Delineating noncovalent interactions between the azinomycins and double-stranded DNA: Importance of the naphthalene substitution pattern on interstrand cross-linking efficiency

Article abstract of DOI:10.1021/ol048653y

(Chemical Equation Presented) Using a series of synthetic azinomycin analogues, it is shown that the efficiency of in vitro DNA interstrand cross-linking is markedly reduced when either the C-5′ methyl group or both the C-5′ methyl and C-3′ methoxy groups

Full text of DOI:10.1021/ol048653y

ORGANIC
LETTERS
2004
Vol. 6, No. 20
3505-3507
Delineating Noncovalent Interactions
between the Azinomycins and
Double-Stranded DNA: Importance of
the Naphthalene Substitution Pattern on
Interstrand Cross-Linking Efficiency
Cyrille A. S. Landreau,† Rachel C. LePla,† Michael Shipman,*,†
Alexandra M. Z. Slawin,‡ and John A. Hartley§
Department of Chemistry, UniVersity of Warwick, Gibbet Hill Road,
CoVentry CV4 7AL, U.K., School of Chemistry, UniVersity of St. Andrews,
Purdie Building, St. Andrews, Fife KY16 9ST, U.K., and CR-UK Drug-DNA
Interactions Research Group, Department of Oncology, Royal Free & UniVersity
College Medical School, UniVersity College London, 91 Riding House Street,
London W1W 7BS, U.K.
m.shipman@warwick.ac.uk
Received July 14, 2004
ABSTRACT
Using a series of synthetic azinomycin analogues, it is shown that the efficiency of in vitro DNA interstrand cross-linking is markedly reduced
when either the C-5
′
methyl group or both the C-5
′
methyl and C-3
′
methoxy groups are deleted from the naphthalene ring.
The antitumor antibiotics azinomycin A (1) and B (2) are
structurally unique natural products containing the strained
1-azabicyclo[3.1.0]hexane ring system (Figure 1).1 These
compounds possess potent in vitro cytotoxic activity and
significant in vivo antitumor activity and appear to act by
disruption of cellular DNA replication by interstrand cross-
link (ISC) formation.1 In 2001, the first total synthesis of
azinomycin A was achieved by Coleman.2 The epoxide and
aziridine are known to be responsible for the cross-linking
process which occurs via N-7 of purine bases two residues
apart on the complementary DNA strands.3 At present, there
is essentially no data available concerning the role of other
functionalities contained within the azinomycin skeleton on
the selectivity of DNA recognition and efficiency of inter-
strand cross-linking. Such information would provide im-
portant insights into the mode of action of the azinomycins
and assist in the design of new DNA targeting agents.
(3) (a) Lown, J. W.; Majumdar, K. C. Can. J. Biochem. 1977, 55, 630.
(b) Armstrong, R. W.; Salvati, M. E.; Nguyen, M. J. Am. Chem. Soc. 1992,
114, 3144. (c) Fujiwara, T.; Saito, I.; Sugiyama, H. Tetrahedron Lett. 1999,
40, 315. (d) Hartley, J. A.; Hazrati, A.; Kelland, L. R.; Khanim, R.; Shipman,
M.; Suzenet, F.; Walker, L. F. Angew. Chem., Int. Ed. 2000, 39, 3467. (e)
Coleman, R. S.; Perez, R. J.; Burk, C. H.; Navarro, A. J. Am. Chem. Soc.
2002, 124, 13008.
† University of Warwick.
‡ University of St Andrews.
§ University College, London.
(1) Hodgkinson, T. J.; Shipman, M. Tetrahedron 2001, 57, 4467.
(2) Coleman, R. S.; Li, J.; Navarro, A. Angew. Chem., Int. Ed. 2001,
40, 1736.
10.1021/ol048653y CCC: $27.50
© 2004 American Chemical Society
Published on Web 09/03/2004

