New dsDNA binding unnatural oligopeptides with pyrimidine selectivity

Article abstract of DOI:10.1016/S0968-0896(02)00268-7

Solid phase peptide library screening followed by extension of a lead recognition element for binding to a dsDNA sequence (NF binding site of IL6) using solution phase screening, delivered a new DNA binding peptide, Ac-Arg-Ual-Sar-Chi-Chi-Tal-Arg-CONH₂. In the present research, the contribution of the different amino acid side chains to the binding strength of the peptide to dsDNA was investigated using an ethidium bromide displacement test. Based on these results, the lead structure was optimized by deconvolution. Eight new unnatural amino acids were evaluated at two positions of the heptapeptide replacing the Ual-Sar fragment. The strongest dsDNA binding was observed using {[(3-chlorophenyl)methyl]amino}acetic acid (Cbg) and β-cyclohexyl-l-alanine (Cha) respectively, at those two positions. A 10-fold increase in affinity compared to the Ual-Sar sequence was obtained. Further enhancement of dsDNA binding was obtained with hybrid molecules linking the newly developed peptide fragment to an acridine derivative with a flexible spacer. This resulted in ligands with affinities in the μM range for the dsDNA target (K_d of 2.1×10^-6 M). DNase I footprinting with the newly developed oligopeptide motifs showed the presence of a pronounced pyrimidine specificity, while conjugation to an intercalator seems to redirect the interaction to mixed sequences. This way, new unnatural oligopeptide motifs and hybrid molecules have been developed endowed with different sequence selectivities. The results demonstrate that the unnatural peptide library approach combined with subsequent modification of selected amino acid positions, is very suited for the discovery of novel sequence-specific dsDNA binding ligands.

Full text of DOI:10.1016/S0968-0896(02)00268-7

Bioorganic & Medicinal Chemistry 10 (2002) 3401–3413
New dsDNA Binding Unnatural Oligopeptides with
Pyrimidine Selectivity
Zhenyu Zhang, Patrick Chaltin, Arthur Van Aerschot, Jeff Lacey, Jef Rozenski,
Roger Busson and Piet Herdewijn*
Laboratory of Medicinal Chemistry, Rega Institute for Medical Research, Katholieke Universiteit Leuven, Minderbroedersstraat 10,
B-3000 Leuven, Belgium
Received 23 January 2002; accepted 3 July 2002
Abstract—Solid phase peptide library screening followed by extension of a lead recognition element for binding to a dsDNA
sequence (NF binding site of IL6) using solution phase screening, delivered a new DNA binding peptide, Ac-Arg-Ual-Sar-Chi-Chi-
Tal-Arg-CONH2. In the present research, the contribution of the different amino acid side chains to the binding strength of the
peptide to dsDNA was investigated using an ethidium bromide displacement test. Based on these results, the lead structure was
optimized by deconvolution. Eight new unnatural amino acids were evaluated at two positions of the heptapeptide replacing the
Ual-Sar fragment. The strongest dsDNA binding was observed using {[(3-chlorophenyl)methyl]amino}acetic acid (Cbg) and
b-cyclohexyl-l-alanine (Cha) respectively, at those two positions. A 10-fold increase in affinity compared to the Ual-Sar sequence
was obtained. Further enhancement of dsDNA binding was obtained with hybrid molecules linking the newly developed peptide
fragment to an acridine derivative with a flexible spacer. This resulted in ligands with affinities in the mM range for the dsDNA
target (Kd of 2.1ꢀ10ꢁ6 M). DNase I footprinting with the newly developed oligopeptide motifs showed the presence of a pro-
nounced pyrimidine specificity, while conjugation to an intercalator seems to redirect the interaction to mixed sequences. This way,
new unnatural oligopeptide motifs and hybrid molecules have been developed endowed with different sequence selectivities. The
results demonstrate that the unnatural peptide library approach combined with subsequent modification of selected amino acid
positions, is very suited for the discovery of novel sequence-specific dsDNA binding ligands.
# 2002 Elsevier Science Ltd. All rights reserved.
Introduction
lot of supplementary information about DNA-ligand
recognition motifs.1,4,9 The involvement of the amide
Nucleic acids are important targets in the development
of new compounds, able to control gene expression. The
armamentarium of nucleic acids recognizing ligands is
increasing exponentially.1ꢁ4 It is well known that most
DNA-targeted drugs are small molecules of limited
specificity and high toxicity.5ꢁ8 The specificity can be
increased by using compounds with multiple interaction
sites with the DNA target, and thus compounds of
medium size. As DNA-protein interaction is the central
event in the regulation of gene expression, it may be
used as a model to design synthetic molecules that
recognize DNA by groove binding.9 The significant
progress in the discovery of new DNA binding agents
within the past decade, particularly in the field of lexi-
tropsins, combilexins and PNAs, has provided us with a
backbone in these three strategies has supported our
approach to synthesize peptide libraries based on mod-
ified amino acids with unique side chains or constrained
backbones. This peptide-based approach offers, like-
wise, the advantage that combinatorial chemistry can be
used to easily generate libraries, representing diverse
recognition patterns.
In previous experiments, a first lead sequence (Arg-Ual-
Sar) was generated using a solid phase screening
method.10 This sequence was extended using solution
phase screening, resulting in several heptapeptide lead
compounds such as Ac-Arg-Ual-Sar-Chi-Chi-Tal-Arg-
CONH2, showing an affinity of 4ꢀ10ꢁ4 M for a dsDNA
14-mer target.11 The structures of Arg, Ual, Sar, Chi
and Tal are displayed in Figure 1. Investigation of the
interaction capacities with dsDNA of the different
amino acids in the peptide lead Ac-Arg-Ual-Sar-Chi-
Chi-Chi-Arg-CONH2, suggested that the uracil (Ual)
*Corresponding author. Tel.: +32-16-337387; fax: +32-16-337340;
e-mail: piet.herdewijn@rega.kuleuven.ac.be
0968-0896/02/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(02)00268-7

3402
Z. Zhang et al. / Bioorg. Med. Chem. 10 (2002) 3401–3413
and sarcosine (Sar) side chains contribute less to the
overall affinity of the peptide than quinazoline-dione
(Chi). Therefore, starting from the lead peptide Ac-Arg-
Ual-Sar-Chi-Chi-Tal-Arg-CONH2, new libraries were
prepared with the aim to select stronger binding residues
for the two positions that initially were fixed during the
solution phase selection round (Ual-Sar), thereby preser-
ving the positively charged arginine (Arg). Eight amino
acid building blocks (Fig. 1) were applied to optimize the
two random positions X and Y in Ac-Arg-Y-X-Chi-Chi-
Tal-Arg-CONH2. Fmoc-strategy-based solid phase pep-
tide synthesis was used to assemble the library. The
dsDNA target used for screening in solution is the labeled
14-mer double stranded DNA, representing the binding
site of the nuclear factor of Interleukine 6, as used in
previous screening assays.10ꢁ12 Gel retardation experi-
ments were performed to evaluate the binding strength of
the ligands and to determine the dissociation constants
of the selected peptides. Footprinting was used to evalu-
ate the sequence specificity of the dsDNA interaction.
may also influence the binding indirectly by its impor-
tance for the conformation of the peptide). This could
yield important information for further modifications of
the oligopeptides to enhance the affinity or selectivity.
The glycine substituted peptides were synthesized on
solid support as described previously.11 Coupling of
Fmoc-glycine was performed with an activation mixture
containing 4 equiv of HOAt, DIC, DIEA and the amino
acid in regard to the free amino groups. Activation
mixtures for N-Fmoc-sarcosine, Nꢀ-Fmoc-b-(uracil-1-
yl)-a-d-alanine,
Nꢀ-Fmoc-NG-2,2,5,7,8-pentamethyl-
chroman-6-sulfonyl-l-arginine and Nꢀ-Fmoc-b-(1H-
quinazolin-2,4-dion-3-yl)-a-d-alanine were as used
before.11 After cleavage from the solid support and
reversed phase HPLC purification, the compounds were
identified by mass spectrometry. All measured masses
corresponded to the calculated ones. The binding
strength of the glycine substituted analogues and the
reference oligopeptide Ac-Arg-Ual-Sar-Chi-Chi-Chi-
Arg-NH2 was compared in an ethidium bromide dis-
placement assay13 using the original 14-mer dsDNA
target. The obtained percentages are relative to controls
without DNA and without ligand addition. Results at
different concentrations are averaged from two to three
measurements and are shown in Table 1.
Results
Contribution of amino acid side chains to dsDNA binding
To determine the contribution to dsDNA binding of the
individual amino acids of the oligopeptide lead struc-
tures selected in previous work, new oligopeptide ana-
logues were prepared and investigated. The middle five
amino acids of Ac-Arg-Ual-Sar-Chi-Chi-Chi-Arg-NH2
were systematically replaced by glycine. This way the
side chains of the different amino acids were removed,
while the peptide backbone was retained. By measuring
the affinity of the different analogues, the contribution
of the amino acid side chains could be evaluated
(although we are aware of the fact that the side chain
The ethidium bromide displacement experiment shows
that replacement of each of the three quinazoline-dione
amino acids with glycine leads to an equally reduced
affinity, corresponding to a ꢂ40% decrease of EtBr-
displacement compared to the reference oligopeptide.
Analogues with glycine substituted for sarcosine and the
uracil amino acid retain, however, a similar affinity for
the dsDNA target as the reference oligopeptide.
These results suggest that the uracil and sarcosine side
chains contribute less to the overall affinity of the pep-
tide Ac-Arg-Sar-Chi-Chi-Chi-Arg-NH2, while the qui-
nazoline-dione side chains actively participate in the
stabilization of the DNA-peptide complex. Further-
more, this highlights again the importance of the selec-
ted quinazoline-dione amino acids (Chi) in the
oligopeptide motif for interaction with dsDNA, as
reported previously.11
Synthesis of the amino acid building blocks
Based on the results obtained with the glycine sub-
stituted oligopeptides, the Ual-Sar positions of the peptide
Table 1. Fluorescence readings of glycine substituted oligopeptides
and the reference peptide Ac-Arg-Ual-Sar-Chi-Chi-Chi-Arg-NH2 at
two different concentrations. The results are reported as % fluores-
cence relative to control wells
Oligopeptide sequences
250 mM
750 mM
Ac-Arg-Ual-Sar-Chi-Chi-Chi-Arg-CONH2
Ac-Arg-Ual-Sar-Chi-Chi-Gly-Arg-CONH2
Ac-Arg-Ual-Sar-Chi-Gly-Chi-Arg-CONH2
Ac-Arg-Ual-Sar-Gly-Chi-Chi-Arg-CONH2
Ac-Arg-Ual-Gly-Chi-Chi-Chi-Arg-CONH2
Ac-Arg-Gly-Sar-Chi-Chi-Chi-Arg-CONH2
42%
87%
79%
77%
46%
45%
27%
67%
66%
62%
30%
20%
Figure 1. Structures of unnatural synthetic amino acids (Ual, PrU,
Nic, Imi, Cha, Cbg, CF3 and Sar) used in the present library approach
to select for position X and Y in Ac-Arg-Y-X-Chi-Chi-Tal-Arg-
CONH2. The structures of Chi, Tal and Arg are also shown.

