Structure-activity relationships of monomeric C2-aryl pyrrolo[2,1- c ][1,4]benzodiazepine (PBD) antitumor agents

Article abstract of DOI:10.1021/jm901722v

A comprehensive SAR investigation of the C2-position of pyrrolo[2,1-c][1,4]benzodiazepine (PBD) monomer antitumor agents is reported, establishing the molecular requirements for optimal in vitro cytotoxicity and DNA-binding affinity. Both carbocyclic and

Full text of DOI:10.1021/jm901722v

J. Med. Chem. 2010, 53, 2927–2941 2927
DOI: 10.1021/jm901722v
Structure-Activity Relationships of Monomeric C2-Aryl Pyrrolo[2,1-c][1,4]benzodiazepine (PBD)
Antitumor Agents
Dyeison Antonow,†,‡ Maciej Kaliszczak,§ Gyoung-Dong Kang,‡ Marissa Coffils,‡ Arnaud C. Tiberghien,‡
Nectaroula Cooper,† Teresa Barata,† Sibylle Heidelberger,† Colin H. James,† Mire Zloh,† Terence C. Jenkins,
Anthony P. Reszka,^ Stephen Neidle,^ Sylvie M. Guichard,§ Duncan I. Jodrell,§,3 John A. Hartley,# Philip W. Howard,*,†,‡ and
David E. Thurston*,†,‡
†Cancer Research UK Gene Targeted Drug Design Research Group, The School of Pharmacy, University of London, 29/39 Brunswick Square,
London WC1N 1AX, U.K., ‡Spirogen Ltd, 29/39 Brunswick Square, London WC1N 1AX, U.K., §Cancer Research UK Pharmacology and
Drug Development Group, University of Edinburgh Cancer Research Centre, Crewe Road, Edinburgh EH4 2XR, U.K., Morvus Technology
Limited, Tyˆ Myddfai, Llanarthne, Carmarthen SA32 8HZ, U.K., ^Cancer Research UK Biomolecular Structure Research Group, The School of
Pharmacy, University of London, 29/39 Brunswick Square, London WC1N 1AX, U.K., and #Cancer Research UK Drug-DNA Interactions
Research Group, UCL Cancer Institute, 72 Huntley Street, London WC1E 6BT, U.K. 3 Current address: Cancer Research UK Cambridge
Research Institute, University of Cambridge.
Received November 21, 2009
A comprehensive SAR investigation of the C2-position of pyrrolo[2,1-c][1,4]benzodiazepine (PBD) mono-
mer antitumor agents is reported, establishing the molecular requirements for optimal in vitro cytotoxicity
and DNA-binding affinity. Both carbocyclic and heterocyclic C2-aryl substituents have been studied
ranging from single aryl rings to fused ring systems, and also styryl substituents, establishing across a library
of 80 analogues that C2-aryl and styryl substituents significantly enhance both DNA-binding affinity and in
vitro cytotoxicity, with a correlation between the two. The optimal C2-grouping for both DNA-binding
affinity and cytotoxicity was found to be the C2-quinolinyl moiety which, according to molecular modeling,
is due to the overall fit of the molecule in the DNA minor groove, and potential specific contacts with
functional groups in the floor and walls of the groove. This analogue (14l) was shown to delay tumor growth
in a HCT-116 (bowel) human tumor xenograft model.
Introduction
alkylating agents, and both the chemical and biological aspects
of these compounds have been explored by several research
groups since the discovery of anthramycin in the 1960s.3-20
Structure-activity relationship (SAR) information for the
PBDs was initially established based on biological data
obtained from natural compounds.21-23 The development
of an understanding of their mode of action prompted the
rational design of synthetic analogues which provided addi-
tional chemical and biological information.1,2 The recent
growth of interest in PBDs is partly due to the progress of
SJG-136, which is the first synthetic PBD-based compound to
enter phase II clinical trials with success.24-27 Joining two
PBD monomeric units together to make PBD dimers such as
SJG-136 results in greater in vitro cytotoxicity and in vivo
antitumor efficacy due to the ability of these molecules to
cross-link DNA.25,28 However, PBD dimer syntheses are
generally lengthier than those for PBD monomers, and this
has limited the scope for structural diversification. Since the
success of SJG-136, there has been interest in progressing a
PBD monomer to clinical trials due to the potential advan-
tages including lower molecular weight, ease of synthesis, and
greater potential for structural diversification. In addition,
compared to PBD dimers, the monomers are poor substrates
for ABC transporters29 which can lead to multidrug resis-
tance in a clinical setting. For example, SJG-136 is a substrate
for the ABC transporter P-glycoprotein (ABCB1/P-gp).30,31
Also, other preclinical data32 indicate that, as a class, C2-
modified PBD monomers have in vivo antitumor activity in
Naturally occurring pyrrolo[2,1-c][1,4]benzodiazepine
(PBD)a compounds are isolated from the fermentation broth
of Streptomyces species and are known for their antibiotic and
antitumor properties.1,2 They can exist as N10-C11 carbino-
lamine methyl ethers (e.g., anthramycin, 1), carbinolamines
(e.g., sibiromycin, 2), or imines (e.g., the synthetic C2-aryl
PBD-imines, 3, 4, and 5), although these different forms can in-
terconvert at the N10-C11 position depending on solvent and
isolation conditions (Figure 1). However, in all cases, the C11-
position is electrophilic and enables the molecules to alkylate
the NH2 group of a guanine in the minor groove of DNA.1 This
reaction is sequence-selective and preferentially targets 50-Pu-
G-Pu sequences. The PBDs have a right-handed twist due to
the “S”-configuration at their C11a-position, which allows
them to fit perfectly within the DNA minor groove. These
properties differentiate PBDs from other families of DNA-
*To whom correspondence should be addressed. Phone: þ44 (0) 207
753 5932. Fax: þ44 (0) 207 753 5965. E-mail: philip.howard@
spirogen.com (P.W.H.); david.thurston@pharmacy.ac.uk (D.E.T.).
a Abbreviations: BAIB, [bis(acetoxy)iodo]benzene; CT-DNA, calf
thymus DNA; ΔTm, enhancement in thermal denaturation temperature
of calf-thymus DNA; EDCI, 1-ethyl-3-(3-dimethylaminopropyl) carbo-
diimide hydrochloride salt; HCT-116, human colorectal tumor cell line;
NaHMDS, sodium bis(trimethylsilyl)amide; 2D NOESY two-dimen-
sional nuclear Overhauser enhancement spectroscopy; PBD, pyrrolo-
[2,1-c][1,4]benzodiazepine; Pu, purine nucleotides; SAR, structure-
activity relationship; TEMPO, 2,2,6,6-tetramethylpiperidine-1-oxyl;
TLC, thin layer chromatography.
r
2010 American Chemical Society
Published on Web 03/10/2010
pubs.acs.org/jmc

2928 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 7
Antonow et al.
mouse models, thus providing the basis for translational
studies toward a phase I clinical trial.
procedure to obtain C1/C2-endo-unsaturated C2-aryl PBD
imines.36
In a previous study of synthetic PBD monomers, we found
that introduction of an aryl group at the 2-position significantly
enhanced in vitro cytotoxicity.33 The role played by the addi-
tional aryl fragment was not fully understood at the time and
prompted the detailed study reported here. In related work, we
established34 that a series of C2-aryl PBD dilactams devoid of
A-ring substituents and lacking a DNA-alkylating N10-C11
imine moiety had enhanced DNA-binding affinity compared to
the parent non-C2-substituted PBD dilactam, elevating the
melting temperature of double-stranded calf thymus DNA by
up to 2.3 ꢀC after 18 h incubation and thus confirming the
importance of the C2-substituent in stabilizing DNA interac-
tion. In a follow-up study we reported the synthesis of a novel
C2-aryl PBD-imine library in which a parallel approach was
employed to obtain 66 structurally unique C2-aryl-substituted
PBDs containing A-ring electron-donating substituents and
N10-C11 imine functionalities.35 We also reported a synthetic
We now report biological (in vitro cytotoxicity and xenograft)
and biophysical (DNA thermal denaturation) data for a library
of 80 C2-aryl-substituted PBD-imine compounds synthesized
via three different routes. The library includes a number of novel
C2-heteroaryl and C2-styryl analogues, and some C1/C2-endo
unsaturated C2-aryl PBD analogues (Tables 1 and 3). A
synthetic route to a novel C-ring saturated C2-aryl PBD (i.e.,
27, Scheme 2) is also described, thus providing an important
“control” molecule along with further insights into the SAR of
PBD monomers (Figure 2). Molecular modeling studies have
been used to examine the DNA-binding of key members of the
series. Furthermore, the water solubility of some analogues has
been dramatically improved by converting to C11-(sodium
bisulfite) adducts (i.e., 6, Figure 1). Finally, one library member
(14l) has been evaluated in a HCT-116 human tumor xenograft
model and shown to have in vivo antitumor activity.
Results and Discussion
Chemistry. Synthesis of C2/C3-endo-Unsaturated C2-
Aryl and C2-Styryl PBD Imines (Scheme 1, and Tables 1
and 3). The C2-aryl PBDs were originally synthesized ac-
cording to the reductive cyclization approach developed by
Leimgruber and co-workers.37 We initially reported33 appli-
cation of the Suzuki reaction to introduce aryl substituents at
the same position employing aryl boronic acids and C2-enol
triflate PBDs. Subsequently, we reported the use of Stille
coupling to install vinyl and heteroaryl substituents at the
C2-position of PBDs38 [The Supporting Information docu-
ment contains full synthetic methodologies and character-
ization data for the Stille and Suzuki coupling products
previously reported only in communications]. More recently,
we reported the parallel synthesis of 66 C2-aryl-modified
PBDs via palladium catalyzed cross-coupling.35 This parallel
approach allowed the rapid synthesis of a set of N10-C11
imine-containing C2-substituted PBDs with stereochemical
integrity at their C11a-position. The same methodology was
used in this study (Scheme 1) for the synthesis of the novel
PBDs included in the sublibraries contained in 8-15
(Table 1)35 and the C2-styryl sublibrary 16a-e (Table 3).
Synthesis of C1/C2-endo-Unsaturated and Fully-Saturated
C-Ring C2-Substituted PBD Imines (Scheme 2, Table 1). The
effect of the location of unsaturation (i.e., C1-C2 or
Figure 1. Structures of the naturally occurring PBD monomers
anthramycin (1), sibiromycin (2), and the novel synthetic C2-sub-
stitued PBD analogues (4-6).
Figure 2. Summary of SARs for the C2-aryl functionalization of PBD monomers based on in vitro cytotoxicity, DNA thermal denaturation,
and NMR data.

