Polyhydroxylated azepanes as new motifs for DNA minor groove binding agents

Article abstract of DOI:10.1016/S0960-894X(01)00719-3

The synthesis of 1,3-bis-[3,4,5,6-tetrahydroxyazepane-N-p-phenoxy] and 1,3-bis-[3,4,5,6-tetrahydroxyazepane-N-p-benzyloxy] propanes is reported. These compounds have been prepared to investigate the potential of incorporating iminosugars as useful recognition elements in DNA minor groove binding agents. The compounds were shown to have very moderate binding affinities for DNA in thermal denaturation and ethidium bromide displacement assays when compared with propamidine. They were also found to possess some in vitro anticancer activity that did not correlate with their DNA binding affinity.

Full text of DOI:10.1016/S0960-894X(01)00719-3

Bioorganic & Medicinal Chemistry Letters 12 (2002) 237–241
Polyhydroxylated Azepanes as New Motifs for
DNA Minor Groove Binding Agents
Heather A. Johnson and Neil R. Thomas*
School of Chemistry, University of Nottingham, Nottingham, NG7 2RD, UK
Received 16 August 2001; revised 26 October 2001; accepted 29 October 2001
Abstract—The synthesis of 1,3-bis-[3,4,5,6-tetrahydroxyazepane-N-p-phenoxy] and 1,3-bis-[3,4,5,6-tetrahydroxyazepane-N-p-ben-
zyloxy] propanes is reported. These compounds have been prepared to investigate the potential of incorporating iminosugars as
useful recognition elements in DNA minor groove binding agents. The compounds were shown to have very moderate binding
affinities for DNA in thermal denaturation and ethidium bromide displacement assays when compared with propamidine. They
were also found to possess some in vitro anticancer activity that did not correlate with their DNA binding affinity. # 2002 Elsevier
Science Ltd. All rights reserved.
The aromatic diamidines 1,5-bis-[4-amidinophenoxy)
pentane (pentamidine) 1 and 1,3-bis-[4-amidinophen-
oxy) propane (propamidine) 2 have been shown to bind
in the minor groove of DNA (Fig. 1).1 Pentamidine is
currently being used for the treatment and prophylaxis
of Pneumocystis carinii pneumonia in patients with
AIDS.2 However, the clinical use of pentamidine and
other aromatic diamidines has been restricted by their
significant toxicity, leading to adverse effects including
hypotension and hypoglycemia, and their limited oral
availability.3 Pentamidine and related drugs have also
been shown to have a high affinity for the imidazoline
I2 receptor found on rat liver membranes sugges-
ting that the amidine groupmay be responsible for the
toxicity.4
ꢀ 10À6 M.6 Inspired by these observations, we have
designed and synthesised ligands to evaluate the tetra-
hydroxyazepane group, an iminosugar, as an alternative
cationic ligand terminator to the amidine groupin
DNA binding ligands.
Polyhydroxyazepanes have several properties that make
them potentially useful as MGBL’s. The flexibility of
the seven-membered ring (compared with five- or six-
membered rings) would allow the hydroxyl groups to
adopt a variety of positions increasing the probability of
them forming hydrogen bonds with the N-3 of the
purine, the urea carbonyl of the pyrimidine bases (H-
bond acceptors), or the 2-amino of guanine (H-bond
donor) which point into the minor groove. The primary
advantage of the level of hydroxylation in poly-
hydroxyazepanes is their high water solubility, allowing
them to circumvent the problem of poor bioavailability
seen with many other MGBL’s. The chirality of the
polyhydroxyazepanes can also be controlled, allowing
access to a range of diastereomers, and possibly leading
to improved sequence selectivity.
A
survey of established DNA-binding antibiotics
reveals that a significant number have minor groove
binding elements composed of oligosaccharides. One
well studied example is the calicheamicin gI1 aryl tetra-
saccharide that has been shown to bind specifically to
homopurine-homopyridine tracts such the TCCT
sequence.5 Kahne and coworkers have demonstrated
that methyl [(1-4)-(3-N-((2-guanidino)acetamido))-2,3-
dideoxy-a-l-fuco-pyranoside]5, a rationally designed
oligosaccharide, functions as a DNA minor groove
binding ligand (MGBL) with a dissociation constant of
In this preliminary study, bis-azepanes that share the
dicationic nature and core structure of propamidine
*Corresponding author. Tel.:+44-115-951-3565; fax:+44-115-951-
3564; e-mail: neil.thomas@nottingham.ac.uk
Figure 1.