Z. Zhang et al. / Bioorg. Med. Chem. 10 (2002) 3401–3413
3403
Ac-Arg-Ual-Sar-Chi-Chi-Tal-Arg-CONH2 were mod-
ified with several new amino acid building blocks to
increase the contribution of these positions to the stabi-
lity of the oligopeptide-DNA complexes. The peptide
Ac-Arg-Ual-Sar-Chi-Chi-Tal-Arg-CONH2 was chosen
to be modified because of its mixed sequence and ease of
synthesis. Looking at the recognition mechanisms
between DNA and proteins, unnatural amino acids
(Fig. 1) were selected for these positions in a way that
they represent both functional and structural diversity.
In previous libraries, the selection of new unnatural
amino acids was predominantly based on the presence
of hydrogen bonding functionalities and intercalating
possibilities. In this research, three new building blocks
(Cha, Cbg and CF3) were selected in order to incorpo-
rate the possibility for hydrophobic interactions, since
these contacts are known to strongly stabilize DNA-
ligand complexes.14 Furthermore, Nic and Imi were
chosen as new amino acid residues to further enlarge the
hydrogen bonding capacities of the peptides. Moreover,
more rigidity was introduced by using the building
block PrU to reduce the entropy loss during binding
(Fig. 1).
Starting from Na-(9-fluorenylmethyloxycarbonyl)-Nb-
tert-butyloxycarbonyl-l-a-2,3-diaminopropionic acid
5,16 compound 6 and 7 were synthesized (Scheme 2).
The Boc-group of 5 was removed with acid and hetero-
cycles were introduced via amide bond formation using
the activated amides 8 and 9 as reagent. The carboxylic
function of 11 and 12 was activated with 1,3-thiazoli-
dine-2-thione, affording a TT-amide.17 These activated
amides occur as yellow crystals, facilitating monitoring
of the reaction with the b-amino function of the diami-
nopropionic acid by disappearance of the yellow color.
Apart from a-amino acids, we also introduced three
peptoid monomers (CF3, Cbg, Sar) in the library. Pre-
parations of CF3 and Cbg proceeded through a reduc-
tive amination by using NaCNBH3 in a mixture of
water and 1,4-dioxane (3:2, v/v)18 thereby reducing the
in situ formed imine and retaining the unreacted alde-
hyde, followed by Fmoc protection (Scheme 3).
Synthesis and screening of the oligopeptide library and
individual peptides
Utilizing a Fmoc-strategy-based solid phase peptide
synthesis19,20 an oligopeptide library was assembled21
with the general structure Ac-Arg-Y-X-Chi-Chi-Tal-
Arg-CONH2, with X and Y representing the eight
amino acid building blocks as shown in Figure 1. The
remaining residues and their sequences were retained
from our previous selection procedures.10,11 The cou-
pling conditions between different residues were opti-
mized by synthesizing several test peptides with random
sequences. The absence of side reactions and racemiza-
tion, was verified by performing HPLC analysis and
mass spectrometry of these test peptides after cleavage
from the solid support.22
In this way, using proline and alanine as scaffold, four
new amino acids were introduced (PrU, Nic, Imi and
Cha) and also two new peptoid monomers (CF3 and
Cbg) were synthesized and used in the library. The Ual
and Sar amino acids were already present in the pre-
vious library and they were added to serve as reference
amino acids. Their synthesis was performed as described
previously.11 Following protection with a N3-benzoyl
group, uracil was reacted with N-Boc-protected trans-l
hydroxyproline methyl ester by Mitsunobu reaction,
similar to the method described by Lowe for the synth-
esis of a thymine derivative (Scheme 1).15 The obtained
cis-l-proline derivative was fully deprotected using
NaOH followed by trifluoroacetic acid. The secondary
amino group of the pyrrolidine ring was protected with
a Fmoc group under Schotten-Baumann conditions
(Scheme 1).
Scheme 2. Synthesis of (2S)-3-{[(1-Boc-1H-benzimidazol-5-yl)carbo-
nyl]amino}-2-Fmoc-amino-propanoic acid (6) and (2S)-2-Fmoc-
amino-3-[(3-pyridinylcarbonyl)amino] propanoic acid (7). (i) 1.2 equiv
(Boc)2O, 10% Na2CO3, 1,4-dioxane; 85%, (ii) 1.2 equiv 1,3-thiazoli-
dine-2-thione, DIC, CH2Cl2; 62% (iii) (a) TFA, CH2Cl2 (1/1, v/v), (b)
0.95 equiv 8 or 9 in saturated NaHCO3, 1,4-dioxane (1/1, v/v); 80%.
Scheme 1. Synthesis of Na-Fmoc-cis-4-uracil-l-proline. (i) 1.2 equiv
(Boc)2O, 10% Na2CO3, 1,4-dioxane; 84%, (ii) CsCO3, 1.2 equiv CH3I
in DMF; 85%, (iii) 1.5 equiv 3-benzoyluracil, PPh3, DEAD in THF;
82%, (iv) (a) 0.1N NaOH, room temperature, (b) TFA/CH2Cl2 (1/1,
v/v); 90%, (v) 1.2 equiv FmocCl, 10% Na2CO3, 1,4-dioxane; 80%.