alcohol 48 with the appropriate acid chlorides9 furnished
esters 5b-d (Scheme 1). Removal of the benzyl ester by
catalytic hydrogenation yielded 6b-d, which were im-
mediately coupled with dehydroamino ester 8, formed by
removal of the N-Boc group from 7,3d to give 9b-d in good
yields. Finally, ring closure to 1-azabicyclo[3.1.0]hexanes
3b-d was accomplished according to methodology originally
devised by Terashima via the corresponding mesylate.10
These ring closures were best performed using TBAF. The
low yields for the ring closure step largely reflect problems
associated with obtaining 3b-d in an acceptable state of
purity.11
The stereochemical assignment about the tetrasubstituted
double bond of 3a was originally based on NOE measure-
ments and correlations with the literature.3d In undertaking
this study, we were fortunate that alcohol 9c devoid of the
methoxy group at C-3′ was highly crystalline, and using a
single crystal grown from EtOAc/n-hexane, we were able
to solve its structure unambiguously using X-ray crystal-
lography (Figure 2).12 As this is the first solid-state structure
Figure 1. Structures of the naturally occurring DNA interstrand
cross-linking azinomycins and functional synthetic analogue 3a.
In systems based on just the “left-hand” epoxide domain,
the naphthoate group plays a critical role in the efficiency
of DNA alkylation,4,5 DNA sequence selectivity,4,5 and
cytotoxicity.5,6 By extrapolation, one can postulate that the
naphthalene ring and its 5-methyl and 3-methoxy substituents
play pivotal roles in noncovalent association between the
azinomycins and double-stranded DNA. However, testing
this hypothesis is difficult because modification of the
aromatic ring within the natural products is precluded because
of their inherent chemical instability. Recently, we described
the synthesis of azinomycin analogue 3a, which acts on
double-stranded DNA in a manner similar to that of the
natural products (Figure 1).3d We reasoned that by studying
a series of derivatives related to 3a in which the naphthoate
fragment was modified, we would be able to probe if the
substituents on the naphthalene ring play a critical role in
DNA binding and ISC efficiency or indeed if they are simply
unnecessary artifacts of the biosynthesis.7 Herein, we report
the preparation of modified azinomycin analogues 3b-d and
demonstrate that the nature of the naphthalene substitution
pattern does indeed influence in vitro DNA ISC efficiency.
The synthesis of epoxy aziridines 3b-d parallels our
earlier synthesis of 3a. Thus, esterification of (2S,3S)-epoxy
Figure 2. X-ray crystal structure of 9c.12
of an “azinomycin-like” material, several features merit
comment. First, the distance between the cross-linking carbon
atoms (C-10 and C-21) is determined to be 8.70 Å in 9c,
Scheme 1
3506
Org. Lett., Vol. 6, No. 20, 2004

