DNA binding, solubility, and partitioning characteristics of extended lexitropsins

Article abstract of DOI:10.1021/jm990620e

Four new ligands that bind to the minor groove of DNA have been designed, synthesized, and evaluated by DNA footprinting. Two of the ligands are polyamides containing central regions with five or six N-methylpyrrole units, conferring hydrophobicity and go

Full text of DOI:10.1021/jm990620e

J . Med. Chem. 2000, 43, 3257-3266
3257
DNA Bin d in g, Solu bility, a n d P a r tition in g Ch a r a cter istics of Exten d ed
Lexitr op sin s
Robert V. Fishleigh,† Keith R. Fox,‡ Abedawn I. Khalaf,§,| Andrew R. Pitt,§ Martin Scobie,§,| Colin J . Suckling,§
J ohn Urwin,§,| Roger D. Waigh,*,| and Stephen C. Young†
Proteus Molecular Design Ltd., Proteus House, Lyme Green Business Park, Macclesfield, Cheshire SK11 0J L, England,
Department of Physiology and Pharmacology, University of Southampton, Bassett Crescent East,
Southampton SO16 7PX, England, Department of Pure and Applied Chemistry, University of Strathclyde,
Glasgow G1 1XL, Scotland, and Department of Pharmaceutical Sciences, Strathclyde Institute for Biomedical Sciences,
27 Taylor Street, Glasgow G4 0NR, Scotland
Received December 20, 1999
Four new ligands that bind to the minor groove of DNA have been designed, synthesized, and
evaluated by DNA footprinting. Two of the ligands are polyamides containing central regions
with five or six N-methylpyrrole units, conferring hydrophobicity and good binding affinity
but without retaining the correct spacing for hydrogen bonding in the base of the minor groove.
The two remaining ligands have central regions which are head-to-head-linked polyamides, in
which the linker is designed to improve the phasing of hydrogen bonding of the ligand with
the floor of the minor groove. The highest affinity was obtained with the two polypyrroles
without headgroup spacers, indicating that H-bond phasing is secondary in determining affinity
compared to the major hydrophobic driving force. With a dimethylaminoalkyl group, represent-
ing a moiety with modest base strength, at both ends, water solubility is good and pH-partition
theory predicts that penetration through lipid membranes will be enhanced, compared to
strongly basic amidine analogues of the alkaloid precursors. All four compounds bind to DNA,
with strong selectivity for AT sequences but some tolerance of GC base pairs and subtle
individual preferences. The data show that very high affinities can be anticipated for future
compounds in this series, but drug design must take account of overall physicochemical
properties as well as the details of hydrogen bonding between ligands and the floor of the
minor groove.
In tr od u ction
An important advance was made in ligand design4
when the N-methylpyrrole units of the natural products
were replaced with other five-membered heterocyclic
rings, in particular imidazole, which changed the se-
quence-binding preference from AT to GC. A combina-
tion of these units offers the possibility of designing
molecules to bind preferentially to any given double-
stranded DNA sequence. Additionally, it is necessary
to extend the length of the molecule to read considerably
more than the four base pairs which bind the natural
products, if a useful level of selectivity is to be achieved,5
which may require the use, for example, of bifunctional
linkers6 to join chains of amide-bonded heterocycles. A
significant variation on this approach allows the chain
to double back on itself.7 Such a ‘hairpin’ structure fills
the minor groove and reads the sequence information
more efficiently than a single chain but requires the
construction of a considerably larger molecule, up to
double the length of the sequence which is being read.
The attractions of drug design where the ligand is
directed at the primary coding sequences of DNA are
many. Such a ligand, should it be sufficiently sequence-
selective, would offer the prospect of gene control in
diseases where genetic malfunction is the primary
cause, such as cancer. In addition, there are exciting
possibilities for the treatment of parasitic diseases,
where the DNA possesses sequences which do not occur
in the host.
Sequence selectivity is not a major feature of the early
type of intercalating ligands, such as ethidium, but
much attention has been devoted in recent years to the
development of ‘lexitropsins’, which bind to the minor
groove of DNA and thus have a greater chance of
reading the information which is provided by the base-
pair sequence as part of the usual process of transcrip-
tion. It was found quite early in these studies that the
natural polypyrroles, such as netropsin (1) and dista-
mycin (2), showed almost total selectivity for AT rather
than GC sequences.1 Very recently, it has been shown
that these compounds show a degree of selectivity for
some specific AT sequences over others.2,3
The central part of the ligand must be hydrophobic,
since the main driving force for binding is the expulsion
of water from the minor groove, itself a largely hydro-
phobic environment with a capacity for hydrogen bond-
ing at the bottom of the groove. A recent thermodynamic
study8 has shown that binding in the minor groove of a
DNA dodecamer by Hoechst 33258 is dominated by
hydrophobic forces; clearly, new ligands must have large
nonpolar regions if they are to show high affinity. A
major concern in drug design, therefore, is the water
* To whom correspondence should be addressed. Tel: 0141-548-
4355. Fax: 0141-552-6443. E-mail: r.d.waigh@strath.ac.uk.
† Proteus Molecular Design Ltd.
‡ University of Southampton.
§ University of Strathclyde.
| Strathclyde Institute for Biomedical Sciences.
10.1021/jm990620e CCC: $19.00 © 2000 American Chemical Society
Published on Web 08/03/2000

3258 J ournal of Medicinal Chemistry, 2000, Vol. 43, No. 17
Fishleigh et al.
solubility of a ligand composed largely of nonpolar
subunits and, if this is overcome by the use of more than
one basic end group, the effect that this will have on
the ability of the molecule to penetrate lipid membranes
by passive diffusion.
The application of pH-partition theory to drugs with
a lipophilic structure and one basic group is well-
established;10 it is a combination of the Henderson-
Hasselbach treatment of the ionization of weak acids
and bases with the concept of the partition coefficient
(P) between two immiscible solvents, giving eq 1, for a
compound with one basic center:
D ) P/(1 + 10(pK -pH)
)
(1)
a
In this equation, D is defined as the ratio of the
concentration of nonionized material in the organic
(lipid) phase to the concentration of the combined
ionized and nonionized material in the aqueous phase,
at equilibrium. Since, for a given compound P and pKa
are constant, it is possible to plot log D against pH.
Where insight is sought into the probable behavior of a
molecule in the body, the region of greatest interest is
that around pH 7.4, the pH of most body tissues. At pH
7.4 molecules such as 4-7 are expected to be doubly
charged, in which case the pH-partition behavior is
described by a special form of eq 1, taking account of
the existence of two basic centers, here assumed to have
the same value of pKa, and a doubly charged species
(eq 2). Where the two pKa values are close, but not
identical, this form of the equation can be regarded as
an adequate approximation:
To test these concepts, we have synthesized four new
ligands 4-7, which combine building blocks known to
give good DNA binding, and have correlated their DNA
binding with their calculated partitioning properties.
Ligands 4 and 5 are extended polyamides, with long
chains of nonpolar N-methylpyrroles rendered more
soluble by dimethylamino groups at both ends. Ligands
6 and 7 are both head-to-head-linked polypyrroles, with
a linker which has been designed to improve H-bonding
in the base of the minor groove: in both cases the linker
is more hydrophilic than the polypyrrole chains. The
dimethylamino end groups of structures 4-7 were
chosen for three reasons. First, they have been used
before9 and have been shown to be consistent with good
DNA binding. Second, the synthetic methods for their
incorporation are well-established, and third, we wished
to use tertiary bases, which would have lower pKa
values than amidines, guanidines, or secondary bases.