3404
Z. Zhang et al. / Bioorg. Med. Chem. 10 (2002) 3401–3413
apparent dissociation constants were determined as the
concentration at which 50% of the target dsDNA was
mobility shifted.24 Although there was not a significant
difference between the affinities of the peptides (Table
2), they did show a binding strength which was more
than one order of a magnitude higher than the reference
peptide Ac-Arg-Ual-Sar-Chi-Chi-Tal-Arg-CONH2 with
a Kd of 4ꢀ10ꢁ4. The peptide Ac-Arg-Cbg–Cha-Chi-Chi-
Tal-Arg-CONH2 was endowed with the lowest dis-
sociation constant of 2.0ꢀ10ꢁ5
.
Scheme 3. Synthesis of (Fmoc-4-(trifluoromethyl)phenyl]methyl}amino)
acetic acid (14) and ([(3-chlorophenyl)methyl]-Fmoc-amino)acetic acid
(16) (i) 1.5 equiv NaCNBH3 in H2O, 1,4-dioxane (3/2, v/v); 60%; (ii)
1.2 equiv FmocCl, 10% Na2CO3,1,4-dioxane; 85%.
Modification of the selected peptide with intercalators
It has been well documented from studies with the
combilexin model that the introduction of intercalating
heteropolyaromatic moieties to a ligand may increase
the binding affinity to dsDNA as well as enhance the
sequence specificity, particularly in the GC rich region.4
This strategy was already applied in previous experi-
ments, where the peptide Ac-Arg-Ual-Sar-Chi-Chi-Tal-
Arg-CONH2 was conjugated to different intercalators
by a variety of linker arms.25 Therefore, the influence on
the binding strength and the sequence selectivity of
coupling intercalating agents to the peptide (Cbg-Cha)
was investigated. As it is not clear yet at which position
the peptides are bound to dsDNA, it is difficult to select
an optimized spacer between peptide and intercalator
based on modeling experiments or previous research.
Therefore we evaluated different spacers to connect the
intercalating agent to the peptide. This could be impor-
tant for an optimal positioning of the intercalator with
respect to the base sequences of dsDNA.26ꢁ28
An iterative deconvolution procedure was used for the
identification of X and Y. The first libraries were syn-
thesized with the aim of optimizing position Y. The
mixture of peptides released from the solid support,
containing eight peptides in every batch, was subjected
to a solution phase screening against the dsDNA target
(50-AGATTGTGCAATGT-30/30-TCTAACACGTTACA-50) by
gel shift experiments.11 Mass spectrometric investiga-
tions, performed as outlined previously,23 proved the
presence of all eight peptides in every sublibrary. These
screening experiments led to determination of Cbg and
Cha as the amino acids possessing the strongest binding
affinity at position Y.
In the next step, 16 peptides with general structures Ac-
Arg-Cbg-X-Chi-Chi-Tal-Arg-CONH2 and Ac-Arg-Cha-
X-Chi-Chi-Tal-Arg-CONH2 were synthesized, with X
representing the amino acids shown in Figure 1, except
Tal, Chi and Arg. Following cleavage from the solid
support, the peptides were purified by reversed phase
HPLC, identified by mass spectrometry (Table 4) and
individually screened for binding to dsDNA. From the
eight peptides with a general structure corresponding to
Ac-Arg-Cbg-X-Chi-Chi-Tal-Arg-CONH2, those with
Cha, Cbg or CF3 at position X revealed as the strongest
binding compounds. In the Ac-Arg-Cha-X-Chi-Chi-Tal-
Arg-CONH2 series, the highest affinity was observed
with position X occupied by Imi, Cbg or CF3.
In the present study, two kinds of intercalators e.g.,
acridine-9-carboxylic acid and 9-fluorenone-2-car-
boxylic acid were evaluated and tethered to the terminal
NH2 function of H2N-Arg-Cbg–Cha-Chi-Chi-Tal-Arg-
CONH2 via four sets of linkers composed of glycine and
b-alanine (Figs 2 and 3). Moreover, two structurally
more flexible linkers having three and five CH2 groups
between the amide functions e.g., g-amino butyric acid
(GABA) and 6-amino caproic acid (CAP) (Fig. 4) were
evaluated by conjugation to acridine, so that we could
compare the effect of the flexibility of the linkers on the
binding strength.
It should be noted that the gel electrophoresis in many
cases does not show one single band as the result of a
gel shift. Often a long smear is seen and it is not clear
yet if this is due to the association of several peptides to
one dsDNA (or the opposite) or to a precipitation of the
dsDNA-peptide complex (change of solubility due to
charge neutralization and increase of size of the com-
plex). Therefore in many of the figures it is difficult to
observe the complex formed between dsDNA and pep-
tide, and the disappearance of the free dsDNA band
was taken as criterion for the selection of dsDNA
binding peptides.11
The amino acid building blocks and linkers were intro-
duced using standard procedures for amide formation.
The carboxylic function of the intercalators were acti-
vated and coupled to the terminal NH2 group of the
peptide linker construct using 4 equiv HATU / DIEA in
DMF. The peptides were purified by reversed phase
HPLC and identified using mass spectrometry (Figs 2–
4).
The solution screening of the ten oligopeptides with
intercalators tethered by an a, b, g or e amino acid lin-
ker, revealed that the acridine-peptide with GABA as
the linker shows a stronger binding than the other
compounds and than the unconjugated peptide. (Figs 5
and 6). Comparison to the dipeptide linkers (Cbg-Cha3,
Cbg-Cha4, Cbg-Cha7 or Cbg-Cha8), indicated that the
removal of the amide bond and the increased flexibility
In order to have an idea of the strength of binding of
these compounds, dissociation constants were deter-
mined for the six selected oligopeptides. Gel retardation
experiments were performed at different concentrations
of the peptides in 0.1 M NaCl PBS buffer, and the

Z. Zhang et al. / Bioorg. Med. Chem. 10 (2002) 3401–3413
3405
Table 2. Dissociation constants of the selected oligopeptide sequences compared to the peptide sequence obtained from previous work [(a) refer-
ence peptide]11
Oligopeptide sequence
Formula
KD
Ac-Arg-Ual–Sar-Chi-Chi-Tal-Arg-CONH2 (a)
Ac-Arg-Cha–Imi-Chi-Chi-Tal-Arg-CoNH2
Ac-Arg-Cha–Cbg-Chi-Chi-Tal-Arg-CONH2
Ac-Arg-Cha–CF3-Chi-Chi-Tal-Arg-CONH2
Ac-Arg-Cbg–Cha-Chi-Chi-Tal-Arg-CONH2
Ac-Arg-Cbg–Cbg-Chi-Chi-Tal-Arg-CONH2
Ac-Arg-Cbg–CF3-Chi-Chi-Tal-Arg-CONH2
C54H68N22O16
C64H81N23O15
4.0ꢀ10ꢁ4
3.4ꢀ10ꢁ5
4.0ꢀ10ꢁ5
2.6ꢀ10ꢁ5
2.0ꢀ10ꢁ5
2.4ꢀ10ꢁ5
4.0ꢀ10ꢁ5
C62H79N20O14Cl1
C63H79N20O14F3
C62H79N20O14Cl1
C62H72N20O14Cl2
C63H72N20O14Cl1F3
of the linker has a beneficial effect on the binding and
allows the intercalator to make a better contact with
dsDNA.
not unrealistic to hypothesize that the lack of efficiency
of the intercalators might be partially due to the diffi-
culty of the polyaromatic rings to be positioned in the
right way for optimal binding to dsDNA, or alter-
natively, optimal binding of the peptide is prohibited by
positioning of the intercalating agent. Nevertheless, it
was important to observe from these comparisons that
the properties of the linkers did have an important
impact on the outcome of the binding affinity.
These results might be explained in a way that, in case
of the a, b and e linkers, there is an absence of a syner-
gistic effect of binding by the peptide and the inter-
calating agent. Both parts of the molecule seem to be
non-compatible for binding using these spacer groups.
As the dipeptide linkers show limited flexibility due to
the presence of two or more amide groups, it is indeed
Figure 4. Selected oligopeptide tethered with acridine by g-amino
butyric acid (GABA) and e-amino caproic acid (CAP).
Figure 5. Gel shift experiments of the selected oligopeptide modified
with the introduction of different intercalators e.g., acridine for Cbg-
Cha 1 to Cbg-Cha 4; fluorenone for Cbg-Cha 5 to Cbg-Cha 8 which
were tethered by different linkers composed of glycine and b-alanine.
The blank lane (Bl) contains no peptide, while the reference lane (Ref)
contained the peptide Ac-Arg-Cbg-Cha-Chi-Chi-Tal-Arg-CONH2.
The same concentrations were used for all peptides.
Figure 2. Selected oligopeptide tethered to acridine by different linkers
and their corresponding codes. Mass spectrometry confirmed the
molecular weight of the isolated compounds.
Figure 6. A comparison of binding affinity of oligopeptides either
without intercalators (Ref) or with intercalators tethered by a dipep-
tide linker (Cbg-Cha 4) or a more flexible spacer (GABA, CAP)
against dsDNA at different ligand/target ratios e.g., 100:1 and 50:1.
The blank lane (Bl) contained no peptide and the reference lane the
unconjugated peptide.
Figure 3. Oligopeptide Arg-Cbg-Cha-Chi-Chi-Tal-Arg-CONH2 teth-
ered with 9-fluorenone by different linkers.

