Estrogen receptor ligands. 12. Synthesis of the major metabolites of an ERα-selective, dihydrobenzoxathiin antagonist for osteoporosis

Article abstract of DOI:10.1021/ol047741f

(Chemical Equation Presented) During the course of drug metabolism studies, a major metabolite of compound 1 was detected in rhesus monkeys and assigned structure 4. The intriguing biotransformation of 1 leading to 4 was confirmed by a 19-step total synth

Full text of DOI:10.1021/ol047741f

ORGANIC
LETTERS
2005
Vol. 7, No. 3
411-414
Estrogen Receptor Ligands. 12.
Synthesis of the Major Metabolites of an
ERr-Selective, Dihydrobenzoxathiin
Antagonist for Osteoporosis
Seongkon Kim,*,‡ Jane Y. Wu,‡ Zhoupeng Zhang,† Wei Tang,† George A. Doss,†
Brian J. Dean,† Frank DiNinno,‡ and Milton L. Hammond‡
Departments of Medicinal Chemistry and Drug Metabolism,
Merck Research Laboratories, P.O. Box 2000 (800B-109), Rahway, New Jersey 07065
seongkon_kim@merck.com
Received November 3, 2004
ABSTRACT
During the course of drug metabolism studies, a major metabolite of compound 1 was detected in rhesus monkeys and assigned structure
4. The intriguing biotransformation of 1 leading to 4 was confirmed by a 19-step total synthesis starting from resorcinol (11), the key feature
of which was the construction of the oxygen bridge utilizing a phenolic oxidation and trapping sequence. In addition, the synthesis of a
related metabolite (5) is described.
In conjunction with a medicinal chemistry program targeting
selective estrogen receptor modulators (SERMs), dihy-
drobenzoxathiin 11 was discovered to be a potent SERAM2
(Selective Estrogen Receptor Alpha Modulator) that exhibited
subnanomolar binding to ERR and 40-fold selectivity. Further
evaluation of 1 revealed its promise for the prevention of
estrogen-deficiency osteopenia in postmenopausal women,
without uterotropism,2b and thereby 1 qualified as a devel-
opmental candidate.3
In preclinical studies in the rhesus monkey, the major
metabolites of 1 included two O-glucuronides (2 and 3)4 and
the bridge compound 4,5 as depicted in Figure 1, which
together accounted for approximately 70% of the metabolism.
Intestinal extraction via glucuronidation, “first-pass effect”,
accounted for the formation of the two O-glucuronides.
However, the intriguing biotransformation leading to 4 was
seemingly unprecedented. A mechanism, which is described
in Scheme 1, may be postulated to partially proceed through
a quinone intermediate 7, generated from a phenolic-radical
fragmentation process embedded in the dihydrobenzoxathiin
† Department of Drug Metabolism.
‡ Department of Medicinal Chemistry.
(1) Kim, S.; Wu, J. Y.; Chen, H. Y.; DiNinno, F. Org. Lett. 2003, 5,
685.
(2) (a) Kim, S: Wu, J. Y.; Birzin, E. T.; Frisch, K.; Chan, W.; Pai, L.
Y.; Yang, Y. T.; Mosley, R. T.; Fitzgerald, P. M. D.; Sharma, N.; Dahlund,
J.; Thorsell, A. G.; DiNinno, F.; Rohrer, S. P.; Schaeffer, J. M.; Hammond,
M. L. J. Med. Chem. 2004, 47, 2171. (b) Kim, S.; Wu, J. Y.; Chen, H. Y.;
Birzin, E. T.; Chan, W.; Yang, Y. T.; Colwell, L.; Li, S.; Dahlund, J.;
DiNinno, F.; Rohrer, S. P.; Schaeffer, J. M.; Hammond, M. L. Bioorg. Med.
Chem. Lett. 2004, 14, 2741. (c) Chen, H. Y.; Kim, S.; Wu, J. Y.; Birzin, E.
T.; Chan, W.; Yang, Y. T.; Dahlund, J.; DiNinno, F.; Rohrer, S. P.;
Schaeffer, J. M.; Hammond, M. L. Bioorg. Med. Chem. Lett. 2004, 14,
2551.
(3) Song, Z. J.; King, A. O.; Waters, M. S.; Lang, F.; Zewge, D.; Bio,
M.; Leazer, J. L., Jr.; Javadi, G.; Kassim, A.; Tschaen, D. M.; Reamer, R.
A.; Rosner, T.; Chilenski, J. R.; Mathre, D. J.; Volante, R. P.; Tillyer, R.
Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 5776.
(4) Independent synthesis confirmed structures of 2 and 3, respectively.
(5) The structure of this unusual oxygen bridge metabolite 4 and 5 was
unambiguously determined by extensive 2D NMR experiments (ROESY,
HMQC, HMBC) and LC/MS/MS. The absolute configuration at C3 was
assumed to be (R) on the basis of the fact that stereocenter at C3 should
remain unchanged during the biotransformation.
10.1021/ol047741f CCC: $30.25
© 2005 American Chemical Society
Published on Web 01/05/2005

