An enantioselective synthesis of benzylidene-protected syn-3,5-dihydroxy carboxylate esters via osmium, palladium, and base catalysis

Article abstract of DOI:10.1021/ol0156188

(matrix presented) The enantioselective syntheses of several protected syn-3,5-dihydroxy carboxylic esters have been achieved from the corresponding achiral 1,3-dieneoates The route relies upon an enantio- and regioselective Sharpless dihydroxylation and

Full text of DOI:10.1021/ol0156188

ORGANIC
LETTERS
2001
Vol. 3, No. 7
1049-1052
An Enantioselective Synthesis of
Benzylidene-Protected syn-3,5-Dihydroxy
Carboxylate Esters via Osmium,
Palladium, and Base Catalysis
Thomas J. Hunter and George A. O’Doherty*
Department of Chemistry, UniVersity of Minnesota, Minneapolis, Minnesota 55455
odoherty@chem.umn.edu
Received January 25, 2001
ABSTRACT
The enantioselective syntheses of several protected syn-3,5-dihydroxy carboxylic esters have been achieved from the corresponding achiral
1,3-dieneoates. The route relies upon an enantio- and regioselective Sharpless dihydroxylation and a palladium-catalyzed reduction to form
δ-hydroxy-1-enoates. The resulting δ-hydroxy-1-enoates are subsequently converted into benzylidene-protected 3,5-dihydroxy carboxylic esters
in one step. The benzylidene-protected 3,5-dihydroxy carboxylic esters are produced in good overall yields (25% to 51%) and high enantiomeric
excesses (80% to >95%).
In the course of a project aimed at a total synthesis of the
potent antifungal and cytotoxic macrolide leucascandrolide
A,1,2 we were interested in a concise route to protected syn-
3,5,7-trihydroxyheptenoic acid ester 2, a potential differenti-
ated synthon of diol 1 (Scheme 1). We envisioned that a
suitably protected and desymmetrized variant of 1 could
become the C-5 through C-11 fragment of leucascandrolide
A. In addition, we planned for the introduction of the C-5,
C-11, and C-12 stereocenters of leucascandrolide A by two
sequential diastereoselective anion additions: allyl at C-5
and then crotyl at C-11.3 Thus, we decided to devise an
efficient and enantioselective route to 2 which would derive
the asymmetry via a catalytic enantioselective asymmetric
reaction.4
Naturally, synthons of 1 would be of general use for the
synthesis of many 1,3-polyol containing natural products.5
Scheme 1
(1) Ambrosio et al. have shown leucacandrolide to have significant
activity against Candida albicans and in vitro IC50 values of 0.05 and 0.25
µg/mL against KB and P388 cell lines, respectively, see: Ambrosio, M.
D.; Guerriero, A.; Debitus, C.; Pietra, F. HelV. Chim. Acta 1996, 79, 51-
59.
(2) Leucascandrolide A has recently been synthesized by the Leighton
group, see: (a) Hornberger, K. R.; Hamblett, C. L.; Leighton, J. L. J. Am.
Chem. Soc. 2000, 122, 12894. For another approach, see: (b) Crimmins,
M. T.; Carroll, C. A.; King, B. W. Org. Lett. 2000, 2, 597-599.
(3) Leighton has demonstrated that this can be accomplished with allyl-
(-)-diisopinocamphenylborane and (E)-crotyl-(-)-diisopinocamphenyl-
borane.
(4) For a nine-step Sharpless asymmetric epoxidation route to a derivative
of 2, see: (a) Miyazawa, M.; Matsuoka, E.; Sasaki, S.; Oonuma, S.;
Maruyam, K. Miyashita, M. Chem. Lett. 1998, 109-110. For a nine-step
resolution approach, see: (b) Evans, D. A.; Trotter, B. W.; Coleman, P. J.;
Cote, B.; Dias, L. C.; Rajpakse, H. A.; Tyler, A. N. Tetrahedron 1999,
55(29), 8671-8726.
