Stereostructure of a novel cytotoxic 18-membered macrolactone antibiotic FD-891.

Article abstract of DOI:10.1021/ol026518k

The absolute stereochemistry of FD-891, a novel cytotoxic 18-membered macrolactone antibiotic, was determined by a synthetic approach as well as X-ray diffraction of degradative derivatives. The absolute configuration of FD-891 turned out to be as shown a

Full text of DOI:10.1021/ol026518k

ORGANIC
LETTERS
2002
Vol. 4, No. 20
3383-3386
Stereostructure of a Novel Cytotoxic
18-Membered Macrolactone Antibiotic
FD-891
,†
Tadashi Eguchi,* Kayako Kobayashi,† Hidehiro Uekusa,† Yuji Ohashi,†
,§
Kazutoshi Mizoue,‡ Yoshitaka Matsushima,§ and Katsumi Kakinuma*
Department of Chemistry and Materials Science, Tokyo Institute of Technology,
O-okayama, Meguro-ku, Tokyo 152-8551, Japan, Research Laboratories, Taisho
Pharmaceutical Co., Ltd., 1-403, Yoshino-cho, Saitama-shi, Saitama 330-8530, Japan,
and Department of Chemistry, Tokyo Institute of Technology, O-okayama, Meguro-ku,
Tokyo 152-8551, Japan
eguchi@cms.titech.ac.jp
Received July 14, 2002
ABSTRACT
The absolute stereochemistry of FD-891, a novel cytotoxic 18-membered macrolactone antibiotic, was determined by a synthetic approach as
well as X-ray diffraction of degradative derivatives. The absolute configuration of FD-891 turned out to be as shown above. The stable conformer
of FD-891 was also discussed with respect to biological activity by comparison with the structurally related concanamycin A on the bases of
molecular mechanics calculations.
FD-891 (1) was isolated from the fermentation broth of
Streptomyces graminofaciens A-8890 and was shown to have
cytotoxic activity in vitro against several tumor cell lines.1
The planar structure of 1 was elucidated to be an 18-
membered macrolactone by spectroscopic means as shown
in Figure 1,2 which appeared to be related to a specific
inhibitor of vacuolar-type H+-ATPase, concanamycin A.3
In recent years, concanamycin A has been shown to
specifically inhibit perforin-dependent cytotoxic T lympho-
cyte (CTL)-mediated cytotoxicity but not affect Fas ligand
† Department of Chemistry and Materials Science, Tokyo Institute of
Technology.
‡ Taisho Pharmaceutical Co., Ltd.
§ Department of Chemistry, Tokyo Institute of Technology.
(1) Seki-Asano, M.; Okazaki, T.; Yamagishi, M.; Sakai, N.; Hanada,
K.; Mizoue, K. J. Antibiot. 1994, 47, 1226.
(2) Seki-Asano, M.; Tsuchida, Y.; Hanada, K.; Mizoue, K. J. Antibiot.
1994, 47, 1234.
(3) (a) Muroi, M.; Shiragami, N.; Nagao, K.; Yamasaki, M.; Takatsuki,
A. Cell Struct. Funct. 1993, 18, 139. (b) Drose, S.; Bindseil, K. U.; Bowman,
E. J.; Siebers, A.; Zeeck, A.; Altendorf, K. Biochemistry 1993, 32, 3902.
Figure 1. Structures of FD-891 (1) and concanamycin A.
10.1021/ol026518k CCC: $22.00 © 2002 American Chemical Society
Published on Web 09/04/2002

(FasL)-dependent CTL-mediated cytotoxicity. These two
cytotoxic pathways play an essential role in the maintenance
of tissue homeostasis.4 In contrast, 1 was found to potently
prevent both perforin- and FasL-dependent CTL-mediated
killing pathways. Moreover, 1 was, in contrast to concana-
mycin A, unable to inhibit vacuolar acidification.5 These
specific biological features of 1 prompted us to study the
chemical basis of its activity in relation to concanamycin A
and related cytotoxic macrolactone compounds. However,
since the stereochemistry of 1 has not been elucidated to
date, we first pursued chemical studies to elucidate the
absolute stereochemistry as well as the stable conformation
of 1.
