Lipase-mediated resolution of 4-TMS-3-butyn-2-ol and use of the mesylate derivatives as a precursor to a highly stereoselective chiral allenylindium reagent

Article abstract of DOI:10.1021/ol016605x

Matrix presented An improved procedure for the resolution of 4-trimethylsilyl-3-butyn-2-ol has been developed. The mesylate derivatives of the resolved alcohols have been found to undergo highly enantio-, regio-, and diastereoselective additions to aldehy

Full text of DOI:10.1021/ol016605x

ORGANIC
LETTERS
2001
Vol. 3, No. 21
3369-3372
Lipase-Mediated Resolution of
4-TMS-3-butyn-2-ol and Use of the
Mesylate Derivatives as a Precursor to a
Highly Stereoselective Chiral
Allenylindium Reagent
,†
James A. Marshall,* Harry R. Chobanian, and Mathew M. Yanik
Department of Chemistry, UniVersity of Virginia, P.O. Box 400319,
CharlottesVille, Virginia 22904
jam5x@Virginia.edu
Received August 17, 2001
ABSTRACT
An improved procedure for the resolution of 4-trimethylsilyl-3-butyn-2-ol has been developed. The mesylate derivatives of the resolved alcohols
have been found to undergo highly enantio-, regio-, and diastereoselective additions to aldehydes, leading to homopropargylic alcohol adducts.
Additions of chiral allenylmetal reagents to aldehydes leading
to enantioenriched syn and anti homopropargylic alcohol
adducts have played a key role in the synthesis of several
polyketide natural products and subunits with potential
medicinal applications.1 One promising variant of the
methodology employs enantioenriched propargylic mesylates
such as 1 as precursors to enantioenriched allenylzinc or
indium reagents.2 These reagents are generated in situ
through “oxidative transmetalation” of transient allenyl-
palladium intermediates (eq 1).3
of these alcohols limits widespread utilization of the meth-
odology. For this reason we have been exploring simple and
cost-effective preparations. A report by Burgess and Jennings
describing the lipase-mediated kinetic resolution of racemic
TMS butynol 5 through enantioselective acetylation seemed
well suited to this goal (Scheme 1).5 However, because of
(1) (a) Zincophorin. Marshall, J. A.; Palovich, M. R. J. Org. Chem. 1998,
63, 3701. (b) Discodermolide. Marshall, J. A.; Johns. B. A. J. Org. Chem.
1998, 63, 7885. (c) Callystatin A. Marshall, J. A.; Fitzgerald, R. A. J. Org.
Chem. 1999, 64, 4477. (d) Aplyronine A. Marshall, J. A.; Johns, B. A. J.
Org. Chem. 2000, 65, 1501. (e) Balifomycin. Marshall, J. A.; Adams, N.
D. Organic Lett. 2000, 2, 2897. (f) Tautomycin. Marshall, J. A.; Yanik, M.
M. J. Org. Chem. 2001, 66, 1373.
(2) (a) Marshall, J. A.; Adams, N. D. J. Org. Chem. 1999, 64, 5201. (b)
Marshall, J. A.; Grant, C. M. J. Org. Chem. 1999, 64, 8214.
(3) Marshall, J. A. Chem. ReV. 1996, 96, 31. (b) Marshall, J. A. Chem.
ReV. 2000, 100, 3163.
(4) Aldrich Chemical Co., Inc. 1001 West St. Paul Avenue, Milwaukee,
WI 53233.
(5) Burgess, K.; Jennings, L. D. J. Am. Chem. Soc. 1991, 113, 6129.
For an application of the lipase resolution methodology to (E)-4-diphenyl-
silyl-3-buten-2-ol and 4-diphenylsilyl-3-butyn-2-ol, see, respectively: Be-
resis, R. T.; Solomon, J. S.; Yang, M. G.; Jain, N. F.; Panek, J. S. Org.
Synth. 1997, 75, 78. Panek, J. S.; Clark, T. D. J. Org. Chem. 1992, 57,
4323.
The alcohol precursors of mesylate (R)-1 and its enanti-
omer are commercially available.4 However, the high cost
† Fax: +01-804-924-7993.
10.1021/ol016605x CCC: $20.00 © 2001 American Chemical Society
Published on Web 09/25/2001

