Diastereoselective palladium-catalyzed formate reduction of allylic carbonates en route to polypropionate systems

Article abstract of DOI:10.1021/jo052267s

Diastereoselective palladium-catalyzed formate reduction of allylic carbonates presents unique opportunities for applications in target-oriented organic synthesis provided that selectivity, in particular stereoselectivity, in the course of this metal-cata

Full text of DOI:10.1021/jo052267s

Diastereoselective Palladium-Catalyzed Formate Reduction of Allylic
Carbonates en Route to Polypropionate Systems
Anh Chau,‡ Jean-Franc¸ois Paquin,§ and Mark Lautens*
DaVenport Research Laboratories, Department of Chemistry, UniVersity of Toronto, Toronto,
Ontario, Canada M5S 3H6
mlautens@chem.utoronto.ca
ReceiVed NoVember 1, 2005
Diastereoselective palladium-catalyzed formate reduction of allylic carbonates presents unique opportunities
for applications in target-oriented organic synthesis provided that selectivity, in particular stereoselectivity,
in the course of this metal-catalyzed reaction can be controlled. This article describes our recent
developments on new and efficient metal-catalyzed processes exploiting resident stereocenters on the
substrates as a means to control stereoselectivity en route to preparing propionate units containing an
array of stereochemical patterns. In particular, the effect of the protecting group, the stereochemistry of
the aldol adduct, neighboring substituents, and the olefin geometry were examined. Strategic choice of
the above parameters provides entry into three of the four possible diastereomeric triads, namely syn-
syn, anti-syn, and anti-anti. Preliminary results indicate that construction of the syn-anti triad is possible,
albeit in moderate diastereoselectivity.
Introduction
developed.3-5 Nevertheless, more efficient and versatile syn-
thetic strategies are required.
Polypropionates (characterized by their alternating methyl-
hydroxy-methyl substituents) embody an important and diverse
class of natural products such as the ionophores and macrolides.1
In recent years, polypropionate systems have stimulated exten-
sive interest due to their association with a broad spectrum of
biologically active targets with proven or potential use in
medicine, including those with antibiotic and anti-cancer proper-
ties.2 Thus the rapid and facile assembly of the polypropionate
framework would be very practical and beneficial. A large
number of methodologies toward these targets have already been
We recently disclosed the application of γ-carboxy-R,â-
unsaturated aldehyde 2 as a synthetic equivalent of â,γ-
unsaturated aldehyde 5 (Figure 1).6 In this strategy, the allylic
(3) For reviews, see (a) Koskinen, A. M. P.; Karisalmi, K. Chem. Soc.
ReV. 2005, 34, 677. (b) Lautens, M.; Fagnou, K.; Hiebert, S. Acc. Chem.
Res. 2003, 36, 48. (c) Meyer, C.; Blanchard, N.; Cossy, J. Acc. Chem. Res.
2003, 36, 766. (d) Paterson, I.; Florence, G. J. Eur. J. Org. Chem. 2003,
2193.
(4) For selected recent approaches to polypropionates units, see (a) Lohse-
Fraefel, N.; Carreira, E. M. Org. Lett. 2005, 7, 2011. (b) Fader, L. D.;
Carreira, E. M. Org. Lett. 2004, 6, 2485. (c) Turks, M.; Fonquerne, F.;
Vogel, P. Org. Lett. 2004, 6, 1053. (d) Chemler, S. R.; Roush, W. R. J.
Org. Chem. 2003, 68, 1319.
(5) For selected total synthesis, see (a) Crimmins, M. T.; Christie, H.
S.; Chaudhary, K.; Long, A. J. Am. Chem. Soc. 2005, 127, 13810. (b)
Paterson, I.; Lyothier, I. J. Org. Chem. 2005, 70, 5494. (c) Takemoto, T.;
Jackson, P. S.; Ley, S. V. Angew. Chem., Int. Ed. 2003, 42, 2521. (d)
Lautens, M.; Colucci, J. T.; Hiebert, S.; Smith, N. D.; Bouchain, G. Org.
Lett. 2002, 4, 1879. (e) Panek, J. S.; Jain, N. F. J. Org. Chem. 2001, 66,
2747. (f) Hu, T.; Tanenaka, N.; Panek, J. S. J. Am. Chem. Soc. 2002, 124,
12806. (g) Mandal, A. K. Org. Lett. 2002, 4, 2043. (h) Nagaoka, H.; Rutsch,
W.; Schmid, G.; Iio, H.; Johnson, M. R.; Kishi, Y. J. Am. Chem. Soc. 1980,
102, 7962.
‡ Present address: Department of Medicinal Chemistry, Merck Frosst Centre
for Therapeutic Research, P.O. Box 1005, Pointe-Claire-Dorval, Que´bec, Canada
H9R 4P8.
§ Present address: De´partement de chimie, Universite´ Laval, Que´bec, Que´bec,
Canada G1K 7P4.
(1) (a) O’Hagen, D. Nat. Prod. Rep. 1995, 12, 1. (b) Dutton, C. J.; Banks,
B. J.; Cooper, C. B. Nat. Prod. Rep. 1995, 12, 165.
(2) For recent examples of biologically active polypropionates, see (a)
Kasai, Y.; Komatsu, K.; Shigemori, H.; Tsuda, M.; Mikami, Y.; Kobayashi,
J. J. Nat. Prod. 2005, 68, 777. (b) Angawi, R. F.; Swenson, D. C.; Gloer,
J. B.; Wicklow, D. T. J. Nat. Prod. 2005, 68, 212. (c) Ho¨ller, U.; Gloer, J.
B.; Wicklow, D. T. J. Nat. Prod. 2002, 65, 876.
10.1021/jo052267s CCC: $33.50 © 2006 American Chemical Society
Published on Web 01/28/2006
1924
J. Org. Chem. 2006, 71, 1924-1933

DiastereoselectiVe Palladium-Catalyzed Reduction
Whereas a number of investigations have focused on under-
standing the palladium-catalyzed formate reduction, a search
through the literature reveals that few of them have been devoted
to this stereoselective approach. While there have been numerous
examples of stereoselective palladium-catalyzed formate reduc-
tions of allylic moieties, they usually rely on the inversion of
an allylic stereocenter10 or the use of achiral11 or racemic12
substrates in conjunction with a chiral ligand. To the best of
our knowledge, there are no examples reported in which the
stereocenters in this class of substrates have been exploited to
influence the differentiation of the two diastereoisomeric
palladium π-allyl complexes that are formed upon oxidative
insertion of Pd(0).13 As such, we hoped to look at diastereose-
lective palladium-catalyzed formate reduction as a stereoselec-
tive method and strategy which would provide efficient routes
to the construction of stereotriad building blocks A-D (Figure
2).14
FIGURE 1. Aldol/palladium-catalyzed formate reduction strategy
ester 3 is obtained from an aldol reaction between the corre-
sponding enol ether 1 and R,â-unsaturated aldehyde 2 followed
by protection of the alcohol functionality. The key step in this
strategy is the palladium-catalyzed formate reduction7 of 3 that
affords our desired products 4 in good yields. Although a direct
aldol reaction between enol ether 1 and the â,γ-unsaturated
aldehyde 5 could be envisaged to access this class of com-
pounds, this has limited applications. In the case when R ) H,
the â,γ-aldehyde 5a tends to isomerize to the more stable R,â-
unsaturated system; therefore, a 5-fold excess of the aldehyde
is required in order to obtain good yields.8 In the case when
R ) Me, the enantiomerically pure aldehyde 5b has not been
reported and the diastereoselectivity with the racemic aldehyde
is modest.9
Results and Discussion
Optimization Studies. Early efforts focused on optimizing
the original reducing system, Pd2(dba)3 (5 mol %), n-Bu3P (20
mol %), HCO2NH4 (2.2 equiv), DMF, 50 °C, using (E)-8 as
the reference substrate,15 by varying different parameters such
as the palladium loading (5-20 mol %), the phosphines (n-
Bu3P, Ph3P, Cy3P, TFP16), the concentration (0.1-0.25 M), the
Pd:P ratio (2:1 and 1:1), and the temperature (50-70 °C) were
unsuccessful. It was surmised that converting to a more reactive
leaving group would lead to enhanced diastereoselectivities
enabling the reactions to be carried out under milder conditions.
