Stereoselective total synthesis of (±)-thielocin A1β

Article abstract of DOI:10.1021/ja0112341

The stereospicific total synthesis of (±)-thielocin A1β has been achieved from the common intermediate ethyl 5-formyl-2,4-dihydroxy-3,6-dimethyl benzoate (8). The racemic synthesis was achieved based on the key reaction of a 4-methyl-3,4-dihydroxy cyclohe

Full text of DOI:10.1021/ja0112341

J. Am. Chem. Soc. 2001, 123, 11381-11387
11381
Stereoselective Total Synthesis of (()-Thielocin Alâ
Yves Ge´nisson,† Peter C. Tyler,§ R. G. Ball,‡ and Robert N. Young*
Contribution from the Merck Frosst Centre for Therapeutic Research, P.O. Box 1005,
Pointe Claire-DorVal, Que´bec, Canada H9R 4P8, and Merck Research Laboratories,
Rahway, New Jersey 07065
ReceiVed May 18, 2001
Abstract: The stereospecific total synthesis of (()-thielocin A1â has been achieved from the common
intermediate ethyl 5-formyl-2,4-dihydroxy-3,6-dimethyl benzoate (8). The racemic synthesis was achieved
based on the key reaction of a 4-methyl-3,4-dihydroxy cyclohexadienone 38 with a quinone methide derived
at low temperature from the fluoride ion catalyzed composition of piperidinium salt 40. The resulting condensate
(31) was homologated by successive esterification with protected monomeric phenol 41 to provide, after careful
removal of the protecting groups, the desired thielocin A1â.
Introduction
Scheme 1
Phospholipases A2 (PLA2’s) catalyze the hydrolysis of fatty
acids from the sn-2 position of aggregated glycerol phospho-
lipids (Scheme 1). PLA2’s have attracted attention because of
their ability to generate arachidonic acid (AA) and lyso-platelet
activating factor (lyso-PAF), two important precursors of
inflammatory lipid mediators such as eicosanoids, leukotrienes,
and PAF.1 Since AA release is thought to be the rate-limiting
step in the biosynthesis of leukotrienes and prostaglandins, it
is likely that PLA2’s play an important regulatory role in the
initiation and exacerbation of the inflammatory response in
animals. Most of the PLA2’s identified so far belong to a family
of low molecular weight secreted enzymes (sPLA2’s) isolated
from various sources such as mammalian pancreas (group I
sPLA2), snake venom (groups I and II sPLA2), bee venom
(group III sPLA2), and human platelets or synovial fluid (group
II sPLA2).2 More recently, a new type of phospholipase A2
(cPLA2) has been discovered: found in the cytosol of cells,
this class of enzymes shows no sequence homology with any
of the secreted enzymes.3 Because of their potential pivotal role
in the formation of a variety of pro-inflammatory lipid media-
tors, both mammalian group II sPLA2 and cPLA2 have been
considered as potential targets for the discovery of antiinflam-
matory agents.
Thielocins
In the beginning of the 90’s a series of unusual polymeric
depsides, named thielocins, have been isolated from culture
broths of ascomycetes ThielaVia terricola RF-143.4 Among
them, thielocin A1â (1b) (Scheme 2) was reported to exhibit
uniquely potent and selective inhibitory activity of group II
secretory phospholipase A2 (sPLA2-II) from rat platelets with
an IC50 of 3.3 nM relative to 21 µM for the rat pancreatic group
I PLA2 (sPLA2-I).4b Later reports indicated that the compound
was a somewhat less potent inhibitor of the corresponding
human sPLA2 (IC50 12 µM for human rheumatoid synovial
sPLA2-II and >100 µM for human pancreatic sPLA2-I).4c
Thielocin A1â has been studied in in ViVo models of
inflammation and has shown activity in a rat carrageenan pleur-
isy assay, thereby lending support the involvement of sPLA2-II
in the pathogenesis of inflammation in this model.5 Its antiin-
flammatory effect associated with PLA2 inhibition has also been
A better understanding of the potential therapeutic and
toxicological implications of inhibition of PLA2’s would help
to define which enzyme is the optimal target for therapeutic
intervention. It is therefore important that potent and specific
inhibitors of the PLA2 subtypes could be evaluated in animal
models of inflammation and, ultimately, in human clinical trials.
* Address correspondence to this author at Merck Frosst Centre for
Therapeutic Research.
† Current address: Laboratoire de Synthe`se et de Physicochimie des
Molecules d’Inte´reˆt Biologique, Campus Paul Sabatier, Bat. 2R1, 118 route
de Narbonne, 31062 Toulouse, Cedex 4, France.
‡ Merck Research Laboratories.
