Design, synthesis, and evaluation of a novel sequence-selective epoxide- containing DNA cross-linking agent based on the pyrrolo[2,1- c][1,4]benzodiazepine system

Article abstract of DOI:10.1021/jm981124d

Synthetic routes have been investigated to prepare a novel C8-epoxide- functionalized pyrrolo[2,1-c][1,4]benzodiazepine 6 as a potential sequence- selective DNA cross-linking agent (Wilson et al. Tetrahedron Lett. 1995, 36, 6333-6336). A successful synthesis was accomplished via a 10-step route involving a pro-N10-Fmoc cleavage method that should have general applicability to other pyrrolobenzodiazepine (PBD) molecules containing acid- or nucleophile-sensitive groups. During the course of this work, a one-pot reductive cyclization procedure for the synthesis of PBD N10-C11 imines from nitro dimethyl acetals was also discovered, although this method results in C11a racemization which can reduce DNA binding affinity and cytotoxicity. The target epoxide 6 was shown by thermal denaturation studies to have a significantly higher DNA-binding affinity than the parent DC-81 (3) or the C8-propenoxy-PBD (15), which is structurally similar but lacks the epoxide moiety. The time course of effects upon thermal denaturation indicated a rapid initial binding phase followed by a slower phase consistent with the stepwise cross-linking of DNA observed for a difunctional agent. This was confirmed by an electrophoretic assay which demonstrated efficient induction of interstrand cross-links in plasmid DNA at concentrations > 1 μM. Higher levels of interstrand cross-linking were observed at 24 h compared to 6 h incubation. A Taq polymerase stop assay indicated a preference for binding to guanine-rich sequences as predicted for bis-alkylation in the minor groove of DNA by epoxide and imine moieties. The pattern of stop sites could be partly rationalized by molecular modeling studies which suggested low-energy models to account for the observed binding behavior. The epoxide PBD 6 was shown to have significant cytotoxicity (45-60 nM) in the A2780, CH1, and CH1cis(R) human ovarian carcinoma cell lines and an IC₅₀ of 0.2 μM in A2780cis(R). The significant activity of 6 in the cisplatin-resistant CH1cis(R) cell line (IC₅₀ = 47 nM) gave a resistance factor of 0.8 compared to the parent cell line, demonstrating no cross-resistance with the major groove cross-linking agent cisplatin.

Full text of DOI:10.1021/jm981124d

4028
J . Med. Chem. 1999, 42, 4028-4041
Design , Syn th esis, a n d Eva lu a tion of a Novel Sequ en ce-Selective
Ep oxid e-Con ta in in g DNA Cr oss-Lin k in g Agen t Ba sed on th e
P yr r olo[2,1-c][1,4]ben zod ia zep in e System
Stuart C. Wilson,† Philip W. Howard,† Stephen M. Forrow,‡ J ohn A. Hartley,‡ Lesley J . Adams,†
Terence C. J enkins,§ Lloyd R. Kelland,∇ and David E. Thurston*,†
CRC Gene Targeted Drug Design Research Group, School of Pharmacy and Biomedical Sciences, University of Portsmouth,
St. Michael’s Building, White Swan Road, Portsmouth, Hants. PO1 2DT, U.K., CRC Drug-DNA Interactions Research Group,
Department of Oncology, University College London Medical School, London W1P 8BT, U.K., School of Chemical and Life
Sciences, University of Greenwich, Wellington Street, Woolwich, London SE18 6PF, U.K., and CRC Centre for Cancer
Therapeutics, Institute of Cancer Research, Sutton, Surrey SM2 5NG, U.K.
Received December 8, 1998
Synthetic routes have been investigated to prepare a novel C8-epoxide-functionalized pyrrolo-
[2,1-c][1,4]benzodiazepine 6 as a potential sequence-selective DNA cross-linking agent (Wilson
et al. Tetrahedron Lett. 1995, 36, 6333-6336). A successful synthesis was accomplished via a
10-step route involving a pro-N10-Fmoc cleavage method that should have general applicability
to other pyrrolobenzodiazepine (PBD) molecules containing acid- or nucleophile-sensitive groups.
During the course of this work, a one-pot reductive cyclization procedure for the synthesis of
PBD N10-C11 imines from nitro dimethyl acetals was also discovered, although this method
results in C11a racemization which can reduce DNA binding affinity and cytotoxicity. The
target epoxide 6 was shown by thermal denaturation studies to have a significantly higher
DNA-binding affinity than the parent DC-81 (3) or the C8-propenoxy-PBD (15), which is
structurally similar but lacks the epoxide moiety. The time course of effects upon thermal
denaturation indicated a rapid initial binding phase followed by a slower phase consistent
with the stepwise cross-linking of DNA observed for a difunctional agent. This was confirmed
by an electrophoretic assay which demonstrated efficient induction of interstrand cross-links
in plasmid DNA at concentrations >1 µM. Higher levels of interstrand cross-linking were
observed at 24 h compared to 6 h incubation. A Taq polymerase stop assay indicated a preference
for binding to guanine-rich sequences as predicted for bis-alkylation in the minor groove of
DNA by epoxide and imine moieties. The pattern of stop sites could be partly rationalized by
molecular modeling studies which suggested low-energy models to account for the observed
binding behavior. The epoxide PBD 6 was shown to have significant cytotoxicity (45-60 nM)
in the A2780, CH1, and CH1cisR human ovarian carcinoma cell lines and an IC50 of 0.2 µM in
A2780cisR. The significant activity of 6 in the cisplatin-resistant CH1cisR cell line (IC50 ) 47
nM) gave a resistance factor of 0.8 compared to the parent cell line, demonstrating no cross-
resistance with the major groove cross-linking agent cisplatin.
In tr od u ction
covalent binding of the N10-C11 imine functionality to
the exocyclic C2-amino group of guanine within the
minor groove of duplex DNA.5,6 The (S)-configuration
at the chiral C11a position has been shown to be
essential for binding activity, as it provides the PBD
molecule with the right-handed twist necessary for fit
within the host minor groove.5 Footprinting-type tech-
niques have established that the PBDs span three base
pairs of DNA with a rank order of preference for Pu-
G-Pu > Pu-G-Py ∼ Py-G-Pu > Py-G-Py motifs (Pu )
purine, Py ) pyrimidine).6 The cytotoxicity of these
agents is thought to be due to their ability to inhibit
cellular processes such as replication and transcription.
Using a bacteriophage T7 RNA polymerase model
system, it has been demonstrated that the PBDs can
inhibit in vitro transcription in a highly sequence-
selective manner.7 In addition, a direct correlation has
been observed between the ability of a series of PBDs
to inhibit the action of the endonuclease enzyme BamHI
in vitro and their cytotoxicity across a number of cell
lines.8
There is interest in discovering and developing small
molecules capable of binding to DNA in a highly
sequence-selective manner.2 Such agents, with the
ability to target and then down-regulate or ablate
individual genes, have potential use as drugs for the
therapy of genetic-based diseases including cancer,3 as
diagnostic agents, and as research tools for use in target
validation and functional genomics studies. The pyrrolo-
[2,1-c][1,4]benzodiazepine (PBD) family of antitumor
antibiotics was first discovered in 1965 when Leimgru-
ber and co-workers isolated anthramycin (1) from Strep-
tomyces refuineus4 (Figure 1). Other compounds such
as tomaymycin (2) and DC-81 (3) were later isolated.5
The PBDs are unique in exerting their activity through
* Correspondence to: Prof. David E. Thurston, University of
Portsmouth. Phone: (+44) 01705-843598 (+voice mail). Fax: (+44)
01705-843573. E-mail: david.thurston@port.ac.uk.
† University of Portsmouth.
‡ University College London Medical School.
§ University of Greenwich.
∇ Institute of Cancer Research.
10.1021/jm981124d CCC: $18.00 © 1999 American Chemical Society
Published on Web 09/16/1999

Novel Epoxide-Containing Pyrrolobenzodiazepine
J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 20 4029
F igu r e 1. Structures of the PBD monomers (1-3), the PBD dimer DSB-120 (4: n ) 3), and the C11a (5) and C8 (6) PBD-
epoxides.
In an attempt to extend the number of base pairs
spanned by these molecules, PBD “dimers” have been
synthesized comprising two PBD units joined through
either their C79 or C810,11 positions via flexible linkers.
Suggs and co-workers have produced C7/C7′-linked
compounds joined through alkanediyldisulfide or al-
kanediyldioxy linkers (some including nitrogen heteroa-
toms) which have been shown to cross-link DNA.9
Thurston and co-workers have produced a series of C8/
C8′-alkanediyldioxy-bridged PBD dimers with linkers
of varying lengths (e.g., 4: n ) 3-6) that are isohelical
with the minor groove of DNA and provide a direct
correlation between DNA interstrand cross-linking ef-
ficiency and cytotoxicity across a number of cell lines.11
With the aim of producing a PBD monomer with DNA
interstrand cross-linking ability, Confalone and co-
workers12 synthesized a PBD analogue with an epoxide
functionality incorporated at the C11a position (5).
Although DNA-binding data were not reported for this
compound, structure-activity considerations suggest5,13
that attachment of the epoxide moiety to the C11a
position may sterically hinder the fit of the PBD within
the minor groove as well as interfere with nucleophilic
attack of a guanine C2-NH2 at the C11 position.
On this basis, we designed a new guanine-specific
PBD-based cross-linking agent by attaching the second
alkylating functionality at the C8 position of the PBD
to overcome any potential problems with steric hin-
drance. Although other types of alkylating moieties were
considered, the epoxide group was chosen as the second
alkylating moiety, because epoxides have an established
base selectivity for guanines, interacting at either the
N7 or C2-NH2 position in the major and minor grooves,
respectively.14 For example, a preference for interaction
with the N7 position of guanine in the major groove is
shared by the epoxide-containing cross-linking antitu-
mor antibiotic carzinophilin (azinomycin),15 certain non-
bay-region diol epoxides of the benzo[a]pyrenes,16 the
activated epoxide-containing form of aflatoxin B1,14 non-
groove-specific compounds such as the 1,2-epoxyal-
kanes,17 and the cross-linking agent diepoxybutane.18
In contrast, other polycyclic aromatic hydrocarbon diol
epoxides derived from benzo[a]pyrene19,20 and benzo[c]-
phenanthrene21 have a preference for the exocyclic C2-
NH2 groups of guanines in the minor groove. Therefore,
to design a bifunctional PBD molecule (i.e., 6) with the
potential to cross-link two separated guanine C2-NH2
groups within the minor groove, the epoxide moiety was
chosen to complement the guanine-specific PBD nucleus.
It was decided to join the epoxide through the C8
position of the PBD in order to retain isohelicity with
the contour of the minor groove of the host DNA
molecule.5
Resu lts a n d Discu ssion
Syn th esis. Numerous synthetic routes to the pyr-
rolobenzodiazepine ring system have been investigated
since Leimgruber reported the first synthesis of anthra-
mycin,22 and these have been reviewed.23 The main
problem associated with the synthesis of PBDs relates
to generation of the relatively labile N10-C11 imine
moiety which is usually introduced at the final step of
the synthetic pathway.23 However, synthesis of the
target molecule 6 was further complicated by the
presence of the C8-epoxide moiety, which is also chemi-
cally reactive. Therefore, a number of different strate-
gies were investigated and are described below.
1. A-Rin g F r a gm en t (Sch em e 1). As epoxides can
be readily synthesized from alkenes, the 4-alkenoxyni-
trobenzoic acid 12 was initially targeted as a suitable
A-ring fragment and epoxide precursor. Vanillic acid (7)
was first converted to the corresponding methyl ester 8
and then alkylated with NaH and allyl bromide to
introduce the propenoxy group (9). Facile nitration at
the C6 position afforded 11 in >90% yield, and the
required allyl ether 12 was obtained by hydrolysis.
Alternatively, vanillic acid (7) could be directly alkylated
to afford 10, which could then be nitrated to 12 with
SnCl4 and fuming HNO3 at -25 °C. Although the first
route requires two additional steps, it results in a higher
yield of 12 (73% compared to 55%) and a shorter overall
reaction time.
2. Attem p ted Ep oxid a tion of th e P r o-C8-P r op e-
n oxy F u n ction a lity in th e P r esen ce of th e Dieth yl
Th ioa ceta l (Sch em e 1). The main difficulty associated
with PBD synthesis is formation of the labile N10-C11
carbinolamine-imine functionality to afford PBDs in
practical yields.23 For the targeted epoxide-PBD 6, this
problem was exacerbated by the requirement to include
an equally labile epoxide group. Initial studies involved
the “thioacetal” route24 which has the advantage of
maintaining stereochemical integrity at the C11a posi-

4030 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 20
Wilson et al.
Sch em e 1a
a
Reagents: (a) MeOH/H2SO4/∆/16 h; (b) DMF/NaH/room temperature/2 h, then allyl bromide/room temperature/18 h; (c) method A,
CH2Cl2/SnCl4/HNO3/-25 °C/30 min or method B, HNO3/0 °C/1 h, room temperature/18 h; (d) THF/aq NaOH/room temperature/24 h; (e)
CH2Cl2/DCC/room temperature/4 h, then 13a /room temperature/16 h; (f) CHCl3/3-CPBA/0 °C/8 h; (g) MeOH/SnCl2‚2H2O/∆/3 h; (h) 14b
or 17/MeCN-H2O/HgCl2/CaCO3/room temperature/24 h; (i) CHCl3/3-CPBA/0 °C/18 h; (j) 14b/dioxane/aq Na2CO3/Fmoc-Cl/0 °C/4 h, then
room temperature/24 h; (k) CH2Cl2/3-CPBA/room temperature/72 h; (l) DMF/Bu4N+F-/room temperature/15 min.
Sch em e 2a
tion of the PBD. The propenoxy thioacetal 14a (Scheme
1) was prepared by coupling the nitro acid A-ring (12)
to the thioacetal C-ring fragment 13a 23,24 using dicy-
clohexylcarbodiimide (DCC) in CH2Cl2 at room temper-
ature. Although peracids are known to oxidize sulfides
to sulfoxides and sulfones at room temperature, selec-
tive oxidation of the allylic functionality of 14a to give
14c was attempted using 3-chloroperbenzoic acid (3-
CPBA) at 0 °C in CHCl3. However, this led to oxidation
of both the propenoxy and thioacetal functionalities to
different extents, affording a mixture of epoxide/prope-
noxy sulfoxides and sulfones as evident from mass
spectral and NMR data.
3. Attem p ted Oxid a tion of th e C8-P r op en oxy
a
Reagents: (a) hydrolysis (H+); (b) method A, MeOH/10% Pd-
C/H2/1 atm/room temperature/16 h or method B, MeOH/10% Pd-
C/1,4-cyclohexadiene/room temperature/18 h or method C, MeOH/
10% Pd-BaCO3/H2/1 atm/room temperature/24 h; (c) MeOH/
SnCl2‚2H2O/∆/4 h; (d) MeOH/SnCl2‚2H2O/∆/2 h; (e) CHCl3/3-
CPBA/room temperature/96 h or Cl(CH2)2Cl/3-tert-butyl-4-hydroxy-
5-methylphenyl sulfide/3-CPBA/∆/2 h; (f) method A, MeOH/10%
Pd-C/H2/1 atm/room temperature/8 h or method B, MeOH/10%
Pd-BaCO3/H2/1 atm/room temperature/16 h or method C, MeOH/
10% Pd-C/1,4-cyclohexadiene/room temperature/18 h or method
D, MeOH/SnCl2‚2H2O/∆/30 min; (g) Amberlite resin.
F u n ction a lity in th e P r esen ce of th e N10-C11
Im in e (Sch em e 1). Even though it is known that
3-CPBA is capable of oxidizing the N10-C11 imine
functionalities of PBDs at room temperature to afford
the corresponding dilactams25 (e.g., structure 16 in
Scheme 1), it was decided to attempt a selective oxida-
tion of the propene moiety of the fully formed C8-
propenoxy-PBD (15) by exploiting either temperature
control or orthogonal protection of the N10-C11 imine.
To produce 15, the nitro thioacetal 14a was reduced to
the amino thioacetal 14b with SnCl2‚2H2O in refluxing
MeOH and then cyclized (49% from 14b) by treatment
with HgCl2 and CaCO3 in MeCN-H2O at room tem-
perature for 48 h. However, oxidation of 15 with 1 equiv
of 3-CPBA in CHCl3 at 0 °C for 18 h led to selective
oxidation of the N10-C11 imine moiety to afford the
novel C8-propenoxy-PBD dilactam (16) in 70% yield.
Attempts to protect the N10-C11 imine functionality of
15 as the C11-propoxy ether prior to oxidation of the
C8-propenoxy moiety failed, affording only the dilactam
16.
4. Attem p ted Red u ction of th e Ar om a tic Nitr o
F u n ction a lity in th e P r esen ce of th e Ep oxid e
(Sch em e 2). Due to the problems associated with the
above approaches, the use of the dimethyl acetal pro-
tecting group was explored to avoid the difficulties