3406
Z. Zhang et al. / Bioorg. Med. Chem. 10 (2002) 3401–3413
The apparent dissociation constant of the peptide teth-
ered with the acridine derivative via a GABA linker was
determined. In comparison with the peptide without
intercalator (Kd=2.0ꢀ10ꢁ5 M), a 10-fold enhancement
of the binding capacity (Kd=2.1ꢀ10ꢁ6 M) was observed
(Table 3). When the binding constant was determined at
higher salt concentration (1.0 M NaCl) the Kd value
increased, which might point to the importance of ionic
interactions between dsDNA and the peptides.
Sequence specificity of the selected oligopeptide and
conjugates
To investigate the influence of the Ual-Sar modification
and conjugation to intercalators on the sequence selec-
tivity of the peptide, DNase I footprinting experiments
were performed. A 271 bp fragment derived from the
cloned plasmid pUC 19 and containing the 14-mer tar-
get used for screening, was applied as substrate, as
described previously.11 A typical autoradiograph of a
sequencing gel used to fractionate the products of par-
tial digestion of the DNA in the presence of the oligo-
peptide Ac-Arg-Cbg-Cha-Chi-Chi-Tal-Arg-CONH2 and
the Acridine-GABA conjugate is presented in Figure 7.
Figure 7. DNase I footprinting with the 271-mer Pvu II–Ase I restriction
fragment of cloned plasmid pUC 19 in the presence of the oligopeptide
Ac-Arg-Cbg-Cha-Chi-Chi-Tal-Arg-CONH2 (left) and 10 mM of the
conjugate oligopeptide-GABA-Acridine (right). Reference track (Ref)
contains dsDNA that was not digested. Control track (C) contained no
product. Guanine-specific sequence markers were run in the lane marked
G. Numbers at the side of the gel refer to the numbering scheme of the
fragment. Preferential binding sites are indicated with gray boxes.
With the products bound, the DNase I cleavage pattern
differs significantly from that seen in the control lane.
Numerous bands in the product-containing lanes are
weaker than the same bands in the product-free lanes,
corresponding to attenuated cleavage, whereas others
display relative enhancement of cutting. To obtain more
detailed information, intensities from selected gel lanes
were quantified by densitometry and converted into a
set of differential cleavage plots (Fig. 8) indicating the
extent to which cleavage at each internucleotide bond is
affected by complexation with the investigated ligands.
The plots in Figure 8 compare the footprinting data
obtained with 50 mM Ac-Arg-Cbg-Cha-Chi-Chi-Tal-Arg-
Table 3. Comparison of the apparent dissociation constants (M) of the initial lead compound (a), the compound obtained from the peptide opti-
mization procedure (b) and the compound developed during the intercalator optimization procedure (c)
Oligopeptide sequences
KD
0.1 M NaCl
4.0ꢀ10ꢁ4
2.0ꢀ10ꢁ5
2.1ꢀ10ꢁ6
(a) Ac-Arg-Ual–Sar-Chi-Chi-Tal-Arg-CONH2
(b) Ac-Arg-Cbg–Cha-Chi-Chi-Tal-Arg-CONH2
(c) Acr-GABA-Arg-Cbg–Cha-Chi-Chi-Tal-Arg-CONH2
Table 4. Oligopeptides synthesized for the selection of amino acid X in the sequence Ac-Arg-Cha/Cbg-X-Chi-Chi-Tal-Arg-CONH2
Oligopeptide sequences
Formula
Exact mass calculated for [M+H]+
Exact mass found
Ac-Arg-Cha–PUal-Chi-Chi-Tal-Arg-CONH2
Ac-Arg-Cha–Ual-Chi-Chi-Tal-Arg-CONH2
Ac-Arg-Cha–Sar-Chi-Chi-Tal-Arg-CONH2
Ac-Arg-Cha–Cha-Chi-Chi-Tal-Arg-CONH2
Ac-Arg-Cha–Imi-Chi-Chi-Tal-Arg-CONH2
Ac-Arg-Cha–Cbg-Chi-Chi-Tal-Arg-CONH2
Ac-Arg-Cha–CF3-Chi-Chi-Tal-Arg-CONH2
Ac-Arg-Cha–Nic-Chi-Chi-Tal-Arg-CONH2
Ac-Arg-Cbg–PUal-Chi-Chi-Tal-Arg-CONH2
Ac-Arg-Cbg–Ual-Chi-Chi-Tal-Arg-CONH2
Ac-Arg-Cbg–Sar-Chi-Chi-Tal-Arg-CONH2
Ac-Arg-Cbg–Cha-Chi-Chi-Tal-Arg-CONH2
Ac-Arg-Cbg–Imi-Chi-Chi-Tal-Arg-CONH2
Ac-Arg-Cbg–Cbg-Chi-Chi-Tal-Arg-CONH2
Ac-Arg-Cbg–CF3-Chi-Chi-Tal-Arg-CONH2
Ac-Arg-Cbg–Nic-Chi-Chi-Tal-Arg-CONH2
C62H80N22O16
C60H78N22O16
C56H76N20O14
C62H86N20O14
C64H81N23O15
C62H79N20O14Cl1
C63H79N20O14F3
C62H80N22O15
C62H73N22O16Cl1
C60H71N22O16Cl1
C56H69N20O14Cl1
C62H79N20O14Cl1
C64H74N23O15Cl1
C62H72N20O14Cl2
C63H72N20O14Cl1F3
C62H73N22O15Cl1
1389.6201
1363.6044
1253.5928
1335.6711
1412.6361
1363.5851
1397.6114
1373.6251
1417.5341
1391.5185
1281.5068
1363.5851
1440.5501
1391.4992
1425.5255
1401.5392
1389.6200
1363.6025
1253.5927
1335.6705
1412.6355
1363.5873
1397.6106
1373.6273
1417.5343
1391.5171
1281.5082
1363.5853
1440.5535
1391.5082
1425.5256
1401.5387

Z. Zhang et al. / Bioorg. Med. Chem. 10 (2002) 3401–3413
3407
Figure 8. Differential cleavage plot comparing the susceptibility of the 30-labeled 271 bp fragment to DNase I cutting in the absence and in the pre-
sence of 50 mM of the oligopeptide Ac-Arg-Cbg-Cha-Chi-Chi-Tal-Arg-NH2 [OP (Cbg-Cha)], 150 mM of Ac-Arg-Ual-Sar-Chi-Chi-Tal-Arg-NH2 [OP
(Ual-Sar)] and 10 mM Acr-GABA-Arg-Cbg-Cha-Chi-Chi-Tal-Arg-NH2 [OP-GABA-Acr]. Negative values correspond to a ligand-protected site and
positive values represent enhanced cleavage. Vertical scale, units of ln(fa)–ln(fc), where fa is the fractional cleavage at any bond in the presence of the
oligopeptide and fc is the fractional cleavage of the same bond in the control. Data are derived from quantitative analysis of several sequencing gels
like the one shown in Figure 7 and must be considered a set of average values.
NH2 [OP (Cbg-Cha)], 10 mM Acr-GABA-Arg-Cbg-
Cha-Chi-Chi-Tal-Arg-NH2 [OP-GABA-Acr] and 150
mM of Ac-Arg-Ual-Sar-Chi-Chi-Tal-Arg-NH2 [Ref-OP
(Ual-Sar)].
Ac-Arg-Ual-Sar-Chi-Chi-Tal-Arg-CONH2. The relative
pyrimidine selectivity of the peptide Ac–Arg–Cbg–Cha–
Chi–Chi–Tal–Arg–CONH2 is lost and the specificity of
the reference peptide is restored.
The differential cleavage plots show that the peptide Ac-
Arg-Cbg-Cha-Chi-Chi-Tal-Arg-NH2 binds to several
sequences of varying length (from two to nine base pairs).
The sites with pronounced inhibited cleavage are situated
at nucleotide positions 95–103 (50-TTCCTGTGT-30) and
155–160 (50-TCCTCT-30). Other sites like 106–107 (50-
CT-30), 113–117 (50-TCATA-30), 127–128 (50-TT-30),
135–136 (50-AT-30), 139–140 (50-CT-30) and 147–149 (50-
TTG-30) have lesser base pairs involved. The sequences
contain many sequential cytosines and thymines, indi-
cating a relative pyrimidine specificity. To some extent,
practically all multiple cytosine and thymine sequences
are targeted. Comparison with the cleavage pattern of
the reference oligopeptide Ac-Arg-Ual-Sar-Chi-Chi-Tal-
Arg-NH2 shows that the pattern of the Cbg-Cha pep-
tide is strongly pronounced showing multiple sites of
increased protection from cleavage, in contrast to the
single low intensity protected site of the reference pep-
tide. Furthermore, pyrimidine specificity can be clearly
identified for the modified oligopeptide in comparison
with the Ual-Sar containing compound. The more
hydrophobic side chains of Cbg and Cha and/or the
different groove geometry and recognition pattern at the
level of polypyrimidine tracts, could account for this
pyrimidine selectivity, opening new possibilities to
investigate the usefulness of analogous compounds in
dsDNA sequence-specific targeting.
Discussion
With the sequencing of the human genome accomplished,
the functions of genes in cellular and in vivo models have
to be determined. Once a target has been validated, drug
discovery efforts will start. The demand for molecules
recognizing sequence-specifically dsDNA or RNA and
which can be used for gene functionalization and target
validation will increase exponentially during coming
years. Moreover, such molecules will provide a basis for
new and more selective therapeutic approaches. The
biochemical and structural studies in the past decade
have already indicated that the details of ligand-DNA
interaction are imposingly complex. The DNA sequence
can be read by recognition of hydrogen bonding pat-
terns along the floor of the grooves. Sequence depen-
dent local deformability of the DNA duplex may be
part of the recognition process. In addition, indirect
readout elements and non-specific interactions may be
involved in positioning the ligand in an entropy favored
orientation for sequence-specific recognition.
The selection of high binding ligands can be accom-
plished by screening large numbers of molecules to dis-
cover an initial lead with modest binding activity
followed by defining the key recognition elements for
maximal activity and the optimization of the lead
structure by synthesis of focussed libraries or of indivi-
dual compounds. The different functionalities of the
library might be placed on a template whose structure is
based upon an important molecular recognition motif in
a biological system. Therefore we started synthesis of
peptide libraries, consisting of modified amino acids
with unique side chains or constrained backbones.
The fact that in some cases only two base pair sequences
seem to be protected from cleavage, might suggest that
in these cases only part of the oligopeptide is bound,
while the other amino acids ensure a right positioning.
The acridine-GABA conjugate shows, however, not the
same inhibited cleavage pattern as with the oligopeptide
containing Cbg-Cha. Only two sites display pronounced
protection against DNase I, namely at nucleotide posi-
tions 105–113 (50-TCATAGCTG-30) and 134–141 (50-
TCTGCATG-30). The latter sequence corresponds to
the main protected site by the reference oligopeptide
In the first study we used solid phase screening to select
an initial lead sequence.10 In the second round, solution
phase screening was used to extend this sequence with
functionalities involved in the recognition of the target,

