γ-butyrolactone synthesis via catalytic asymmetric cyclocarbonylation

Full text of DOI:10.1021/ja005568m

J. Am. Chem. Soc. 2001, 123, 6457-6458
6457
Scheme 1. The Hetero-Pauson-Khand Reaction
γ-Butyrolactone Synthesis via Catalytic Asymmetric
Cyclocarbonylation
Sunil K. Mandal, Sk. Rasidul Amin,1 and William E. Crowe*
Department of Chemistry, Louisiana State UniVersity
Baton Rouge, Louisiana 70803
Scheme 2
ReceiVed September 1, 2000
ReVised Manuscript ReceiVed May 17, 2001
The butyrolactone ring is an integral building block of many
natural products.2 An atom economic route toward the formation
of this important skeleton is a formal [2+2+1] cycloaddition of
an alkene, a carbonyl, and carbon monoxide. The first example
of this Hetero-Pauson-Khand reaction was reported from our
laboratory where δ,ꢀ-unsaturated ketones and aldehydes were
converted to fused, bicyclic γ-butyrolactone products via the
cyclocarbonylation reaction shown in Scheme 1.3 Although
Buchwald4 has recently reported a catalytic version of this
transformation using Cp2Ti(PMe3)2 or Cp2Ti(CO)2, this method
is limited to aryl ketone substrates (Scheme 1). Development of
a general catalytic protocol and an asymmetric version of this
transformation would be a significant improvement. Herein we
report a general catalytic cyclocarbonylation of enals and enones
using a chiral titanocene catalyst that also affords the first example
of the asymmetric version of this reaction.5
Cp2Ti(CO)2 (formed under catalytic conditions) does not react
with most δ,ꢀ-unsaturated ketone and aldehyde substrates, render-
ing cyclocarbonylation reactions stoichiometric. Ansa-metallocene
complexes often exhibit markedly different reactivity than their
unbridged counterparts. In particular, Brintzinger has shown that
ansa-titanocene dicarbonyl complexes are substantially less stable
than Cp2Ti(CO)2.6 We therefore reasoned that the ansa-metal-
locene (EBTHI)Ti(CO)2 (1, Scheme 2) might be more reactive
than Cp2Ti(CO)2 toward cyclocarbonylation substrates and tested
the utility of 1 as a catalyst. An attractive feature of catalyst 1 is
that it can be generated in situ from the air stable precursor
(EBTHI)TiMe2 (Scheme 2). In analogy to the Petasis olefination
reagent,7 Cp2TiMe2, (EBTHI)TiMe2 can be easily handled as a
solid on the benchtopswithout making special provisions to
exclude oxygen or moisture.
γ-butyrolactones in very good to excellent yield without any side
reaction (eq 1, Table 1).
Cyclocarbonylation of trifluoromethyl ketone 211 or tert-butyl
ketone 3 was not successful; no significant lactone formation was
observed even after 4 days under standard reaction conditions.
The attempted cyclocarbonylation of δ,ꢀ-unsaturated aryl ketone
4 also did not afford the desired lactone product. Instead this
substrate underwent a vinylogous pinacol coupling reaction
(stoichiometric in titanium) to produce a highly substituted
cyclopentanol derivative 5.12
Diastereofacial selectivity was investigated using racemic 1 to
cyclocarbonylate substrates possessing a preexisting stereocenter.
A mixture of diastereomeric lactones was obtained for substrates
possessing a â-carbon stereocenter (entries 1-4), and essentially
perfect diastereofacial selectivity was observed for substrates
possessing a stereocenter at the R- or γ-carbon (entries 5 and 6).
Observed diastereoselectivities are markedly higher than those
obtained in corresponding room temperature, stoichiometric
reactions reported previously.3 This may be attributable to the
elevated reaction temperature used in the catalytic reactions.13
In a typical experiment a mixture of substrate (see Table 1)
and catalyst 18 in toluene was heated at 100 °C, under CO
pressure,9 in the presence of excess PMe3.10,17 The catalyst system
worked well both for enal and enone substrates forming fused
(1) Present address: Albany Molecular Research, Inc.: Albany, NY 12203.
