Synthesis of NF-κB activation inhibitors derived from epoxyquinomicin C

Article abstract of DOI:10.1016/S0960-894X(00)00114-1

In order to develop new inhibitors of NF-κB activation, we designed and synthesized dehydroxymethyl derivatives of epoxyquinomicin C, namely, DHM2EQ and its regioisomer DHM3EQ. These derivatives were synthesized from 2,5-dimethoxyaniline in 5 steps. Since DHM2EQ was more active and less toxic than DHM3EQ, its stereochemical configuration was determined by X-ray crystallographic analysis. Each enantiomer of the protected DHM2EQ was separated by a chiral column and deprotected. DHM2EQ inhibited TNF-α-induced activation of NF-κB in human T cell leukemia cells, and also inhibited collagen-induced arthritis in a rheumatoid model in mice. (C) 2000 Elsevier Science Ltd. All rights reserved.

Full text of DOI:10.1016/S0960-894X(00)00114-1

Bioorganic & Medicinal Chemistry Letters 10 (2000) 865±869
Synthesis of NF-ꢀB Activation Inhibitors Derived from
Epoxyquinomicin C
Naoki Matsumoto, a Akiko Ariga, b Sakino To-e, b Hikaru Nakamura, a Naoki Agata, c
Shin-ichi Hirano, c Jun-ichiro Inoue d and Kazuo Umezawa b,*
aInstitute of Microbial Chemistry, 3-14-23 Kamiosaki, Shinagawa-ku, Tokyo 141-0021, Japan
bDepartment of Applied Chemistry, Faculty of Science and Technology, Keio University, 3-14-1 Hiyoshi,
Kohoku-ku, Yokohama 223-0061, Japan
cCentral Research Laboratories, Mercian Corp., 4-9-1 Johnan, Fujisawa, Kanagawa 251-0057, Japan
dInstitute of Medical Science, University of Tokyo, 4-6-8 Shirokanedai, Minato-ku, Tokyo 108-0071, Japan
Received 22 November 1999; accepted 11 February 2000
AbstractÐIn order to develop new inhibitors of NF-kB activation, we designed and synthesized dehydroxymethyl derivatives of
epoxyquinomicin C, namely, DHM2EQ and its regioisomer DHM3EQ. These derivatives were synthesized from 2,5-dimethoxy-
aniline in 5 steps. Since DHM2EQ was more active and less toxic than DHM3EQ, its stereochemical con®guration was determined
by X-ray crystallographic analysis. Each enantiomer of the protected DHM2EQ was separated by a chiral column and deprotected.
DHM2EQ inhibited TNF-a-induced activation of NF-kB in human T cell leukemia cells, and also inhibited collagen-induced
arthritis in a rheumatoid model in mice. # 2000 Elsevier Science Ltd. All rights reserved.
Introduction
additional hydroxymethyl group compared with pane-
poxydone that inhibits NF-kB. Therefore, we designed
and synthesized 5-dehydroxymethyl derivatives of
epoxyquinomicin C.
NF-kB is a transcription factor that mediates the
expression of a variety of cellular genes regulating the
in¯ammatory response. For example, the expression of
various cytokines such as IL-1, IL-2, IL-8 and TNF-a is
regulated by NF-kB. NF-kB is located in the cytoplasm
along with its endogenous inhibitor, I-kB. Stimulatory
signals such as TNF-a and phorbol esters induce
degradation of I-kB, and NF-kB consequently activated
enters the nucleus to bind to the kB site of DNA. Inhi-
bitors of this process of NF-kB activation are likely to
become new anti-in¯ammatory and anti-rheumatoid
agents.1 Recently, panepoxydone2 and cycloepoxydon3
were reported to inhibit NF-kB activation, both of
which have the 4-hydroxy-5,6-epoxycyclohexenone
structure. We previously isolated four novel 5,6-epox-
ycyclohexenone compounds named epoxyquinomicins
from Amycolatopsis sp. MK299-95F4 as antibiotics and
anti-in¯ammatory agents.4,5 Epoxyquinomicin C was
the simplest among them, having a 4-hydroxy-5,6-
epoxycyclohexenone structure, however, it did not inhi-
bit activation of NF-kB. Epoxyquinomicin C has an
Synthesis
The synthetic route for DHM2EQ and DHM3EQ is
outlined in Scheme 1. We employed the Wipf method6
with modi®cations for the preparation of the quinone
monoketal structure. Commercially available 2,5-dime-
thoxyaniline, 3, and acetylsalicyloyl chloride were cou-
pled in pyridine to give salicylamide 4, which was
subsequently oxidized by iodobenzenediacetate in
methanol to yield quinone monoketal 5, in 50% yield.
Epoxidation of 5 with alkaline hydrogen peroxide in
aqueous THF gave epoxide 6 in 53% yield con-
comitant with deprotection of the phenolic acetyl
group. Although the condition was strongly basic, no
*Corresponding author. Tel.: +81-45-563-1141; fax: +81-45-562-
7625; e-mail: umezawa@applc.keio.ac.jp
0960-894X/00/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(00)00114-1