Novel Epoxide-Containing Pyrrolobenzodiazepine
J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 20 4031
Ta ble 1.a Thermal Denaturation Data with CT-DNA at a [PBD]/[DNA] Molar Ratio of 1:5 and in Vitro Cytotoxicity Data in the
A2780/A2780cisR, CH1/CH1cisR, and SKOV-3 Cell Lines for the PBD Epoxide 6, C8-O-Allyl Precursor 15, DC-81 (3), and DSB-120 (4)
∆Tm (°C) after incubation for
IC50 (µM)b,c
compd
0 h
4 h
18 h
36 h
SKOV-3
A2780
A2780cisR
RFd
CH1
CH1cisR
RFd
6e
15
0.5
0.4
0.3
0.8
0.7
0.5
1.3
0.8
0.7
1.5
-
1.3
3.3
-
0.74
4.2
0.045
1.05
-
0.0072
0.16
0.2
1.65
-
0.21
6.1
4.4
1.6
-
29.2
38.1
0.059
0.16
0.10
0.033
0.029
0.047
0.135
-
0.022
0.17
0.8
0.8
-
0.7
5.9
DC-81 (3)
DSB-120 (4)
cisplatin
0.7
10.2
13.1
15.1
15.4
a
Determined for CT-DNA (100 µM) at pH 7.0 with a fixed [PBD]/[DNA] molar ratio of 1:5 (see text). For native CT-DNA, Tm ) 67.83
b
(0.06 °C. Drug-induced ∆Tm values represent the mean from at least 3 determinations (values (0.1-0.2 °C). Dose of compound required
to inhibit cell growth by 50% compared to untreated cell controls. Cells were incubated with the drug for 96 h at 37 °C. In vitro cytotoxicity
data for cisplatin are included for comparative purposes. c A2780cisR and CH1cisR are cisplatin-resistant counterparts of the A2780 and
CH1 cell lines, respectively. Relative resistance factor given by IC50(resistant)/IC50(parent). e Evaluated as a diastereomeric mixture
d
due to the racemic epoxide moiety.
relating to sulfur oxidation. For this route, the acetal
epoxide 21a is the key intermediate that can be reduced
to the amine 21b and then cyclized with weakly acidic
Amberlite resin23,26 to afford the required epoxide PBD
6. Therefore, the propenoxy nitro acid 12 (Scheme 1)
was coupled23 to the pyrrolidine fragment 13b (Scheme
1), either by initial conversion to its acid chloride or by
using DCC, the two methods affording 20a in overall
yields of 53% and 19%, respectively. 20a was then
treated with 3-CPBA in CHCl3 at room temperature.
However, even after 96 h reaction was incomplete and
workup afforded 21a in only 10% yield. In an attempt
to improve the yield, radical inhibitors such as 3-tert-
butyl-4-hydroxy-5-methylphenyl sulfide were investi-
gated so that higher temperatures could be used in the
presence of the oxidizing agent.27 This approach suc-
ceeded, and use of the sulfide inhibitor with 3-CPBA in
Cl(CH2)2Cl under reflux for 2 h afforded the novel 21a
in 85% yield.
The penultimate step of this route involved selective
reduction of the nitro group of 21a in the presence of
the epoxide functionality to give the amine 21b. How-
ever, hydrogenation for 8 h at 1 atm pressure with 10%
Pd-C in MeOH also opened the epoxide to give the
novel amino alcohol 23 in good yield. Three milder
reduction methods were investigated, including hydro-
genation in the presence of a “poisoned” catalyst (10%
Pd-BaCO3/MeOH), phase-transfer hydrogenation with
10% Pd-C in 1,4-cyclohexadiene and MeOH at room
temperature, and reduction with SnCl2‚2H2O under
reflux in MeOH. However, in each case quantitative
conversion to 23 occurred within 16 h, 18 h, and 30 min,
respectively.
5. Attem p ted Red u ction of th e Nitr o F u n ction a l-
ity P r ior to Ep oxid a tion (Sch em e 2). Due to the
problems encountered with synthesis of the epoxide
amino acetal 21b, reduction of the nitro group of 20a
prior to epoxidation was attempted. However, hydro-
genation with either 10% Pd-C or 10% Pd-BaCO3 in
MeOH at 1 atm pressure for 16 or 24 h, respectively, or
10% Pd-C in cyclohexadiene and MeOH at room
temperature for 18 h resulted in reduction of both the
propenoxy and nitro groups to give the novel propanoxy
amino acetal 22 rather than 20b. To avoid reduction of
the alkene moiety, SnCl2‚2H2O in refluxing MeOH was
investigated next. Surprisingly, instead of forming the
expected propenoxy amino acetal 20b, simultaneous
nitroreduction and cleavage of the acetal occurred after
4 h to afford the propenoxy imine 15 in 73% yield. This
one-pot reductive cyclization reaction represents a novel
method of PBD synthesis, particularly as previous
attempts23,28 to cyclize nitro aldehyde intermediates of
type 20c using a number of different reagents have
usually led to low yields of cyclized product due to
overreduction to the corresponding secondary amine
(e.g., 15 where N10-C11 is -NH-CH2-). Interestingly,
in this case, the acetal moiety was shown to be neces-
sary for smooth conversion to the imine, as treatment
of the propoxy nitro aldehyde 20c (prepared by hydroly-
sis of acetal 20a ) with SnCl2‚2H2O in MeOH for 2 h at
reflux caused complete conversion to two new products
identified as the C8-propenoxy imine 15 (20% yield) and
the secondary amine overreduction product (not iso-
lated).
6. Syn th esis of 6 via Ep oxid a tion of th e N10-
F m oc-P r otected In ter m ed ia te 18 (Sch em e 1). Due
to the problems encountered with the previous synthetic
routes, it was decided to utilize the intermediate 14b
in a Fukuyama-type approach.29,30 It had been reported
that the 9-fluorenylmethyloxycarbonyl (Fmoc) group can
be used to protect amines in high yields (88-98%)31 and
can then be easily removed by cleavage with Bu4N+F-
(TBAF) at room temperature.32 Other reports suggested
that an epoxide group should be relatively stable to F-
ions at this temperature.33 Therefore, it was decided to
attempt epoxidation of the N10-Fmoc intermediate 18
in order to provide the required target molecule 6 after
deprotection.
Fmoc protection of the propenoxy amino thioacetal
14b using 9-fluorenylmethyl chloroformate gave 17 in
88% yield. Cleavage of the diethyl thioacetal group with
HgCl2 and CaCO3 in MeCN-H2O effected ring closure
to afford the required N10-Fmoc-protected carbinola-
mine 18 in 95% yield. The next step involved epoxida-
tion of the C8-propenoxy group with 3-CPBA to afford
a diastereomeric mixture of the epoxides (19) in 55%
yield. Finally, removal of the Fmoc protecting group
with TBAF in DMF produced the N10-C11 imine while
leaving the epoxide intact, thus affording the target
epoxide-PBD 6 in 79% yield.
In summary, these results highlight the advantages
of using Fmoc as an N10-protecting group in a Fuku-
yama-type approach to PBD synthesis; these include
rapid N10 deprotection under mild conditions and in
high yield. This strategy should be generally applicable
to the synthesis of other PBD analogues, particularly
those containing acid- or nucleophile-sensitive func-
tional groups.
DNA-Bon d in g Activity. 1. Th er m a l Den a tu r a tion
Stu d ies. The DNA-binding activities of the PBD com-

4032 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 20
Wilson et al.
F igu r e 2. Autoradiograph of a neutral agarose gel showing induction of interstrand DNA cross-links by epoxide-PBD 6 with
linear 32P-end-labeled pBR322 DNA. Reactions (6 and 24 h at 37 °C) for concentrations of 1-1000 µM of 6 were in 25 mM
triethanolamine/1 mM EDTA, pH 7.2 buffer with 10 ng of DNA in a final volume of 50 µL. Reaction was terminated by the
addition of an equal volume of 0.6 M NaOAc, 20 mM EDTA, and 100 mg/mL tRNA, followed by precipitation of the DNA with
EtOH. Dried pellets were taken up in strand separation buffer (30% DMSO in 1 mM EDTA), followed by denaturation for 2 min
at 90 °C and then immediate chilling in an ice-water bath. Electrophoresis was carried out on submerged horizontal 0.8% agarose
gels at 40 V for 16 h with Tris-acetate running buffer. Single-stranded (Cd,SS) and double-stranded (C,DS) controls were included,
and both types of DNAs were quantitated by scanning laser densitometry.
pounds produced were first assessed from the stabiliza-
tion afforded to duplex-form calf thymus DNA toward
thermal denaturation.5,8,10a,11,34 Melting studies (Table
1) show that these agents stabilize the thermal helix f
coil transition (Tm) for the CT-DNA duplex at pH 7.0,
with the epoxide PBD 6 showing greater activity than
either the structurally similar C8-propenoxy-PBD ana-
logue 15 or the less substituted DC-81 (3) molecule, both
of which monoalkylate DNA. Interestingly, the induced
∆Tm shifts are dependent on the length of incubation
at 37 °C, although each compound can effect a similar
level of stabilization without significant DNA-drug
contact (i.e., 0 h incubation). Time course plots of ∆Tm
for 3, 6, and 15 are similar for incubation periods up to
4 h (data not shown), although values for both 3 and
15 reach plateau levels after 4-18 h, whereas the ∆Tm
for 6 continues to increase up to 36 h, giving an ultimate
3-fold increase in stabilization. In contrast, the sym-
metric PBD dimer DSB-120 (4: n ) 3) is much more
effective at short incubation times although there is also
a strong time-dependent component to the binding
profile.10a,11
For many cross-linking agents, it is known that the
formation of DNA interstrand cross-links involves initial
covalent reaction of the agent with a nucleophilic site
on the first strand to form a monoadduct, followed by a
subsequent reaction with a nucleophile on the opposite
DNA strand. This “second-arm” reaction is generally
slow compared to initial monoadduct formation, result-
ing in a diagnostic “biphasic” DNA-binding time course
in DNA-binding assays. It is also known that the
covalent binding of epoxides to DNA can be slow. For
example, the 7-(2,3-epoxypropoxy) analogue of actino-
mycin D has been shown35 to reversibly intercalate
through the minor groove of DNA where the epoxide
group covalently binds to guanine, taking up to 6 h for
complete reaction. On this basis, the results shown in
Table 1 can be interpreted as rapid formation of an
initial monoadduct of 6 with DNA involving reaction of
the PBD N10-C11 imine with the C2-NH2 of a guanine
base, as suggested by the similar early reaction rates
for the monofunctional agents 3 and 15. This is presum-
ably followed by a slower time-dependent covalent
interaction of the epoxide group of DNA-bound 6 to form
either inter- or intrastrand DNA cross-links, manifest-
ing as an increasing ∆Tm value after further incubation.
The slow second-arm alkylation reaction may also
involve an unbonding/bonding-type redistribution of the
covalently bound monoadduct to more favorable or
competitive cross-linking sites in the host DNA mol-
ecule.
2. DNA Cr oss-Lin k in g Assa y. To confirm that the
epoxide-PBD 6 causes DNA cross-linking, an agarose
gel electrophoresis assay was performed,36 the results
of which are shown in the autoradiograph in Figure 2.
The results clearly demonstrate that 6 forms inter-
strand cross-links with DNA efficiently. However, as
with the DNA thermal denaturation studies, the results
also reveal the relatively slow nature of the process.
Following a 6 h incubation, cross-links are detectable
at doses as low as 1-10 µM and increase dose-
dependently reaching almost 100% at 500 µM or above.
Following a 24 h incubation, the level of cross-linking
was higher at each dose, with 100% cross-linking
observed by 250 µM. These results clearly demonstrate
the formation of interstrand DNA cross-links; however,
the presence (and ratio) of intrastrand cross-links and/
or epoxide-DNA or imine-DNA monoadducts cannot
be determined from this experiment.
3. Ta q P olym er a se Assa y. The sequence selectivity
of the covalent interaction of 6 with DNA was investi-
gated using a Taq polymerase stop assay.37 The auto-
radiograph shown in Figure 3 indicates that 6 produces
discrete “stop” sites in a concentration-dependent man-
ner. Furthermore, using the sequencing markers to
establish base positions, it is clear that the stop sites
occur at guanine residues and preferentially within
guanine-rich sequences. This result compares favorably
with the benzo[a]pyrene diol epoxides which also bind
to the C2-NH2 of guanine in the minor groove with a
pronounced sequence preference for guanines immedi-
ately flanked by other guanines and with G3-5 tract
sequences binding 3-4-fold more readily than isolated
guanines.19,20 In the case of the benzo[a]pyrene diol
epoxides, this selectivity can be explained by calculating
the molecular electrostatic potential at the N2 position
of central guanines within different base triplets20,38
which suggests that central guanines have the highest
potentials. Guanines neighboring one other guanine
base have the next highest potentials, and isolated

