A 3,4-Epoxypiperidine structure as a novel and simple DNA-cleavage unit

Article abstract of DOI:10.1016/S0960-894X(02)00085-9

Based on the 4-hydroxy-1-azabicyclo[3.1.0]hexane structure of azinomycin, a 3,4-epoxypiperidine structure was designed as a novel and simple alkylating molecular unit, and some 3,4-epoxypiperidine derivatives were found to show DNA-cleavage activity, the structural requirements for which were revealed.

Full text of DOI:10.1016/S0960-894X(02)00085-9

Bioorganic & Medicinal Chemistry Letters 12 (2002) 1075–1077
A 3,4-Epoxypiperidine Structure as a Novel and Simple
DNA-Cleavage Unit
Kazuyuki Miyashita, Myunji Park, Sayako Adachi, Sayori Seki, Satoshi Obika and
Takeshi Imanishi*
Graduate School of Pharmaceutical Sciences Osaka University, 1-6 Yamadaoka, Suita, Osaka 565–0871, Japan
Received 26December 2001; accepted 26January 2002
Abstract—Based on the 4-hydroxy-1-azabicyclo[3.1.0]hexane structure of azinomycin, a 3,4-epoxypiperidine structure was designed
as a novel and simple alkylating molecular unit, and some 3,4-epoxypiperidine derivatives were found to show DNA-cleavage
activity, the structural requirements for which were revealed. # 2002 Elsevier Science Ltd. All rights reserved.
An alkylating agent is an attractive group of biologi-
cally active compounds such as an anticancer agent or
as an enzyme inhibitor. Although a number of mol-
ecules causing alkylation of nucleophiles in a biological
system have been evaluated, it is still required to create
a molecular unit having a more potent alkylating
ability.1
As some examples of ring contraction from a piperidine
ring system having a leaving group at C-3 to a 1-azabi-
cyclo[3.1.0]hexane ring system via an intramolecular
substitution reaction have been reported,7 it occurred to
us that the neighboring hydroxyl group of the aziridine
ring of 1 might promote the ring opening to transform
into an epoxypiperidine as shown in Figure 2. We
assumed, if such promotion occurs, the 3,4-epoxypiper-
idine could also work as an alkylating agent. Based on
this hypothesis, we synthesized 3,4-epoxypiperidine
derivatives 2 and studied their DNA-cleavage activity.
In order to design a novel molecule having DNA-alky-
lating ability, we focused on azinomycins (Fig. 1) which
are an antitumor antibiotic isolated from Streptomyces
sp.2 and are known to work as a DNA cross-linking
agent. As is easily expected from the structure, 1-azabi-
cyclo[3.1.0]hexane and oxirane ring units are respon-
sible for the activity and were found to alkylate a
guanosine and/or an adenosine residue of a genomic
DNA duplex.3 Because of its important biological
activity and its characteristic structure, a number of
synthetic studies on azinomycin and its analogues have
been conducted,4 and the first total synthesis of azino-
mycin has been achieved recently by Coleman’s group.5
Through these studies, it has become known that the 4-
hydroxy-1-azabicyclo[3.1.0]hexane ring system is highly
unstable: the hydroxyl protected compounds are stable
enough to isolate, but deprotection of the hydroxyl
group greatly lowers the stability.6 However, no infor-
mation regarding the decomposition mechanism and the
decomposition product has ever been reported.
The epoxides cis- and trans-2 were synthesized as shown
in Scheme 1. Unsaturated alcohol 38 was epoxidized
with m-CPBA to give cis-epoxide exclusively, sub-
Figure 1. Structures of azinomycins A and B.
*Corresponding author. Tel.: +81-6-6879-8200; fax: +81-6-6879-
8204; e-mail: imanishi@phs.osaka-u.ac.jp
Figure 2. Design concept for 3,4-epoxypiperidine.