Carbonates are known to be excellent leaving groups in π-allyl
Pd chemistry,17 and applying the original conditions in the
presence of this moiety ((E)-9) gave incomplete conversion with
a disappointing dr of 75:25. A survey of different solvent
systems18 revealed that acetonitrile was the solvent of choice,
Early experiments in our laboratory demonstrated the potential
of this strategy since the reduction of syn-aldol adduct 6 under
standard palladium-catalyzed reduction conditions furnished
terminal olefin 7 in 95% isolated yield as the sole regioisomer
(eq 1).6b While this result was encouraging, the effect of an
alkyl substituent at the vinylic position on the regioselectivity
and the ability of such a group to induce diastereoselectivity
remained to be investigated. The palladium reduction of this
substrate becomes attractive now because it involves the
formation of a new stereocenter, affording the possibility of a
substrate-controlled diastereoselective transformation. This meth-
odology would prove useful as a new efficient approach to the
assembly of stereotriad fragments. In addition, these subunits
are particularly versatile as the terminal olefin of these adducts
can be further elaborated, both functionally and for chain
extension by a variety of protocols. In light of these results,
allylic acetate (E)-8 was subjected to similar reduction conditions
resulting in alkene syn-10 in good yield (75%) and as a
diastereomeric mixture of 5.7:1 in favor of the syn isomer, along
with 10% of the internal olefin 11 (eq 2).
(10) Ba¨ckwall, J.-E. Acc. Chem. Res. 1983, 16, 335.
(11) Hayashi, T.; Iwamura, H.; Naito, M.; Matsumoto, Y.; Uozumi, Y.;
Miki, M.; Yanagim, K. J. Am. Chem. Soc. 1994, 116, 775.
(12) Hayashi, T.; Kawatsura, M.; Iwamura, H.; Yamaura, Y.; Uozumi,
Y. J. Chem. Soc., Chem. Commun. 1996, 1767.
(13) Tsuji has reported a few examples where reduction of allylic
carbonates or formates gave a mixture of diastereomers with modest to no
selectivity, see Mandai, T.; Suzuki, S.; Murakami, T.; Fujita, M.; Kawada,
M.; Tsuji, J. Tetrahedron Lett. 1992, 33, 2987.
(6) (a) Lautens, M.; Paquin, J.-F. Org. Lett. 2003, 5, 3391. (b) Hughes,
G.; Lautens, M.; Wen, C. Org. Lett. 2000, 2, 107.
(14) Review: Hoffman, R. W. Angew. Chem., Int. Ed. Engl. 1987, 26,
489.
(7) (a) Tsuji, J.; Mandai, T. Synthesis 1996, 1 and references therein.
(b) Tsuji, J. Minami, I.; Shimizu, I. Synthesis 1986, 623. (c) Mandai, T.;
Matsumoto, T.; Kawada, M.; Tsuji, J. Tetrahedron 1993, 49, 5483. (d) Tsuji,
J.; Yamakawa, T. Tetrahedron Lett. 1979, 7, 613. (e) Hayashi, T. Acc. Chem.
Res. 2000, 33, 354. (f) Hayashi, T. J. Organomet. Chem. 1999, 576, 195.
(8) For aldol reaction with 5a, see (a) Crimmins, M. T.; Choy, A. L. J.
Am. Chem. Soc. 1999, 121, 5653. (b) Paquette, L. A.; Braun, A. Tetrahedron
Lett. 1997, 38, 5119. (c) Crimmins, M. T.; Kirincich, S. J.; Wells, A. J.;
Choy, A. L. Synth. Commun. 1998, 28, 3675.
(15) All substrates used in this study are racemic unless noted otherwise.
For preparation of substrates see Supporting Information for details.
(16) TFP ) tri-2-furylphosphine. For a review on the use of TFP in
catalysis, see Anderson, N. G.; Keay, B. A. Chem. ReV. 2001, 101, 997.
(17) (a) Tsuji, J. Palladium Reagents and Catalysts. InnoVations in
Organic Synthesis; John Wiley & Sons: Chichester, 1995; pp 290-422.
(b) Pfaltz, A.; Lautens, M. ComprehensiVe Asymmetric Catalysis; Jacobsen,
E. N., Pfaltz, A., Yamamoto, H., Eds.; Springer: New York, 1999; Vol. 2,
pp 833-884. (c) Paquin, J.-F.; Lautens, M. ComprehensiVe Asymmetric
Catalysis; Jacobsen, E. N., Pfaltz, A., Yamamoto, H., Eds.; Springer-
Verlag: Berlin, 2004; pp 73-95.
(9) For aldol reaction with 5b, see (a) Roush, W. R. J. Org. Chem. 1991,
56, 4151. (b) Ahmar, M.; Bloch, R.; Mandville, G.; Romain, I. Tetrahedron
Lett. 1992, 33, 2501.
J. Org. Chem, Vol. 71, No. 5, 2006 1925

Chau et al.
TABLE 1. Optimized Catalyst System
FIGURE 2. Stereotriad building blocks A-D
increasing the propensity of the starting material to undergo
reduction with good conversion and dr > 91:9 at 40 °C or room
temperature.
conversiona
(°C) 10 11 (E)-9
Pd loading
(mol %)
[conc]
(M)
T
entryb
Pd:P
drc
To improve the efficiency of the reaction, as well as resolve
some variability problems, alternative palladium and phosphine
sources, Pd(OAc)2 and [n-Bu3PH]BF4, were investigated,
respectively.19-21 These modifications permitted the number and
the nature of the ligands on the Pd center to be controlled,22 as
well as avoiding the use of air-sensitive n-Bu3P. The tetrafluo-
roborate salt is an air-stable precursor of the phosphine and can
be easily liberated under the reaction conditions in the presence
of a base. Under the modified conditions, employing 10 mol %
Pd(OAc)2, 10 mol % [n-Bu3PH]BF4, and 3 equiv of HCO2H/
Et3N23 (1:2) at 40 °C in CH3CN for 5 min gave complete
conversion with less than 3% of the internal olefin 11 in a dr
of 93:7. A summary of the optimization studies on (E)-9 is
presented in Table 1.24 Using 20 mol % phosphine led to lower
conversion with a dr of 83:17 along with 17% of 11. The
absence of a phosphine ligand effectively shuts down the
reaction (entry 3), and thus future optimization experiments were
performed using a 1:1 ratio of Pd:P. The conversion and the dr
were identical whether the reaction was conducted at room
temperature or 40 °C (entries 1 and 4). However, running the
reaction at 0-5 °C led to incomplete conversion although an
excellent dr was obtained (entry 5). Diluting the reaction had a
beneficial effect on the diastereoselectivity of the reaction since
a ratio of >94:6 was obtained (entry 6). The palladium loading
could be decreased to 2.5 mol % (entries 6, 7, 9), consequently
increasing the reaction time from 1 to 3 h. The order of addition
had some effect since an inverse addition (addition of the
catalyst to the reaction mixture) gave good conversion with a
dr of 92:8 along with a substantial amount of 11 (entry 8).
Furthermore, addition of 1 equiv of Et3N prior to the formate
source and (E)-9 had no effect (entry 10).
1
2
3
10
10
10
10
10
10
5
5
2.5
2.5
1:1
1:2
1:0
1:1
1:1
1:1
1:1
1:1
1:1
1:1
0.1
0.1
0.1
0.1
40
40
40
20
97 <3
0
0
100
0
47
0
0
0
0
0
93:7
83:17
83 17
0
0
4
97 <3
93:7
5d
6
0.1
0-5 50 <3
>94:6
>94:6
>94:6
92:8
>94:6
>94:6
0.05
0.05
0.05
0.05
0.05
20
20
20
20
20
96 ∼4
96 ∼4
83 17
97 <3
97 <3
7
8e
9
10f
a Estimated by 1H NMR spectroscopy of crude mixture after workup.
b 0.1 mmol scale, solution of HCO2H/Et3N (1:2) in CH3CN. c See ref 24.
d Reaction stopped after 8 h. e Inverse addition: solution of Pd(OAc)2 and
[n-Bu3PH]BF4 in CH3CN was added to a solution of (E)-9 and HCO2H/
Et3N (1:2) in CH3CN; f Successive addition of Et3N (1 equiv) and a solution
of (E)-9 and HCO2H/Et3N (1:1) in CH3CN to a solution of Pd(OAc)2 and
[n-Bu3PH]BF4 in CH3CN.
Influence of the Leaving Group and Protecting Group.
To examine the influence of the protecting group and leaving
group on the reduction, a series of compounds were prepared
with varying steric requirements and leaving group ability (Table
2).27 Reduction of allylic benzyl carbonate (E)-12 under the
optimized conditions proceeded to give the desired product 10
with results identical to those of the ethyl carbonate (entry 2).
Alternative leaving groups such as acetate or formate28 resulted
in incomplete conversion (although the dr was >94:6) and no
reaction, respectively (entry 3 and 4). The observation that
changing the leaving group does not affect the diastereoselec-
tivity supports the reasonable proposal that the ionization is not
the selectivity-determining step.
When the reduction was conducted in the presence of a TES
or TBS protecting group, the reduced adducts (10 and 19) were
formed in good yield and high diastereoselectivity (entries 1
and 5). In the case of compound (E)-15 (R ) Ac), the desired
product 20 was isolated in a moderate yield with a low dr (entry
6), but with selective ionization of the carbonate over the acetate.