0960-894X/02/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(01)00719-3

238
H. A. Johnson, N. R. Thomas / Bioorg. Med. Chem. Lett. 12 (2002) 237–241
have been prepared, allowing a direct comparison of
their DNA binding abilities and biological activity.
propylidene-d-mannitol 5 was treated with trimethyl-
orthoacetate in the presence of pyridinium para-toluene
sulphonate. The orthoesters produced were then con-
verted to a-bromoacetates by treatment with acetyl-
bromide and triethylamine in dichloromethane. The
acetyl esters were then hydrolyzed with potassium car-
bonate, causing spontaneous cyclisation to give the bis-
epoxide 6a. Overall this gave 1,2:5,6-dianhydro-3,4-
O-isopropylidene-d-mannitol 6a in 72% yield from
d-mannitol. The alternative C2-symmetric bis-epoxide,
1,2:5,6-dianhydro-3,4-O-isoproptylidene-l-iditol 6b was
prepared from d-mannitol following the route described
by Depezay et al. (Scheme 2).12 Conversion of the pri-
mary hydroxyls of 5 to their benzoyl esters and the sec-
ondary alcohols to their tosylates was followed by ester
hydrolysis and cyclisation with inversion at C-2 and C-4
to give 6b in an overall yield of 47% from d-mannitol.
Synthesis
Two families of ligands have been prepared based on
1,3-bis-[alkyl-N-p-phenoxy] and 1,3-bis-[alkyl-N-p-ben-
zyloxy]propanes. These would be expected to differ sig-
nificantly in the pKa of their azepane nitrogens (pKa
5.5–6.0 for a N,N-dialkylphenylamine; 7.3–7.8 for a
N,N-dialkyl-p-methoxybenzylamine; compared to 11.4
for p-methoxybenzamidine) and the linear flexibility of
the ligands. The synthetic route initially investigated
(Scheme 1) was a highly convergent one in which the
central 1,3-bis-[p-aminophenoxy] or 1,3-bis-[4-amino-
methylphenoxy] propane core 3 or 4 was reacted with
the appropriate bis-epoxide derived from d-mannitol or
l-iditol to produce the required bis-azepane.
The bis-epoxides 6a and 6b were then reacted with
either 1,3-bis-(4-aminophenoxy)propane 3 or 1,3-bis-(4-
aminomethyl phenoxy)propane 4. However, in neither
case was the required bis-azepane isolated. Rather than
undergoing the second epoxide opening through a 7-
endo-tet-type aminocyclisation to generate the azepane,
the second epoxide appeared to have reacted with
another 1,3-bis-(4-aminoaryloxy)propane molecule lead-
ing to polymer formation. Changes in the reaction con-
ditions and the ratio of the bis-epoxide to bis-amine did
not give any of the required product. An alternative
strategy as shown in Scheme 3 was therefore investigated.
3,4,5,6-Tetrahydroxyazepanes 7a and 7b were synthe-
sized from 4-acetoxyphenylamine and 1,2:5,6-dianhydro-
3,4-O-isopropylidene-d-mannitol 6a or d-iditol 6b,
respectively, by heating in water. The protected azepanes
were then treated with 2 M NaOH in methanol to
remove the acetyl group. The bis-azepanes were then
obtained by treatment of the phenol 8a or 8b with
potassium carbonate in dry ethanol followed by addi-
tion of 1,3-dibromopropane. The resulting mixture was
heated under reflux for 4 days and the required di-iso-
propylidene bis-azepane 9a or 9b was obtained by column
chromatography in 40 or 27% yield, respectively. The
isopropylidene groups were removed by treatment with
0.5 M hydrochloric acid giving 1,3-bis-[(3R,4R,5R,6R)-
3,4,5,6 - tetrahydroxy - azepane - N - p - phenoxy]propane
dihydrochloride 10a, and the (3S,4R,5R,6S)-diastereo-
isomer 10b as beige solids.