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 7 2929
Table 1. Cytotoxicity Data for the C2-Aryl PBD Librarya

2930 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 7
Table 1. Continued
Antonow et al.
a NCI 60 cell-line panel. Figures correspond to the mean values (MG-MID) calculated from the 60 individual cell lines (see Supporting Information).
b K562 human chronic myeloid leukemia cell line (96 h incubation, see Experimental Section).
C2-C3) in the PBD C-ring on biological potency has been
unclear to date. Therefore, we prepared examples of both for
evaluation. We found that a C1/C2-enol-triflate analogue of
7 could be obtained under kinetic conditions which allowed
further elaboration to C2-aryl PBDs (e.g., 29, Table 1)
via Suzuki coupling.36 To fully understand the impact of

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 7 2931
Table 3. Cytotoxicity Data for the C2-Styryl Analogues
Table 2. Selected Thermal Denaturation (ΔTm) Values for Representa-
tive PBD Imine Library Members
Compd
Induced ΔTm (ꢀC)a
no.
C2-substituent
0 h
4 h
18 h
12bb
9b
(phenyl) 2,6-CH3
(phenyl) 4-NH2
(phenyl) 2-CH3
(phenyl) 4-CH(CH3)2
(phenyl) 4-N(CH3)2
(phenyl) 4-Ph
0.5c
5.8
0.5 c
6.2
0.5c
6.5
11a
9j
6.2
6.5
6.8
6.7
6.9
6.9
9a
6.9
8.0
7.6
8.1
8.0
8.4
9l
9g
(phenyl) 4-C(CH3)3
(phenyl) 4-CHCH2
(phenyl) 4-CH2CH3
(phenyl) 3-OCH3
(phenyl) 4-Cl
8.7
8.5
9.0
9.0
9.2
9.2
9k
9i
9.0
9.8
9.4
9.7
10b
9o
10.1
9.9
10.2
10.2
10.3
10.1
10.6
10.6
10.7
10.8
10.8
11.0
11.9
12.1
13.0
15.7
18.2
20.5
13.0
16.3
9.6
9.9
14i
9h
(phenyl) 3,4-OCH2O
(phenyl) 4-CH3
(phenyl) 4-OCH3
(phenyl) 4-F
10.3
9.7
9.4
9e
9n
10.2
10.1
10.4
10.4
10.5
10.7
11.4
11.5
12.0
14.8
17.7
19.1
9.4
10.5
10.4
10.5
10.6
10.6
10.9
11.7
11.9
12.7
15.4
18.1
19.9
11.2
16.0
9m
9q
(phenyl) 4-SCH3
(phenyl) 4-CHO
(phenyl) 4-H
8
9s
(phenyl) 4-CN
(phenyl) 4-CF3
(phenyl) 4-OPh
2-thiophenyl
9r
9f
14a
13a
14k
14l
a NCI 60 cell-line panel. Figures correspond to the mean values (MG-
MID) calculated from the 60 individual cell lines (see Supporting
Information). b K562 human chronic myeloid leukemia cell line (96 h
incubation, see Experimental Section).
2-naphthyl
3-quinolinyl
6-quinolinyl
anthramycin methyl ether (1)47
sibiromycin (2)47
15.7
bisulfite) adducts (28a-f, Scheme 3) were prepared to improve
the water solubility of the C2-aryl PBDs. PBD-bisulfite adducts
were originally reported by Kaneko and co-workers40 and were
shown to have greater chemical stability than their parent
N10-C11 imines. Later, Langlois and co-workers reported the
synthesis of an anthramycin-related analogue containing a C11-
bisulfite moiety.41 However, in neither case was the stereochem-
istry of the C11-SO3-Naþ linkage assigned.
a See Experimental Section for details. b This compound was synthe-
sized according to ref 33. c Values for this compound were measured in a
different laboratory using 10 mM NaH2PO4/Na2HPO4 and 1 mM
Na4EDTA as buffer, rather than the 10 mM NaH2PO4/Na2HPO4 and
1 mM Na2EDTA used for studies with all other compounds. The higher
[Naþ] and increased ionic strength for this buffer is reflected in a higher
Tm value for the native CT-DNA (i.e., Tm (DNA) = 70.63 ꢀC rather
than 67.82 ꢀC). Further, 20 µM 12b and higher 200 µM DNA (in DNAp,
equivalent to 100 µM in base pairs [bp]) was used. This 1:10 [ligand]/
[DNAp] molar ratio compares with the 1:5 ratio used for all other
compounds. Using this high-[Naþ] buffer and a 1:10 molar ratio,
anthramycin methyl ether (1) gave corresponding ΔTm shifts of 13.70,
14.17 and 14.08 ꢀC, at 0, 4 and 18 h, respectively.
We initially utilized Langlois’ method but with modified
reaction conditions in terms of time and solvent composi-
tion, which proved to be efficient for the synthesis of C2-aryl
PBD C11-sodium bisulfite adducts. The straightforward
workup, involving simple extraction into deionized water,
gave the adducts as solids following lyophilization. Because
the stereochemistry of C11-bisulfite adducts had not been
studied by either Kaneko or Langlois, we carried out a
comprehensive NMR investigation in both DMSO-d6 and
D2O solutions. Using compound 28d as an example, we
found that the C11S-(sodium bisulfite) adduct is the major
species and that its formation is affected by solvent composi-
tion and reaction time [NOTE: The 1H NMR in DMSO-d6
showed a doublet (J = 10.4 Hz) at 3.96 ppm corresponding
to the H11-proton, the strong coupling indicating that this
proton was positioned in an “anti” orientation relative to
H11a.1 The C11R-(sodium bisulfite) addition product was
not observed by NMR under these conditions (see Support-
unsaturation in the C-ring, we also synthesized the novel
fully saturated C2-aryl PBD-imine 27 (Scheme 2). The
synthesis started from the C-ring building block 17,39 which
was oxidized to the C2-ketone (18) using TEMPO and BAIB.
The enol-triflate 19 was obtained in excellent yield (91%) via
kinetic enolisation of 18 using NaHMDS. Suzuki coupling
was performed next on the C-ring building block before
attachment to the nitrobenzoyl fragment (A-ring precursor).
This strategy allowed the use of hydrogenation for selective
reduction of the endo-unsaturation in the pyrrolidine ring.
The cross-coupled product 20 was exposed to H2 in
the presence of catalytic amounts of palladium on charcoal
to simultaneously deprotect the cyclic amine and reduce
the endo-unsaturation in stereospecific fashion. Next, the
C-ring building block 21 was coupled to 2-nitroveratric acid,
and the product further elaborated to the N10-Alloc-pro-
tected PBD 26 using standard synthetic methods. Final
deprotection with Pd(PPh3)4 afforded the novel C2-
substituted C-ring saturated PBD-imine 27 which was char-
acterized by 2D-NOESY NMR (see Supporting In-
formation).
1
ing Information for H- and 2D-NMR spectra)]. Further
NMR analysis including COSY and NOESY experiments
indicated the presence of the N10-C11 hydrolysis product
(i.e., B-ring opened aldehyde) (Figure 3) [NOTE: There is a
significant change in the chemical environment of the H1
protons upon B-ring closure. As a consequence, H1R and
H1β become less shielded in the PBD ring system]. Interest-
ingly, the ratio of products changed when a mixture of
2-propanol-d8/D2O (2:1 v/v) was used as solvent.8 In this
Synthesis and C11-Stereochemistry of C2-Aryl PBD C11-(So-
dium Bisulfite) Adducts (Scheme 3, Table 4). C11-(Sodium

2932 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 7
Antonow et al.
Scheme 1. C2/C3-Endo-Unsaturated C2-Aryl PBDs Obtained via the Parallel Approach.a
a Reagents and conditions: (i) PS-PPh3Pd (0.01 equiv), boronic acid 1.1 equiv), TEA (6.0 equiv), 100 ꢀC (microwave radiation), 10 min; then 10% Cd/
Pb (8.0 equiv), 1 M NH4OAc/THF (2:3), rt, 0.5 to 2.5 h, 4 to 67%. See ref 35 for full analytical data and experimental details for 8-15. Analytical data for
novel compounds 14k, 14l, and 16 are provided in Supporting Information.
case, formation of the C11S-bisulfite was maximized, and
only a negligible amount of B-ring-opened product was
observed. 1H NMR time-course experiments were also car-
ried out using the biphasic system D2O/CD2Cl2 (1:1 v/v) as
solvent (see Supporting Information). The results suggested
that, under the acidic conditions (pH 5) of the aqueous
sodium bisulfite layer, the PBD-imine initially dissolved in
the organic phase and then became protonated to form an
iminium species which reacted with excess water to give the
N10-C11 carbinolamine as the predominant species. Car-
binolamines of this type have been reported to exist in
equilibrium with their equivalent ring-opened amino-alde-
hyde forms.42 This aldehyde species can in turn react with
bisulfite anion to give the B-ring-opened bisulfite adduct,
although this product was only observed at early reaction
times (i.e., 1-5 min, see Supporting Information) and pre-
sumably rapidly interconverts back to the cyclized C11S-
bisulfite product according to the equilibrium shown in
Figure 3. The presence of these interconverting forms was
further supported by “trapping” the PBD-imine form in the
organic layer. This was carried out by shaking a D2O
solution of an imine-free lyophilized sample of the isolated
PBD-bisulfite adduct 28d with CD2Cl2, followed by phase
separation and 1H NMR analysis, which clearly showed the
presence of the PBD-imine species (i.e., 1H NMR spectrum
identical to 10b).
Biological and Biophysical Data. In Vitro Cytotoxicity. All
library members were evaluated for cytotoxicity across the
NCI 60 cell-line panel (see Supporting Information) and the
K562 leukemia cell line. The NCI screen was repeated after a
period of 2-3 months with the majority of compounds
retaining their cytotoxicity properties, thus providing an
indication of their chemical stability. The most noticeable
observation from the NCI data was the selective cytotoxicity
of the C2-aryl PBD imines towards melanoma cell lines.
Compounds with C1/C2-endo unsaturation (e.g., 29) and or
a fully saturated C-ring (i.e., 27) were either significantly less
cytotoxic than the equivalent C2/C3-endo-unsaturated mo-
lecules or completely inactive (Table 1). Molecular modeling
DNA Thermal Denaturation Studies. Reactivity of the C2-
aryl PBD monomers toward double-stranded DNA was
assessed by measuring their induced effects on the melting
behavior (Tm) of duplex-form calf thymus DNA.44,45 Ther-
mal denaturation data for selected members of the PBD
imine library are shown in Table 2. C2-Substituted PBD
imines were found to shift the DNA melting curve signifi-
cantly (stabilization) compared to PBD dilactams,34,44 high-
lighting the importance of the N10-C11 imine functionality,
the 7- and 8-OCH3 A-ring substituents, and the C2-aryl
functionalities on DNA-binding affinity. A number of trends
are apparent from Table 2 for simple C2-phenyl substituted
compounds. First, the addition of two ortho-methyl substit-
uents to the simple C2-phenyl group (12b) reduces DNA sta-
bilization significantly (ΔTm = 0.5 ꢀC). On the other hand,
electron-withdrawing groups appear to be more favorable
then electron-donators in enhancing binding affinity in
the para-substituted analogues (See Table S1, Supporting
Information). For example, 9r (4-CF3, ΔTm=11.9 ꢀC) and
9s (4-CN, ΔTm =11.0 ꢀC) provide greater DNA stabiliza-
tion than 9h (4-CH3, ΔTm=10.1 ꢀC) or 9e (4-OCH3, ΔTm=
10.6 ꢀC). However, fused carbocyclic and heterocyclic
six-membered ring systems provide the greatest degree of
stabilization, with 13a (2-naphthyl), 14k (3-quinolinyl), and
14l (6-quinolinyl), stabilizing the double helix by 15.7, 18.2,
and 20.5 ꢀC, respectively. These high values are remarkable
in that they are in the range of ΔTm values associated with
PBD dimer molecules such as DSB-120 (ΔTm=15.1 ꢀC, un-
der identical conditions),46 which covalently cross-link the
two DNA strands in an interstrand manner.
Structure-Activity Relationships (SARs). The importance
of C2-aryl-substituents and a sp2 carbon at the C2-position
can be seen in naturally occurring PBDs with the C2-
functionalizations of anthramycin (1) and sibiromycin (2),
two of the most potent PBD natural products known. It is
also evident from the range of known natural products
that these changes to the C-ring can lead to greater
cytotoxicity enhancement than A-ring modifications.47 With
C2-sp2 compared to C2-sp3 functionalization, the mole-
cules adopt a flatter overall structure and a conforma-
tion more favorable for interacting in the DNA minor
groove through maximal intermolecular hydrogen bond-
ing, electrostatic, and van der Waals interactions. We
have previously synthesized38,48,49 C2/C3-endo-unsaturated
PBD-imines with average in vitro potencies higher than
anthramycin.
analysis involving
a conformational search (Macro-
Model, version 9.6) indicated that 27 and 29 are unlikely to
bind effectively to double-stranded DNA (a model 14-mer
DNA duplex43 was studied; data not shown) due to their
overall shape and lack of isohelicity with the DNA minor
groove. In addition, it was observed that 29 is prone to
undergo C-ring aromatization36 which may also reduce its
overall cytotoxicity. In vitro data generated for the parallel-
synthesized C2/C3-endo-unsaturated C2-aryl and C2-styryl
PBD imines, and the PBD C11-(sodium bisulfite) adducts
are collected in Tables 1, 3, and 4).
The high potency associated with many of the compounds in
this study was “off scale” in the NCI’s high-throughput cyto-
toxicity assay. Hence, we focused our SAR analysis mainly on
the IC50 values obtained from the K562 cell line. Cytotoxicity