This is highly significant when pH-partition theory is
applied to these molecules.
D ) P/(1 + 2(10(pK -pH)) + 10(2(pK -pH))
)
(2)
a
a
Since the introduction of two charged groups might
influence DNA binding, we compared the binding char-
acteristics of compounds 4-7 using the footprinting
technique.
Discu ssion
DNA Bin d in g. We used DNase I and hydroxyl
radical footprinting to examine the interaction of these
ligands with natural and synthetic DNA fragments.
With TyrT DNA (Figure 1), a natural DNA fragment
which has been widely used in footprinting studies, we
found that each of the compounds altered the digestion

Characteristics of Extended Lexitropsins
J ournal of Medicinal Chemistry, 2000, Vol. 43, No. 17 3259
lower concentrations. The pattern produced by 1 µM 5
was similar to that generated by 4 at low concentrations,
except that the footprint around position 65 was more
clearly defined, while the footprint around position 25
extended over one or two more bases in the 3′-direction,
consistent with its larger size. At higher concentrations
(3 µM and above) hydroxyl radical cleavage was almost
completely abolished. An example of a differential
cleavage plot, derived from densitometer scans of the
data for 1 µM compound 5 only, is presented in Figure
2A, revealing three distinct regions of protection be-
tween positions 20-70. Three clear footprints were also
evident with 0.3 µM 7 around positions 25-30, 46-50,
and 81-86, as observed with the other ligands, except
that there was little protection around position 65. With
1 µM of ligand 7 new sites of protection were evident,
corresponding to the pattern produced with higher
concentrations of 4. Compound 7 showed five binding
sites between positions 20-77. The first and third
(around positions 25 and 45) correspond to long AT
tracts, while the latter (around position 65-70) contains
the sequence TGTAAA. The other two sites are harder
to define rigorously but appear to contain the sequences
AACTGGTT (positions 37-44) and TTAGCTT (55-61).
The changes with 1 µM 6 were less clear, with only one
clear footprint at positions 27-31, with attenuated
cleavage around positions 46-50 and 81-86.
F igu r e 1. Sequences of the various DNA fragments used in
this work. Only the strand bearing the radioactive label
(underlined) is shown. TyrT is an EcoR1-Ava1 fragment. The
numbering system used for this DNA is the same as that used
in previous publications.11-3 For p(A/T)6-12, pA3-6, p(A/T)10, and
p(A/T)6 only the sequence of the insert (which was cloned into
the BamH1 site of pUC18) is shown. Every tenth base of the
TyrT fragment is numbered for easy reference.
These results confirm that the compounds are indeed
AT-selective, with larger binding sites than the parent
antibiotic distamycin. In addition they suggest that at
moderate concentrations these compounds can bind to
some sites which contain 1 or 2 GC base pairs and that
the various derivatives have subtly different sequence
binding properties. We have therefore investigated their
interaction with several synthetic DNA fragments con-
taining different length AT tracts and various arrange-
ments of AT base pairs.
pattern at concentrations of 0.1-3 µM. In each case
reductions in cleavage were seen around positions 25,
60, and 80, each of which corresponds to an AT-rich site
which has previously been shown to be affected by other
minor groove-binding ligands. These patterns were not
easy to interpret since DNase I cleavage in the control
was very uneven and several regions of poor cleavage
in the control were located in the AT-rich regions around
positions 25-30 and 45-50, which are expected to be
good ligand-binding sites. We therefore repeated the
experiments using hydroxyl radicals as the footprinting
probe. These produced a much more even ladder of
cleavage products in the control, enabling a more precise
determination of the ligand-binding sites. Low concen-
trations of compound 4 (0.3 and 1 µM) produced attenu-
ated cleavage around positions 25-30, 46-50, 61-65,
and 81-86. The first two are located toward the 3′-side
of long AT tracts (7 and 6 bases, respectively), while the
third contains only 4 consecutive AT pairs. However, it
may be significant that this lies within an AT-rich
region; the 13 AT pairs between positions 59-71 contain
only 3 GC pairs. It is possible that this represents ligand
binding to a longer site (possibly TGTAAA at 65-70),
tolerating a single guanine residue. Each of these
footprints was about 6 bases long, compared with 3-4
bases observed with distamycin,11 confirming the larger
ligand-binding site. At higher concentrations (3 µM and
above) the footprinting pattern changed and more
regions were protected from cleavage. Most notable was
a new footprint around position 40 (AACTGGTT). The
regions protected around positions 60 and 80 also
increased in size and appeared to split into two foot-
prints located on either side of the ones produced at
Fragment pAAD (not shown) contains 5 (A/T)4 sites,
each separated by 6 GC base pairs, and has previously
been used to determine AT selectivity of several minor
groove-binding ligands.2 None of compounds 4-7 af-
fected either DNase I or hydroxyl radical cleavage of
this fragment at concentrations up to 3 µM, though
higher concentrations abolished cleavage throughout the
fragment. This confirms that, in this fragment, 4 con-
secutive AT base pairs were not sufficient to constitute
a specific ligand-binding site.
Figure 3 shows DNase I cleavage of fragment pA3-6
which contains 4 An/Tn tracts of various lengths (n )
3-6). It can be seen that, with the exception of 3 µM 7,
cleavage of T3 was not affected by the ligands and
cleavage at T4 was only attenuated at the highest ligand
concentrations. Cleavage of T5 and T6 was inhibited at
lower ligand concentrations with T6 as the better site
in each case. Both 4 and 7 protected T6 at the lowest
concentration (0.3 µM), while T5 was not fully protected
below 1 µM. Similarly 5 protected T6 at 1 µM, while T5
required 3 µM. For 6, T6 was protected at the highest
concentration, while the other sites were hardly affected.
It therefore appears that the ligands bind best to sites
containing 6 consecutive AT base pairs, although there
is a significant interaction with sites containing 5 and
4 AT pairs. Hydroxyl radical cleavage patterns with this
fragment are not easy to interpret since digestion of the

3260 J ournal of Medicinal Chemistry, 2000, Vol. 43, No. 17
Fishleigh et al.
F igu r e 2. Differential cleavage plots showing the interaction of 1 µM 5 with the 5 different DNA fragments: A, TyrT DNA; B,
pA3-6; C, p(A/T)6-12; D, p(A/T)6; E, p(A/T)10. The data were obtained from hydroxyl radical cleavage patterns and were analyzed
as described in the text.
DNA was attenuated in each Tn tract in the absence of
the ligand, consistent with previous suggestions that
these sequences possess narrower than average minor
grooves. A differential cleavage plot comparing the
pattern in the presence of 1 µM 5 with that in the
control is shown in Figure 2B. With this compound
footprints were only produced at T5 and T6, each
covering 6-7 bases. 6 also produced clear footprints at
T6 and T5, and the protection at T4 and T3 coalesced into
a single footprint. Compounds 4 and 7 produced foot-
prints at all the Tn sites, each of which covers 5-6 bases.
However, examination of the pattern at T3 reveals that
the apparent protection is far to the 3′-side of the actual
site, suggesting that there was something unusual about
this interaction.
The results with pA3-6 confirmed that the compounds
bind best to longer AT tracts and showed that the
binding site, as revealed by hydroxyl radicals, covered
about 6 base pairs. We therefore extended these studies
to a fragment containing longer AT tracts. With hy-
droxyl radical cleavage of p(A/T)6-12, for which the
sequence13 is given in Figure 1, compounds 4 and 5
produced clear footprints covering 6-7 base pairs at the
sites containing 8, 10, and 12 base pairs but had little
effect on the upper site containing 6 AT base pairs
(AATATT) (Figure 3). A differential cleavage plot show-
ing the results for 5 is shown in Figure 2C. In contrast
7 produced footprints at all 4 sites, including AATATT.
