Diastereoselective Palladium-Catalyzed Formate Reduction of Allylic Carbonates as a New Entry into Propionate Units

Article abstract of DOI:10.1021/ol0350238

(Equation presented) The diastereoselective palladium-catalyzed formate reduction of allylic carbonates is described. Reduction of allylic carbonates under mild conditions (Pd(OAc)₂ (2.5-5 mol %), [n-Bu ₃PH]BF₄ (2.5-5 mol

Full text of DOI:10.1021/ol0350238

ORGANIC
LETTERS
2003
Vol. 5, No. 19
3391-3394
Diastereoselective Palladium-Catalyzed
Formate Reduction of Allylic Carbonates
as a New Entry into Propionate Units
Mark Lautens* and Jean-Franc¸ois Paquin
DaVenport Chemical Laboratories, Department of Chemistry, UniVersity of Toronto,
80 St. George Street, Toronto, Ontario, Canada M5H 3H6
mlautens@chem.utoronto.ca
Received June 6, 2003
ABSTRACT
The diastereoselective palladium-catalyzed formate reduction of allylic carbonates is described. Reduction of allylic carbonates under mild
conditions (Pd(OAc)2 (2.5−5 mol %), [n-Bu3PH]BF4 (2.5−5 mol %), HCO H/Et3N (1:2) (3 equiv), CH CN (0.05M), 40 °C) affords the terminal olefin
2
3
as the syn isomer in good yields and modest to excellent diastereoselectivity. These compounds, which are useful building blocks for the
synthesis of polypropionate units, are the synthetic equivalent of the products obtained from an aldol reaction of an r-methyl-â,γ-unsaturated
aldehyde.
We recently reported the use of γ-carboxy-R,â-unsaturated
aldehyde 2 as synthetic equivalent of â,γ-unsaturated alde-
hyde 5 (Figure 1).1 In this strategy, the allylic ester 3 is
obtained from an aldol reaction between the enol ether 1
and aldehyde 2 followed by protection of the alcohol
functionality. The key step of this strategy is the palladium-
catalyzed formate reduction2 of 3 that affords 4 in good
yields. Product 4b is of interest to us, since a wide variety
of natural products contain propionate fragments3,4 and 4b
is a versatile building block that possesses functional groups
useful for further elaboration. Furthermore, direct access by
(1) Hughes, G.; Lautens, M.; Wen, C. Org. Lett. 2000, 2, 107.
(2) (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.
(3) Davies-Coleman, M. T.; Garson, M. J. Nat. Prod. Rep. 1998, 15,
477.
(4) For recent syntheses of propionates containing natural products in
our laboratory, see (a) Lautens, M.; Colucci, J. T.; Hiebert, S.; Smith, N.
Figure 1. Aldol/Pd-Catalyzed Formate Reduction Strategy
D.; Bouchain, G. Org. Lett. 2002, 4, 1879. (b) Lautens, M.; Stammers, T.
A. Synthesis 2002, 1993.
10.1021/ol0350238 CCC: $25.00 © 2003 American Chemical Society
Published on Web 08/27/2003