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 7 2933
Scheme 2. Fully Saturated C2-Aryl PBD (27) Obtained via Suzuki Coupling between the Pyrrolidinic Enol-triflate 19 and
4-Methoxyphenylboronic Acida
a Reagents and conditions: (i) TEMPO (0.1 equiv), BAIB (1.1 equiv), CH2Cl2, rt, 2 h, 48%; (ii) NaHMDS (1.1 equiv), THF, -78 ꢀC, 1.5 h, then N,N-
bis(trifluoromethanesulfonyl)anilide (1.1 equiv), 91%; (iii) 4-methoxyphenyl boronic acid (1.0 equiv), Pd(PPh3)4 (0.01 equiv), aq Na2CO3, toluene, rt, 36
h, 45%; (iv) 10% Pd/C, EtOH, H2, rt 12 h, 99%; (v) 4,5-dimethoxy-2-nitrobenzoic acid, EDCI, 1.2 equiv), HOBt (1.2 equiv), CH2Cl2, 0 ꢀC, 2 h, 74%; (vi)
10% Pd/C, EtOH, H2, rt 12 h, 72%; (vii) Alloc-Cl (1.5 equiv), pyridine (2.2 equiv), CH2Cl2, rt, 6 h, 82%; (viii) TBAF (1.2 equiv), THF, rt 3 h, 96%; (ix)
TEMPO (0.1 equiv), BAIB (1.1 equiv), CH2Cl2, rt, 5 h, 65%; (x) Pd(PPh3)4 (0.01 equiv), pyrrolidine (1.1 equiv), CH2Cl2, rt, 30 min, 58%.
Scheme 3. Synthesis of C2-Aryl PBD C11-(Sodium Bisulfite) Adductsa
a Reagents and conditions: (i) CH2Cl2/water (1:1 v/v), aq NaHSO3 (1.1 equiv), rt, 5 to 80 min; (ii) 2-propanol/water (2:1 v/v), aq NaHSO3 (1.1 equiv),
rt, 5 min (See Supporting Information). b Yields refer to reaction conditions (i).
values for the C2-aryl and styryl PBDs (Tables 1 and 3)
exhibited a wide range from <0.1 to 920 nM. In general,
compounds with the highest ΔTm values had the lowest IC50
values, however this correlation was not perfect due to the
and 10d-e) which may reflect the electronegativity or bulki-
ness of the substituents. However, introduction of long
aliphatic moieties protruding from the para or meta positions
(e.g., 9x-z and 10l and 10m) substantially reduced cytotoxi-
city. This may be due to compacting or “folding” of these
chains in the minor groove leading to a reduced fit. Com-
pounds with cyclic C2-substituents such as 4-(4-methyl)-
piperazinyl (9a0) and 4-morpholinyl (9b0) are exceptions with
very low IC50 values. One possible explanation is that
entropic costs are minimized with cyclic C2-substituents.
With the C2-pyridinyl compounds 14d-e, the position of
the nitrogen in the ring does not appear to significantly affect
cytotoxicity. Likewise, exchange of the heteroatom between
2-benzofuranyl (14g) and 2-benzothiophenyl (14h) analo-
gues does not enhance activity. However, the presence of
a N-methyl group in the 4-pyrazolyl conjugates (14b-c)
wide structural diversity of the compounds (r2=0.33, Pvalue
=
0.0014). Several physicochemical properties were investi-
gated using Chem3D 11.0 (CambridgeSoft) in an attempt
to identify parameters that might better correlate with
cytotoxicity. Interestingly, we found a statistically significant
correlation between solvent accessible surface area (SASA)
and cytotoxicity (r2=0.81, Pvalue<0.0001).
C2-Aryl PBD Imines. There is no significant difference
between the IC50 values of compounds in the C2-phenyl
series (Table 1) for different substituents in the ortho, meta,
and para positions, although a clear trend is noticeable
within the para- and meta-halogenated series (i.e., 9n-o

2934 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 7
Antonow et al.
improves cytotoxicity. Finally, carboxylic acid groups as in
9t and 10i completely deplete cytotoxicity, possibly due to
repulsion between the negative charges encountered in the
DNA and the substituents in these specific PBDs.
Insertion of a naphthyl substituent at the C2-position
dramatically increases both the DNA binding affinity and
cytotoxicity of PBD monomers, as can be seen from the ΔTm
and IC50 values for the C2-(2-naphthyl)-substituted com-
pound 13a (IC50=0.5724 μM, ΔTm=15.8 ꢀC). On the basis
of the decreased cytotoxicity of the C2-(1-naphthyl) analo-
gue 13d, it would appear that the best position for joining
fused ring systems to the C2-position of a PBD is furthest
from the bridge carbons (i.e., the 2-position using the naph-
thalene ring numbering system). In support of this, NMR
studies have revealed that for the C2-(2-naphthyl) system,
the 2-naphthyl ring lies along the minor groove extending the
isohelical shape created by the A-, B-, and C-rings of the
PBD skeleton.43 For the C2-(1-naphthyl) configuration, the
naphthyl ring can take up two possible conformations, one in
which it clashes with the floor of the minor groove or another
in which it points out of the groove, neither of which are
conducive to a good overall fit of the PBD molecule. This
concept is also supported by the decreased cytotoxicity of
the 4-dibenzothiophenyl (15a), 4-dibenzofuranyl (15b) and
1-thianthrenyl (15c) analogues).
Table 4. Cytotoxicity Data for the C2-Aryl PBD C11-(Sodium
Bisulfite) Adducts
The C2-heteroaryl sublibrary contained the most active
compounds, with the C2-quinolinyl-substituted analogues
14k and 14l providing greater DNA stabilization than the
corresponding C2-(2-naphthyl) analogue 13a (Table 2). The
C2-(6-quinolinyl) PBD (14l) has the highest ΔTm value of all
the compounds examined (ΔTm = 20.5 ꢀC), which is also
reflected in its high cytotoxicity (IC50 =1.4 nM in K562).
Modeling studies with four 1:1 (ligand/DNA) complexes
with the 9-mer DNA duplex sequence d(50-CGCA-
GACGC-30).d(50-GCGTCTGCG-30) (Figure 4A) provided
estimates of differences in relative stabilization energies.
These were calculated relative to the value for the 9-mer
DNA adduct with compound 8:
8-ð9-mer DNAÞ : 0:0 kcal mol-1
12b-ð9-mer DNAÞ : þ 18:2 kcal mol-1
13a-ð9-mer DNAÞ : -0:4 kcal mol-1
14l-ð9-mer DNAÞ : -2:2 kcal mol-1
These values are in good qualitative agreement with the
rank order of ΔTm values and show that the C2-quinolynyl
substituent in 14l is the best in energetic terms. The out-
standingly higher energy of the dimethyl adduct (12b) is a
consequence of the reduction in nonbonded interactions
between the DNA minor groove walls and the phenyl group
as a result of steric clashes involving the methyl groups.
Figure 4A shows that the C2-substituent is forced out of the
groove as a consequence. It is also notable that compound
a NCI 60 cell-line panel. Figures correspond to the mean values (MG-
MID) calculated from the 60 individual cell lines (see Supporting
Information). b K562 human chronic myeloid leukemia cell line (96 h
incubation, see Experimental Section).
Figure 3. Possible interconvertible forms present in an aqueous solution of 28d based on NMR. (Note: Structures in brackets are proposed
based on sequential addition/elimination mechanisms and have not been observed by NMR.)

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 7 2935
Figure 4. (A) Overlay of the final minimized simulated structures of the covalent complexes between the duplex DNA sequence
d(50-CGCAGACGC-30).d(50-GCGTCTGCG-30) and compounds 12b and 14l. The 2,6-dimethylphenyl group in the former is colored magenta
and the quinolinyl ring in the latter is cyan. (B) Snapshot of a model of part of the (14l)2-DNA adduct [d(50-AATCTTTAAAGATT-30)2]
with the C2-quinolinyl substituent of 14l in the “inward conformation”. Conformational search (MacroModel, version 9.6) predicted indirect
H2O-bridged H-bonding interactions between the C2-quinolinyl nitrogen atom and backbone atoms of the noncovalently modified DNA
strand. Direct H-bonding interactions between the nitrogen atom and DNA bases were not predicted from these studies.
12b has correspondingly poor cytotoxicity compared to
other compounds in the series, confirming that biological
potency is related to minor groove binding.
duplex upon adduct formation and may explain the higher
ΔTm value of this compound compared to 13a.
C2-Styryl PBD Imines. To further explore the SAR of C2-
aryl PBDs, we designed a set of C2-styryl analogues (Table 3)
in which the sp2 C2-carbon is distanced further from the aryl
substituent by a conjugated double bond. The cytotoxicity
data for 16a-e show that this arrangement still enhances
potency, not only compared to C2-aryl analogues but also to
the C2-vinyl compound 32. The unsubstituted styryl deriva-
tive 16a and the 4-fluoro-substituted 16c were highly cyto-
toxic in both the NCI and K562 cell lines. The (4-CH3)-
substituted styryl 16b gave an intermediate ΔTm value of
14.0 ꢀC, which ranks between the average ΔTm for the C2-
phenyl PBDs (∼10.0 ꢀC) and the ∼20.0 ꢀC observed for the
C2-quinolinyl PBDs.
C2-Aryl PBD C11-(Sodium Bisulfite) Adducts. All C11-
bisulfite library members (28a-f) are freely soluble in water
as opposed to their N10-C11 imine precursors. However,
with the exception of the C2-naphthyl 28e (more active) and
C2-quinolinyl 28f (equipotent) adducts (Table 4), all the
remaining C11-(sodium bisulfite) adducts were less cytotoxic
in both the NCI panel and the K562 leukemia cell line in
comparison to the equivalent parent PBD N10-C11 imines
from which they derive. To verify that DNA interaction was
occurring in a similar manner to the parent N10-C11
imines, MS-MALDI experiments (see Supporting In-
formation) were carried out to confirm covalent interaction
and identical stoichiometry after reaction of both the PBD
imine 13a and its C11-bisulfite derivative (28e) with a 14-mer
DNA duplex. This provides further evidence that PBD C11-
bisulfites are acting as “prodrugs” in solution to generate a
low but steady concentration of the equivalent PBD imine
species, thus leading to a slower reaction with DNA com-
pared to the parent PBD imines.8 As the imine species react
with DNA, albeit slowly, the equilibrium presumably adjusts
until eventually all C11-bisulfite is consumed. This also
explains the higher IC50 values compared to their PBD imine
counterparts.
The improved energy of 14l relative to the 2-naphthyl
derivative (13a) is due, at least in part, to the polarization
induced by the ring nitrogen atom which enhances the elec-
trostatic contribution to the overall energy. Another possible
explanation relates to the hydrogen bonding potential of the
ring nitrogen in 14l. We initially assumed that the nitrogen
atom in the quinolinyl ring may interact with DNA bases via
hydrogen bonding interactions (H-bonds). However, initial
molecular modeling experiments failed to reveal any direct H-
bonds between 14l and DNA, although it was established that
the quinolinyl ring of 14l could adopt two different conforma-
tions in the minor groove both potentially conducive to
bonding. In one case, it is aligned along the minor groove
(the “inward conformation”, see Figure 4B) and in the other
the ring is protruding slightly from the groove (the “outward
conformation”).43 Preliminary energy minimization experi-
ments using implicit solvent were carried out with both
conformations and 16 different 8-mer DNA duplex sequences
(i.e., d(50-TAXXGATT-30).d(50-AATCYYTA-30) where X =
G, A, T or C) in order to account for all possible permutations
for the base pairs in close proximity to the naphthyl and
quinolinyl rings. The preliminary results showed that, as
anticipated, the total energies for the two different types of
conformations across all 16 adducts had an average difference
of 24.5 kcal mol-1 in favor of the inward conformation. This
resultsupportsrecent NMR studies ona (13a)2-DNA adduct
which demonstrated that the C2-naphthyl ring of 13a lies
along the hydrophobic floor of the minor groove.43 However,
no discrete H-bonds were observed in any of the 16 models,
and so we next performed conformational searches on a
selected 14-mer (14l)2-DNA adduct [d(50-AATCTTTAAA-
GATT-30)2]43 in the inward conformation in the presence of
water molecules using MacroModel (version 9.6) (Figure 4B).
The results ofthis computational analysis suggested that14l in
the “inward conformation” can interact with DNA through
H-bonds involving the ring nitrogen of the C2-quinolinyl
substituent, water molecules, and the backbone atoms of the
noncovalently modified DNA strand (Figure 4B). These
additional interactions could further stabilize the DNA
In Vivo Studies. Evaluation of 14l in a HCT-116 Human
Tumour Xenograft Model. On the basis of the unusually high
DNA-binding affinity of the C2-(6-quinolinyl) PBD mono-
mer 14l, it was progressed into an in vivo human tumor

