JM-PHOS ligands: Second-generation phosphine oxazolines for asymmetric catalysis

Article abstract of DOI:10.1021/ol991008k

(formula presented) A small library of phosphine oxazollne ligands 2 was prepared and tested in palladium-mediated allylation processes (reactions 1 and 2). They were found to be superior to the first-generation ligands 1 and as effective as the well-know

Full text of DOI:10.1021/ol991008k

ORGANIC
LETTERS
1999
Vol. 1, No. 11
1745-1747
JM-PHOS Ligands: Second-Generation
Phosphine Oxazolines for Asymmetric
Catalysis
Duen-Ren Hou and Kevin Burgess*
Department of Chemistry, Texas A&M UniVersity, Box 30012,
College Station, Texas 77842-3012
burgess@mail.chem.tamu.edu
Received September 3, 1999
ABSTRACT
A small library of phosphine oxazoline ligands 2 was prepared and tested in palladium-mediated allylation processes (reactions 1 and 2). They
were found to be superior to the first-generation ligands 1 and as effective as the well-known phosphine oxazolines 3.
High throughput screening methods for discovery and
optimization of asymmetric reactions may involve large
libraries of easily accessible materials or smaller collections
focused on lead structures.1 Our group is particularly
interested in the latter approach. Divergent syntheses of
small, but carefully designed, ligand sets are critical for that
objective. Earlier contributions from these laboratories
described syntheses of the phosphine oxazolines 1 and their
performance in the asymmetric allylation reaction 1.2 The
best enantiomeric excess obtained in that work was 93%.
At that time, we speculated that chelate size was important
and that the homologous ligands 2 would give higher
enantioselectivities. This Letter describes a synthesis of a
small ligand set having the generic structure 2 and tests of
our hypothesis.
(1) (a) Weinberg, W. H.; Jandeleit, B.; Self, K.; Turner, H. Curr. Opin.
Solid State Mat. Sci. 1998, 3, 104. (b) Jandeleit, B.; Weinberg, W. H. Chem.
Ind. 1998, 795. (c) Jandeleit, B.; Turner, H. W.; Uno, T.; vanBeek, J. A.
M.; Weinberg, W. H. CATTECH 1998, 2, 101.
(2) Porte, A. M.; Reibenspies, J.; Burgess, K. J. Am. Chem. Soc. 1998,
120, 9180.
(3) (a) Dawson, G. J.; Frost, C. G.; Williams, J. M. J. Tetrahedron Lett.
1993, 34, 3149. (b) Sprinz, J.; Helmchen, G. Tetrahedron Lett. 1993, 34,
1769. (c) Sprinz, J.; Kiefer, M.; Helmchen, G.; Reggelin, M.; Huttner, G.;
Walter, O.; Zsolnai, L. Tetrahedron Lett. 1994, 35, 1523. (d) Matt, P. V.;
Loiseleur, O.; Koch, G.; Pfaltz, A. Tetrahedron 1994, 5, 573.
(4) Ksander, G. M.; Jesus, R. D.; Yuan, A.; Ghai, R. D.; Trapani, A.;
McMartin, C.; Bohacek, R. J. Med. Chem. 1997, 40, 495.
Ligands 2 and the well-known ligands 3 both form six-
membered ring chelates when bound to a metal. Conse-
quently, a goal of this work was to determine if ligands 2
could give enantiomeric excesses that are comparable with
those obtained using ligands 3.3
10.1021/ol991008k CCC: $18.00 © 1999 American Chemical Society
Published on Web 10/24/1999