(4) (a) Yoshida, T.; Nakamoto, S.; Sakazaki, R.; Matsumoto, K.; Terui,
Y.; Sato, T.; Arita, H.; Matsutani, S. J. Antibiot. 1991, 44, 1467-1470. (b)
Tanaka, K.; Matsutani, S.; Matsumoto, K.; Yoshida, T. J. Antibiot. 1992,
45, 1071-1078. (c) Tanaka, K.; Matsutani, S.; Kanda, A.; Kato, T.; Yoshida,
T. J. Antibiot. 1994, 47, 631-638. (d) Matsumoto, K.; Tanaka, K.;
Matsutani, S.; Sakazaki, R.; Hinoo, H.; Uotani, N.; Tanimoto, T.; Kawamura,
Y.; Nakamoto, S.; Yoshida, T. J. Antibiot. 1995, 48, 106-112. (e) Teshirogi,
I.; Matsutani, S.; Shirahase, K.; Fujii, Y.; Yoshida, T.; Tanaka, K.; Ohtani,
M. J. Med. Chem. 1996, 39, 5183-5191.
§ Current address: Industrial Research Ltd., P.O. Box 31-310, Lower
Hut, New Zealand.
(1) Chang, J.; Musser, J. H.; McGregor Biochem. Pharmacol. 1987, 135,
117.
(2) Mayer, R. J.; Marshall, L. A. FASEB J. 1993, 7, 339.
(3) Clarke, J. D.; Milona, N.; Knopf, J. L. Proc. Natl. Acad. Sci. U.S.A.
1990, 87, 7708. Kramer, R. M.; Roberts, E. F.; Marretta, J.; Putnam, J. E.
J. Biol. Chem. 1991, 5268.
10.1021/ja0112341 CCC: $20.00 © 2001 American Chemical Society
Published on Web 10/26/2001

11382 J. Am. Chem. Soc., Vol. 123, No. 46, 2001
Ge´nisson et al.
Scheme 2
Scheme 3
demonstrated in a model of secretory PLA2-induced edema in
mouse paw.6 Thielocin A1â may thus provide an important lead
for further chemical modifications and serve as a useful tool
for the pharmacological evaluation of selective inhibition of
group II sPLA2 in models of shock and inflammation.
tion of a hydroxy dienone 5, a quinone methide 6, and a benzoic
acid derivative 7, all derived from the common phenolic
intermediate ethyl 5-formyl-2,4-dihydroxy-3,6-dimethyl ben-
zoate 88 (Scheme 3).
The initial report of isolation and structure elucidation of
thielocins Al (1a,b) described the selective saponification of
the central R,â-unsaturated ester bond.4a The reaction, performed
on the diastereomeric mixture of orthoformate dimethylester
derivatives 2, provided a dimeric depside 3 along with the two
ortho esters 4 (Scheme 2). Crystallization of one of these
diastereomers allowed structural assignment and designation of
the relative stereochemistry of the central core of thielocins A1.
The lability of the R,â-unsaturated ester unit together with the
relative resistance of the phenolic esters to hydrolysis indicated
that any synthetic approach would require intermediates bearing
readily and selectively removable benzoic acid protecting
groups. Thielocin A1â can be considered to be formed from
coupling of trimeric and dimeric depsides, each composed of a
common monomer 2,4-dihydroxy-3,5,6-trimethyl benzoic acid.
Further oxidation of the central phenolic unit and stereospecific
formation of a carbon-carbon bond is apparently involved in
the biosynthesis of 1a,b. The exact sequence of events is
undetermined but it is interesting to note that closely related
trimeric depsides, weak group II sPLA2’s inhibitors named
thielavins, are known to co-occur in the same fermentation broth
of ThielaVia terricola RF-143.4d
Preliminary Model Studies
The required hydroxy dienone 5 was considered derivable
from a 2,4-dihydroxy-3,5,6-trimethylbenzoate such as 9 through
oxidation, by analogy with the work of Barton et al.9 Treatment
of 9a with a variety of oxidative reagents such as CeO2/H2O2,
K2Fe(CN)6, Co(OAc)2/H2O2, (PhSeO)2O, Tl(NO3)3, PhI(OAc)2/
AcOH, or Pb(OAc)4 gave either no reaction or extensive
decomposition. However, reactions on the silyl-monoprotected
analogue 9b were more successful: Action of Pb(OAc)4 in
benzene gave 60% yield of an acetoxy dienone 10 which proved
to be isomeric with the desired product. However, oxidation of
9b with PhI(OAc)2 in acetic acid gave the desired acetoxy
dienone 11 in 60% yield10 (Scheme 4). The only isolable side
product was benzylic acetate 12 (10-20% isolated yield), likely
resulting from the trapping of a quinone methide intermediate
by acetate anion.11 Unfortunately, all attempts to suppress this
competing pathway were unsuccessful. Desilylation of 12 (HF,
CH3CN) gave 13. Careful hydrolysis of 11 (LiOH, EtOH)
directly afforded the hydroxy dienone 14.
