Fluorescent 7-diethylaminocoumarin pyrrolobenzodiazepine conjugates: Synthesis, DNA interaction, cytotoxicity and differential cellular localization

Article abstract of DOI:10.1016/j.bmcl.2008.01.083

The pyrrolo[2,1-c][1,4]benzodiazepines (PBDs) are a class of DNA minor groove binding agents that react covalently with guanine bases, preferably at Pu-G-Pu sites. A series of three fluorescent PBD-coumarin conjugates with different linker architectures has been synthesized to probe correlations between DNA binding affinity, cellular localization and cytotoxicity. The results show that the linker structure plays a critical role for all three parameters.

Full text of DOI:10.1016/j.bmcl.2008.01.083

Available online at www.sciencedirect.com
Bioorganic & Medicinal Chemistry Letters 18 (2008) 2147–2151
Fluorescent 7-diethylaminocoumarin pyrrolobenzodiazepine
conjugates: Synthesis, DNA interaction, cytotoxicity
and differential cellular localization
Geoffrey Wells,a Marie Suggitt,b Marissa Coffils,b Mirza A. H. Baig,a
Philip W. Howard,a,b Paul M. Loadman,b John A. Hartley,b,c
Terence C. Jenkinsd and David E. Thurstona,b,
*
aCancer Research UK Gene Targeted Drug Design Research Group, School of Pharmacy, University of London, 29/39
Brunswick Square, London WC1N 1AX, UK
bSpirogen Ltd, 29/39 Brunswick Square, London WC1N 1AX, UK
cCancer Research UK Drug–DNA Interactions Research Group, UCL Cancer Institute, Paul O’Gorman Building,
Huntley Street, London WC1E 6BT, UK
d
ˆ
Morvus Technology Limited, Ty Myddfai, Llanarthne, Carmarthen SA32 8HZ, UK
Received 3 December 2007; revised 21 January 2008; accepted 22 January 2008
Available online 30 January 2008
Abstract—The pyrrolo[2,1-c][1,4]benzodiazepines (PBDs) are a class of DNA minor groove binding agents that react covalently
with guanine bases, preferably at Pu-G-Pu sites. A series of three fluorescent PBD–coumarin conjugates with different linker archi-
tectures has been synthesized to probe correlations between DNA binding affinity, cellular localization and cytotoxicity. The results
show that the linker structure plays a critical role for all three parameters.
Ó 2008 Published by Elsevier Ltd.
The design and synthesis of sequence-selective DNA-
interactive agents are currently of interest for therapeu-
tic applications, particularly in view of the recent
sequencing of the human genome and the growing
understanding of genetic changes associated with serious
diseases such as cancer. A number of strategies have
been exploited to develop ligands which bind to and dis-
rupt DNA functioning, such as triplex-forming oligonu-
cleotides, DNA intercalators, and major and minor
groove binding agents.1 Of the minor groove-targeted
agents, the polyamides, especially those with a hairpin
structure, are notable for their sequence discriminating
properties and high affinity for DNA.2 While the cell-
free activity of these compounds was promising, early
studies suggested that this could not be reliably trans-
lated into cellular responses.3
Understanding the cellular distribution of DNA-target-
ing molecules is particularly important as they have the
potential to interact off-target which may diminish their
ability to reach the genome and elicit the desired re-
sponse. For these reasons, experiments with fluoro-
phore-conjugated polyamides have been carried out to
investigate their lack of cellular activity. Fluorescent
BODIPY–polyamide conjugates were shown to retain
affinity towards naked DNA but confocal microscopy
demonstrated that they localized predominantly in the
cytoplasm rather than the nucleus for >20 different cell
lines (with the exception of certain T cell lines).3,4 Fur-
ther experiments showed that BODIPY conjugates accu-
mulated in acidic cytoplasmic vesicles in rheumatoid
synovial fibroblasts and HCT-116 tumor cells, whereas
fluorescein coupled polyamides could enter the nuclei
of HCT-116 cells.5 A more comprehensive study of fluo-
rophore-conjugated polyamides indicated that composi-
tion, charge, size, fluorophore type and linker structure
all influenced nuclear accumulation in various different
cell types.6 Other types of DNA-interactive agents have
been attached to fluorescent markers; for example, a
Keywords: Pyrrolobenzodiazepines; PBDs; DNA-binding; DNA seq-
uence-selective; Fluorescence conjugates; Nuclear penetration; Anti-
cancer agents; Antitumour agents.