a phenomenon we ascribe to partial racemization of the
allenylpalladium precursor of the chiral allenylzinc reagent.6
Normant and co-workers have prepared racemic R-TMS
allenylzinc bromide reagents and found that they add to chiral
imines with high diastereoselectivity and enantiodiscrimi-
nation (kinetic resolution).7 They have found these reagents
to be configurationally stable below -10 °C.
Scheme 1
It was also of interest to explore additions employing the
allenylindium reagent which is derived from mesylate (R)-9
through oxidative transmetalation of the aforementioned
allenylpalladium intermediate with InI.2b Initial findings with
Pd(dppf)Cl2 as the catalyst precursor for the transmetalation
were not promising. Only a small amount of adduct was
obtained from cyclohexanecarboxaldehyde, even at room
temperature over 12 h. In contrast, use of Pd(OAc)2‚PPh3 at
0 °C as the catalyst precursor for the InI exchange reaction
led to a remarkable improvement, affording anti adducts with
er values of 99:1 or higher and with virtually no trace of
syn adducts (Table 2). This selectivity was observed not only
the high volatility of TMS butynol (S)-5 and its enantiomeric
acetate 6, especially the former, recovery of material after
separation by chromatography was low (27% for (S)-5).
Nonetheless, the high efficiency of the resolution (>97:3 er
for both (S)-5 and 6 encouraged our investigation of the
methodology. To minimize losses of products during initial
solvent removal, we conducted the resolution in pentanes
rather than hexanes as previously reported.5 Also, to cir-
cumvent material losses and the additional expense of
chromatography, we treated the mixture of acetate 6 and
alcohol (S)-5 with succinic anhydride to afford a ca. 1:1
mixture of succinate 7 and unchanged acetate 5 which was
extracted with NaHCO3 to remove the succinate. Once
separated, the esters 6 (ca. 82% yield) and 7 (86% yield)
were converted to alcohols (R)-5 and (S)-5 by treatment with
excess DIBAL-H in hexanes followed by distillation. The
er values of the two alcohols were determined by GC analysis
of the acetate derivatives on a â-Dex column.
Table 2. Silylaed Allenyulindium Additions to Achiral
Aldehydes
Rather than desilylate alcohols (R)- and (S)-5 to the volatile
and water-soluble parent 3-butyn-2-ols, we explored the use
of the mesylate derivative (S)-9 for the synthesis of the anti
homopropargylic alcohols 10a-c through application of the
allenylzinc methodology (Table 1).2a We were particularly
interested in the regio- and enantioselectivity of these
reactions. In fact, the additions proceeded with excellent
diastereoselectivity and only slight loss of enantioselectivity,
a GC analysis on a â-Dex column after cleavage of the silyl group with
TBAF. b GC analysis on an R-Dex column after cleavage of the silyl group
with TBAF.
with branched aldehydes 8a and 8b but also with unbranched
aldehydes 8c and 8d, as well as the conjugated enal 8e. By
comparison, additions of allenylzinc and indium reagents
derived from 3-butyn-2-ol mesylate or the 4-acetoxymethyl
derivative to unbranched and conjugated aldehydes typically
afford anti adducts with 70-85% ds.2,3 An addition of the
TMS reagent to 2-octynal was less satisfactory. A 2:1 mixture
of anti and syn diastereomers was obtained in 45% yield
along with a chromatographically separable allenylcarbinol
isomer in 30% yield. An investigationof this intriguing
discrepancy is under investigation. Close inspection of the
1H NMR spectra of adducts 10a-e revealed signals for traces
(5% or less) of comparable adducts.
Table 1. Silylated Allenylzinc Additions to Achiral Aldehydes
(6) Chiral allenylpalladium compounds are racemized by Pd(0) catalysts.
Elsevier, C. J.; Vermeer, P. J. Org. Chem. 1985, 50, 3042. See also:
Marshall, J. A.; Wolf, M. A.; Wallace, E. M. J. Org. Chem. 1997, 62, 367.
(7) Poisson, J.-F.; Chemla, F.; Normant, J.-F. Synlett 2001, 305.
a GC analysis on a â-Dex column after cleavage of the silyl group with
TBAF. b Corrected for the er of the resolved alcohol (S)-5 (97:3).
3370
Org. Lett., Vol. 3, No. 21, 2001