10.1021/ol0156188 CCC: $20.00 © 2001 American Chemical Society
Published on Web 03/06/2001

The syn-1,3-diol structural unit is a common motif in many
natural products with a wide range of biological activities.6
Due to the wide distribution of the syn-1,3-polyol motif in
natural products, many synthetic methodologies have been
employed to synthesize this core structure. These range from
the use of stoichiometric chiral reagents such as enolates7
and allyl anions8 to the more elegant use of catalytic
reagents.9 Perhaps the most successful application of enan-
tioselective catalysis is Rychnovsky’s5,6 use of the Noyori
hydrogenation.10 More recently, Leighton has developed two
carbonylation approaches to 1,3-syn-diols of type 1.11
Recently, we developed several effective routes to various
carbohydrate-based natural products using the Sharpless
dihydroxylation and aminohydroxylation to establish the
absolute stereochemistry.12 Continuing our investigations on
the utility of these transformations, we decided to investigate
its merit toward the synthesis of the syn-1,3-diol structural
motif. We envisioned establishing 3,5-dihydroxy carboxylic
esters from δ-hydroxy-1-enoates 4, which Evans has shown
can be converted into a benzylidene-protected 3,5-dihydroxy
carboxylic ester 3 in a single step (Scheme 2).13
of a 3,4-dihydroxy-1-enoate, which in turn could be prepared
via an asymmetric dihydroxylation of a dienoate such as 6.15
Miyashita has developed a similar reduction strategy to
prepare δ-hydroxy-1-enoates from γ,δ-epoxy acrylates.4a
Herein, we describe our approach to the synthesis of these
key building blocks via an efficient asymmetric and dia-
stereoselective reaction sequence.
To test the feasibility of this sequence, we decided to start
with commercially available ethyl sorbate (6a). Following
Sharpless’s protocol for the dihydroxylation of 6a, diol 5a
was provided in good yields and enantiomeric excesses
(Scheme 3).16,17 Either enantiomer of diol 5a was obtained
Scheme 3
Scheme 2
with enantiomeric excesses on the order of 80% from the
(DHQ)2PHAL ligand system and > 90% from the (DHQD)2-
PHAL ligand.
Diol 5a was conveniently converted into bis-benzoate 7
or bis-ethyl carbonate 8 by treatment with benzoyl chloride
or ethyl chloroformate (57 and 81% yields, respectively).
At this stage the two functional groups were readily
differentiated by taking advantage of the fact that allyl
benzoates and carbonates are good leaving groups for the
formation of π-allyl palladium complexes.18
Thus, treatment of 7 with a catalytic amount of a
palladium(0) source and triphenylphosphine (2.5% Pd2(dba)3‚
CHCl3/6.3% PPh3) and a mild hydride source (3 equiv, Et3N/
HCO2H) gave the reduced product 9 in low yield (43%)
Thus, the problem was reduced to an efficient asymmetric
synthesis of δ-hydroxy-1-enoates.14 We hoped these δ-hy-
droxy-1-enoates could be prepared by the selective reduction
(5) Rychnovsky, S. D. Chem. ReV. 1995, 95, 2021-2040.
(6) (a) Rychnovsky, S. D.; Hoye, R. C. J. Am. Chem. Soc. 1994, 116,
1753-1765. (b) Rychnovsky, S. D.; Khire, U. R.; Yang, G. J. Am. Chem.
Soc. 1997, 119, 2058-2059.
(7) (a) Evans, D. A. Aldrichimica Acta. 1982, 2, 23. (b) Evans, D. A.;
Nelson, J. V.; Taber, T. R. Top. Stereochem. 1982, 13, 1.
(8) For good reviews see: (a) Hoffmann, R. W. Pure Appl. Chem. 1988,
60(1), 123-30. (b)Yamamoto, Y.; Asao, N. Chem. ReV. 1993, 93, 2207-
2293.