Since NMR analysis, including various NOE experiments,
was not informative for the configurational analysis of 1 and
modification into a crystalline derivative was not successful
either, we undertook degradation studies of 1. First, 1 was
benzoylated in a conventional way to the corresponding
tetrabenzoate in high yield, which was then subjected to
ozonolyis. After reductive workup with NaBH4 of the
resulting ozonides, two fragments were isolated and found
to be the C5-C12 fragment 2 and the C19-C26 fragment 3
by spectroscopic analysis as shown in Scheme 1.6 Interest-
these conditions. The reason such an unusual oxidative
cleavage occurred is not clear. Moreover, a fragment
comprised of C15-C26 was not obtained at all under several
reaction conditions.
For the unambiguous assignment of the absolute stereo-
chemistry of these fragments, we attempted chemical trans-
formation of 2 and 3 into crystalline derivatives for X-ray
analysis. Fortunately, the dicamphanic ester derivative 4,
which was obtained by treatment of 2 with (-)-camphanic
chloride in pyridine in the presence of 4-(dimethylamino)-
pyridine, formed suitable crystals for X-ray analysis by
recrystallization from hexanes-ethyl acetate. The absolute
stereochemistry of the C6-C10 of 1 was clearly determined
as shown in Figure 2 simply on the basis of the absolute
Scheme 1a
Figure 2. ORTEP drawings of 4 and 5 (thermal ellipsoids at the
25% probability level) and the absolute stereostructures of 4 and
5.
structure of (-)-camphanic ester structure. As to the fragment
3, hydrolysis with MeOK-MeOH, followed by esterification
similarly with (-)-camphanic chloride gave crystalline
derivative 5 by recrystallization again from hexanes-ethyl
acetate. Thus, the absolute stereochemistry of the C21-C25
portion of 1 was also determined by X-ray crystallography
as shown in Figure 2.
At this stage, the remaining stereogenic centers to be
determined were C17 and C18. Because a fragment containing
C17 and C18 was not available from ozonolysis of the
tetrabenzoate derivative as described above, we modified a
substrate of ozonolysis. Thus, 1 was first subjected to
reduction with DIBAL, followed by benzoylation with
benzoyl chloride in pyridine, to give pentabenzoate 6 and
the corresponding hexabenzoate. The minor product 6 turned
out to be crucial. A fragment 7 containing C17 and C18
stereogenic centers was obtained in 54% yield as an unusual
oxidative cleavage and benzoyl migrated product, only when
a Reaction conditions: (a) BzCl, DMAP/py; (b) O3/MeOH, and
then NaBH4; (c) (-)-camphanic chloride, DMAP/py; (d) MeOK-
MeOH; (e) DIBAL/CH2Cl2; (f) BzCl/py; (g) (R)-(-)-MTPACl,
DMAP/py.
ingly, both benzoyl groups were migrated from secondary
hydroxy groups to primary ones in the fragment 2. Unex-
pectedly further, the fragment 3 appeared to be the product
by unusual oxidative cleavage of the C18-C19 bond under
(4) Kataoka, T.; Shinohara, N.; Takayama, H.; Takaku, K.; Kondo, S.;
Yonehara, S.; Nagai, K. J. Immunol. 1996, 156, 3678.