However the footprint at the longest site was much
broader with this ligand affecting 8-9 base pairs. This
may be because the ligand does not adopt a unique
position within this long AT site but can assume one of
several locations each of which does not cover the entire
site. Surprisingly 1 µM 6 failed to show appreciable
protection at any of the sites. It should be emphasized
that each of these hydroxyl radical footprints is different
from that produced by the parent antibiotic distamycin,
which produces two distinct regions of protection within
the 8-, 10-, and 12-base pair sites, consistent with the
suggestion that two ligand molecules can simulta-
neously bind within each site.13
Previous studies have shown that, although distamy-
cin and other minor groove-binding ligands can bind to
all arrangements of 4 consecutive AT base pairs, some
sites are better than others.2 AATT is a particularly
good binding site, while the ligands bind less well to

Characteristics of Extended Lexitropsins
J ournal of Medicinal Chemistry, 2000, Vol. 43, No. 17 3261
with the larger binding site size for these ligands, so
that only one molecule is bound in each AT tract. Once
again, between 5-7 bases are protected at each (A/T)10
site, with much weaker binding at AATT. DNase I
cleavage of this fragment shows protection at all the
(A/T)10 sites but reveals that the ligands do not dis-
criminate between the different arrangements of AT
base pairs.
Analogues 4 and 5 both have central regions which
are simply extended polypyrroles. Above three pyrrole
units, the hydrogen-bonding ability of the amide links
ceases to be in-phase with the spacing on the floor of
the minor groove,6,14 resulting in a limit to the reading
ability. The good affinity of these two analogues is
therefore a reflection of the extra capacity for hydro-
phobic bonding in the minor groove. Their greater
length, compared to the natural ligands, accounts for
their preference for extended A/T base pair sequences.
The glycine linkers in 6 are spaced appropriately for
restoration of phasing with the H-bonding opportunities
in the minor groove and were designed to be tolerant
of, and perhaps selective toward, GC base pairs.15 In
this case, however, the hydrophilic nature of the amide
links may affect binding to DNA and also would
prejudice the chances of such a molecule reaching the
site of action in vivo. Compound 6 is capable of weakly
protecting short AT runs (not shown in the figures)
which may indicate that the molecule is only binding
at one end, with the central amide groups extending out
into water, or that the molecule is behaving as a
‘hairpin’ and folding back to give side-by-side filling of
the minor groove. To do this, at least one of the charges
on the end amino groups would have to be lost by
deprotonation.
F igu r e 3. DNase I footprinting patterns for the interaction
of compounds 4-7 with the fragments derived from pA3-6 and
p(AT)6-12. The concentration of the ligand (µM) is shown at
the top of each gel lane. ‘Con’ indicates the control, in the
absence of added ligand. Tracks labeled ‘GA’ are Maxam-
Gilbert sequencing lanes specific for purines. The position of
the various Tn blocks is indicated at the side of the autoradio-
graph.
Analogue 7 is a dimethylamino version of a head-to-
head-linked netropsin analogue previously reported to
display bidentate binding in (AT)n-rich sequences.16 This
relatively rigid linker has a length of 4.2 Å, close to the
4.4 Å required to maintain the ideal phasing of ligand
amide hydrogens. Although the central fumaramide unit
has the right length, modeling indicates a distortion in
the helicity of the molecule in the center, possibly
accounting for the slightly weaker binding observed
with this analogue.
sites containing TpA steps, especially TTAA and TATA.
We therefore performed similar experiments with a
fragment containing four different arrangements of 6
AT base pairs (A3T3, T3A3, (TA)3, and TAATTA) together
with AATT for comparison. The results of hydroxyl
radical footprinting experiments with 5 on this fragment
are shown in Figure 2D. Compounds 4, 5, and 7
produced DNase I footprints at each of the (A/T)6 sites,
while cleavage at AATT was not affected. The footprints
at these various sites appeared at different ligand
concentrations: cleavage at T3A3 and TAATTA was
abolished at 0.2 µM 7, while (TA)3 required higher
ligand concentrations (1 µM) to produce a footprint. No
data could be obtained for A3T3 since this region showed
no cleavage in the drug-free controls. The rank order of
binding sites for these ligands is T3A3 > TAATTA >
(TA)3. It therefore appears that the binding strength
decreases with increasing numbers of TpA steps. Com-
pound 6 only affected enzyme cleavage at concentrations
of 3 µM and above, though once again protection of (TA)3
was weaker. A differential cleavage plot for hydroxyl
radical cleavage of this fragment in the presence of
1 µM 5 is shown in Figure 2D. This confirms the
binding sites and reveals that between 5-7 residues are
protected at each position. Figure 2E shows a similar
differential cleavage plot for hydroxyl radical cleavage
of the fragment containing blocks of 10 AT base pairs
in the presence of 1 µM 5. These results are in contrast
to those with distamycin,13 which produces two distinct
footprints at each AT site. This difference is consistent
Ch em istr y. For the construction of polypyrroles with
a dimethylaminoalkyl end group, we used stepwise
addition of N-methyl-4-nitro-2-trichloroacetylpyrrole
units to amino-substituted mono- or polypyrroles with
dimethylaminopropylamino tails (Scheme 1). The amino
groups were generated by catalytic hydrogenation of the
nitro group of the growing chain, an approach which
was used to generate oligomers up to four rings in
length. Addition of further rings resulted in a rapid
decline in yield.
The syntheses of 4 and 5 were therefore based on the
coupling of a two-ring and a three- or four-ring unit in
a convergent fashion, in good yield (Scheme 1). This
route was designed to allow a flexible approach to the
synthesis of longer and more complex lexitropsins in the
future. Analogue 4 was constructed by coupling a three-
ring amine (Scheme 1) to a two-ring acid (Scheme 2),
using HBTU/Et3N, followed by HPLC purification. The
hexapyrrole 5 followed similarly, from a four-ring amine
and a two-ring acid.

3262 J ournal of Medicinal Chemistry, 2000, Vol. 43, No. 17
Fishleigh et al.
Sch em e 1a
a
Reagents: (i) NH2(CH2)3NMe2, rt; (ii) Pd/C/H2; (iii) fumaric acid, HBTU, NMM; (iv) HBTU, Et3N, DMF; (v) HBTU, NMM, DMF.
Sch em e 2a
Ta ble 1. Input Data and Calculated log Distribution
Coefficients for Extended Lexitropsins
b
compd
log Pa
pKa
log D (pH 7.4)c
1
2
3
4
5
6
7
-3.29
-0.87
2.17
3.89
4.68
12.4
12.4
9.5
9.5
9.5
-13.29
-5.87
0.07
-0.32
0.47
-2.74
-1.03
1.47
3.18
9.5
9.5
a
b
Calculated using the program in ref 17. Assumed figures.
c Calculated from eqs 1 and 2.
barriers. Calculated values of log P, assumed values of
pKa, and the resulting calculated values of log D at pH
7.4 for compounds 1-7 are given in Table 1.
If the amidine is replaced with a dimethylaminoalkyl
substituent, giving 3, which has been found to be
consistent with maintenance of DNA binding,19 the pKa
is reduced by about 2 log units, by analogy with a range
of compounds such as imipramine20 and log P is
increased by ca. 3 log units, giving the curve in Figure
4. Inspection of the curve shows that the behavior in
the critical region about pH 7.4 is markedly affected:
no longer is there such a strong preference for the
aqueous phase. Such an analogue would be expected to
show much improved bioavailability.