(2) Hoffmann, H. M. R.; Rabe, J. Angew. Chem., Int. Ed. Engl. 1985, 24,
94-110.
(3) Crowe, W. E.; Vu, A. T. J. Am. Chem. Soc. 1996, 118, 1557-1558.
(4) (a) Kablaoui, N. M.; Hicks, F. A.; Buchwald, S. L. J. Am. Chem. Soc.
1996, 118, 5818-5819. (b) Kablaoui, N. M.; Hicks, F. A.; Buchwald, S. L.
J. Am. Chem. Soc. 1997, 119, 4424-4431.
Preliminary studies using enantiomerically pure (r,r) or (s,s)
catalyst (Scheme 2) revealed a substrate dependent enantioselec-
(5) Some preliminary experiments were run at Emory University, Atlanta,
GA 30322.
(6) Smith, J. A.; Brintzinger, H. H. J. Organomet. Chem. 1981, 218, 159-
167.
(11) Trifluoromethyl ketone on treatment with a stoichiometric amount of
catalyst under similar conditions resulted in a complex mixture of products
(TLC). In this case the highly electrophilic carbonyl carbon of the substrate
probably participated in the reaction with the cyclopentadienyl ring of
titanocene. For a precedence of such a reaction see: Gleiter, R.; Wittwer, W.
Chem. Ber. 1994, 127, 1797.
(12) Unsaturated aryl ketones were reported to react with Cp2Ti(CO)2 to
generate highly substituted cyclopentanol in good yield. See: Schobert, R.;
Maaref, F.; Durr, S. Synlett 1995, 83.
(13) Buchwald (ref 4) reported high diastereoselectivity for cyclocar-
bonylation reactions run at 70 °C and attributed the selectivity enhancement
to reversible metallacycle formation allowing equilibration of thermodynami-
cally favored isomers.
(7) Petasis, N. A.; Bzowej, E. I. J. Am. Chem. Soc. 1990, 112, 6392.
(8) (EBTHI)Ti(CO)2 was generated in situ from (EBTHI)TiMe2 following
the procedure reported in: Hicks, F. A.; Buchwald, S. L. J. Am. Chem. Soc.
1996, 118, 11688-11689. For a similar synthesis of (CH2)2(C5H4)2Ti(CO)2,
see: Smith, J. A.; Brintzinger, H, H. J. Organomet. Chem. 1981, 218, 159.
(9) CO pressure of the reaction vessel was set to 50 psi at the inception of
the reaction and generally increased to 60-65 psi during reaction at 100 °C.
(10) 100% conversion occurred in the presence of excess PMe3, whereas
without excess PMe3 only 50-60% conversion was observed. For a similar
effect of PMe3 on related reaction see ref 3b. Addition of excess PMe3 was
also found to decrease the amount of catalyst (from 20 mol % to 10 mol %)
necessary for complete conversion.
10.1021/ja005568m CCC: $20.00 © 2001 American Chemical Society
Published on Web 06/09/2001

6458 J. Am. Chem. Soc., Vol. 123, No. 26, 2001
Table 1. Catalytic Cyclocarbonylation Reactions
Communications to the Editor
Figure 1. Potential stereocontrol elements.
three reacted with high enantioselectivity (89-90% ee), three
reacted with moderate enantioselectivity (38-60% ee), and two
reacted with no enantioselectivity (0% ee).
Metallacycle intermediates corresponding to major reaction
products are shown in Figure 1 where the conformations of the
fused cyclopentane and oxatitanacycle rings are based on our
previously reported crystal structure.3 For diastereoselective
reactions of chiral substrates (substrate control, structure A), the
major diastereomer formed corresponds to the metallacycle
intermediate where the R-, â-, or γ-carbon substituent adopts a
pseudoequatorial alignment (RR, Râ, or Rγ * H). For enantiose-
lective reactions of achiral substrates, the major enantiomer
formed corresponds to metallacycle B (in the six reactions which
proceed selectively). The most selective reactions (entries 10-
12) correspond to substrates which place methyl groups in one
or both of the positions (R and/or R′ ) Me) which would seem
to maximize the steric difference between the diastereomeric
metallacycles B (favored) and C (disfavored). Results obtained
with other substrates (entries 9, 13, 14) show, however, that
enantioselectivity is also influenced by factors not evident from
inspection of structures B and C.