3408
Z. Zhang et al. / Bioorg. Med. Chem. 10 (2002) 3401–3413
yielding lead compounds as the peptide Ac-Arg-Ual-
Sar-Chi-Chi-Tal-Arg-CONH2.11 In the present study we
further optimized the binding making use of a library
deconvolution technique. We selected eight new unna-
tural amino acid residues containing a diversity of
functional groups and introduced them at the two ran-
dom positions of Ac-Arg-Y-X-Chi-Chi-Tal-Arg-
CONH2, as Ual and Sar seemed to contribute to a lesser
extent to the interaction capacity of the oligopeptide.
Since in water, correct hydrogen bond complementa-
rities may not add much to the stability of the complex
in contrast to hydrophobic interactions, several of the
selected amino acids had an apolar side chain.
well for tethering with the acridine intercalator. Based
upon gel retardation experiments, out of these ten new
peptides, a beneficial effect on the binding affinity was
shown upon connecting acridine to the peptide with the
GABA linker. The peptide Acr-GABA-Arg-Cbg-Cha-
Chi-Chi-Tal-Arg-CONH2 demonstrated an improve-
ment on the binding strength (Kd=2.1ꢀ10ꢁ6 M) by one
order of magnitude compared to the unconjugated pep-
tide (Kd=2ꢀ10ꢁ5 M). Hybrid molecules containing the
lead peptide structure Ac-Arg-Ual-Sar-Chi-Chi-Tal-
Arg-CONH2, were also endowed with affinities in the
mM range. These affinities are comparable to those
described for known DNA binding ligands. 9-Aminoa-
cridine-4-carboxamide, an intercalator with a limited
extended aliphatic chain, demonstrates a Kd value of
4.5ꢀ10ꢁ6 M for poly (dA-dT) and of 2.0ꢀ10ꢁ6 M for
poly (dG–dC).35 A synthetic combilixin, NetGA, com-
posed of netropsin and acridine, was reported to have a
Kd value of 1.1ꢀ10ꢁ6 M.26 Thus modification of the
lead peptide by introducing acridine with GABA as the
linker delivered a DNA binding agent with binding
strength similar to that of known compounds.
Out of a pool of 64 peptides (8 sublibraries of 8 com-
pounds) followed by the evaluation of 16 individual
peptides, 6 peptides were selected and their binding
strength was evaluated by measuring dissociation con-
stants in 0.1 M NaCl at pH 7.4. The peptide Ac-Arg-
Cbg-Cha-Chi-Chi-Tal-Arg-CONH2 demonstrated the
lowest Kd value, which was one order of magnitude
lower than the initial lead compound. The exchange of
Ual and Sar at position X and Y by, respectively, Cbg
and Cha enhanced the binding affinity of the peptide to
dsDNA. The new residues are more likely involved in
hydrophobic interactions due to the presence of cyclo-
hexane and a substituted benzene ring, whilst the former
residues possessing a uracil base are expected to be more
involved in hydrogen bonding interaction. Although the
importance of hydrogen bonding interactions in the
recognition of dsDNA has been widely accepted, their
significance for the determination of binding affinity
might be overestimated. Geometrical and electrostatic
factors might be the prime recognition system super-
imposed by hydrogen bonding and van der Waals
interaction.26 Recent studies within the combilexin
model demonstrated that the introduction of an inter-
calating heteropolyaromatic moiety to molecules might
increase the binding affinity to dsDNA.29 Moreover,
DNA binding molecules with a hybrid structure con-
taining a peptide fragment and an intercalating frag-
ment are well known e.g., actinomycin D, echinomycin,
triostin and luzopeptin. Previous research showed
already that conjugation of intercalators to the peptide
lead structures has a beneficial effect on the binding
strength.25 With this in mind, an intercalating chromo-
phore was tethered on our selected peptide. Because of
its synthetic accessibility, acridine is extensively con-
sidered as the prototype of the intercalators.30,31 Due to
the unique AT specificity of tilorone, different from
most intercalators with GC specificity,32 its core struc-
ture, fluorenone was applied as second intercalator for
evaluation. The spacer between the intercalator and the
peptide is very important and should be designed so
that it can deliver and optimally orient the intercalating
agent (without being too long) so that the molecule can
be bound to DNA with a minimal increase in
entropy.33,34 Hence, four sets of linkers composed of
glycine and b-alanine were connected to two inter-
calators and their influence on the binding affinity of the
reference peptide to the same dsDNA was determined.
Moreover, two flexible linkers with g-amino butyric acid
(GABA) and e-amino caproic acid (CAP) were used as
The sequence specificity of the strongest binding peptide
Ac-Arg-Cbg-Cha-Chi-Chi-Tal-Arg-NH2 was investi-
gated and compared with the reference compound Ac-
Arg-Ual-Sar-Chi-Chi-Tal-Arg-NH2. The footprints
showed a pronounced pyrimidine sequence selectivity,
which was not observed with the Ual-Sar containing
peptide. Probably the apolar side chains of Cbg and
Cha are important for this interaction. Footprinting
experiments with the newly developed conjugate acri-
dine-GABA-OP revealed that the pyrimidine selectivity
is lost, with the conjugate protecting analogous sites as
the reference peptide containing Ual-Sar. Perhaps the
intercalator is hampering interaction of the hydro-
phobic side chains of the Cbg and Cha amino acids,
probably responsible for the CT-specificity, restoring
the selectivity of the other part of the peptide, Chi-Chi-
Tal-Arg. Further optimization of the structure of these
peptides and conjugates and investigations of their
binding mode are the subject of present research.
Conclusion
In our search for sequence-specific dsDNA binding
ligands, the oligopeptide Ac-Arg-Ual-Sar-Chi-Chi-Tal-
Arg-NH2 with moderate affinity and selectivity was
modified. The amino acids Ual and Sar, displaying less
interaction capacities than the other side chains
involved, were replaced with new unnatural building
blocks. The oligopeptides with more hydrophobic side
chains on these positions were endowed with a 10-fold
increased affinity. The sequence specificity of the stron-
gest binding peptide Ac-Arg-Cbg-Cha-Chi-Chi-Tal-
Arg-NH2 showed a pronounced pyrimidine sequence
selectivity, which was not observed with the Ual-Sar
containing peptide. Conjugation of an intercalator
destroyed this specificity and restored interaction with
the originally protected sequence by the Ual-Sar con-
taining compound. As such, new dsDNA ligands, based
on combinations of unnatural amino acids and inter-