*
Corresponding author. Tel.: +44 2077535932; fax: +44
2077535964; e-mail: david.thurston@pharmacy.ac.uk
0960-894X/$ - see front matter Ó 2008 Published by Elsevier Ltd.
doi:10.1016/j.bmcl.2008.01.083

2148
G. Wells et al. / Bioorg. Med. Chem. Lett. 18 (2008) 2147–2151
Fmoc-linked analogue of distamycin was used to show
that it localizes in cytoplasmic organelles.7
occurring DC-81 (1). These are based on a fluorophore
tethered to the C8-hydroxyl group of the PBD through
different linkers. 7-Diethylaminocoumarin was selected
as the fluorophore due to its chemical stability and
favorable optical characteristics.
To date there have been relatively few studies of the cel-
lular localization of agents that interact covalently with-
in the minor groove of DNA where off-target chemical
reactions with nucleophiles such as glutathione and var-
ious proteins may be problematic. So far, only Fmoc-la-
beled polyamide–seco-cyclopropabenzindole conjugates
have been studied and appear to localize in the cytoplas-
mic organelles of treated EMT-6 cells.8
The coumarin fragment was prepared by Knoevenagel
condensation of benzaldehyde 4 with diethyl malonate
to give ester 5, hydrolysis of which produced the acid
6. Conversion to the acid chloride followed by reaction
with mono-Boc-protected 1,4-diaminobutane gave the
N-Boc-butylamino derivative 7. Deprotection with
95% TFA provided the free amine 8 (Scheme 1). The
simplest conjugate 11, where the PBD is joined to the
coumarin fragment via a C3-linker, was synthesized by
reaction of the previously reported N-10-Alloc-pro-
tected 8-(3-aminopropyl)-PBD (9)10 with the coumarin
acid (6), to give the coupled intermediate 10. Alloc
deprotection using Pd[PPh3]4 gave the desired conjugate
119 (Scheme 2). The equivalent C8-linked homologue 17
was synthesized in a similar manner (Scheme 3). The
protected PBD 1211 was O–debenzylated and the result-
Our own recent studies have focused on sequence-selec-
tive DNA-targeting agents based on the pyrrolo[2,1-
c][1,4]benzodiazepine (PBD) structure (see Fig. 1). These
molecules have innate DNA sequence-recognition prop-
erties, binding covalently via their electrophilic N10–
C11 imine moiety to embedded purine–guanine–purine
sites in the minor groove. In order to evaluate the cellu-
lar localization of PBDs and investigate how this corre-
lates with their DNA-binding and cytotoxicity, we have
synthesized a number of analogues of the naturally
Figure 1. Examples of naturally occurring PBD molecules.
O
O
CHO
OH
NHR
iii., iv.
i.
OR
N
H
Et2N
Et2N
O
O
Et2N
O
O
5: R = Et
7: R = Boc
8
: R = H
4
v.
ii.
6
: R = H
Scheme 1. Reagents and conditions: (i) diethyl malonate, EtOH, reflux, 48%; (ii) NaOH, H2O, MeOH, 97%; (iii) (COCl)2, DMF, CH2Cl2; (iv)
H2N(CH2)4NHBoc, Et3N, CH2Cl2, 70% (2 steps); (v) 95% aq TFA, 92%.
Et2N
O
O
O
Alloc
Alloc
OH
H
OH
H
H
N
N
N
H2N
O
O
i.
N
N
MeO
MeO
O
O
9
10
ii.