Novel Epoxide-Containing Pyrrolobenzodiazepine
J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 20 4033
result from intrastrand cross-linked adducts and/or
monoadducts. In addition, data are only available from
one strand of the DNA duplex host. Despite these
limitations, a number of consensus sequences such as
5′-GG, 5′-GC, 5′-CGG, and 5′-GGG are clearly discern-
ible.
4. Molecu la r Mod elin g. Modeling experiments to
rationalize the interstrand cross-linking results started
with the assumption that 6 bonds to separated guanines
in the minor groove due to the known base preference
of imine and epoxide moieties and the known groove
preference of the PBDs.5 Preliminary studies of the
interaction of 6 with model DNA duplex sequences
suggested that upon bonding of the PBD N10-C11 imine
to a guanine C2-NH2 in the minor groove (e.g., G3 in
Figure 4a), the unopened pendant epoxide moiety
should be approximately positioned between the neigh-
boring base pair (i.e., X2‚Y2) and the next adjacent base
pair (i.e., G1‚C1). Assuming that this could lead to either
intra- or interstrand cross-links with either of the two
base pairs implicated, four distinct models were envis-
aged (Figure 4a-d), each with the benzenoid A-ring of
6 oriented toward the 5′-end of the PBD-modified strand
with (S)-stereochemistry at the C11 position of the PBD
(designated “5S”). Four equivalent models were also
examined (not shown) with the adducts in the reverse
orientation with the A-rings orientated toward the 3′-
end of the covalently modified strand and with C11(S)
stereochemistry (designated “3S”). Models with the C11-
(R)-configuration were not considered, as this configu-
ration is known to be of significantly higher energy for
structures based on the DC-81 (3) core.5 However, as
both (R)- and (S)-stereochemistries are possible for the
C8 chiral epoxide residue, 16 models in total were
examined for DNA-6 cross-link formation.
Molecular mechanics (MM) and dynamics (MD) simu-
lations using the CHARMm force field39 were carried
out on these models, and the results are collected in
Table 2. The 9-mer duplex DNA sequences for the
simulations (see Table 2) used a core 5′-AGA motif (the
preferred PBD binding sequence5) wherever possible
(i.e., not in the case of the nonadjacent models [e.g.,
Figure 4a,b]) and were terminated with flanking G‚C
base pairs to improve helical stability (see Experimental
Section). The results shown in Table 2 indicate three
energetically favored models. The lowest-energy model
F igu r e 3. Autoradiograph showing the results of the Taq
polymerase stop assay for the epoxide-PBD 6. The agent was
incubated at 37 °C with linearized calf thymus DNA for 24 h.
Lanes a-i correspond to increasing concentrations of 6 (0, 1,
10, 25, 50, 100, 250, 500, and 1000 µM), and lane j is 10 µM
cisplatin after incubation with the DNA at room temperature
for 1 h. Lanes T, G, C, and A correspond to thymine, guanine,
cytosine, and adenine sequencing lanes, respectively.
(depicted in Figure 5, Einteraction ) -22.1 kcal mol-1
)
involves an interstrand cross-link between adjacent
guanines on opposite strands (i.e., model 4d), with (3S)-
orientation of the PBD and (S)-stereochemistry for the
epoxide. The next most favored models also involve
interstrand cross-links but to nonadjacent guanines
separated by one base pair (i.e., model 4b) and with (5S)-
alignment of the PBD and (R)- or (S)-stereochemistries
for the epoxide (-18.8 and -19.3 kcal mol-1, respec-
tively). A second adjacent-base interstrand cross-linking
model (Figure 4d) is the next preferred model (-17.3
kcal mol-1) with (3S)-orientation of the PBD and an (R)-
configuration for the epoxide. The first intrastrand
cross-link model involves nonadjacent guanines (model
4a, (3S)-PBD, (R)-epoxide) and has an energy of -15.6
guanines are associated with the lowest potentials.
On the basis of the known preference for both epoxide
and imine alkylating moieties to target guanine bases,
and the minor groove preference and base-pair span of
the PBDs, the interstrand adduct of the bifunctional
molecule 6 is thought to involve interaction with either
two adjacent guanines on opposite strands or an analo-
gous site separated by one spacing base pair [i.e., (G/
C)(C/G) or (G/C)N(C/G), respectively; N ) any other
base pair], with the alkylating units bonded to the
guanine C2-NH2 functions. However, a more detailed
analysis of the sequence selectivity is restricted by the
fact that some of the stop sites shown in Figure 3 may
kcal mol-1
.
The most energetically preferred model (Figure 4d,
(3S)-PBD, (S)-epoxide, -22.1 kcal mol-1) implicates a

4034 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 20
Wilson et al.
models (i.e., 4a : 5′-GNG and 4c: 5′-GG) can also be
found in Figure 3. Therefore, until further experimental
data become available, the modeling results can be used
only in a predictive sense.
5. In Vitr o Cytotoxicity of th e Ep oxid e-P BD 6.
The cytotoxic potencies of the epoxide-PBD 6, the C8-
propenoxy-PBD (15), and DC-81 (3) were determined
in five cell lines, and the results are shown in Table 1.
Data for the cross-linking agents DSB-120 (4: n ) 3)
and cisplatin (which forms intra- and interstrand cross-
links in the major groove of DNA) are included for
comparison. Using the median of the cytotoxicity values
for each compound across the five cell lines, the DNA
cross-linking epoxide PBD 6 (median ) 0.059 µM) is
∼18-fold more cytotoxic than 15 (median ) 1.05 µM)
which is only capable of monoalkylating DNA. The
median IC50 for 6 is only slightly higher than that for
the established interstrand minor groove cross-linking
agent DSB-120 (median ) 0.033 µM), also supporting
an interstrand cross-linking mechanism of action. A2780
(followed closely by CH1cisR) is the most sensitive cell
line, in which 6 (IC50 ) 0.045 µM) is ∼23-fold more
cytotoxic than 15, suggesting that these cells may be
more prone to the effects of cross-linking, as with DSB-
120 which is significantly more cytotoxic (IC50 ) 0.0072
µM) in A2780 compared to the other four cell lines.
Furthermore, the significant potency of 6 in both the
cisplatin-resistant cell line CH1cisR and its parent CH1
(giving a resistance factor of 0.8) indicates a lack of
cross-resistance with cisplatin. As with previous studies
on series of PBD compounds,11 the higher cytotoxicity
of 6 compared to 15 is reflected in the thermal dena-
turation data for these compounds (see Table 1). Finally,
it is important to note that 6 was evaluated as a mixture
of diastereomers due to the chirality of the epoxide
moiety. Although the molecular modeling studies sug-
gest that, in some cases, there is little difference in
binding energy between DNA adducts with either the
(R)- or (S)-epoxide configuration, it is possible that
greater cytotoxicity may be realized with single diaster-
eomers of 6.
Con clu sion s
The first example (6) of a pyrrolo[2,1-c][1,4]benzo-
diazepine molecule with an epoxide moiety at the C8
position has been synthesized via a novel pro-N10
deprotection procedure utilizing a variation of a method
originally pioneered by Fukuyama and co-workers.30
The procedure adopted here differed in two key as-
pects: (i) the B-ring of 6 was closed via cleavage of a
thioacetal functionality rather than through an oxida-
tive cyclization process, and (ii) the N10-Alloc protecting
group originally used by Fukuyama was replaced with
a Fmoc moiety which offers significant advantages
including rapid deprotective cleavage under mild condi-
tions. Therefore, this new strategy should have general
applicability to the synthesis of other PBD analogues,
particularly those containing acid- or nucleophile-sensi-
tive functional groups.
F igu r e 4. Possible models for covalent interaction between
the epoxide-PBD 6 and the 9-mer DNA duplexes shown in
Table 2 (column 1), with the difunctional molecule bound to
nonadjacent (a or b) or adjacent (c or d) guanine bases, leading
to either intrastrand (a or c) or interstrand (b or d) DNA cross-
links. The PBD is in the favored C11(S)-configuration for all
models shown, with the aromatic A-ring oriented toward the
5′-end (“5S”) of the covalently modified strand. The epoxide
moiety may have either the (R)- or (S)-configuration prior to
ring opening.
5′-GC sequence. In support of this model, analysis of
the stop sites shown in Figure 3 reveals that 8 of the
17 marked sites include 5′-GC sequences. The next-
favored models (Figure 4b, (5S)-PBD, (R/S)-epoxide)
require a 5′-CNG sequence, and 4 of the 17 observed
stop sites include this motif. Although the modeling
suggests that intrastrand cross-linking is less energeti-
cally favorable, potential stop sites relating to the
sequence requirements for both of the intrastrand
On the basis of the results of the electrophoretic assay
shown in Figure 2, 6 induces interstrand DNA cross-
links at low micromolar concentrations, a result sup-
ported by both the DNA thermal denaturation studies
and the enhanced cytotoxicity profile compared to

Novel Epoxide-Containing Pyrrolobenzodiazepine
J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 20 4035
Ta ble 2. Interaction Energies of the Epoxide-PBD 6 with the 9-mer DNA Duplex Sequences (shown in column 1)a
sequence of 9-mer
bonding site
of PBD
bonding site
of epoxide
PBD
orientation
epoxide
configuration
cross-linking
mode
DNA-PBD
model Figure
Einteracton
duplex (5′)
(kcal mol-1
)
GCGAG5AG7CG
GCGAG5AC7CG
GCGAG5AG7CG
GCGAG5AC7CG
GCG3AG5AGCG
GCC3AG5AGCG
GCG3AG5AGCG
GCC3AG5AGCG
GCGAG5G6GCG
GCGAG5C6GCG
GCGAG5G6GCG
GCGAG5C6GCG
GCGG4G5AGCG
GCGC4G5AGCG
GCGG4G5AGCG
GCGC4G5AGCG
G5
G5
G5
G5
G5
G5
G5
G5
G5
G5
G5
G5
G5
G5
G5
G5
G7
G7
G7
G7
G3
G3
G3
G3
G6
G6
G6
G6
G4
G4
G4
G4
3S
3S
3S
3S
5S
5S
5S
5S
3S
3S
3S
3S
5S
5S
5S
5S
R
R
S
S
R
R
S
S
R
R
S
S
R
R
S
S
intra
inter
intra
inter
intra
inter
intra
inter
intra
inter
intra
inter
intra
inter
intra
inter
4a
4b
4a
4b
4a
4b
4a
4b
4c
4d
4c
4d
4c
4d
4c
4d
-15.6
-12.0
-10.0
-14.4
-15.4
-18.8
-6.0
-19.3
+2.6
-17.3
-0.5
-22.1
-5.0
-10.2
-8.9
-11.1
a
The results of 16 molecular mechanics/dynamics simulations are shown (column 8), with the molecule bonding either across adjacent
G‚C base pairs (i.e., models 4c or d) or those one base pair removed (i.e., models 4a or b) (columns 2, 3). All models have the C11 position
of the PBD in the (S)-configuration, with 6 orientated toward either the 5′ (5S) or 3′ (3S) end of the covalently modified strand (column
4). The epoxide moiety was modeled in both the (R)- and (S)-configurations (column 5), both of which could give rise to intra- or interstrand
DNA cross-links (column 6).
(Figure 3) demonstrate that 6 produces concentration-
dependent sequence-selective covalent adducts in gua-
nine-rich regions of DNA. Molecular modeling studies
have suggested low-energy models for such adducts
where 6 spans either two adjacent [i.e., (GC)‚(CG)] base
pairs or an analogous arrangement with a further
spanned base pair [i.e., (GC)‚N‚(GC)] with the electro-
philic epoxide and imine functionalities alkylating
guanine residues on opposite strands. However, from
the data available so far, it is not possible to rule out
the formation of intrastrand cross-links and/or monoalky-
lated adducts, even though the molecular modeling
studies suggest that interstrand cross-linking leads to
the most energetically favored adducts.
The epoxide PBD 6 has significant cytotoxicity across
the five cell lines examined, with activity in the 45-60
nM range in the A2780, CH1, and CH1cisR cell lines.
The similar potency in both the cisplatin-resistant cell
line CH1cisR and the parent CH1 (resistance factor )
0.8) indicates that 6 is not cross-resistant with cisplatin.
This suggests that, as is the case with the symmetric
PBD dimers [e.g., DSB-120 (4: n ) 3)],10,11 the bifunc-
tional epoxide PBD 6 may have potential application
in the treatment of certain cisplatin-resistant tumors.
Future studies will include the synthesis of 6 in its
diastereomerically discrete forms, each of which may
possess greater cytotoxicity and DNA cross-linking
efficiency than the racemic mixture examined here, and
may provide more interpretable patterns of stop sites
in the Taq polymerase assay.
Exp er im en ta l Section
Syn th etic Ch em istr y. Melting points (mp) were deter-
F igu r e 5. Energy-minimized representation of the adduct of
mined on a Gallenkamp P1384 digital melting point apparatus
epoxide-PBD 6 in the minor groove of the d(GCCAGAGCG)‚
and are uncorrected. Infrared (IR) spectra were recorded using
d(CGCTCTGGC) duplex according to the interstrand cross-
1
a Perkin-Elmer 297 spectrophotometer. All H and 13C NMR
linking model shown in Figure 4b. The epoxide and N10-C11
imine functionalities are covalently bound to the exocyclic C2-
spectra were recorded in CDCl3 solution (unless stated other-
wise) using a J EOL GSX 270 MHz FT-NMR instrument
NH2 groups of guanines on opposite strands separated by one
operating at 20 ( 1 °C. Chemical shifts are reported in parts
per million (δ) downfield from Me4Si. Spin multiplicities are
described as: s (singlet), bs (broad singlet), d (doublet), dd
(doublet of doublets), t (triplet), q (quartet), or m (multiplet).
Mass spectra (MS) were recorded using a J EOL J MS-DX 303
GC-mass spectrometer (EI mode: 70 eV, source 117-147 °C).
base pair. The molecule is oriented with the epoxide moiety
toward the 5′-end of the covalently modified strand.
monofunctional PBD analogues of similar structure.
Furthermore, the results of the Taq polymerase assay