2936 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 7
Antonow et al.
Figure 5. Left: 14l inhibits the growth of HCT-116 tumors in nude mice. Randomized HCT-116 tumor-bearing mice were injected iv with
either a single bolus injection of 300 μg/kg or daily 120 μg/kg injections for 5 days. The change in tumor volume at each time point was
determined for each animal (as tumor volume relative to tumor volume at the beginning of treatment - day 1). The line graphs show the mean
relative tumor volumes for the animals in each treatment group. Asterisks depict the statistical significance of the differences between the
control group and each of the 14l treatment groups (using Student’s t-test). Right: Animal weights are provided as percentage change relative to
animal weights at day 1. Error bars in both panels represent standard error.
xenograft mouse model inoculated with the HCT-116 cell
line, a widely used xenograft model. Both irinotecan and
capecitabine have been reported to induce significant growth
delays and regression in this model.50,51 Two schedules of
administration were evaluated: a single iv injection of 300
μg/kg, and five consecutive daily iv injections of 120 μg/kg
(1 cycle from day 1 to day 5). The results are statistically
significant (Student’s t-test) and demonstrate not only the
ability of 14l to delay tumor growth in this model but also the
prolonged benefit of treatment (Pvalue <0.01) (Figure 5, left).
Crucially, at these doses and different schedules of administra-
tion, there was no apparent toxicity as judged by a lack of
significant adverse effects or reduction of animal weights
throughout the treatment (Figure 5, right).
Overall, this study has provided a substantial body of SAR
data (Figure 2) relating to C2-substituted PBD monomers,
demonstrating that some substituents (e.g., C2-quinolinyl) can
enhance both cytotoxicity and DNA-binding affinity to a simi-
lar level of potency as that observed for PBD dimers which can
form covalent interstrand cross-links between DNA strands.
Finally, the ability of 14l to delay tumor growth at doses not
associated with significant weight loss or other adverse effects
in the HCT-116 colon cancer xenograft model suggests that
C2-aryl PBD monomers may have potential in the treatment of
human disease.
Experimental Section
Chemistry. General Synthetic Methods. Optical rotations
were measured on an ADP 220 polarimeter (Bellingham Stanley
Ltd.) and concentrations (c) are given in g/100 mL. Melting points
were measured using a digital melting point apparatus (Electro-
thermal). IR spectra were recorded on a Perkin-Elmer Spectrum
1000 FT IR spectrometer. 1H and13C NMR spectra were acquired
at 300 K using a Bruker Avance NMR spectrometer at 400 and
100 MHz, respectively, unless stated otherwise. Chemical shifts are
reported relative to TMS (δ = 0.0 ppm), and signals are desig-
nated as s (singlet), br s (broad singlet), d (doublet), t (triplet), q
(quadruplet), qu (quintuplet), h (heptuplet), dt (double triplet), dd
(doublet of doublets), ddd (double doublet of doublets), or m
(multiplet), with coupling constants given in Hertz (Hz). A pro-
PBD numbering system has been used for carbon and proton
assignments for synthetic intermediates (i.e., based on the final
tricyclic PBD ring system). Mass spectroscopy data were collected
using a Waters Micromass ZQ instrument coupled to a Waters
2695 HPLC with a Waters 2996 PDA. Waters Micromass ZQ
parameters used were: capillary, 3.38 kV; cone, 35 V; extractor, 3.0
V; source temperature, 100 ꢀC; desolvation temperature, 200 ꢀC;
cone flow rate, 50 L/h; desolvation flow rate, 250 L/h. High-
resolution mass spectroscopy data were recorded on a Waters
Micromass QTOF Global in positive W-mode using metal-coated
borosilicate glass tips to introduce samples into the instrument.
Thin-layer chromatography (TLC) was performed on silica gel
aluminum plates (Merck 60, F254), and flash chromatography
utilized silica gel (Merck 60, 230-400 mesh ASTM). All chemicals
and anhydrous solvents were purchased from Sigma-Aldrich.
Compounds 8-16 (Tables 1 and 3) were synthesized according
to previously described methodology.35 The novel PBD-imine C2-
quinolinyl (14k-l) and C2-styryl (16a-e) compounds reported
here were synthesized as part of this study and purified accord-
ingly. All compounds used in the biophysical and biological
evaluations were >95% pure as determined by previously de-
scribed HPLC methodology.35 Analytical data for these com-
pounds are collected in Supporting Information.
Conclusions
Novel synthetic methodologies have been applied to advance
understanding of the SAR profile of PBD monomers. An
80-member library containing a wide variety of substituents
at the C2-position of the PBD skeleton has been successfully
prepared using three different approaches involving the palla-
dium-catalyzed cross-coupling of enol-triflate substrates. The
cytotoxicity data demonstrated a wide range of activities across
the library. In the simple C2-phenyl-substituted series, activity
does not appear to depend significantly on the position of
substituents in the benzene ring, although there is limited
evidence that electron-withdrawing substituents are favorable
for optimal activity. However, there is clear evidence that fused
aromatic ring systems (e.g., 2-naphthyl) markedly improve
both DNA binding and cytotoxicity, although for optimal
activity the point of attachment of the fused ring system to
the C2-position of the PBD must not be immediately adjacent
to the bridge atoms of the fused ring system to avoid clashes
with the walls of the host DNA minor groove. Moreover,
insertion of heteroatoms into a C2-naphthyl moiety can sub-
stantially enhance both cytotoxicity and DNA-binding affinity
(i.e., the 3- and 6-quinolinyl analogues). Interestingly, C2-styryl
compounds were found to be significantly more cytotoxic
in the K562 cell line compared to other C2-substituted sub-
library members. Finally, we have shown that formation of
C11S-(sodium bisulfite) adducts dramatically improves the
water solubility of C2-aryl PBD-imines, although surprisingly
they are less potent than their PBD-imine counterparts. This
may be due to these adducts acting as “prodrugs”, with slow
hydrolysis resulting in release of the DNA-reactive N10-C11
imine species.8