Z. Zhang et al. / Bioorg. Med. Chem. 10 (2002) 3401–3413
3409
calators and showing medium affinities and variable
sequence selectivities have been developed. This
approach of modifying dsDNA binding oligopeptides
opens the possibility to develop, by combinatorial
selection, a new class of dsDNA binding agents with a
wide range of sequence selectivity.
piperidine, trifluoroacetic acid (TFA), diisopropylcarbo-
diimide (DIC) and 1-hydroxybenzotriazole (HOBt) were
supplied by ACROS (Geel, Belgium). Na-(9-Fluor-
enylmethyloxycarbonyl)-Nb-tert-butyloxycarbonyl-l-a-
2,3-diaminopropionic acid was synthesized as descri-
bed.16 Fmoc protection of the primary amine of both
GABA and CAP was realized under standard Schotten–
Baumann conditions. The two oligonucleotides which
served as target were synthesized on an Applied Biosys-
tem392 DNA synthesizer with phosphoramidites from
Applied Biosystems and were worked up as described
previously.36 T4 polynucleotide kinase and DNase I
were purchased from Gibco BRL and [g-32P]ATP was
supplied by ICN. NAP-51 columns were obtained from
Pharmacia.
Experimental
General methods
1
The H and 13C NMR spectra were recorded with a
Varian Gemini 200 spectrometer in the solvent indi-
1
cated. For H spectra taken in CDCl3 tetramethylsilane
(TMS) was used as reference. For spectra taken in
DMSO-d6, the solvent signal at 2.5 ppm was used as
reference. For the 13C spectra, the solvent peaks were
used as reference e.g., 77.00 ppm for CDCl3 and 39.60
ppm for DMSO-d6. Exact mass measurements were
performed on a quadrupole/orthogonal-acceleration
time-of-flight (Q/oaTOF) tandem mass spectrometer
(qTof 2, Micromass, Manchester, UK) equipped with a
standard electrospray ionization (ESI) interface. Sam-
ples were infused in a 2-propanol/water (1:1) mixture at
3 mL/min. In some cases, exact mass data were acquired
on a Kratos Concept IH double focussing mass spec-
trometer (Kratos, Manchester, UK) with a MASPEC 2
data system (MSS Ltd., Manchester, UK). Samples
were dissolved in a suitable matrix and ionized by liquid
secondary ion mass spectrometry (LSIMS). Melting
point was measured with a Buchi SMP-20. Precoated
Machery-Nagel Alugram1 SilG/UV 254 plates were
used for TLC and spots were examined with UV light,
sulfuric acid/anisaldehyde spray, or ninhydrine spray.
Column chromatography was performed on Acros silica
gel (0.06–200 nm). Anhydrous solvents were obtained as
follows: pyridine and N,N-diisopropylethylamine
(DIEA) were refluxed overnight over potassium
hydroxide and distilled; N,N-dimethylformamide
(DMF) was stored on molecular sieves for 3 days and
was tested for the absence of dimethylamine by the
bromophenol test prior to use. Tetrahydrofuran (THF)
was refluxed over sodium/benzophenone under nitrogen
and distilled. CH3CN for HPLC was purchased from
Rathburn (grade S) and water for HPLC purification
was distilled twice. Rink amide MBHA resin was sup-
plied by Novabiochem (Laufelfingen, Swizerland). Di-
chloromethane (DCM), N,N-dimethylformamide, acetic
anhydride (Ac2O) and pyridine were obtained from
BDH (Poole, UK). 1-Hydroxy-7-azabenzotriazole
(HOAt), Fmoc-b-alanine (b-Ala), and Fmoc-glycine
(Gly) were purchased from Advanced ChemTech
(Louisville, Kentucky). trans-4-Hydroxy-l-proline, di-
tert-butyldicarbonate [(Boc)2O], iodomethane, uracil,
benzoyl chloride, diethylazodicarboxylate (DEAD), tri-
phenylphosphine (PPh3), 9-fluorenylmethyloxycarbonyl
chloride, 1,3-thiazolidine-2-thione (TT), 1H-benzimida-
zole-5-carboxylic acid, 3-pyridine carboxylic acid,
3-chlorobenzaldehyde, 4-(trifluoromethyl)benzaldehyde,
sodium cyanoborohydride (NaCNBHf>3), acridine-9-
carboxylic acid, 9-fluorenone-2-carboxylic acid, g-amino
butyric acid (GABA) and 6-amino caproic acid (CAP),
Synthesis of the unnatural amino acids. N-tert-Butoxy-
carbonyl-cis-4-(N3-benzoyluracil-1-yl)-L-proline methyl
ester (3). Protection of the a-amino function of trans-4-
hydroxy-l-proline with a Boc group and protection of
the carboxylic function as methyl ester was described
previously.14 To a suspension of the alcohol 2 (0.245 g,
1.0 mmol), N3-benzoyluracil (0.216 g, 1.0 mmol) and
triphenylphosphine (294 mg, 1.1 mmol) in dry THF (10
mL) was added DEAD (182 mL, 1.1 mmol) dropwise at
ꢁ15 ꢃC. The reaction mixture was stirred at room tem-
perature overnight. The clear solution was evaporated
to dryness and the residue was purified by column
chromatography (CH2Cl2–acetone 95:5) to give N-tert-
butoxycarbonyl-cis-4-(N3-benzoyluracil-l-yl)-l-proline
methyl ester 3 as a foam (0.363 g, 0.82 mmol, 82%).
1H NMR (CDCl3) d 1.41 (s, 9H, Boc–CH3), 2.00–2.25,
2.65–2.95 (m, 2H, 3-H), 3.60–3.70, 3.85–4.10 (m, 2H, 5-
H), 3.72 (s, 3H, Me ester), 4.25–4.45, 5.10–5.30 (m, 2H,
2-H, 4-H), 7.20–8.00 (m, 7H, aromatics) ppm.
13C NMR (CDCl3) d 172.88 (ester–CO), 168.60 (Bz–
CO), 161.71 (uracil C-4), 156.70 (Boc-CO), 149.93 (ura-
cil C-2), 140.49 (uracil C-6), 135.15 (benzoyl p-CH),
131.39 (benzoyl-C), 130.45 (benzoyl m-CH), 129.17
(benzoyl o-CH), 102.80 (uracil C-5), 81.28 (Boc-C),
58.60 (C-2), 52.96 (C-5), 52.41 (ester CH3), 49.86 (C-4),
36.90, 34.38 (C-3), 28.07 (Boc–CH3) ppm. Exact mass
(LSIMS, thioglycerol) calcd for C22H25N3O7 [M+H]+
444.1770; found 444.1795.
Nꢀ-(9-Fluorenylmethyloxycarbonyl)-cis-4-uracil-L-proline
(4). To a solution of compound 3 (0.85 g, 1.92 mmol) in
100 mL of water and 1,4-dioxane (1/1, v/v) was added
sodium hydroxide (0.38 g, 9.60 mmol). After 1.5 h stir-
ring at room temperature, the solution was acidified
with 0.1 N HCl to pH 3 and the white suspension was
extracted three times with ethyl acetate. The combined
organic layer was dried on anhydrous MgSO4, filtered
and evaporated to dryness. The solid was treated with a
mixture of TFA and CH2Cl2 (1/1, v/v) at room tem-
perature over 2 h. After evaporation, the resulting foam
was dissolved in 10% Na2CO3 (5 mL). To this solution
was added a solution of 9-fluorenylmethyloxycarbonyl
chloride (0.60 g, 2.30 mmol) in 5 mL 1,4-dioxane at
0 ꢃC. The reaction mixture was stirred overnight and
then poured into 100 mL of H2O and the solution was