866
N. Matsumoto et al. / Bioorg. Med. Chem. Lett. 10 (2000) 865±869
Scheme 1. Reagents and conditions: (a) acetylsalicyloyl chloride, pyridine, 0 ꢀC, 30 min, quant.; (b) PhI(OAc)2, MeOH, 0 ꢀC, 1 h, 50%; (c) 30%
H2O2, 1N NaOH, 0 ꢀC, 2 h, 53%; (d) NaBH4, 0 ꢀC, 10 min; (e) p-TsOH, RT, 1 h; (f) p-TsOH, 60 ꢀC, 1.5 h, 81%; (g) NaBH(OAc)3, RT, 86%, 10 min;
(h) TBDMS-Cl, N,N-diisopropylethylamine, 62%; (i) optical resolution; (j) HF, RT, 2 hr, 75% for ( )-1, 79% for (+)-1.
signi®cant N-desalicylation was observed in 5, unlike in
the case for the acetamide reported by Wipf et al.6
Reduction of epoxide 6 with NaBH4 gave 7 in 64%
yield and its epimer 8 in 10% yield after isolation.
Deprotection of the major isomer 7 with p-TsOH at
0 ꢀC smoothly proceeded and a€orded DHM3EQ (2) in
74% yield. Similarly, minor product 8 was deprotected
to 4-epiDHM3EQ (12). Dimethylketal 6 was treated
with p-toluenesulfonic acid at 70 ꢀC to give epoxy-
quinone 9 in 81% yield. Reduction of 9 with NaBH
(OAc)3 regioselectively occurred, and DHM2EQ (1) was
obtained in 86% yield. The relative stereochemistry of
DHM2EQ was determined by X-ray crystallographic
analysis. Although the crystals of 1 were too ®ne, after
monosilylation of 1, satisfactory crystals of 11 for X-ray
analysis were grown from the ethanol solution.
at a temperature of 20‹1 ꢀC using the o-2y scan tech-
nique to a maximum 2y value of 130.1ꢀ. Of the 3527
re¯ections collected, 3256 were unique. The data were
corrected for Lorentz and polarization e€ects. The
structure was solved by direct methods (SIR92)7 and
expanded using Fourier techniques.8 The non-hydrogen
atoms were re®ned anisotropically, whereas hydrogen
atoms were not. The ®nal cycle of full-matrix least-
squares re®nement was based on 2386 observed re¯ec-
tions (I>2.00s(I)) and 235 variable parameters and
converged with unweighted and weighted agreement
factors of R=0.040 and Rw=0.056. The maximum and
minimum peaks on the ®nal di€erence Fourier map
corresponded to 0.21 and 0.18e /A3, respectively.
All calculations were performed using the teXsan
crystallographic software package of Molecular Struc-
ture Corporation. From these results, orientation
between the 4-hydroxyl and the epoxide in 11 was cis as
displayed in Figure 1.
A colorless prism crystal of 11 having approximate
dimensions of 0.06Â0.10Â0.15 mm was mounted on a
glass ®ber. All measurements were made on a Rigaku
AFC7R di€ractometer with graphite monochromated
Cu-Ka radiation. The following crystal data were
obtained: empirical formula, C19H25NO5Si; formula
weight, 375.50; crystal system, triclinic; space group,
P1(#2); lattice parameters, a=9.9408(6)A, b=13.604(1)A,
c=7.6298(6)A, a=98.075(7)ꢀ, b=90.118(6)ꢀ, g=69.577
(5)ꢀ, V=956.2(1)A3; Z value, 2; Dcalc, 1.304 g/cm3;
m(CuKa), 13.38 cm 1. The re¯ection data were collected
The NMR spectroscopic studies failed to determine the
relative stereochemistry of 11 or that of 1 because of the
lack of de®nitive NOE between H-4 and H-6. We did
not determine the stereochemistry of 2 because we
found a cytotoxicity as described bellow. Next we
focused on enantiomeric separation of 1. The optical
resolution of 11 was successful by passage of 11 through
a chiral stationary phase column, Daicel Chiralpak AS,