Treatment of the benzoate ((E)-16) system under the reaction
conditions proceeded in moderate yield and diastereoselectivity
(entry 7). Protection as the pivalate ester ((E)-17), furnished an
Upon scaling up of the reaction, at a 2.5 mol % catalyst
loading, heating to 40 °C was necessary to achieve complete
conversion within a reasonable time,25 fortunately without
compromising the levels of diastereoselectivity. Optimized
conditions, Pd(OAc)2 (2.5 mol %), [n-Bu3PH]BF4 (2.5 mol %),26
HCO2H/Et3N (1:2) (3 equiv), CH3CN (0.05 M), 40 °C, were
used for the rest of the study.
(18) Other solvents, including THF, DMF, or a combination of the two,
gave lower conversion with low dr.
(19) Mandai, T.; Matsumoto, T.; Tsuji, J.; Saito, S. Tetrahedron Lett.
1993, 34, 2513.
(20) Netherton, M. R.; Fu, G. C. Org. Lett. 2001, 3, 4295.
(21) The phosphonium salt can be easily synthesized (see ref 20) or,
alternatively, is available from Strem Chemicals.
(22) It was demonstrated that the dba ligands are not fully displaced by
n-Bu3P from Pd2(dba)3 (see ref 19) or by Ph3P from Pd(dba)2; see (a)
Amatore, C.; Jutand, A. Coord. Chem. ReV. 1998, 178-180, 511. (b)
Amatore, C.; Jutand, A.; Khalil, F.; M’Barki, M. A.; Mottier, L. Organo-
metallics 1993, 12, 3168.
(25) On a 0.2 mmol scale, the reaction with 2.5 mol % Pd at room
temperature led to almost complete conversion after 3 days, compared to
<1 h at 40 °C. On a 0.5 mmol scale, the reaction with 2.5 mol % Pd is
done in <2 h at 40 °C.
(26) A brief screening of ligands ([Me3PH]BF4, [Et3PH]BF4, [i-Pr3PH]-
BF4, [t-Bu3PH]BF4, and [Cy3PH]BF4) was carried out to assess the
effectiveness of alkyl phosphines in controlling the diastereoselectivity. All
gave excellent conversion and dr’s, along with variable amounts of the
internal olefin. [n-Bu3PH]BF4 consistently led to the cleanest reactions, and
therefore was used throughout the rest of the study.
(27) Stereochemistry of the products was determined by comparison with
authentic samples or similar or previously reported compounds; see
Supporting Information for details.
(28) Tsuji has shown that the reduction of formate ester was highly
sensitive to the purity of the phosphine and the Pd catalyst, see ref 7c or
Mandai, T.; Matsumoto, T.; Kawada, M.; Tsuji, J. J. Org. Chem. 1992, 57,
1326. It is possible that the presence of amine salts shuts down the reaction.
(23) Formate salts (HCO2K, HCO2Cs) were not as effective, probably
due to their low solubility.
(24) The dr (ratio of syn/anti) of all reactions was estimated by 1H NMR
spectroscopy. For the purpose of this study, >94:6 dr indicates that the
other diastereomer was present in less than 6% (typically 4-6%), whereas
>99:1 dr specifies that the other diastereomer was not detectable.
1926 J. Org. Chem., Vol. 71, No. 5, 2006

DiastereoselectiVe Palladium-Catalyzed Reduction
TABLE 2. Influence of the Protecting Group and Leaving Group
TABLE 3. Influence of Protecting Group
entry
R
substrate
product
yielda (%)
drb
entry
R
R′
Et
Bn
Ac
CHO
Et
Et
Et
Et
Et
substrate
product
yielda (%)
drb
1
2
3
4
5
TES
TBS
TBS
Ac
(E)-24
(E)-25
(E)-25
(E)-26
(E)-27
28
29
29
30
31
84
46
89
51
63
92:8
1
2
3
4
5
6
7
8
9
TES
TES
TES
TES
TBS
Ac
Bz
Piv
H
(E)-9
(E)-12
(E)-8
(E)-13
(E)-14
(E)-15
(E)-16
(E)-17
(E)-18
10
10
10
10
19
20
21
22
23
85
85
69
n/r
85
72
48d
42
60
>94:6
>94:6
>94:6
ndc
>94:6
58:42
80:20e
>99:1f
73:27
>99:1
>99:1c
29:71
20:80d
H
a Isolated yield. b The dr of the isolated product; see ref 24. c Reaction
was performed using 5 mol % Pd(OAc)2 and 5 mol % [n-Bu3PH]BF4 since
some starting carbonate (20-30%) was observed by 1H NMR when using
the standard conditions. d Product was contaminated with 25% of the internal
olefin.32
a Isolated yield. b The dr of the isolated product; see ref 24. c Not
determined. d The crude product was contaminated with an unidentified
product (18%) and internal olefin (7%). e The dr was estimated by 13C NMR
spectroscopy. f The dr of the crude mixture was 89:11.
could be removed by flash chromatography to give 29 in
moderate yield (46%) as a single diastereomer along with
unreacted (E)-25 (entry 2). The yield could be improved to 85%
by increasing the catalyst loading to 5 mol % with equal
selectivities (entry 3). This permits access to the anti-syn
isomeric sequence B. Interestingly, when (E)-26 or (E)-27 was
reacted under the optimal conditions, the anti diastereomers anti-
30 and anti-31 were formed in moderate yield and low dr (entry
4 and 5). This result is particularly motivating since it opens
the way to the synthesis of the anti-anti triad for which there
are limited known methods.31
Consequently, a number of different reaction parameters33
were examined including the palladium:phosphine ratio, con-
centration, solvent, and nature of the ligand used to optimize
for the anti-anti triad using the free alcohol (E)-27. Formation
of the anti-substrate would require delivery of the hydride from
the same face as the hydroxy group, perhaps through pre-
coordination. It was hoped that, similar to the well-established
hydroxy-directed rhodium-catalyzed hydrogenation of allylic
alcohols, substrate direction could provide useful levels of
diastereoselectivity.34 Carrying the reaction out in a nonpolar
solvent, such as toluene, indeed gave slightly higher levels of
diastereoselectivity (14:86 dr). Unfortunately this was also
accompanied by predominant formation of the internal olefin
(86%) selectively. Interestingly, the diastereoselectivity was
slightly increased by slow addition of the formate source over
a period of 8 h to give the desired anti-anti triad, anti-31, in
63% yield (13:87 dr) with 35% of the internal olefin. These
results suggest that the slow addition extends the lifetime of
the palladium π-allyl intermediates to provide a chance for syn-
anti isomerization to occur.
89:11 dr in the crude mixture. Fortunately, the major isomer
could be isolated by flash chromatography to give 22 in 42%
yield and >99:1 dr. The electronic nature of the protecting group
seems to have some influence since although it was possible to
achieve good diastereoselectivity with some non-silicon based
protecting groups, the isolated yields were systematically lower
due to side-products formed or unreacted starting material.
Subjecting the free hydroxyl under the reaction conditions
afforded a lower yield and a modest dr (entry 9).29,30 Surpris-
ingly, the reduction of the same compound bearing an acetate
instead of an ethyl carbonate under the originally reported
conditions gave a 1:1 mixture of diastereomers,6b whereas a
slight preference for the syn isomer is observed under our
optimized system. From this study, it appears that the protecting
group plays an important role in influencing the diastereose-
lectivity in the course of the formate reduction of these
compounds but the cause of this behavior is not fully understood.
The syn-selectvity of the reaction correlated approximately with
the steric demand of the protecting group. Reactions employing
bulky silyl protecting groups afforded the highest selectivity,
and the dr decreased with diminishing the protecting group size
until reaching synthetically useless levels with the free hydroxyl
group. In conjunction with a bulky silyl protecting group, such
as a tert-butyldimethylsilyl, the syn-syn triad A can be accessed
in synthetically useful levels of diastereoselectivity using this
methodology.
anti-syn Triad. The aforementioned results led us to inves-
tigate the effect of the relative stereochemistry on the starting
allylic carbonates. The results for the reduction of the anti aldol
adducts (E)-24-27 under standard reduction conditions are
shown in Table 3. Exposure of (E)-24 with formic acid and
triethylamine in the presence of 2.5 mol % of the active
palladium catalyst generated from Pd(OAc)2 and [n-Bu3PH]BF4,
proceeded to give 28 in good yield (84%) albeit the dr was
slightly lower than that for the syn diastereomer (E)-9 (entry
1). Congruent with the syn-syn triad, the use of a large TBS
protecting group bestowed better selectivity with a dr of >94:6
as observed from the crude mixture. The minor diastereomer
Investigation of the Effect of Olefin Geometry. To date
we are able to prepare the stereotriads in greater than 94:6 dr.
However, the methodology is restricted to the construction of
syn-syn A and anti-syn B stereotriads. Nevertheless, preliminary
results indicated that the preparation of anti-anti stereotriads is
also possible for cases where R ) H or Ac, albeit in low dr
(31) Hoffmann, R. W.; Dahmann, G.; Andersen, M. W. Synthesis 1994,
629 and references therein.