4036 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 20
Wilson et al.
Accurate molecular masses (HRMS) were determined by peak
matching using perfluorokerosene (PFK) as an internal mass
marker. Column chromatography was performed using chro-
matography grade silica gel 400 (Aldrich). Thin-layer chro-
matography (TLC) was performed using GF254 silica gel
including fluorescent indicator on glass plates. All solvents and
reagents, unless otherwise stated, were purchased from Ald-
rich and were used as supplied. Anhydrous solvents were
prepared by distillation under dry N2 in the presence of the
appropriate drying agent and were stored over 4-Å molecular
sieves or sodium wire. Intermediates and final products did
not provide useful combustion analysis data as they were
either oils and/or were found to be prone to either oxidation
(e.g., amines such as 14b, 20b, 21b, 22, and 23) and/or water
absorption. Therefore, high resolution mass spectrometry data
were obtained throughout the study, and many intermediates
were used directly in subsequent steps.
11 as a yellow solid (4.92 g, 93%): mp 82-83 °C; IR (Nujol) v
2924, 2854, 1724 (CdO), 1606 (CdC), 1540 (NO2), 1458, 1377,
1295, 1212, 1138, 985, 870, 814; H NMR δ 3.91 (s, 3H, H-8),
1
3.98 (s, 3H, H-9), 4.68-4.71 (m, 2H, H-10), 5.36-5.50 (m, 2H,
H-12), 5.99-6.14 (m, 1H, H-11), 7.08 (s, 1H, H-6), 7.47 (s, 1H,
H-3); 13C NMR δ 53.3 (C-8), 56.7 (C-9), 70.3 (C-10), 108.5 (C-
3), 110.9 (C-6), 119.5 (C-12), 121.8 (C-1), 131.5 (C-11), 140.9
(C-2), 149.2 (C-5), 152.8 (C-4), 166.4 (C-7); MS m/z (relative
intensity) 267 ([M]+, 100), 236 (24), 226 (67), 210 (13), 195 (11),
181 (10), 165 (14), 121 (30), 109 (11), 93 (11), 77 (9), 65 (8), 59
(22), 51 (5). HRMS: calcd for C12H13O6N (267.0743), found
267.0817.
5-Meth oxy-2-n itr o-4-(2-p r op en oxy)ben zoic Acid (12). A
mixture of ester 11 (2.2 g, 8.2 mmol) and aqueous NaOH (25
mL of 1 M) in THF (25 mL) was stirred at room temperature
for 24 h. The THF was evaporated in vacuo and the aqueous
layer extracted with EtOAc (3 × 20 mL) which was then
discarded. The aqueous phase was acidified to pH 1 (concen-
trated HCl) and extracted with EtOAc (3 × 25 mL). The
combined organic layer was washed with water (4 × 20 mL)
and brine (20 mL), dried (MgSO4), and evaporated in vacuo to
give the nitrobenzoic acid 12 as a yellow solid (1.98 g, 95%):
mp 118-119 °C; IR (Nujol) v 3287-2010, 1684 (CdO), 1597
(CdC), 1539 (NO2), 1459, 1377, 1275, 1217, 919, 870, 760, 722;
1H NMR (CDCl3-DMSO-d6) δ 3.97 (s, 3H, H-8), 4.68-4.70 (m,
2H, H-9), 5.35-5.49 (m, 2H, H-11), 5.99-6.14 (m, 1H, H-10),
7.16 (s, 1H, H-6), 7.41 (s, 1H, H-3); 13C NMR (CDCl3-DMSO-
d6) δ 56.5 (C-8), 70.1 (C-9), 108.3 (C-3), 111.1 (C-6), 119.3 (C-
11), 122.6 (C-1), 131.6 (C-10), 141.1 (C-2), 149.0 (C-5), 152.5
(C-4), 167.3 (C-7); MS m/z (relative intensity) 253 ([M]+, 97),
236 (5), 212 (20), 208 (45), 195 (30), 167 (41), 121 (63), 111
(10), 95 (14), 77 (11), 65 (7), 51 (18), 41 (100). HRMS: calcd
for C11H11O6N (253.0586), found 253.0603.
Met h yl 4-H yd r oxy-3-m et h oxyb en zoa t e (8). Concen-
trated H2SO4 (16.56 g, 0.16 mol, 0.5 equiv) was added dropwise
to a stirred solution of vanillic acid (7) (50.44 g, 0.30 mol) in
MeOH (250 mL) at room temperature, and the mixture was
then heated at reflux for 16 h. The solvent was removed in
vacuo and the residue dissolved in EtOAc (200 mL). This
solution was washed with saturated NaHCO3 (3 × 100 mL),
water (2 × 100 mL), and brine (100 mL), then dried (MgSO4),
and evaporated in vacuo to give ester 8 as a beige solid (50.94
g, 93%): mp 69-70 °C; IR (Nujol) v 3536 (OH), 2924, 2854,
1699 (CdO), 1600 (CdC), 1512, 1462, 1376, 1284, 1220, 1188,
1
1113, 1026, 984, 872, 768; H NMR δ 3.89 (s, 3H, H-8), 3.91
(s, 3H, H-9), 6.31 (bs, 1H, H-10), 6.94 (d, 1H, J ) 8.3 Hz, H-5),
7.55 (d, 1H, J ) 1.8 Hz, H-2), 7.63 (dd, 1H, J ) 1.8 & 8.3 Hz,
H-6); 13C NMR δ 52.0 (C-8), 56.1 (C-9), 111.8 (C-2), 114.2 (C-
5), 122.2 (C-1), 124.2 (C-6), 146.3 (C-3), 150.12 (C-4), 167.0 (C-
7); MS m/z (relative intensity) 182 ([M]+, 74), 167 (7), 151 (100),
139 (4), 123 (17), 108 (7), 77 (3), 65 (4), 51 (6). HRMS: calcd
for C9H10O4 (182.0579), found 182.0583.
(2S)-N-(5-Met h oxy-2-n it r o-4-[2-p r op en oxy]b en zoyl)-
p yr r olid in e-2-ca r boxa ld eh yd e Dieth yl Th ioa ceta l (14a ).
1,3-Dicyclohexylcarbodiimide (815 mg, 4 mmol) was added to
a solution of benzoic acid 12 (1 g, 4 mmol) in dry CH2Cl2 (50
mL), and the mixture was stirred for 4 h at room temperature
under a N2 atmosphere. A solution of thioacetal 13a 23,24 (811
mg, 4 mmol) in dry CH2Cl2 (10 mL) was then added and the
mixture stirred for 16 h at room temperature. The mixture
was filtered, and the filtrate evaporated in vacuo to afford the
coupled benzamide thioacetal 14a as a pale yellow oil (1.73 g,
99%): IR (neat) v 2924, 2854, 1634 (CdO), 1580 (CdC), 1524
(NO2), 1462, 1377, 1328 (NO2), 1276, 1215, 1060, 981, 850, 825,
Meth yl 3-Meth oxy-4-(2-p r op en oxy)ben zoa te (9). A solu-
tion of 8 (25 g, 137 mmol) in dry DMF (50 mL) was added
dropwise to a stirred suspension of NaH (5 g, 208 mmol, 1.5
equiv) in dry DMF (200 mL) at 0 °C under a N2 blanket, and
stirring was continued for 2 h at room temperature. Allyl
bromide (33.2 g, 274 mmol, 2 equiv) was added dropwise at 0
°C, and stirring continued for 18 h at room temperature. Brine
(300 mL) and EtOAc (200 mL) were added to the mixture with
shaking and the phases allowed to separate. The aqueous
phase was extracted with EtOAc (2 × 200 mL) and the
combined organic phase was washed with brine (2 × 200 mL),
then dried (MgSO4), and evaporated in vacuo. Toluene (100
mL) was added to the residue and the mixture evaporated in
vacuo. This process was repeated a further three times. The
resulting waxy solid was triturated with hexane (20 mL) and
filtered, and this process was repeated five times. The residue
was dried for 16 h at 50 °C to give allyl ether 9 as a white
solid (28.06 g, 92%): mp 54-55 °C; IR (Nujol) v 2924, 2854,
1718 (CdO), 1588 (CdC), 1508, 1460, 1377, 1284, 1217, 1170,
1128, 998, 947, 870, 759; 1H NMR δ 3.89 (s, 3H, H-8), 3.93 (s,
3H, H-9), 4.65-4.68 (m, 2H, H-10), 5.27-5.49 (m, 2H, H-12),
6.01-6.15 (m, 1H, H-11), 6.88 (d, 1H, J ) 8.4 Hz, H-5), 7.56
(d, 1H, J ) 2.0 Hz, H-2), 7.64 (dd, 1H, J ) 2.0 & 8.4 Hz, H-6);
13C NMR δ 52.0 (C-8), 56.0 (C-9), 69.7 (C-10), 111.9 (C-2), 112.2
(C-5), 118.5 (C-12), 122.8 (C-1), 123.4 (C-6), 132.5 (C-11), 148.9
(C-3), 151.9 (C-4), 166.9 (C-7); MS m/z (relative intensity) 222
([M]+, 100), 207 (3), 191 (17), 181 (98), 163 (8), 153 (11), 149
(20), 121 (19), 107 (4), 94 (9), 79 (12), 65 (4), 59 (14), 51 (9).
1
761, 668; H NMR δ 1.31-1.39 (m, 6H, H-18 & H-20), 1.62-
2.36 (m, 4H, H-9 & H-10), 2.59-2.87 (m, 4H, H-17 & H-19),
3.19-3.34 (m, 2H, H-8), 3.96 (s, 3H, H-12), 4.42-4.46 (m, 1H,
H-11), 4.69-4.71 (m, 2H, H-13), 4.94 (d, 1H, J ) 3.9 Hz, H-16),
5.36-5.51 (m, 2H, H-15), 6.01-6.15 (m, 1H, H-14), 6.84 (s, 1H,
H-6), 7.69 (s, 1H, H-3); 13C NMR δ 15.0 & 15.1 (C-18 & C-20),
24.9 (C-9), 26.3 & 27.2 (C-17 & C-19), 26.6 (C-10), 50.2 (C-8),
52.8 (C-16), 56.5 (C-12), 61.0 (C-11), 70.2 (C-13), 108.7 (C-3),
109.3 (C-6), 119.4 (C-15), 128.3 (C-1), 131.6 (C-14), 137.2 (C-
2), 147.8 (C-5), 154.4 (C-4), 166.5 (C-7); MS m/z (relative
intensity) 440 ([M]+, 9), 411 (5), 379 (3), 351 (8), 334 (2), 305
(44), 252 (7), 236 (100), 195 (6), 174 (3), 135 (38), 121 (8), 70
(8). HRMS: calcd for C20H28O5N2S2 (440.1440), found 440.1509.
(2S)-N-(2-Am in o-5-m eth oxy-4-[2-p r op en oxy]ben zoyl)-
p yr r olid in e-2-ca r boxa ld eh yd e Dieth yl Th ioa ceta l (14b).
A solution of nitro thioacetal 14a (1 g, 2.3 mmol) and SnCl2‚
2H2O (2.56 g, 11.4 mmol, 5 equiv) in MeOH (100 mL) was
refluxed for 3 h. After solvent removal in vacuo and digestion
in CHCl3 (300 mL), saturated aqueous NaHCO3 (300 mL) was
added and the mixture was stirred for 16 h at room temper-
ature. The mixture was then filtered through Celite and the
organic layer separated. The aqueous layer was extracted with
CHCl3 (2 × 100 mL) and the combined organic layers were
washed with water (3 × 200 mL), dried (MgSO4), and evapo-
rated in vacuo to give the amine 14b as a yellow oil (746 mg,
80%): IR (neat) v 3440 (NH2), 2923, 2854, 1701 (CdO), 1614
(CdC), 1501, 1451, 1402, 1381, 1340, 1255, 1228, 1191, 1168,
1100, 1030, 998, 870, 820, 755, 701; 1H NMR δ 1.12-1.37 (m,
6H, H-19 & H-21), 1.67-2.31 (m, 4H, H-9 & H-10), 2.65-2.74
Meth yl 5-Meth oxy-2-n itr o-4-(2-pr open oxy)ben zoate (11).
Allyl ether 9 (4.4 g, 19.8 mmol) was added slowly over 15 min
to a stirred solution of HNO3 (70% w/w, 47.22 g, 0.52 mol) at
0 °C. The mixture was stirred at 0 °C for 1 h and then allowed
to return to room temperature over 45 min, before pouring
onto ice and stirring at room temperature for 18 h. The
reaction mixture was then extracted with EtOAc (3 × 50 mL),
and the extracts were washed with saturated aqueous NaH-
CO3 (5 × 100 mL), water (200 mL), and brine (100 mL) and
then dried (MgSO4). Solvent removal in vacuo gave nitro ester