structure assignments, as well as to further evaluate the
metabolites, we targeted the synthesis of all of the different
stereoisomers and now report the details of this effort.
Our initial goal was to secure the carbon framework as
shown in Scheme 2. The synthesis began with the known
Scheme 2a
Figure 1.
core, followed by a nucleophilic addition of a phenolic-
radical in the pendant aromatic ring to the resultant benzo-
quinone moiety in the A-ring to provide the intermediate 8.
In turn, this initially formed intermediate would eventually
yield metabolite 4 via sequential glucuronidation and bridge
ether formation of the tautomeric phenol 10. Intuitively, the
isolation of 5 and 66 from the in vitro incubation of 1 with
monkey liver microsomes provided support for the plausibil-
ity of the above metabolic pathway. The possibility that
metabolite 5 would be a biosynthetic precursor of 4, however,
could not be ruled out. Although mechanistic probing awaits
further study, the unique structural features of compound 4,
combined with the fascinating biotransformaton, prompted
us to attempt the total synthesis of 4 and 5. To confirm the
a Key: (a) (i) 2-bromo-5-methoxybenzoic acid, CuSO4, NaOH,
(ii) BnBr, Cs2CO3, quant; (b) Me3Al, Et2NHCl, 80 °C, 90%; (c)
chlorodimethyl thiocarbamate, NaH, DMAP, 75%; (d) Ph2O, 280
°C, 4 h, 75%; (e) NaOH in MeOH, then 1 N HCl, quant; (f) (i)
BBr3, 0 °C, (ii) TIPSCl, 70% for two steps, (iii) BBr3, rt, 80%; (g)
(i) LAH, 0 °C, (ii) TFA, Et3SiH, 80% for two steps, (iii) TBSCl,
Et3N, quant yield.
compound 12,7 which was easily synthesized from resorcinol
11 in two steps. The resulting coumarin 12 was converted
into the corresponding amide 13 in excellent yield by reaction
with methylchloroaluminum amide.8
Scheme 1
A low yield was observed when dimethylamine was used
as the amidating reagent, due in part to concomitant reclosure
to the starting material 12. The phenol 139 was immediately
allowed to react with dimethylthiocarbamoyl chloride to
afford the O-aryl dimethylthiocarbamate 14 in 75% yield,
which upon pyrolysis gave the S-aryl dimethylthiocarbamate
15 as the product of the Newman-Kwart rearrangement.10
A temperature of 280 °C was necessary for clean conversion;
(6) We isolated 6 as a single isomer, whose stereochemistry was believed
to be (S), the result of a retention of the stereochemistry of C-2 in compound
1. However, we did not confirm the stereochemistry.
(7) (a) Devlin, J. P. Can. J. Chem. 1975, 53, 343. (b) Lederer, E.;
Polonsky, J. Bull. Soc. Chim. Fr. 1948, 831.
(8) Levin, J. I.; Turos, E.; Weinreb, S. M. Synth. Commun. 1982, 12,
989.
(9) After overnight at room temperature, ca. 20% of the product was
converted back to the starting material 12.
(10) Villemin, D.; Hachemi, M.; Lalaowi, M. Synth. Commun. 1996,
26, 2461.
412
Org. Lett., Vol. 7, No. 3, 2005