When the reaction was carried out in alcoholic solvents such
as in methanol or (trimethylsilyl)ethanol, the corresponding
alkoxy dienones 15a,b could be obtained (30-40% yield), but
all attempts to directly introduce a hydroxyl group by performing
As the isolation yield of thielocins A1 from culture was poor
and variable and to make available analogues and synthons for
further study of structure activity relationships, we undertook
synthetic efforts directed toward these natural compounds. We
wish to present here the full report of the first total synthesis of
(()-thielocin A1â.7 Our approach is based on the efficient,
highly stereoselective, and hypothetically biomimetic condensa-
(8) Obtained from ethyl 2,4-dihydroxy-3,6-dimethylbenzoate (Elix, J. A.;
Norfolk, S. Aust. J. Chem. 1975, 28, 1113-1124) by formylation (Pulgarin,
C.; Gunzinger, J.; Tabacchi, R. HelV. Chim. Acta 1985, 68, 1948-1951).
(9) Barton, D. H. R.; Magnus, R. D.; Quinney, J. C. J. Chem. Soc., Perkin
Trans. 1 1975, 1610.
(5) Tanaka, K.; Kato, T.; Matsumoto, K.; Yoshida, T. Inflammation 1993,
17, 107.
(6) Tanaka, K.; Matsutani, S.; Matsumoto, K.; Yoshida, T. Eur. J. Pharm.
1995, 279, 143-148.
(7) Ge´nisson, Y.; Tyler, P. C.; Young, R. N. J. Am. Chem. Soc. 1994,
116, 759.
(10) Assignment of the regiochemistry in structures 9 and 10 relied on
extensive NMR analysis (NOE experiments): a representative example is
detailed for synthetic intermediate 23, the (trimethylsily)ethyl ester analogue
of 10a (see ref 14).
(11) Tamura, Y.; Yakura, T.; Haruta, J.; Kita, Y.; J. Org. Chem. 1987,
52, 3927. Pelter, A.; Elgendy, S. Tetrahedron Lett. 1988, 29, 677.

StereoselectiVe Total Synthesis of (()-Thielocin Alâ
J. Am. Chem. Soc., Vol. 123, No. 46, 2001 11383
Scheme 6a
Scheme 4a
a Reagents: (a) BnBr, t-BuOH, t-BuOK, 70 °C, 50%; (b) 18,
t-BuOH, t-BuOK, 35 °C, 12%.
a Reagents: (a) Pb(OAc)4, benzene, 60%; (b) Phl(OAc)2, AcOH;
(c) aqueous HF, CH3CN, 71%; (d) LiOH, EtOH, 53%.
Scheme 7a
Scheme 5a
a Reagents: (a) PhB(OH)2, (CHO)n, propionic acid, benzene, 4,
100%.
a Reagents: (a) Phl(OAc)2, ROH, ca. 40%; (b) 15b, H2O, AcOH,
THF, 61%.
Scheme 8
the reaction in aqueous media were unsuccessful (Scheme 5).
It was possible to cleanly cleave the silyl enol ethers in 15a or
15b (HF in CH3CN) to obtain the corresponding dienones 16a
and 16b.
Although our initial synthetic plan was to condense 13 or 14
with a quinone methide electrophile, some model studies were
undertaken to determine the regio- and stereochemical outcome
of alkylation of 13, 14, or 16b with a benzylic halide. Reactions
of 13 with benzyl bromide under a variety of basic conditions
yielded either exclusive or a preponderance of O-benzylated
products. However, using the (trimethylsilyl)ethoxy analogue
16b in t-BuOK/t-BuOH, it was possible to obtain the C-alkylated
derivative 17 (along with some of its O-alkylated congener).
However, when the same conditions were applied to 16b and a
more elaborate benzyl bromide 18 (derived in five steps from
ethyl 2-methoxy-3,6-dimethyl-4-hydroxybenzoate) the major
product resulted from O-alkylation and only 12% yield of
C-alkylated material 19 was obtained (Scheme 6). Most dis-
couragingly, this C-alkylated derivative was predominantly the
undesired stereoisomer wherein the newly formed carbon-
carbon bond was introduced anti to the tertiary trimethylsilyl-
ethyloxy group. We therefore returned our focus to the prep-
aration of suitably substituted quinone methides as originally
envisaged.
Quinone methides of type 6 have been prepared in a number
of ways. We first attempted the condensation of 13 with the
quinone methide generated in situ from the cyclic boronate 20
(Scheme 7) formed from reaction of ethyl 2-methoxy-3,6-
dimethyl-4-hydroxybenzoate, phenylboronic acid, and formal-
dehyde.12 However, attempts under different Lewis acid con-
ditions (Et2O‚BF3 or TiCl4) yielded no useful product mixture.