Et2N
O
O
O
H
N
N
O
H
N
MeO
O
11
Scheme 2. Reagents: (i) 6, EDCI, DMAP, DMF, 81%; (ii) Pd[PPh3]4, pyrrolidine, CH2Cl2, 89%.

G. Wells et al. / Bioorg. Med. Chem. Lett. 18 (2008) 2147–2151
2149
O
Boc
Boc
Boc
OTHP
H
OTHP
H
OTHP
H
N
N
N
RO
N
X
O
H2N
X
O
ii.
iii.
O
N
N
N
MeO
MeO
MeO
O
O
O
X = -[CH2]6-
14
X = -[CH2]6-
12: R = Bn
13
i.
: R = H
15
iv.
Et2N
O
O
O
Et2N
O
O
Boc
OTHP
H
H
N
H
N
N
v.
N
X
O
X
O
H
O
N
N
MeO
MeO
O
O
X = -[CH2]6-
X = -[CH2]6-
17
16
Scheme 3. Reagents and conditions: (i) H2, 10% Pd/C, EtOH, 100%; (ii) N-8-bromooctylphthalimide, K2CO3, DMF, 100%; (iii) N2H4, MeOH; (iv) 6,
EDCI, DMAP, DMF, 50% (2 steps); (v) 95% aq TFA, 95%.
ing phenol (13) treated with N-8-bromooctylphthalimide
to give 14. Hydrazinolytic cleavage gave the PBD-octyl-
amine 15, which was coupled to the coumarin acid (6) to
provide conjugate 16. Treatment with 95% TFA effected
removal of both the Boc and THP protecting groups to
provide the C8-linked PBD–coumarin conjugate 17.12
Finally, for the PBD–coumarin conjugate with an amide
linker (20), the PBD acid 18 was coupled to the couma-
rin fragment 8 to give conjugate 19. Removal of the Boc
group using 95% TFA afforded the PBD–coumarin con-
jugate 2013 (Scheme 4).
cence (e.g., absorption 320 nm, emission 420 nm for
the C-8 methyl ether of 1), their direct visualization in
microscopy applications is less practicable due to the
requirement for excitation by a UV laser rather than
light at visible wavelengths.
DNA thermal denaturation15,16 results for the com-
pounds are presented in Table 1. The data show that
the simple coumarin 5 is a poor stabilizing ligand for
calf thymus (CT) DNA and there is no time-dependent
increase in DTm during the 72 h exposure. In contrast,
the hybrid agents 11 and 17 are effective at stabilizing
CT-DNA. Their DTm values increase with exposure
time, confirming the occurrence of time-dependent cova-
lent binding typical for PBD agents.17 Interestingly, the
short-chain homologue 11 (i.e., 3 methylenes in linker) is
more potent and reacts faster than its longer-chain
counterpart 17 (i.e., 8 methylenes). Compound 20 is
considerably less effective than the related conjugate 17
The conjugates had similar fluorescence wavelengths to
the free coumarin 5 (i.e., kmax absorption ꢀ410–
420 nm, emission 455–460 nm). However, fluorescence
was significantly more intense for the conjugated agents,
which was attributed to the presence of an amide rather
than an ester linkage at the coumarin 3-position. While
the PBD molecules alone showed some innate fluores-
Et2N
O
O
O
Boc
Boc
OH
H
OH
H
H
N
N
HO
O
N
X
O
i.
O
N
N
MeO
MeO
O
O
X = -[CH2]3NHCOCH2-
18
19
ii.
Et2N
O
O
H
N
N
X
O
H
O
N
MeO
O
X = -[CH2]3NHCOCH2-
20
Scheme 4. Reagents: (i) 8, EDCI, DMAP, DMF, 98%; (ii) 95% aq TFA, 89%.