(32) Evans, D. A.; Nelson, J. V.; Vogel, E.; Taber, T. R. J. Am. Chem.
Soc. 1981, 103, 3099.
(29) The diastereomers are separable and the pure syn isomer (30%) could
be isolated by flash chromatography.
(30) The lack of selectivity observed for the reduction of the free hydroxyl
group substrate (E)-23 can be overcome using the chiral (R)-MeO-MOP
ligand (>96:4 dr).
(33) Solvent: CH3CN, DMF, THF, DME, dioxane, and toluene.
Ligands: [n-Bu3PH]BF4, [Me3PhH]BF4, [t-Bu3PH]BF4, PPh3, and TFP.
Concentration: 0.025-0.1 M.
(34) Hoveyda, A. H.; Evans, D. A.; Fu, G. C. Chem. ReV. 1993, 93,
1307.
J. Org. Chem, Vol. 71, No. 5, 2006 1927

Chau et al.
SCHEME 1. Effect of Double-Bond Geometry
SCHEME 2. Effect of r-Substitution and Geometry of
Double Bond
(vide supra). Access to stereotriads C and D is also important
by virtue of their frequent occurrence in polyketide natural
products of synthetic interest such as tirandamycin,35 rifamy-
cin,36 ionomycin,37 and denticulatin.38
To correct this deficiency, the effect of the olefin geometry
on the course of the reaction was probed. From our earlier work,
the reduction of TES protected (E)-9 gave >94:6 dr (Table 2,
entry 1). Using the conditions applied to (E)-9, reduction of
(Z)-24 gave the propionate system in good yield as a 33:67
diastereomeric mixture (Scheme 1). It is interesting to note that
reduction of the (Z)-carbonate gives opposite selectivity to that
of the (E)-carbonate, thereby favoring formation of the anti
isomer. Gratifyingly, subjecting the free alcohol (Z)-27 to
standard formate reduction conditions furnished the anti-anti
triad (anti-31) in good yield with 6:>94 diastereoselectivity with
no trace of the internal olefin isomer, according to analysis of
isolated in 69% with a 80:20 dr. Complete conversion could be
achieved by using 5 mol % of the catalyst, and 10 was obtained
in 89% yield with the same diastereoselectivity (Scheme 1).
Interestingly, when (Z)-15 and (Z)-18 were submitted to the
reaction conditions, the anti diastereomer was formed in
moderate yield and dr.39,40 In an effort to enhance the stereo-
selectivity, several different permutations of concentration,
solvent, formate (1-10 equiv), phosphine ligands, palladium:
phosphine ratio (1:0-1:2), palladium sources (Pd(OAc)2, Pd2-
(dba)3), and temperature (-10-80 °C) were evaluated.33
However, it proved to be unsuccessful. The syn-anti triad
remains the most elusive to obtain in the course of our study.
Following this work, the highest selectivity that could be
obtained remained a dr of 20:80.
Effect of Substitution. The effects of substitution at the R
position, relative to the ketone, on the regioselectivity and
diastereoselectivity were next investigated (Scheme 2). Unfor-
tunately, for less sterically biased substrates, low diastereomeric
ratios were obtained in the reduction of the free alcohol (E)-33
(58:42 dr). A slight increase in the selectivity was obtained by
incorporating a bulky triethylsilyl protecting group which gave
34 in 87% yield with a dr of 75:25.41 Reduction of (Z)-32 and
(Z)-33 gave similar diastereoselectivities to their (E)-configured
analogues. In both scenarios, irrespective of olefin geometry,
the nature of the protecting group had little effect in the absence
of the R substituent.
1
the H NMR spectrum. Furthermore, the selectivity could be
increased to at least 1:>99 by slow addition of the formate
source over 6 h. Under the same reaction conditions, reduction
of the (E)-configured carbonate (E)-27 gave a dr of 73:27, but
was compromised with considerable formation of the internal
olefin (Table 2, entry 9). This result further highlights the
dramatic influence of olefin geometry in the palladium reduction
reaction, whereby incorporation of the (Z)-configured olefin into
the anti-aldol adduct leads to the anti-anti isomeric sequence
C.
Accordingly, the (Z)-configured syn aldol adduct was next
investigated. The anti-selective reduction of syn-aldol adducts
is much harder to achieve as these substrates have an intrinsic
preference to provide the syn triad as a result of steric effects.
Diastereoselective formate reduction of syn aldol adducts proved
more challenging than the previous substrate. The reduction of
the (Z)-allylic carbonate (Z)-9 using 2.5 mol % of the catalyst
system gave incomplete conversion, although 10 was still
Introduction of a gem-dimethyl substituent in the R position
significantly increased the dr to >99:1 in the presence of the
bulky silyl protecting group ((E)-36) albeit in moderate yield
due to unreacted starting material. The conversion was improved
by employing 5 mol % of the catalyst system whereby 38 was
obtained in 81% as the syn isomer (>99:1 dr). In the absence
of this steric bulk ((E)-33), the dr dramatically decreased to 55:
45.42 These results suggest that the presence of adjacent methyl
(35) Duchamp, D. J.; Branfman, A. R.; Button, A. C., Jr.; Rinehard, K.
L. J. Am. Chem. Soc. 1973, 95, 4077.
(36) (a) Brufani, M.; Cerrini, S.; Fedeli, W.; Vaciago, A. J. Mol. Biol.
1974, 87, 409. (b) Hanessian, S.; Wang, W.; Yonghua, G.; Olivier, E. J.
Am. Chem. Soc. 1997, 119, 10034 and references therein.
(37) Toeplitz, B. K.; Cohen, A. I.; Funke, P. T.; Parker, W.; Gougoutas,
J. Z. J. Am. Chem. Soc. 1979, 1015, 3344.
(38) Hochlowski, J. E.; Coll, J. C.; Faulkner, D. J.; Biskupiak, J. E.;
Ireland, C. M.; Qu-tai, Z.; Cun-heng, H.; Clardy, J. J. Am. Chem. Soc. 1983,
105, 7413.
(39) Compound 20 was contaminated with 9% of the internal olefin.
(40) Compound 23 was contaminated with 20% of the internal olefin
(E/Z 1:1).
1928 J. Org. Chem., Vol. 71, No. 5, 2006

DiastereoselectiVe Palladium-Catalyzed Reduction
SCHEME 3. Other Substrates
used with NMR spectroscopy to get more insight on biological
systems,45 the preparation of isotopically labeled propionate units
was attempted.
The reaction of (E)-9 using the optimized conditions and
DCO2D in place of HCO2H gave the desired labeled compound
44 in good yield, excellent diastereoselectivity, and deuterium
incorporation (eq 3). The deuterium atom was incorporated only
1
at the allylic position, which is evident by H and 13C NMR
substituents and a bulky silyl protecting group exerts a
significant influence on the stereoselectivity in the reduction.
The (Z)-double bond geometry, which had been observed to
select for the anti isomer in previous cases, indeed exhibited a
reversal in selectivity. Reaction of a TES protected (Z)-36
yielded a diastereometric mixture of 33:67 in favor of the anti
isomer. Interestingly, reduction of the free hydroxyl derivative
(Z)-37 provided a significantly higher dr of 7:93 in favor of the
expected isomer anti-39. This observation is consistent with the
reduction of (Z)-27 above, which also favored the anti isomer.
These results indicate the crucial role of the R substituent on
the stereoinduction process. Interestingly, diastereoselectivities
are almost independent of the olefin geometry in the absence
of an R substituent. Along the same lines, increasing the steric
bulk of the protecting group favored the syn isomer while
decreasing the size of the protecting group favored the anti
isomer. This is operative in all the scenarios discussed. Last,
there exists a synergistic relationship between the protecting
group and olefin geometry i.e., a large protecting group and
(E)-olefin geometry is required to select for the syn isomer and
vice versa.
and HRMS.46 This example shows that the use of formic acid-
d2 allows for the rapid and stereocontrolled preparation of
deuterium-labeled propionate fragments.
SCHEME 4. Working Model
Preparation of the Other Substrates. All the substrates
prepared so far possess a methyl group at the vinylic position
in order to afford a propionate moiety after reduction. However,
to expand the scope and test the generality of the reaction, we
elected to prepare substrates with alternative groups at this
position. With the requisite substrates in hand, the effect of the
substitution was evaluated, and the results are presented in
Scheme 3.43 Replacing the ethyl group with a phenyl ((E)-40)
has little effect on the reduction, and the selectivity remained
at >94:6. Furthermore, the minor diastereomer could be
removed by flash chromatography to give 41 in 86% yield as
a single diastereomer. The reaction of (E)-42 gave a modest
Mechanistic Insights. It is not possible with the current
information available to propose a definitive model for the
reactivity observed. However, some comments about the pos-
sible equilibration of the Pd complexes can be made. The
working model illustrated in Scheme 4 is based upon proposals
described in the literature for this class of substrate.17b-c,47 The
oxidative addition of the palladium catalyst to the allylic
carbonates (E)-45 would result in the formation of the π-allyl
complexes 46 and 47. The oxidative addition might occur
preferentially from one side of the allylic carbonate. However,
this should not affect the outcome since it is believe that 46
and 47 are in rapid equilibration.48As first proposed by Hayashi,
1
diastereoselectivity of 89:11 as estimated by H NMR of the
crude mixture. However, similarly, the minor isomer could be
removed by flash chromatography to give 43 in 58% yield as
a single diastereomer.