N. Matsumoto et al. / Bioorg. Med. Chem. Lett. 10 (2000) 865±869
867
with methanol as an eluent. Figure 2 shows the HPLC
pro®le of 11. Twenty mg of racemic 11 was subjected to
the chiral column in 20 portions to give 9.5 and 9.7 mg
of the corresponding (+)- and ( )-enantiomers,
respectively. The former showed optical rotation of [a]d23
+198ꢀ(c 0.8 methanol); and the latter, optical rotation
of [a]2d4 201ꢀ(c 0.8 methanol). Enantiomeric purities of
both separated enantiomers were >98% based on
HPLC analyses. After desilylation of both enantiomers
with HF in AcCN, (+)-DHM2EQ and ( )-DHM2EQ
were obtained with optical rotation of [a]2d4 +224ꢀ(c 0.1
methanol) and [a]2d4 242ꢀ(c 0.1 methanol), respectively.
tern of (+)-DHM2EQ was also close to that of epoxy-
don reported previously.9,10 These indicate that the CD
spectrum pattern of (+)-DHM2EQ is closely related to
that of 1-epiepoxyquinomicin C. Therefore, the stereo-
chemistry of C-4, C-5 and C-6 in (+)-DHM2EQ was
assigned as R, R, and R, respectively.
Biological activity
Human T cell leukemia Jurkat cells were used for the
assay of NF-kB activity. The cells were transfected with
2mg of DNA by the DEAE-Dextran method. The trans-
fected cells were seeded into 12-well plates at 1Â106/well.
Chemicals dissolved in DMSO and TNF-a were added
at 14 and 16 h, respectively. Six hours after this TNF-a
addition, the cells were harvested and lysed; and the
lysate was used for the luciferase assay with luciferin
and ATP (Promega Luciferase Assay Systems). Lumi-
nescence was measured with a Lumat 9501 (Berthold).
Each value was corrected by the transfection eciency
obtained from the b-galactosidase assay.
Absolute stereochemistry was determined by comparing
the circular dichroism of (+)-DHM2EQ with that of 1-
epiepoxyquinomicin C, which was obtained by reduc-
tion of natural epoxyquinomicin B with NaBH(OAc)3.
(+)-DHM2EQ possessed one peak at 340 nm
(ÁE=+9.54) with a positive Cotton e€ect and one peak
at 295 nm (ÁE= 6.76) with a negative one (c 0.01
dioxane). While 1-epiepoxyquinomicin C showed one
peak at 347 nm (ÁE=+12.8) with a positive Cotton
e€ect and one peak at 296 nm (ÁE= 10.3) with a
negative one (c 0.005 dioxane). The CD spectrum pat-
TNF-a induces degradation of I-kB and activation of
NF-kB in human T cell leukemia Jurkat cells. The
activity of NF-kB can be detected by transient trans-
fection of the reporter DNA having the binding
sequence for NF-kB and the luciferase gene. Figure 3
shows inhibition of TNF-a-induced NF-kB activation
by the racemic DHM2EQ, DHM3EQ and the
DHM3EQ epimer. DHM2EQ was more potent and less
toxic than DHM3EQ. 4-epiDHM3EQ showed inhibi-
tion of NF-kB activation, but it was even more toxic
than DHM3EQ. After enantiomeric separations, ( )-
DHM2EQ was slightly more e€ective than (+)-
DHM2EQ, as shown in Figure 4. Thus, DHM2EQ's
were found to be unique inhibitors of NF-kB activation.
Figure 1. ORTEP drawing of 11.
Figure 3. E€ect of racemic DHM2EQ and its isomers on TNF-a-
induced activation of NF-kB in Jurkat cells. The cells were treated
with DHM2EQ 1 (*), DHM3EQ 2 (&) or 4-epiDHM3EQ 10 (&)
with or without 20 ng/mL TNF-a. DHM3EQ and 4-epiDHM3EQ
were toxic at 10 mg/ml. The values indicate the fold of luciferase
activity induced by TNF-a, and are means ‹SD of triplicate deter-
minations.
Figure 2. HPLC pro®le of 11. Column: Daicel Chiralpak AS, column
size: 10 mm I.D.Â250 mm; eluent: methanol; ¯ow rate: 1.0 mL/min;
detection: 260 mm,.