(9) (a) Carreira, E. M.; Singer, R. A.; Lee, W. J. Am. Chem. Soc. 1994,
116, 8837-8838. (b) Evans, D. A.; Murry, J. A.; Kozlowski, M. C. J. Am.
Chem. Soc. 1996, 118, 5814-5815. (c) Evans, D. A.; Coleman, P. J.; Cote,
B. J. Org. Chem. 1997, 62, 788-789. (d) Paterson, I.; Oballa, R. M.;
Norcross, R. D. Tetrahedron Lett. 1996, 37, 8581-8584. (e) Paterson, I.;
Gibson, K. R.; Oballa, R. M. Tetrahedron Lett. 1996, 37, 8585-8588.
(10) Noyori, R.; Hashiguchi, S. Acc. Chem. Res. 1997, 30, 97-102.
(11) (a) Leighton, J. L.; O’Neil, D. N. J. Am. Chem. Soc. 1997, 119,
11118-11119. (b) Sarraf, S. T.; Leighton, J. L. Org. Lett. 2000, 2, 3205-
3208. (c) Leighton, J. L.; Chapman, E. J. Am. Chem. Soc. 1997, 119,
12416-12417.
(14) For vinylogous aldol approaches to δ-hydroxy-1-enoates, see: (a)
Fleming, I. Bull. Soc. Chem. Fr. 1981, 2, 7-13. (b) Barloy-Da Silva, C.;
Benkouider, A.; Pale, P. Tetrahedron Lett. 2000, 41, 3077-3081. (c)
Albaugh-Robertson, P.; Katzenellenbogen, J. A. J. Org. Chem. 1983, 48,
5288-302. For aldol/Wittig approaches, see: ref 4 and (d) Keck, G. E.;
Palani, A.; McHardy, S. F. J. Org. Chem. 1994, 59, 3113. (e) Solladie, G.;
Gressot, L.; Colobert, F. Eur. J. Org. Chem. 2000, 357-364.
(15) Previously Sharpless had shown that the AD mix reagent dihy-
droxylates simple dienoates 6a and 6c with good enantio- and diastereo-
control, see: (a) Xu, D.; Crispino, G. A.; Sharpless, K. B. J. Am. Chem.
Soc. 1992, 114, 7570-7571. (b) Becker, H.; Soler, M. A.; Sharpless, K. B.
Tetrahedron 1995, 51, 1345-76.
(12) (a) Harris, J. M.; O’Doherty, G. A. Org. Lett. 2000, 2, 2983-2986.
(b) Harris, J. M.; Keranen, M. D.; Nguyen, H.; Young, V. G.; O’Doherty,
G. A. Carbohydr. Res. 2000, 328, 17-36. (c) Balachari, D.; O’Doherty,
G. A. Org. Lett. 2000, 2, 863-866. (d) Harris, J. M.; Keranen, M. D.;
O’Doherty, G. A. J. Org. Chem. 1999, 64, 2982-2983. (e) Haukaas, M.
H.; O’Doherty, G. A. Org. Lett. 2001, 3, 401-404.
(16) All levels of enantioinduction were determined by HPLC analysis
(8% IPA/Hexane, Chiralcel OD) and/or Mosher ester analysis. (a) Sullivan,
G. R.; Dale, J. A.; Mosher, H. S. J. Org. Chem. 1973, 38, 2143. (b)
Yamaguchi, S.; Yasuhara, F.; Kabuto, K. T. Tetrahedron 1976, 32, 1363.
(17) All new compounds were identified and characterized by 1H NMR,
13C NMR, FTIR, HRMS, and/or elemental analysis.
(13) See: ref 4b and Evans, D. A. Gauchet-Prunet, J. A. J. Org. Chem.
1993, 58, 2446-2453.
(18) (a) Tsuji, J.; Minami, I. Acc. Chem. Res. 1987, 20, 140. (b) Hughes,
G.; Lautens, M.; Wen, C. Org. Lett. 2000, 2, 107-110.