We therefore investigated alternative methods for generating
the desired quinone methide under milder conditions.
It was envisaged that decomposition of an activated derivative
21, where X represents a good leaving group such as a tertiary
amine N-oxide or a quaternary ammonium salt, could provide
the desired quinone methide at low temperature (Scheme 8).
The amine precursor 22 could be readily prepared from
Mannich condensation of 2-methoxy-3,6-dimethyl-4-hydroxy-
benzoate with piperidine and formaldehyde. Standard oxidation
with m-CPBA provided the N-oxide 23. Treatment of 23 under
either basic conditions (NaH, NaCN, THF) or Lewis acid
conditions (TMSCN, CH2Cl2) yielded only a rearrangement (to
24) where the in situ generated quinone methide had been
trapped by the N-hydroxypiperidine formed in the reaction.
Silylation of the phenol 22 followed by quaternization of the
amine with methyl triflate furnished salt 25. Treatment of 25
with TBAF at -20 °C in dichloromethane led to instantaneous
conversion to the quinone methide, which was stable below -20
°C, and slowly underwent [2+4] cycloaddition to yield dimer
26 when warmed to room temperature. In a model reaction,
treatment of 25 with TBAF in the presence of 2-methylcyclo-
hexane-1,3-dione at -30 °C in THF followed by warming to
room temperature yielded the desired tricyclic structure 27 (43%
yield). Reaction with acetyl-protected dienone 13 under similar
conditions gave the expected hemiketal 28 (62% yield) as a
mixture of all four possible isomers (Scheme 9).
(12) Chambers, J. D.; Crawford, J.; Williams, H. W. R.; Dufresne, C.;
Scheigetz, J.; Bernstein, M. A.; Lau, C. K. Can. J. Chem. 1992, 70, 1717.
This methodology has been applied in our group to the total synthesis of
thielocin B3 (Ge´nisson, Y.; Young, R. Tetrahedron Lett. 1994, 35, 7747-
7750).

11384 J. Am. Chem. Soc., Vol. 123, No. 46, 2001
Ge´nisson et al.
Scheme 9a
Figure 1.
Figure 2.
a Reagents: (a) piperidine, (CHO)n, EtOH, 4, 85%; (b) mCPBA,
CHCl3, -10 °C, 100%; (c) NaH, NaCN, -60 °C to room temperature,
THF, 50%; (d) (1) TBDMSCl, NaH, THF, RT, 3h, (2) methyl triflate,
ether, -15 °C to room temperature, crude product used directly in next
reaction; (f) 2-methyl-1,3-cyclohexandione, TBAF, -30 °C to room
temperature, 43% for three steps; (g) 13, TBAF, -30 °C to room
temperature, 62% for three steps.
The reasons for the exceptional stereocontrol of this coupling
reaction in contrast to the reaction of 13, which yielded a mixture
of isomers, are not clearly understood. The results strongly
suggest a critical influence of the free tertiary hydroxyl group
in 14 on the stereochemical outcome, likely through a stabilizing
hydrogen bond with the quinone methide carbonyl in the
transition state (Figure 2). The free diol 29 proved to be sensitive
to basic hydrolysis conditions while the methyl orthoformate
derivatives 30 were surprisingly resistant to saponification. It
was therefore evident that for a successful synthesis of thielocin
A1, the protecting esters used must be carefully chosen so as
to allow smooth and selective removal.
Scheme 10
Synthetic Strategy
The overall synthetic strategy was derived from the assump-
tion that the phenolic esters of thielocin A1 would be introduced
after the formation of the central core of the molecule. This
implied that different acid protecting groups would have to be
introduced on the hydroxy dienone and on the quinone methide
precursors. We envisaged that the R,â-unsaturated ester of the
central dimeric unit must be hydrolyzed selectively and then
esterified with a phenolic monomer bearing the same protecting
group as that on the aromatic function originating from the
quinone methide component. Both of these esters could then
be simultaneously cleaved to yield a dicarboxylic acid from
which the pentameric skeleton of thielocin A1 could be directly
available via concomitant esterification with the phenolic
monomer unit. Final deprotection would then lead to the target
natural product. Based on model studies, it was decided to
prepare the dienone component 5 protected as (trimethylsilyl)-
ethyl (TMSE) ester and the phenol 7 as well as the quinone
methide 6 both protected as 2,2,2-trichloroethyl (TCE) esters.
However, when the same reaction was carried out with the
free hydroxy analogue 14, the desired product 29 was obtained
in essentially quantitative yield (as a dynamic mixture of
hemiketals) and with specific control of the relative stereo-
chemistry (Scheme 10). No O-alkylation or traces of the other
isomers could be observed. Indications that the relative stereo-
chemistry corresponded to the target natural compound were
obtained by NMR studies on the mixture of methyl orthofor-
mates 30 derived from the R isomer.13 Interestingly, crystal-
lization of 29 (from ethyl acetate) selectively provided the R
isomer through dynamic equilibrium. X-ray diffraction analysis
of this crystalline material confirmed the relative stereochemistry
indicated from NMR structural assignment (Figure 1).