In conclusion, we have developed a general catalytic protocol
for the intramolecular Hetero-Pauson-Khand cyclization. Notable
features of this new catalytic reaction are the operational simplicity
of the procedure (air stable catalyst precursor and mild reaction
conditions) and the remarkable effect of the ansa bridge in
promoting catalytic activity. Although the use of ansa-bridging
to specifically tune the steric environment of metallocene
complexes has been widely utilized, other reactivity changes
promoted by ansa bridging have received far less attention.15 The
ansa effect responsible for the enhanced reactivity of 1 is a
previously unexploited design feature for metallocene catalyzed
reactions; we are currently attempting to determine its precise
nature.16
Acknowledgment. Support of this work by funds provided by the
National Institutes of Health (R01 GM53172) and Louisiana State
University is gratefully acknowledged. We thank Drs. Dale Treleaven
and Frank Zhou for assistance with 2D NMR experiments.
Supporting Information Available: Experimental procedures and
spectral data for all compounds (PDF). This material is available free of
charge via the Internet at http://pubs.acs.org.
a Reaction conditions were as follows (see eq 1): 50 psig CO, 100
°C, 36-40 h, toluene. Unless otherwise indicated, all reactions were
run with 10 mol % 1 (generated in situ) and 30 mol % PMe3. b Racemic
1 was used. c Ratio refers to the mixture of epimers at C2, C3, or C4.
All products were exclusively cis-fused within the limits of detection.
d 50 mol % PMe3 was used. e 20 mol % 1 and 80 mol % PMe3 were
used. f Only one diastereomer detected by 1H NMR. g Product config-
uration depicted is for the reaction run with (S,S)-1. h Reactions run
with the (R,R)-1 catalyst gave a product configuration enantiomeric to
that depicted in the table for the (S,S)-1 catalyst.
JA005568M
(15) (a) Churchill, D.; Shin, J. H.; Hascall, T.; Hahn, J. M.; Bridgewater,
B. M.; Parkin, G. Organometallics 1999, 18, 2403. (b) Lee, H.; Desrosiers,
P. J.; Guzei, I.; Rheingold, A. L.; Parkin, G. J. Am. Chem. Soc. 1998, 120,
3255. (c) Lee, H.; Bonanno, J. B.; Hascall, T.; Cordaro, J.; Hahn, J. M.; Parkin,
G. J. Chem. Soc., Dalton Trans. 1999, 1365. (d) Shin, J. H.; Parkin, G. Chem.
Commun. 1999, 887.
(16) We believe that the ansa bridge of the (EBTHI)Ti(CO)2 complex
provides a driving force for the reductive cyclization in the form of ansa strain
which is released upon formation of the metallacycle intermediate.
(17) Experimental procedure: A dry Fisher-Porter bottle was charged with
substrate, (EBTHI)TiMe2, PMe3 and toluene, then attached to a Schlenk line,
evacuated, and backfilled with 50 psig of CO. The reaction was heated at
100 °C for 36-40 h, then cooled to room temperature and CO pressure
cautiously released inside a fume hood. The reaction mixture was filtered
(silica gel), washed (Et2O), concentrated under reduced pressure, and purified
by flash column chromatography.
tivity as summarized in Table 1 (entries 7-14).14 Of the eight
substrates examined, four aldehydes and four methyl ketones,
(14) Product absolute stereochemistry was assigned using the Pirkle chiral
solvating agent (S)-(+)-2,2,2-trifluoro-1-(9-anthryl)ethanol to induce chemical
shift inequivalence in the R proton signals of enantiomeric lactone products:
Pirkle, W. H.; Sikkenga, D. L.; Pavlin, M. S. J. Org. Chem. 1977, 42, 384.