otherwise only starting material or complex mixtures were
obtained. Alkaline hydrolysis of compound 15 followed by
simple addition of acid (dilute HCl) promoted smooth
cyclization to the thiocoumarin 16 in high yield.
5. To our delight, chromatographic separation of both
diastereomers, as well as the enantiomers, was realized by
HPLC using a chiracel OD column. Separately, all four
isomers underwent cleavage of the silicon protecting groups
with TBAF at room temperature to afford the respective
stereoisomers of 5: (+)-5, (-)-5, (+)-25, and (-)-25.
Among them, compound (-)-5 was found to be identical in
all respects to an authentic sample of the metabolite (1H
NMR, 13C NMR, MS, HPLC,14 [R]D ) -148° in MeOH, c
) 0.466).
The configurations of the newly created stereogenic centers
at C2,3 were assigned as follows. For 23 and 24, the coupling
constant of H2,3, from the observed AX pattern, were large
at J ) 9.2 and 9.6 Hz, respectively, suggesting that the two
protons adopted an antiperiplanar conformation. In this way,
both isomers 23 and 24 would have minimal gauche
interactions. However, most striking was the chemical shift
of Ho in the B ring of threo 24, which resonated at δ 5.98
ppm. In contrast, the same proton of the erythro isomer 23
was found at δ 6.86 ppm. The unusual, higher field chemical
shift of the threo isomer was attributed to the anisotropic
effect of circulating π-electrons of the phenyl group of the
side chain. Accordingly, the configuration at C2,3 in 5 was
designated erythro, 23.15
With the enantiomer (-)-5 in hand, we continued to pursue
the synthesis of 4 starting with enantiomer (-)-23, whose
stereochemistry at C2,3 was believed to be related to the
metabolite, based primarily on the mechanism by which
metabolite 4 could be derived from 5. Toward this end, the
benzylic alcohol (-)-23 was converted to the TES ether 26
in 90% yield after treatment with TESOTf. Selective
deprotection of the phenolic TBS ether 26 was effected by
exposure to mild basic medium16 to afford the monophenol
27 in 75% yield. More basic conditions, such as TBAF or
Cs2CO3, tended to give a mixture of phenols in a nonselective
manner. Subsequent glycosylation was achieved in 70% yield
under Schmidt conditions, wherein the phenol 27 in CH2Cl2
was treated with chloroimidate B and catalytic BF3-etherate,
at -10 °C.17 This coupling provided a single product 28 with
the desired â-stereochemistry at the anomeric center,18 as
present in the metabolite 4. Exhaustive desilylation with
TBAF in the presence of HOAc delivered the penultimate
intermediate 29 in 80% yield (Scheme 4).19
At this juncture, the synthetic plan required a suitable
protecting group strategy for the installation of the glucu-
ronide in the B ring, and phenolic oxidation in the A-ring
that would build the ether bridge framework. The benzyl
group was selectively removed with 1.5 equiv of boron
tribromide at 0 °C to the corresponding phenol 17, which
without purification was transformed to the TIPS-protected
compound 18. Then selective deprotection of methyl group
was realized with an excess of boron tribromide at room
temperature to afford the desired compound 19 in quantitative
yield. This operation was dictated by our inability to cleave
the methoxy group at a latter part of the synthesis due to the
acid lability11 of the final compound 4. Partial reduction of
compound 19 with LAH in THF at 0 °C gave the 10-hydroxy
thiopyran 20, which was further reduced to compound 21
with TFA/Et3SiH. It should be noted that this reduction
initially gave the dimeric ether (not shown)12 upon treatment
with 10-15 equiv of TFA; however, an excess amount of
acid was required to drive the reduction to completion. The
protection of the phenol with TBSCl was uneventful, which
ultimately set the stage for the completion of target com-
pounds 4 and 5.
Having secured a viable sequence to the core 22, we then
focused on the installation of the pyrrolidine side chain,
which should then directly provide metabolite 5. Treatment
of thiopyran 22 with n-BuLi at -78 °C resulted in a deep-
reddish-colored lithiated complex. This R-thio anion13 seemed
stable at -78 °C but slowly decomposed at higher temper-
atures (> -40 °C), and reaction with the aldehyde A
provided a mixture of products (()-erythro 23 and (()-threo
24, in high yield, with 1:1 diasteroselectivity, as determined
1
by H NMR (Scheme 3). No further attempt was made to
Scheme 3
(12) (a) Hori, M.; Kataoka, T.; Shimizu, H.; Komatsu, O.; Hamada, K.
J. Org. Chem. 1987, 52, 3668. (b) Ridley, D. D.; Smal, M. A. Aust. J.
Chem. 1983, 36, 795.
(13) The R-thio anion seemed to be quite sensitive to oxygen, and
therefore THF was rigorously degassed by N2 prior to use. Unterhalt, B.;
Bruening, S. Sci. Pharm. 1997, 65, 1.
(14) The four isomers of 5 were well separated on a Chiracel OD-RH
column (4.6 mm × 150 mm, 3 µm).
(15) For the isothiochromane system, see: (a) Tomooka, K.; Wang, L.
F.; Okazaki, F.; Nakai, T. Tetrahedron Lett. 2000, 6121. (b) Bohme, H.;
Sutoyo, P. N. Phosphorous Sulfur Silicon Relat. Elem. 1982, 13, 235.
(16) Wilson, N. S.; Keay, B. A. Tetrahedron Lett. 1997, 38, 187.
(17) (a) Schmidt, R. R. Pure Appl. Chem. 1989, 61, 1257. Chloro imidate
B was prepared in 3 steps from the commercially available methyl-1,2,3,4-
tetra-O-acetyl-D-glucopyranuronate.
improve the stereoselectivity, since both diastereomers were
needed for structural elucidation of both metabolites 4 and
(18) For 28, a 9.2 Hz trans diaxial coupling constant was observed
between 1′-H and 2′-H in the glycoside. For comparison, it was 7.6 Hz in
metabolite 4.
(19) Addition of HOAc was required because of the basic lability of the
acetate group in the glycosidic moiety.
(11) An independent stability study of 4 showed that only 60% of the
parent compound remained after 3 days at room temperature, in pH 7 buffer
solution. However, 4 was relatively stable at -10 °C. The instability of the
bridge ether may be due to the inherent strain in the ring system.
Org. Lett., Vol. 7, No. 3, 2005
413

