Concise total synthesis and structure assignment of TAN-1085

Article abstract of DOI:10.1002/anie.200453801

Rapid access to the tetracyclic core of TAN-1085 (1) was possible by exploiting three efficient processes: a tandem electrocyclic reaction (A), a pinacol cyclization (B), and a regioselective monobenzoylation of a 1,2-diol (C), which allowed the first total synthesis of 1.

Full text of DOI:10.1002/anie.200453801

Angewandte
Chemie
Natural Products Synthesis
Concise Total Synthesis and Structure Assignment
of TAN-1085**
Ken Ohmori, Keiji Mori, Yuji Ishikawa,
Hideyuki Tsuruta, Shunsuke Kuwahara,
Nobuyuki Harada, and Keisuke Suzuki*
TAN-1085 (1), an antibiotic produced by Streptomyces sp.
S-11106,[1] has attracted considerable synthetic interest, not
only for its important biological activities, angiogenesis and
aromatase inhibition, but also for its unique structural
features, that is, a curved tetracyclic chromophore with a
vicinal diol, one of which is glycosylated with a rhodinose
unit.[1,2] The relative and absolute stereochemistries of this
compound, however, remained to be assigned.
Herein we report the first totalsynthesis of 1 and its 5,6-
bis-epimer, thereby finally establishing the remaining stereo-
chemicalquestions. The synthesis features three key trans-
formations: 1) a tandem electrocyclic reaction for the facile
construction of the aglycon precursor, 2) its stereoselective
conversion into a trans-diolby SmI -mediated pinacolcycil-
2
zation, and 3) discrimination of the resulting C5/C6 hydroxy
groups by in situ benzoylation. Furthermore, the stereochem-
icalassignment of 1 was effected by exploiting the CD exciton
chirality method.[3]
Scheme 1 outlines our retrosynthetic analysis, predicated
upon the construction of tetracyclic aglycon I from biaryl
[*] Dr. K. Ohmori, K. Mori, Y. Ishikawa, Dr. H. Tsuruta, Prof. Dr. K. Suzuki
Department of Chemistry
Tokyo Institute of Technology, CREST-JST Agency
2-12-1 O-okayama, Meguro-ku, Tokyo 152–8551 (Japan)
Fax: (+81)3-5734-2788
E-mail: ksuzuki@chem.titech.ac.jp
S. Kuwahara, Prof. Dr. N. Harada
Institute of Multidisciplinary Research for Advanced Materials
Tohoku University
2-1-1 Katahira, Aoba-ku, Sendai 980–8577 (Japan)
[**] This work was partially supported by the 21 st Century COE program
(Tokyo Institute of Technology). We thank Takeda Chemical
Industries, Ltd. for providing an authentic sample of natural TAN-
1085.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
Angew. Chem. Int. Ed. 2004, 43, 3167 –3171
DOI: 10.1002/anie.200453801
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3167

Communications
[4]
dialdehyde II by pinacolcycizlation.
We
hoped that the sterically encumbered biaryl
structure in II (or the precursor II’) could be
accessible by the ring enlargement of benzo-
cyclobutene III by way of tandem thermal
electrocyclic reactions via A.[5,6] The key
intermediate III could be obtained by assem-
bly of styryl anion IV and the selectively
protected benzocyclobutenedione V, which in
turn is readily available from silyl enol ether VI
and benzyne VII.[7]
Scheme 2 illustrates the preparation of 9
for the key electrocyclizations. Known bro-
mide 2[8] was converted into aldehyde 3 by
successive treatment with MeLi and nBuLi[9]
followed by DMF. After silylation, the MOM
group proximalto the carbonylfunction was
selectively removed, giving phenol 4 in good
yield. Benzylation of 4 was followed by dibro-
moolefination[10] to give 5 in 85% yield, which
was successively treated with nBuLi and para-
formaldehyde to afford propargyl alcohol 6 in
91% yield. Hydroalumination of 6 with Red-
Al followed by quenching with iodine gave,
after silylation, vinyl iodide 7. Union of
benzocyclobutenone 8[7] and the vinyllithium
species generated from 7 followed by treat-
ment with methyltrifal te in situ gave 9, ready
for the key tandem electrocyclic reaction.
