Enantioselective syntheses of colletodiol, colletol, and grahamimycin A.

Article abstract of DOI:10.1021/ol0269502

[reaction: see text] The enantioselective synthesis of colletodiol has been achieved in 11 steps from methyl 1,3,5-octatrienoate and 16 total steps from both ethyl sorbate and methyl 1,3,5-octatrienoate. The route relies upon an enantio- and regioselective Sharpless dihydroxylation and a palladium-catalyzed reduction to form a 5-hydroxy-1-enoate and an 7-hydroxy-1,3-dienoate. These esters were further functionalized, coupled, and macrolactonized to provide colletodiol after deprotection. Grahamimycin A and colletol were synthesized in one and two steps, respectively, from colletodiol.

Full text of DOI:10.1021/ol0269502

ORGANIC
LETTERS
2
002
Vol. 4, No. 25
447-4450
Enantioselective Syntheses of
Colletodiol, Colletol, and
Grahamimycin A
4
†
Thomas J. Hunter and George A. O’Doherty*
Department of Chemistry, West Virginia UniVersity, Morgantown, West Virginia 26506
george.o’doherty@mail.wVu.edu
Received September 23, 2002
ABSTRACT
The enantioselective synthesis of colletodiol has been achieved in 11 steps from methyl 1,3,5-octatrienoate and 16 total steps from both ethyl
sorbate and methyl 1,3,5-octatrienoate. The route relies upon an enantio- and regioselective Sharpless dihydroxylation and a palladium-
catalyzed reduction to form a 5-hydroxy-1-enoate and an 7-hydroxy-1,3-dienoate. These esters were further functionalized, coupled, and
macrolactonized to provide colletodiol after deprotection. Grahamimycin A and colletol were synthesized in one and two steps, respectively,
from colletodiol.
Colletodiol (1) is a 14-membered bis-macrolactone, which
was isolated as a metabolite of the plant fungi Colletotrichum
capsici and Chaetomium funicola in 1966 and 1969,
algae, green algae, and five fungi. The absolute and relative
stereochemistry of grahamimycin A 5 was established by a
1
1
2
combination of X-ray crystalography and total synthesis of
6
respectively. In 1968, MacMillan isolated and proposed
structures for three other 14-membered bis-macrolactones
from C. capsici metabolites colletol 2, colletoallol 3, and
the enantiomer from tartaric acid. Similarly, the absolute
and relative stereochemistry of colletodiol 1 was established
via total synthesis by both Seebach and Mitsunobu, although
both routes suffered from low yields of macrolactonization.
The macrolactonization problems for colletodiol 1 and
3
colletoketol 4 (Figure 1). More recently, two related 14-
membered bis-lactones were isolated from the aerobic
fermentation of cultures of Cytospora sp. ATCC 20502, these
being the structurally isomeric grahamimycin A 4 and
4
grahamimycin A
1
5. It was later realized that the structures
5
of colletoketol and grahamimycin A were identical.
While colletodiol 1 exhibits only mild antibiotic activity,
grahamimycin A 4 (colletoketol) and grahamimycin A 5
1
have shown significant activity against various pathogenic
microorganisms. In particular, grahamimycin A 4 exhibited
potent activity against several species of bacteria, blue-green
†
University of Minnesota.
(1) Grove, J. F.; Speake, R. N.; Ward, G. J. Chem. Soc. C 1966, 230.
(2) Powell, J. W.; Whalley, W. B. J. Chem. Soc. C 1969, 911.
(3) (a) MacMillan, J.; Pryce, R. J. Tetrahedron Lett. 1968, 5497. (b)
MacMillan, J.; Simpson, T. J. J. Chem. Soc., Perkin Trans. 1 1973, 1487-
1
493.
(4) Gurusiddaiah, S.; Ronald, R. C. Antimicrob. Agents Chemother. 1981,
1
9, 153-165.
(
5) (a) O’Neill, J. A.; Simpson, T. J.; Willis, C. L. J. Chem. Soc., Chem.
Commun. 1993, 738-740. (b) Keck, G. E.; Boden, E. P.; Wiley, M. R. J.
Org. Chem. 1989, 54, 896-906.
Figure 1.
1
0.1021/ol0269502 CCC: $22.00 © 2002 American Chemical Society
Published on Web 11/16/2002