2150
G. Wells et al. / Bioorg. Med. Chem. Lett. 18 (2008) 2147–2151
Table 1. Cytotoxicity and DNA thermal denaturation data for 5 and PBD conjugates 11, 17, and 20
a
Compound
Cell line IC50 (lM)
Induced DTm (°C)b,c after incubation at 37 °C for
OVCAR-5d
LOX IMVId
0 he
4 he
18 he
72 he
5
>10
>10
0.1
2.9
1.1
0.3
0.2
3.4
1.4
0.7
0.2
3.6
2.0
0.9
0.2
3.9
2.9
1.0
11
17
20
3.2 1.04
7.8 0.33
>10
1.4 0.24
4.5 0.22
>10
a Determined by Alamar blue assay (96 h drug exposure).14 Each value represents the mean standard deviation of three independent experiments.
b For calf thymus (CT) DNA alone, DTm = 67.82 0.07 °C (mean from >110 experiments).15,16 All DTm values are 0.08–0.12 °C.
c For a fixed 1:5 molar ratio of [ligand]/[DNA], DNA concentration = 50 lM (bp) and ligand concentration = 20 lM, in aqueous sodium phosphate
buffer [10 mM sodium phosphate + 1 mM EDTA, pH 7.00 0.01]. Solutions were incubated at 37 0.1 °C for the time shown.
d 8-Methyl DC-81 was inactive at 1 lM or below in this assay.
e Values of 0.3, 0.5, 0.7, and 0.8 °C were determined for DC-81 (1) at the same 1:5 ligand/DNA molar ratio after incubation for 0, 4, 18, and 72 h,
respectively.
despite the similar linker length. It behaves in a compa-
rable manner to the simple PBD monomer DC-81 (1).
(from 1 h at 1 lM concentration; data not shown).
Compound 20 showed a very weak cytoplasmic localiza-
tion, similar to the fluorophore 5 alone, at a concentra-
tion of 10 lM suggesting poor cellular penetration. In
contrast, conjugate 11 could be observed to enter cells
and accumulate in the nucleus of both cell lines after
1 h exposure at concentrations above 1 lM. Compound
17 could be observed in the cytoplasm of both cell lines
(1–10 lM) and appeared to be present in the nucleus
after 24 h (10 lM).
The responses of OVCAR-5 and LOX IMVI tumor cells
to 5 and the conjugates are also presented in Table 1.
These cell lines were selected on the basis of intrinsic
sensitivity (LOX IMVI melanoma) and known resis-
tance (OVCAR-5 ovarian carcinoma) to PBD molecules
in xenograft studies.18 The free coumarin 5 was inactive
in both cell lines at the highest concentration tested. In
contrast, the PBD–coumarin conjugates had effects that
were dependent on their linker structure. The C3-linked
compound 11 was the most active agent tested, with
IC50 values in the two cell lines of 3.2 lM and 1.4 lM
after incubation for 96 h, with cytotoxic effects evident
in the more sensitive melanoma cell line after 24 h
(IC50 1.5 lM). The C8-linked compound 17 was less ac-
tive, with IC50 values of 7.8 lM and 4.5 lM after 96 h.
However, incorporation of an amide unit within the
C8 linker (i.e., 20) resulted in reduced activity.
These results demonstrate the critical effect of the linker
structure in governing the behavior of these conjugates.
The simple coumarin 5 has no intrinsic DNA-binding
ability, little tendency to accumulate in the nucleus of
cells, and negligible cytotoxic and cell membrane-dam-
aging effects (data not shown). However, greater fluores-
cence intensity is seen in the cytoplasm of treated LOX
IMVI melanoma cells relative to OVCAR-5. The conju-
gates 11, 17, and 20 differ significantly in their cellular
distribution properties, DNA interaction behavior and
cytotoxic potencies. Compound 11, with the short C3
linker exhibits the greatest stabilizing effect on DNA
melting and the fastest DNA binding kinetics, presum-
ably through additional hydrogen bonding and van
der Waals interactions with groups on the floor and
walls of the DNA minor groove. Localization in the nu-
clei of both treated cell lines is similarly rapid and occurs
during the first 1–4 h of treatment. Nuclear condensa-
tion (confocal microscopy) and cell killing (trypan blue
exclusion assays, data not shown) evident at longer
exposure times are consistent with the cytotoxicity of
the agent as determined using the Alamar blue assay.