a direct aldol between 1 and R-methyl-â,γ-unsaturated
aldehyde 5b is complicated since the enantiomerically pure
aldehyde has not been reported and the diastereoselectivity
with the racemic aldehyde is modest.5
and [n-Bu3PH]BF4 as the palladium and phosphine sources,
respectively.11-13 These changes allowed us to control the
number and the nature of the ligands on the Pd center,14 as
well as avoid the use of air-sensitive n-Bu3P since the
tetrafluoroborate salt is an air-stable precursor of the phos-
phine. Under the modified conditions, using 10 mol % Pd-
(OAc)2, 10 mol % [n-Bu3PH]BF4, and 3 equiv of HCO2H/
Et3N15 (1:2) at 40 °C in CH3CN gave, after 5 min, complete
conversion with less than 3% of the internal olefin 8 in a dr
of 13:1. A summary of the optimization studies on 6b is
presented in Table 1.16 Using 20 mol % phosphine led to
lower conversion with a dr of 5:1 along with 17% of 8. The
use of a phosphine is required since no reaction was observed
in its absence (entry 3), and a 1:1 ratio of Pd:P was used
throughout the rest of the study. The conversion and the dr
were identical between room temperature and 40 °C.
However, running the reaction at 0-5 °C led to incomplete
conversion but an excellent dr (entry 5). Diluting the reaction
had a beneficial effect on the diastereoselectivity of the
reaction since a ratio of >15:1 was obtained (entry 6). The
palladium loading could be decreased to 2.5 mol % (entry
6-8), causing the reaction time to increase from 1 to 3 h.
When the scale was increased, it was found that for a
palladium loading of 2.5 mol % running the reaction at 40
°C was required in order to obtain complete conversion in a
reasonable time,17 although no effect on the dr was observed.
Optimized conditions [Pd(OAc)2 (2.5 mol %), [n-Bu3PH]-
BF4 (2.5 mol %), HCO2H/Et3N (1:2) (3 equiv), CH3CN (0.05
M), 40 °C] were used for the rest of the study. This study
represents the first use of phosphonium salts as a precursor
of phosphines in allylic substitution reactions.
Early experiments indicated the potential of our strategy
since reduction of 6a gave the desired product 7 in good
yield (75%) and moderate selectivity (5.7:1 dr) in favor of
the syn isomer along with 10% of the internal olefin 8 (eq
1).1 In this case, an R-substituted-γ-carboxy-R,â-unsaturated
aldehyde is used as a synthetic equivalent of an R-substituted-
â,γ-unsaturated aldehyde, which are rarely used in total
syntheses6 primarily because of olefin isomerization.7
Even though this strategy allows access to products of
broad interest, to be synthetically useful, higher selectivity
is needed as well as a determination of the scope of the
reaction. We have improved the diastereoselectivity (up to
>20:1 dr) by optimizing the reaction conditions, studied the
influence of protecting group, leaving group, olefin geometry,
and nearby substitution, and report the results of our
investigation herein.
The effects of the leaving group and protecting group were
also examined (Table 2).18 Under the optimized conditions,
the model substrate 6b gave 84% yield with a dr of >15:1
with less than 3% of 8. The substrate bearing a benzyl
carbonate (9) gave identical results.19 Using a bulkier TBS
group allowed the reaction to proceed in similar yield and
diastereoselectivity (entry 3). When R ) Ac, the desired
product 14 was obtained in lower yield with a low dr (entry
Initial efforts to optimize the original system (Pd2(dba)3
(5 mol %), n-Bu3P (20 mol %), HCO2NH4 (3 equiv), DMF,
50 °C) using 6a8 were unsuccessful. We turned our attention
to a more reactive leaving group hoping to improve the
diastereoselectivity by using milder conditions. Carbonates
are known to be an excellent leaving group in π-allyl Pd
chemistry,9 and applying the original conditions to 6b gave
incomplete conversion with a disappointing dr of 3:1.
Examination of different media10 showed that acetonitrile
was the solvent of choice and gave 7 with good conversion
with dr > 10:1 at 40 °C or room temperature.
(11) Mandai, T.; Matsumoto, T.; Tsuji, J.; Saito, S. Tetrahedron Lett.
1993, 34, 2513.
(12) Netherton, M. R.; Fu, G. C. Org. Lett. 2001, 3, 4295. The free
phosphine is liberated by the action of a base, Et3N in our case.
(13) Phosphonium salt can be easily synthesized (see ref 12) or,
alternatively, is available from Strem Chemicals.
To make the reaction simpler to run as well as resolve
some variability problems, we decided to employ Pd(OAc)2
(14) It was demonstrated that the dba ligands are not fully displaced by
n-Bu3P from Pd2(dba)3 (see ref 11) or by Ph3P from Pd(dba)2; see: Amatore,
C.; Jutand, A. Coord. Chem. ReV. 1998, 178-180, 511. Amatore, C.; Jutand,
A.; Khalil, F.; M’Barki, M. A.; Mottier, L. Organometallics 1993, 12, 3168.
(15) Formate salts (HCO2K, HCO2Cs) were not as effective, probably
due to their low solubility.
(5) (a) Roush, W. R. J. Org. Chem. 1991, 56, 4151. (b) Ahmar, M.;
Bloch, R.; Mandville, G.; Romain, I. Tetrahedron Lett. 1992, 33, 2501.
(6) For relevant examples of the use of R-substituted-â,γ-unsaturated
aldehyde in total synthesis, see: (a) Liu, P.; Jacobsen, E. N. J. Am. Chem.
Soc. 2001, 123, 10772. (b) Salamonczyk, G. M.; Han, K.; Guo, Z.-W.;
Sih, C. J. J. Org. Chem. 1996, 61, 6893. (c) Evans, D. A.; DiMare, M. J.
Am. Chem. Soc. 1986, 108, 2476.
(7) For a Lewis-acid-mediated approach to â,γ-unsaturated aldehydes
and their reactions in situ, see: (a) Lautens, M.; Maddess, M. L.; Sauer, E.
L. O.; Ouellet, S. G. Org. Lett. 2002, 4, 83. (b) Lautens, M.; Ouellet, S. G.;
Raeppel, S. Angew. Chem., Int. Ed. 2000, 39, 4079.
(16) The dr (ratio of syn/anti) of all reactions was estimated by 1H NMR
spectroscopy. For the purpose of this study, >15:1 dr indicates that the
other diastereomer was present in less than 6% (typically 4-6%), whereas
>20:1 dr specifies that the other diastereomer was not detectable.
(17) 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.
(8) All substrates used in this study are racemic.
(9) (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.
(18) Stereochemistry of the products was determined by comparison with
authentic samples or similar or previously reported compounds; see
Supporting Information for details.
(19) Under the optimized conditions, other leaving groups such as acetate
or formate (see ref 1c) gave incomplete conversion and no reaction,
respectively.
(10) Other solvents, including THF, DMF, or a combination of the two,
gave lower conversion with low dr.
3392
Org. Lett., Vol. 5, No. 19, 2003

