic substitution has been demonstrated. Ongoing work is directed towards the synthesis of improved ligands based on this class for a variety of transition metal-catalysed asymmetric transformations. We thank the EPSRC and Glaxo-Wellcome Research and Development for support of an Industrial CASE studentship (to H. T.) and Dr J.Ballantine of the EPSRC National Mass Spectroscopic service (Swansea) for HRMS analysis of certain compounds. Footnote * E-mail: m.wills@warwick.ac.uk References 1 R. Noyori, Asymmetric Catalysis in Organic Synthesis, Wiley, NY, 1994; I. Ojima, Catalytic Asymmetric Synthesis, VCH, Berlin, 1993. 2 B. M. Trost and D. L. Van Vranken, Chem. Rev., 1996, 96, 395; J. M. J. Williams, C. G. Frost and J. Howarth, Tetrahedron: Assymmetry, 1992, 3, 1089. 3 G. Brenchley, E. Merifield, M.Wills and M. Fedouloff, Tetrahedron Lett., 1994, 35, 2791; G. Brenchley, M. Fedouloff, M. F. Mahon, K. C. Molloy and M. Wills, Tetrahedron, 1995, 51, 10 581; G. Brenchley, M. Fedouloff, E. Merifield and M. Wills, Tetrahedron: Asymmetry, 1996, 7, 2809. 4 A. Alexakis, S. Mutti and P. Mangeney, J. Org. Chem., 1992, 57, 1224; R. Hulst, N. Keon de Vries and B. L. Feringa, Tetrahedron: Asymmetry, 1994, 5, 699; R. Hulst, R. M. Kellogg and B. L. Feringo, Recl. Trav. Chim. Pays-Bas, 1995, 114, 115; P.G. Devitt, M. C. Mitchell, J. M. Weetman, R. J. Taylor and T. Kee, Tetrahedron: Asymmetry, 1995, 6, 2039. 5 G. Buono and P. Cros, Nouv. J. Chim., 1987, 11, 573; W. J. Richter, Chem. Ber., 1984, 117, 2328. 6 K. Drewelies and H. P. Latscha, Angew. Chem., Int. Ed. Engl., 1982, 21, 638. 7 T. Hayashi, H. Iwamura, Y. Uozumi, Y. Matsumoto and F. Ozawa, Synthesis, 1994, 526; T. Hayashi, H. Iwamura, M. Naito, Y. Matsumoto, Y. Uomuzi, M. Miki and K. Yanagi, J. Am. Chem.Soc., 1994, 116, 775. 8 A. Alexakis, N. Lensen, J.-P. Tranchier, P. Mangeney, J. Feneau- Dupont and J. P. Declercq, Synthesis, 1995, 1038; S. E. Denmark and R. L. Dorow, J. Am. Chem. Soc., 1990, 112, 864; S. E. Denmark, P. C. Miller and S. R. Wilson, J. Am. Chem. Soc., 1991, 113, 1468; C. J. Cramer, S. E. Denmark, P. C. Miller, R. L. Dorow, K. A. Swiss and S. R. Wilson, J. Am. Chem. Soc., 1994, 116, 2437; S. Hanessian and S. Beaudoin, Tetrahedron Lett., 1992, 33, 7655. 9 N.Alcock, H. Tye and M. Wills, unpublished result. 10 P. Von Matt, O. Lieseleur, G. Koch, A. Pfaltz, C. Lefeber, T. Feucht and G. Helmchen, Tetrahedron: Asymmetry, 1995, 5, 573; R. Jumnah, A. C. Williams and J. M. J. Williams, Synlett, 1995, 821. Received in Cambridge, UK, 19th March 1997; Com. 7/01922J Table 1 Asymmetric addition of dimethyl malonate to 3 catalysed by 1andash;i Ligand ee ()a Ligand Pd (mol) (mol) t/h Yield () (config.) 1a 10 20 16 97 28 (S) 1b 10 20 16 71 23 (S) 1c 10 20 16 97 89 (R) 1c 10 10 20 89 83 (R) 1c 5 10 12 85 85 (R) 1c 2 4 20 94 82 (R) 1d 10 20 24 90 78 (R) 1e 10 20 16 84 59 (R) 1f 10 20 16 91 56 (R) 1g 10 20 24 85 66 (R) 1h 10 20 24 84 51 (R) 1i 10 20 20 92 35 (R) a Enantiomeric excess determined by 1H NMR spectroscopy using (+)-Eu(hfc)3 as a chiral shift reagent. Fig. 1 Scheme 3 Reagents and conditions: i, 1c, Pd(h-C3H5)Cl2, R1R2NH, NaOAc, THF or CH2Cl2 Table 2 Asymmetric addition of nitrogen nucleophiles catalysed by 1c eea () Product Solvent t/h Yield () (config.) 4a CH2Cl2 42 68 78 (S) 4a THF 20 49 73 (S) 4a THF 5 77 68 (S) 4b CH2Cl2 168 67 78 (S) 4b THF 8 51 61 (S) 4c THF 8 67 61 (S) 4d CH2Cl2 64 45 76 (S) 4e CH2Cl2 41 48 68 (S) 4f CH2Cl2 16 63 58 (S) a Enantiomeric excess determined by chiral HPLC using a chiralcel OD column. 1054 Chem. Commun., 1997