(5) Kataoka, T.; Yamada, A.; Bando, M.; Honma, T.; Mizoue, K. Nagai,
K. Immunology 2000, 100, 170.
3384
Org. Lett., Vol. 4, No. 20, 2002

6 was used as a substrate of ozonolysis (Scheme 1).6 The
reason is not clear again. Since the fragment 7 included only
two stereogenic centers, we took advantage of the chemical
synthesis of the fragment 7 in a stereochemically defined
manner. The relative stereochemistry of the fragment 7 was
(R)-(-)-MTPA chloride to give di-(-)-MTPA ester 14. The
1H NMR spectrum of naturally derived (-)-MTPA ester 14
was quite different from that of the synthetic (-)-MTPA ester
13b but completely identical to that of (+)-MTPA ester 13a
as shown in Figure 3. Consequently, the absolute stereo-
1
envisioned to be syn (S,S or R,R), because the H-1H
coupling constant between two stereogenic centers of 7 was
relatively small (J ) 3.2 Hz). Consequently, we synthesized
the authentic sample from the known (R,R)-diol 87 as shown
in Scheme 2. Protection of hydroxy groups of 8 with a TBS
Scheme 2a
a Reaction conditions: (a) TBSCl, imidazole/DMF; (b) H2, Pd-
C/EtOH; (c) MsCl, Et3N/CH2Cl2; (d) NaCN/DMSO; (e) DIBAL/
toluene; (f) NaBH4/MeOH; (g) 2 N HCl/THF; (h) (S)-(+)-MTPACl,
DMAP/py or (R)-(-)-MTPACl, DMAP/py.
Figure 3. Partial 1H NMR spectra (500 MHz, CDCl3) of (A) (+)-
MTPA ester 13a, (B) naturally derived (-)-MTPA ester 14, and
(C) (-)-MTPA ester 13b.
group and deprotection of the benzyl group by hydrogenation
afforded 9 in high yield. The primary hydroxy group of 9
was displaced with a cyano group via its mesylate to give
10. Reduction of 10 with DIBAL, followed further by NaBH4
reduction, gave alcohol 11 in 65% yield. Acidic hydrolysis
of 11 with 2 N HCl yielded triol 12.
chemistry of C17-C18 of 1 was proven to be enantiomeric
to the synthetic triol 12. In conclusion, by combining all the
data discussed above, the absolute stereochemistry of 1 was
unequivocally determined as shown in Figure 4.
The triol 12 was converted into di-(+)- or (-)-MTPA
esters 13a or 13b, respectively, in order to be compared with
the corresponding naturally derived MTPA ester. Thus,
the above-mentioned fragment 7 was hydrolyzed with
MeOK-MeOH, and the resulting triol was then treated with
(6) Spectroscopic data of compounds 2, 3, and 7. Compound 2: 1H NMR
(500 MHz, CDCl3) δ 1.07 (d, J ) 7.3 Hz, 3H), 1.98-2.12 (m, 2H), 2.21-
2.26 (m, 1H), 3.14 (dd, J ) 2.3, 3.2 Hz, 1H), 3.23 (dd, J ) 2.3, 4.2 Hz,
1H), 3.81 (m, 1H), 3.98 (m, 1H), 4.26 (dd, J ) 5.9, 11.1 Hz, 1H), 4.44 (dd,
J ) 8.0, 11.0 Hz, 1H), 4.51 (dt, J ) 11.3, 5.7 Hz, 1H), 4.56 (ddd, J ) 5.3,
8.4, 11.2 Hz, 1H), 7.42 (m, 4H), 7.56 (m, 2H), 8.02 (m, 4H); 13C NMR
(126 MHz, CDCl3) δ 10.9, 33.8, 35.9, 57.1, 58.0, 61.3, 66.4, 67.2, 68.3,
128.4, 128.4, 128.5, 129.6, 129.6, 130.0, 133.1, 133.1, 166.7, 166.7.