Novel Epoxide-Containing Pyrrolobenzodiazepine
J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 20 4037
(m, 4H, H-18 & H-20), 3.59-3.71 (m, 2H, H-8), 3.79 (s, 3H,
H-12), 4.57-4.69 (m, 4H, H-11, H-13 & H-17), 5.28-5.44 (m,
2H, H-15), 5.99-6.14 (m, 1H, H-14), 6.26 (s, 1H, H-3), 6.83 (s,
1H, H-6); 13C NMR δ 14.9 & 15.2 (C-19 & C-21), 25.2 (C-9),
26.5 & 27.2 (C-18 & C-20), 26.6 (C-10), 51.6 (C-8), 53.2 (C-17),
56.8 (C-12), 60.9 (C-11), 69.5 (C-13), 102.1 (C-3), 110.3 (C-1),
112.8 (C-6), 118.2 (C-15), 132.8 (C-14), 141.1 (C-2), 147.4 (C-
5), 150.9 (C-4), 169.8 (C-7); MS m/z (relative intensity) 410
([M]+, 57), 381 (8), 330 (6), 275 (36), 248 (10), 206 (100), 191
(14), 180 (8), 166 (17), 135 (15), 107 (4), 94 (6), 83 (3), 70 (8),
56 (3). HRMS: calcd for C20H30O3N2S2 (410.1698), found
410.1739.
H-15), 6.02-6.14 (m, 1H, H-14), 6.92 (s, 1H, H-6), 7.31-7.79
(m, 8H, H-21-24 & H-27-30), 7.87 (s, 1H, H-3), 8.99 (bs, 1H,
H-16); 13C NMR δ 14.9 & 15.2 (C-34 & C-36), 25.3 (C-9), 26.5
& 27.3 (C-33 & C-35), 26.8 (C-10), 47.1 (C-19), 52.2 (C-8), 53.2
(C-32), 56.4 (C-12), 61.3 (C-11), 67.1 (C-18), 69.8 (C-13), 105.5
(C-3), 111.7 (C-6), 118.8 (C-15), 120.1 (C-1), 120.7, 125.2, 127.1,
& 127.8 (C-21-24 & C-27-30), 132.4 (C-14), 141.2 (C-2), 141.3
(C-25 & C-26), 143.7 (C-20 & C-31), 143.8 (C-5), 150.2 (C-4),
153.8 (C-17), 168.9 (C-7); MS m/z (relative intensity) 395 (9),
366 (5), 260 (43), 191 (100), 149 (8), 122 (6).
10-(9-F lu o r e n y lm e t h o x y c a r b o n y l)-1,2,3,10,11,11a -
h exa h yd r o-11-h yd r oxy-7-m et h oxy-8-(2-p r op en oxy)-5H -
p yr r olo[2,1-c][1,4]ben zod ia zep in -5-on e (18). A solution of
thioacetal 17 (316 mg, 0.50 mmol), HgCl2 (298 mg, 1.10 mmol,
2.2 equiv), and CaCO3 (125 mg, 1.25 mmol, 2.5 equiv) in
MeCN-H2O (6 mL of 4:1 v/v) was stirred for 24 h at room
temperature. The mixture was filtered through Celite and the
filtrate evaporated in vacuo to give a dark yellow oil. Purifica-
tion by column chromatography (CHCl3) afforded benzodiaz-
epinone 18 as a pale yellow solid (250 mg, 95%): mp 75-76
°C; IR (Nujol) v 3458 (OH), 2923, 2854, 1709 (CdO), 1603 (Cd
(11aS)-7-Meth oxy-8-(2-pr open oxy)-1,2,3,11a-tetr ah ydr o-
5H-p yr r olo[2,1-c][1,4]ben zod ia zep in -5-on e (15). A suspen-
sion of amino thioacetal 14b (110 mg, 0.27 mmol), HgCl2 (160
mg, 0.59 mmol, 2.2 equiv), and CaCO3 (67 mg, 0.67 mmol, 2.5
equiv) in MeCN-H2O (4:1 v/v, 5.5 mL) was stirred for 24 h at
room temperature. The mixture was filtered through Celite
and the filtrate evaporated in vacuo to give a dark yellow oil
which was purified by gradient chromatography (100% EtOAc,
followed by 100:0f95:5 v/v CHCl3-MeOH) to afford the title
compound 15 as a pale yellow oil (47 mg, 61%): IR (neat) v
2923, 2854, 1630 (CdO), 1508 (CdC), 1451, 1263, 1150, 1015,
1
C), 1520, 1453, 1330, 1216, 915, 758; H NMR (CDCl3-CD3-
OD) δ 1.61-2.12 (m, 4H, H-1 & H-2), 3.37-3.62 (m, 2H, H-3),
3.90 (s, 3H, H-12), 4.00-4.69 (m, 7H, H-11a, H-13, H-17, H-18
& H-31), 5.23-5.45 (m, 2H, H-15), 5.63 (d, 1H, J ) 9.5 Hz,
H-11), 6.00-6.14 (m, 1H, H-14), 6.74 (s, 1H, H-6), 6.83-7.69
(m, 9H, H-9, H-20-23 & H-26-29); 13C NMR (CDCl3-CD3-
OD) δ 23.1 (C-2), 28.7 (C-1), 46.6 (C-3), 46.9 (C-18), 56.3 (C-
12), 60.1 (C-11a), 68.5 (C-17), 70.0 (C-13), 85.8 (C-11), 111.2
(C-9), 114.7 (C-6), 118.5 (C-5a), 118.7 (C-15), 120.0, 120.5,
125.1, 125.4, 127.1, 127.2, 127.8 & 127.9 (C-20-23 & C-26-
29), 132.4 (C-14), 141.3 (C-9a), 143.2 (C-24 & C-25), 143.5 (C-
19 & C-30), 143.8 (C-7), 150.3 (C-8), 154.1 (C-16), 167.6 (C-5);
MS m/z (relative intensity) 330 (13), 315 (20), 300 (25), 285
(56), 261 (18), 239 (36), 191 (100), 178 (44), 151 (26), 119 (44),
97 (37), 85 (44), 71 (63), 57 (91).
1
780, 755, 715; H NMR δ 2.02-2.11 (m, 2H, H-2), 2.29-2.37
(m, 2H, H-1), 3.53-3.87 (m, 3H, H-3 & H-11a), 3.96 (s, 3H,
H-12), 4.60-4.72 (m, 2H, H-13), 5.31-5.47 (m, 2H, H-15),
6.02-6.16 (m, 1H, H-14), 6.82 (s, 1H, H-6), 7.52 (s, 1H, H-9),
7.68 (d, 1H, J ) 4.4 Hz, H-11); 13C NMR δ 24.2 (C-2), 29.6
(C-1), 46.6 (C-3), 53.7 (C-11a), 56.1 (C-12), 69.7 (C-13), 110.8
(C-6), 111.5 (C-9), 118.7 (C-15), 120.4 (C-5a), 132.3 (C-14),
140.4 (C-9a), 147.7 (C-7), 150.2 (C-8), 162.4 (C-11), 164.6 (C-
5); MS m/z (relative intensity) 286 ([M]+, 100), 271 (18), 245
(14), 230 (11), 217 (49), 203 (11), 176 (13), 160 (3), 121 (5), 93
(9), 70 (11). HRMS: calcd for C16H18O3N2 (286.1317), found
286.1292.
(11a S)-1,2,3,10,11,11a -Hexa h yd r o-7-m eth oxy-8-(2-p r o-
p en oxy)-5H -p yr r olo[2,1-c][1,4]b en zod ia zep in e-5,11-d i-
on e (16). A solution of 3-chloroperoxybenzoic acid (50-60%
w/w, 60 mg, 0.17 mmol) and 15 (50 mg, 0.17 mmol) in CHCl3
(10 mL) was stirred for 18 h at 0 °C, after which TLC (CHCl3-
MeOH, 95:5 v/v) indicated that reaction was complete. The
solvent was evaporated in vacuo and the residue purified by
gradient chromatography (100:0f85:15 v/v CHCl3-MeOH) to
afford the dilactam 16 as a yellow solid (36 mg, 70%): mp 114-
115 °C; IR (Nujol) v 3500 (NH), 2923, 2854, 1693 & 1653 (Cd
(11a S)-8-(2,3-Ep oxyp r op oxy)-10-(9-flu or en ylm eth oxy-
ca r bon yl)-1,2,3,10,11,11a -h exa h yd r o-11-h yd r oxy-7-m eth -
oxy-5H-p yr r olo[2,1-c][1,4]ben zod ia zep in -5-on e (19). A so-
lution of allyl ether 18 (160 mg, 0.30 mmol) and 3-chloroperoxy-
benzoic acid (50-60% w/w, 149 mg, 0.43 mmol, 1.4 equiv) in
CH2Cl2 (10 mL) was stirred for 72 h at room temperature. The
reaction mixture was then evaporated in vacuo and the residue
purified by gradient chromatography (100:0f96:4 v/v CHCl3-
MeOH) to give epoxide 19 as a yellow oil (90 mg, 55%): IR
(neat) v 3459 (OH), 3040, 2926, 2855, 1711 (CdO), 1605 (Cd
C), 1520, 1455, 1340, 1217, 1060, 990, 914, 810, 755; 1H NMR
δ 1.83-2.17 (m, 4H, H-1 & H-2), 2.69-2.74 (m, 1H, H-15),
2.83-2.87 (m, 1H, H-15), 3.34-3.36 (m, 1H, H-14), 3.38-3.87
(m, 2H, H-3), 3.91 (s, 3H, H-12), 3.99-4.44 (m, 7H, H-11a,
H-13, H-17, H-18 & H-31), 5.64 (d, 1H, J ) 9.5 Hz, H-11), 6.75
(s, 1H, H-6), 6.84-7.70 (m, 9H, H-9, H-20-23 & H-26-29);
13C NMR δ 23.0 (C-2), 28.7 (C-1), 44.7 (C-15), 46.5 (C-3), 46.6
(C-18), 49.9 (C-14), 56.3 (C-12), 60.1 (C-11a), 68.5 (C-17), 70.0
(C-13), 86.0 (C-11), 111.2 (C-9), 114.8 (C-6), 118.4 (C-5a), 120.0,
125.0, 125.4, 127.0, 127.1, 127.7 & 127.9 (C-20-23 & C-26-
29), 141.2 (C-9a), 143.2 (C-24 & C-25), 143.5 (C-19 & C-30),
143.7 (C-7), 149.9 (C-8), 154.1 (C-16), 166.9 (C-5); MS m/z
(relative intensity) 307 (8), 279 (19), 251 (8), 202 (4), 167 (3),
149 (100), 94 (3), 71 (5), 57 (8).
(11a S)-8-(2,3-E p oxyp r op oxy)-7-m et h oxy-1,2,3,11a -t et -
r a h yd r o-5H-p yr r olo[2,1-c][1,4]ben zod ia zep in -5-on e (6). A
solution of Fmoc-protected derivative 19 (70 mg, 0.13 mmol)
and Bu4N+ F- (1 M solution in THF; 289 mg, 0.32 mmol, 2.5
equiv) in dry DMF (5 mL) was stirred for 15 min at room
temperature. The reaction mixture was then quenched with
MeOH (5 mL) and evaporated in vacuo, after which the residue
was dissolved in EtOAc (25 mL) and extracted with water (3
× 25 mL) and the organic layer discarded. Brine (20 mL) was
added to the aqueous extract, which was then extracted with
CHCl3 (3 × 25 mL). The combined organic phase was dried
(MgSO4) and the solvent evaporated in vacuo. Gradient
chromatography (EtOAc, followed by 100:0f94:6 v/v CHCl3-
MeOH) provided the title epoxide PBD 6 as a pale yellow oil
1
O), 1614 (CdC), 1500, 1462, 1377, 1257, 1217, 1018, 722; H
NMR δ 1.92-2.36 (m, 4H, H-1 & H-2), 3.45-3.92 (m, 3H, H-3
& H-11a), 3.93 (s, 3H, H-12), 4.66-4.76 (m, 2H, H-13), 5.33-
5.52 (m, 2H, H-15), 6.03-6.18 (m, 1H, H-14), 6.56 (s, 1H, H-9),
7.51 (s, 1H, H-6), 9.70 (bs, 1H, H-10); 13C NMR δ 29.7 (C-2),
32.1 (C-1), 46.7 (C-3), 56.3 (C-12), 56.4 (C-11a), 69.8 (C-13),
105.5 (C-9), 108.0 (C-6), 113.5 (C-5a), 119.0 (C-15), 132.0 (C-
14), 144.4 (C-9a), 148.9 (C-7), 153.8 (C-8), 158.5 (C-11), 160.3
(C-5); MS m/z (relative intensity) 302 ([M]+, 64), 272 (100),
257 (7), 231 (47), 217 (17), 203 (51), 192 (9), 156 (33), 139 (25),
111 (9), 91 (5), 70 (32). HRMS: calcd for C16H18N2O4 (302.1267),
found 302.1227.
(2S)-N-(2-[9-Flu or en ylm eth oxycar bon yl]am in o-5-m eth -
oxy-4-[2-p r op en oxy]b en zoyl)p yr r olid in e-2-ca r b oxa ld e-
h yd e Dieth yl Th ioa ceta l (17). 9-Fluorenylmethyl chlorofor-
mate (584 mg, 2.26 mmol, 2.4 equiv) in dioxane (2.4 mL) was
added dropwise to a stirred solution of 14b (387 mg, 0.94
mmol) and aqueous Na2CO3 (2.5 mL of 10% w/v) in dioxane
(1.2 mL) at 0 °C. Stirring was continued for 4 h at 0 °C then
for 24 h at room temperature, after which water (2 mL) was
added and the solvent removed in vacuo. Gradient chroma-
tography (20:80f50:50 v/v EtOAc-hexane) of the residue
afforded the Fmoc-protected thioacetal 17 as a yellow oil (523
mg, 88%): IR (neat) v 3432 (NH), 2926, 2855, 1730 (CdO),
1598 (CdC), 1522, 1450, 1403, 1333, 1263, 1203, 1117, 1029,
914, 869, 757; 1H NMR δ 1.18-1.43 (m, 6H, H-34 & H-36),
1.68-2.27 (m, 4H, H-9 & H-10), 2.62-2.79 (m, 4H, H-33 &
H-35), 3.44-3.65 (m, 2H, H-8), 3.85 (s, 3H, H-12), 4.26-4.77
(m, 7H, H-11, H-13, H-18, H-19 & H-32), 5.29-5.48 (m, 2H,