Lengthening the linker between the PBD and the cou-
marin from 3 to 8 methylene units (i.e., 11 ! 17) has a
detrimental effect on the interaction of the conjugate
with DNA; the kinetics of association are slower and
the ultimate effect on thermal denaturation is smaller.
Cellular distribution of this compound was also differ-
ent. It clearly localized within the cytoplasm of both cell
lines after 1 h, with a more uniform distribution, includ-
ing the nuclear compartment, after 24 h. Cytotoxicity
was also diminished relative to compound 11. Interest-
ingly, the introduction of an amide unit into the linker
(i.e., 20) abolished most of the interaction with DNA.
The rigidity introduced into the chain by the amide
functionality may prevent the molecule from adopting
a favorable DNA-binding conformation. The potential
Cellular uptake and localization of the fluorescent PBD
conjugates within the OVCAR-5 and LOX IMVI cells
are shown in Figure 2. The detection limit for all com-
pounds was >0.1 lM. The free coumarin 5 was localized
and concentrated in the cytoplasm of both cell lines
Cpd
Cell Linea
OVCAR-5
1 h 24 h
LOX IMVI
1 h 24 h
5
11
17
20
Figure 2. Fluorescence microscopy of cells treated with 5 and PBD
conjugates 11, 17, and 20. aCells were treated for 48 h with 1 lM or
10 lM of agent. Data are shown for 1 h and 24 h time points for a
10 lM concentration.

G. Wells et al. / Bioorg. Med. Chem. Lett. 18 (2008) 2147–2151
2151
8. Sharma, S. K.; Morrissey, A. T.; Miller, G. G.; Gmeiner, W.
H.; Lown, J. W. Bioorg. Med. Chem. Lett. 2001, 11, 769.
9. Analytical data for compound 11. 1H NMR (CDCl3,
400 MHz) 8.89 (1H, br s), 8.68 (1H, s), 7.65 (1H, d,
J = 4.4 Hz), 7.50 (1H, s), 7.42 (1H, d, J = 9.0 Hz), 6.82
(1H, s), 6.64 (1H, dd, J = 9.0, 2.4 Hz), 6.49 (1H, d,
J = 2.3 Hz), 4.16 (2H, m), 3.93 (3H, s), 3.85–3.54 (5H, m),
3.45 (4H, q, J = 7.1 Hz), 2.32–2.28 (2H, m), 2.23–2.16 (2H,
m), 2.08–2.02 (2H, m), 1.23 (6H, t, J = 7.1 Hz); HRMS
([M+H]Æ+) C30H35N4O6 calc. 547.2551 found 547.2539.
10. Calderon, L. G. PhD Thesis, University of Portsmouth,
1999.
for the amide to generate additional H-bonding interac-
tions with the floor of the host minor groove did not ap-
pear to help in this case, although other placements in
the linker might be more effective. As anticipated, based
on its poor DNA binding, 20 is less cytotoxic and shows
poor cellular penetration, with minor accumulation in
the cytoplasmic compartment in both cell lines.
In summary, as demonstrated by Dervan and co-work-
ers for the hairpin polyamide family of DNA-interactive
agents,6 these data reveal the important influence of the
linker within these PBD conjugates on the extent and
time course of DNA binding, in vitro cytotoxic potency
and cellular distribution. This information provides a
useful starting point for the design of new PBD conju-
gates with optimized linkers and fluorophores that
may be of interest for biophysical optical assays and
for studying various mechanisms of inherent and ac-
quired resistance (e.g., drug efflux, DNA repair, gluta-
thione conjugation) or sub-cellular distribution in
intact cells.
11. Howard, P. W.; Kang, G.-D. Intl. Patent Appl.
WO2005085259, 2005.