Relatively small molecules such as 1-3 would be
expected to be sufficiently water-soluble to be capable
of formulation as drugs. Longer DNA reading frames,
however, require long stretches of hydrophobic bases,
which by definition are likely to be less-water-soluble.
Introduction of a second dimethylaminoalkyl group at
the opposite end of the molecule to the first one, as in
4-7, offers the possibility of improvement in water
solubility.
a
Reagents: (i) Pd/C/H2; (ii) HOOC(CH2)3NMe2/HBTU/Et3N/
DMF; (iii) NaOH.
The isophthalic diamide 6 was synthesized in good
yield, by first treating isophthaloyl chloride with glycine
methyl ester to give 18, followed by hydrolysis and
coupling to 10 (R ) NH2), using HBTU/NMM (Scheme
1). Derivative 7 came directly from the coupling of the
two-ring amine 10 (R ) NH2), used in the previous
syntheses, with fumaric acid (Scheme 1).
P a r tition in g. Application of eq 1 to distamycin (2),
which has expected pKa ca. 12.4 and calculated17 log P
value of -0.87, gives the curve shown in Figure 4,
showing the effect of the very strongly basic amidine
moiety, cf. acetamidine.18 At pH 7.4, distamycin is
expected to favor the aqueous phase, with potential
difficulty in passing through a series of lipid membrane
The two basic centers are sufficiently far apart to
behave as individual units, with the same or very

Characteristics of Extended Lexitropsins
J ournal of Medicinal Chemistry, 2000, Vol. 43, No. 17 3263
in water (A) and acetonitrile 90%, water 10%, TFA 0.1% (B).
Elution commenced with 95% A, 5% B and finished with 100%
B. Accurate mass measurements of compounds 4-7 were
carried out at the EPSRC National Mass Spectrometry Service
Centre, University of Wales, Swansea, U.K.
3-(1-Met h yl-4-n it r op yr r ole-2-ca r b oxa m id o)d im et h yl-
a m in op r op a n e, 9. A literature procedure21 gave 9 in 90%
yield: mp 130-131 °C (lit.21 mp 129-130 °C).
3-[1-Meth yl-4-(1-m eth yl-4-n itr opyr r ole-2-car boxam ido)-
p yr r ole-2-ca r boxa m id o]d im eth yla m in op r op a n e, 10. A
literature procedure21 gave 10 in 78% yield: mp 191-192 °C
(lit.21 mp 190-191 °C).
3-{1-Meth yl-4-[1-m eth yl-4-(1-m eth yl-4-n itr op yr r ole-2-
ca r boxa m id o)p yr r ole-2-ca r boxa m id o]p yr r ole-2-ca r box-
a m id o}d im eth yla m in op r op a n e, 11. A literature proce-
dure21 gave 11 in 63% yield: mp 198 °C (lit.21 mp 136-137
°C; lit.22,23 mp 205-206 °C).
(2E)-N1,N4-Bis[5-({[5-({[3-dim eth ylam in opr opyl]am in o}-
ca r b on yl)-1-m et h yl-1H -p yr r ol-3-yl]a m in o}ca r b on yl)-1-
m et h yl-1H -p yr r ol-3-yl]-2-b u t en ed ia m id e
Bist r iflu or -
a ceta te, 7. 10, R ) NO2 (200 mg, 0.53 mmol), and Pd/C-10%
(200 mg) were suspended in 2-propanol (25 mL) and hydro-
genated for 4 h at room temperature. Filtration of the catalyst
over Kieselguhr under N2 followed by removal of the solvent
in vacuo gave the amine 10, R ) NH2, as an off-white solid
that was used without further purification. Fumaric acid (25
mg, 0.22 mmol), HBTU (190 mg) and N-methylmorpholine (70
mL) were dissolved in DMF (1 mL, dry), with stirring at room
temperature and then left for 30 min. The amine was dissolved
in DMF (1.5 mL, dry) and added to the reaction mixture at
room temperature with stirring. The reaction mixture was left
stirring at room-temperature overnight, then purified by
HPLC to give 7 as the bis-TFA salt (75 mg, 35% yield): 1H
NMR δ (DMSO-d6) 1.24-1.84 (4H, p, 2CH2), 2.79 (3H, s, NMe),
2.80 (3H, s, NMe), 3.05-3.10 (4H, m, 2CH2), 3.24-3.27 (4H,
m, 2CH2), 3.82 (6H, s, 2NMe, pyrrole), 3.68 (6H, s, 2NMe,
pyrrole), 6.93 and 6.94 (2H, d), 6.94 and 6.95 (2H, d), 7.09 (2H,
s), 7.16 and 7.17 (2H, d), 8.13-8.16 (2H, t, 2CONH exchange-
able), 9.32 (2H, broad 2TFA exchangeable), 9.92 (2H, s,
2CONH exchangeable), 10.47 (2H, s, 2CONH exchangeable);
ES-MS found (M + 1) 773.4216, calcd for C38H53N12O6 773.4211.
3-[1-Met h yl-4-(1-m et h yl-4-(1-m et h yl-4-{1-m et h yl-4-n i-
tr op yr r ole-2-ca r boxa m id o}p yr r ole-2-ca r boxa m id o)p yr -
r ole-2-ca r b oxa m id o)p yr r ole-2-ca r b oxa m id o]d im et h yl-
a m in op r op a n e, 12. A mixture of 11 (1.00 g, 2.01 mmol) and
Pd/C-10% (1.00 g) in ethanol (100 mL) was hydrogenated at
room temperature for 20 h. The catalyst was removed by
filtration through Kieselguhr and the filtrate evaporated under
reduced pressure to give the crude amine as a pale yellow solid
residue (772 mg, 82%). This amine (722 mg, 1.6502 mmol) was
dissolved in dry DMF (3.0 mL) and the solution cooled in an
ice bath and stirred while a solution of 8 (447 mg, 1.65 mmol)
in dry DMF (1.0 mL) was added slowly during a period of 5
min. The resultant mixture was allowed to come to room
temperature and stirred for 21 h. Water (30 mL) was added
and the oily precipitate was scratched until it solidified. The
solid was washed with water and, after drying in vacuo,
amounted to 745 mg. Crystallization from aqueous methanol
gave 12 (425 mg, 34% yield) as a yellow microcrystalline
solid: mp 192-195 °C (lit.24 mp 195 °C); ES-MS [M + 1] found
622.6, calcd for C29H36N10O6 621.7; 1H NMR (DMSO-d6) δ 1.61
(2H, m, CH2-CH2-CH2), 2.13 (6H, s, N(CH3)2), 2.24 (2H, t,
CH2-CH2-NMe2), 3.21 (2H, m, CONH-CH2-CH2), 3.80 (3H,
s, NMe), 3.85 (3H, s, NMe), 3.87 (3H, s, NMe), 3.97 (3H, s,
NMe), 6.83 (1H, d, ArH), 7.06 (2H, d, 2ArH), 7.19 (1H, d, ArH),
7.26 (1H, d, ArH), 7.29 (1H, d, ArH), 7.60 (1H, d, ArH), 8.09
(1H, t, CO-NH-CH2, exchangeable), 8.20 (1H, d, ArH), 9.91
(1H, s, CONH, exchangeable), 10.01 (1H, s, CONH, exchange-
able), 10.31 (1H, s, CONH, exchangeable).