Synthesis of Key Intermediates 35, 40, and 41
(13) The key data were the following (acetone-d6, 400 MHz, δ in ppm):
the product was a mixture of ca. 2:1 diastereoisomers. The highest field
methyl singlets (δ 1.22 and 1.20) were assigned to the angular methyl groups
R to the enone carbonyl based on NOEs to the central methylene protons
signal (δ 2.90). Both of these methyl signals displayed NOEs to the methyls
in γ to the enone carbonyl singlets (δ 1.87 and 1.81). Finally, irradiation
of δ 1.87 methyl singlet (major isomer) induced an NOE in the δ 5.90
orthoformate proton signal, whereas no similar NOE was observed for the
δ 1.81 signal (minor isomer).
Conversion of the pivotal intermediate 8 to the hydroxy
dienone trimethylsilylethyl ester 35 was effected as follows
(Scheme 10). Clemensen reduction provided 31 (88% yield)
which on transesterification with sodium (trimethylsilyl)ethoxide
in benzene gave the TMSE ester 32 (63% yield). After selective
protection of the more reactive para hydroxyl group as the tert-

StereoselectiVe Total Synthesis of (()-Thielocin Alâ
J. Am. Chem. Soc., Vol. 123, No. 46, 2001 11385
Scheme 11a
with triethylsilane in trifluoroacetic acid, which cleanly afforded
the desired reduced compound as its TCE ester 41 (96% yield).15
Final Assembly of the Units
We were pleased to observe that, in keeping with our model
studies, fluoride ion catalyzed coupling of 35 and 40 (TBAF,
-20 °C to room temperature) proceeded in essentially quantita-
tiVe yield and with complete stereoselectiVity to afford the
differentially protected tricyclic compound 42 as a thermody-
namic mixture of two hemiketals (Scheme 12).
Considering the difficulties anticipated in dealing with a
mixture of hemiketals and the potential reactivity of the masked
phenol in the subsequent steps of the synthesis, it was decided
to protect the two free hydroxyl groups. Treatment of the
mixture of hemiketals with carbonyldiimidazole and triethyl-
amine led to a complete transformation to the cis form and for-
mation of a single carbonate 43 (98% yield).
Selective hydrolysis of the TMSE ester 43 with fluoride ion
(TBAF in DMF) gave the mono acid 44 (83% yield) (Scheme
7). This was converted to the acid chloride with dichlorometh-
ylmethyl ether in refluxing dichloromethane16 and then reacted
with phenol 41 (Et3N, CH2C12) to yield the trimeric ester 45
(89% combined yield).
a Reagents: (a) Zn, HgCl2, HCl, EtOH, 4, 91%; (b) Me3CH2CH2ONa,
benzene, 4, 63%; (c) TBDPSiCl, Et3N, DMAP, CH2Cl2, 91%; (d)
Phl(OAc)2, AcOH (57%); (e) LiOH, THF, H2O, 25 °C, 70%; (f)
Cl3CH2OH, concentrated H2SO4, 0 to 25 °C, 61%; (g) (MeO)2SO2,
K2CO3, acetone, 50 °C, 90%; (h) piperidine, Na(CN)BH3, CH3CN, 83%;
(i) NaH, THF, TBDMSiCl, 95%; (j) CF3SO2OMe, CH2Cl2, 0 °C, 99%;
(k) Et3SiH, TFA, 96%.
At this point, we were disappointed when attempted reductive
cleavage of the TCE esters under standard conditions (Zn(0) in
AcOH) gave complex reaction mixtures. This is although model
studies on several analogous highly substituted aromatic TCE
esters proceeded cleanly to give the desired benzoic acids. These
results suggested that the R,â-unsaturated ketoester moiety in
the central core of the molecule was sensitive to reducing
conditions. However, concomitant removal of the two TCE
esters could be smoothly achieved by utilizing milder conditions
of Cd(0) in DMF/AcOH17 to give the dicarboxylic acid 46 (84%
yield). Bis-esterification of 46 with the phenol 41, utilizing a
process previously reported for the preparation of depsides
(TFAA, benzene),18 afforded the protected thielocin A1R (47)
(83% yield). Removal of the TCE esters again proceeded
smoothly with Cd(0) in DMF/AcOH to provide the penultimate
carbonate 48 (80% yield) (Scheme 12).