The planned reactions of 9, however, were
Scheme 2. Preparation of 9. a) MeLi (1.1 equiv), THF, ꢀ788C, 10 min; then nBuLi
(1.2 equiv), DMF, ꢀ788C, 5 min (97%); b) TBDPSCl, imidazole, DMF, 08C!RT, 78%;
c) Montmorillonite K-10, benzene, room temperature, 11 h, 408C, 50 min; d) BnBr,
NaH, DMF, 08C!RT, 10 h, 75% (two steps); e) CBr4, PPh3, C HCl2, 08C, 20 min, 85%;
2
f) nBuLi (2.0 equiv), THF; then (HCHO)n, ꢀ788C!RT, 91%; g) Red-Al, Et2O,
ꢀ78!ꢀ108C, 50 min; then EtOAc, ꢀ108C, 15 min; then I2, THF, ꢀ78!08C , 1 h;
h) TBDMSCl, imidazole, DMF, 08C!RT, 1.5 h, 93%; i) 7 (1.0 equiv), tBuLi (2.1 equiv),
Et2O, ꢀ788C, 20 min; then 8 (1.5 equiv), THF, ꢀ788C, 25 min; then MeOTf (3.0 equiv),
ꢀ78!08C, 1.5 h, 85%. DMF=N,N-dimethylformamide, TBDPS=tert-butyldiphenylsilyl,
TBDMS=tert-butyldimethylsilyl, Red-Al=NaAlH2(OCH2CH2OCH3)2, Tf=trifluoro-
methanesulfonyl.
uniformly unsuccessful under various conditions owing to the
prevalence of another pericyclic process, the 1,7-hydrogen
shift shown in B (Scheme 3). For example, although the
starting material 9 was completely consumed upon refluxing
in toluene for 4 h, only silyl ether 11 was obtained.
Scheme 3. Attempts at thermal electrocyclic reactions.
After considerable experimentation, we discovered a
highly efficient solution to this problem [Eq. (1)]. Simply by
raising the oxidation level of C6 in 9, the 1,7-shift was
completely suppressed, and the desired 6p process was
remarkably facilitated. Thus, after removal of the two silyl
groups in 9, the resulting diol 12 was subjected to Swern
oxidation. Importantly, when the Swern oxidation mixture
was warmed to 258C and kept standing for 8 h, the ring
opening of C and subsequent 6p-closure occurred, thereby
directly giving the biaryl dialdehyde 13, the key intermediate
in our synthetic plan. This remarkable rate enhancement of
the pericyclic processes could be rationalized by the cooper-
Scheme 1. Retrosynthetic analysis of 1.
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3168
www.angewandte.org
Angew. Chem. Int. Ed. 2004, 43, 3167 –3171

Angewandte
Chemie
ative effect of the electron-donating (2 ” OMe) and the
electron-withdrawing (enal) substituents of the four-mem-
bered ring toward the required conrotation.[11]
The biaryldiadlehyde
13 was subjected to a pinacol
cyclization to the tetracyclic structure (Scheme 4). Treatment
[12]
of 13 with SmI2 gave the corresponding diols, trans-14 and
Scheme 4. Pinacol cyclization and discrimination of the two hydroxy
cis-14 (9:1), which were separable by column chromatogra-
phy.[13,14] For the regioselective introduction of the sugar
moiety, discrimination of the C5/C6 hydroxy groups in trans-
14 was required, a challenging task owing to the local C2
symmetry. All attempts at the monofunctionalization of diol
14 were unfruitful; no regioselectivity and/or double reaction
at the C5/C6 diols occurred.
Gratifyingly, a felicitous solution to this issue was offered
by direct quenching of the pinacolcycilzation (see above).