F igu r e 4. pH-partition behavior of compounds 2, 3, 5, and 6
calculated from eqs 1 and 2.
similar pKa values, which by analogy20 we expect to be
in the range 9-10. Figure 4 shows the pH-partition
behavior of the most (5) and least (6) lipophilic of the
dibasic analogues described in the present study, cal-
culated from eq 2. Comparison of the data in Figure 4
shows the very marked effect of the second basic center,
in 5 and 6, in the region about pH 7.4. It is apparent
that such analogues can tolerate much larger lipophilic
‘loads’ in the central region; of these analogues, the
polypyrrole 5, with the highest calculated value of log
P, promises the kind of partitioning behavior consistent
with good bioavailability. All of the dibasic compounds
in the present study were readily water-soluble.
Overall, the significance of these results lies not only
with the sequence-reading ability of these particular
molecules but in the finding that water solubility can
be a designed attribute, without compromising the
physicochemical characteristics theoretically required
for tissue penetration and maintaining affinity for DNA.
It is becoming increasingly apparent, especially with the
larger molecules required for extended sequence recog-
nition in DNA, that drug design must take account of
solubility and tissue penetration as primary features,19
alongside receptor affinity and selectivity. In this con-
text, the insertion of extra amide links to change
hydrogen-bond phasing, as in 6, must be approached
with caution, since an amide makes a strong negative
contribution to the log P. However, if attention is paid
to the pH-partitioning characteristics of designed mol-
ecules, there is no reason lexitropsins cannot be de-
signed to read long lengths of DNA and also have
adequate bioavailability.
Exp er im en ta l Section
Ch em istr y. 1H NMR spectra were recorded at 400 MHz
on a Bruker AMX spectrometer. Electrospray mass spectra
were recorded on a Fisons VG Platform Benchtop LC-MS and
EIMS on a J EOL J MS-AX505HA mass spectrometer. HPLC
purifications were effected on a Vydac Protein and Peptide C18
column using a solvent gradient with two solvents: 0.1% TFA
E t h yl 1-Met h yl-4-(1-m et h yl-4-n it r op yr r ole-2-ca r b ox-
a m id o)p yr r ole-2-ca r boxyla te, 14. The nitro compound 13
(2.057 g, 10.39 mmol) was hydrogenated over 10% Pd/C (805
mg) in ethanol (200 mL) at room temperature for 2 h. The
catalyst was removed by filtration through Kieselguhr and the

3264 J ournal of Medicinal Chemistry, 2000, Vol. 43, No. 17
Fishleigh et al.
filtrate evaporated under reduced pressure to leave the crude
amine as a gray oil. This oil was dissolved in an ice-water
bath, while a solution of 19 (10.39 mmol, prepared24 from 1.766
g of the carboxylic acid, using 2.0 mL of thionyl chloride, and
8 mL of DME) in dry DME (10 mL) was added slowly over a
period of 5 min. The resultant mixture was stirred and allowed
to come to room temperature overnight then it was evaporated
under reduced pressure. The residue was stirred for 1 h with
N HCl (50 mL) and the solid filtered, washed well with ethanol
and dried in vacuo to give 14 as a yellow powder (1.377 g, 41%
yield): mp 232-234 °C (lit.26 mp 239-240 °C); IR (KBr) νmax
3358, 3131, 1692, 1663, 1572, 1437, 1391, 1317, 1253, 1121
ylation occurred to give 17. This decarboxylation product had
a very close HPLC retention time to that of 16, but separation
could be accomplished using flash column chromatography
over silica using EtOAc/MeOH/NH4OH as eluant as described
above. Compound 17 was thus isolated from one experiment
as an amber gum: ES-MS 332.1, calcd for C17H25N5O2 332.4
[M + 1]; 1H NMR (DMSO-d6) δ 1.79 (2H, m, CH2-CH2-CH2),
2.28 (2H, m, CH2), 2.34 (6H, s, NMe2), 3.38 (2H, m, CH2), 3.55
(3H, s, pyrrole NMe) 3.80 (3H, s, pyrrole NMe), 6.05 (1H, m,
ArH), 6.52 (1H, m, ArH), 6.81 (1H, d, ArH), 7.08 (1H, m, ArH),
7.13 (1H, d, ArH), 9.71 (1H, s, exchangeable, CONH), 9.86 (1H,
s, exchangeable, CONH).
cm-1; H NMR (DMSO-d6) δ 1.24-1.30 (3H, t, CH2CH3), 3.84
1
3-(1-Met h yl-4-(1-m et h yl-4-(1-m et h yl-4-(1-m et h yl-4-(1-
m eth yl-4-(4-d im eth yla m in obu tyr yl)a m in op yr r ole-2-ca r -
b oxa m id o)p yr r ole -2-ca r b oxa m id o)p yr r ole -2-ca r b ox-
am ido)pyr r ole-2-car boxam ido)pyr r ole-2-car boxam ido)di-
m eth yla m in op r op a n e Bistr iflu or a ceta te, 4. A mixture of
11 (151 mg) and Pd/C-10% (164 mg) in EtOH (15 mL) was
hydrogenated at room temperature for 13 h. The catalyst was
removed by filtration through Kieselguhr, and the filtrate
evaporated under reduced pressure to leave the crude amine
(128.5 mg, 90% yield) as a pale yellow solid. This material was
used at once without further purification. A mixture of the
carboxylic acid 16 (68.6 mg, 0.18 mmol, 1 equiv), HBTU (104.1
mg, 0.275 mmol, 1.5 equiv), Et3N (76 mL, 55.5 mg, 0.55 mmol,
3 equiv) and dry DMF (1.5 mL) was placed under N2 and
stirred at room temperature for 30 min. To the resultant clear
solution was added a solution of the above crude amine (128.5
mg, 0.275 mmol, 1.5 equiv) in dry DMF (1.0 mL), and the
mixture stirred at room temperature for 8 h. The resultant
suspension was purified by reverse-phase HPLC. Fractions
containing the product were frozen immediately after collection
and then freeze-dried to give the bis-TFA salt 4 as a fawn
powder (65.3 mg, 34% yield): ES-MS found [M + 1] 826.4476,
(3H, s, NMe), 3.94 (3H, s, NMe), 4.16-4.24 (2H, q, CH2CH3),
6.90-6.91 (1H, d), 7.42-7.43 (1H, d), 7.54-7.55 (1H, d), 8.18-
8.19 (1H, d), 10.26 (1H, s, CONH exchangeable).
Eth yl 1-Meth yl-4-(1-m eth yl-4-(4-dim eth ylam in obu tyr yl)-
a m in o)p yr r ole-2-ca r boxyla te, 15. A mixture of 14 (605 mg)
and Pd/C-10% (627 mg) in ethanol (60 mL) was hydrogenated
at room temperature for 16 h. The catalyst was removed by
filtration through Kieselguhr and the filtrate was evaporated
under reduced pressure to leave the crude amine as a buff
colored glassy foam (456 mg, 83% yield) which was used
without further purification. A mixture of 4-dimethylamino-
butyric acid hydrochloride (527 mg, 3.15 mmol, 2 equiv), HBTU
(1.1923 g, 3.15 mmol, 2 equiv), Et3N (1.31 mL, 6 equiv) and
dry DMF (7.0 mL) was placed under N2 and stirred at room
temperature for 30 min. Then to the stirred mixture was slowly
added a solution of the above amine (456 mg, 1.57 mmol, 1
equiv) in dry DMF (2.5 mL) over a period of 5 min. After
stirring at room temperature for 3 h the resultant mixture
was diluted with EtOAc (250 mL) then extracted with 10%
Na2CO3 (100 mL). The organic layer was washed with brine,
dried (MgSO4), and evaporated under reduced pressure to
leave a red-brown oil (1.036 g). The crude product was purified
by flash column chromatography over silica, using MeOH
containing 3% of concentrated ammonia as eluant, to give 15
as an amber oil (487 mg, 77% yield): ES-MS [M + 1] found
404.2, calcd for C20H29N5O4 404.5; 1H NMR (DMSO-d6) δ 1.24
(3H, t, O-CH2-CH3), 1.66 (2H, m, CH2-CH2-CH2), 2.08 (6H,
s, NMe2), 2.20 (4H, m, CH2-CH2-CH2), 3.77 (3H, s, pyrrole
NMe), 3.79 (3H, s, pyrrole NMe), 4.17 (2H, q, O-CH2-CH3),
6.80 (1H, d, ArH), 7.12 (1H, d, ArH), 7.36 (1H, d, ArH), 9.80
(1H, s, CONH, exchangeable), 9.90 (1H, s, CONH, exchange-
able).