1050
Org. Lett., Vol. 3, No. 7, 2001

and a mild hydride source (Et3N/HCO2H) provided the
desired δ-hydroxy ester 4a in good yield (70%), with no
loss of enantiomeric excess (Scheme 6). Varying the
Scheme 4
Scheme 6
(Scheme 4). More rigorous examination of the reaction
mixture led to the identification of less than 10% of ethyl
sorbate 6a. Unfortunately, varying the substrate to the bis-
ethyl carbonate 8 did not change the yield of reduced product
10 (47%) or the relative amount of ethyl sorbate (4:1). To
determine if the ethyl sorbate was being formed from an
elimination reaction, both palladium reaction products were
reexposed to the reaction conditions without palladium
present. Both enoates 9 and 10 were completely stable to
refluxing THF solutions of Et3N and Et3N/HCO2H. No ethyl
sorbate was detected even as the reaction was concentrated
by solvent distillation, indicating that ethyl sorbate was a
direct product from the palladium reaction conditions.19
The formation of ethyl sorbate must have occurred via an
alternative reaction pathway available to the π-allyl palladium
intermediate 11. Presumably 11, in addition to the normal
hydrogen migration to the γ-carbon, can undergo a loss of
a proton and concomitant â-elimination of the C-5 ethyl
carbonate or benzoate to give palladium-bound ethyl sorbate
13 (path B, Scheme 5).20 To minimize the â-elimination side
substitution on the dienoate from methyl to propyl, 6b, or
phenyl, 6c, had no detrimental effects on the yields of this
reaction sequence and inevitably improved the enantiomeric
excess of the asymmetric dihydroxylation. Both diols 5b and
5c were formed from 6b and 6c in 80% and 79% yields and
greater than 95% enantiomeric excesses. Similarly, 5b and
5c were converted into cyclic carbonates 14b (94%) and 14c
(91%) and then reduced upon exposure to the palladium
procedure to form the homoallylic alcohols 4b and 4c (80%
and 72% yields, respectively). The methyl-, propyl-, and
phenyl-substituted dienoates 6a-c were converted into the
corresponding δ-hydroxy enoates in good overall yields (47-
67%) in three steps and with no loss of enantiomeric excess
(80-95% ee).
Scheme 5
Having established a general procedure for the enantio-
selective synthesis of homoallylic alcohols 4, we turned our
attention to the base-catalyzed acetal/conjugated addition
reaction. Exposing the three δ-hydroxy enoates 4a-c to the
Evans procedure (1.1 equiv of benzaldehyde and 0.1 equiv
of KOt-Bu, 0 °C, repeat 3-4 times every 15 min)13 led to
good yields of the benzylidene-protected 3,5-dihydroxy
carboxylic esters 3a-c (Scheme 7). Three stereochemically
Scheme 7
reaction, we decided to make the C-5 substituent a poorer
leaving group (i.e., negatively charged). This was easily
accomplished by activating the allylic alcohol functionality
as a cyclic carbonate. The cyclic carbonate 14a was prepared
by treating a pyridine/CH2Cl2 solution of diol 5a with
triphosgene, providing 14a in excellent yield (87%). Treat-
ment of 14a with a catalytic amount of palladium/triphenyl-
phosphine (2.5% Pd2(dba)3‚CHCl3 or (allylPdCl)2/6.3% PPh3)
enriched esters 3a, 3b, and 3c were formed as a single
diastereomer (>95%) and in good yields (60%, 60%, 50%,
respectively).
With the success of the model system, we decided to test
the compatibility of the reaction sequence with the oxygen-
(19) Furthermore, the ratios of 9/6a and 10/6a for both reactions were
equivalent at both low and high conversion.
(20) These results do not preclude the mechanistic possibility for the
formation of ethyl sorbate via a â,γ-unsaturated regioisomer of 9 and 10
and subsequent base- and/or palladium-catalyzed elimination.