4038 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 20
Wilson et al.
(31 mg, 79%): IR (neat) ν 3040, 2924, 2854, 1628 (CdO), 1512
(CdC), 1452, 1270, 1215, 1015, 870, 813, 780, 755; 1H NMR δ
2.02-2.11 (m, 2H, H-2), 2.28-2.37 (m, 2H, H-1), 2.75-2.80
(m, 1H, H-15), 2.90-2.94 (m, 1H, H-15), 3.40-3.43 (m, 1H,
H-14), 3.53-3.87 (m, 3H, H-3 & H-11a), 3.95 (s, 3H, -OCH3),
3.99-4.22 & 4.25-4.38 (m, 2H, H-13), 6.85 (s, 1H, H-9), 7.53
(s, 1H, H-6), 7.67 (d, 1H, J ) 4.4 Hz, H-11); 13C NMR δ 24.1 &
24.2 (C-2), 29.6 & 29.7 (C-1), 44.9 (C-15), 46.7 (C-3), 49.8 &
49.9 (C-14), 53.7 (C-11a), 56.1 & 56.2 (C-12), 69.6 & 70.1 (C-
13), 111.0 & 111.1 (C-6), 111.8 (C-9), 121.0 (arom.), 140.5
(arom.), 147.8 (arom.), 150.2 (arom.), 162.6 (C-11), 164.5 (C-
5); MS m/z (relative intensity) 302 ([M]+, 100), 273 (5), 259
(4), 245 (16), 217 (3), 70 (9), 57 (8). HRMS: calcd for
C16H18N2O4 (302.1266), found 302.1209.
24.1 (C-9), 24.7 (C-10), 44.6 (C-15), 48.9 (C-8), 49.7 (C-14), 56.5
(C-12), 56.6 & 57.7 (C-17 & C-18), 59.0 (C-11), 70.4 (C-13),
104.8 (C-16), 109.0 (C-3), 109.4 (C-6), 128.8 (C-1), 137.3 (C-2),
147.9 (C-5), 154.6 (C-4), 166.7 (C-7); MS m/z (relative intensity)
396 ([M]+, 1), 380 (2), 365 (5), 321 (5), 296 (4), 267 (6), 252
(49), 236 (4), 222 (4), 207 (6), 149 (6), 122 (6), 75 (100), 57 (8).
HRMS: calcd for C18H24O8N2 (396.1533), found 396.1541.
(2S)-N-(2-Am in o-5-m eth oxy-4-p r op oxyben zoyl)p yr r o-
lid in e-2-ca r boxa ld eh yd e Dim eth yl Aceta l (22). Meth od
A: A mixture containing nitro acetal 20a (350 mg, 0.92 mmol),
10% Pd-C (20 mg), and MeOH (20 mL) was hydrogenated for
16 h at room temperature and 1 atm pressure until TLC (99:
1, v/v CHCl3-MeOH) indicated that reaction was complete.
The catalyst was removed by filtration through Celite, and the
filtrate evaporated in vacuo to afford amino acetal 22 as a
yellow oil (311 mg, 96%).
Meth od B: A mixture of nitro acetal 20a (300 mg, 0.79
mmol) and 10% Pd-BaCO3 (60 mg) in MeOH (10 mL) was
hydrogenated for 24 h at room temperature and 1 atm pressure
until TLC (99:1, v/v CHCl3-MeOH) indicated that reduction
was complete. The catalyst was removed by filtration through
Celite, and the solvent evaporated in vacuo to give 22 as a
yellow oil (272 mg, 92%).
Meth od C: A mixture of nitro acetal 20a (400 mg, 1.05
mmol), 1,4-cyclohexadiene (8.83 g, 0.11 mol), and 10% Pd-C
(20 mg) in MeOH (40 mL) was stirred for 18 h at room
temperature until TLC (99:1 v/v CHCl3-MeOH) indicated that
reduction was complete. The catalyst was removed by filtration
through Celite, and the solvent evaporated in vacuo to furnish
propyl ether 22 as a yellow oil (332 mg, 90%): IR (neat) v 3520
(NH2), 3477 (OH), 1632 (CdO), 1576 (CdC), 1428, 1277, 1219,
1182, 1062, 980, 865, 755; 1H NMR δ 1.17 (t, 3H, J ) 6.8 Hz,
H-15), 1.79-2.27 (m, 6H, H-9, H-10 & H-14), 3.51-3.64 (m,
10H, H-8, H-13, H-18 & H-19), 3.81 (s, 3H, H-12), 4.56-4.61
(m, 1H, H-11), 4.66 (d, 1H, J ) 2.7 Hz, H-17), 6.32 (s, 1H, H-3),
6.78 (s, 1H, H-6); 13C NMR δ 10.1 (C-15), 24.1 (C-14), 24.4 (C-
9), 24.8 (C-10), 50.2 (C-8), 56.7 & 57.7 (C-18 & C-19), 57.0 (C-
12), 59.1 (C-11), 75.5 (C-13), 102.1 (C-3), 105.2 (C-17), 110.3
(C-1), 112.7 (C-6), 141.2 (C-2), 146.8 (C-4), 147.9 (C-5), 169.7
(C-7); MS m/z (relative intensity) 352 ([M]+, 1), 321 (36), 290
(21), 277 (92), 234 (8), 208 (100), 203 (4), 165 (31), 144 (7), 121
(48), 82 (28), 75 (89). HRMS: calcd for C18H28O5N2 (352.1998),
found 352.1967.
(2S)-N-(2-Am in o-4-[2-h yd r oxyp r op oxy]-5-m eth oxyben -
zoyl)p yr r olid in e-2-ca r boxa ld eh yd e Dim eth yl Aceta l (23).
Meth od A: Epoxide acetal 21a (40 mg, 0.10 mmol) was
dissolved in dry MeOH (10 mL) and 10% Pd-C (20 mg) was
added. The mixture was hydrogenated for 8 h at room
temperature and 1 atm pressure, after which TLC (97:3 v/v
CHCl3-MeOH) indicated that the reaction was complete. The
catalyst was removed by filtration through Celite, and the
solvent evaporated in vacuo to give the 2-hydroxypropyl ether
23 as a yellow oil (38 mg, 99%).
(2S)-N-(5-Met h oxy-2-n it r o-4-[2-p r op en oxy]b en zoyl)-
p yr r olid in e-2-ca r boxa ld eh yd e Dim eth yl Aceta l (20a ).
Meth od A: DMF (1 drop) was added to a stirred solution of
benzoic acid 12 (88 mg, 0.35 mmol) and oxalyl chloride (53
mg, 0.42 mmol, 1.2 equiv) in dry THF (10 mL), and stirring
was continued for 3 h at room temperature. Volatiles were
removed in vacuo, and the resulting oil was dissolved in dry
THF (5 mL) and then added dropwise over 5 min to a stirred
suspension of the pyrrolidine dimethyl acetal 13b (50 mg, 0.34
mmol), triethylamine (71 mg, 0.70 mmol, 2 equiv), and water
(1 drop) at 0 °C. The reaction mixture was warmed to room
temperature and then stirred for an additional 16 h. After
concentration in vacuo, the residue was purified by gradient
chromatography (100:0f95:5 v/v CHCl3-MeOH) to give the
coupled nitro acetal 20a as a yellow oil (113 mg, 85%).
Meth od B: 1,3-Dicyclohexylcarbodiimide (705 mg, 3.42
mmol) was added to a solution of 12 (866 mg, 3.42 mmol) in
dry CH2Cl2 (50 mL) and the mixture stirred for 4 h at room
temperature under a N2 atmosphere. A solution of 13b (497
mg, 3.42 mmol) in dry CH2Cl2 (20 mL) was then added, and
stirring was continued for 16 h. The reaction mixture was then
filtered and the filtrate evaporated in vacuo. Gradient chro-
matography (50:50f100:0 v/v EtOAc-hexane) afforded nitro
acetal 20a as a pale yellow oil (407 mg, 31%): IR (neat) v 2939,
1635 (CdO), 1579 (CdC), 1522 (NO2), 1427, 1336 (NO2), 1276,
1218, 1180, 1060, 981, 864, 790, 757; 1H NMR δ 1.74-2.22
(m, 4H, H-9 & H-10), 3.11-3.27 (m, 2H, H-8), 3.57 & 3.59 (s,
6H, H-17 & H-18), 3.98 (s, 3H, H-12), 4.41-4.46 (m, 1H, H-11),
4.69-4.71 (m, 2H, H-13), 4.95 (d, 1H, J ) 2.4 Hz, H-16), 5.36-
5.51 (m, 2H, H-15), 6.01- -6.15 (m, 1H, H-14), 6.76 (s, 1H, H-6),
7.71 (s, 1H, H-3); 13C NMR (CDCl3) δ 24.1 (C-9), 24.7 (C-10),
48.9 (C-8), 56.7 (C-12), 56.8 & 57.8 (C-17 & C-18), 59.0 (C-11),
70.2 (C-13), 104.8 (C-16), 108.7 (C-3), 109.2 (C-6), 119.4 (C-
15), 128.3 (C-1), 131.6 (C-14), 137.3 (C-2), 147.9 (C-5), 154.6
(C-4), 166.5 (C-7); MS m/z (relative intensity) 380 ([M]+, 2),
349 (29), 305 (89), 273 (25), 236 (100), 220 (17), 191 (49), 162
(30), 121 (63), 112 (45), 89 (45), 75 (99). HRMS: calcd for
C18H24O7N2 (380.1584), found 380.1525.
(2S)-N-(4-[2,3-E p oxyp r op oxy]-5-m et h oxy-2-n it r ob en -
zoyl)pyr r olidin e-2-car boxaldeh yde Dim eth yl Acetal (21a).
A solution of 3-chloroperoxybenzoic acid (50-60% w/w, 64 mg,
0.18 mmol, 1.4 equiv), nitro acetal 20a (50 mg, 0.13 mmol),
and 3-tert-butyl-4-hydroxy-5-methylphenyl sulfide (1 mg, 3
µmol, 0.02 equiv) in Cl(CH2)2Cl (10 mL) was refluxed at 85 °C
for 2 h, after which TLC (75:25 v/v EtOAc-hexane) indicated
that reaction was complete. The mixture was neutralized with
an aqueous solution of NaHCO3 (92 mg, 1.10 mmol in 5 mL).
The organic layer was separated then washed with water (4
× 20 mL) and dried (MgSO4) and the solvent removed in vacuo
to give a yellow oil. Gradient chromatography (50:50f100:0
v/v EtOAc-hexane) gave epoxide acetal 21a as a yellow oil
(44 mg, 85%): IR (neat) v 3050, 2924, 2854, 1721 & 1632 (Cd
O), 1579 (CdC), 1522 (NO2), 1462, 1377 (NO2), 1276, 1220,
1185, 1060, 980, 862, 791, 756, 722; 1H NMR δ 1.69-2.24 (m,
4H, H-9 & H-10), 2.81 (dd, 1H, J ) 2.5 & 4.7 Hz, H-15), 2.96
(dd, 1H, J ) 4.0 & 4.5 Hz, H-15), 3.09-3.26 (m, 2H, H-8), 3.41-
3.43 (m, 1H, H-14), 3.56 & 3.59 (s, 6H, H-17 & H-18), 3.97 (s,
3H, H-12), 4.01-4.16 (m, 1H, H-13), 4.2-4.33 (m, 1H, H-11),
4.44 (dd, 1H, J ) 2.8 & 11.4 Hz, H-13), 4.94 (d, J ) 1.7 Hz,
H-16), 6.76 (s, 1H, H-6), 7.77 (s, 1H, H-3); 13C NMR (CDCl3) δ
Meth od B: A mixture containing 21a (60 mg, 0.15 mmol)
and 10% Pd-BaCO3 (40 mg) in MeOH (10 mL) was hydroge-
nated for 16 h at room temperature and 1 atm pressure until
TLC (97:3 v/v CHCl3-MeOH) indicated that reaction was
complete. The catalyst was removed by filtration through
Celite, and the solvent evaporated in vacuo to afford 23 as a
yellow oil (55 mg, 95%).
Meth od C: A mixture of 21a (245 mg, 0.62 mmol), 1,4-
cyclohexadiene (5.16 g, 64.4 mmol), and 10% Pd-C (20 mg)
in MeOH (25 mL) was stirred for 18 h at room temperature
until TLC (97:3 v/v CHCl3-MeOH) indicated that the reaction
was complete. The catalyst was removed by filtration through
Celite, and the solvent evaporated in vacuo to afford the
2-hydroxypropyl ether acetal derivative 23 as a yellow oil (214
mg, 90%).
Meth od D: A mixture of 21a (200 mg, 0.50 mmol) and
SnCl2‚2H2O (226 mg, 1.0 mmol, 2 equiv) in MeOH (50 mL)
was refluxed at 60 °C for 30 min; then the solvent was removed
in vacuo. The residue was purified by gradient chromatography
(100:0f97:3 v/v CHCl3-MeOH) to afford 23 as a yellow oil
(136 mg, 71%): IR (neat) v 3542 (OH), 3480 (NH2), 1726 (Cd