1
calcd for C41H56N13O6 826.4476; H NMR (DMSO-d6) δ 1.82-
1.93 (4H, 2 × overlapping quintets, 2 × CH2-CH2-CH2), 2.35
(2H, t, CH2-CH2-CO), 2.79 (12H, m, 2 × NMe2), 3.07 (4H,
m, 2 × CH2-CH2-NMe2), 3.24 (2H, m, NH-CH2-CH2), 3.72-
3.86 (15H, 5 × pyrrole NMe), 6.87-7.22 (10H, 10 × ArH), 8.14
(1H, t, exchangeable CONH-CH2), 9.2-9.5 (2H, broad, ex-
changeable, 2 × TFA), 9.89-9.94 (5H, exchangeable, 5 ×
CONH).
3-(1-Met h yl-4-(1-m et h yl-4-(1-m et h yl-4-(1-m et h yl-4-(1-
m et h yl-4-(1-m et h yl-4-(4-d im et h yla m in ob u t yr yl)a m in o-
p yr r ole-2-ca r boxa m id o)p yr r ole-2-ca r boxa m id o)p yr r ole-
2-car boxam ido)pyr r ole-2-car boxam ido)pyr r ole-2-car box-
a m id o)p yr r ole-2-ca r b oxa m id o)d im et h yla m in op r op a n e
Bistr iflu or a ceta te, 5. The nitro compound 12 (1125 mg) was
hydrogenated in a mixture of MeOH (10.5 mL) and 0.1 N HCl
(4.5 mL) over Pd/C-10% (125 mg) as described in the previous
experiment above, to afford, after freeze-drying, the crude
amine (112.5 mg, 84%). This amine was mixed with the
carboxylic acid 16 (42.4 mg, 0.11 mmol), HBTU (64.2 mg, 0.17
mmol), NMM (50 mL) and DMF(1.0 mL, dry), and the
resultant mixture was stirred at room temperature for 24 h.
The product was isolated by means of reverse-phase HPLC.
Fractions containing the product were frozen immediately on
collection then freeze-dried to give the bis-TFA salt of 5 as a
buff-colored powder (58.3 mg, 37% yield) which had no distinct
melting point: ES-MS found [M + 1] 948.4964, calcd for
1-Meth yl-4-(1-m eth yl-4-(4-d im eth yla m in obu tyr yl)a m i-
n o)p yr r ole-2-ca r boxa m id o)p yr r ole-2-ca r boxylic Acid , 16.
A solution of 15 (465 mg, 1.15 mmol) in EtOH (5 mL) and
NaOH (1.73 mmol) was heated under reflux for 1 h and then
concentrated under reduced pressure. The residue was dis-
solved in MeOH (5 mL) containing 3% of concentrated NH4-
OH, and the resultant solution was applied to a flash column
of silica. The column was eluted first with a mixture of EtOAc
(60 parts) and MeOH containing 3% concentrated NH4OH (40
parts), to elute any unreacted ester 15 and any decarboxylation
product 17, then MeOH containing 3% of concentrated NH4-
OH was used to elute the carboxylic acid 16. After evaporation
water was removed from the residue by coevaporation with
EtOH, and finally the residue was triturated with Et2O until
it solidified. After drying in vacuo the product was obtained
as a cream-colored powder (258 mg, 60% yield): mp 138 °C
(gas evolution); ES-MS [M + 1] found 376.2, calcd for
1
C47H62N15O7 948.4956; H NMR (DMSO-d6) δ 1.82-1.93 (4H,
2 × overlapping quintets, 2 × CH2-CH2-CH2), 2.35 (2H, t,
CH2-CH2-CO), 2.79 (12H, m, 2 × NMe2), 3.07 (4H, m, 2 ×
CH2-CH2-NMe2), 3.24 (2H, m, NH-CH2-CH2), 3.82-3.87
(18H, 6 × pyrrole NMe), 6.87-7.22 (12H, 12 × ArH), 8.15 (1H,
t, exchangeable, CONH-CH2), 9.30 (1H, broad, exchangeable,
TFA), 9.50 (1H, broad, exchangeable, TFA), 9.90-9.94 (6H,
exchangeable, 6 × CONH).
C
18H25N5O4 376.4; IR (KBr) νmax 1639, 1617, 1566 cm-1 1H
;
NMR (DMSO-d6) δ 1.69 (2H, m, CH2-CH2-CH2), 2.14 (6H, s,
NMe2), 2.26 (4H, m, CH2-CH2-CH2; after the addition of D2O
this became 2H, t, and a new triplet for 2H appeared at 2.85),
3.81 (6H, s, 2 × pyrrole NMe), 6.78 (1H, d, ArH), 6.85 (1H, d,
ArH), 7.15 (1H, d, ArH), 7.37 (1H, d, ArH), 9.82 (1H, s,
exchangeable, CONH), 9.83 (1H, s, exchangeable, CONH). The
CO2H signal was not observed.
Isop h th a la m id od ia cetic Acid , 18. A stirred suspension
of isophthaloyl chloride (200 mg, 1.00 mmol) and glycine
methyl ester hydrochloride (275 mg, 2.20 mmol) in anhydrous
methylene chloride (20 mL) was treated dropwise with a
solution of triethylamine (0.66 g, 6.60 mmol) in dry methylene
chloride (5 mL). The resulting solution was then stirred at
Note: It was inadvisable to acidify the product of this
hydrolysis. Whenever this was done, either by using HCl, or
by attempting to isolate the product by reverse-phase HPLC
using an acidic (TFA) solvent system, considerable decarbox-

Characteristics of Extended Lexitropsins
J ournal of Medicinal Chemistry, 2000, Vol. 43, No. 17 3265
room temperature for 1.5 h. The solvent was evaporated and
the residue treated with 2 M HCl (5 mL) then extracted with
methylene chloride. Evaporation of the extract gave the
required product as a pale yellow gum (210 mg, 75% yield):
1H NMR (DMSO-d6) δ 3.66 (6H, s, 2 × OMe), 4.04 (4H, d, J 6
Hz, 2 × NHCH2, collapses to a singlet on addition of D2O),
7.62 (1H, t, J 8 Hz, ArH), 8.03 (2H, dd, J o 8 Hz, J m 2 Hz, 2 ×
ArH), 8.38 (1H, m, ArH), 9.12 (2H, t, J 6 Hz, COHN); EI-MS
found 308.1026, calcd for C14H16N2O6 308.2220. A solution of
the dimethyl ester (690 mg, 2.30 mmol) in ethanol (30 mL)
was treated with 2 M NaOH (15 mL) and the resulting solution
stirred at room temperature for 1 h. The mixture was
concentrated by evaporation to a total volume of 10-15 mL
then acidified with 2 M HCl and extracted with ethyl acetate.