Initial attempts to hydrolyze the carbonate by heating in
aqueous pyridine gave some of the expected diol (12% yield)
along with compounds where the central nonaromatic ester had
been cleaved to yield the phenolic dimer 49 (45% yield) and
the decarboxylated trimeric component 50 (57% yield) (Scheme
13). This decarboxylation was thought to result from deconju-
gation of the R,â-unsaturated keto ester function, either through
addition of a nucleophile (such as the phenolate anion derived
from opening of the hemiketal) or base-catalyzed enolization
(vide infra). The R-keto ester thus liberated could subsequently
undergo hydrolytic decarboxylation and either re-elimination
of the nucleophile or isomerization would then yield the
decarboxylated product. Careful monitoring of this final step
appeared to be necessary to avoid any side reactions.
butyldiphenysilyl (TBDPS) ether to give 33 (91% yield),
regioselective oxidation with phenyliodonium diacetate in acetic
acid provided the acetoxy dienone 34 (57% yield).14 Carefully
optimized hydrolysis of both acetate and TBDPS enol ether
(LiOH, 12-crown-4) gave the desired hydroxy dienone 35 (76%
yield).
Synthesis of the key piperidinium salt 40 was achieved as
follows (Scheme 11). Transesterificat¨ıon of 8 with trichloroeth-
anol in sulfuric acid yielded the TCE ester 36 (61% yield).
Treatment of the latter with dimethyl sulfate and potassium
carbonate selectively provided the monomethyl ether 37 (90%
yield), where methylation occurred exclusively ortho to the ester
function. It is interesting to note that no selectivity was observed
when the same conditions were applied to the ethyl ester
analogue 8. Reductive amination (piperidine, NaCNBH3) then
cleanly afforded the benzylic amine 38 (83% yield). Protection
of the free hydroxyl group as the tert-butyldimethylsilyl ether
39 (95% yield) followed by quaternization of the amino group
with methyl triflate provided the stable quaternary ammonium
salt 40 (99% yield).
The phenolic monomer 41 required for the completion of the
synthesis was thought to be readily derivable from aldehyde
37 (Scheme 11). However, preliminary attempts to effect this
conversion under hydrogenation conditions (40 psi H2, Pd(OH)2,
AcOH) unexpectedly yielded the 2-methoxy analogue of 31,
as a result of the concomitant reduction of the TCE ester into
an ethyl ester. This could be circumvented by treatment of 37
After exploration of various conditions, hydrolysis of the
carbonate 48 was found to proceed smoothly upon treatment
with 1 N sodium hydroxide in dioxane at -10 °C. Monitoring
by reverse-phase HPLC showed almost instantaneous conversion
of the starting carbonate to a more retained intermediate, likely
the enol 51 (treatment of 51 with diazomethane led to the
(14) The structure assignment of 34 was based on the following NMR
observations (acetone-d6, 400 MHz, δ in ppm): the methyl group proximal
to the ester group was easily established by irradiation of the singlet at δ
1.94 and observing an NOE in the ester methylene protons signal (δ 4.28).
Irradiation of the methyl singlets at δ 2.04 and 1.81 displayed NOEs in the
methyl signal shown to be ortho to the ester group. The singlet at δ 1.81
showed a strong NOE and was assigned to the quaternary methyl group
while the δ 2.04 singlet showed a comparatively weak NOE, which together
with its chemical shift led us to ascribe it to the acetoxy group. Finally,
irradiation of the acetoxy methyl signal resulted in a small NOE in the δ
1.81 methyl. The remaining δ 1.18 methyl singlet was rationalized to be
ortho to the ketone functionality because of its relatively shielded resonance
shift and a small 1,3 interaction with the quaternary methyl signal observed
in the NOE experiment.
(15) West, C. T.; Donnelly, S. J.; Kooistra, D. A.; Doyle, M. P. J. Org.
Chem. 1973, 38, 2675.
(16) Rieche, A.; Gross, H. Chem. Ber. 1959, 92, 83.
(17) Hancock, G.; Galpin, I. J.; Morgan, B. A. Tetrahedron Lett. 1982,
23, 249.
(18) Parish, R. C.; Stock, L. M. J. Org. Chem. 1965, 30, 927.

11386 J. Am. Chem. Soc., Vol. 123, No. 46, 2001
Ge´nisson et al.
Scheme 12a
Scheme 14a
a Reagents: (a) TBAF, CH2Cl2, THF, -20 °C to room temperature,
95%; (b) CDI, Et3N, DMF, 98%; (c) TBAF, DMF, 83%; (d) (1)
Cl2CHOCH3, CH2Cl4, 44, 4, 3 h, (2) 41, Et3N, room temperature, 89%
overall; (e) Cd0, AcOH, DMF, 84%; (f) TFAA, 41, benzene, 83%; (g)
Cd0, AcOH, DMF, 80%.
Scheme 13a
a Reagents: (a) 1 N NaOH, dioxane, -20 °C, 5 min; (b) CH2N2,
100%; (c) 1 N NaOH, dioxane, 8-10 h; (d) 2.5 N H2PO4, workup,
HPLC, 59% overall.