Novel Epoxide-Containing Pyrrolobenzodiazepine
J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 20 4039
O), 1585 (CdC), 1530, 1461, 1376, 1271, 1220, 1053, 976, 865,
720; H NMR δ 1.21 (t, 3H, J ) 6.6 Hz, H-15), 1.69-2.34 (m,
aqueous phase was extracted twice with CHCl3/isoamyl alcohol
(24:1 v/v); then aqueous 4 M NaOAc was added to give a 0.3
M solution, followed by EtOH (3 volumes). After chilling and
centrifugation, the supernatant was removed and the pellet
lyophilized.
1
4H, H-9 & H-10), 3.05-3.32 (m, 2H, H-8), 3.56 & 3.59 (s, 6H,
H-19 & H-20), 3.97 (s, 3H, H-12), 4.02-4.20 (m, 3H, H-13 &
H-14), 4.40-4.45 (m, 1H, H-11), 4.95 (d, 1H, J ) 2.4 Hz, H-18),
6.29 (s, 1H, H-3), 7.30 (s, 1H, H-6); 13C NMR δ 8.5 (C-15), 26.0
(C-9), 26.6 (C-10), 49.8 (C-8), 56.3 (C-12), 56.5 & 56.7 (C-19 &
C-20), 59.4 (C-11), 69.8 (C-14), 72.8 (C-13), 104.7 (C-18), 101.4
(C-3), 113.5 (C-6), 118.7 (C-1), 139.3 (C-2), 147.2 (C-5), 153.5
(C-4), 169.9 (C-7); MS m/z (relative intensity) 368 ([M]+, 2),
337 (21), 310 (15), 306 (35), 293 (76), 236 (100), 224 (96), 209
(4), 166 (13), 121 (45), 82 (17), 70 (9). HRMS: calcd for
5. 5′-En d La belin g. The resultant dephosphorylated DNA
was 5′-labeled with [γ-32P]ATP (5000 Ci/mmol; Amersham plc).
A mixture of the DNA (10 µg in 30 µL water), Forward
Reaction Buffer (9 µL [300 mM Tris-HCl, pH 7.8, 75 mM
2-thioethanol, 50 mM MgCl2, 1.65 µM ATP]), [γ-32P]ATP (3 µL,
30 µCi), and PNK (4 µL, 24 units) was incubated at 37 °C for
30 min. Aqueous NH4OAc (7.5 M, 0.5 volume) and EtOH (2
volumes) were added, and the mixture was chilled and
centrifuged (13 000 rpm/10 min). The supernatant was re-
moved and the pellet resuspended in 20 µL of Tris/EDTA buffer
(10 mM Tris-HCl, pH 7.5, 1 mM EDTA, 0.3 M NaOAc). After
reprecipitation with EtOH (60 µL) and centrifugation, the
supernatant was removed and the pellet washed with EtOH
(20 µL) and lyophilized. The DNA was then resuspended in
sterile double-distilled water at 1 µg/µL. Approximately 10 ng
of labeled DNA was used for each experiment.
6. Rea ction P r otocols. Reactions were performed with
epoxide-PBD 6 at concentrations of 1, 10, 25, 50, 100, 250, 500,
and 1000 µM in 25 mM triethanolamine/1 mM EDTA at pH
7.2 and 37 °C. In separate experiments, reactions were
terminated after either 6 or 24 h by the addition of an equal
volume of stop solution (0.6 M NaOAc/20 mM EDTA/100 µg/
mL tRNA). After precipitation of the DNA by addition of EtOH
(3 volumes) followed by centrifugation and removal of super-
natant, the pellet was dried by lyophilization.
7. Electr op h or esis. Samples were dissolved in strand
separation buffer (10 µL of 30% DMSO/1 mM EDTA/0.04%
bromophenol blue/0.04% xylene cylanol), heated at 90 °C for
2 min, and then chilled immediately in an ice-water bath prior
to loading. Control nondenaturated samples were dissolved in
6% sucrose/0.04% bromophenol blue (10 µL) and loaded
directly. Electrophoresis was performed on a 20-cm horizontal
submerged 0.8% agarose gel at 40 V for 16 h using a 40 mM
Tris/20 mM HOAc/2 mM EDTA gel running buffer.
8. Au tor a d iogr a p h y. Gels were dried at 80 °C between one
layer of Whatman 3 mm and one layer of DE81 filter paper
within a Bio-Rad model 583 gel drier connected to a vacuum.
Autoradiography was performed with Hyperfilm MP (Amer-
sham plc) for 4 h at -70 °C using a DuPont-Cronex lighting-
plus intensifying screen.
9. Ta q DNA P olym er a se Assa y. The Taq DNA polymerase
assay37 was carried out on 6 using pBR322 DNA, T4 poly-
nucleotide kinase (PNK), and the HindIII enzyme (Northum-
bria Biological Ltd.; used as supplied). [γ-32P]ATP (5000 Ci/
mmol) and Amplitaq recombinant Taq DNA polymerase were
obtained from Amersham plc and Perkin-Elmer Cetus, respec-
tively, and used as supplied. Linearization of pBR322 DNA
and PBD treatments were carried out as described above.
10. Ta q P olym er a se Con d ition s. The PBD-bound DNA
(0.5 µg), prepared as above, was combined with Taq DNA
polymerase (1 unit), 32P-5′-end-labeled oligonucleotide primer
(5 pmol, 5′-TATGCGACTCCTGCATTAGG-3′), dideoxynucleo-
side triphosphate mixture (62.5 nM), gelatin (0.01%), (NH4)2-
SO4 (20 µM), and MgCl2 (2.5 µM) in reaction buffer (10 µL,
Tris [pH 9, 75 µM]/Tween 20 [0.01%]) to give a total volume
of 100 µL. Thermal cycling was then carried out as follows:
94 °C for 1 min (denaturation), 58 °C for 1 min (annealing),
and 72 °C for 1 min (primer extension); this cycle was repeated
for a total of 30 cycles, with the annealing step being extended
by 1 s for each new cycle. The products were then chilled on
ice, extracted with CHCl3/isoamyl alcohol (24:1 v/v), precipi-
tated with EtOH (95% v/v, 3 volumes), washed with EtOH
(70% v/v, 100 µL), and dried by lyophilization.
C
18H28O6N2 (368.1947), found 368.1902.
(11a RS)-7-Meth oxy-8-(2-p r op en oxy)-1,2,3,11a -tetr a h y-
d r o-5H-p yr r olo[2,1-c][1,4]ben zod ia zep in -5-on e (15). A so-
lution of nitro acetal 20a (200 mg, 0.53 mmol) and SnCl2‚2H2O
(593 mg, 2.63 mmol, 5 equiv) in MeOH (50 mL) was refluxed
for 4 h. After solvent removal in vacuo the residue was
dissolved in CHCl3 (50 mL), then saturated aqueous NaHCO3
solution (50 mL) added, and the mixture stirred for 18 h at
room temperature. The reaction mixture was then filtered
through Celite and the organic layer separated. The aqueous
layer was extracted with CHCl3 (2 × 50 mL) and the combined
organic phase washed with water (3 × 50 mL), dried (MgSO4),
and evaporated in vacuo. Gradient chromatography (100%
EtOAc, then 100:0f94:6 v/v CHCl3-MeOH) gave imine 15 as
a yellow oil (111 mg, 73%). Analytical data were identical to
those for 15 prepared by the route shown in Scheme 1.
DNA Bin d in g Stu d ies. 1. Th er m a l Den a tu r a tion . Com-
pounds were subjected to thermal denaturation studies with
duplex-form calf thymus DNA (CT-DNA) using an adaptation
of a reported procedure.34 Working solutions in aqueous buffer
(10 mM NaH2PO4/Na2HPO4, 1 mM Na2EDTA, pH 7.00 ( 0.01)
containing CT-DNA (100 µM in phosphate) and the PBD (20
µM) were prepared by addition of concentrated PBD solutions
in MeOH to obtain a fixed [PBD]/[DNA] molar ratio of 1:5.
The DNA-PBD solutions were incubated at 37 °C for 0, 4,
18, or 36 h prior to analysis. Samples were monitored at 260
nm using a Varian-Cary 1E spectrophotometer fitted with a
Peltier heating accessory, and heating was applied at 1 °C
min-1 in the 40-98 °C range. Optical data were imported into
the Origin 5.0 computer package (MicroCal Inc., Northampton,
MA) for analysis, and DNA helix f coil transition tempera-
tures (Tm) were obtained from the maxima in the d(A260)/dT
derivative plots. Results are given as the mean ( SD from
three determinations and are corrected for the effects of MeOH
cosolute using a linear correction term.34b Drug-induced al-
terations in DNA melting behavior are given by: ∆Tm
)
Tm(DNA + PBD) - Tm(DNA alone), where the Tm value for
the PBD-free CT-DNA is 67.83 ( 0.06 °C (from 50 determina-
tions). The fixed [PBD]/[DNA] ratio used did not result in
binding saturation of the host DNA duplex for any compound
examined.
2. DNA Cr oss-Lin k in g Assa y. The agarose gel electro-
phoresis assay36 was carried out with 6 using pBR322 DNA
and T4 polynucleotide kinase (PNK) purchased from Northum-
bria Biological Ltd. and was used as supplied. Bacterial
alkaline phosphatase (BAP) and HindIII enzymes were pur-
chased from BRL and used as supplied.
3. Lin ea r iza tion of p BR322 DNA. A mixture of circular
pBR322 DNA (10 µg in 20 µL), React 2 aqueous buffer (3 µL
[500 mM Tris-HCl, pH 8, 100 mM MgCl2, 500 mM NaCl]),
HindIII (3 µL, 30 units), and water (4 µL) was incubated at
37 °C for 1 h. Aqueous 4 M NaOAc was added to give a 0.3 M
solution, then EtOH (95% v/v, 3 volumes) was added, and the
mixture chilled. After centrifugation and removal of superna-
tant, the pellet was lyophilized.
4. Dep h osp h or yla tion of Lin ea r ized DNA. A mixture of
linearized DNA in water (15 µL), BAP buffer (4 µL [50 mM
Tris-HCl, pH 8, 600 mM NaCl]), and BAP (1 µL, 100 units)
was incubated at 65 °C for 1 h. After extraction with Tris/
EDTA buffer (10 mM Tris-HCl, pH 7.5, 1 mM EDTA)/saturated
phenol (1:1 v/v), the aqueous layer was collected and the
phenolic layer back-extracted with buffer. The combined
11. Electr op h or esis. Products were separated on 0.4-mm
thick 6% denaturing polyacrylamide at 3000 V and 55 °C for
approximately 3 h using a Tris/boric acid/EDTA running
buffer. Autoradiography was carried out as described above.
Cytotoxicity Stu d ies. In vitro cytotoxicity was evaluated
using the human ovarian carcinoma cell lines SKOV-3, A2780,

4040 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 20
Wilson et al.
and CH1, and the cisplatin-resistant counterparts of the A2780
and CH1 lines (A2780cisR and CH1cisR, respectively). Viable
cells were seeded in growth medium (160 µL) into 96-well
microtiter plates and allowed to attach overnight. The PBDs
were dissolved in DMSO (to give a 20 mM concentration in
each case) immediately prior to adding to the cells in quadru-
plicate wells. Final PBD concentrations in the wells ranged
from 100 µM to 2.5 nM as follows: 100, 25, 10, 2.5, 1 µM; 250,
100, 25, 10, 2.5 nM. This was achieved by diluting the PBDs
in growth medium and then adding 40 µL to the existing well
volume of 160 µL to give the final concentrations stated above.
After 4 days (96 h) the medium was removed and the
remaining cells were fixed using 10% trichloroacetic acid on
ice for 30 min. Wells were then washed 3-4 times with tap
water and air-dried overnight, and 100 µL of 0.4% sulfor-
hodamine B (dissolved in 1% glacial HOAc) was added to each
well. Staining was allowed to continue for 10-15 min; the wells
were then washed 3-4 times with 1% acetic acid and air-dried,
and Tris base (100 µL of 10 mM) was added to each one. The
plates were then shaken and absorbance readings taken at
540 nm using a plate reader. From plots of concentration
versus percentage absorbance (compared to 8 untreated wells),
IC50 values were calculated using the Quattro-Pro software
package.
snapshots was carried out to a gradient tolerance change of
0.1 kcal mol-1
.
Angle constraints of -110° (purines) and -115° (pyrim-
idines) were applied to the glycosidic torsions in the B-type
DNA, using a force constant of 500 kcal mol-1 during stages
(ii)-(iv). Planar restraints were not required for the bases,
although atoms C1′ and P of all terminal bases were fixed
during MD stages (iii) and (iv) to prevent the short DNA
duplex from unravelling or “fraying”. Nonbonded interactions
were included up to 11.5 Å; the switching function was
employed between 10.5 and 9.5 Å, and a factor of 0.4 was used
to dampen 1,4-electrostatic interactions. Hydrogen-bonded
contacts were included up to 3.5 Å, with a switching function
used between 4.0 and 4.5 Å. Solvent and counterion effects
were simulated using a distance-dependent dielectric con-
stant42 with ꢀ ) crij, where c ) 4.0 during the MM stages (ii)
and (v) and c ) 1.0 during the MD stages (iii) and (iv). Covalent
interaction energies were calculated using Einteraction ) Ecomplex
- [EDNA + EPBD], where Ecomplex is the energy of the minimized
average structure for the DNA-PBD adduct, and the EDNA and
EPBD component energies were determined from separate
single-point energy calculations on this complex after removal
of either the ligand or DNA and replacement of all required
protons (i.e., guanine C2-NH2 or H11 for the PBD). In the case
where the PBD is covalently bound to the DNA through C11
but the epoxide is in the ring-opened form, the alcohol was
built and used instead of the intact epoxide ring.
Molecu la r Mod elin g Stu d ies. The strategy used for the
molecular modeling was developed from recent studies of
DNA-PBD adduct formation.10b,40 Host DNA duplexes for
energy calculations were selected with embedded 5′-(G/C)A0-1
-
Ack n ow led gm en t. This work was supported in part
by the Cancer Research Campaign UK (Grant SP1938/
0301 to D.E.T., T.C.J ., and J .A.H. and Grant SP1938/
0401 to D.E.T.). J ane Rimington is thanked for her help
with the preparation of this manuscript.
(G/C) target sequences flanked by strand-terminating G‚C base
pairs to prevent strand fraying during MD simulations (Table
2). Additional flanking adenine bases were incorporated, where
appropriate, to provide the favored 5′-AGA motifs for alkyla-
tion by the electrophilic PBD moiety. For each DNA sequence,
simulations were initially carried out with covalent PBD
adducts formed with C11(S)-stereochemistry (see earlier) and
alignment of the aromatic A-ring toward either the 5′- or 3′-
end of the covalently modified strand. Models were examined
with the epoxide in either the bound or unbound states, and
in each possible (i.e., R or S) configuration. This required 8
simulations for the monoadducts and 16 simulations for the
cross-linked adducts.
Refer en ces
(1) Wilson, S. C.; Howard, P. W.; Thurston, D. E. Design and
Synthesis of a Novel Epoxide-Containing Pyrrolo[2,1-c][1,4]-
benzodiazepine (PBD) via a New Cyclization Procedure. Tetra-
hedron Lett. 1995, 36, 6333-6336.
(2) (a) Thurston, D. E. Nucleic Acid Targeting: Therapeutic Strate-
gies for the 21st Century. Br. J . Cancer 1999, 80 (Suppl. 1), 65-
85. (b) Trauger, J . W.; Baird, E. E.; Dervan, P. B. Recognition of
16 Base Pairs in the Minor Groove of DNA by a Pyrrole-
Imidazole Polyamide Dimer. J . Am. Chem. Soc. 1998, 120, 3534-
3535. (c) White, S.; Szewczyk, J . W.; Turner, J . M.; Baird, E. E.;
Dervan, P. B. Recognition of the Four Watson-Crick Base Pairs
in the DNA Minor Groove by Synthetic Ligands. Nature 1998,
391, 468-471. (d) Kielkopf, C. L.; White, S.; Szewczyk, J . W.;
Turner, J . M.; Baird, E. E.; Dervan, P. B.; Rees, D. C. A
Structural Basis for Recognition of A‚T and T‚A Base Pairs in
the Minor Groove of B-DNA. Science 1998, 282, 111-115. (e)
Dervan, P. B. Gene-Specific Transcription Inhibition In Vivo by
Designed Ligands. FASEB J . 1997, 11, 2546. (f) Turner, J . M.;
Baird, E. E.; Dervan, P. B. Recognition of Seven Base Pair
Sequences in the Minor Groove of DNA by Ten-Ring Pyrrole-
Imidazole Polyamide Hairpins. J . Am. Chem. Soc. 1997, 119,
7636-7644. (g) Wemmer, D. E.; Dervan, P. B. Targeting the
Minor Groove of DNA. Curr. Opin. Struct. Biol. 1997, 7, 355-
361. (h) Gottesfeld, J . M.; Neely, L.; Trauger, J . W.; Baird, E.
E.; Dervan, P. B. Regulation of Gene Expression by Small
Molecules. Nature 1997, 387, 202-205. (i) Szewczyk, J . W.;
Baird, E. E.; Dervan, P. B. Sequence-Specific Recognition of DNA
by a Major and Minor Groove Binding Ligand. Angew. Chem.,
Int. Ed. Engl. 1996, 35, 1487-1489.
Using the Quanta 4.1 program (Molecular Simulations Inc.,
San Diego, CA), versions of compound 6 were constructed with
each epoxide configuration based on the crystal coordinates
of anthramycin methyl ether.41 ESP charges were then calcu-
lated using the MNDO Hamiltonian in MOPAC 6, and
coordinates for the DNA duplexes were generated using the
nucleic acid sequence builder module in Quanta 4.1. Guanine
bases used to form covalent bonds with either the PBD or
epoxide moieties were built by deleting the outer-facing
exocyclic H22 proton followed by rationalization of the residual
charge over the remaining atoms.10b A similar procedure was
adopted for the loss of a proton at the C11 position of the bound
PBD molecule. As nucleophilic ring opening of the epoxide
would form a secondary alcohol, this species was also con-
structed.
The following stepwise protocol using the CHARMm 23
program39 was adopted to dock and minimize the ligand into
the minor groove of DNA: (i) the PBD was manually docked
in the minor groove with the C11 atom positioned 1.45 Å from
the N2 of G5 and in the plane of the base, thereby effectively
replacing the outer-facing guanine C2-NH2 proton not involved
in G‚C base pairing. (ii) Powell-type MM minimization for 1000
steps, or to a gradient tolerance change of 0.1 kcal mol-1, was
carried out to remove bad contacts from the initial docking.
(iii) The system was heated from 0 to 300 K for 20 ps (0.2-fs
time step) and the equilibrium assessed by monitoring the
temperature every 200 steps and uniformly reassigning veloci-
ties from the averaged velocities at that time step if more than
(10 K from the target 300 K. (iv) MD was continued at 300 K
for a further 25 ps with a 1-fs time step. The coordinates were
saved after every 100 steps to give a total of 500 coordinate
sets. (v) Final Powell MM minimization of the rms-averaged
(3) (a) Neidle, S.; Puvvada, M. S.; Thurston, D. E. The Relevance of
Drug-DNA Sequence-Specificity to Antitumour Activity. Eur.
J . Cancer 1994, 30A, 567-568. (b) Neidle, S.; Thurston, D. E.
DNA Sequences as Targets for New Anticancer Agents. In New
Targets for Cancer Chemotherapy; Kerr, D. J ., Workman, P.,
Eds.; CRC Press Ltd.: Boca Raton, FL, 1994; pp 159-175.
(4) Leimgruber, W.; Stefanovic, V.; Shenker, F.; Karr, A.; Berger,
J . Isolation and Characterization of Anthramycin,
a New
Antitumor Antibiotic. J . Am. Chem. Soc. 1965, 87, 5791-5793.
(5) Thurston, D. E. Advances in the Study of Pyrrolo[2,1-c][1, 4]-
benzodiazepine (PBD) Antitumour Antibiotics. In Molecular
Aspects of Anticancer Drug-DNA Interactions; Neidle, S., War-
ing, M. J ., Eds.; Macmillan Press: London, 1993; pp 54-88.
(6) (a) Hertzberg, R. P.; Hecht, S. M.; Reynolds, V. L.; Molineux, I.
J .; Hurley, L. H. DNA Sequence Specificity of the Pyrrolo[1,4]-
benzodiazepine Antitumor Antibiotics. Methidiumpropyl-EDTA-
Iron(II) Footprinting Analysis of DNA Binding Sites for Anthra-