colletol 2 were skirted by Keck’s use of DCC/DMAP5b,7
conditions or by Kobayashi’s use of Yamaguchi conditions.
tion coupling of alcohol 7 and hexenoic acid 8. Our
retrosynthetic plan diverges from others in our desire to
establish the absolute and relative stereochemistry of colle-
8
There have been several synthetic efforts toward colleto-
9
10
11
13
diol 1, colletol 2, and grahamimycin A
1
5. Surprisingly,
todiol by a series of selective dihydroxylation reactions.
there has been only one route to the most biologically active
structure, grahamimycin A (colletoketol) 4, by Kobayashi,
who also reconfirmed its stereoselective conversion into
Thus, sequential application of our catalytic Os/Pd-oxidation/
reduction sequence on methyl octa-1,2,3-trienoic acid 9
should allow for the installation of the triol stereochemistry
of hydroxyester 7. A similar approach should allow for the
preparation of the hexenoic acid 8 from ethyl sorbate 10
(Scheme 1).
8
4
colletodiol with NaBH . This reduction was first reported
by Simpson5 and Ronald, although a procedure was not
a
4
reported.
1
2
As part of our efforts for the use of asymmetric catalysis
for the enantioselective synthesis of antimicrobial lactones,
we were interested in an enantioselective synthesis of
colletodiol 1. We targeted colletodiol because of its inter-
mediary oxidation state; thus, we thought 1 would be a
convenient precursor to the reduced isomers colletol 2 and
colletoallol 3, as well as the oxidized isomers grahamimycin
Following our reported protocol, the commercially
1
4
available ethyl sorbate 10 was converted into alcohol 12
in three steps and a 57% overall yield (Scheme 2). The
Scheme 2
1
A 4 and grahamimycin A 5. We were additionally interested
in colletodiol because we thought it was an ideal substrate
to test our catalytic Os/Pd-oxidation/reduction approach to
1
,2- and 1,3-diols.12
Scheme 1
Sharpless dihydroxylation of ethyl sorbate 10 using (DHQD)
PHAL yielded the 4,5-diol in good yield (71%) and enan-
2
-
1
5,16
tiomeric excess (>90%).
The diol hydroxyl groups were
readily differentiated by taking advantage of π-allyl pal-
1
7
ladium chemistry. The diol was converted into cyclic
carbonate 11 by treating a CH Cl /pyridine solution of the
diol with triphosgene (87%). Treatment of 11 with a catalytic
amount of palladium and triphenylphosphine (2.5% Pd
2
2
2
-
1
8
(
dba)
3
‚CHCl
2
3 3 3
/6.3% PPh ) and a mild hydride source (Et N/
Similar to the approaches of others, our retrosynthetic
HCO
H) afforded an excellent yield (92%) of the desired
analysis of colletodiol envisioned a macrolactonization of
hydroxyacid 6, which could be prepared from an esterifica-
δ-hydroxyester 12. The synthesis of the hexenoic acid
fragment 13 was completed by first protection of the
hydroxyl group as the TBS-ether (TBSCl/Et
lowed by ester hydrolysis (LiOH, 88%).
With the TBS-protected hydroxy acid 13 in hand, we next
investigated the viability of this Os/Pd approach for the
enantioselective conversion of trienoate 9 to the 7-hydroxy
dienoate 16 by an identical three-step reaction sequence.
3
N, 96%) fol-
(
(
(
(
6) Seidel, W.; Seebach, D. Tetrahedron Lett. 1982, 23, 159-162.
7) Keck, G. E.; Murry, J. A. J. Org. Chem. 1991, 56, 6606-6611.
8) Kobayashi, Y. Matsuumi, M. J. Org. Chem. 2000, 65, 7221-7224.
9) For the synthesis of colletodiol, see refs 5 and 8 and: (a) Tsutsui,
H.; Mitsunobu, O. Tetrahedron Lett. 1984, 25, 2159-2162. (b) Tsutsui,
H.; Mitsunobu, O. Tetrahedron Lett. 1984, 25, 2163-2166. (c) Schnurren-
berger, P.; Hungerbuhler, E.; Seebach, D. Liebigs Ann. Chem. 1987, 733-
7
44. (d) Fujiji, M.; Omura, T.; Furuyama, M.-A. Shimizu, I. Tennen Yuki
Kagobutsu Toronkai Koen Yoshishu 1997, 376-371.
10) For the synthesis of colletol, see the work of Keck et al., ref 7, and:
a) Shimizu, I.; Omura, T. Chem. Lett. 1993, 1759-1760. (b) Sharma, G.
(
(13) For a related but less efficient approach using the Sharpless
epoxidation, see refs 9d and 10a.
(
V. M.; Raja Rao, A. V. S.; Murthy, V. S. Tetrahedron Lett. 1995, 36, 4117-
(14) The Aldrich Chemical Co. sells ethyl sorbate for $0.30/g.
(15) 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.
4
3
120. (c) Sollaoie, G.; Gressot, L. Colobert, F. Eur. J. Org. Chem. 2000,
57-364.
(
11) For the synthesis of grahamimycin A1, see ref 6 and: (a) Ghiring-
helli, D. Tetrahedron Lett. 1983, 24, 287-290. (b) Hillis, L. R.; Ronald,
R. C. J. Org. Chem. 1985, 50, 470-473. (c) Bestmann, H.-J.; Schobert, R.
Tetrahedron Lett. 1987, 28, 6587-6590. (d) Ohta, K.; Miyagawa, O.;
Tsutsui, H.; Mitsunobu, O. Bull. Chem. Soc. Jpn. 1993, 66, 523-535.
1
(16) All new compounds were identified and characterized by H NMR,
C NMR, FTIR and HRMS.
13
(17) (a) Tsuji, J.; Minami, I. Acc. Chem. Res. 1987, 20, 140. (b) Hughes,
(
12) (a) Hunter, T. J.; O’Doherty, G. A. Org. Lett. 2001, 3, 1049-1052.
b) Hunter, T. J.; O’Doherty, G. A. Org. Lett. 2001, 3, 2777-2780. (c)
Garaas, S.; Hunter, T. J.; O’Doherty, G. A. J. Org. Chem. 2002, 67, 2682-
685.
G.; Lautens, M.; Wen, C. Org. Lett. 2000, 2, 107-110.
(
(18) This lower than normal (2/1 phosphine to palladium) ratio gave
higher yields and faster reaction times, whereas the use of ligandless
conditions lead to the hydrogenation of the C-C double bond.
2
4448
Org. Lett., Vol. 4, No. 25, 2002