Table 1. Selected Results for the Optimization of the Reduction of Allylic Carbonates
conversionb
(%)
Pd loading
(mol %)
concentration
(M)
entrya
Pd:P
T °C
drc
7
8
6b
1
2
3
4
5d
6
7
8
10
10
10
10
10
10
5
1:1
1:2
1:0
1:1
1:1
1:1
1:1
1:1
0.1
0.1
0.1
0.1
40
40
40
20
0-5
20
20
20
97
83
0
97
50
96
96
97
<3
17
0
<3
<3
∼4
∼4
<3
0
13:1
5:1
0
100
0
47
0
13:1
>15:1
>15:1
>15:1
>15:1
0.1
0.05
0.05
0.05
0
0
2.5
1
a Scale ) 0.1 mmol. b Estimated by H NMR spectroscopy of the crude mixture after workup. c See ref 16. d Reaction stopped after 8 h.
4) but with selective ionization of the carbonate over the
acetate. The substrate having a free hydroxyl group was
reduced under the reaction conditions although in lower yield
and a modest dr (entry 5).20 Surprisingly, the reduction of
6a under the originally reported conditions gave a 1:1 mixture
of diastereomers, whereas a slight preference for the syn
isomer is observed under our optimized system. The protect-
ing groups play an important role in influencing the diaste-
reoselectivity, but the cause of this behavior is not fully
understood at this point. Fortunately, the use of a silyl
protecting group opens the way to the preparation of the
syn-syn triad in synthetically useful levels of diastereo-
selectivity.
with a phenyl had little effect on the reduction since the
selectivity was >15:1. In addition, the minor diastereomer
could be removed by flash chromatography to give 17 in
Scheme 1. Effect of Substitution
Table 2. Effect of the Leaving Group and Protecting Group
entry
substrate
product
yielda (%)
drb
1
2
3
4
5
6b
9
10
11
12
7
7
13
14
15
84
84
85
72
60
>15:1
>15:1
>15:1
1.4:1c
2.7:1
86% yield and >20:1 dr. Similarly, the reduction of the anti
diastereomer 18 proceeded in >15:1 selectivity and 19 was
isolated as a single diastereomer in 86% yield, opening the
way to the synthesis of the anti-syn triad.21 The substrate
lacking a methyl substituent R to the ketone (20a) was
a Isolated yield. b See ref 16. c The dr was estimated by 13C NMR
spectroscopy.
(20) Diastereomers are separable, and the pure syn isomer (30%) could
be isolated by flash chromatography.
(21) 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.
The effect of substitution on the substrate was also
evaluated as shown in Scheme 1.18 Replacing the ethyl group
Org. Lett., Vol. 5, No. 19, 2003
3393