1H, 13C NMR, melting point, mass, and IR spectra. The
inhibitory profile of thielocin A1â for type II secretory phos-
pholipase A2 was confirmed by utilizing purified human synovial
fluid enzyme (IC50 ) 2.4 µM) as well as semipurified enzyme
from rat paw (IC50 ) 2 nM).
a Reagents: (a) H2O, pyridine, 90 to 110 °C, 10 h.
formation of the enol ether 52, Scheme 14).19 The latter was
then progressively (8-10 h) transformed under hydrolytic
conditions into a less retained precursor, assumed to be the
enolized thielocin A1 (53).20 Quenching of the reaction with
phosphate buffer and extraction followed by evaporation did
not affect the composition of the reaction mixture. However,
subsequent purification of this precursor by reverse-phase HPLC
followed by extractive isolation was accompanied by exclusive
isomerization to the desired thielocin A1â that could be thus
Conclusion
We have thus achieved a total synthesis of (()-thielocin A1â
and confirmed the biological activity of the synthetic material.
The three key synthons, 35, 40, and 41, are all derived from a
common intermediate ethyl 5-formyl-2,4-dihydroxy-3,6-di-
methyl benzoate (8). It is interesting to speculate as to whether
the biosynthesis of thielocin A1â may also derive from the
condensation of a quinone methide with a suitably substituted
hydroxy dienone. Similar condensation has been hypothesized
for the formation of the complex metabolites derived from
vitamin E.22 The high degree of regio- and stereoselectivity ob-
served in this condensation may suggest that similar forces are
involved in the biosynthesis of thielocin A1â itself.
21
isolated with 59% yield (Scheme 14).
Thielocin A1â gave physical and structural data essentially
identical to those reported for the natural product as judged by
(19) Rapid quench of the reaction after 5 min with diluted phosphoric
acid followed by extractive workup yielded a crude mixture consisting
essentially of this intermediate, as judged by HPLC analysis. Characteristic
IR absorptions at 1820 cm-1 indicated the continued presence of a carbonate
unit in the molecule. The 1H NMR spectrum (CDCl3, 400 MHz, δ in ppm)
of this material showed two characteristic apparent singlets in the olefinic
region (δ 5.84 and 5.96) and an exchangeable broad signal (δ 8.54-8.77).
Attempts to further purify this crude mixture on HPLC were accompanied
by partial isomerization back to the starting diacid carbonate 48. Neverthe-
less, treatment of the crude material with diazomethane gave a stable
Experimental Section
Diacid 48. Cadmium powder (100 mesh; 500 mg, 4.45 mmol) was
added to a stirred suspension of 47 (117.2 mg, 0.091 mmol) in a
dimethylformamide/acetic acid (1:1) mixture (2 mL) at 25 °C. This
mixture was vigorously stirred at the same temperature for 15 h. Water
(20 mL) and ethyl acetate (20 mL) were then added. Insoluble cadmium
excess was decanted. The aqueous layer was separated and extracted
with ethyl acetate (20 mL). The combined organic extracts were washed
with brine (2 × 20 mL), dried (MgSO4), filtered, and concentrated in
vacuo. Precipitation in CH3OH of the residue obtained gave 48 (75
mg, 80%) as a yellow solid: mp: 268-271 °C; IR (KBr) υmax 3700-
2300 (br), 1825, 1740, 1700, 1570 cm-1; 1H NMR (400 MHz, CDCl3-
CD3OD (1:1)) δ 3.87 (s, 3H, OMe), 3.86 (s, 3H, OMe), 3.84 (s, 6H,
1
compound isolable by silica gel chromatography. The H NMR spectrum
(CDCl3, 400 MHz, δ in ppm) of this derivative showed the continued
presence of two singlets (in the δ 5.70-6.00 region) plus the presence of
three extra methyls in the methoxy region, including the two expected methyl
esters. Mass spectrum (FAB) showed an (MH)+ peak at m/z 1065 confirming
the incorporation of three methyls. These spectral data are in agreement
with the structure of the methyl enol ether 52 (Scheme 14).
(20) This assumption is based on the fact that the 1H NMR spectrum
(CDCl3, 400 MHz, δ in ppm) of the worked up crude material after
completion of the hydrolysis but before isomerization to the final product
showed two recognizable singlets in the olefinic region (δ 5.70 and 5.74)
along with other characteristic signals for thielocins A1.
(21) The reaction provided the thermodinamically more stable â isomer.
Thielocins A1R and â are known to form an equilibrium largely in favor
of the â isomer (see ref 4).
(22) Suarna, C.; Craig, D. C.; Cross, K. J; Southwell-Keely, P. T. J.
Am. Chem. Soc. 1988, 53, 12.