Article
Synthesis of the Fully Saturated C-Ring C2-Aryl PBD Imine
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 7 2937
(C10 and Ar-CBz), 125.9-128.5 (C20, C60, and Ar-CBz),
120.5 (C1), 114.0 (C30 and C50), 67.1 (CH2-CBz), 66.3 (C11a),
64.9 (C11), 54.6 (C3), 55.3 (40-OCH3), 25.8 (C(CH3)3), 18.2
(Cquat), -5.4 (SiCH3)2).
(27). (S)-Benzyl 2-((tert-Butyldimethylsilyloxy)methyl)-4-oxo-
pyrrolidine-1-carboxylate (18). TEMPO (0.1 g, 0.68 mmol, 0.1
equiv) and BAIB (2.43 g, 7.53 mmol, 1.1 equiv) were added to a
solution of Cbz-protected amine alcohol 1739 (2.5 g, 6.84 mmol,
1.0 equiv) in CH2Cl2 (50 mL), and the mixture stirred for 2 h
when TLC (EtOAc/hexane 1:1) indicated complete consump-
tion of starting material. The reaction mixture was diluted with
CH2Cl2 (100 mL) and washed with saturated aqueous Na2S2O3
(60 mL). The aqueous layer was extracted with CH2Cl2 (2 ꢀ 50
mL), and the combined organic layer was washed with brine (50
mL) and dried over MgSO4. Removal of excess solvent under
reduced pressure afforded a solid that was purified by flash
column chromatography (EtOAc/hexane 3:7 v/v) to afford 18
(1.2 g, 48%). 1H NMR (CDCl3) rotamers: δ -0.02-0.06 (m, 6
H, Si(CH3)2), 0.83 (s, 9 H, C(CH3)3), 2.45 (d, 1 H, J = 17.9 Hz,
H1), 2.67 (m, 1 H, H1), 3.56 (d, 1 H, J = 10.3 Hz, H3), 3.74 (d, 1
H, J = 18.4 Hz, H11), 3.72-4.07 (m, 2 H, H3 and H11),
4.39-4.48 (m, 1 H, H11a), 5.18 (s, 2 H, CH2-CBz), 7.33-
7.39 (m, 5 H, Ar-CBz). 13C NMR (CDCl3): δ 209.9 (C2), 154.4
(NCdO), 136.4 (Ar-CBz), 127.9-128.5 (Ar-CBz), 67.3 (C11),
64.9 (CH2-CBz), 55.8 (C11a), 53.7 (C3), 40.1 (C1), 25.7
(C(CH3)3), 18.0 (Cquat), -5.8 (SiCH3)2).
(2S,4R)-2-((tert-Butyldimethylsilyloxy)methyl)-4-(4-methoxy-
phenyl)pyrrolidine (21). A catalytic amount of 10% palladium
on carbon (50 mg) was added to a solution of Cbz-protected
compound 20 (0.5 g, 1.10 mmol) in absolute EtOH (30 mL). The
reaction mixture was hydrogenated under a H2 balloon for 12 h.
When reaction was complete as indicated by TLC (EtOAc/
hexane 15:85 v/v), the reaction mixture was filtered through
celite and excess solvent removed under reduced pressure to
afford the amine 21 (0.36 g, 99%), which was used in the
subsequent reaction without further purification. MS (ESIþ)
m/z (relative intensity): 322.1 ([M þ H]•þ, 8%), 364.2 ([M þ
CH3CN þ H]•þ, 100%).
((2S,4R)-2-((tert-Butyldimethylsilyloxy)methyl)-4-(4-methoxy-
phenyl)pyrrolidin-1-yl)(4,5-dimethoxy-2-nitrophenyl)methanone
(22). EDCI (0.26 g, 1.35 mmol, 1.2 equiv) was added to a stirred
solution of the carboxylic acid (0.31 g, 1.35 mol, 1.2 equiv) in
anhydrous CH2Cl2 (30 mL) under a nitrogen atmosphere at
0 ꢀC. After stirring for 10 min, the mixture was treated with
HOBt (0.18 g, 1.35 mmol, 1.2 equiv) and a few drops of DMF,
and the resulting mixture allowed to warm to room temperature
and stirred for 2 h. The mixture was again cooled to
0 ꢀC and treated dropwise with a solution of amine 21 (0.36 g,
1.12 mmol, 1.0 equiv) in anhydrous CH2Cl2 (10 mL). When
reaction was complete, as indicated by TLC (EtOAc/hexane 1:1
v/v), the reaction mixture was diluted with CH2Cl2 (30 mL),
washed with 1 N HCl (20 mL), saturated aqueous NaHCO3
(20 mL), and brine (20 mL), and then dried (MgSO4), filtered,
and evaporated in vacuo. The title compound 22 was purified by
flash column chromatography (EtOAc/hexane 3:7 v/v) to pro-
(S)-Benzyl 2-((tert-Butyldimethylsilyloxy)methyl)-4-(trifluo-
romethylsulfonyloxy)-2,5-dihydro-1H-pyrrole-1-carboxylate (19).
A cold (-78 ꢀC) solution of ketone 18 (800 mg, 2.20 mmol, 1.0
equiv) in THF (20 mL) was added slowly to a stirred solution of
sodium bis(trimethylsilyl)amide (2.4 mL, 2.4 mmol, 1.1 equiv) in
THF (20 mL) at -78 ꢀC. After 1 h, a cold (-20 ꢀC) solution of N,
N-bis(trifluoromethanesulfonyl)anilide (0.96 g, 2.4 mmol, 1.1
equiv) in THF (20 mL) was cannulated into the flask and the
reaction mixture was allowed to warm slowly to room tempera-
ture. After 4 h, TLC (EtOAc/hexane 3:7 v/v) indicated complete
consumption of starting material. The excess solvent was then
removed under reducedpressure, andthe residue was subjected to
flash chromatography purification (EtOAc/hexane 1:9 v/v) to
vide a brown solid (0.44 g, 74%). [R]22 = -0.26ꢀ (c = 0.20,
D
CHCl3). IR (film, νmax/cm-1): 1644, 1580, 1516, 1429, 1337,
1277. 1H NMR (CDCl3): δ 0.13 (s, 6 H, Si(CH3)2), 0.95 (s, 9 H,
C(CH3)3), 2.26-2.48 (m, 2 H, H1), 3.16-3.40 (m, 3 H, H2 and
H3), 3.76 (s, 3 H, 40-OCH3), 3.94 (s, 6 H, 7,8-OCH3), 3.90 (br s, 1
H, H11), 4.25 (br s, 1 H, H11), 4.47 (br s, 1 H, H11a), 6.77 (s, 1 H,
H6), 6.80 (d, 2 H, J = 8.6 Hz, H30 and H50), 7.11 (d, 2 H, J = 8.6
Hz, H20 and H60), 7.67 (s, 1 H, H9). 13C NMR (CDCl3) rotamers:
δ 166.5 (C5), 158.6 (C40), 154.1 (C7), 148.8 (C8), 137.6 (C9a),
131.9 (C10), 128.2 (C5a), 128.1 (C20 and C60), 114.0 (C30 and C50),
109.2 (C6), 107.2 (C9), 62.7 (C11), 58.8 (C11a), 57.3 (C3), 56.6,
56.5 (7-OCH3 and 8-OCH3), 55.3 (40-OCH3), 43.2 (C2), 34.5 (C1),
25.8 (C(CH3)3), 18.2 (Cquat), -5.5 (SiCH3)2). MS (ESIþ) m/z
(relative intensity): 531.3 ([M þ H]•þ, 100%).
afford the enol-triflate 19 (1.0 g, 2.02 mmol, 91%). [R]21
=
D
-0.25ꢀ (c=0.18, CHCl3). IR (film, νmax/cm-1): 1716, 1429, 1328,
1219, 1140, 837. 1H NMR (CDCl3) rotamers: δ -0.04-0.02 (m, 6
H, Si(CH3)2), 0.86 (s, 9 H, C(CH3)3), 3.66-3.88 (m, 2 H, H11),
4.20-4.41 (m, 2 H, H3), 4.61-4.69 (m, 1 H, H11a), 5.18 (s, 2 H,
CH2-CBz), 5.74 (d, 1 H, J = 8.4 Hz, H1), 7.31-7.40 (m, 5 H,
Ar-CBz). 13C NMR (CDCl3) rotamers: δ 154.1 (NCO-CBz),
143.5 (C2), 139.2 (Ar-CBz), 128.1-128.6 (Ar-CBz), 115.2 (C1),
118.5 (q, J=256 Hz from 13C19F3 coupling, CF3), 67.3 (CH2-
CBz), 63.6 (C11), 62.6 (C11a), 50.1 (C3), 25.7 (C(CH3)3), 18.1
(Cquat), -5.6 (SiCH3)2).
(S)-Benzyl 2-((tert-Butyldimethylsilyloxy)methyl)-4-(4-meth-
oxyphenyl)-2,5-dihydro-1H-pyrrole-1-carboxylate (20). 4-Meth-
oxyphenylboronic acid (0.37 g, 2.42 mmol, 1.0 equiv) and
Pd(PPh3)4 (0.03 g, 0.024 mmol, 0.01 equiv) were added to a
solution of enol-triflate 19 (1.2 g, 2.42 mmol, 1.0 equiv) in
toluene (50 mL) at room temperature. After stirring for 5 min,
a 2.0 M aqueous solution of Na2CO3 was added to the mixture.
The reaction mixture was stirred at room temperature for 36 h
when TLC (EtOAc/hexane 3:7 v/v) indicated complete con-
sumption of starting material. Excess solvent was removed, and
the residue was dissolved in EtOAc (50 mL), washed with water
(50 mL) and brine (50 mL), and dried over MgSO4. Filtration
and evaporation of solvent afforded the crude product which
was purified by flash column chromatography (hexane/EtOAc
(2-Amino-4,5-dimethoxyphenyl)((2S,4R)-2-((tert-butyldimethyl-
silyloxy)methyl)-4-(4-methoxyphenyl)pyrrolidin-1-yl)methanone
(23). A catalytic amount of 10% palladium on carbon (25 mg)
was added to a solution of 22 (0.25 g, 0.47 mmol) in absolute
EtOH (30 mL). The reaction mixture was hydrogenated under a
H2 balloon for 12 h, and when reaction was complete as
indicated by TLC (EtOAc-hexane 1:1 v/v), the reaction mixture
was filtered through celite and excess solvent removed under
reduced pressure to afford the amine 23 which was used in the
subsequent reaction without further purification (170 mg, 72%).
[R]18 = -0.10ꢀ (c = 0.22, CHCl3). MS (ESIþ) m/z (relative
D
intensity): 501.3 ([M þ H]•þ,100%) (see Supporting Informa-
tion for IR and NMR data).
Allyl 2-((2S,4R)-2-((tert-Butyldimethylsilyloxy)methyl)-4-(4-
methoxyphenyl)pyrrolidine-1-carbonyl)-4,5-dimethoxyphenylcar-
bamate (24). Pyridine (0.11 mL, 1.36 mmol, 2.2 equiv) was added
to a stirred solution of the amine 23 (0.31 g, 0.62 mmol, 1.0
equiv) in CH2Cl2 (20 mL) at 0 ꢀC under a N2 atmosphere. The
cooled mixture was then treated dropwise with a solution of allyl
chloroformate (0.10 mL, 0.93 mmol, 1.5 equiv) in CH2Cl2 (10
mL). After 6 h stirring at room temperature, the mixture was
diluted with CH2Cl2 (30 mL), washed with 1 N HCl (20 mL),
water (20 mL) and brine (20 mL), and then dried (MgSO4),
1:9 v/v) to afford 20 as a white glass (0.5 g, 45%). [R]22
=
D
-0.32ꢀ (c = 0.20, CHCl3). IR (film, νmax/cm-1): 1710, 1609,
1515, 1416, 1257, 1112, 836. 1H NMR (CDCl3) rotamers: δ
0.0-0.07 (m, 6 H, Si(CH3)2), 0.89 (s, 9 H, C(CH3)3), 3.62-4.04
(m, 2 H, H11), 3.82 (s, 3 H, 40-OCH3), 4.45-4.67 (m, 2 H, H3),
4.63-4.79 (m, 1 H, H11a), 5.19-5.29 (m, 2 H, CH2-CBz), 6.09
(d, 1 H, J = 4.4 Hz, H1), 6.89-6.93 (m, 2 H, H30 and H50),
7.30-7.46 (m, 7 H, Ar-CBz, H20 and H60). 13C NMR (CDCl3)
rotamers: δ 159.5 (C40), 154.5 (NCO-CBz), 136.6-137.1

