Asymmetric molybdenum-catalyzed alkylations [18]

Full text of DOI:10.1021/ja973298a

1104
J. Am. Chem. Soc. 1998, 120, 1104-1105
Asymmetric Molybdenum-Catalyzed Alkylations
Initial experiments focused on the use of tungsten with chiral
ligands 1, 2, and 3. The catalyst was generated by stirring a
8
9
10
Barry M. Trost* and Iwao Hachiya
Department of Chemistry, Stanford UniVersity
Stanford, California 94305-5080
ReceiVed September 19, 1997
Interest in molybdenum-1,2 and tungsten-catalyzed reactions
of allyl substrates with nucleophiles stemmed from the issue of
regioselectivity. Most notably, with aryl-substituted allyl systems,
palladium-catalyzed reactions normally provide products from
attack at the less substituted terminus, even with chiral ligands
3
(
eq 1, path a).4 On the other hand, molybdenum and tungsten
2 5 3 3
1:1.5 mixture of (C H CN) W(CO) :ligand in THF at 60 °C. The
test reaction employed the cinnamate 4 (Ar ) Ph) with dimethyl
catalysts favor attack at the more substituted terminus (eq 1, path
b). The value of products of the latter type as building blocks
makes a highly regio- and enantioselective method desirable. A
recent report utilizing tungsten with phosphinooxazoline ligands
with only dimethyl malonate as the pronucleophile proved
promising, whereas the isostructural molybdenum complex was
described as “not useful as a catalyst”.5 Early studies in our
laboratories utilizing a variety of chiral nitrogen-based ligands
for molybdenum failed to give any appreciable asymmetric
induction.6 In both cases, ligation of the octahedral complexes
involves structures such as I. Creating helical-like chiral
sodiomalonate (5a) in THF at reflux. With ligands 1a, 1b, and
2, the reaction either failed or generated only trace amounts of
products. These results mirror our earlier results which suggested
that phosphines deactivated the tungsten catalyst compared to
pyridines.3 On the other hand, the Pfaltz5 ligand system involves
both a phosphine and an imine. The first sign of some success
came in the use of the nitrogen ligand 3 whereby, with 5 mol %
3
of catalyst, a low yield of a 19:1 ratio of 6 :7 (Ar ) Ph, R ) H),
where 6 had an enantiomeric excess (ee) of 98% was observed.
Increasing the catalyst to 15 mol % increased the yield to 55%
and the 6:7 ratio to 49:1 with 6 still having 98% ee.
,7
In contrast to the results of Pfaltz, switching to the molybdenum
system proved better. As summarized in Table 1 (entry 1),
repeating the above but replacing the tungsten complex by (C
CN) Mo(CO) gave an 88% yield of a 97:3 ratio of 6:7 (Ar )
2 5
H -
3
3
Ph, R ) H) with 6 having an ee of 99%. Lowering the
temperature to room temperature (entry 2) still provided a good
yield and somewhat improved regioselectivity while maintaining
a high ee.
A key question relates to the mechanism of the asymmetric
induction. Since 4 is achiral, the catalyst may differentiate the
enantiotopic faces leading preferentially to π-allylmolybdenum
complex 9 or 10 and then onto product. Alternatively, the two
diastereomeric π-allylmolybdenum complexes 9 and 10 may be
in dynamic equilibrium (eq 3) in which the enantiodifferentiation
complexes represented by II may create a more effective chiral
environment. An alternative might be to bridge in a trans fashion
as in III. We report our recent preliminary observations designed
toward complexes of these latter types that have led to reactions
according to path b of eq 1 with high regioselectivity as well as
enantioselectivity.
(
1) Trost, B. M.; Lautens, M. J. Am. Chem. Soc. 1982, 104, 5543; 1987,
09, 1469. Trost, B. M.; Lautens, M. Tetrahedron 1987, 43, 4817 and
references therein. Trost, B. M.; Merlic, C. A. J. Am. Chem. Soc. 1990, 112,
590.
2) For early work on stoichiometric π-allylmolybdenum alkylations, see:
1
9
(
Adams, R. D.; Chodosh, D. F.; Faller, J. W.; Rosan, A. M. J. Am. Chem. Soc.
occurs by preferential nucleophilic attack on 9 or 10. Starting
from the chiral substrate 8 (Ar ) Ph) differentiates these two
possibilities. If the first scheme operates, either a kinetic
resolution or racemic product would be observed. In the latter
instance, we should see results similar to those obtained with the
achiral allylic substrate 4. As shown in entries 3 and 4, the results
starting with 8 (Ar ) Ph) mostly mirror those starting from 4
(Ar ) Ph) (entries 1 and 2). While the observed small differences
1
979, 101, 2570. McCleverty, J. A.; Murray, A. J. J. Organomet. Chem. 1978,
1
49, C29. For more recent work, see: Rubio, A.; Liebeskind, L. S. J. Am.