The formation of tetrahydroxyazepanes from bis-epox-
ides by a nucleophilic opening/aminocyclisation has
previously been described by Depezay,7 Lohray8 and
Wong.9,10 To investigate the effects of different stereo-
chemistry on DNA binding two different C2-symmetric
bis-epoxides were selected. The common intermediate,
3,4-O-isopropylidene-d-mannitol 5, was prepared by
converting d-mannitol to its triacetonide using dimeth-
oxypropane and a catalytic amount of toluene sul-
fonic acid, followed by removal of the terminal
acetonides using acidic aqueous methanol.10 We found
that 1,2:5,6-dianhydro-3,4-O-isopropylidene-d-manni-
tol 6a required for the formation of the bis-azepanes 10a
and 13a was most efficiently produced using a route
devised by Lohray et al. (Scheme 2).11 3,4-O-Iso-
The benzyl bis-azepane homologues, 1,3-bis-[(3R,4R,
5R,6R)-3,4,5,6-tetrahydroxyazepane-N-p-benzyloxy]-
Scheme 1. Proposed route to bis-azepanes.
Scheme 2. (a) 2,2-DMP, pTsOH, 95%; (b) 1:10 MeOH/H2O, pTsOH, 75%; (c) (i) MeC(OMe)3, PPTS, DCM; (ii) Et3N, AcBr, DCM; (iii) K2CO3,
MeOH, 95%. (d) BzCl, pyr, DCM, 72%; (e) TsCl, pyr, 100%; (f) K2CO3, MeOH, DCM, 85%.

H. A. Johnson, N. R. Thomas / Bioorg. Med. Chem. Lett. 12 (2002) 237–241
239
propane dihydrochloride 13a and the (3S,4R,5R,6S)-
diastereoisomer 13b were produced by a similar route
except that 4-allyloxybenzylamine was used in place of
4-acetoxyphenylamine because the reactions with the
bis-epoxides were cleaner and the allyl groups could be
removed easily prior to coupling with 1,3-dibromopro-
pane as shown in Scheme 4. All intermediates and final
products were fully characterized by NMR (1H, 13C,
DEPT), HRMS and elemental analysis.
1. MCF7/ADR is a mutant form of the breast cancer
cell line that is resistant to doxorubicin.
The results shown in Table 1 demonstrate that three of
the bis-azepane derivatives are more effective than
chlorambucil, of similar activity to propamidine and
much less effective than doxorubicin in inhibiting the
growth of the colon cancer cell lines. One of the aceto-
nide protected bis-azepanes 12a is a better growth inhi-
bitor than any of the drugs in the doxorubicin resistant
breast cancer cell line, although all compounds tested
only had modest activity in this case. As the bis-azepane
with the best DNA binding affinity 13a has the poorest
inhibitory activity of those tested, it is unlikely that this
biological activity is due to DNA binding, unless the
acetonide groups on 12a aid transport across the cell
membrane and are then hydrolyzed within the cell to give
compound 13a. In comparison, compound 13a may have
very poor cellular uptake due to its increased polarity.
Biological Activity
3,4,5,6-Tetrahydroxyazepanes have been shown to be
good competitive inhibitors of glycosidase enzymes
from plant origin,13 their 2,5-dialkyl derivatives to have
some activity as competitive inhibitors of HIV and FIV
proteases,14 malto-oligosaccharides containing tetra-
hydroxyazepanes have been shown to inhibit human
a-amylase.15 Lohray et al. have shown that
(3S,4R,5R,6S)-tetrahydroxyazepane and N-benzyl (3R,
4R,5S,6R)-tetrahydroxyazepane have modest activities
against several cancer cell lines (GI90 20À90 mM) with
the highest activity being observed against the SNB-75
CNS cancer cell line.16 The site of action in this case has
not been not determined.
Binding Studies
To determine the effectiveness of bis-azepanes as DNA
binding agents, both thermal denaturation18 and ethidium
bromide displacement studies19 have been conducted.