0960-894X/02/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(02)00085-9

1076
K. Miyashita et al. / Bioorg. Med. Chem. Lett. 12 (2002) 1075–1077
Figure 3. DNA cleavage activity for trans-2, cis-2, 6, (S)-trans-2, and
(R)-trans-2.11 Form I: supercoiled pBR 322 DNA; Form II: relaxed
DNA, lanes 1 and 2: trans-2; lanes 3 and 4: cis-2; lanes 5 and 6: 6; lanes
7 and 8: (S)-trans-2; lanes 9 and 10: (R)-trans-2; lane 11: control.
Concentrations of the compounds are 100 and 10 mM, respectively.
Figure 4. Structures of racemic trans-derivatives 8–12.
Scheme 1. Reagents and conditions: (i) m-CPBA, CH2Cl2 (88%); (ii)
BnBr, NaH, THF (91%); (iii) MeLi, THF then H2O (quant); (iv)
BnBr, NaH, THF (95%); (v) m-CPBA, CH2Cl2 (trans/cis=ca. 4:1,
trans-4, 78%); (vi) MeLi, THF then H2O (quant); (vii) MeLi, THF
then H2O (quant); (viii) O-tert-butyldimethylsilyl-(S)-mandelic acid,
sequent benzylation of which afforded cis-4. On the other
hand, 3 was benzylated, then oxidized with m-CPBA to
give an isomeric epoxide trans-4 as a main product. The
ethoxycarbonyl group of cis- and trans-4 was removed
by treatment with methyllithium to afford cis- and trans-
2,9 respectively, in almost quantitative yield.10 In order
to investigate the role of the epoxide structure, olefinic
compound 6 was also prepared similarly.
.
EDC HCl, DMAP, THF (81%); (ix) silica gel chromatography; (x)
0.2 M NaOH–EtOH (90%).
hexane derivative 8 did not show the activity at all
(lanes 3–6). In contrast, the activity of N-methyl deri-
vative 9 was lower than that of trans-2, but remained
(lanes 7 and 8). N,N-Dimethylammonium derivative 10
also showed faint activity (lanes 9 and 10), suggesting
that, on DNA-cleavage, the amino group could be
transformed to the corresponding ammonium form to
interact with a phosphate linkage. Taking account of
the fact that the ammonium derivative 10 has the activ-
ity, it is obvious that the alkylation of the DNA takes
place via a nucleophilic cleavage of the epoxide form,14
but the possibility still remains that the alkylation
involves two pathways, via an epoxide form and an
aziridine form, as the activity of 10 was much weaker
than that of trans-2.
At first, the activity of cis-2, trans-2 and 6 with respect
to DNA was examined by a relaxation assay of super-
coiled plasmid DNA.11 As shown in Figure 3, relaxation
was observed with trans-2 at the concentration of 100
mM (lane 1), while weak activity was observed on treat-
ment with cis-2 at the same concentration (lane 3). In
contrast, the olefinic compound 6 was inactive (lanes 5
and 6). These results clearly show that the 3,4-epoxy-
piperidine structure has DNA-cleavage activity as
expected, and also indicate the important role of the
epoxide structure and the stereochemistry.
The optically active compounds, (S)-trans-2 and (R)-
trans-2, were also obtained by optical resolution of the
corresponding mandelate ester 7 as shown in Scheme 1.
The resulting (R)- and (S)-312 were transformed to (S)-
and (R)-trans-2, respectively. The (S)-isomer was found
to be more active than the (R)-isomer (lanes 7 and 8
versus lanes 9 and 10) (Fig. 4).
Compound 11 in which the benzyl group of trans-2 was
replaced with a cyclohexylmethyl group completely lost
the activity (lanes 11 and 12), while 12 having a 2-
naphthylmethyl group showed greatly increased activity
(lanes 13 and 14). These results strongly suggest that the
aromatic ring works as an intercalator.