Scheme 2. One-Step Route to Oxazolines 2
Aspartic acid was the starting material used for synthesis
of ligands 2a-d (Scheme 1). This amino acid was esterified,
BOC-protected, and then reduced via literature procedures4
that can be performed on a large scale without resorting to
chromatographic separations. Conversion of diol 4 to the
desired oxazolidine 5 is also a known procedure.4 Product 5
was contaminated with the undesired oxazine 6, but the
oxazolidine could be isolated by fractional crystallization
from heptane or via flash chromatography. Incorporation of
the phosphine moiety was achieved via tosylate 7, nucleo-
philic displacement, and protection in situ by formation of
phosphine borane 8.2 Compound 8 is the hub of our divergent
ligand synthesis.
Ligands 2b-d (and several others, data not shown) were
readily obtained via acylation/mesylation (Scheme 3). The
procedures outlined in Schemes 1-3 facilitate preparations
of ligands 2 in gram quantities.
Scheme 3. Two-Step Route to Oxazolines 2
Scheme 1. Synthesis of the Key Intermediate 8
Ligands 2 (called JM-PHOS ligands after the major
sponsor for this research) were tested in the allylation reaction
1 using a multiwell reactor as described previously.2 In three
of the four experiments the enantiomeric excess of the
allylation product was greater than 98%, as determined by
HPLC analysis on a Chiralcel OD column (Figure 1a).
Reaction 2 provides a stringent test of ligand performance.
In one isolated example, an enantiomeric excess of 92% was
reported for this reaction,5 but values of less than 74% are
the norm under these conditions.6 The best ee obtained in
this reaction for the first-generation ligands 1 was 75%. Data
collected for the JM-PHOS series (Figure 1b) gave an
optimum (80% ee) that compares favorably with those in
the previous studies.
Different methods were used for conversion of oxazolidine
8 into the required ligands. Phenyl-substituted oxazoline 2a
was prepared by condensation of an imidate with amino
alcohol 9 (Scheme 2).
(5) Trost, B. M.; Krueger, A. C.; Bunt, R. C.; Zambrano, J. J. Am. Chem.
Soc. 1996, 118, 6520.
(6) (a) Matt, P. V.; Pfaltz, A. Angew. Chem., Int. Ed. Engl. 1993, 32,
566. (b) Ahn, K. H.; Cho, C.-W.; Park, J.; Lee, S. Tetrahedron: Asymmetry
1997, 8, 1179.
(7) Shitangkoon, A.; Vigh, G. J. Chromatogr. A 1996, 738, 31.
(8) Helmchen, G.: Kudis, AS.; Sennhenn, P.; Steinhagen, H. Pure Appl.
Chem. 1997, 69, 513.
Ligands 2 give more enantioselective catalysts (Figure 1)
than our first design 1. This supports our previous assertion
that equilibria of the type shown in Figure 2 are more
1746
Org. Lett., Vol. 1, No. 11, 1999

Figure 2. Postulate to explain why superior enantioselectivities
were observed using ligand 2 compared with 1. The presumed
positions of the equilibria are accentuated to highlight the postulated
differences.
complexes could adversely affect the enantiomeric excesses
obtained. We postulate that these diphosphine complexes are
more prevalent for ligands 1 than for 2; hence the latter give
more enantioselective catalysts. Unfortunately, however, we
were unable to measure equilibrium constants for these
situations because the diphosphine complexes could not be
isolated or observed under nonequilibrating conditions.
Factors other than the one outlined in Figure 2 must also
have an impact on asymmetric allylations involving ligands
1 and 2. 31P NMR studies of [(η3-PhCHCHCHPh)Pd-
(ligand)]+ complexes revealed syn:anti ratios in CDCl3 at
25 °C of approximately 3:1 for ligand 1 and 4:1 for 2,
whereas that reported for 3 was 9:1.8 These data provide
insights into stereoelectronic properties of the ligand metal
complexes. Implications of these observations and some
further applications of the JM-PHOS ligands will be reported
in the full account of this work.
Figure 1. Chemical and optical yield data for (a) reaction 1 and
(b) reaction 2. Chemical yields were determined by adding an
internal standard. Enantioselectivity data shown in Figure 1b were
determined via GC. In both reactions 1 and 2, the experiments were
repeated at least two times, and good reproducibility was observed.
Acknowledgment. We thank Drs. Tom Colacot and
Richard Teichman (JM) for helpful discussions. Johnson
Matthey and The Robert Welch Foundation provided finan-
cial support.
significant for ligands 1. Indeed, a crystallographic analysis
of complex 11 (PF6 salt) that shows the P-Pd-N angle is
90.5 °, close to the ideal for a square planar complex, and
implying a lack of ring strain in the chelate of ligand 2.
Details of this crystallographic analysis will be reported
elsewhere. The implication of this observation is that
complexes of ligand 2 are less likely to ring-open, giving
diphosphine complexes. Small amounts of the diphosphine
Supporting Information Available: Experimental pro-
cedures for preparation of compounds 2-10. This material
is available free of charge via the Internet at http://pubs.acs.org.
OL991008K
Org. Lett., Vol. 1, No. 11, 1999
1747