Thus, after treatment of 13 with SmI2 as above, benzoyl
chloride (1.5 equiv) was directly added to give C5-benzoate
trans-15 in 86% yield. Neither the C6 benzoate nor the
groups. a) SmI2, THF, 08C, 2 min, 96%, trans/cis=9:1; b) SmI2, THF,
08C, 2 min; then BzCl (1.5 equiv), 86%.
dibenzoate was detected, which could be rationalized by the
difference in the nucleophilicity of two oxygen functions: The
C6 samarium alkoxide is chelated as in I, and is therefore less
reactive than the C5 alkoxide.[15]
Construction of the aglycon portion was completed in
three steps to give methylester 17, primed for introduction of
the sugar moiety (Scheme 5). Tetracyclic 17 was glycosylated
by reaction with l-rhodinosylacetate 19[16,17] in the presence
Scheme 5. Construction of the aglycon 17 and its glycosylation with l-rhodinoside 19. a) H2SO4 (aq., 0.5m), DME, 608C, 12 h, 91%; b) PhNTf2,
K2CO3, acetone, 08C, 3 h, 98%; c) CO (3 atm), Pd(OAc)2 (30 mol%), dppp (30 mol%), Et3N, MeOH, DMF, 658C , 30 h, 91%; d) H, Pd/C, MeOH,
2
room temperature, 30 min; e) HCl (2m), AcOH, THF (1:1:1), room temperature, 7 h, 80% (two steps); f) Ac2O, pyridine, room temperature,
30 min, 94%, a/b=1:1.4; g) BF3·OEt2 (1.1 equiv), 19 (2.0 equiv, a/b=1:1.4), molecular sieves (4 Š), CH2Cl2, 95%, 1:1 mixture of diastereomers;
h) DIBAL, CH2Cl2, ꢀ78!ꢀ208C, 1 h, 98%. dppp=1,3-bis(diphenylphosphanyl)propane, DIBAL=diisobutylaluminum hydride.
Angew. Chem. Int. Ed. 2004, 43, 3167 –3171
www.angewandte.org
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3169

Communications
of BF3·OEt2 to give the a glycoside 20 as an inseparable
mixture of diastereomers (1:1). These isomers were the
diastereomers arising from the racemic nature of the aglycon
relative to the l-rhodinoside. The anomeric stereochemistries
1
of the diastereomers were both shown to be a by H NMR
spectroscopic analysis.[18] Without separation, methylester 20
was treated with DIBAL, which led to the simultaneous
removalof the two benzoylgroups to give diastereomeric
products, which were separable by column chromatography
(CHCl3/EtOAc = 1:30); 21 (Rf = 0.10), 22 (Rf = 0.17).
In spite of considerable efforts, we were frustrated by the
inability to directly assign the stereochemistry of these
diastereomers, and decided to rely on indirect methods.
Thus, the sugar was detached from the more-polar isomer 21
to obtain the “resolved” aglycon 23 (with unknown absolute
stereochemistry at this stage), which was converted into
benzoate 24[19] for measuring the CD exciton chirality effects
(Figure 1a).[3]
Scheme 6. Construction of 1. a) CAN, H2O, CH3CN, 08C, 20 min,
b) H2, Pd/C, MeOH (53% for 25 from 21, 56% for 26 from 22).
CAN=ceric ammonium nitrate.
The CD spectrum of alcohol 23 showed strong and
complicated Cotton effects because of the markedly twisted
p-electron system in the dihydrobenz[a]anthracene skeleton
(blue line, Figure 1b). The CD spectrum of benzoate 24 also
showed intense and complicated Cotton effects similar to
those of 23 (red line). Therefore, to detect the genuine exciton
CD Cotton effects due to the interaction between the
long axes of naphthalene and benzoate chromo-
phores, the difference CD curve was calculated:
DDe = De(24)ꢀDe(23). As shown in Figure 1c, the
difference CD curve shows a typicalpattern of exciton
split CD, DDe = ꢀ32 at 236 nm (the first Cotton
effect) and DDe = + 35 at 224 nm (the second Cotton
effect), A = DDe (1st)ꢀDDe (2nd) = ꢀ67; the negative
sign of the A value proved the counterclockwise
relationship between these chromophores (Fig-
ure 1d). The 5S,6S configuration was thus unambig-
uously assigned.[20]
With the structures secured, diastereomeric triols
21 and 22 were converted into the finalproducts by
oxidation with CAN followed by deprotection
(Scheme 6). Of the diastereomers 25 (from 21) and
26 (from 22), the former coincided with the authentic
specimen of 1.[21]
In conclusion, the first synthesis of TAN-1085 was
completed through the concise and efficient construc-
tion of the aglycon. The stereochemistry of natural
product was concluded to be that of 25.