Compound 3: 1H NMR (500 MHz, CDCl3) δ 1.04 (d, J ) 7.0 Hz, 3H),
1.15 (d, J ) 6.2 Hz, 3H), 1.19 (d, J ) 7.1 Hz, 3H), 1.80-1.81 (m, 1H),
1.93-2.00 (m, 1H), 2.01-2.04 (m, 1H), 2.26 (double quintet, J ) 2.0, 7.1
Hz, 1H), 3.19 (s, 3H), 3.24 (dq, J ) 6.1, 6.2 Hz, 1H), 3.49 (m, 1H), 3.65
(m, 1H), 5.33 (dd, J ) 4.1, 7.6 Hz, 1H), 5.52 (dt, J ) 9.7, 2.3 Hz, 1H),
7.26 (dd, J ) 7.7, 7.8 Hz, 2H), 7.36 (t, J ) 7.7 Hz, 2H), 7.44 (t, J ) 7.3
Hz, 1H), 7.51 (dd, J ) 8.3, 1.2 Hz, 1H), 7.86 (dd, J ) 8.3, 1.2 Hz, 2H),
7.94 (d, J ) 8.2 Hz, 2H); 13C NMR (126 MHz, CDCl3) δ 9.7, 11.4, 16.2,
36.0, 39.0, 40.2, 56.4, 58.4, 69.9, 76.1, 79.0, 128.1, 128.3, 129.5, 129.6,
130.2, 130.3, 132.6, 132.9, 166.1, 167.3. Compound 7: 1H NMR (500 MHz,
CDCl3) δ 0.95 (d, J ) 6.9 Hz, 3H), 1.55 (ddt, J ) 4.7, 14.3, 6.4, Hz, 1H),
1.74 (ddt, J ) 6.2, 14.5, 7.2 Hz, 1H), 1.84 (double sextet, J ) 3.2, 6.8 Hz,
1H), 1.90 (m, 2H), 3.67 (ddd, J ) 4.6, 7.6, 10.9 Hz, 1H), 3.76 (m, 2H),
4.44 (ddd, J ) 5.6, 5.6, 11.2 Hz, 1H), 4.61 (dt, J ) 11.3, 6.8 Hz, 1H), 7.45
(t, J ) 7.8 Hz, 2H), 7.57 (m, 1H), 8.04 (m, 2H); 13C NMR (126 MHz,
CDCl3) δ 14.4, 32.8, 36.0, 36.4, 60.7, 62.7, 71.4, 128.4, 129.6, 133.0, 167.1.
Figure 4. Absolute stereochemistry of FD-891 (1).
As described above, FD-891 (1) appeared to be structurally
related to concanamycin A. However, the biological char-
acteristics of these two natural products are fairly different.
These findings raised a question of how 1 and concanamycin
A exert their effect at the molecular level. As the first step,
we initiated detailed conformational and structure-activity
investigations of 1 and concanamycin A. Molecular modeling
approach has been frequently used to determine the lower
(7) (a) Arai, H.; Matsushima, Y.; Eguchi, T.; Shindo, K.; Kakinuma, K.
Tetrahedron Lett. 1998, 39, 3181. (b) Ireland, R. E.; Thaisrivongs, S.;
Dussault, P. H. J. Am. Chem. Soc. 1988, 110, 5768.
Org. Lett., Vol. 4, No. 20, 2002
3385

energy conformations of a molecule of interest. We took
advantage of recent advances in molecular modeling tech-
niques to carry out conformational analysis of 1 and
concanamycin A.8 The calculated most stable conformation
of 1 is shown in Figure 5.
arguments of these two compounds. An interesting feature
found in the structure of concanamycin A was a hydrogen-
bond network through lactone carbonyl/19-hydroxy group/
six-membered hemiketal hydroxy group, which could restrict
the side chain conformation. The structure-activity relation-
ship studies of concanamycins previously revealed that the
18-membered lactone ring conformation and the hydrogen-
bond network were crucial for the inhibitory effects on
vacuolar-type H+-ATPase activity as well as lysosomal
acidification, but the carbohydrate residue was dispensable
for these activities.3b,15
As can be seen in the superimposed structures of 1 and
concanamycin A (Figure 6), although the hydrogen bond
Figure 5. Most stable conformation of FD-891. Values are the
1
calculated H-1H coupling constants (in hertz) deduced from the
Karplus-Altona equation with Boltzmann averaging.13 Experi-
mentally observed data are given in parentheses.