StereoselectiVe Total Synthesis of (()-Thielocin Alâ
J. Am. Chem. Soc., Vol. 123, No. 46, 2001 11387
2×OMe), 3.08 (ABq, J ) 16 Hz, ∆ν ) 84 Hz, 2H, CH2), 2.45 (s, 3H,
Me), 2.43 (s, 6H, 2×Me), 2.32 (s, 3H, Me), 2.31 (s, 3H, Me), 2.30 (s,
6H, 2×Me), 2.27 (s, 9H, 3×Me), 2.26 (s, 6H, 2×Me), 2.13 (s, 3H,
Me), 1.45 (s, 3H, Me); 13C NMR (100 MHz, CDCl3-CD3OD (1:1) δ
192.5, 170.3, 165.9, 165.8, 162.0, 155.1, 153.8, 152.6, 150.6, 149.9,
148.8, 148.6, 148.5, 148.3, 133.4, 133.1, 131.6, 130.4, 127.9, 126.3,
125.5, 125.1, 123.1, 121.8, 121.4, 121.4, 117.5, 114.2, 104.2, 84.2,
61.5, 61.4, 61.1, 43.8, 27.6, 19.8, 16.4, 16.3, 15.7, 15.4, 15.1, 12.3,
12.1, 9.59, 9.29, 8.26; FAB MS (NBA/NaCl) m/z 1045 (M + Na+);
FAB HRMS (Glycerol) m/z calcd for C55H59O19 (M + H+) 1023.3654,
found 1023.3650.
reverse-phase HPLC (column: µBondapak C18 (125 Å, 10 µm, 25 ×
100 mm); eluant: CH3CN/0.1% H3PO4 (65:35); flow: 15 mL/min;
detection: 220 nm) allowed isolation of fractions corresponding to the
precursor 53 and eventually the desired compound 1b itself (rt 10.5
and 7.3 min, respectively). These fractions were combined and most
of the CH3CN evaporated off in vacuo. The resulting aqueous phase
was then extracted with ethyl acetate (3 × 10 mL). The combined
organic extracts were washed with brine (2 × 5 mL), dried (MgSO4),
filtered, and concentrated in vacuo to give fully isomerized thielocin
A1â (1b) (11.2 mg, 59%) as a white solid: mp 198-201 °C (ether)
(lit.4a mp 190-194 °C); our synthetic racemic material gave otherwise
1
Thielocin A1 â (1b). NaOH solution (1 N, 200 µL) was added to a
well-stirred solution of carbonate 48 (20.1 mg, 0.019 mmol) in dioxane
(200 µL) at 25 °C under Ar. This mixture was instantaneously cooled
to -10 °C and strictly maintained at this temperature with vigorous
stirring until quenching. Reaction was then monitored by analytical
reverse-phase HPLC (column: Nova-Pak C18 (60 Å, 4 µm, 3.9 × 150
mm); eluant: CH3CN/0.1% H3PO4 (60:40); flow: 1 mL/min; detec-
tion: 220 nm): aliquots (1 µL) of the reaction mixture were acidified
with 2.5 N H3PO4 solution (20µL) and diluted with CH3CN (40µL)
before injection. HPLC showed instantaneous conversion of the starting
carbonate (48, rt 9 min) into a slower intermediate (enol 51, rt 19 min).
Slow transformation of the latter into a less retained precursor (enolized
thielocin 53, rt 6.4 min) was then observed. Only traces of the desired
â isomer (1b, rt 3.4 min) could be detected at that point. On completion
of the reaction (8 to 10 h) 2.5 N H3PO4 solution (1 mL) was added at
-10 °C. The resulting milky aqueous phase was diluted with water (1
mL) and extracted with ethyl acetate (3 × 5 mL). The combined organic
extracts were washed with brine (2 × 5 mL), dried (MgSO4), filtered,
and concentrated in vacuo. This treatment did not affect the composition
of the crude as indicated by HPLC. Purification on semipreparative
physical and spectral data essentially identical (IR, H and 13C NMR,
MS) to those reported for the natural chiral material.
Acknowledgment. This work was partially supported by a
grant to Y.G. (1992-1993) from the “Fondation pour la
Recherche Medicale” (Paris, France). We are indebted to D.
A. Evans for useful suggestions. P Weech and N. Tremblay
are acknowledged for biochemical experiments. We are also
thankful to C. Li, J. Yergey for mass spectra, M. Bernstein and
L. Trimble for NMR analysis. We would like to thank D.
Delorme, R. Ruel and Y. Han for helpful scientific discussions.
Supporting Information Available: Complete experimental
details (PDF). This material is available free of charge via the
Internet at http://pubs.acs.org. The structure of compound 29
has been deposited with the Cambridge Crystallographic Data
Centre. Refer to deposition number CCDC168787.
JA0112341