While Sharpless has shown the AD-mix reagent system
selectively oxidizes triene 9 to diol 14 with no over
oxidation,19 we were concerned with the possibility of
forming regioisomers and/or double-bond isomers during the
palladium reaction.
Scheme 4
Treatment of triene 9 with asymmetric dihydroxylation
conditions gave diol 14 in an excellent yield (82%) and
enantiomer excess (95% ee).1
5,19,20
The cyclic carbonate was
prepared by treating a CH
2
Cl
2
/pyridine solution of diol 14
with triphosgene, providing 15 in an 88% yield (Scheme 3).
Scheme 3
NaHCO
3
, afforded triol 17.22 Isolation of triol 17 proved to
be difficult; therefore, the crude reaction mixture was filtered
through a pad of Celite, followed by a pad of Florisil, and
directly subjected to acetonide protection conditions to yield
acetonide 7 in modest overall yield over two steps (54%).
A higher yielding and more reproducible procedure
(
Scheme 5) for the preparation of 7 evolved after first
protection of alcohol 16 as a TBS ether to yield dienoate 18
98%). The asymmetric dihydroxylation of 18 using the same
(
2
2
buffered conditions produced TBS-protected triol 19 in
good yield (71%). One-pot protection as the acetonide and
23
removal of the TBS group produced acetonide 7 (88%).
Scheme 5
Treatment of 15 with a catalytic amount of palladium and
triphenylphosphine (1% Pd
2
(dba)
3
‚CHCl
2
3 3
/2.5% PPh ) and a
H) afforded an excellent
mild hydride source (Et N/HCO
3
yield (85%) of the desired δ-hydroxy ester 16 as a 3:1 (trans/
trans to trans/cis) mixture of γ,δ-double bond isomers. This
3:1 mixture of double bond isomers was easily converted
into a 9:1 mixture by exposure to a catalytic amount of I in
2
the presence of light (3500 Å). However, the photochemical
isomerization was not needed when triphenylphosphine was
not used in the palladium-catalyzed reduction. Running the
reaction in the absence of triphenylphosphine (1% Pd
CHCl , 3 equiv of Et N/HCO H), trans,trans-7-hydroxydi-
2 3
(dba) ‚
3
3
2
enoate 16 was produced as a 7:1 mixture of double-bond
isomers in excellent yield (93%).21
With the C-7 hydroxyl and diene functionality stereose-
lectively installed, we next addressed the selective introduc-
tion of the remaining C-4 and C-5 hydroxyl groups by means
of another asymmetric dihydroxylation reaction (Scheme 4).
Treatment of diene 16 with typical asymmetric dihydroxy-
lation conditions, including the addition of 3 equiv of
Having synthesized the two halves of colletodiol, we
turned our attention to the fragment coupling (alcohol 7 plus
carboxylate 8) and the ensuing conversion of 20 to colletodiol
1
(Scheme 6). Alcohol 7 was esterified with acid 8 using
2
,4,6-trichlorobenzoyl chloride in 77% yield. Removal of
(
19) Sharpless has shown that the AD-mix reagent dihydroxylates simple
the TBS group was accomplished using TBAF to yield
compound 20 in 79% yield. The addition of 1 equiv of LiOH
selectively hydrolyzed the methyl ester of 20 yielding
dienoates and trienoates with good enantio- and diastereocontrol; see: (a)
Xu, D.; Crispino, G. A.; Sharpless, K. B. J. Am. Chem. Soc. 1992, 114,
7
1
570-7571. (b) Becker, H.; Soler, M. A.; Sharpless, K. B. Tetrahedron
995, 51, 1345-76.
24,25
hydroxy acid 6 in good yield (79%).
(20) An interesting aspect of this reaction is that the stereochemical
configuration, as well as, the electron withdrawing nature of the 6-hydroxy
group of 14 protect the dienoate against further oxidation to a tetraol. This
is evident from the observation that once the 6-hydroxyl group of 14 was
removed (as in 16) the substrate readily oxidizes (16 to 17) under the AD-
mix conditions.
(22) Sharpless has suggested that buffering the AD-mix reaction with
NaHCO3 can help prevent base-catalyzed side reactions; see: Kolb, H. C.;
VanNieuwenhze, M. S.; Sharpless, K. B. Chem. ReV. 1994, 94, 2483-
2547.
(23) Diol 19 was isolated in a 71% yield along with ∼10% of the
regioisomeric diol product and ∼5% of the recovered trans/cis diene isomer
of 16.
(21) This mixture could be used in the next step without separation of
the double-bond isomers because the trans/cis diene isomer of 16 dihy-
droxylates at a much slower rate, see ref 23.
Org. Lett., Vol. 4, No. 25, 2002
4449