12. Analytical data for compound 17. H NMR (DMSO-d6,
1
400 MHz) 8.65 (1H, s), 8.63 (1H, br s), 7.77 (1H, d,
J = 4.4 Hz), 7.68 (1H, d, J = 9.1 Hz), 7.32 (1H, s), 6.80
(1H, dd, J = 2.5, 8.9 Hz), 6.80 (1H, s), 6.61 (1H, d,
J = 2.2 Hz), 4.03 (2H, m), 3.81 (3H, s), 3.72 (1H, m), 3.67
(1H, m), 3.48 (2H, q, J = 7.1 Hz), 3.40 (1H, m), 3.30 (2H,
m), 2.24 (2H, m), 1.94 (2H, m), 1.74 (2H, m), 1.52–1.32
(10H, m), 1.14 (6H, t, J = 7.0 Hz); HRMS ([M+H]Æ+)
C35H45N4O6 calc. 617.3334 found 617.3317.
13. Analytical data for compound 20. 1H NMR (CDCl3,
400 MHz) 8.82 (1H, br s), 8.68 (1H, s), 7.65 (1H, d,
J = 4.3 Hz), 7.50 (1H, s), 7.42 (1H, d, J = 9.0 Hz), 6.84
(1H, s), 6.64 (1H, dd, J = 9.0, 2.5 Hz), 6.58 (1H, br s), 6.49
(1H, d, J = 2.3 Hz), 4.33 (2H, m), 3.92 (3H, s), 3.83–3.69
(2H, m), 3.48–3.31 (9H, m), 2.32–2.28 (2H, m), 2.07–2.03
(2H, m), 1.68–1.59 (4H, m), 1.24 (6H, t, J = 7.1 Hz);
HRMS ([M+H]Æ+) C34H42N5O7 calc. 632.3079 found
632.3055.
14. O’Brien, J.; Wilson, I.; Orton, T.; Pognan, F. Eur. J.
Biochem. 2000, 267, 5421.
15. Mcconnaughie, A. W.; Jenkins, T. C. J. Med. Chem. 1995,
38, 3488.
References and notes
1. Thurston, D. E. Chemistry and Pharmacology of Antican-
cer Drugs; CRC Press (Taylor & Francis): Boca Raton,
FL, USA, 2006, pp 281.
2. Dervan, P. B.; Edelson, B. S. Curr. Opin. Struct. Biol.
2003, 13, 284.
3. Belitsky, J. M.; Leslie, S. J.; Arora, P. S.; Beerman,
T. A.; Dervan, P. B. Bioorgan. Med. Chem. 2002,
10, 3313.
4. Dudouet, B.; Burnett, R.; Dickinson, L. A.; Wood, M. R.;
Melander, C.; Belitsky, J. M.; Edelson, B.; Wurtz, N.;
Briehn, C.; Dervan, P. B.; Gottesfeld, J. M. Chem. Biol.
2003, 10, 859.
5. Crowley, K. S.; Phillion, D. P.; Woodard, S. S.; Schweit-
zer, B. A.; Singh, M.; Shabany, H.; Burnette, B.; Hippen-
meyer, P.; Heitmeier, M.; Bashkin, J. K. Bioorg. Med.
Chem. Lett. 2003, 13, 1565.
6. Best, T. P.; Edelson, B. S.; Nickols, N. G.; Dervan, P. B.
Proc. Natl. Acad. Sci. U.S.A. 2003, 100, 12063.
7. Kumar, R.; Rai, D.; Marcus, S. L.; Ko, S. C. C.; Lown, J.
W. Lett. Org. Chem. 2004, 1, 154.
16. Antonow, D.; Jenkins, T. C.; Howard, P. W.; Thurston,
D. E. Bioorgan. Med. Chem. 2007, 15, 3041.
17. 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. J. Med. Chem.
1999, 42, 1951.
18. 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.
Cancer Res. 2004, 64, 6693.