3410
Z. Zhang et al. / Bioorg. Med. Chem. 10 (2002) 3401–3413
extracted three times with diethyl ether. The aqueous
layer was acidified with 2 N HCl and the white suspen-
sion was extracted three times with ethyl acetate. The
combined organic layer was dried on anhydrous
MgSO4, filtered and evaporated to afford compound 4
(0.6 g, 1.34 mmol, 70%).
13C NMR (CDCl3) d 202.20 (TT-20), 171.27 (CO), 147.60
(OCON), 143.83, 143.53 (C-5), 134.54 (C-8), 129.84 (C-
9), 127.59, 127.10 (C-2), 126.85, 125.99 (C-6), 122.83 (C-
2), 120.43 (C-4), 116.73, 114.33 (C-7), 86.35 (Boc-C),
56.57 (TT-40), 29.71 (TT-50), 27.89 (Boc-CH3) ppm.
Exact mass (ESI-MS) calcd for C16H17N3O3S2
[M+H]+ 364.0789; found 364.0797.
1H NMR (CDCl3) d 1.90–2.45 (m, 2H, 3-H), 3.29–3.90
(m, 2H, 5-H), 4.10–4.60 (m, 5H, 2-H, 4-H, Fmoc-CH2,
Fmoc-90-H), 7.00–7.90 (m, 10H, aromatics) ppm.
(2S)-3-{[(1-{[(1,1-Dimethylethyl)oxy]carbonyl}-1H-ben-
zimidazol-5-yl)carbonyl]amino}-2-({[9H-fluoren-9-ylme-
thyloxy]carbonyl}amino)propanoic acid (6). Na-(9-
Fluorenylmethyloxycarbonyl)-Nb-tert-butyloxycarbonyl-l-
a-2,3-diaminopropionic acid (2.0 g, 4.7 mmol) was
treated with TFA in CH2Cl2 (1/1, v/v) at room tem-
perature over 2 h. After evaporation, the resulting foam
was dissolved in saturated NaHCO3 (10 mL). To this
solution was added a solution of compound 8 (1.62 g,
4.47 mmol) in 10 mL of 1,4-dioxane. The reaction
mixture was stirred overnight at room temperature and
the yellow color disappeared. The suspension was acid-
ified with 2 N HCl to pH 3 and extracted three times
with ethyl acetate. The combined organic layer was dried
over anhydrous MgSO4, filtered and evaporated. The
crude product was purified by column chromatography
(CH2Cl2–MeOH 95:5) and crystallized from diethyl
ether to afford compound 6 (2.0 g, 3.58 mmol, 80%).
13C NMR (CDCl3) d 172.90 (COOH), 163.37 (uracil C-
4), 154.15 (OCONH), 151.20 (uracil C-2), 143.89
(Fmoc-100), 142.64 (uracil C-6), 140.91 (Fmoc-110),
127.95 (Fmoc-30), 127.41 (Fmoc-20), 125.46 (Fmoc-10),
120.33 (Fmoc-40), 101.55 (uracil C-5), 67.10 (Fmoc-
CH2), 57.45 (C-2), 53.65 (C-5), 48.52 (C-4), 46.70
(Fmoc-90), 34.08, 32.92 (C-3) ppm.
Exact mass (ESI-MS) calcd for C24H21N3O6 [M+H]+
448.1508; found 448.1490; Elem. anal. for C24H21N3O6
C4H8O2 H2O calcd C 60.75, H 5.64, N 7.59; found C
60.77, H 5.36, N 7.33.
1-{[(1,1-Dimethylethyl)oxy]carbonyl}-1H-benzimidazole-
5-carboxylic acid (11). To a solution of 6.5 g (40 mmol)
of 1H-benzimidazole-5-carboxylic acid 10 in 100 mL of
10% Na2CO3 were added 60 mL of 1,4-dioxane and
10.5 g (48 mmol) of di-tert-butyldicarbonate at 0 ꢃC.
The reaction mixture was stirred overnight at room
temperature and was washed three times with diethyl
ether. The aqueous layer was acidified with 2 N HCl and
the white suspension was extracted three times with
ethyl acetate. The combined organic layer was dried
over anhydrous MgSO4, filtered and evaporated to
afford compound 11 (8.9 g, 34 mmol, 85%).
1H NMR (DMSO-d6) d 1.65 (s, 9H, Boc-CH3), 3.50–
3.80 (br, 2H, b-H), 4.10–4.40 (br, 4H, a-H, Fmoc-CH2,
Fmoc-90-H), 7.45–8.80 (m, 14H, aromatics, 2ꢀNH,
COOH) ppm.
13C NMR (DMSO-d6) d 172.45 (COOH), 166.68
(NHCO), 156.30 (Fmoc-OCONH), 147.65 (Boc-
OCONH), 144.89–114.05 (aromatic-C), 86.01 (Boc-C),
65.88 (Fmoc-CH2), 53.99 (C-a), 46.73 (Fmoc-90), 40.88
(C-b), 27.61 (Boc-CH3) ppm.
1H NMR (DMSO-d6) d 1.65 (s, 9H, Boc-CH3), 7.80–
8.80 (m, 4H, aromatics) ppm.
Exact mass (LSIMS, diethanolamine) calcd for
C31H30N4O7 [M-H]ꢁ 569.2036; found 569.2052; Elem.
.
anal. for C31H30N4O7 C4H8O2 calcd C 63.80, H 5.82, N
8.51; found C 63.76, H 6.11, N 8.58.
13C NMR (DMSO-d6) d 167.44 (COOH), 147.53, 147.01
(OCONH), 144.56, 143.74 (C-5), 134.30 (C-8), 131.08 (C-
9), 126.41, 125.40 (C-6), 121.67 (C-2), 120.15 (C-4), 115.99,
114.23 (C-7), 86.07 (Boc-C), 27.58 (Boc-CH3) ppm.
3-(3-Pyridinylcarbonyl)-1,3-thiazolidine-2-thione
(9).
Exact mass (ESI-MS) calcd for C13H14N2O4 [M+H]+
263.1031; found 263.1027.
Diisopropylcarbodiimde (1.9 mL, 12 mmol) was added
to a stirred solution of 3-pyridine carboxylic acid 12
(1.23 g, 10 mmol) and 1,3-thiazolidine-2-thione (1.43 g,
12 mmol) in CH2Cl2 (50 mL). The mixture was stirred
overnight at room temperature and a precipitate (diiso-
propylurea) was filtered off. Evaporation of the filtrate
under reduced pressure left an oily residue which was
purified by column chromatography (DCM–acetone
95:5) to give compound 9 as yellow powder (1.4 g, 6.2
mmol, 62%).
1,1-Dimethylethyl 5-[(2-thiono-1,3-thiazolidin-3-yl)carbo-
nyl]-1H-benzimidazole-1-carboxylate (8). Diisopro-
pylcarbodiimide (5 mL, 33 mmol) was added to a stirred
solution of 11 (8.6 g, 33 mmol) and 1,3-thiazolidine-2-
thione (4.7 g, 40 mmol) in CH2Cl2 (150 mL). The mix-
ture was stirred overnight at room temperature and a
precipitate (diisopropylurea) was filtered off. Evapora-
tion of the filtrate under reduced pressure left an oily
residue which was purified by column chromatography
(CH2Cl2–acetone 98:2) to give compound 8 as yellow
powder (7.2 g, 19.8 mmol, 60%).
1H NMR (CDCl3) d 3.47–3.55 (t, 2H, J=7.4 Hz, TT-
CH2), 4.55–4.62 (t, 2H, J=7.4 Hz, TT-CH2), 7.29–8.90
(m, 4H, aromatics) ppm.
1H NMR (CDCl3) d 1.71 (s, 9H, Boc-CH3), 3.47–3.55 (t,
2H, J=7.4 Hz, TT-CH2), 4.54–4.61 (t, 2H, J=7.4 Hz,
TT-CH2), 7.80–8.60 (m, 4H, aromatics) ppm.
13C NMR (CDCl3) d 202.37 (TT-20), 169.33 (CO),
152.57 (C-2), 149.84 (C-4), 136.55 (C-6), 129.99 (C-3),
123.07 (C-5), 56.09 (TT-40), 29.56 (TT-50) ppm.

Z. Zhang et al. / Bioorg. Med. Chem. 10 (2002) 3401–3413
3411
Exact mass (ESI-MS) calcd for C9H8N2O1S2 [M+H]+
225.0156; found 225.0145.
9-fluorenylmethyloxycarbonyl chloride at 0 ꢃC. The
reaction mixture was stirred overnight at room tem-
perature and was poured into 120 mL of H2O after
which the mixture was washed three times with diethyl
ether. The aqueous layer was acidified with 2 N HCl and
the white suspension was extracted three times with
ethyl acetate. The combined ethyl acetate layer was
dried on anhydrous MgSO4, filtered and evaporated to
dryness. The crude product was crystallized from metha-
nol to afford compound 14 (2.7 g, 5.95 mmol, 85%).
(2S)-2-({[9H-Fluoren-9-ylmethyloxy]carbonyl}amino)-3-
[(3-pyridinylcarbonyl)amino] propanoic acid (7). Na-(9-
Fluorenylmethyloxycarbonyl)-Nb-tert-butyloxycarbonyl-
l-a-2,3-diaminopropionic acid 5 (1.84 g, 4.33 mmol)
was treated with TFA in CH2Cl2 (1/1, v/v) at room
temperature for 2 h. After evaporation, the resulting
foam was dissolved in a saturated solution of NaHCO3
(10 mL). Hereto was added a solution of compound 9
(0.92 g, 4.12 mmol) in 10 mL of 1,4-dioxane. The reac-
tion mixture was stirred overnight at room temperature
and the yellow color disappeared. The suspension was
acidified with 2 N HCl to pH 3 and extracted three times
with ethyl acetate. The combined organic layer was
dried on anhydrous MgSO4, filtered and evaporated.
The crude product was purified by column chroma-
tography (CH2Cl2–MeOH 90:10) and crystallized from
diethyl ether to afford compound 7 (1.42 g, 3.30 mmol,
80%).
1H NMR (DMSO-d6) d 3.92 (s, 2H, a-CH2), 4.10–4.60
(m, 5H, Fmoc-CH2, Fmoc-90-H, Bn-CH2), 7.15–7.90
(m, 12H, aromatics) ppm.
13C NMR (DMSO-d6) d 171.08, 170.90 (COOH),
156.15, 155.91 (OCON), 143.95–120.24 (aromatic-C),
67.34, 67.01 (Fmoc-CH2), 51.10, 50.69 (Bn-CH2), 49.19,
48.55 (C-a), 46.85, 46.64 (Fmoc-90) ppm.
Exact mass (LSIMS, thioglycerol: NaOAc) calcd for
[M+2Na-H]+
C25H20N1O4F3
500.1062;
found
1H NMR (DMSO-d6) d 3.50–3.80 (br, 3H, b-H, a-H),
4.10–4.40 (br, 3H, Fmoc-CH2, Fmoc-90-H), 7.45–8.90
(m, 12H, aromatics) ppm.
500.1087; Elem. anal. for C25H20N1O4F3 calcd C 65.93,
H 4.43, N 3.08; found C 65.74, H 4.37, N 2.71.
{[(3-Chlorophenyl)methyl]amino}acetic acid (15). To a
suspension of 3-chlorobenzaldehyde (2.83 mL, 25
mmol) and glycine (1.5 g, 20 mmol) in 50 mL of dioxane
and H2O (1/1, v/v) was added sodium cyanoborohy-
dride (2.0 g, 30 mmol). The pH of the reaction mixture
was adjusted to 5–6 with 2 N HCl. After overnight stir-
ring at room temperature, the mixture was acidified to
pH 3–4 and evaporated to dryness. Hot MeOH was
added to the residue and the suspension was filtered.
The filtrate was absorbed on silica gel and purified by
column chromatography (CH2Cl2/MeOH/H2O 60:35:5)
yielding compound 15 (2.4 g, 12 mmol, 60%).
13C NMR (DMSO-d6) d 172.17 (COOH), 164.47
(NHCO), 156.33 (OCONH), 149.53–120.33 (aromatic-
C), 65.91 (Fmoc-CH2), 53.59 (C-a), 46.73 (Fmoc-90),
40.87 (C-b) ppm.
Exact mass (LSIMS, diethanolamine) calcd for
C24H21N3O5 [M-H]ꢁ 430.1402; found 430.1439; Elem.
anal. for C24H21N3O5 0.5H2O calcd C 65.45, H 5.03, N
9.54; found C 65.92, H 5.03, N 9.43.
({[4-(Trifluoromethyl)phenyl]methyl}amino)acetic
acid
(13). To a suspension of 4-(trifluoromethyl)benzaldehyde
(4mL,30mmol)andglycine (1.8 g, 24 mmol) in 50 mL of
dioxane and H2O (1/1, v/v) was added sodium cyano-
borohydride (2.4 g, 36 mmol). The pH of the reaction
mixture was adjusted to 5–6 with 2 N HCl. After over-
night stirring at room temperature, the mixture was
acidified to pH 3–4 and evaporated to dryness. Hot
MeOH was added to the residue and the suspension was
filtered. The filtrate was absorbed on silica gel and sub-
jected to column chromatography (CH2Cl2/ MeOH/H2O
60:35:5) yielding compound 13 (2.8 g, 12 mmol, 50%).
1H NMR (D2O) d 3.86 (s, 2H, a-CH2), 4.24 (s, 2H, Bn-
CH2), 7.30–7.55 (m, 4H, aromatics) ppm.
13C NMR (D2O) d 169.78 (COOH), 134.95 (C-5),
132.75 (C-1), 131.42 (C-4), 130.53 (C-6), 130.48 (C-2),
128.99 (C-3), 50.72 (Bn-CH2), 47.45 (C-a) ppm.
Exact mass (ESI-MS) calcd for C9H10N1O2Cl1
[M+H]+ 200.0478; found 200.0489.
([(3-Chlorophenyl)methyl]{[(9H-fluoren-9-ylmethyl)oxy]-
carbonyl}amino)acetic acid (16). To a solution of 0.42 g
(2.1 mmol) of 15 in 15 mL of 10% Na2CO3 was added
1H NMR (D2>O) d 3.95 (s, 2H, a-CH2), 4.36 (s, 2H,
Bn-CH2), 7.50–7.90 (d, 2H, J=8 Hz; d, 2H, J=8 Hz,
aromatics) ppm.
10 mL of 1,4-dioxane and 0.65
g (2.5 mmol)
of 9-fluorenylmethyloxycarbonyl chloride at 0C. The
reaction mixture was stirred overnight at room tem-
perature and was poured into 120 mL of H2O after
which the mixture was washed three times with diethyl
ether. The aqueous layer was acidified with 2 N HCl
and the white suspension was extracted three times with
ethyl acetate. The combined ethyl acetate layer was
dried on anhydrous MgSO4, filtered and evaporated
to dryness. The crude product was crystallized from
methanol to afford compound 16 (0.75 g, 1.78 mmol,
85%).
13C NMR (D2O) d 171.68 (COOH), 137.04–129.03
(aromatics), 52.60 (Bn-CH2), 49.57 (C-a) ppm.
Exact mass (ESI-MS) calcd for C10H10N1O2F3
[M+H]+ 234.0742; found 234.0744.
({[9H-Fluoren-9-ylmethyloxy]carbonyl}-({[4-(trifluorome-
thyl)phenyl]methyl}amino) acetic acid (14). To a solution
of 1.62 g (7.0 mmol) of 13 in 50 mL of 10% Na2CO3
was added 30 mL of 1,4-dioxane and 2.0 g (7.6 mmol) of