Evaporation of the extract gave 18 as a colorless solid (260
different arrangements of (A/T)10 sites together with the
sequence AATT, each separated by 4 GC base pairs (CGCG).
Polylinker DNA fragments containing each of these synthetic
sequences were obtained by digesting with HindIII, labeling
with [R-32P]dATP using reverse transcriptase and digesting
again with either EcoR1 (p(A/T)6-12 and pA3-6) or Sac1 (p(A/
T)10 and pAAD1). The labeled DNAs of interest were separated
from the remainder of the plasmid on 6-8% polyacrylamide
gels and dissolved in 10 mM Tris-HCl pH 7.5 containing 0.1
mM EDTA.
DNa se I footp r in tin g: Radiolabeled DNA (1.5 µL) was
mixed with 1.5 µL of ligand (dissolved in 10 mM Tris-HCl pH
7.5 containing 10 mM NaCl) and left to equilibrate at room
temperature for at least 30 min. 2 µL of DNase I (ap-
proximately 0.01 units/mL diluted in 20 mM NaCl containing
2 mM MgCl2 and 2 mM MnCl2) was then added. The reaction
was stopped after 1 min by adding 4 µL of formamide
containing 10 mM EDTA and 0.1% (w/v) bromophenol blue.
Hyd r oxyl r a d ica l footp r in tin g: Radiolabeled DNA (2 µL)
was mixed with 2 µL of ligand (dissolved in 10 mM Tris-HCl
pH 7.5 containing 10 mM NaCl) and left to equilibrate at room
temperature for at least 30 min. Digestion was initiated by
adding 6 µL of a freshly prepared solution containing 30 µM
ferrous ammonium sulfate, 300 µM EDTA, 3 mM ascorbic acid
and 0.1% H2O2. The reaction was stopped after 10 min by
adding 8 µL of 0.75 M sodium acetate and 60 µL of ethanol.
The DNA was precipitated, washed with 70% ethanol, dried
and redissolved in 8 mL of formamide containing 10 mM EDTA
and 0.1% (w/v) bromophenol blue.
Gel electr op h or esis: Samples were denatured by boiling
for 3 min, cooled on ice and loaded onto denaturing polyacry-
lamide gels (6-10% w/v) containing 8 M urea. Gels (40 cm
long) were run at 1500 V for about 2 h before fixing in 10%
acetic acid (v/v), transferring to Whatman 3MM paper, drying
under vacuum at 80 °C and exposing to autoradiography at
-70 °C with an intensifying screen. Autoradiographs were
scanned with a Hoefer 365W scanning microdensitometer. The
intensity of each band was estimated using the Hoefer
software. Differential cleavage plots, calculated from these
scans, represent the intensity of each band in the drug-treated
lane divided by the intensity of the corresponding band in the
drug-free control.
1
mg, 41% yield): mp 152-154 °C; H NMR (DMSO-d6) δ 3.94
(4H, d, J 6 Hz, CH2 collapses to a singlet on addition of D2O),
7.62 (1H, d, J 8 Hz, ArH), 8.03 (2H, dd, J o 8 Hz, J m 2 Hz, 2 ×
ArH), 8.37 (1H, s, ArH), 8.98 (2H, t, J 6 Hz, 2 × CONHCH2
exchangeable), 13.12 (2H, broad, CO2H exchangeable); ES-MS
found 279.01 [M - 1], calcd for C12H12N2O6 [M - 1] 279.17.
N1,N3-Bis(2-{[5-({[5-({[3-d im eth yla m in op r op yl]a m in o}-
ca r b on yl)-1-m et h yl-1H -p yr r ol-3-yl]a m in o}ca r b on yl)-1-
m e t h yl-1H -p yr r ol-3-yl]a m in o}-2-oxoe t h yl)isop h t h a l-
a m id e Bistr iflu or a ceta te, 6. A solution of the dicarboxylic
acid 18 (28 mg, 0.1 mmol), NMM (30 mg, 0.3 mmol) and HBTU
(165 mg, 0.3 mmol) in DMF (1 mL, dry) was stirred at room
temperature for 1 h. The mixture was then treated with a
solution of freshly prepared 10, R ) NH2 [prepared beforehand
by hydrogenating a suspension of 10, R ) NO2 (165 mg, 0.3
mmol) with Pd/C-10% (165 mg) in 2-propanol (20 mL) for 6 h]
in dry DMF (1 mL). The resulting solution was stirred at room
temperature overnight. The mixture was then purified by
HPLC to give the required product (60 mg, 52% yield) as a
pale yellow solid (bis-TFA salt) with no distinct melting
point: 1H NMR (DMSO-d6) δ 1.83 (4H, m, J 7 Hz, CH2CH2-
CH2), 2.78 (12H, d, J 4 Hz, 4 × NHMe2), 3.07 (4H, m,
CONHCH2, collapses to a triplet on addition of D2O), 3.24 (4H,
m, 2 × Me2NHCH2, collapses to a triplet on addition of D2O),
3.81 (6H, s, 2 × NMe), 3.83 (6H, s, 2 × NMe), 4.04 (4H, d, J 6
Hz, NHCH2CO, collapses to a singlet on addition of D2O),
6.91-6.93 (4H, m, 2 × ArH), 7.15 (2H, d, J 2 Hz, ArH), 7.17
(2H, d, J 2 Hz, ArH), 7.61 (1H, t, J o 8 Hz, ArH), 8.06 (2H, dd,
J o 8 Hz, J m 2 Hz, ArH), 8.14 (2H, t, J 6 Hz, 2 × NHCH2,
exchangeable), 8.43 (1H, s, ArH), 8.91 (2H, t, J 6 Hz, 2 ×
CONH, exchangeable), 9.27 (2H, broad, 2 × NHMe2, exchange-
able), 9.85 (2H, s, 2 × CONH, exchangeable), 9.96 (2H, s, 2 ×
CONH, exchangeable); ES-MS found [M + 1] 937.4804, calcd
for C46H61N14O8 937.4797.
Refer en ces
(1) Wartel, R. M.; Larson, R. M.; Wells, R. D. A specific probe for
A-T regions of duplex deoxyribonucleic acid. J . Biol. Chem.
1974, 249, 6719-6731.
(2) Abu-Daya, A.; Brown, P. M.; Fox, K. R. DNA sequence prefer-
ences of several AT-selective minor groove binding ligands.
Nucleic Acids Res. 1995, 23, 3385-3392.
(3) Hamdan, I. I.; Skellern, G. G.; Waigh, R. D. Ligand binding to
oligonucleotides. Methods Mol Biol., in press.
F ootp r in tin g Stu d ies. Liga n d s: The ligands were dis-
solved in 10 mM Tris-HCl pH 7.5 containing 10 mM NaCl at
a concentration of 10 mM and stored at -20 °C. The com-
pounds were diluted to working concentrations in the same
buffer immediately before use. Oligonucleotides for preparing
the DNA fragments containing synthetic DNA inserts were
purchased from Oswel. DNase I was purchased from Sigma
and stored at -20 °C at a concentration of 7200 units/mL.
(4) For example: Rao, K. E.; Shea, R. G.; Bathini, Y.; Lown, J . W.