Isotopic Labeling. The use of formic acid-d2 in the pal-
ladium-catalyzed formate reduction of allylic systems offers an
entry into labeled products having deuterium at the allylic
position.44 Since propionates are an important class of biological
molecules and isotopically labeled compounds are frequently
(45) For examples, see (a) Anglister, J. Quart. ReV. Biophys. 1990, 23,
175 and references therein. (b) Wade, D. Chem.-Biol. Interact. 1999, 117,
191 and references therein. (c) Tsuzuki, H.; Tsukinoki, T.; Mataka, S.;
Fukata, G.; Ishimoto, K.; Tashiro, M. Radioisotopes 1995, 44, 929.
(46) The stereochemistry of 44 was assigned based on the selectivity
observed with (E)-9 using HCO2H.
(47) (a) Oshima, M.; Sakamoto, T.; Maruyama, Y.; Ozawa, F.; Shimizy,
I.; Yamamoto, A. Bull. Chem. Soc. Jpn. 2000, 73, 453. (b) Oshima, M.;
Shimizu, I.; Yamamoto, A.; Ozawa, F. Organometallics 1991, 10, 1221.
(c) Hutchins, R. O.; Learns, L. J. Org. Chem. 1982, 47, 4380.
(41) Derivatives of this class of compounds have attracted some interest,
see (1) (a) Flippin, L. A.; Brown, P. A.; Jalali-Araghi, K. J. Org. Chem.
1989, 54, 3588. (b) Alemany, C.; Bach, J.; Garcia, J.; Lo´pez, M.; Rodr´ıguez,
A. B. Tetrahedron 2000, 56. (c) Hirama, M.; Nakamine, T.; Itoˆ, S.
Tetrahedron Lett. 1986, 27, 5281.
(42) The product 39 was contaminated with 22% of the internal olefin.
(43) The dr given is for the isolated product, see ref 24.
(44) Hayashi, T.; Iwamura, H.; Uozumi, Y.; Matsumoto, Y.; Ozawa, F.
Synthesis 1994, 526.
J. Org. Chem, Vol. 71, No. 5, 2006 1929

Chau et al.
SCHEME 5. Strategy
the stereochemical outcome (syn-48 or anti-48) would be
determined by the thermodynamic stability of the epimeric
π-allyl palladium intermediates (46 and 47).49 Based on this
understanding, since the product syn-48 is the major isomer
observed, complex 46 would be the more thermodynamically
stable complex.50 The scenario with the olefinic isomer (Z)-45
could be explained by a similar argument where a mixture of
π-allyl palladium complexes would probably be formed upon
oxidative addition. However, based on Hayashi’s work, syn-
anti isomerization with this class of substrate would be slow
compared to reduction.48 This notion is supported by the fact
that the reduction of E-olefinic isomers gave products with
different diastereoselectivity than the reduction of the analogous
Z-olefinic isomers. On the basis of our results, we reasoned that
when R is a large silyl protecting group, the steric bulk shields
one face of the π-allyl complex, forcing the palladium to
coordinate from the opposite face and leading to the syn
diastereoisomer as the major product. When the size of the
protecting group is decreased, intermediate 50 may be more
favorable due to precoordination of the palladium followed by
hydride delivery from the same face of the π-allyl complex to
give to anti isomer. Although we suggest internal participation
of the free hydroxy group, its critical role in the reduction step
still remains unclear.
TABLE 4. Selected Results from Optimization of Double
Palladium-Catalyzed Formate Reduction
Pd loading
(mol %)
Pd:P
ratio
formate
equiv
[conc]
(M)
entrya
drb
1
2
3
2.5
2.5
2.5
10
2.5
5.0
5.0
2.5
2.5
2.5
2.5
1:1
1:1
1:2
1:1
1:1
1:1
1:1
1:1
1:1
1:1
1:1
3.0
0.05
0.05
0.05
0.05
0.05
0.05
0.05
0.025
0.025
0.025
0.0125
77:23
86:14
76:24
76:24
n/r
71:29
60:40
83:17
83:17
64:36
77:23
3.0c
3.0c
3.0c
3.0c
3.0
At this stage, it is not possible for us to confidently rationalize
the sense of stereoinduction that we observe especially since
the exact nature of the reductive decarboxylation is not known.51
However, one thing that is clear to us is that the reaction itself
is multi-faceted and that there is no single controlling element.
Our results suggest that the relative configuration of the methyl-
hydroxy array obtained from the reduction is controlled by the
initial stereogenic centers in the aldol adduct, the nature of the
protecting group, nearby substitution, and the geometry of the
double bond. These factors all play an important role in
determining the conformation52 of the palladium π-allyl inter-
mediate that subsequently undergoes reduction.
4
5d
6
7
8
9
10
11
6.0
3.0
3.0c
6.0
3.0
a See ref 24. b Complete conversion was obtained for all the entries except
entry 2 (62%). c Slow formate addition to substrate and catalyst over 8 h.
d Room temperature.
reduction of the ketone moiety permits formation of a seventh
stereocenter en route to polypropionate systems.
Double Diastereoseletive Palladium-Catalyzed Formate
Reduction. With our diastereoselective palladium-catalyzed
formate reduction in hand, we considered whether this meth-
odology could serve as a key step toward the construction of
building blocks for expedient polypropionate synthesis. We
envisaged the target molecule 53 being generated by sequential
chain extension in two directions, followed by a double
diastereoselective palladium-catalyzed formate reduction (Scheme
5). One of the advantages that would accrue from the develop-
ment of such a process would be the simultaneous formation
of two new stereocenters. These products are synthetically
interesting because the terminal olefin functionality provides a
handle for additional carbon skeleton-expanding operations and
With the assembly of the bis aldol adducts,53 double pal-
ladium-catalyzed formate reductions were carried out and a
number of different reaction parameters were examined includ-
ing catalyst, ligand, concentration, and formate equivalents to
optimize the diastereoselectivies. Bis carbonate 54 was initially
chosen as the reference substrate, as its reduction leads to a
meso substrate, to simplify our analysis. We felt that the
presence of the two functional groups remote from one another
would not alter the reaction course. On the basis of this
reasoning, we expected nearly no formation of the anti-syn-
syn-syn-anti isomer. Table 4 highlights the preliminary results
for the optimization of the all syn stereotriad syn-55. The
reduction proceeded smoothly to give the diolefin 55 in 86%
yield and with 77:23 diastereoselectivity for the establishment
of two new stereocenters (entry 1). Slow formate addition was
also found to improve the dr to 86:14, but gave low conversions
(entry 2). Unfortunately, attempts to stabilize the catalyst by
increasing the phosphine:ligand ratio (entry 3) or increasing the
catalyst loadings (entry 4) had a negative effect on the dr.
Increasing the equivalents of formate (6.0 equiv, entry 7) had
negative effects on the dr. Concentration was found to be
important as diluting the reaction by a factor of 2 increased the
(48) It has been shown by Hayashi that the model complex PdCl(η3-
1,1-dimethylallyl)((R)-MOP) existed as a mixture of isomers which are in
an equilibrium state between -60 °C and 20 °C, see ref 11.
(49) (a) Hayashi, T.; Iwamura, H.; Uozumi, Y. Tetrahedron Lett. 1994,
35, 4813. (b) Kawatsura, M.; Uozumi, Y.; Ogasawara, M.; Hayashi, T.
Tetrahedron 2000, 56, 2247. Also see refs 11, 12, 44.
(50) This argument assumes similar rates as in Hayashi’s chemistry. If
the rate of epimerization is much faster than the reduction, the Curtin-
Hammet principle is applicable. In such a case, the syn/anti ratio is not
determined by the relative stability of the π-allyl complexes 46 and 47 but
by the relative rate of product formation from each intermediate.
(51) See ref 47a,b and Guibe´, F. Tetrahedron 1998, 54, 2967 and
references therein.
(52) For a review of the conformational design, see Hoffmann, R. W.
Angew. Chem., Int. Ed. 2000, 39, 2054 and references therein.
(53) Substrates are nonracemic. For preparation of bis aldol adducts, see
Supporting Information for details.