Scheme 4a
a Key: (a) TESOTf, Et3N, CH2Cl2, 90%; (b) K2CO3, EtOH, H2O, 75%; (c) Schmidt reagent B, CH2Cl2, BF3OEt2, 70%; (d) TBAF,
HOAc, rt, 80%; (e) Pb[OAc]4, CH2Cl2, NaHCO3, 5 min, rt; (f) Et3N/MeOH/H2O ) 1:1:4, THF, 40% for two steps.
At this point in the synthetic plan, it was envisioned that
nucleophilic attack by the hydroxyl group to the arene-
oxonium ion20 in intermediate 30, generated from a two-
electron oxidation of 29, would lead to the desired six-
membered bridge ether 31. To this end, a number of
oxidants21 were employed (DDQ, PIDA, and ferric isocy-
anate), but these attempts produced only a small amount of
the desired product 31, along with complex mixtures.
Interestingly, a careful examination of the crude material
identified the presence of aldehyde A (see Scheme 3), which
may have resulted from a retro-Aldol fragmentation of 32,11
product 31, which without further purification underwent
hydrolysis under basic conditions.22 Purification of the crude
product by reverse-phase HPLC provided the desired zwit-
terionic compound (+)-4, in 40% yield for the two steps.23
To our delight, the synthetic material was found to be
identical in all respects to the metabolite (1H NMR, 13C
NMR, MS, coelution on HPLC, [R]D ) +87.5° in 1:2 H2O/
AcCN, c ) 0.8). The other stereoisomers, (+)-23, (+)-24,
and (-)-24, were also converted into the analogous deriva-
tives by application of the same procedures. Upon HPLC
comparison,24 it was found that they were all distinct from
the metabolite.
In summary, the confirmation of the structural assignments
of the metabolites 4 and 5 of the dihydrobenzoxathiin
derivative 1 was accomplished by an independent, 19-step
total synthesis.
Acknowledgment. We gratefully acknowledge Professor
Barry M. Trost for helpful discussions during his scientific
consultations.
an intermediate that could rapidly degrade from the bridge
ether 31 in the acidic medium. The putative companion
intermediate (not shown) was not isolable. After extensive
experimentation, clean oxidation was accomplished by brief
exposure of 29 to Pb[OAc]4 buffered with either NaHCO3
or pyridine. This protocol consistently gave rise to the desired
Supporting Information Available: Characterization (1H
NMR, 2D NMR, 13C NMR, and HPLC) of synthetic and
authentic 4 and 5 and spectroscopic data for new compounds.
This material is available free of charge via the Internet at
http://pubs.acs.org.
OL047741F
(20) (a) Quideau, S.; Lebon, M.; Lamidey, A. Org. Lett. 2002, 4, 3975.
(b) Quideau, S.; Pouysegu, L.; Oxoby, M.; Looney, M. A. Tetrahedron
2001, 57, 319.
(21) (a) Venkateswarlu, R.; Kamakshi, C.; Moinuddin, S. G. A.; Subhash,
P. V.; Ward, R. S.; Pelter, A.; Coles, S. J.; Hursthouse, M. B.; Light, M. E.
Tetrahedron 2001, 57, 273 (b) Wipf, P.; Km, Y. J. Org. Chem. 1993, 58,
1649.
(22) Vaccaro, W.; Davis, H. R., Jr. Bioorg. Med. Chem. Lett. 1998, 8,
313.
(23) Metachem Basic C8 4.6 mm × 30 mm, 5 µm, H2O/CH3CN.
(24) HPLC conditions: Phenomenex-Luna 5 µM Phenyl-Hexyl, 4.6 mm
× 250 mm, H2O/CH3CN.
414
Org. Lett., Vol. 7, No. 3, 2005