which is close to the estimated linear distance of 8.5 Å
between the N-7’s of the two reacting guanines in B-form
DNA.13 Second, the lone pairs of N-9 and N-16 are nearly
orthogonal (CO-N-CdC torsional angle ) 84°), presum-
ably to minimize electron repulsion. Finally, it is apparent
that there are differences between the solid-state conforma-
tion of 9c and that of azinomycin B (2) derived by
mechanism filtered molecular modeling.13 These potentially
important conformational differences warrant scrutiny in the
future.
The DNA interstrand cross-linking activities of epoxy
aziridines 3b-d were compared with resynthesized 3a using
an agarose gel assay using the pUC18 linearized plasmid.14
The 32P 5′-end-labeled duplex was incubated with 3a-d
(1.0-200 µM) for 1 h at 37 °C prior to denaturing with alkali
and subsequent gel electrophoresis. The extent of cross-
linking was determined by quantifying the relative amounts
of double-stranded and single-stranded DNA by storage
phosphorimage analysis. The experiment was performed four
times, and a representative gel is presented (Figure 3).
of ISC efficiency follows the trend 3a ≈ 3c > 3b > 3d
with ca. 18-fold difference in efficiency across the group.
At 10 µM, 3a induces more cross-links than 3c, indicating
that it is a little more potent. Because 3c and 3a possess the
5′-methyl group and 3b and 3d do not, we conclude that
this substituent plays an important role in noncovalent
association between 3 and the DNA duplex. Interestingly,
the 3′-methoxy group, which is proposed to be introduced
relatively late in the biosynthesis,7 has less influence upon
the efficiency of ISC formation (3a cf. 3c), although the
deletion of both substituents does have an appreciable effect
(3a cf. 3d).
How the 5′-methyl promotes binding remains unclear,
although we believe that it is most probably enhancing
hydrophobic interactions between the DNA duplex and the
epoxy aziridines. Currently, we do not know whether this
arises from specific contacts between the 5′-methyl and the
DNA backbone or a more general enhancement of the
hydrophobicity of the drug. Mechanisms based on more
active involvement of the naphthalene ring such as intercala-
tion are most likely not operating, as extensive efforts by
Coleman to detect intercalative binding using azinomycin
B have not yet been successful.3e
To conclude, we have established that the 5′-methyl group
is an important element in noncovalent association between
DNA and azinomycin-like compounds. Future efforts will
be focused on using this strategy to identify other important
noncovalent binding interactions in this class of agent.
Figure 3. Agarose cross-linking gel for epoxy aziridines 3a-d.
Plasmid DNA was treated with the agents at the concentrations
shown for 1 h prior to alkali denaturation and gel electrophoresis.
Cn and Cd are control nondenaturated and denatured samples,
respectively. DS and SS indicate the positions of double and single
stranded DNA, respectively. At .100% cross-linking, the intensity
of the DS band reduces as other DNA products begin to appear
(not shown).
Acknowledgment. We are indebted to BBSRC (B15997)
for financial support of this work and to Dr. Kevin Moffat
of the University of Warwick for his help and assistance.
Supporting Information Available: 1H and 13C NMR
spectra for compounds 3b-d and crystallographic data in
CIF format. This material is available free of charge via the
Internet at http://pubs.acs.org.
Although all of the epoxy aziridines produce DNA cross-
links at micromolar concentrations of drug, the efficiency
of cross-linking varied with the naphthalene substitution
pattern. At 50 µM, 3a produced 95.0 ( 1.9% cross-links
(mean ( SD); 3b, 27.1 ( 3.8% cross-links; 3c, 95.1 ( 3.9%
cross-links; and 3d, 5.1 ( 1.6% cross-links. Thus, the order
OL048653Y
(11) Epoxy aziridines 3a-d are unstable and cannot be readily purified
by chromatography. After extensive experimentation, we have devised a
purification protocol that involves, after aqueous workup, selective precipita-
tion of the impurities by addition of n-hexane to a chloroform solution of
the crude product. After removal of the precipitate using a centrifuge, the
mother liquor provides 3a-d in a good state of purity (ca. 80-90%) as
1
judged by H and 13C NMR spectroscopy (see Supporting Information).
(4) Zang, H.; Gates, K. S. Biochemistry 2000, 39, 14968.
(5) Coleman, R. S.; Burk, C. H.; Navarro, A.; Brueggemeier, R. W.;
Diaz-Cruz, E. S. Org. Lett. 2002, 4, 3545.
(6) Hodgkinson, T. J.; Kelland, L. R.; Shipman, M.; Suzenet, F. Bioorg.
Med. Chem. Lett. 2000, 10, 239.
(7) Corre, C.; Lowden, P. A. S. Chem. Commun. 2004, 990.
(8) Bryant, H. J.; Dardonville, C. Y.; Hodgkinson, T. J.; Hursthouse,
M. B.; Malik, K. M. A.; Shipman, M. J. Chem. Soc., Perkin Trans. 1 1998,
1249.
(9) 5-Methylnaphthoyl chloride was made from ethyl 3-hydroxy-5-
methylnaphthoate8 in 4 steps [(i) Tf2O, Et3N, CH2Cl2; (ii) cat. Pd(dppf)-
Cl2, Et3SiH, DMF, 60 °C; (iii) LiOH, MeOH, H2O; (iv) (COCl)2, cat. DMF,
CH2Cl2, 4 h, 54% over 4 steps]. 3-Methoxynaphthoic acid was made
according to published methods. See: Horii, Z.; Matsumoto, Y.; Momose,
T. Chem. Pharm. Bull. 1971, 19, 1245. Newman, M. S.; Sankaran, V.; Olsen,
D. R. J. Am. Chem. Soc. 1976, 98, 3237 and references therein. Ester 5d
has been made previously (see ref 6).
(12) Crystallographic data for 9c: X-ray diffraction studies on a
colorless crystal grown from EtOAc/n-hexane were performed at 125 K
using a Bruker SMART diffractometer with graphite-monochromated Mo
KR radiation (λ ) 0.71073 Å). The structure was solved by direct methods.
C26H30N2O7, M ) 482.52, monoclinic, space group P21, a ) 9.010(1), b )
12.978(2), c ) 10.722(1) Å, V ) 1181.1(3) Å3, Z ) 2, Dc ) 1.357 M
gm-3, µ ) 0.099 mm-1, F(000) ) 512, crystal size ) 0.18 × 0.1 × 0.1
mm3. Flack parameter 0.0(9). Of 5101 measured data, 3105 were unique
(Rint ) 0.0155) and 2810 observed (I > 2σ(I)]) to give R1) 0.035 and wR2
) 0.0821. All non-hydrogen atoms were refined with anisotropic displace-
ment parameters; the NH and OH protons were located from ∆F maps and
allowed to refine isotropically subject to a distance constraint (O-H/N-H
) 0.98 Å). All remaining hydrogen atoms bound to carbon were idealized.
Structural refinements were by the full-matrix least-squares method on F2.
(13) (a) Alcaro, S.; Coleman, R. S. J. Org. Chem. 1998, 63, 4620. (b)
Alcaro, S.; Coleman, R. S. J. Med. Chem. 2000, 43, 2783. (c) Alcaro, S.;
Ortuso, F.; Coleman, R. S. J. Med. Chem. 2002, 45, 861.
(10) Hashimoto, M.; Matsumoto, M.; Yamada, K.; Terashima, S.
Tetrahedron Lett. 1994, 35, 2207.
(14) Hartley, J. A.; Berardini, M. D.; Souhami, R. L. Anal. Biochem.
1991, 193, 131.
Org. Lett., Vol. 6, No. 20, 2004
3507