Molecular recognition between oligopeptides and nucleic acids
- DNA-sequence specificity and binding properties of thiazole-
lexitropsins incorporating the concepts of base site acceptance
and avoidance. Anti-Cancer Drug Des. 1990, 5, 3-20. Dwyer,
T. J .; Geierstanger, B. H.; Bathini, Y.; Lown, J . W.; Wemmer,
D. E. Design and binding of
d(CGCAAGTTGGC)‚d(GCCAACTTGCG)
a
distamycin-A analogue to
synthesis, NMR
-
DNA fr a gm en ts: The sequences of the five most significant
DNA fragments used in this work are shown in Figure 1. The
TyrT fragment, derived25 from plasmid pKM∆-98, was ob-
tained by digesting the plasmid with EcoR1 and Ava1 and
labeled at the 3′-end of the EcoR1 site with [R-32P]dATP using
reverse transcriptase. Plasmid pAAD1 was prepared as previ-
ously described.2 The other plasmids were prepared by cloning
oligonucleotides into the SmaI site of pUC18 as previously
described. pAAD12 contains an insert which possesses 5
different (A/T)4 tracts, separated by 6 GC base pairs (CGCGCG).
Plasmid pA3-6 contains 4 Tn × An tracts of varying length (n
) 3-6) separated by 4 GC base pairs (CGCG). p(A/T)6-12
contains 4 AT tracts of 6, 8, 10, and 12 bases long with the
sequence AnTATn (n ) 2-5) separated by 4 GC base pairs
(CCGG). Plasmid p(AT)6 contains 4 different arrangements of
(A/T)6 sites together with the sequence AATT, each separated
by 4 GC base pairs (CGCG). Plasmid p(A/T)10 contains 3
studies and implications for the design of sequence-specific
minor groove binding oligopeptides. J . Am. Chem. Soc. 1992,
114, 5912-5919. Geierstanger, B. H.; Mrksich, M.; Dervan,
P. B.; Wemmer, D. E. Design of a G. C specific DNA minor-
groove binder. Science 1994, 266, 646-650.
(5) Turner, J . M.; Baird, E. E.; Dervan, P. B. Recognition of seven
base pair sequences in the minor groove of DNA by ten-ring
pyrrole-imidazole polyamide hairpins. J . Am. Chem. Soc. 1997,
119, 7636-7644.
(6) For example: Guo, D.; Gupta, R.; Lown, J . W. DNA sequence-
selective binding of head-to-tail linked bis-lexitropsins - relation
of phasing to cytotoxic potency. Anti-Cancer Drug Des. 1993, 8,
369-397.
(7) Trauger, J . W.; Baird, E. E.; Dervan, P. B. Recognition of DNA
by designed ligands at subnanomolar concentrations. Nature
1996, 382, 559-561. Herman, D. M.; Baird E. E.; Dervan, P. B.
Stereochemical control of the DNA binding affinity, sequence
specificity and orientation preference of chiral hairpin poly-
amides in the minor groove. J . Am. Chem. Soc. 1998, 120,
1382-1391.

3266 J ournal of Medicinal Chemistry, 2000, Vol. 43, No. 17
Fishleigh et al.
(8) Haq, I.; Ladbury, J . E.; Chowdrhy, B. Z.; J enkins, T. C.; Chaires,
J . B. Specific binding of Hoechst 33258 to the (CGCAAATT-
TGCG)2 duplex: calorimetric and spectroscopic studies. J . Mol.
Biol. 1997, 271, 244-257.
(9) For example: Singh, M. P.; Plouvier, B.; Hill, G. C.; Gueck, J .;
Pon, R. T.; Lown, J . W. Isohelicity and strand selectivity in the
minor groove binding of chiral (1R,2R)-bis(netropsin)-1,2-cyclo-
propanedicarboxamide and (1S,2S)-bis(netropsin)-1,2-cyclopro-
panecarboxamide ligands to duplex DNA. J . Am. Chem. Soc.
1994, 116, 7006-7020. Wang, A. H. J .; Cottens, S.; Dervan, P.
B.; Yesinowski, J . P.; Vandermarel, G. A.; Vanboom, J . H.
(17) Leo, A. The octanol-water partition coefficient of aromatic
solutes - the effect of electronic interactions, alkyl chains,
hydrogen bonds and ortho-substitution. J . Chem. Soc., Perkin
Trans. 2 1983, 825-838 and references cited therein.
(18) Sykes, P. A Guidebook to Mechanism in Organic Chemistry, 4th
ed.; Longman: London, 1975; p 68.
(19) Zhao, R.; Lown, J . W. Synthesis of
a novel N-methoxy-
methylpyrrole-containing DNA minor-groove binding oligo-
peptide related to distamycin. Heterocycles 1995, 41, 337-344.
(20) Green, A. L. Ionization constants and water solubilities of some
aminoalkylphenothiazine tranquillizers and related compounds.
J . Pharm. Pharmacol. 1967, 19, 10-16.
(21) Nishiwaki, E.; Tanaka, S.; Lee, H.; Shibuya, M. Efficient
synthesis of oligo-N-methylpyrrolecarboxamides and related
compounds. Heterocycles 1988, 27, 1945-1952.
(22) Glibin, E. N.; Tsukerman, B. V.; Ginzburg, O. F. Analogue of
distamycin A with a dimethylamino group. J . Org. Chem. USSR
(Engl. Transl.) 1977, 13, 2082-2083.
(23) Glibin, E. N.; Tsukerman, B. V.; Ginzburg, O. F. Synthesis of
actinomycin analogues. X. Polyfunctional DNA complexing
agents. J . Org. Chem. USSR (Engl. Transl.) 1981, 17, 568-572.
(24) Arcamone, F.; Nerviano, M.; Mongelli, N.; Penco, S. Poly-4-
aminopyrrole-2-carboxamide derivatives, procedures for their
preparation, and antiviral and antineoplastic pharmaceuticals
containing them. Ger. Offen. DE 3,623,880, J an 29, 1987.
(25) Drew, H. R.; Travers, A. A. DNA structural variations in the E.
coli tyrT promoter. Cell 1984, 37, 491-502.
(26) Turchin, K. F.; Grokhovskii, S. L.; Zhuze, A. L.; Gottikh, B. P.
Ligands possessing affinity for definite pairs of DNA bases. II.
A study of the stereochemistry of the chromophore of distamycin
A by 1H NMR spectroscopy. Sov. J . Bioorg. Chem. (Engl. Transl.)
1978, 4, 780-790.
Interactions between
a symmetrical minor groove binding
compound and DNA oligonucleotides - H-1 and F-19 NMR
studies. J . Biomol. Struct. Dyn. 1989, 7, 101-117.
(10) Florence, A. T.; Attwood, D. Physicochemical Principles of
Pharmacy, 3rd ed.; Macmillan: London, 1998; p 185.
(11) Fox, K. R.; Waring, M. J . DNA structural variations produced
by actinomycin and distamycin as revealed by DNAase
I
footprinting. Nucleic Acids Res. 1984, 12, 9271-9285.
(12) Portugal, J .; Waring, M. J . Hydroxyl radical footprinting of the
sequence-selective binding of netropsin and distamycin to DNA.
FEBS Lett. 1987, 225, 195-200.
(13) Abu-Daya, A.; Fox, K. R. Interaction of minor groove binding
ligands with long AT-tracts. Nucleic Acids Res. 1997, 25, 4962-
4969 and unpublished observations.
(14) Lown, J . W. Lexitropsins - rational design of DNA-sequence
reading agents as novel anti-cancer agents and potential cellular
probes. Anti-Cancer Drug Des. 1988, 3, 25-40.
(15) Waszkowycz, B. Unpublished data.
(16) Rao, K. F.; Zimmerman, J .; Lown, J . W. Sequence-selective DNA-
binding by linked bis-N-methylpyrrole dipeptides - an analysis
by MPE footprinting and force-field calculations. J . Org. Chem.
1991, 56, 786-797.
J M990620E