1930 J. Org. Chem., Vol. 71, No. 5, 2006

DiastereoselectiVe Palladium-Catalyzed Reduction
SCHEME 6. Double Palladium-Catalyzed Formate
Reduction
colorless liquid after purification by flash chromatography using
10% Et2O/pentane. H NMR (400 MHz, CDCl3) δ 8.02 (2H, m),
1
7.56 (1H, m), 7.45 (2H, m), 5.84 (1H, ddd, J ) 17.2, 10.4, 8.8 Hz,
anti isomer), 5.76 (1H, ddd, J ) 17.2, 10.0, 8.2 Hz, syn isomer),
5.50 (1H, dd, J ) 8.8, 4.8 Hz, anti isomer), 5.46 (1H, dd, J ) 8.0,
4.8 Hz, syn isomer), 5.12 (2H, m), 2.95 (1H, m), 2.70-2.41 (3H,
m), 1.16 (3H, d, J ) 7.2 Hz, syn isomer), 1.13 (3H, d, J ) 7.2 Hz,
anti isomer), 1.07-1.01 (6H, m); 13C NMR (100 MHz, CDCl3) δ
211.8, 166.0, 139.6, 139.1, 133.0, 130.0, 129.7, 128.4, 116.3, 116.1,
76.3, 48.2, 48.0, 41.1, 35.1, 34.4, 17.5, 16.3, 12.9, 10.6, 7.8; IR
(neat) ν ) 3072, 2977, 2939, 1720, 1452, 1272, 1110, 712 cm-1
;
HRMS-ES calcd for C17H22O3 [M + Na]+ 297.1461, found
297.1447. The spectral data were identical to an authentic sample
(syn isomer only) prepared from protection of syn-23 (BzCl,
pyridine, CH2Cl2, 0 °C to rt) obtained from deprotection of syn-10
(pTsOH‚H2O, H2O, THF).
(4R*,5S*,6R*)-4,6-Dimethyl-5-pivaloyloxy-7-octen-3-one (syn-
22). Following the general procedure on a 0.20 mmol scale using
(E)-17, the crude product (89:11 dr by 1H NMR spectroscopy) was
purified by flash chromatography using 5% Et2O/pentane to give
dr to 83:17 (entry 8). Slow formate addition under optimized
concentration (0.025 M) had no effect on the dr (entry 9).
Further dilution did not further improve the dr (entry 11).
The three diastereomeric bis aldol adducts were treated under
the optimized conditions (i.e. 2.5 mol % Pd(OAc)2, 2.5 mol %
[n-Bu3PH]BF4, and 3 equiv of HCO2H/Et3N (1:2) at 40 °C in
CH3CN (0.025 M)) (Scheme 6). The reactions proceeded
smoothly to give the reduced adducts in good yields and
diastereoselectivities for the establishment of two new stereo-
centers. Although we have not yet pursued further derivitization
of our doubly reduced adducts, additional two-directional
applications can be envisioned.
Conclusions. A new approach to propionate units using a
diastreoselective palladium-catalyzed formate reduction of allylic
carbonates has been developed and can serve as a useful strategy
to construct building blocks for polypropionate synthesis. To
the best of our knowledge, stereocenters in this class of
substrates have never been exploited to control the stereochem-
ical outcome in allylic substitution reactions. Our results indicate
the stereoinduction is multi-faceted and is dependent on a
number of factors including the nature of the protecting group,
nearby substitution, stereochemistry of the aldol adduct, and
olefin geometry. Our methodology allows for the preparation
of syn-syn, anti-syn, and anti-anti triads in synthetically useful
diastereoselectivities. Our preliminary results indicate that the
synthesis of the syn-anti triad is also possible (dr up to 20:80).
Future efforts to devise better solutions to the selectivity
problems of the syn-anti triad in the reduction reaction seem
very much worthwhile.
1
syn-22 (22 mg, 42%, >99:1 dr by H NMR spectroscopy) as a
1
colorless liquid. H NMR (300 MHz, CDCl3) δ 5.70 (1H, ddd,
J ) 17.1, 10.2, 8.4 Hz), 5.18-5.05 (3H, m), 2.83 (1H, m), 2.63
(1H, m), 2.50-2.32 (2H, m), 1.18 (9H, s), 1.07-0.98 (9H, m);
13C NMR (74.5 MHz, CDCl3) δ 211.7, 177.7, 139.6, 116.0, 75.0,
47.8, 40.7, 39.0, 34.3, 27.2, 16.3, 9.8, 7.8; IR (neat) ν ) 3078,
2978, 2935, 2874, 1729, 1481, 1460 1281, 1154 cm-1; HRMS-ES
calcd for C15H26O3 [M + Na]+ 277.1774, found 277.1780. The
spectral data were identical to those of an authentic sample (syn
isomer only) prepared from protection of syn-23 (PivCl, pyridine,
DMAP, CH2Cl2, 0 °C to rt) obtained from deprotection of syn-10
(pTsOH‚H2O, H2O, THF).
(4S*,5S*,6R*)-4,6-Dimethyl-5-triethylsiloxy-7-octen-3-one (syn-
28). Following the general procedure on a 0.20 mmol scale using
1
(E)-24, syn-28 (48 mg, 92:8 dr by H NMR spectroscopy) was
obtained by flash chromatography using 5% Et2O/hexane as a
colorless liquid. 1H NMR (300 MHz, CDCl3) δ 5.82 (1H, m), 5.06
(1H, dt, J ) 7.8, 1.5 Hz), 5.01 (1H, d, J ) 1.2 Hz), 3.94 (1H, dd,
J ) 7.8, 3.6 Hz), 2.74 (1H, m), 2.49 (1H, m), 2.31 (1H, m), 0.95
(18H, m), 0.55 (6H, q, J ) 7.5 Hz); 13C NMR (74.5 MHz, CDCl3)
δ 214.0, 141.8, 114.5, 77.7, 50.1, 40.7, 36.7, 13.5, 13.1, 7.3, 6.9,
5.2; IR (neat) ν ) 3078, 2958, 2874, 1721, 1461, 1378, 1170, 1006
cm-1; MS-EI (m/z) 255 [M - Et]+, 57 [C3H5O]+. The stereochem-
istry was established by conversion of syn-28 to syn-29 by
deprotection (pTsOH‚H2O, H2O, THF) and reprotection (TBSOTf,
2,6-lutidine, CH2Cl2, 0 °C to rt).
(4S*,5S*,6S*)-5-Acetoxy-4,6-dimethyl-7-octen-3-one (syn- and
anti-30). Following the general procedure on a 0.20 mmol scale
using (E)-26, the inseparable mixture of diastereomers (22 mg, 51%,
29:71 dr by 1H NMR spectroscopy) was obtained by flash
chromatography using 20% Et2O/pentane as a colorless liquid. 1H
NMR (400 MHz, CDCl3) δ 5.74 (1H, m), 5.14-5.02 (3H, m), 2.85
(1H, m), 2.55-2.36 (3H, m), 2.00 (3H, s), 1.11 (3H, d, J ) 6.8
Hz, syn isomer), 1.06 (3H, d, J ) 7.2 Hz, anti isomer), 1.01 (6H,
m); 13C NMR (100 MHz, CDCl3) δ 212.5, 170.1, 170.0, 139.8,
138.3, 116.1, 115.4, 77.2, 76.8, 48.2, 48.0, 40.0, 38.9, 34.9, 34.7,
20.8, 20.7, 17.5, 13.7, 13.4, 13.2, 7.5; IR (neat) ν ) 3072, 2976,
2935, 1746, 1718, 1460, 1374, 1233, 1020 cm-1; HRMS-ES calcd
for C12H20O3Na [M + Na]+ 235.1304, found 235.1293. The
stereochemistry was established by comparison with authentic syn
compound prepared from protection of syn-31 (AcCl, pyridine, CH2-
Cl2, 0 °C to rt) obtained from deprotection of syn-28 (pTsOH‚H2O,
H2O, THF).
Experimental Section
General Experimental Methods. See the Supporting Informa-
tion.
General Procedure for Diastereoselective Palladium-Cata-
lyzed Formate Reduction. To a solution of Pd(OAc)2/[n-Bu3PH]-
BF4 (1:1) (2.5 mol %, 0.06 M/CH3CN) under nitrogen was added
a solution of allylic carbonate (1 equiv) and HCO2H/Et3N (1:2) (3
equiv, 1.0 M/CH3CN) in CH3CN (0.05 M) via cannula. The
resulting mixture was heated to 40 °C until the apparition of black
particles (∼1 h). H2O was added, and the aqueous layer was
extracted with Et2O. The combined organic layers were washed
with H2O and brine and dried over MgSO4, and the solvent was
evaporated. The crude product was purified by flash chromatog-
raphy using Et2O/hexane as an eluant.
(4S*,5S*,6S*)-5-Hydroxy-4,6-dimethyl-7-octen-3-one (syn-31).
Following the general procedure on a 0.21 mmol scale using (E)-
1
(4R*,5S*,6R*)-5-Benzoyloxy-4,6-dimethyl-7-octen-3-one (syn-
and anti-21). Following the general procedure on a 0.20 mmol
scale using (E)-16, the inseparable mixture of diastereomers (26
27, 31 (23 mg, 63%, 20:80 dr by H NMR spectroscopy) was
obtained by flash chromatography using 30% Et2O/ pentane as a
colorless liquid. Careful flash chromatography using 30% Et2O/
1
1
mg, 48%, 80:20 dr by H NMR spectroscopy) was isolated as a
hexane allows for isolation of a clean sample of syn-31. H NMR
J. Org. Chem, Vol. 71, No. 5, 2006 1931

Chau et al.