2938 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 7
Antonow et al.
filtered, and evaporated in vacuo. The crude residue was pur-
ified by flash chromatography (EtOAc/hexane 2:8 v/v) to afford
the alloc-protected product 24 as a pale orange solid (0.30 g,
Hz, H3R), 6.81 (s, 1 H, H9), 6.89 (d, 2 H, J = 8.6 Hz, H30 and
H50), 7.20 (d, 2 H, J = 8.6 Hz, H20 and H60), 7.54 (s, 1 H, H6),
7.64 (d, 1 H, J = 4.1 Hz, H11). 13C NMR (CDCl3): δ 164.8 (C5),
162.7 (C11), 158.8 (C40), 151.5(C8), 147.7 (C7), 140.8 (C9a), 131.5
(C10), 128.0 (C20 and C60), 119.5 (C5a), 114.1 (C30 and C50), 111.6
(C6), 109.7 (C9), 56.1 (7,8-OCH3), 55.3 (40-OCH3), 54.0 (C11a),
52.6 (C3), 41.4 (C2), 37.7 (C1). MS (ESIþ) m/z (relative intensity):
365.3 ([M - H]•-, 40%).
82%). MS (ESIþ) m/z (relative intensity): 585.3 ([M þ H]•þ
,
100%).
Allyl 2-((2S,4R)-2-(Hydroxymethyl)-4-(4-methoxyphenyl)py-
rrolidine-1-carbonyl)-4,5-dimethoxyphenylcarbamate (25). To a
solution of 24 (0.3 g, 0.51 mmol, 1.0 equiv) in THF (30 mL) at
0 ꢀC under a N2 atmosphere, a solution of TBAF (0.61 mL of a 1
M solution in THF, 0.61 mmol, 1.2 equiv) was added. The
mixture was allowed to warm to room temperature and stirring
continued for a further 3 h when TLC (EtOAc/hexane 1:1 v/v)
revealed complete consumption of starting material. The excess
solvent was removed, the residue dissolved in EtOAc (30 mL),
and the solution washed with saturated aqueous NH4Cl (10 mL)
and brine (10 mL), and then dried (MgSO4). Evaporation of the
solvent in vacuo afforded the crude product that was purified by
flash column chromatography (EtOAc/hexane 1:1 v/v) to afford
General Methodology for the Synthesis of C2-Aryl PBD
C11-(Sodium Bisulfite) Adducts. A solution of C2-aryl PBD-
imine (0.06 mmol, 1.0 equiv) in CH2Cl2 (1.0 mL) was added to a
freshly prepared aqueous solution of NaHSO3 (1.0 mL; 0.06
mmol, 1.1 equiv) in a test tube, and the two-phase system was
vigorously stirred for 4 h at room temperature. The work-up
involvedwashing withCH2Cl2 (3ꢀ 1 mL) using a phase-separator
cartridge pre-conditioned with CH2Cl2 (1.0 mL). The organic
layers were removed, and the remaining aqueous solution
lyophilized to afford the bisulfite adducts 28a-f (yield and
analytical data for 28a are provided below; see Supporting
Information for 28b-f) [NOTE 1: The purity of the C2-aryl
bisulfite adducts ranged from 86% to 97% as determined by
previously described HPLC methodology.35 However, analysis
of LC-MS traces relating to the additional UV-detected peaks
gave molecular ions ([M þ H]•þ) for the corresponding PBD
imines. For example, LC-MS profile [UV-peak (min); (ESIþ)
m/z, relative intensity] for compound 28d (two peaks): 7.02 min;
447.48 (C11-bisulfite: [M - Na þ H]•þ, 100%) and 7.23 min;
365.59 (N10-C11 imine: [M - SO3-Naþ þ H]•þ, 30%), and
383.56 (C11-carbinolamine ([M + H2O þ H]•þ, 100%). The
same sample showed only the C11(S)-(sodium bisulfite) species
25 as a white solid (0.23 g, 96%). [R]20 = 0.17ꢀ (c = 0.19,
D
CHCl3). MS (ESIþ) m/z (relative intensity): 471.2 ([M + H]•þ
70%). (See Supporting Information for IR and NMR data).
,
(2R,11S,11aS)-Allyl 11-Hydroxy-7,8-dimethoxy-2-(4-methoxy-
phenyl)-5-oxo-2,3,11,11a-tetrahydro-1Hpyrrolo[2,1-c][1,4]benzo-
diazepine-10(5H)-carboxylate (26). BAIB (0.17 g, 0.54 mmol, 1.1
equiv) and TEMPO (0.01 g, 0.05 mmol, 0.1 equiv) were added to
a solution of the alcohol 25 (0.23 g, 0.49 mmol, 1.0 equiv) in
CH2Cl2 (10 mL), and the mixture stirred at room temperature
for 5 h. When reaction was complete as indicated by TLC
(EtOAc), the mixture was diluted with CH2Cl2 (20 mL) and
washed with saturated Na2S2O3 (10 mL). The aqueous layer was
extracted with CH2Cl2 (2 ꢀ 30 mL) and the combined organic
layer washed with brine (20 mL) and dried (MgSO4). Evapora-
tion in vacuo afforded a solid that was purified by flash column
chromatography (EtOAc/hexane 1:1 v/v) to afford 26 as a white
1
by H-NMR (D2O and d6-DMSO). Taken together with the
1H-NMR data, these results indicate an equilibrium between the
imine and bisulfite species. It is probable that the low pH of
the HPLC mobile phase displaced the equilibrium towards the
imine form, which might be expected to be present in non-
detectable amounts in the neutral solutions employed for NMR
analyses]. A more efficient conversion to C11S-(sodium bisul-
phite) adducts can be achieved by vigorously vortexing the
emulsion as fully described in the time-course 1H-NMR experi-
ments in Supporting Information.
18
solid (0.15 g, 65%). [R]D = þ0.40ꢀ (c = 0.16, CHCl3). IR
(film, νmax/cm-1): 3351, 1709, 1605, 1516, 756. 1H NMR
(CDCl3): δ 2.11-2.17 (m, 1 H, H1), 2.72 (qu, 1 H, J = 7.4
Hz, H1), 3.27-3.36 (m, 2 H, H2 and H3), 3.62-3.69 (m, 2 H,
H11a and OH), 3.80 (s, 3 H, 40-OCH3), 3.87 (s, 3 H, 7-OCH3),
3.94 (s, 3 H, 8-OCH3), 4.50 (br s, 1 H, CH2-Alloc), 4.55 (dd, 1
H, J = 7.4, 13.9 Hz, H3), 4.69 (dd, 1 H, J = 6.6, 13.9 Hz,
CH2-Alloc), 5.15 (d, 2 H, J = 11.8 Hz, CHdCH2-Alloc), 5.73
(br s, 1 H, H11), 5.81 (br s, 1 H, CHdCH2-Alloc), 6.69 (s, 1 H,
H9), 6.88 (d, 2 H, J = 10.1 Hz, H30 and H50), 7.18 (d, 2 H, J =
10.1 Hz, H20 and H60), 7.27 (s, 1 H, H6). 13C NMR (CDCl3): δ
167.3 (C5), 158.7 (C40), 156.0 (CdO-Alloc), 150.8 (C9a), 148.4
(C8), 131.8 (C7 and C10), 128.5 (CHdCH2-Alloc), 128.0 (C20
and C60), 124.8 (C5a), 118.1 (CHdCH2-Alloc), 114.2 (C30 and
C50), 112.3 (C9), 110.8 (C6), 88.0 (C11), 66.8 (CH2-Alloc), 59.7
(C11a), 56.1 (7,8-OCH3), 55.3 (40-OCH3), 51.8 (C3), 42.1 (C2),
38.4 (C1). MS (ESIþ) m/z (relative intensity): 469.2 ([M þ H]þ,
100%).
Representative Compound. (11S,11aS)-7,8-Dimethoxy-2-(phe-
nyl)-5-oxo-5,10,11,11a-tetrahydro-1H-pyrrolo[2,1-c][1,4]benzodia-
zepine-11-sulfonic Acid Sodium Salt (28a). Yield 79%, 8 mg. H
1
NMR (DMSO): δ 3.26 (ddd, 1 H, J = 2.04, 10.23, 16.85 Hz, H1R),
3.52 (dd, 1 H, J= 2.51, 17.10 Hz, H1β), 3.71 (s, 3 H, 7-OCH3), 3.77
(s, 3 H, 8-OCH3), 3.96 (d, 1 H, J = 10.36 Hz, H11), 4.32 (td, 1 H,
J = 3.70, 10.47 Hz, H11a), 5.25 (s, 1 H, NH), 6.46 (s, 1 H, H9), 7.06
(s, 1 H, H6), 7.21 (t, 1 H, J = 7.31 Hz, H40), 7.34 (t, 2 H, J = 7.56
Hz, H30 and H50), 7.44 (d, 2 H J = 7.19 Hz, H20 and H60), 7.57 (s, 1
H, H3) (see Supporting Information for NMR data of 28b-f).
Biology. K562 Leukemia Cytotoxicity Assay. K562 human
chronic myeloid leukemia cells were maintained in RPMI
medium supplemented with 10% fetal calf serum and 2 mM
glutamine. Then, 190 μL of K562 suspension (104 cells/mL) was
added to each well of columns 2-11 of a 96-well plate, and
190 μL of medium was added to each well of columns 1 and 12.
The compound solution (in 100% DMSO) was serially diluted
across a 96-well plate. Each resulting point was then further
diluted 1/10 into cell culture medium, and each point was added
in triplicate to the cell plate (5% v/v). To cell -ve blanks and
compound -ve control wells, 10% DMSO was added at 5% v/v.
Assay plates were incubated for 96 h at 37 ꢀC in a humidified
atmosphere containing 5% CO2. Following incubation, 20 μL
of Alamar Blue (1 μM, final concentration) was added to each
well. The plates were then kept for a further 4 h at 37 ꢀC in a
humidified atmosphere containing 5% CO2 and were read at
580-620 nm fluorescence emission using an Envision plate
reader. The data were analyzed using GraphPad Prism, and
the IC50 values were read as the dose required to reduce the final
optical density to 50% of the control value.
(2R,11aS)-7,8-Dimethoxy-2-(4-methoxyphenyl)-2,3-dihydro-
1H-pyrrolo[2,1-c][1,4]benzodiazepin-5(11aH)-one (27). A cataly-
tic amount of tetrakis(triphenylphosphine)palladium (3.7 mg,
3.0 μmol, 0.01 equiv) was added to a stirred solution of the alloc-
protected carbinolamine 26 (150 mg, 0.32 mmol, 1.0 equiv) and
pyrrolidine (30 μL, 0.35 mmol, 1.1 equiv) in CH2Cl2 (20 mL).
After stirring for 30 min at room temperature under a N2
atmosphere, TLC (EtOAc) indicated complete consumption
of starting material. The solvent was evaporated in vacuo and
the crude product purified by flash chromatography (EtOAc) to
afford 27 as a colorless glass which was repeatedly evaporated
from CHCl3 in vacuo to provide the imine 27 as a light-
yellow solid (68 mg, 58%). [R]D19 = þ1.76ꢀ (c = 0.26, CHCl3).
IR (film, νmax/cm-1): 1603, 1514, 1434, 1251, 1034, 758. 1H
NMR (CDCl3): δ 2.36-2.45 (m, 1 H, H1β), 2.74-2.83 (m, 1 H,
H1R), 3.39-3.49 (m, 1 H, H2), 3.61 (dd, 1 H, J = 9.2, 12.0 Hz,
H3β), 3.80 (s, 3 H, 40-OCH3), 3.92 (s, 3 H, 7-OCH3), 3.96 (s, 3 H,
8-OCH3), 3.85-4.00 (m, 1 H, H11a), 4.34 (dd, 1 H, J = 7.5, 12.1

Article
DNA Thermal Denaturation Assay. All compounds were
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 7 2939
suite of programs (CambridgeSoft), exporting to a PDB format
for subsequent conversion to a mol2 format for AMBER using
the “antechamber” routine. Gasteiger charges were used, and
missing parameters added by means of the “parmchk” program.
Covalent binding of each ligand to the C2-NH2 functional
group of the guanine residue of DNA was performed graphi-
cally using AMBER “Xleap” utilizing “parm99” and the general
Amber force field parameters53,55,56 (gaff), maintaining the (S)-
configuration at the C11-position of the central PBD ring at the
point of attachment to the guanine N2-position. DNA/ligand
constructs were exported for subsequent minimization using
“Sander” during which the DNA alone was initially restrained
with a high force constant, allowing the ligand to adjust to the
DNA environment. Further minimization steps were performed
while gradually reducing the restraints to zero. The generalized
Born/surface area (GB/SA) implicit solvent model was used
with monovalent electrostatic ion screening simulated with
SALTCON set to 0.2 M. A long-range nonbonded cutoff of
stored at -20 ꢀC under anhydrous conditions prior to use.
Stock ligand solutions were freshly prepared in HPLC-grade
DMSO, and working solutions were produced by appropriate
dilution in aqueous buffer as required. The protocol used to
determine thermal denaturation temperatures (Tm) for double-
stranded calf thymus DNA and ligand-induced shifts (ΔTm)
has been previously described, together with the analytical proce-
dure.44,46,52 Test compounds were subjected to DNA thermal
melting (denaturation) studies using calf thymus DNA (CT-
DNA, type-I, highly polymerized sodium salt; 42% G þ C
[Sigma]) at a fixed 100 μM (in DNAp, equivalent to 50 μM in base
pairs [bp]) concentration, quantitated using an extinction coeffi-
cient of 6600 (M phosphate)-1 cm-1 at 260 nm. This gives a 1:5
[ligand]/[DNAp] ratio. Solutions were prepared in pH 7.00 ( 0.01
aqueous buffer containing 10 mM NaH2PO4/Na2HPO4 and 1 mM
Na2EDTA (all AnalaR grade). Working assay mixtures containing
CT-DNA and the test compound (0-20 μM, as required) were
incubated at 37.0 ( 0.1 ꢀC for 0-72 h using a Grant GD120 water
bath. Samples were monitored at 260 nm using a Cary 4000
UV-visible spectrophotometer fitted with a Peltier heating acces-
sory. A precision probe calibrated to (0.01 ꢀC in the -10 to þ120
ꢀC range was used for temperature measurements. Heating was
applied at a rate of 1 ꢀC min-1 in the 50-98 ꢀC range, with optical
data sampling at 0.10 ꢀC intervals. A separate experiment was
carried out using buffer alone, and this baseline was subtracted
from each DNA melting curve before data treatment. Optical data
were imported into the Origin 5 program (MicroCal Inc., North-
ampton, MA) for analysis. DNA helix f coil transition tempera-
tures (Tm) were determined at the midpoint of the normali-
zed melting profiles. Results for each compound are shown as the
mean ( standard deviation from at least three determinations.
Ligand-induced alterations in DNA melting behavior (ΔTm) are
given by ΔTm = Tm(DNA þ ligand) - Tm(DNA), where the Tm
value determined for native CT-DNA is 67.82 ( 0.07 ꢀC (averaged
from ∼110 runs). Working solutions with the candidate PBDs
contained e0.3% v/v DMSO, and Tm results were corrected for the
effects of DMSO cosolvent using a previously described linear
correction term determined for calibration mixtures.44,52 For ki-
netic experiments, working solutions of DNA-PBD mixtures at
the fixed 5:1 molar ratio were incubated at 37 ꢀC and evaluated after
fixed time intervals of 0 (i.e., no incubation), 4 and 18 h. Data were
also acquired after incubation for 72 h (not shown).
˚
100 A was used. To examine the sequence variations of the
genomic DNA which would be encountered by the ligand in calf
thymus DNA (1:1 adduct), all 16 possible permutations of the
two bases upstream of the covalent bond in the guanine of d(50-
TAXXGATT-30).d(50-AATCYYTA-30 were modeled.
Conformational analysis (Figure 4B) with a 14-mer DNA-
(14l)2 adduct [d(50-AATCTTTAAAGATT-30)2] was carried out
using a previously established procedure.43 Briefly, the confor-
mational search was carried out using the Monte Carlo multiple
minimum (MCMM) method and AMBER force field implemen-
ted in Macromodel v9.6207.57 Two thousand MCMM steps were
carried out for each adduct, and the resulting structures analyzed
using XCluster and Maestro v8.5207. The position of water
molecules around the DNA-(14l)2 adduct was also studied using
conformational search. The lowest energy NMR-based structure
of the C2-(2-naphthyl) PDB adduct (PDB entry 2k4l) was used as
a starting structure for the 1:2 adduct by substituting the relevant
C atom on the naphthyl ring with a N atom. The initial structure
was solvated using the soak command in the Impact v5.208
software, and the solvated system was subjected to 2000 steps
of conformational search as described above.
In Vivo Human Tumour Xenograft Studies. Human tumor
xenograft experiments were carried out under a project license
issued by the UK Home Office, and UKCCCR guidelines were
followed rigorously.58 HCT 116 cells (107 cells) were injected
subcutaneously into the flank of female nu/nu-BALB/c athymic
nude mice. Treatment started when xenografts reached 50-100
mm3. 14l was prepared in a vehicle of 1% DMSO/saline solution,
and control mice were injected with the vehicle only. Animals
received 14l intravenously (iv) at 300 μg/kg as a single dose or
120 μg/kg/d day for five days based on the dosing schedule
defined for SJG-136 by Alley and co-workers.24 Weights were
monitored daily from the start of treatment (i.e., day 1) to the end
of study (i.e., 12 days). Xenografts tumors were measured 3 times
a week, and tumor volumes were calculated from calliper mea-
surements as width2 ꢀ length/2. Relative tumor volumes (RTVs)
were calculated for each tumor by dividing the tumor volume at a
specific number of days after the start of treatment by the tumor
volume at the start of treatment (day 1) multiplied by 100%.
Results are mean ( SE from 5-10 experiments.
Molecular Modeling. Molecular modeling studies were per-
formed on a representative set of complexes built using the 9-
mer DNA duplex of sequence d(50-CGCAGACGC-30).d(50-
GCGTCTGCG-30) (Figure 4A) in canonical B-DNA form.
Structures for compounds 8, 12b, 13a, and 14l were built using
the HYPERCHEM package (Hyper Inc.), with point-centered
charges calculated by a reparameterized version of the AM1
semiempirical method, as implemented in HYPERCHEM.
Complexes were built with a covalent bond to the N2 atom of
the central guanine nucleotide on the first strand and were
initially subjected to molecular-mechanics minimization using
the AMBER9953 force field as implemented in HYPERCHEM
and a distance-dependent dielectric constant. Sodium counter-
ions were then added to each complex, which were then placed in
a TIP3P water box54 that extended ∼10 A from the complex,
˚
and further minimizations were performed until a gradient of
<0.1 kcal A-1 mol-1 was reached. Molecular dynamics simula-
˚
Acknowledgments. Cancer Research UK (C180/A1060 to
D.E.T.; C96/A5008 to D.I.J.; C96/A6670 to M.K.) and Spirogen
Ltd (U.K.) are thanked for financial support of this work, and
the Engineering and Physical Sciences Research Council (GR/
R47646/01) is acknowledged for providing a grant for the NMR
and MS instrumentation. The authors are grateful to Dr. Emma
Sharp for her expert help in preparation of the manuscript.
tions were then performed under these explicit solvent condi-
tions, with a production run of 5 ns for each complex, during
which the total energy was judged to have reached stability.
A final round of molecular mechanics minimization was per-
formed on each complex at the end of the dynamics, and the
resulting relative energies qualitatively compared with reference
to the unsubstituted 2-phenyl compound 8.
Preliminary energy minimization studies with the 14l complex
were carried out as follows. The 6-quinolinyl substituted ligand
(14l) was constructed and minimized using the ChemBioOffice
Supporting Information Available: Full synthetic methodol-
ogies, yields, and spectroscopic details for compounds pre-
viously published in communication form; synthetic methods,