In conclusion, although details of the mechanism for the
DNA cleavage of the 3,4-epoxypiperidines are still
unclear, the present results showed that 3,4-epoxypiper-
idines can work as a novel and potent DNA-cleavage
unit, and also revealed the structural requirements for
the activity. As is obvious from the results, the free
amino group and the aromatic group play an important
role in the DNA-cleavage. As the urethane derivative
trans-4 was inactive, a trigger function, activating the
Focusing on the other two functional groups, the amino
group and the benzyloxy group, in the molecule of
trans-2, we further examined the DNA-cleavage activity
of other racemic trans-derivatives, trans-4 and 8–12,13 to
investigate the structure–activity relationship (Fig. 5).
The amino group was found to be essentially important,
because ethoxycarbonyl derivative trans-4 and cyclo-
Figure 5. DNA cleavage activity for trans-2, 4 and 8–12.11 Form I: supercoiled pBR 322 DNA; Form II: relaxed DNA, lanes 1 and 2: trans-2; lanes
3 and 4: 4; lanes 5 and 6: 8; lanes 7 and 8: 9; lanes 9 and 10: 10; lanes 11 and 12: 11; lanes 13 and 14: 12; lane 15: control. Concentrations of the
compounds are 100 and 10 mM, respectively.

K. Miyashita et al. / Bioorg. Med. Chem. Lett. 12 (2002) 1075–1077
1077
molecule under particular conditions, would be easily
introduced by choosing a proper N-protecting group.
Although the DNA cleavage activity of the present
compounds are not high compared with that of clini-
cally employed DNA cleavage agents, it is also antici-
pated to increase the activity by combination with a
proper target recognition site because of the simple
structure. Further studies on the mechanism and on its
application are in progress in our laboratory.
Hz, JAB=15 Hz, C2-H), 3.25 (1H, dd, J=3, 4 Hz, C3-H), 3.41
(1H, t-lile, J=4 Hz, C4-H), 3.73 (1H, ddd, J=4, 5, 5 Hz, C5-
H), 4.70 (2H, AB, J=12 Hz, PhCH2O–), 7.35–7.40 (5H, m,
aromatic H). trans-2 (CDCl3) d: 2.47 (1H, dd, J=5, 14 Hz,
C6-H), 3.01 (1H, dd, J=7, 14 Hz, C6-H), 3.10, 3.19 (2H, AB
in ABX, JAX=2, JBX=ca. 0 Hz, JAB=16Hz, C2-H), 3.13
(1H, d, J=2 Hz, C3-H), 3.27 (1H, d, J=5 Hz, C4-H), 3.62
(1H, dd, J=5, 7 Hz, C5-H), 4.66 (2H, AB, J=12 Hz,
PhCH2O–), 7.31–7.36(5H, m, aromatic H).
References and Notes
1. For example, Kauffman, J. E.; Foye, W. O. In Principles of
Medicinal Chemistry, 2nd ed.; Foye, W. O., Ed.; Lea & Febi-
ger: Philadelphia, 1981: p 837. Rajski, S. R.; Williams, R. M.
Chem. Rev. 1998, 98, 2723.
2. (a) Yokoi, K.; Nagaoka, K.; Nakashima, T. Chem. Pharm.
Bull. 1986, 34, 4554. (b) Armstrong, R. W.; Moran, E. J. Tet-
rahedron Lett. 1991, 57, 2208.
Figure 6. NOE correlation observed in 13 and trans-2.
10. The epoxides cis- and trans-2 were stable in a solution of
usual organic solvents, but sometimes underwent decomposi-
tion after concentration, suggesting that polymerization took
place. Therefore, cis- and trans-2 were prepared from cis- and
trans-4 before use and were employed immediately.