Org. Lett., Vol. 3, No. 7, 2001
1051

ated dienoates 16a and 16b. These dieneoates were readily
prepared by a PPh3/PhOH-catalyzed deconjugation of ynoates
15a and 15b (80% and 95% yields, respectively).21 Both
ynoates 15a and 15b were readily prepared from 5-hexyn-
1-ol and 4-pentyn-1-ol in two steps, in 93% and 97% overall
yields, respectively (TBSCl, Et3N, 0 °C; n-BuLi then ClCO2-
CH2CH3, 0 °C). The exposure of 16a and 16b to the
dihydroxylation conditions gave diols 17a and 17b in 82%
and 91% yields and ∼90% and >95% ee, respectively
(Scheme 8). These diols were subsequently converted into
benzylidene acetal formation reaction. Exposure of the
hexenoate 19a to the Evans’ procedure resulted in a TBS-
group migration/cyclization reaction, yielding the tetra-
hydrofuran product 20 as a mixture of diastereoisomers (88%
yield). However, exposure of the PMB-protected variant 19c
to the same conditions led to a clean conversion into the
desired benzylidene acetal 21 in a 61% yield. In contrast,
the homologous TBS-protected substrate 19b did not undergo
a similar silyl-group migration and subsequent rearrangement
under the same conditions. Treatment of 19b to the sequential
KOt-Bu/benzaldehyde procedure22 yielded the benzylidene-
protected syn-1,3-diol 2 in good yield (68%) (Scheme 9).
Scheme 8
Scheme 9
In conclusion, this highly enantio- and diastereocontrolled
route to benzylidene-protected syn-1,3-diols illustrates the
utility of the asymmetric dihydroxylation/palladium-catalyzed
reduction reaction sequence. This methodology provides
rapid and enantioselective access to densely functionalized
molecules starting from commercially available and inex-
pensive starting materials, as well as, providing an alternative
to vinylogous aldol strategies. Further studies on the use of
these chiral building blocks toward the synthesis syn-1,3-
polyol-containing natural products will be reported in due
course.
carbonates 18a (94%) and 18b (95%) upon treatment of a
CH2Cl2/pyridine solution with triphosgene. Similarly, the
carbonates 18a and 18b were reductively opened to form
the homoallylic alcohols 19a (88%) and 19b (87%) in good
yields. Because of the instability of the primary TBS group
in later sequences, the TBS group of 18a was converted into
a PMB group by TBS deprotection (5% HF/AN, 93% yield)
and PMB protection (1% CSA, Cl3C(NH)OPMB, 61% yield)
forming carbonate 18c in a 57% overall yield. Similarly,
cyclic carbonate 18c underwent palladium-catalyzed reduc-
tion to give the homoallylic alcohol 19c, but with a slightly
lower efficiency (66%).
Acknowledgment. We thank the Arnold and Mabel
Beckman Foundation for their generous support of our
program.
With access to the oxygenated substrates 19a-c, we were
ready to address the formation of the desired synthon 2.
Unfortunately, the vicinal tert-butyldimethylsiloxy group in
19a was incompatible with the basic conditions of the
Supporting Information Available: Complete experi-
mental procedures and spectral data for all new compounds.
This material is available free of charge via the Internet at
http://pubs.acs.org.
OL0156188
(21) We found the Rychnovsky phenol variant of the Trost procedure
was the best for these substrates. (a) Rychnovsky, S. D.; Kim, J. J. Org.
Chem. 1994, 59, 2659-60. For the original Trost procedure, see: (b) Trost,
B.; Kazmaier, U. J. Am. Chem. Soc. 1992, 114, 7933-35.
(22) Benzaldehyde (1.1 equiv) with 0.1 equiv of KOt-Bu; after 15 min
more benzaldehyde (1.1 equiv) and KOt-Bu (0.1 equiv).
1052
Org. Lett., Vol. 3, No. 7, 2001