868
N. Matsumoto et al. / Bioorg. Med. Chem. Lett. 10 (2000) 865±869
Figure 4. Inhibition of TNF-a-induced activation of NF-kB by ( )- and (+)-DHM2EQ in Jurkat cells. The cells were treated with the chemical
with (white column) or without (dark column) 20 ng/ml TNF-a. The values are means ‹SD of triplicate determinations. The ®gures indicate the
fold increase in luciferase activity.
Figure 5. Inhibition of type-II collagen-induced rheumatoid arthritis in mice. Mice were immunized by intravenous administration of type-II
collagen on Day 0. The collagen was intraperitoneally injected on Day 21. DHM2EQ or DHM3EQ was administered intraperitoneally at 0(*),
2(&), or 4(^) mg/kg chemicals three times a week. The grade of rheumatoid arthritis was scored as described in ref 8.
We further examined the antiarthritic e€ects of DHM2EQ
and DHM3EQ on type-II collagen-induced arthritis in
DBA1/J mice. This animal model is widely used for
evaluation of antirheumatic drugs because of its patholo-
gical similarities to human rheumatoid arthritis. The
arthritis was elicited and scored as described previously.5
As shown in Figure 5, DHM2EQ markedly inhibited
type-II-collagen-induced arthritis in mice, whereas
DHM3EQ tended to only slightly inhibit it. The order for
the potency of the antiarthritic e€ect was parallel to that
for the inhibitory e€ect on NF-kB activation, suggesting
that the observed antiarthritic action by DHM2EQ
may be, at least in part, due to the inhibition of NF-kB
activation.
Acknowledgements
The authors wish to thank Ms. K. Kameo and Ms. M.
Endoh, Mercian Corporation, for measuring the phy-
sico-chemical data. This work was ®nancially supported
in part by the Special Coordination Funds for Promo-
tion of Science and Technology from the Science and
Technology Agency, and by a grant from the Ministry
of Education, Science, Culture, and Sports of Japan
(Academic Frontier Promotion Project).
References and Notes
1. Baeuerle, P. A.; Baichwal, V. R. Adv. Immunol. 1997, 65, 111.

N. Matsumoto et al. / Bioorg. Med. Chem. Lett. 10 (2000) 865±869
869
2. Erkel, G.; Anke, T.; Sterner, O. Biochem. Biophys. Res.
Commun. 1996, 226, 214.
Giacovazzo, C.; Guagliardi, A.; Polidori, G. J. Appl. Crystal-
logr. 1994, 27, 435.
3. Gehrt, A.; Erkel, G.; Anke, Timm; Sterner, O. J. Antibiotics
1998, 51, 455.
4. Matsumoto, N.; Tsuchida, T.; Umekita, M.; Kinoshita, N.;
Iinuma, H.; Sawa, T.; Hamada, M.; Takeuchi, T. J. Anti-
biotics 1997, 50, 900.
8. Beurskens, P. T.; Admiraal, G.; Beurskens, G.; Bosman,
W. P.; de Gelder, R.; Israel, R.; Smits, J. M. M. The DIRDIF-
94 program system, Technical Report of the Crystallography
Laboratory, University of Nijmegen, The Netherlands,
1994.
5. Matsumoto, N.; Iinuma, H.; Sawa, T.; Takeuchi, T.; Hirano,
S.; Yoshioka, T.; Ishizuka, M. J. Antibiotics 1997, 50, 906.
6. Wipf, P.; Kim, Y.; Jahn, H. Synthesis 1995, 1549.
7. Altomare, A.; Burla, M. C.; Camalli, M.; Cascarano, M.;
9. Crosse, A.; Mauli, R.; Sigg, H. P. Helv. Chim. Acta 1966,
49, 204.
10. Sakamura, S.; Niki, H.; Obata, Y.; Sakai, R.; Matsumoto,
T. Agr. Biol. Chem., 1969, 698