We suspected that the allenyl TMS substituent was
responsible for the remarkable selectivity of the allenylindium
reagent (2-octynal notwithstanding) but could not exclude
the possibility that the Pd(OAc)2‚PPh3 catalyst precursor was
a contributing factor. We therefore examined additions of
allenylindium reagents derived from mesylate (R)-1 to
representative aldehydes with this catalyst precursor (Table
3). In fact, the product ratios were virtually identical to those
previously observed with the Pd(dppf) catalyst.2b
Table 3. Allenylindium Additions to Achiral Aldehydes
“Hoffmann test” 9 with aldehyde (S)-12 and racemic mesylate
(R,S)-9 (eq 4). In fact, a 50:47:3 mixture of diastereomeric
adducts 14:13a:13b was produced, indicating that no race-
mization of the reagent had taken place.
Additions of the (R)-9-derived reagent to the R-methyl
â-OTBS aldehyde (R)-15 afforded the anti,anti adduct 16
as the sole detectable product. Additions of the enantiomeric
reagent (S)-9 were equally selective, affording adduct 17
exclusively after correcting for the er (97:3) of the allenyl-
indium precursor (eqs 5 and 6).
a GC analysis on a â-Dex column. b Corrected for the er of the resolved
alcohol (R)-5 (97:3). c GC analysis on an R-Dex column.
Thus we surmise that the allenyl R-TMS substituent must
impose a significant directing effect in these additions.
Furthermore, the amount of racemization is lower with the
Pd(OAc)2‚PPh3 catalyst system than that of the Pd(dppf)
catalyst (<1% vs ∼5% with cyclohexanecarboxaldehyde)
previously studied.2c These results suggest that the former
catalyst decreases the rate of allenylpalladium racemization7
and/or increases the rate of Pd-In exchange to yield a
configurationally stable allenylindium reagent. This latter
postulate would be consistent with the observed rate differ-
ence between additions performed with PdCl2(dppf) vs
Pd(OAc)2‚PPh3 as the transmetalation catalysts.
With possible applications to natural product synthesis in
mind, we examined additions of allenylindium reagents
derived from (R)- and (S)-9 to R-oxygenated and R-methyl
aldehydes (eqs 2 and 3). We have previously found that
nonsilylated allenylindium reagents show significant match-
ing/mismatching with the former but not the latter aldehydes.8
In fact, addition of the (R)-9-derived reagent to the TBS ether
(S)-12 of (S)-lactic aldehyde afforded a 90:10 mixture of the
anti adduct 13 and the corresponding syn,syn diastereomer
3b (not shown). This compares to an 80:20 mixture of the
corresponding adducts from the mesylate of 3-butyn-2-ol.8
As expected, the (S)-9-derived reagent gave only the anti,-
anti adduct 14 with no detectable syn isomer. In both
additions, a trace (5%) of the allenylcarbinol byproduct could
be seen in the 1H NMR spectrum. To test the configurational
stability of the allenylindium reagent, we performed a
The enhanced diastereoselectivity of the R-TMS-substi-
tuted allenylindium reagents, particularly with regard to the
remarkably high levels observed with unbranched and
conjugated aldehydes, is suggestive of a tight transition state
for these additions. Such a transition state would place the
CH3 and the nearly eclipsed aldehyde substituent R in close
proximity for the syn pathway, whereas these groups are at
some distance in the anti array (Figure 1).10 The precise role
(steric or electronic) of the TMS substituent in effecting this
perturbation is unclear at present.
The present findings reveal an unprecedented and unex-
pected TMS effect for enhancing the diastereoselectivity of
(9) Hoffman, R. W.; Lanz, J.; Hoppe, D.; Metternich, R.; Tarrana, G.
Angew. Chem., Int. Ed. Engl. 1987, 26, 1145.
(10) Support for the eclipsed transition states has been obtained for the
analogous allenylzinc additions through ab initio calculations at the 3-21G*
basis set level. Details of these calculations will be described in a
forthcoming publication.
(8) Marshall, J. A.; Chobanian, H. R. J. Org. Chem. 2000, 65, 8357.
Org. Lett., Vol. 3, No. 21, 2001
3371