Received: January 20, 2004 [Z53801]
Keywords: antibiotics · electrocyclic reactions ·
.
glycosylation · structure elucidation · total synthesis
[1] T. Kanamaru, Y. Nozaki, M. Muroi (Kokai Tokkyo
Koho), JP 02-289-532/1990, 1991 [Chem. Abstr. 1991,
115, 47759n].
[2] For synthetic studies, see: a) D. Mal, H. N. Roy, N. K.
Hazra, S. Adhikari, Tetrahedron 1997, 53, 2177 – 2184;
b) D. S. Larsen, M. D. OꢀShea, J. Org. Chem. 1996, 61,
5681 – 5683; c) F. M. Hauser, W. A. Dorsch, D. Mal,
Org. Lett. 2002, 4, 2237 – 2239.
Figure 1. Application of the exciton chirality method. a) Conformations of tetracycle 23
and 24, b) CD and UV spectra of 23 and 24, c) difference CD curve between 24 and 23:
DDe=De(24)ꢀDe(23), d) Anticlockwise relationship between the long axis of the naph-
thalene and the benzoate chromophores.
3170
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
Angew. Chem. Int. Ed. 2004, 43, 3167 –3171

Angewandte
Chemie
[3] N. Harada, K. Nakanishi, Circular Dichroic Spectroscopy
Exciton Coupling in Organic Stereochemistry, University Science
Books, Mill Valley, 1983.
[4] a) K. Ohmori, M. Kitamura, K. Suzuki, Angew. Chem. 1999, 111,
1304 – 1307; Angew. Chem. Int. Ed. 1999, 38, 1226 – 1229; b) M.
Kitamura, K. Ohmori, T. Kawase, K. Suzuki, Angew. Chem.
1999, 111, 1308 – 1311; Angew. Chem. Int. Ed. 1999, 38, 1229 –
1232.
[5] For leading references, see: a) D. K. Jackson, L. Narasimhan,
J. S. Swenton, J. Am. Chem. Soc. 1979, 101, 3989 – 3990; b) L. S.
Liebeskind, S. Iyer, C. F. Jewell, Jr., J. Org. Chem. 1986, 51,
3065 – 3066; c) S. T. Perri, L. D. Foland, O. H. W. Decker, H. W.
Moore, J. Org. Chem. 1986, 51, 3067 – 3068; d) D. N. Hickman,
T. W. Wallace, J. M. Wardleworth, Tetrahedron Lett. 1991, 32,
819 – 822.
[6] a) T. Matsumoto, T. Hamura, M. Miyamoto, K. Suzuki, Tetrahe-
dron Lett. 1998, 39, 4853 – 4856; b) T. Hamura, M. Miyamoto, T.
Matsumoto, K. Suzuki, Org. Lett. 2002, 4, 229 – 232; c) T.
Hamura, M. Miyamoto, K. Imura, T. Matsumoto, K. Suzuki,
Org. Lett. 2002, 4, 1675 – 1678.
[17] B. Renneberg, Y.-M. Li, H. Laatsch, H.-H. Fiebig, Carbohydr.
Res. 2000, 329, 861 – 872.
[18] For both diastereomers, the coupling constants between the
anomeric proton and the methylenic protons at C2’ were small
(< 3.2 Hz), suggesting the a-stereochemistry of these anomeric
centers. The conclusions were further supported by the obser-
vation of 13C–1H coupling constants at anomeric centers, that is,
1JC(1’),H(1’) = 167.4 and 165.8 Hz; see: K. Bock, C. Pedersen, J.
Chem. Soc. Perkin Trans. 2 1974, 293 – 297.