(400 MHz, CDCl3) δ 5.75 (1H, ddd, J ) 16.8, 10.4, 7.6 Hz), 5.04
(2H, m), 3.45 (1H, m), 2.81 (2H, m), 2.61-2.41 (2H, m), 2.30
(1H, quintet, J ) 6.8 Hz), 1.17 (3H, d, J ) 7.6 Hz), 1.07-1.02
(6H, m); 13C NMR (100 MHz, CDCl3) δ 217.6, 141.6, 115.1, 77.4,
47.2, 42.0, 36.2, 14.9, 14.6, 7.3; IR (neat) ν ) 3498, 3078, 2976,
2936, 1712, 1459, 1374, 973 cm-1; HRMS-EI calcd for C10H19O2
[M + H]+ 171.1385, found 171.1377. The stereochemistry was
established by comparison with authentic syn compound prepared
from deprotection of syn-28 (pTsOH‚H2O, H2O, THF).
(3H, s), 0.98 (12H, m), 0.63 (6H, q, J ) 7.6 Hz); 13C NMR (75.4
MHz, CDCl3) δ 177.2, 143.0, 113.4, 79.7, 60.2, 48.4, 41.4, 24.7,
29.0, 16.2, 14.0, 7.1, 5.5; IR (neat) ν ) 2958, 2878, 1735, 1463,
1264, 1113, 1085, 1006 cm-1; HRMS-ES calcd for C17H34SiO3Na
[M + Na]+ 337.2169, found 337.2174. The stereochemistry was
established by comparison with an authentic sample prepared (see
Supporting Information).
Ethyl (3S*,4R*)-3-Hydroxy-2,2,4-trimethyl-5-hexenoate (anti-
39). Following the general procedure on a 0.20 mmol scale using
1
(4S*,5S*,6R*)-5-Hydroxy-4,6-dimethyl-7-octen-3-one (anti-
31). Following the general procedure (with slow addition of formate
over 6 h) on a 0.28 mmol scale using (Z)-27, anti-31 (39 mg, 82%,
1:>99 dr by 1H NMR spectroscopy) was obtained by flash
chromatography using 20% Et2O/pentane as a colorless liquid.
Rf ) 0.29 (70% hexanes:ether); 1H NMR (400 MHz, CDCl3) δ 5.83
(1H, ddd, J ) 17.0, 10.6, 8.4 Hz), 5.07 (2H, m), 3.62 (1H, ddd,
J ) 6.2, 3.7, 1.1 Hz), 2.70 (2H, app dq, J ) 7.1, 7.1 Hz), 2.62-
2.41 (3H, m), 2.33 (1H, m), 1.11 (3H, d, J ) 2.6 Hz), 1.09 (3H, d,
J ) 2.4 Hz), 1.04 (3H, t, J ) 7.3 Hz); 13C NMR (100 MHz, CDCl3)
δ 216.7, 138.9, 115.7, 77.1, 49.0, 40.9, 35.9, 17.6, 14.1, 7.4; IR
(E)-37, 39 (29 mg, 70%, 55:45 dr by H NMR spectroscopy) was
isolated as a colorless liquid after purification by flash chroma-
tography using 20f30% Et2O/pentane. Careful flash chromatog-
raphy using 20f30% Et2O/pentane allows for isolation of a clean
1
sample of anti-39. IR H NMR (300 MHz, CDCl3) δ 5.75 (1H,
ddd, J ) 17.4, 10.2, 9.3 Hz), 5.06-4.97 (2H, m), 4.09 (2H, q, J )
7.2 Hz), 3.41 (1H, dd, J ) 8.7, 3.3 Hz), 3.17 (1H, d, J ) 8.7 Hz),
2.45 (1H, m), 1.28 (3H, s), 1.26 (3H, t, J ) 7.2 Hz), 1.19 (3H, s),
1.10 (3H, d, J ) 6.9 Hz); 13C NMR (75.4 MHz, CDCl3) δ 178.0,
139.2, 115.6, 80.8, 60.6, 45.6, 40.8, 24.1, 22.1, 19.5, 13.9; (neat)
ν ) 3479, 3072, 2979, 2928, 1726, 1261, 1138 cm-1; HRMS-ES
calcd for C11H21O3 [M + H]+ 201.1485, found 201.1490. The
stereochemistry was established by comparison with an authentic
sample prepared (see Supporting Information).
(neat) ν ) 3462, 3073, 2975, 2937, 1708, 1459, 1376, 971 cm-1
;
HRMS-EI calcd for C10H19O2 [M + H]+ 171.1386, found 171.1398.
The stereochemistry was established by comparison of spectral data
above.
(1R*,2R*)-1-Triethylsiloxy-2-vinylcyclohexane (syn-43). Fol-
lowing the general procedure on a 0.20 mmol scale using (E)-42,
Ethyl (3R*,4R*)-4-Methyl-3-triethylsiloxy-5-hexenoate (syn-
and anti-34). Following the general procedure on a 0.20 mmol
scale using (E)-32, the inseparable mixture of diastereomers (50
mg, 87%, 75:25 dr 13C NMR spectroscopy) was isolated as a
colorless liquid after purification by flash chromatography using
3% Et2O/hexane. IR (neat) ν ) 3077, 2858, 2880, 1739, 1462, 1377,
1180, 1085, 1012 cm-1; 1H NMR (300 MHz, CDCl3) δ 5.81 (1H,
m), 5.04 (2H, m), 4.13 (3H, m), 2.38 (3H, m), 1.26 (3H, t, J ) 7.2
Hz), 0.98 (12H, m), 0.60 (6H, q, J ) 7.8 Hz); 13C NMR (100 MHz,
CDCl3) δ 172.2, 140.2, 139.8, 115.3, 114.9, 72.9, 72.4, 60.3, 43.8,
43.7, 40.0, 39.4, 14.9, 14.1, 6.8, 5.0, 4.9; HRMS-EI calcd for C13H25-
SiO3 [M - Et]+ 257.1573, found 257.1572. The relative configu-
ration was unambiguously deduced after compound 34 was reduced
and deprotected (DIBAL, CH2Cl2, -78 °C then 10% HCl). 1H NMR
of the crude mixture showed a 3:1 dr and the chemical shift of the
internal vinyl proton of the minor diastereomer was identical to
the previously reported anti diastereomer,54 thus confirming the syn
configuration for the major diastereomer.
Ethyl (3R*,4R*)-3-Hydroxy-4-methyl-5-hexenoate (syn- and
anti-35). Following the general procedure on a 0.21 mmol scale
using (E)-33, the inseparable mixture of diastereomers (17 mg, 46%,
60:40 dr 1H NMR spectroscopy) was isolated as a colorless liquid
after purification by flash chromatography using 40% Et2O/hexane.
1H NMR (400 MHz, CDCl3) δ 5.85-5.70 (1H, m), 5.12-5.06 (2H,
m), 4.21-4.15 (2H, m), 3.93 (1H, m, anti isomer), 3.87 (1H, m,
syn isomer), 2.92 (1H, d, J ) 4.4 Hz, syn isomer), 2.81 (1H, d,
J ) 3.6 Hz, anti isomer), 2.57-2.26 (3H, m), 1.27 (3H, t, J ) 7.2
Hz), 1.08 (3H, m); 13C NMR (100 MHz, CDCl3) δ 173.3, 173.2,
140.1, 139.6, 116.1, 115.7, 71.2, 71.1, 60.7, 43.3, 43.2, 38.7, 15.7,
15.4, 14.1; IR (neat) ν ) 3498, 3078, 2979, 2933, 1736, 1374,
1183, 1031, 918 cm-1; HRMS-ES calcd for C9H16O3Na [M + Na]+
195.0991, found 195.0982. The stereochemistry was established
by comparison with authentic compound prepared from deprotection
of 34 (pTsOH‚H2O, H2O, THF).