We have tested propamidine 2 together with the bis-
azepanes and their acetonide precursors against a num-
ber of cell lines in in vitro biological tests. Literature
GI50 data are also given on doxorubicin (intercalator
which induces topoisomerase II mediated cleavage) and
chlorambucil (a DNA alkylator) as benchmark
agents.17 Data for the compounds found to be active
against HCT116 and HT29 human colon cancer, and
MCF7 human breast cancer cell lines are given in Table
Thermal denaturation
A comparison of thermal denaturation profiles of DNA
in the presence and absence of ligand provides the sim-
plest means of detecting binding, and also ascertaining
relative binding strength. Melting curves were measured
at 260 nm following a modification of the method
described by Cory for the investigation of pentamidine
analogues.18 Whilst amidine groups would be protonated
Scheme 3. (a) H2O, 95 ꢁC, 40%; (b) 2 M NaOH, MeOH, 75%; (c) Br(CH2)3Br, K2CO3, EtOH, reflux, 4 days, 40%; (d) 0.5 M HCl, 95%.
Scheme 4. (a) H2O, 60 ꢁC, 69%; (b) t-BuOK, DMSO 100 ꢁC, 81%; (c) HgO, HgCl2 in 10:1 acetone/water, 87%; (d) K2CO3, Br(CH2)3Br, EtOH,
reflux for 4 days, 31%; (e) 0.5 M HCl, 100%.

240
H. A. Johnson, N. R. Thomas / Bioorg. Med. Chem. Lett. 12 (2002) 237–241
Table 1. Activities against colon and breast cancer cell linesa
Compd
HCT116 GI50 (mM)b
HT29 GI50 (mM)b
MCF7 GI50 (mM)b
MCF7/ADR GI50 (mM)b
Propamidine 2
Doxorubicinc
Chlorambucilc
9a
12a
13a
11 (Æ2.8)
0.023
105
32 (Æ5.0)
6 (Æ0.6)
>50
>50
0.056
129
40 (Æ1.0)
23 (Æ2.5)
24 (Æ0.6)
9 (Æ3.6)
0.013
30
33 (Æ2.1)
16.7
56
>50
24 (Æ9.0)
>50
10 (Æ1.5)
>50
>50
aValues are means of three experiments, standard deviation is given in parentheses (50 mM was maximum concn tested).
bIn vitro GI50 (concn required to inhibit growth by 50%).
cValues from ref 17.
and, therefore, positively charged at pH 7.50, this would
not be true of the phenylamine derivatives 10a/b
and possibly not of the benzylamine derivatives 13a/b
used in our study. We therefore conducted the thermal
denaturation studies at both pH 7.5 and pH 5.0. Calf
thymus DNA (50 mM) (Sigma D-1501) was taken upin
a buffer solution composed of 5 mM TRIS (pH 7.5) or
10 mM sodium citrate/sodium phosphate (pH 5.0), 50
mM EDTA, 10 mM sodium chloride and 0.25% DMSO.
This gave melting temperatures of the DNA of 69.5 ꢁC
at pH 7.5 and 67.2 ꢁC at pH 5.0. Table 2 summarizes the
changes observed on adding either propamidine or one
of the ligands under investigation (5 mM). It can be seen
that propamidine stabilises the DNA duplex to give an
increase in the melting temperature of+8.2 ꢁC at p H
7.5, falling to+6.3 ꢁC at pH 5.0, whilst none of the bis-
azepanes gave any apparent stabilisation at pH 7.5
within experimental error, and only one, 13a gave a
slight increase at pH 5.0. Even at this lower pH, it is
possible that the majority of nitrogens of the phenyl-
amine ligands are still unprotonated. The studies were
repeated leaving the ligands to equilibrate for 12 h with
the DNA prior to determining the denaturation tem-
perature. In this case, several of the ligand–DNA mix-
ture melting temperatures were elevated indicating that
some stabilisation of the DNA duplex structure was
occurring, but again, not to the extent of that seen with
propamidine. The polyhydroxylated bis-azepanes are
significantly broader than propamidine, and they may
exhibit slower DNA binding kinetics (milliseconds
rather than microseconds). This would not fully account
for the increases in melting temperature observed. Most
minor groove binding ligands including propamidine
and pentamidine have a preference for binding AT rich
DNA sequences.1 It was therefore decided to repeat the
thermal denaturation studies using both poly(-
dA).poly(dT) and poly(dG–dC)2 at pH 5.0. Whilst
poly(dG–dC)2 did not melt below 95 ꢁC in our buffer
system, poly(dA).poly(dT) had a melting temperature of
53.3 ꢁC. The increase in melting temperature on addi-
tion of ligand were found to be +25.7 ꢁC for
propamidine;+0.8 ꢁC for 13a; and +0.5 ꢁC for 13b;
other compounds did not affect the melting tempera-
ture. In this case, preincubation of the ligand with the
DNA for 12 h did not effect the DNA melting tem-
perature.