2940 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 7
Antonow et al.
analytical data: IR, 1H-, 13C-, 2D-NMR, MS, and HRMS data;
optical rotation and HPLC analytical data for all novel C2-
substituted PBD-imine compounds reported in the main text;
(19) Wang, J.-J. Synthesis of Pyrrolo[2,1-c][1,4]benzodiazepine Analo-
gues. U.S. Patent 6,660,856 B2, 2003.
(20) Zhou, Q.; Duan, W.; Simmons, D.; Shayo, Y.; Raymond, M. A.;
Dorr, R. T.; Hurley, L. H. Design and Synthesis of a Novel
DNA-DNA Interstrand Adenine-Guanine Cross-Linking
Agent. J. Am. Chem. Soc. 2001, 123, 4865–4866.
(21) 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.; Johnson, R. K. Pyrrolo[1,4]Benzodiazepine
Antitumor Antibiotics;Relationship of DNA Alkylation and
Sequence Specificity to the Biological-Activity of Natural and
Synthetic Compounds. Chem. Res. Toxicol. 1988, 1, 258–268.
(22) Hurley, L. H.; Thurston, D. E. Pyrrolo(1,4)Benzodiazepine Anti-
tumor Antibiotics;Chemistry, Interaction with DNA, and Bio-
logical Implications. Pharm. Res. 1984, 52–59.
(23) Thurston, D. E.; Hurley, L. H. A Rational Basis for the Devolop-
ment of Antitumour Agents in the Pyrrolo[1,4]benzodiazepine
Group. Drugs Future 1983, 8, 957–971.
(24) Alley, M. C.; Hollingshead, M. G.; Pacula-Cox, C. M.; Wand, W.
R.; Hartley, J. A.; Howard, P. W.; Gregson, S. J.; Thurston, D. E.;
Sausville, E. A. SJG-136 (NSC 694501), a Novel Rationally De-
signed DNA Minor Groove Interstrand Cross-Linking Agent with
Potent and Broad Spectrum Antitumor Activity. Part 2: Efficacy
Evaluations. Cancer Res. 2004, 64, 6700–6706.
(25) Hartley, J. A.; Spanswick, V. J.; Brooks, N.; Clingen, P. H.;
McHugh, P. J.; Hochhauser, D.; Pedley, R. B.; Kelland, L. R.;
Alley, M. C.; Schultz, R.; Hollingshead, M. G.; Schweikart, K. M.;
Tomaszewski, J. E.; Sausville, E. A.; Gregson, S. J.; Howard, P. W.;
Thurston, D. E. SJG-136 (NSC 694501), a Novel Rationally
Designed DNA Minor Groove Interstrand Cross-Linking Agent
with Potent and Broad Spectrum Antitumor Activity. Part 1:
Cellular Pharmacology, in Vitro and Initial in Vivo Antitumor
Activity. Cancer Res. 2004, 64, 6693–6699.
(26) Hochhauser, D.; Meyer, T.; Spanswick, V. J.; Wu, J.; Clingen, P.
H.; Loadman, P.; Cobb, M.; Gumbrell, L.; Begent, R. H.; Hartley,
J. A.; Jodrell, D. Phase I Study of Sequence-Selective Minor
Groove DNA Binding Agent SJG-136 in Patients with Advanced
Solid Tumors. Clin. Cancer Res. 2009, 15, 2140–2147.
1
time-course H NMR analyses and MS-MALDI experiments
with the sodium bisulfite adducts and DNA-adducts; NCI 60
cell line data for all library members. This material is available
free of charge via the Internet at http://pubs.acs.org.
References
(1) 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.; Waring,
M. J., Eds.; The Macmillan Press Ltd.: London, 1993; Vol. 1, pp 54-88.
(2) Thurston, D. E.; Bose, D. S. Synthesis of DNA-Interactive
Pyrrolo[2,1-c][1,4]benzodiazepines. Chem. Rev. 1994, 94, 433–465.
(3) Al-Said, N. H. Facile synthesis of pyrrolo[2,1-c][1,4]benzodia-
zepine-5,11-dione dimer. J. Heterocycl. Chem. 2006, 43, 1091–1093.
(4) Baraldi, P. G.; Balboni, G.; Cacciari, B.; Guiotto, A.; Manfredini,
S.; Romagnoli, R.; Spalluto, G.; Thurston, D. E.; Howard, P. W.;
Bianchi, N.; Rutigliano, C.; Mischiati, C.; Gambari, R. Synthesis,
in vitro Antiproliferative Activity, and DNA-Binding Properties of
Hybrid Molecules Containing Pyrrolo[2,1-c][1,4]benzodiazepine
and Minor-Groove-Binding Oligopyrrole Carriers. J. Med. Chem.
1999, 42, 5131–5141.
(5) Carlier, P. R.; Zhao, H.; MacQuarrie-Hunter, S. L.; DeGuzman, J.
C.; Hsu, D. C. Enantioselective Synthesis of Diversely Substituted
Quaternary 1,4-Benzodiazepin-2-ones and 1,4-Benzodiazepine-
2,5-diones. J. Am. Chem. Soc. 2006, 128, 15215–15220.
(6) Fukuyama, T.; Liu, G.; Linton, S. D.; Lin, S. C.; Nishino, H. Total
Synthesis of (þ)-Porothramycin-B. Tetrahedron Lett. 1993, 34,
2577–2580.
(7) Gauzy, L.; Zhao, R.; Deng, Y.; Li, W.; Bouchard, H.; Chari, R. V.
J.; Commercon, A. Cytotoxic Agents Comprising New Tomaymy-
cin Derivatives and Their Therapeutic Use. International Patent
WO 2007/085930 A1, 2007.
(27) Janjigian, Y. Y.; Lee, W.; Kris, M. G.; Miller, V. A.; Krug, L. M.;
Azzoli, C. G.; Senturk, E.; Calcutt, M.; Rizvi, N. A.; Phase, I Trial
of SJG-136 (NSC/694501) in Advanced Solid Tumors. Cancer
Chemother. Pharmacol. 2009, 0.
(28) Rahman, K. M.; Thompson, A. S.; James, C. H.; Narayanaswamy,
M.; Thurston, D. E. The Pyrrolobenzodiazepine Dimer SJG-136
Forms Sequence-Dependent Intrastrand DNA Cross-Links and
Monoalkylated Adducts in Addition to Interstrand Cross-Links.
J. Am. Chem. Soc. 2009, 131, 13756–13766.
(29) Kaliszczak, M.; Antonow, D.; Patel, K.; Jodrell, D.; Howard, P.;
Thurston, D.; Guichard, S. In Pyrrolo[2,1-c][1,4]benzodiazepine
(PBD) analogues show differential substrate specificity to ABC
transporters. AACR Meeting Abstracts, April 12, 2008, Abstract
No 3226.
(30) Aird, R. E.; Thomson, M.; Macpherson, J. S.; Thurston, D. E.;
Jodrell, D. I.; Guichard, S. M. ABCB1 Genetic Polymorphism
Influences the Pharmacology of the New Pyrrolobenzodiazepine
Derivative SJG-136. Pharmacogenomics J. 2008, 8, 289–296.
(31) Guichard, S. M.; Macpherson, J. S.; Thurston, D. E.; Jodrell, D. I.
Influence of P-glycoprotein Expression on in vitro Cytotoxicity
and in vivo Antitumour Activity of the Novel Pyrrolobenzodiaze-
pine Dimer SJG-136. Eur. J. Cancer 2005, 41, 1811–1818.
(32) Burger, A. M.; Loadman, P. M.; Thurston, D. E.; Schultz, R.;
Fiebig, H.; Bibby, M. C. Preclinical Pharmacology of the Pyrrolo-
benzodiazepine (PBD) Monomer DRH-417 (NSC 709119).
J. Chemother 2007, 19, 66–78.
(8) Howard, P. W.; Chen, Z. Z.; Gregson, S. J.; Masterson, L. A.;
Tiberghien, A. C.; Cooper, N.; Fang, M.; Coffils, M. J.; Klee, S.;
Hartley, J. A.; Thurston, D. E. Synthesis of a Novel C2/C2 0-Aryl-
Substituted Pyrrolo[2,1-c][1,4]benzodiazepine Dimer Prodrug with
Improved Water Solubility and Reduced DNA Reaction Rate.
Bioorg. Med. Chem. Lett. 2009, 19, 6463–6466.
(9) Hu, W.-P.; Liang, J.-J.; Kao, C.-L.; Chen, Y.-C.; Chen, C.-Y.; Tsai,
F.-Y.; Wu, M.-J.; Chang, L.-S.; Chen, Y.-L.; Wang, J.-J. Synthesis
and Antitumor Activity of Novel Enediyne-Linked Pyrrolo[2,1-
c][1,4]benzodiazepine Hybrids. Bioorg. Med. Chem. 2009, 17, 1172–
1180.
(10) Kamal, A.; Rajender; Reddy, D. R.; Reddy, M. K.; Balakishan, G.;
Shaik, T. B.; Chourasia, M.; Sastry, G. N. Remarkable Enhance-
ment in the DNA-Binding Ability of C2-Fluoro Substituted
Pyrrolo[2,1-c][1,4]benzodiazepines and their Anticancer Potential.
Bioorg. Med. Chem. 2009, 17, 1557–1572.
(11) Kraus, G. A.; Melekhov, A. Synthesis of Pyrrolobenzodiazepines
via the PIFA Oxidation of Amines. Synthesis of 8-Deoxy DC-81.
Tetrahedron 1998, 54, 11749–11754.
(12) Kumar, R.; Lown, J. W. Recent Developments in Novel Pyrrolo-
[2,1-c][1,4]Benzodiazepine Conjugates: Synthesis and Biological
Evaluation. Mini-Rev. Med. Chem. 2003, 3, 323–339.
(13) Mayer, J. P.; Zhang, J. W.; Bjergarde, K.; Lenz, D. M.; Gaudino,
J. J. Solid phase synthesis of 1,4-benzodiazepine-2,5-diones. Tetra-
hedron Lett. 1996, 37, 8081–8084.
(14) Moroder, L.; Lutz, J.; Grams, F.; RudolphBohner, S.; Osapay, G.;
Goodman, M.; Kolbeck, W. A New Efficient Method for the
Synthesis of 1,4-Benzodiazepine-2,5-dione Diversomers. Biopoly-
mers 1996, 38, 295–300.
(33) Cooper, N.; Hagan, D. R.; Tiberghien, A.; Ademefun, T.; Matthews,
C. S.; Howard, P. W.; Thurston, D. E. Synthesis of Novel C2-Aryl
Pyrrolobenzodiazepines (PBDs) as Potential Antitumour Agents.
Chem. Commun. 2002, 1764–1765.
(15) O’Neil, I. A.; Thompson, S.; Kalindjian, S. B.; Jenkins, T. C. The
Synthesis and Biological Activity of C2-Fluorinated Pyrrolo[2,1-
c][1,4]benzodiazepines. Tetrahedron Lett. 2003, 44, 7809–7812.
(16) Purnell, B.; Sato, A.; O’Kelley, A.; Price, C.; Summerville, K.;
Hudson, S.; O’Hare, C.; Kiakos, K.; Asao, T.; Lee, M.; Hartley,
J. A. DNA Interstrand Crosslinking Agents: Synthesis, DNA
Interactions, and Cytotoxicity of Dimeric Achiral Seco-Amino-
CBI and Conjugates of Achiral Seco-Amino-CBI with Pyrrolo-
benzodiazepine (PBD). Bioorg. Med. Chem. Lett. 2006, 16, 5677–81.
(17) Rojas-Rousseau, A.; Langlois, N. Synthesis of a New Structural
Analogue of (þ)-Porothramycin. Tetrahedron 2001, 57, 3389–3395.
(18) Tercel, M.; Stribbling, S. M.; Sheppard, H.; Siim, B. G.; Wu, K.;
Pullen, S. M.; Botting, K. J.; Wilson, W. R.; Denny, W. A. Un-
symmetrical DNA Cross-Linking Agents: Combination of the CBI
and PBD Pharmacophores J. Med. Chem. 2003, 46, 2132–2151.
(34) Antonow, D.; Jenkins, T. C.; Howard, P. W.; Thurston, D. E.
Synthesis of a Novel C2-Aryl Pyrrolo[2,1-c][1,4]benzodiazepine-
5,11-dione Library: Effect of C2-Aryl Substitution on Cytotoxicity
and Non-Covalent DNA Binding. Bioorg. Med. Chem. 2007, 15,
3041–3053.
(35) Antonow, D.; Cooper, N.; Howard, P. W.; Thurston, D. E. Parallel
Synthesis of a Novel C2-Aryl Pyrrolo[2,1-c][1,4]benzodiazepine
(PBD) Library. J. Comb. Chem. 2007, 9, 437–445.
(36) Kang, G. D.; Howard, P. W.; Thurston, D. E. Synthesis of a Novel
C2-Aryl Substituted 1,2-Unsaturated Pyrrolobenzodiazepine.
Chem. Commun. 2003, 1688–1689.
(37) Leimgruber, W.; Batcho, A. D.; Czajkowski, R. C. Total Synthesis
of Anthramycin. J. Am. Chem. Soc. 1968, 90, 5641–5643.
(38) Tiberghien, A. C.; Hagan, D.; Howard, P. W.; Thurston, D. E.
Application of the Stille Coupling Reaction to the Synthesis of