11. To a solution of supercoiled pBR322 DNA (17 mg) in pH
7.4 TE buffer (9 mL) was added a DMSO solution of the
compounds (1 mL, 1 mM and 100 mM), and the whole was
incubated for 24 h at 37 ꢀC to complete strand cleavage.16 The
resulting DNAs were analyzed by electrophoresis on 0.7%
native agarose gel at 7.4 v/cm for 20 min. Piperidine-treatment
for 30 min at 90 ꢀC in place of the incubation for 24 h at 37 ꢀC
was also examined to complete strand cleavage, but showed
the same result.
3. Armstrong, R. W.; Salvati, M. E.; Nguyen, M. J. Am.
Chem. Soc. 1992, 114, 3144. Alcaro, S.; Coleman, R. S. J. Org.
Chem. 1998, 63, 4620. Fujiwara, T.; Sato, I. Tetrahedron Lett.
1999, 40, 315. Hartley, J. A.; Hazrati, A.; Kelland, L. R.;
Khanim, R.; Shipman, M.; Suzenet, F.; Walker, L. F. Angew.
Chem. Int. Ed. 2000, 39, 3467. Alcaro, S.; Coleman, R. S. J.
Med. Chem. 2000, 43, 2783.
4. For review: Hodgkinson, T. J.; Shipman, M. Tetrahedron
2001, 57, 4467.
5. Coleman, R. S.; Li, J.; Navarro, A. Angew. Chem. Int. Ed.
2001, 40, 1736.
6. Coleman, R. S.; Kong, J.-S. J. Am. Chem. Soc. 1998, 120,
3538. Coleman, R. S.; Richardson, T. E.; Carpenter, A. J. J.
Org. Chem. 1998, 63, 5738. Hashimoto, M.; Terashima, S.
Heterocycles 1998, 47, 59.
12. The absolute configurations of 3 were determined as
shown by Mosher’s method.
7. Banks, R. E.; Besheesh, M. K.; Lawrence, N. J.; Pritchard,
R. G.; Tovell, D. J. Chem. Commun. 1999, 47. Tehrani, K. A.;
Syngel, K. V.; Boelens, M. J.; Contreras, J.; Kimpe, N. D.;
Knight, D. W. Tetrahedron Lett. 2000, 41, 2507.
8. Imanishi, T.; Shin, H.; Hanaoka, M.; Momose, T.; Ima-
nishi, I. Chem. Pharm. Bull. 1982, 30, 3617.
13. Cyclohexane derivative 8 was prepared from 2-cyclo-
hexen-1-ol similarly to trans-2. N-Methyl derivatives, 9 and 10,
were prepared by methylation of trans-2. Alkoxy derivatives,
11 and 12, were prepared by the same method as that for
trans-2.
14. Some typical exmples of epoxide derivatives which alky-
late DNA: Benasutti, M.; Ejadi, S.; Whitlow, M. D.; Loechler,
E. L. Biochemistry 1988, 27, 472. Hara, M.; Yoshida, M.;
Nakano, H. Biochemistry 1990, 29, 10449. Chan, K. L.;
Sugiyama, H.; Saito, I.; Hara, M. Tetrahedron Lett. 1991, 32,
7719.
9. It has been reported that, in the 1-azabicyclo[3.1.0]hexane
structure 13, NOE correlation is observed between the hydro-
gens shown below.2a,15 1H NMR study of trans-2 taken in
CDCl3 revealed that the hydrogen at C-5 has NOE correlation
only with the hydrogens at C-4 and C-6, showing that trans-2
1
has an epoxide structure at least in CDCl3 (Fig. 6). H NMR
15. Armstrong, R. W.; Moran, E. J. J. Am. Chem. Soc. 1992,
114, 371.
16. Nakatani, K.; Okamoto, A.; Saito, I. Angew. Chem., Int.
Ed. Engl. 1997, 36, 2794.
data for cis-2 and trans-2 are as follows. cis-2 (CDCl3) d: 2.66,
2.84 (2H, AB in ABX, JAX=5 Hz, JBX=5 Hz, JAB=14 Hz,
C6-H), 3.01, 3.18 (2H, AB in ABX, JAX=3 Hz, JBX=ca. 0