Overall, the thermal denaturation studies indicated that
one of the five bis-azepane ligands, 13a had a stabilizing
affect on the DNA, and this was very weak compared
with that of propamidine 2.
Ethidium bromide displacement
Using a procedure developed by Cain et al.19 it is pos-
sible to measure the relative binding affinity of a ligand
for DNA by determining the C50 value of the ligand.
This is the concentration of the ligand required to dis-
place 50% of the ethidium from a preformed, inter-
calated complex. It is also possible to calculate the
apparent binding constant for a ligand given that the
binding constant for ethidium [Ke=9.5Â10À6 M(bp)À1
]
is known under the conditions used. Table 3 sum-
marizes the results for those ligands that did affect the
fluorescence due to the intercalated ethidium at pH 7.0.
One of the acetonide-protected bis-N-phenylazepanes 9a
had a low affinity for the DNA, whilst the two unpro-
tected N-benzylazepanes 13a and 13b displayed apparent
binding affinities of the same order of magnitude as that
of propamidine. None of the other compounds tested,
9b, 10a/b or 12a/b, caused a change in fluorescence.
Table 2. Change in melting temperature of calf thymus DNA
Table 3. Ethidium bromide displacement assays
Ligand
pH 5.0a,b,c
(ꢁC)
pH 5.0a,b,d
(after 12 h) (ꢁC)
pH 7.5a,b,c
(ꢁC)
Ligand
C50 (mM)a,c
Kapp (MÀ1 b
)
4.7Â105
1.7Â104
1.9Â105
1.2Â105
6.7Â106
Propamidine 2
9a
12a
10a
13a
13b
+6.3
+0.1
+0.1
0.0
+0.3
+0.2
+6.8
+0.4
+0.5
+0.7
+1.2
—
+ 8.2C
- 0.1
+ 0.1
+ 0.1
+ 0.1
—
Propamidine 2
9a
13a
13b
Distamycind
25.6 (Æ1.5)
703 (Æ161)
61.6 (Æ5.3)
97.0 (Æ7.9)
1.8
aValues are means of three experiments, standard deviation is given in
parentheses.
aValues are means of three experiments.
b50 mM DNA, 5 mM ligand, 10 mM NaCl, 50 mM EDTA, 0.25%
DMSO.
bKapp=(1.26ÂKe)/C50: ethidium, Ke=9.5Â10À6 M(bp)À1
.
c1 mM DNA, 1.26 mM ethidium, 2 mM HEPES, 9.4 mM NaCl, 10 mM
EDTA.
c5 mM Tris.
d10 mM sodium phosphate–sodium citrate.
dFrom ref 20.

H. A. Johnson, N. R. Thomas / Bioorg. Med. Chem. Lett. 12 (2002) 237–241
241
Conclusions
4. Wood, D. H.; Hall, J. E.; Rose, B. G.; Tidwell, R. R. Eur.
J. Pharmacol. 1998, 353, 97.
5. (a) Walker, S. L.; Andreotti, A. H.; Kahne, D. E. Tetra-
hedron 1994, 50, 1351. (b) Bifulco, G.; Galeone, A.; Nicolaou,
K. C.; Chazin, W. J.; Gomez-Paloma, L. J. Am. Chem. Soc.
1998, 120, 7183.