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 7 2941
C2-Substituted endo-exo Unsaturated Pyrrolo[2,1-c][1,4]-
benzodiazepines (PBDs). Bioorg. Med. Chem. Lett. 2004, 14,
5041–5044.
Unsaturation on the Cytotoxicity and DNA-Binding Reactivity
of Pyrrolo[2,1-c][1,4]benzodiazepines. Bioorg. Med. Chem. Lett.
2000, 10, 1849–1851.
(39) Gregson, S. J.; Howard, P. W.; Corcoran, K. E.; Barcella, S.;
Yasin, M. M.; Hurst, A. A.; Jenkins, T. C.; Kelland, L. R.;
Thurston, D. E. Effect of C2-exo Unsaturation on the Cytotoxicity
and DNA-Binding Reactivity of Pyrrolo[2,1-c][1,4]benzodia-
zepines. Bioorg. Med. Chem. Lett. 2000, 10, 1845–1847.
(40) Kaneko, T.; Wong, H.; Doyle, T. W.; Rose, W. C.; Bradner, W. T.
Bicyclic and Tricyclic Analogs of Anthramycin. J. Med. Chem.
1985, 28, 388–392.
(41) Langlois, N.; Rojas-Rousseau, A.; Gaspard, C.; Werner, G. H.;
Darro, F.; Kiss, R. Synthesis and Cytotoxicity on Sensitive and
Doxorubicin-Resistant Cell Lines of New Pyrrolo[2,1-c][1,4]ben-
zodiazepines Related to Anthramycin. J. Med. Chem. 2001, 44,
3754–3757.
(42) Thurston, D. E.; Langley, D. R. Synthesis and Stereochemistry of
Carbinolamine-Containing Pyrrolo[1,4]Benzodiazepines by Re-
ductive Cyclization of N-(2-Nitrobenzoyl)Pyrrolidine-2-Carboxal-
dehydes. J. Org. Chem. 1986, 51, 705–712.
(50) Guichard, S.; Arnould, S.; Hennebelle, I.; Bugat, R.; Canal, P.
Combination of oxaliplatin and irinotecan on human colon cancer
cell lines: activity in vitro and in vivo. Anti-Cancer Drugs 2001, 12,
741–751.
(51) Guichard, S. M.; Macpherson, J. S.; Mayer, I.; Reid, E.; Muir, M.;
Dodds, M.; Alexander, S.; Jodrell, D. I. Gene expression predicts
differential capecitabine metabolism, impacting on both pharma-
cokinetics and antitumour activity. Eur. J. Cancer 2008, 44, 310–
317.
(52) McConnaughie, A. W.; Jenkins, T. C. Novel Acridine-Triazenes as
Prototype Combilexins - Synthesis, DNA-Binding, and Biological
Activity. J. Med. Chem. 1995, 38, 3488–3501.
(53) Cornell, W. D.; Cieplak, P.; Bayly, C. I.; Gould, I. R.; Merz, K. M.;
Ferguson, D. M.; Spellmeyer, D. C.; Fox, T.; Caldwell, J. W.;
Kollman, P. A. A 2nd Generation Force-Field for the Simulation
of Proteins, Nucleic-Acids, and Organic-Molecules. J. Am. Chem.
Soc. 1995, 117, 5179–5197.
(43) Antonow, D.; Barata, T.; Jenkins, T. C.; Parkinson, G. N.;
Howard, P. W.; Thurston, D. E.; Zloh, M. Solution Structure of
a 2:1 C2-(2-Naphthyl) Pyrrolo[2,1-c][1,4]benzodiazepine DNA
Adduct: Molecular Basis for Unexpectedly High DNA Helix
Stabilization. Biochemistry 2008, 47, 11818–11829.
(44) Jones, G. B.; Davey, C. L.; Jenkins, T. C.; Kamal, A.; Kneale, G.
G.; Neidle, S.; Webster, G. D.; Thurston, D. E. The Noncovalent
Interaction of Pyrrolo[2,1-c][1,4]Benzodiazepine-5,11-Diones with
DNA. Anti-Cancer Drug Des. 1990, 5, 249–264.
(45) Perry, P. J.; Gowan, S. M.; Reszka, A. P.; Polucci, P.; Jenkins,
T. C.; Kelland, L. R.; Neidle, S. 1,4- and 2,6-Disubstituted
Amidoanthracene-9,10-dione Derivatives as Inhibitors of Human
Telomerase. J. Med. Chem. 1998, 41, 3253–3260.
(46) Gregson, S. J.; Howard, P. W.; Gullick, D. R.; Hamaguchi, A.;
Corcoran, K. E.; Brooks, N. A.; Hartley, J. A.; Jenkins, T. C.;
Patel, S.; Guille, M. J.; Thurston, D. E. Linker Length Modulates
DNA Cross-Linking Reactivity and Cytotoxic Potency of C8/C80
Ether-Linked C2-exo-Unsaturated Pyrrolo[2,1-c][1,4]benzodia-
zepine (PBD) Dimers. J. Med. Chem. 2004, 47, 1161–1174.
(47) Thurston, D. E.; Bose, D. S.; Howard, P. W.; Jenkins, 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 Behavior and Cytotoxicity of Pyrrolo[2,1-c]
[1,4]benzodiazepines. J. Med. Chem. 1999, 42, 1951–1964.
(48) Chen, Z. Z.; Gregson, S. J.; Howard, P. W.; Thurston, D. E. A
Novel Approach to the Synthesis of Cytotoxic C2-C3 Unsaturated
Pyrrolo[2,1-c][1,4]benzodiazepines (PBDs) with Conjugated Acry-
lyl C2-Substituents. Bioorg. Med. Chem. Lett. 2004, 14, 1547–1549.
(49) Gregson, S. J.; Howard, P. W.; Barcella, S.; Nakamya, A.; Jenkins,
T. C.; Kelland, L. R.; Thurston, D. E. Effect of C2/C3-Endo
(54) Jorgensen, W. L.; Chandrasekhar, J.; Madura, J. D.; Impey, R. W.;
Klein, M. L. Comparison of Simple Potential Functions for
Simulating Liquid Water. J. Chem. Phys. 1983, 79, 926–935.
(55) Case, D. A.; Cheatham, T. E.; Darden, T.; Gohlke, H.; Luo, R.;
Merz, K. M.; Onufriev, A.; Simmerling, C.; Wang, B.; Woods, R. J.
The Amber Biomolecular Simulation Programs. J. Comput. Chem.
2005, 26, 1668–1688.
(56) Pearlman, D. A.; Case, D. A.; Caldwell, J. W.; Ross, W. S.;
Cheatham, T. E.; Debolt, S.; Ferguson, D.; Seibel, G.; Kollman,
P. Amber, a Package of Computer-Programs for Applying
Molecular Mechanics, Normal-Mode Analysis, Molecular-Dy-
namics and Free-Energy Calculations to Simulate the Structural
and Energetic Properties of Molecules. Comput. Phys. Commun.
1995, 91, 1–41.
(57) Mohamadi, F.; Richards, N. G. J.; Guida, W. C.; Liskamp, R.;
Lipton, M.; Caufield, C.; Chang, G.; Hendrickson, T.; Still, W. C.
Macromodel;An Integrated Software System for Modeling
Organic and Bioorganic Molecules Using Molecular Mechanics.
J. Comput. Chem. 1990, 11, 440–467.
(58) Workman, P.; Twentyman, P.; Balkwill, F.; Balmain, A.; Chaplin,
D.; Double, J.; Embleton, J.; Newell, D.; Raymond, R.; Stables, J.;
Stephens, T.; Wallace, J. United Kingdom Co-Ordinating Com-
mittee on Cancer Research (UKCCCR) Guidelines for the Welfare
of Animals in Experimental Neoplasia (2nd Edition). Br. J. Cancer
1998, 77, 1-10.
(59) Masterson, L. A.; Spanswick, V. J.; Hartley, J. A.; Begent, R. H.;
Howard, P. W.; Thurston, D. E. Synthesis and Biological Evalua-
tion of Novel Pyrrolo[2,1-c][1,4]benzodiazepine Prodrugs for Use
in Antibody-Directed Enzyme Prodrug Therapy. Bioorg. Med.
Chem. Lett. 2006, 16, 252–256.