[19] The more-polar isomer 21 was easily converted into alcohol 23
(1. H2, Pd/C, MeOH; 2. NaH, MeI; 3. TsOH·H2O, MeOH) in
87% overall yield. Alcohol 23 was further converted into
benzoate 24 (BzCl, pyridine) in 95% yield. The coupling
constants between 5- and 6-H (3.5 Hz for 23, 3.4 Hz for 24 in
CD3OD) suggest that the C5 and C6 substituents of both
compounds adopted pseudoaxialorientations (Figure 1a), a
prerequisite for determining the stereochemistry by CD.
[20] This conclusion was further confirmed by the 1H NMR aniso-
tropy method (see Supporting Information).
[21] All physical data of synthetic 25 (1H and 13C NMR, IR, [a]D)
were identicalto those of an authentic specimen of 1.
[7] T. Hamura, T. Hosoya, H. Yamaguchi, Y. Kuriyama, M. Tanabe,
M. Miyamoto, Y. Yasui, T. Matsumoto, K. Suzuki, Helv. Chim.
Acta 2002, 85, 3589 – 3604, and references therein.
[8] C. A. Townsend, S. G. Davis, S. B. Christensen, J. C. Link, C. P.
Lewis, J. Am. Chem. Soc. 1981, 103, 6885 – 6888.
[9] Simple treatment of 2 with 2 equivalents of nBuLi was ineffec-
tive and gave considerable amounts of the debrominated
product. This difficulty could be ascribed to the extreme rapidity
of the halogen–lithium exchange with nBuLi, which competes
with the deprotonation of the hydroxy group, thereby inducing
the partialquenching of the arylanion by internalproton
transfer, see: S. B. Rosenblum, R. Bihovsky, J. Am. Chem. Soc.
1990, 112, 2746 – 2748. For the effectiveness of MeLi to secure
the Li alkoxide formation prior to the halogen–lithium
exchange, see: T. Obitsu, K. Ohmori, Y. Ogawa, H. Hosomi, S.
Ohba, S. Nishiyama, S. Yamamura, Tetrahedron Lett. 1998, 39,
7349 – 7352.
[10] E. J. Corey, P. L. Fuchs, Tetrahedron Lett. 1972, 3769 – 3772.
[11] a) C. W. Jefford, G. Bernardinelli, Y. Wang, D. C. Spellmeyer, A.
Buda, K. N. Houk, J. Am. Chem. Soc. 1992, 114, 1157 – 1165;
b) K. Shishido, A. Yamashita, K. Hiroya, K. Fukumoto, T.
Kametani, Chem. Lett. 1987, 2113 – 2116; c) K. Rudolf, D. C.
Spellmeyer, K. N. Houk, J. Org. Chem. 1987, 52, 3708 – 3710;
d) S. Ingham, R. W. Turner, T. W. Wallace, J. Chem. Soc. Chem.
Commun. 1985, 1664 – 1666.
[12] For the preparation of SmI2/THF, see: P. Girard, J. L. Namy,
H. B. Kagan, J. Am. Chem. Soc. 1980, 102, 2693 – 2698.
[13] The major diastereomer of diol 14 was acetylated to give a
crystalline derivative, whose X-ray crystallographic analysis
confirmed the trans relationship at C5/C6. CCDC-225281
contains the supplementary crystallographic data for this
paper. These data can be obtained free of charge via
www.ccdc.cam.ac.uk/conts/retrieving.html(or from the Cam-
bridge Crystallographic Data Centre, 12, Union Road, Cam-
bridge CB21EZ, UK; fax: (+ 44)1223-336-033; or deposit@
ccdc.cam.ac.uk).
[14] Although the C5/C6 relative stereochemistry of 1 was unknown,
we decided to employ trans-14 for further transformation into 1
based on analogy to a related natural product, PD-116740: J. H.
Wilton, D. C. Cheney, G. C. Hokanson, J. C. French, J. Org.
Chem. 1985, 50, 3936 – 3938.
[15] A small amount (8%) of nonbenzoylated product, trans-14, was
also obtained, and interestingly, cis-14 (2%) was obtained
without benzoylation. The latter fact could be rationalized by
considering that the cis isomer forms bicyclic samarium chelate
II, in which both oxygen functions are poorly reactive.
[16] Prepared from known compound 18[17] in the three steps.
Angew. Chem. Int. Ed. 2004, 43, 3167 –3171
www.angewandte.org
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3171