The Boltzmann averaged 1H-1H coupling constants
deduced from the Karplus-Altona equation13 are also
summarized in Figure 5. These values are in good accordance
with the observed coupling constants, which supports well
the reliability of these calculations. As to concanamycin A,
the most stable conformation is quite similar to the crystalline
state structure.14 These results secure the conformational
Figure 6. Superimposed view of FD-891 and concanamycin A.
(Red) FD-891, (blue) concanamycin A. Dashed lines indicate a
hydrogen-bond network.
(8) A conformational search and energy minimization of FD-891 and
concanamycin A were performed using the MMFF force field9 in Macro-
model version 6.5.10 All calculations were performed using the implicit
CHCl3 GB/SA solvation model of Still et al.11 A conformational search
was performed using the Monte Carlo method of Goodman and Still.12 For
each search, 20 000 starting structures were generated and minimized to
an energy convergence of 0.05 (kJ/mol)/(Å/mol) using the full matrix
Newton-Raphson method implemented in Macromodel. Duplicated struc-
tures and those greater than 50 kJ/mol above the global minimum were
discarded.
(9) (a) Halgren, T. A. J. Comput. Chem. 1996, 17, 490. (b) Halgren, T.
A. J. Comput. Chem. 1996, 17, 520. (c) Halgren, T. A. J. Comput. Chem.
1996, 17, 553. (d) Halgren, T. A. J. Comput. Chem. 1996, 17, 587.
(10) Macromodel Version 6.5; Department of Chemistry, Columbia
University: New York.
(11) (a) Hasel, W.; Hendrickson, T. F.; Still, W. C. Tetrahedron Comput.
Methodol. 1988, 1, 103. (b) Still, W. C.; Tempczyk, A.; Hawley, R. C.;
Hendrickson, T. J. Am. Chem. Soc. 1990, 112, 6127.
(12) (a) Chang, G.; Guida, W. C.; Still, W. C. J. Am. Chem. Soc. 1989,
111, 5239. (b) Goodman, J. M.; Still, W. C. J. Comput. Chem. 1991, 12,
1110. (c) Guida, W. C.; Bohacek, R. S.; Erion, M. D. J. Comput. Chem.
1992, 13, 214. (d) Kolossvary, I.; Guida, W. C. J. Chem. Inf. Comput. Sci.
1992, 32, 191. (e) Kolossvary, I.; Guida, W. C. J. Comput. Chem. 1993,
14, 691. (f) Kolossvary, I.; Guida, W. C. J. Am. Chem. Soc. 1993, 115,
2107.
network exists in the most stable conformation of 1, the
conformation of macrolactone ring and the direction of the
side-chain are quite different from each other. It appears
therefore that, although the major chemical difference
between FD-891 and concanamycin A is obviously found
in their side-chain structure, the macrolactone conformation
is also significantly different, which may well affect their
biological activity. Further structure-activity relationship
study of this class of antibiotics is underway.
Supporting Information Available: Experimental pro-
cedure of compounds 2-12 and X-ray crystallographic data
of 4 and 5. This material is available free of charge via the
Internet at http://pubs.acs.org.
OL026518K
(13) Haasnoot, C. A. G.; De Leeuw, F. A. M.; Altona, C. Tetrahedron
1980, 36, 2783.
(14) Nakai, H.; Matsuhashi, S. Acta Crystallogr. 1992, C48, 1519.
(15) (a) Droese, S.; Boddien, C.; Gassel, M.; Ingenhorst, G.; Zeeck, A.;
Altendorf, K. Biochemistry 2001, 40, 2816. (b) Woo, J. T.; Shinohara, C.;
Sakai, K.; Hasumi, K.; Endo, A. J. Antibiot. 1992, 45, 1108.
3386
Org. Lett., Vol. 4, No. 20, 2002