6. Xuereb, H.; Maletic, M.; Glidersleeve, J.; Pelczer, I.;
Kahne, D. J. Am. Chem. Soc. 2000, 122, 1883.
7. Poitout, L.; Le Merrer, Y.; Depezay, J.-C. Tetrahedron
Lett. 1994, 35, 3293.
8. Lohray, B. B.; Jayamma, Y.; Chatterjee, M. J. Org. Chem.
1995, 50, 5958.
9. Qian, X.; Moris-Varas, F.; Wong, C.-H. Bioorg. Med.
Chem. Lett. 1996, 6, 1117.
This preliminary study has demonstrated that bis-tetra-
hydroxyazepanes do have modest growth inhibitory
activity in cancer cell lines and are capable of binding to
DNA at low pH, although not with the affinity of
ligands such as propamidine. Future research will
involve conducting DNA footprinting studies on com-
pound 13a, and the preparation of bis-azepanes with a
reduced number of hydroxyl groups, or with the
hydroxyls modified as their methyl ethers, which may
exhibit better minor groove binding abilities.
10. Qian, X.; Moris-Varas, F.; Fitzgerald, M. C.; Wong, C.-
H. Bioorg. Med. Chem. 1996, 4, 2055.
11. Lohray, B. B.; Jayamma, Y.; Chatterjee, M. Synth. Com-
mun. 1997, 27, 1711.
Acknowledgements
H.A.J. would like to thank the EPSRC for a student-
ship. N.R.T. would like to thank the Royal Society for a
University Research Fellowship. Chetna Modi and
Charles S. Matthews of the Cancer Research Labora-
tory at Nottingham are thanked for their assistance
with the thermal denaturation and biological activity
studies, respectively.
12. Le Merrer, Y.; Dureault, A.; Greck, C.; Micas-Languin,
D.; Gravier, C.; Depezay, J.-C. Heterocycles 1987, 25, 541.
13. Le Merrer, Y.; Poitout, L.; Depezay, J.-C.; Dosbaa, I.;
Geoffroy, S.; Foglietti, M.-J. Bioorg. Med. Chem. 1997, 5, 519.
14. Qian, X.; Moris-Varas, F.; Fitzgerald, M. C.; Wong, C.-
H. Bioorg. Med. Chem. 1996, 4, 2055.
15. Uchida, R.; Nasu, A.; Tokutake, S.; Kasai, K.; Tobe, K.;
Yamaji, N. Chem. Pharm. Bull. 1999, 47, 187.
16. Lohray, B. B.; Bhushan, V.; Prasuna, G.; Jayamma, Y.;
Raheem, M. A.; Papireddy, P.; Umadevi, B.; Premkumar, M.;
Lakshmi, N. S.; Narayanareddy, K. Indian J. Chem., Sect. B
1999, 38, 1311.
References and Notes
1. (a) Henderson, D.; Hurley, L. H. Nature Med. 1995, 1, 525.
(b) Thurston, D. E. Br. J. Cancer 1999, 80, 65.
2. (a) Reddy, B. S. P.; S. Sondhi, M.; Lown, J. W. Pharma-
col. Ther. 1999, 84, 1. (b) Neidle, S. Nat. Prod. Rep. 2001, 18,
291.
3. Hopkins, K. T.; Wilson, W. D.; Bender, B. C.; McCurdy,
D. R.; Hall, J. E.; Tidwell, R. R.; Kumar, A.; Bajic, M.; Boy-
kin, D. W. J. Med. Chem. 1995, 38, 912.
17. AntiCancer Agent Mechanism Database: http://dtp.nci.-
nih.gov/docs/cancer/searches/standard_mechanism.html
18. Cory, M.; McKee, D. D.; Kagan, J.; Henry, D. W.;
Miller, J. A. J. Am. Chem. Soc. 1985, 107, 2528.
19. Cain, B. F.; Baguley, B. C.; Denny, W. A. J. Med. Chem.
1978, 19, 1101.
20. Braithwaite, A. W.; Baguley, B. C. Biochemistry 1980, 19,
1101.