Novel Epoxide-Containing Pyrrolobenzodiazepine
J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 20 4041
mycin and Related Drugs. Biochemistry 1986, 25, 1249-1258.
(b) Hurley, L. H.; Reck, T.; Thurston, D. E.; Langley, D. R.;
Holden, K. G.; Hertzberg, R. P.; Hoover, J . R. E.; Gallagher, G.;
Faucette, L. F.; Mong, S.-M.; J ohnson, R. K. Pyrrolo[1,4]-
benzodiazepine Antitumour Antibiotics; Relationship of DNA
Alkylation and Sequence Specificity to the Biological Activity
of Natural and Synthetic Compounds. Chem. Res. Toxicol. 1988,
1, 258-268.
(20) Pullman, A.; Zakrzewska, C.; Perahia, D. Molecular Electrostatic
Potential of the B-DNA Helix. I. Region of the Guanine-Cytosine
Base Pair. Int. J . Quantum Chem. 1979, 16, 395-403.
(21) Cheh, A. M.; Yagi, H.; J erina, D. M. Effect of DNA-Base
Sequence on the Configuration of Deoxyadenosine Adducts
Formed by the Fjord Region Diol Epoxide, (+)-(1R,2S,3R,4S)-
3,4-Dihydroxy-1,2-Epoxy-1,2,3,4-Tetrahydrobenzo[c]phenan-
threne. Biochemistry 1994, 33, 12911-12919.
(22) Leimgruber, W.; Batcho, A. D.; Czajkowski, R. C. Total Synthesis
of Anthramycin. J . Am. Chem. Soc. 1968, 90, 5641-5643.
(23) Thurston, D. E.; Bose, D. S. Synthesis of DNA-Interactive
Pyrrolo[2,1-c][1,4]benzodiazepines. Chem. Rev. 1994, 94, 32-
87.
(24) Langley, D. R.; Thurston, D. E. A Versatile and Efficient
Synthesis of Carbinolamine-Containing Pyrrolo[1,4]benzo-
diazepines via the Cyclization of N-(2-Aminobenzoyl)pyrrolidine-
2-carboxaldehyde Diethylthioacetals: Total Synthesis of Proth-
racarcin. J . Org. Chem. 1987, 52, 91-97.
(25) Tozuka, Z.; Takaya, T. Studies on Tomaymycin. I. The Structure
Determination of Tomaymycin on the Basis of NMR Spectra. J .
Antibiot. 1983, 36, 142-146.
(26) Bose, D. S.; J ones, G. B.; Thurston, D. E. New Approaches to
Pyrrolo[2,1-c][1,4]benzodiazepines: Synthesis, DNA-Binding and
Cytotoxicity of DC-81. Tetrahedron 1992, 48, 751-758.
(27) Kishi, Y.; Aratani, M.; Tanino, H.; Fukuyama, T.; Goto, J . New
Epoxidation with m-Chloroperbenzoic Acid at Elevated Tem-
peratures. J . Chem. Soc. Chem. Commun. 1972, 2, 64-65.
(28) Thurston, D. E.; Langley, D. R. Synthesis and Stereochemistry
of Carbinolamine-Containing Pyrrolo[1,4]benzodiazepines by
Reductive Cyclization of N-(2-Nitrobenzoyl)pyrrolidine-2-car-
boxaldehydes. J . Org. Chem. 1986, 51, 705-712.
(29) Deziel, R. Mild Palladium(0)-Catalyzed Deprotection of Allyl
Esters - A Useful Application In The Synthesis of Carbapenems
and Other â-Lactam Derivatives. Tetrahedron Lett. 1987, 28,
4371-4372.
(30) Fukuyama, T.; Liu, G.; Linton, S. D.; Lin, S. C.; Nishino, H. Total
Synthesis of (+)-Porothramycin-B. Tetrahedron Lett. 1993, 34,
2577-2580.
(31) Carpino, L. A.; Han, G. Y. The 9-Fluorenylmethoxycarbonyl
Amino-Protecting Group. J . Org. Chem. 1972, 37, 3404-3409.
(32) Ueki, M.; Amemiya, M. Removal of 9-Fluorenylmethyloxycar-
bonyl (Fmoc) Group with Tetrabutylammonium Fluoride. Tet-
rahedron Lett. 1987, 28, 6617-6620.
(7) Puvvada, M. S.; Hartley, J . A.; Gibson, I.; Stephenson, P.;
J enkins, T. C.; Thurston, D. E. Inhibition of Bacteriophage T7
RNA Polymerase In Vitro Transcription by the DNA-Binding
Pyrrolo[2,1-c][1,4]benzodiazepine (PBD) Antitumour Antibiotics.
Biochemistry 1997, 36, 2478-2484.
(8) Puvvada, M. S.; Hartley, J . A.; J enkins, T. C.; Thurston, D. E.
A Quantitative Assay to Measure the Relative DNA-Binding
Affinity of Pyrrolo[2,1-c][1,4]benzodiazepine (PBD) Antitumour
Antibiotics Based on the Inhibition of Restriction Endonuclease
BamHI. Nucleic Acids Res. 1993, 21, 3671-3675.
(9) (a) Farmer, J . D.; Rudnicki, S. M.; Suggs, J . W. Synthesis and
DNA Cross-linking Ability of a Dimeric Anthramycin Analogue
Tetrahedron Lett. 1988, 29, 5105-5108. (b) Farmer, J . D.;
Gustafson, G. R.; Conti, A.; Zimmt, M. B.; Suggs, J . W. DNA
Binding Properties of a New Class of Linked Anthramycin
Analogues. Nucleic Acids Res. 1991, 19, 899-903.
(10) (a) Bose, D. S.; Thompson, A. S.; Ching, J .; Hartley, J . A.;
Berardini, M. D.; J enkins, T. C.; Neidle, S.; Hurley, L. H.;
Thurston, D. E. Rational Design of a Highly Efficient Irreversible
DNA Interstrand Cross-Linking Agent based on the Pyrroloben-
zodiazepine Ring System. J . Am. Chem. Soc. 1992, 114, 4939-
4941. (b) J enkins, T. C.; Hurley, L. H.; Neidle, S.; Thurston, D.
E. Structure of a Covalent DNA Minor Groove Adduct with a
Pyrrolobenzodiazepine Dimer: Evidence for Sequence-Specific
Interstrand Cross-Linking. J . Med. Chem. 1994, 37, 4529-4537.
(11) (a) Bose, D. S.; Thompson, A. S.; Smellie, M.; Berardini, M. D.;
Hartley, J . A.; J enkins, T. C.; Neidle, S.; Thurston, D. E. Effect
of Linker Length on DNA-Binding Affinity, Cross-Linking
Efficiency and Cytotoxicity of C8-Linked Pyrrolobenzodiazepine
Dimers. J . Chem. Soc. Chem. Commun. 1992, 14, 1518-1520.
(b) Thurston, D. E.; Bose, D. S.; Thompson, A. S.; Howard, P.
W.; Leoni, A.; Croker, S. J .; J enkins, T. C.; Neidle, S.; Hartley,
J . A.; Hurley, L. H. Synthesis of Sequence-Selective C8-Linked
Pyrrolo[2,1-c][1,4]benzodiazepine DNA Interstrand Cross-Link-
ing Agents. J . Org. Chem. 1996, 61, 8141-8147. (c) Gregson, S.
J .; Howard, P. W.; J enkins, T. C.; Kelland, L. R.; Thurston, D.
E. Synthesis of a Novel C2/C2′-exo Unsaturated Pyrrolobenzo-
diazepine Cross-linking Agent with Remarkable DNA Binding
Affinity and Cytotoxicity. J . Chem. Soc. Chem. Commun. 1999,
9, 797-798.
(33) Bonini, C.; Righi, G. Regioselective- and Chemoselective Syn-
thesis of Halohydrins by Cleavage of Oxiranes with Metal-
Halides. Synthesis 1994, 3, 225-238.
(12) Confalone, P. N.; Huie, E. M.; Ko, S. S.; Cole, G. M. Design and
Synthesis of Potential DNA Cross-Linking Reagents Based on
the Anthramycin Class of Minor Groove Binding Compounds.
J . Org. Chem. 1988, 53, 482-487.
(34) (a) J ones, G. B.; Davey, C. L.; J enkins, T. C.; Kamal, A.; Kneale,
G. G.; Neidle, S.; Webster, G. D.; Thurston, D. E., The Non-
Covalent Interaction of Pyrrolo[2,1-c][1,4]benzodiazepine-5, 11-
diones with DNA. Anti-Cancer Drug Des. 1990, 5, 249-264. (b)
McConnaughie, A. W.; J enkins, T. C. Novel Acridine-Triazenes
as Prototype Combilexins: Synthesis, DNA-Binding, and Bio-
logical Activity. J . Med. Chem. 1995, 38, 3488-3501.
(35) Sengupta, S. K.; Blondin, J .; Szabo, J . Covalent Binding of
Isomeric 7-(2,3-Epoxypropoxy)actinomycin D to DNA. J . Med.
Chem. 1984, 27, 1465-1470.
(36) Hartley, J . A.; Berardini, M. D.; Souhami, R. L. An Agarose Gel
Method for the Determination of DNA Interstrand Cross-Linking
Applicable to the Measurement of the Rate of Total and “Second-
Arm” Cross-link Reactions. Anal. Biochem. 1991, 193, 131-134.
(37) Ponti, M.; Forrow, S. M.; Souhami, R. L.; D′Incalci, M.; Hartley,
J . A. Measurement of the Sequence Specificity of Covalent DNA
Modification by Antineoplastic Agents using Taq DNA Poly-
merase. Nucleic Acids Res. 1991, 19, 2929-2933.
(38) Pullman, B. Molecular Mechanisms of Specificity in DNA-
Antitumour Drug Interactions. Adv. Drug Res. 1989, 18, 1-113.
(39) Brooks, B. R.; Bruccoleri, R. E.; Olafson, B. D.; States, D. J .;
Swaminathan, S.; Karplus, M.; CHARMM: A Program for
Macromolecular Energy Minimization and Dynamics Calcula-
tions. J . Comput. Chem. 1983, 4, 187-217.
(40) Adams, L. J .; Morris, S. J .; Banting, L.; J enkins, T. C.; Thurston,
D. E. Molecular Mechanics Study of the Stereochemistry of
Formation of Covalent Pyrrolobenzodiazepine-DNA Adducts.
Pharm. Sci. 1995, 1, 151-154.
(13) (a) Thurston, D. E.; Hurley, L. H. A Rational Basis for the
Development of Antitumour Agents in the Pyrrolo[1,4]-
benzodiazepine Group. Drugs Future 1983, 8, 957-971. (b)
Hurley, L. H.; Thurston, D. E. Pyrrolo[1,4]benzodiazepine An-
titumour Antibiotics: Chemistry, Interaction with DNA and
Biological Implications. Pharm. Res. 1984, 2, 51-59. (c) Thur-
ston, D. E.; Bose, D. S.; Howard, P. W.; J enkins, T. C.; Leoni,
A.; Baraldi, P. G.; Guiotto, A.; Cacciari, B.; Kelland, L. R.;
Foloppe, M.-P.; Rault, S. Effect of A-Ring Modifications on the
DNA-Binding Behaviour and Cytotoxicity of Pyrrolo[2,1-c][1,4]-
benzodiazepines. J . Med. Chem. 1999, 42, 1951-1964.
(14) Warpehoski, M. A.; Hurley, L. H. Sequence Selectivity of DNA
Covalent Modification. Chem. Res. Toxicol. 1988, 1, 315-333.
(15) Armstrong, R. W.; Salvati, M. E.; Nguyen, M. Novel Interstrand
Cross-Links Induced by the Antitumour Antibiotic Carzinophilin
Azinomycin-B. J . Am. Chem. Soc. 1992, 114, 3144-3145.
(16) Macleod, M. C.; Evans, F. E.; Lay, J .; Chiarelli, P.; Geactinov,
N. E.; Powell, K. L.; Daylong, A.; Luna, E.; Harvey, R. G.
Identification of a Novel N7-Deoxyguanosine Adduct as the
Major DNA Adduct Formed by a Non-Bay-Region Diol Epoxide
of Benzo[a]pyrene with Low Mutagenic Potential. Biochemistry
1994, 33, 2977-2987.
(17) Hemminki, A.; Vayrynen, T.; Hemminki, K. Reaction-Kinetics
of Alkyl Epoxides with DNA and Other Nucleophiles. Chem.-
Biol. Interact. 1994, 93, 51-58.
(18) Cochrane, J . E.; Skopek, T. R. Mutagenicity of Butadiene and
its Epoxide Metabolites. 1. Mutagenic Potential of 1,2-Epoxy-
butene, 1,2,3,4-Diepoxybutane and 3,4-Epoxy-1,2-butanediol in
Cultured Human Lymphoblasts. Carcinogenesis 1994, 15, 713-
717.
(19) De Los Santos, C.; Cosman, M.; Hingerty, B. E.; Ibanez, V.;
Margulis, L. A.; Geacintov, N. E.; Broyde, S.; Patel, D. J .
Influence of Benzo[a]pyrene Diol Epoxide Chirality on Solution
Conformations of DNA Covalent Adducts - The (-)-trans-anti-
[BP]G‚C Adduct Structure and Comparison with the (+)-trans-
anti-[BP]G‚C Enantiomer. Biochemistry 1992, 31, 5245-5252.
(41) Arora, S. K. Structural Investigations of Mode of Action of Drugs.
2. Molecular Structure of Anthramycin Methyl Ether Monohy-
drate. Acta Crystallogr. 1979, B35, 2945-2948.
(42) Orozco, M.; Laughton, C. A.; Herzyk, P.; Neidle, S. Molecular
Mechanics Modelling of Drug-DNA Structures: The Effects of
Differing Dielectric Treatment on Helix Parameters and Com-
parison with a Fully Solvated Structural Model. J . Biomol.
Struct. Dyn. 1990, 8, 359-373.
J M981124D