Acknowledgment. This research was supported by
research grants from the National Science Foundation (CHE-
9901319) and the National Institutes of Health (R01
CA090383). We thank Professor James Panek for advice and
encouragement in connection with the lipase resolution
methodology.
Supporting Information Available: Experimental pro-
cedures for all new compounds and selected 1H NMR
spectra. This material is available free of charge from the
Internet at http://pubs.acs.org.
Figure 1. Transition states for additions of allenylindium reagents
to aldehydes.
allenylindium additions to aldehydes. The finding that
unbranched and conjugated aldehydes respond to this effect
considerably broadens the synthetic scope of the methodol-
ogy. The superiority of the Pd(OAc)2‚PPh3 catalyst to the
previously employed Pd(dppf)Cl2 precursor is also of
considerable practical import. An improved, efficient, and
inexpensive resolution of the 4-TMS-3-butyn-2-ol precursor
of the allenylindium reagents has also been developed.
Finally, it should be noted that the alkynyl TMS adducts
can be readily desilylated to synthetically useful terminal
alkynes1 with TBAF or used directly in CuCl2-promoted
coupling reactions with vinyl and aryl halides, thus further
enhancing the utility of the present discovery.11 Additional
seamless methodologies for directly integrating the TMS
alkynyl moiety into useful synthetic sequences are currently
under investigation.12
OL016605X
(9.1 mL, 65.0 mmol), DMAP (92 mg, 0.75 mmol), and succinic anhydride
(4.15 g, 41.1 mmol) successively. The mixture was heated to reflux for 4
h, cooled, and quenched with 30 mL of NaHCO3. The solution was stirred
vigorously for 1 h and diluted with Et2O. The Et2O solution was separated
and washed with 10% HCl and brine. The Et2O solution was dried over
MgSO4, filtered, and concentrated by rotary evaporation. The resulting oil
was purified by bulb-to-bulb distillation (50 °C at 0.5 mmHg) to yield 5.31
g (82%) of acetate 6 as a clear oil: [R]20 +117.5 (c ) 2.30, CHCl3); H
1
D
NMR (300 MHz, CDCl3) δ 0.15 (s, 9H), 1.45 (d, J ) 7.2 Hz, 3H), 2.08 (s,
3H), 5.42 (q, J ) 7.2 Hz, 1H). Analysis by GC on a â-Dex column showed
a single peak. The aqueous phase was carefully acidified with 12.0 M HCl
to pH ∼1.0 and extracted with EtOAc. The extracts were combined, dried
over MgSO4, filtered, and concentrated by rotary evaporation, affording
7.35 g (86%) of the succinate 7 as a light yellow oil which solidified upon
cooling to 0 °C. 1H NMR (300 MHz, CDCl3): δ 0.15 (s, 9H), 1.45 (d, J )
7.2 Hz, 3H), 2.62 (m, 4H), 5.47 (q, J ) 7.2 Hz, 1H). This acid was carried
on without further purification. To a solution of acetate 6 (4.22 g) in 30
mL of hexanes at -78 °C was added 35 mL of DIBAL-H (1.0 M in
hexanes). The solution was stirred for 10 min and poured into a rapidly
stirred mixture of 300 mL of aqueous Rochelle’s salt and 200 mL of Et2O.
Once the Et2O layer clarified, it was separated, dried over MgSO4, filtered,
and concentrated at atmospheric pressure to yield an oil. The oil was purified
by bulb-to-bulb distillation (65 °C at 0.5 mmHg) to yield 2.70 g (83%) of
(11) Cf. Nishihara, Y.; Ikegashira, K.; Hirabayashi, K. Ando, J.-i.; Mori,
A.; Hiyama, T. J. Org. Chem. 2000, 65, 1780. Results from our laboratory
will be described in a forthcoming publication.
alcohol (R)-5: [R]20 +23.8 (c ) 2.02, CHCl3); IR (film) υ 3335, 2173
D
1
cm -1; H NMR (300 MHz, CDCl3) δ 0.15 (s, 9H), 1.44 (d, J ) 6.6 Hz,
(12) Resolution of 4-TMS-3-butyn-2-ol. To a solution of racemic
4-trimethylsilyl-3-butyn-2-ol (10.0 g, 69.9 mmol) in pentane (250 mL) were
added Amano AK Lipase (2.0 g, Aldrich), freshly distilled vinyl acetate
(50 mL), and 1 g of pulverized, activated 4-Å molecular sieves. The reaction
progress was monitored by GC (Carbowax, 110 °C, 1 °C ramp/min; acetate,
5.75 min; alcohol, 8.22 min). After 72 h, the ratio of acetate to alcohol was
48:52. The mixture was filtered through a medium scintered glass funnel,
washed with additional pentane, and concentrated via rotary evaporation,
affording 11.5 g of a nearly 1:1 mixture of alcohol and acetate by 1H NMR
analysis. To this mixture of 6 and (S)-5 in THF (50 mL) were added Et3N
3H), 4.51 (q, J ) 6.6 Hz, 1H); 13C NMR (75 MHz, CDCl3) δ 107.6, 88.4,
58.7, 24.2, -0.96. To a solution of (S)-succinate 7 (2.92 g) in 30 mL of
CH2Cl2 at -78 °C was added 28 mL of DIBAL-H (1.0 M in hexanes). The
solution was stirred for 10 min, and the product was isolated as described
above. The resulting oil was purified by bulb-to-bulb distillation (65 °C/
0.5 mmHg) to yield 1.49 g (87%) of alcohol (S)-5: [R]20 -22.6 (c )
D
2.29, CHCl3); lit. (-25.9, c ) 3.12).5 A portion of the alcohol was acetylated
with excess acetic anhydride, NEt3, and DMAP (catalytic) in CH2Cl2. GC
analysis of the crude acetate on a â-Dex column indicated a 97:3
enantiomeric ratio.
3372
Org. Lett., Vol. 3, No. 21, 2001