of colletodiol with triphosgene yielded the cyclic carbonate
in excellent yield (90%). The palladium-catalyzed reduction
on this cyclic carbonate produced colletol 2, and, unfortu-
nately, its double bond isomer 22 in a 1:1 mixture (82%
combined yield). On the basis of the results for the
conversion of 15 to 16, the reaction was attempted in the
absence of triphenylphosphine; however, no product was
obtained as the palladium precipitated out of solution. To
determine whether the double bond isomer 22 was formed
from the π-allyl reaction or from a subsequent double bond
isomerization of 1 under the reaction conditions, both 2 and
Scheme 6
2
2 were reexposed to the reaction conditions and no double
bond isomerization occurred. In fact, all attempts to isomerize
the cis-double bond to the trans-double bond (I , hV; PhSH,
2
AIBN) were unsuccessful.
It was similarly envisioned that grahamimycin A 4 could
be synthesized by a selective oxidation of the allylic alcohol
of colletodiol 1 (Scheme 8). Attempts at this selective
oxidation using MnO , DDQ, and Dess-Martin periodinane
2
were unsuccessful. While grahamimycin A could be observed
from the Dess-Martin conditions, a clean conversion
resulted upon the use of a TEMPO oxidation. Thus, treatment
of colletodiol with 2 equiv of TEMPO and p-TsOH produced
grahamimycin A in good yield (76%).
The spectral data for hydroxy-acid 6 matched the data
obtained by Keck.5b Following Keck’s procedure for mac-
rolactonization (DCC/DMAP, 72%) and deprotection (Dowex/
MeOH, 86%), 6 was cleanly converted into colletodiol in
good overall yield (62%). The synthetic material had spectral
1
data that matched that recorded for the isolated material ( H
Scheme 8
NMR, 13C NMR, IR, and optical rotation). This enantiose-
lective route to colletodiol required only 11 steps and
produced sufficient quantities (12% overall yield) for further
synthetic studies (vide infra).
Scheme 7
In conclusion, a short and enantioselective synthesis of
colletodiol has been developed in which all the stereocenters
were established by the Sharpless asymmetric dihydroxyla-
tion. The overall route is comparable in length and efficiency
with previous routes, providing colletodiol in 11 longest
linear steps (16 total steps). The synthesis provides colletodiol
in 12% overall yield (87% average yield) and is amenable
for the preparation of either enantiomer. Colletol and its cis-
double-bond isomer 22 were synthesized from colletodiol
using the same palladium-catalyzed reduction. Grahamimycin
A was synthesized from colletodiol using a TEMPO oxida-
tion. Further studies on the use of this catalytic Os/Pd-
oxidation/reduction sequence for the preparation of other
macrolactone antibiotics as well as colletodiol analogues will
be reported in due course.
Because one of the two hydroxyl groups in colletodiol is
allylic, it was envisioned that colletol 2 could be synthesized
using the same palladium-catalyzed reduction (Scheme 7).
Efforts to achieve this transformation started with the
formation of the cyclic carbonate 21 of colletodiol. Treatment
Acknowledgment. We thank the Arnold and Mabel
Beckman Foundation for their generous support of our
program.
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.
(
24) For our route, this selective hydrolysis of the methyl ester was a
significant improvement. Most previous syntheses of colletodiol used either
a thioester, allyl ester, or â-sulfonylethyl ester to differentiate the two
carboxylates, which we believed would not be compatible with our approach.
(25) The methyl ester could also be deprotected with excess LiI in
refluxing pyridine.
OL0269502
4450
Org. Lett., Vol. 4, No. 25, 2002