Chem. Soc. 1993, 115, 891.
(3) Trost, B. M.; Hung, M.-H. J. Am. Chem. Soc. 1983, 105, 7757. Trost,
B. M.; Tometzki, G. B.; Hung, M.-H. J. Am. Chem. Soc. 1987, 109, 2176.
(4) Godleski, S. A. In ComprehensiVe Organic Synthesis; Trost, B. M.,
Fleming, I., Semmelhack, M. F., Eds.; Pergamon Press: Oxford, U.K., 1991;
Vol. 4, pp 585-662. (b) However, for a recent interesting exception with Pd
catalysis, see: Hayashi, T.; Kawatsura, M.; Uozumi, Y. Chem. Commun. 1997,
5
61.
(
5) Lloyd-Jones, G. C.; Pfaltz, A. Angew. Chem., Int. Ed. Engl. 1995, 34,
4
62.
(8) Trost, B. M.; Van Vranken, D. L. Angew. Chem., Int. Ed. Engl. 1992,
31, 228. Trost, B. M.; Van Vranken, D. L.; Bingel, C. J. Am. Chem. Soc.
1992, 114, 9327.
(9) Richter, W. Unpublished work in these laboratories.
(10) Chapman, R. L.; Vagg, R. S.; Watton, E. C.; Barnes, D. J. J. Chem.
Eng. Data 1978, 23, 349. Also see: Adolfsson, H.; Moberg, C. Tetrahedron:
Asymmetry 1995, 6, 2023.
(
(
6) Merlic, C. A. Ph.D. Thesis, University of Wisconsin, 1988.
7) For asymmetric alkylations with enantiomerically pure stoichiometric
complexes, see: Faller, J. W.; Chao, K.-H. J. Am. Chem. Soc. 1983, 105,
3
893. Faller, J. W.; Chao, K.-H. Organometallics 1984, 3, 927. For an
overview, see: Faller, J. W.; Mazzieri, M. R.; Nguyen, J. T.; Parr, J.;
Tokunaga, M. Pure Appl. Chem. 1994, 66, 1463.
S0002-7863(97)03298-8 CCC: $15.00 © 1998 American Chemical Society
Published on Web 01/23/1998

Communications to the Editor
J. Am. Chem. Soc., Vol. 120, No. 5, 1998 1105
Table 1. Mo-Catalyzed Asymmetric Allylic Alkylationsa
entry
4 or 8, Ar
4, Ph
4, Ph
8a, Ph
8a, Ph
8a, 2-thienyl
8a, 2-pyridyl
8a, 1-naphthyl
4, Ph
4, 2-furyl
8b, 2-furyl
8b, 2-furyl
8a, 2-pyridyl
8a, 2-thienyl
8b, 2-furyl
5 R
T, °C
time, h
% yieldb
6:7
% eec,d of 6
1
2
3
4
5
6
7
8
9
1
1
1
1
1
H
H
H
H
H
H
H
CH
CH
CH
CH
CH
CH
CH
reflux
rt
reflux
rt
reflux
reflux
reflux
reflux
reflux
reflux
rt
reflux
reflux
rt
3
3
3
3
2
2
2
4
2
2
18
2
2
12
88
70 (90)
70
61 (68)
78
69 (82)
82
32:1
49:1
13:1
32:1
19:1
8:1
99:1
24:1
32:1
32:1
99:1
5:1
99
99
92
97
88
96
87
d
3
3
3
3
3
3
2
67
71
65
54
71
71
50
98
97
87
95
94
75
0
1
2
3
4
13:1
99:1
e
CHdCH
2
98
a
2 5 3 3
All reactions were performed with 10 mol % (C H CN) Mo(CO)
and 15 mol % 3 in THF at 0.1 M. b Isolated yields. Yields in parentheses
c
are based upon recovered starting material. Determined by chiral HPLC using Daicel Chiralcel OD and Chiralpak AD eluting with heptane-2-
propanol. The major isomer assigned as S by comparison to the literature for 6 (Ar ) Ph, R ) H) and 6 (Ar ) 1-naphthyl, R ) H) and the rest
d
by analogy. Because of priorities, the same absolute configuration of the product derived from the substituted malonate with Ar ) Ph corresponds
e
to the R configuration. The ee was determined on the subsequent transformation product.
suggest that both paths may be occurring to some extent, the major
path involves nucleophilic attack on equilibrating π-allylmolyb-
denum complexes. Thus, either allylic regioisomeric starting
material can be used with a preference for the achiral isomer.