1
the crude product (8:1 dr by H NMR spectroscopy) was purified
by flash chromatography using pentane as eluant to give syn-43
1
(28 mg, 52%, >99:1 dr by H NMR spectroscopy) as a colorless
liquid. 1H NMR (400 MHz, CDCl3) δ 5.92 (1H, m), 4.98 (2H, m),
3.86 (1H, m), 2.10 (1H, m), 1.74-1.59 (4H, m), 1.47-1.22 (4H,
m), 0.95 (9H, t, J ) 8.0 Hz), 0.57 (6H, q, J ) 8.0 Hz); 13C NMR
(100 MHz, CDCl3) δ 141.8, 113.6, 70.9, 46.9, 33.5, 26.7, 24.5,
20.8, 6.9, 5.0; IR (neat) ν ) 3074, 2935, 2877, 1459, 1238, 1100,
1021, 906, 741 cm-1; HRMS-EI calcd for C12H23SiO [M - Et]+
211.1518, found 211.1521. The stereochemistry was established
by deprotection (pTsOH‚H2O, H2O, THF) and comparison of the
1
crude H NMR with known compound.55
(4R*,5R*,6R*)-6-Deuterio-4,6-dimethyl-5-triethylsiloxy-7-
octen-3-one (syn-44). Following the general procedure on a 0.20
mmol scale using (E)-9 and DCO2D, syn-44 (44 mg, 75%, >94:6
dr by 1H NMR spectroscopy, >97% D) was isolated as a colorless
liquid after purification by flash chromatography using 5% Et2O/
1
pentane. H NMR (300 MHz, CDCl3) δ 5.76 (1H, dd, J ) 17.1,
10.5 Hz), 5.05-4.98 (2H, m), 3.96 (1H, d, J ) 5.4 Hz), 2.66 (1H,
m), 2.48 (2H, q, J ) 7.8 Hz), 1.09 (3H, d, J ) 6.9 Hz), 1.03 (3H,
t, J ) 7.8 Hz), 0.97 (3H, s), 0.95 (9H, t, J ) 8.1 Hz), 0.60 (6H, q,
J ) 8.1 Hz); 13C NMR (74.5 MHz, CDCl3) δ 214.0, 141.5, 114.5,
76.2, 49.9, 42.5 (very weak t, J ) 20 Hz), 35.1, 15.4, 12.3, 7.7,
7.0, 5.3; IR (neat) ν ) 3075, 2958, 2871, 1711, 1462, 1098, 739
cm-1; HRMS-ES calcd for C16H31DSiO2 [M + Na]+ 308.2126,
found 308.2142.
(2R,4R,5R,7R,8S,9S)-4,8-Bis-(tert-butyldimethylsiloxy)-3,5,7,9-
tetramethyl-undeca-1,10-dien-6-one (syn-55). Following the gen-
eral protocol (0.025 M) on a 0.304 mmol scale using 54, 55 was
isolated by flash chromatography using 2% Et2O/hexane as a
colorless oil (126 mg, 86%, mixture of diastereomers 83:17 dr).
1
Rf ) 0.85 (90% hexanes:ether); H NMR (400 MHz, CDCl3) δ
5.78 (2H, ddd, J ) 15.0, 9.9, 7.3 Hz), 5.01 (2H, dd, J ) 2.4, 1.3
Hz), 4.97 (2H, dd, J ) 1.8, 1.1 Hz), 3.96 (2H, dd, J ) 4.9, 4.9
Hz), 2.81 (2H, app dq, J ) 7.1, 4.6 Hz), 2.20 (2H, m), 1.09 (6H,
d, J ) 7.1 Hz), 0.98 (6H, d, J ) 6.8 Hz), 0.88 (18H, s), 0.07 (6H,
s), -0.01 (6H, s); 13C NMR (100 MHz, CDCl3) δ 215.0, 141.7,
114.4, 74.9, 48.5, 26.1, 18.4, 15.2, 12.5, -3.8, -4.2; IR (film)
ν ) 3079, 2958, 2886, 2709, 1834, 1706, 1640, 1472, 1414, 1380,
Ethyl (3S*,4R*)-2,2,4-Trimethyl-3-triethylsilyloxy-5-hexenoate
(syn-38). Following the general procedure on a 0.20 mmol scale
using (E)-36 and a 5 mol % catalyst system, syn-38 (51 mg, 81%,
1
>99:1 dr by H NMR spectroscopy) was isolated as a colorless
liquid after purification by flash chromatography using 10% Et2O/
1
pentane. H NMR (400 MHz, CDCl3) δ 5.73 (1H, ddd, J ) 17.2,
10.4, 8.0 Hz), 4.95 (1H, dt, J ) 17.2, 1.6 Hz), 4.90 (1H, ddd, J )
10.4, 1.6, 0.8 Hz), 4.07 (2H, q, J ) 7.2 Hz), 3.95 (1H, d, J ) 4.4
Hz), 2.28 (1H, m), 1.24 (3H, t, J ) 7.2 Hz), 1.17 (3H, s), 1.12
(55) (a) Schlosser, M.; Franzini, L.; Bauer, C.; Leroux, F. Chem. Eur. J.
2001, 7, 1909. (b) Kocovsky, P.; Ahmed, G.; Srogl, J.; Malkov, A. V.;
Steele, J. J. Org. Chem. 1999, 64, 2765.
(54) Drouet, K. E.; Theodorakis, E. A. Chem. Eur. J. 2000, 6, 1987.
1932 J. Org. Chem., Vol. 71, No. 5, 2006

DiastereoselectiVe Palladium-Catalyzed Reduction
1255, 1106 cm-1; HRMS-EI calcd for C23H47O3Si2 [M - C4H7]+
427.3072, found 427.3064.
2.38-2.17 (2H, m), 1.14 (3H, d, J ) 1.1 Hz), 1.04-0.99 (9H, m),
0.87 (9H, s), 0.86 (9H, s), 0.10 (3H, s), 0.05 (3H, s), -0.03 (3H,
s), -0.05 (3H, s); 13C NMR (75 MHz, CDCl3) δ 214.0, 142.2,
141.9, 114.2, 76.8, 73.8, 49.8, 49.0, 43.2, 41.2, 26.2, 26.1, 18.5,
18.2, 15.9, 14.6, 12.5, 11.3, -3.8, -4.0, -4.3; IR (film) ν ) 3078,
(2S,4S,5R,7S,8R,9R)-4,8-Bis-(tert-butyldimethylsiloxy)-3,5,7,9-
tetramethyl-undeca-1,10-dien-6-one (syn-57). Following the gen-
eral protocol (0.025 M) on a 0.16 mmol scale using 56, 57 was
isolated by flash chromatography using 2% Et2O/ hexane as a
colorless oil (70 mg, 89%, <5% of minor diastereomer by 1H and
2957, 2886, 2858, 1711, 1640, 1472, 1413, 1377, 1254, 1104 cm-1
;
HRMS-EI calcd for C23H47O3Si2 [M - C4H7]+ 427.3072, found
427.3064.
1
13C NMR). Rf ) 0.85 (90% hexanes:ether); H NMR (400 MHz,
CDCl3) δ 5.75 (2H, ddd, J ) 17.4, 10.3, 7.7 Hz), 4.99 (4H, m),
3.94 (2H, dd, J ) 5.3, 5.3 Hz), 2.80 (2H, app dq, J ) 7.1, 1.5 Hz),
2.34 (2H, app dq, J ) 7.0, 6.8 Hz), 1.14 (6H, d, J ) 7.1 Hz), 1.02
(6H, d, J ) 6.8 Hz), 0.89 (18H, s), 0.07 (6H, s), 0.02 (6H, s); 13C
NMR (75 MHz, CDCl3) δ 214.9, 142.4, 114.5, 77.7, 50.4, 41.9,
26.3, 18.5, 15.9, 12.6, -3.8, -4.1; IR (film) ν ) 3077, 2957, 2858,
Acknowledgment. We gratefully acknowledge Merck Frosst
Canada and the Natural Sciences and Engineering Research
Council (NSERC) for an Industrial Research Chair and the
University of Toronto for financial support of this work. A.C.
thanks NSERC and the University of Toronto for funding in
the form of postgraduate scholarships. J.-F. P. thanks NSERC,
the Fonds que´be´cois de la recherche sur la nature et les
technologies (FQRNT), the Walter C. Sumner Foundation, and
the University of Toronto for postgraduate scholarships.
2709, 1834, 1700, 1640, 1472, 1413, 1377, 1255, 1078 cm-1
;
HRMS-EI calcd for C23H47O3Si2 [M - C4H7]+ 427.3072, found
427.3064.
(2S,4S,5R,7S,8S,9S)-4,8-Bis-(tert-butyldimethylsiloxy)-3,5,7,9-
tetramethyl-u ndeca-1,10-dien-6-one (syn-59). Following the
general protocol (0.025 M) on a 0.14 mmol scale using 58, 59 was
isolated by flash chromatography using 2% Et2O/ hexane as a
Supporting Information Available: Experimental procedures,
1
isolation and spectroscopic information, and H and 13C spectra
1
colorless oil (53 mg, 81%, 85:15 dr by H and 13C NMR). Rf )
for the new compounds prepared. This material is available free of
charge via the Internet at http://pubs.acs.org.
0.85 (90% hexanes:ether); 1H NMR (300 MHz, CDCl3) δ 5.76 (2H,
m), 5.48 (4H, m), 4.07 (1H, dd, J ) 6.1, 3.2 Hz), 3.97 (1H, dd,
J ) 6.6, 4.1 Hz), 2.88 (1H, m), 2.65 (1H, dq, J ) 7.2, 3.1 Hz),
JO052267S
J. Org. Chem, Vol. 71, No. 5, 2006 1933