reduced in 87% yield but with a modest 3:1 dr,22 whereas
substrate 20b gave >20:1 dr in 81% yield.21 These results
suggest that substituents on the carbon chain play a role
determining the conformation23 of the π-allyl complex
undergoing reduction.
Finally, the geometry of the double bond seems to play
an important role since the reaction of the (Z)-isomer 22 had
to be conducted using 5 mol % catalyst to ensure complete
conversion21 and 7 was isolated in 89% yield with a
diastereoselectivity of 4:1 in favor of the syn isomer (Scheme
2).
In conclusion, we have shown that the diastereoselective
palladium-catalyzed formate reduction of allylic carbonates
provides access to propionate units including the syn-syn
and anti-syn triads. The reaction occurs under mild condi-
tions, and the desired products are obtained in good yields
with modest to excellent diastereoselectivity. Our work also
shows that R-substituted-γ-carboxy-R,â-unsaturated aldehyde
can be used as a synthetic equivalent of R-substituted-â,γ-
unsaturated aldehydes.
Extension of this methodology and application to the
synthesis of natural products, as well as mechanistic studies,
are underway and will be reported in due course.
Acknowledgment. We gratefully acknowledge the fi-
nancial support of the Natural Sciences and Engineering
Research Council (NSERC) of Canada, the Ontario Research
and Development Challenge Fund (ORDCF), and the Uni-
versity of Toronto. J.-F.P. acknowledges NSERC, the Fonds
que´be´cois de la recherche sur la nature et les technologies,
the Walter C. Sumner Foundation, and the University of
Toronto for PGS scholarships. We thank Prof. James D.
White for kindly providing us with NMR spectra of 13.
Scheme 2. Effect of Double-Bond Geometry
Supporting Information Available: Experimental pro-
cedures for the synthesis of the allylic carbonates, general
experimental procedures for the formate reduction, isolation
This work1 represents, to the best of our knowledge, the
first practical example of diastereoselective palladium-
catalyzed formate reduction.24 Previous stereoselective ap-
proaches have been based on the use of chiral phosphine
with achiral25 or racemic26 allylic acetates or carbonates, or
inversion of the allylic stereocenter.2c,27
1
and spectroscopic information, and H and 13C spectra for
the new compounds prepared. This material is available free
of charge via the Internet at http://pubs.acs.org.
OL0350238
Although the precise mechanism for the hydride transfer
has not been clearly established,28 it is apparent from our
work that changes in either the protecting groups or the
substitution pattern of the substrates influence this crucial
step, and studies are currently underway in order to under-
stand this phenomenon.
(25) Hayashi, T.; Iwamura, H.; Naito, M.; Matsumoto, Y.; Uozumi, Y.
J. Am. Chem. Soc. 1994, 116, 775.
(26) Hayashi, T.; Kawatsura, M.; Iwamura, H.; Yamaura, Y.; Uozumi,
Y. J. Chem. Soc., Chem. Commun. 1996, 1767.
(27) (a) Oshima, M.; Yamazaki, H.; Shimizu, I.; Nisar, M.; Tsuji, J. J.
Am. Chem. Soc. 1989, 111, 6280. (b) Shimizu, I.; Aida, F. Chem. Lett.
1988, 601. (c) Shimizu, I.; Oshima, M.; Nisar, M.; Tsuji, J. Chem. Lett.
1986, 1775. (d) Tsuji, J. Shimizu, I.; Minami, I. Chem. Lett. 1984, 1017.
(e) Greenspoon, N.; Keinan, E. J. Org. Chem. 1988, 53, 3723. (f) Ba¨ckwall,
J.-E. Acc. Chem. Res. 1983, 16, 335. (g) Mandai, T.; Matsumto, T.; Kawada,
M.; Tsuji, J. J. Org. Chem. 1992, 57, 1326.
(22) The dr was estimated by 13C NMR spectroscopy
(23) Hoffmann, R. W. Angew. Chem., Int. Ed. 2000, 39, 2054 and
references therein.
(24) 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.
(28) (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.
3394
Org. Lett., Vol. 5, No. 19, 2003