Having established the ability to generate 6 (Ar ) Ph) with
excellent regio- and enantioselectivity, we examined variation of
reactions with dimethyl methylmalonate and cinnamyl substrates
normally led to attack at the less substituted allyl terminus even
when catalyzed with molybdenum complexes.1 With chiral ligand
3, equally good selectivity for attack at the more substituted
terminus occurs for both dimethyl malonate and its substituted
analogues 5b,c. The reactivity should be noted compared to other
molybdenum catalysts bearing nitrogen ligands according to
11a
the aromatic ring. An electron-rich thiophene ring (entry 5)
and an electron-deficient pyridine ring (entry 6)11a gave good
selectivities. A bulkier naphthalene substrate (entry 7)1 still gave
good results.
structure I. These reactions typically require heating at reflux
1b
1,6
for 24 h or more.
In contrast, the reactions reported herein
typically proceed in 2-3 h at reflux and even room temperature
suffices. We suggest that these results are more in accord with
complexes of type III than of type II. Furthermore, molecular
modeling indicates that complexes such as III are strongly
preferred over I and II for ligand 3.15 In the case of type III
complexes, the bridging ligand would promote ligand dissociation
and would be coordinatively unsaturated. As a result, rates of
ligand exchange (i.e., rates of substrates and products going on
and off molybdenum) should be faster. In addition, a coordina-
tively unsaturated π-allylmetal complex should exhibit higher
reactivity toward nucleophiles. Clearly, any further discussion
should await structural characterization of the active complex.
Trying to understand the sense of the stereochemical induction
becomes very difficult here since it is not clear that this
molybdenum reaction proceeds by a double-inversion or double-
retention mechanism. With achiral molybdenum complexes, some
evidence supports the notion that it proceeds via a double-retention
Increasing the steric bulk of the nucleophile by using 5b still
gave excellent results with both carbocyclic and heterocyclic
substrates (entries 8,1 9, 10, 11, and 12 ). Only in the
case of the thiophene substrate did we see a deterioration of the
ee to 75% for the product12 (entry 13), although no attempt has
been made to optimize this reaction. The regioselectivity was
still good. Increasing the steric bulk further by using the
allylmalonate nucleophile 5c did slow the reaction (entry 14),13
but the regio- and enantioselectivities were still excellent. In this
case, heating the alkylation product at 80 °C in 5:2 water:ethanol
gave the diastereomeric Diels-Alder adducts 11 and 12 in a 3:1
ratio, each of which had an ee of 98% (eq 4) as determined by
chiral HPLC. For the furan substrate bearing the leaving group
at the secondary carbon, the acetate 8b rather than carbonate 8a
ester was employed.
1a
11a
11a
11a
12
pathway rather than the more common double-inversion path-
way.1
4,16
At this stage, this catalytic system represents an excellent
approach for regio- and enantioselective alkylations at the more
substituted terminus of cinnamyl-like systems. It also suggests
that molybdenum-catalyzed processes at least complement pal-
ladium-catalyzed processes. Considering the lower cost of molyb-
denum compared to palladium, the prospects that it may be more
generally useful will also be an important direction to pursue.
This asymmetric alkylation provides the first example of a
practical catalytic system for molybdenum.14 The remarkable
insensitivity of the high selectivity over the temperature range
Acknowledgment. We thank the National Science Foundation and
the National Institutes of Health, General Medical Sciences, for their
generous support of our programs and JSPS for a postdoctoral fellowship
for I.H. Mass spectra were obtained from the Mass Spectrometry Facility,
University of San Francisco, supported by the NIH Division of Research
Resources.
20-65 °C may suggest a fairly rigid chiral active site. While
we have focused on reactions that cannot be achieved efficiently
with palladium,4b this new catalytic system may be more generally
useful. The regioselectivity observed here for attack at the more
substituted terminus is generally significantly higher than with
our earlier achiral molybdenum catalysts. For example, previous
Supporting Information Available: General experimental procedures
and characterization data for 6 (Ar ) Ph, 2-furyl, 2-thienyl, 2-pyridyl,
(
11) (a) For the products of these reactions, see ref 3. (b) For the products
of these reactions, see ref 5.
12) All new products have been fully characterized spectroscopically and
the elemental compositions established by combustion analysis.
13) Trost, B. M.; Lautens, M.; Hung, M.-H.; Carmichael, C. S. J. Am.
Chem. Soc. 1984, 106, 7641.
3 2 2
and 1-naphthyl, R ) H, CH , or CH CHdCH ) (6 pages). See any current
(
masthead page for ordering and Web access instructions.
JA973298A
(
(14) In a footnote, asymmetric induction in a catalytic process involving a
(15) A CaChe molecular modeling system was employed for these
calculations.
(16) Faller, J. W.; Linebarrier, D. Organometallics 1988, 7, 1670.
meso-π-allylmolybdenum fragment is noted. See: Dvor a´ k, D.; Stary, I.;
Kocovsky, P. J. Am. Chem. Soc. 1995, 117, 6130.