3412
Z. Zhang et al. / Bioorg. Med. Chem. 10 (2002) 3401–3413
1H NMR (DMSO-d6) d 3.91 (s, 2H, a-CH2), 4.15–4.55
(m, 5H, Fmoc-CH2, Fmoc-90-H, Bn-CH2), 7.15–7.90
(m, 12H, aromatics) ppm.
EtBr solution and a varying volume of oligopeptides
(individual or mixtures) to obtain the necessary con-
centrations. The appropriate volume of a Tris/NaCl
buffer (10 mM Tris–10 mM NaCl pH, 7.4) was added to
obtain a total volume of 100 mL per well. Before adding
the DNA in the wells, it was rendered double stranded
by taking equal amounts of both complementary
strands and heating them for 3 min at 80 ꢃC, at room
temperature for 5 min and at 4 ꢃC for 20 min. After
incubation at room temperature for 30 min, each well
was read on a FL600 Microplate Fluorescence reader,
with 530/25 nm as excitation wavelength and 590/35 nm
as the emission detection wavelength. Two control wells
(no agent=100% fluorescence, no DNA=0% fluores-
cence) were used per 12 samples. Fluorescence readings
are reported as% fluorescence relative to the control
wells. Generally two to three sets of measurements were
performed to calculate average values.
13C NMR (DMSO-d6) d 171.08, 170.87 (COOH),
156.09, 155.84 (OCON), 143.92–120.27 (aromatic-C),
67.37, 67.13 (Fmoc-CH2), 50.92, 50.50 (Bn-CH2), 49.01,
48.40 (C-a), 46.82, 46.64 (Fmoc-90) ppm.
Exact mass (ESI-MS) calcd for C24H20N1O4Cl1
[M+Na]+ 444.0979; found 444.0995; Elem. anal. for
C24H20N1O4Cl1 calcd C 68.33, H 4.78, N 3.32; found C
68.05, H 4.88, N 2.71.
Solid phase peptide synthesis of the library and
individual oligopeptides
The glycine substituted oligopeptides were prepared on
solid support as described previously.11 Fmoc-glycine (4
equiv) was coupled with the use of 4 equiv HOAt, DIC
and DIEA in DMF.
Solution phase screening of the library and individual
oligopeptides
The library and individual peptides corresponding to the
general structure Arg-Y-X-Chi-Chi-Tal-Arg-CONH2
were likewise assembled on solid support and the mix and
split method was applied.21 An amount of 10 mg (5.4
mmol) solid support (Rink amide MBHA resin) was
applied per peptide for library synthesis and a cocktail
of 4 equiv amino acid, 4 equiv HOAt, 4 equiv DIC and
4 equiv DIEA in DMF was used for couplings. For
reactions of amino acids with the peptoid monomers
Cbg and CF3, 2 equiv PyBop was added to the mixture.
Procedures used during the peptide synthesis cycle were
as described.11 Before each coupling, the unreacted amino
groups were capped using a mixture of pyridine/Ac2O/N-
methylimidazole 4:1:0.5 over 10 min while the Fmoc-pro-
tecting group was removed by 15 min treatment with 20%
piperidine in DMF. In the case of modification by, the
coupling cocktail was composed of 4 equiv intercalator, 4
equiv HATU, 4 equiv DIEA and DMF. The linkers g-
amino butyric acid (GABA) and 6-amino caproic acid
(CAP) were introduced using 4 equiv HOAt, 4 equiv
DIC, 4 equiv DIEA and 2 equiv PyBop in DMF. After
the final coupling cycle, the Fmoc-protecting group was
removed and the terminal amino group was acetylated.
The oligopeptides were released from the bead by treat-
ment with TFA/H2O 95:5 in the presence of 5% thioanisol
for two h at room temperature. The mixture was filtered
and the filtrate was evaporated and co-evaporated with
toluene, affording the peptides as solid material. In the
case of individual oligopeptides, purification by reversed
phase preparative HPLC (PLRP-S1 column) was imple-
mented. The peptides were eluted from the column using a
linear gradient from 40 to 100% B in 30 min, with a flow
rate of 3 mL/min [elute A: 0.1% TFA in CH3CN–H2O (5/
95); elute B: 0.1% TFA in CH3CN–H2O (80/20)], with
UV detection at 220 nm. The major peak was isolated and
analyzed by high resolution mass spectrometry.
The solution screening process was performed by gel
shift experiments as previously outlined.11 Both 32P-
radiolabeled strands of the target oligonucleotide were
separately dissolved in PBS-buffer (137 mM NaCl, 2.7
mM KCl, 10 mM Na2HPO4, pH 7.4) at a 2 mM con-
centration. Equal volumes of both solutions were mixed
and treated to allow hybridization of the DNA-strands
(heating and cooling). The oligopeptides (individual or
mixtures) and the conjugates were dissolved in an
appropriate volume of PBS. In a total reaction volume
of 5 mL (PBS), 1 mL of the oligonucleotide solution
(containing 1 pmol dsDNA) was mixed with the volume
of conjugate or oligopeptide solution, necessary to
acquire the desired concentrations. The mixture was
stored at 4 ꢃC for 2 h. To the ‘blank’, no peptide was
added. The mixtures were resolved at 10 ꢃC on a 15%
native polyacrylamide gel with TBE pH 7.4 running-
buffer at 2 W/gel over 3 h. The gels were quantitatively
imaged using a Cyclone Phosphorimager (Packard).
The degree of complex formation was quantified by
measuring the residual amount of free DNA using
OptiquantTM
.
Apparent dissociation constant determination by gel
mobility-shift assay
The dissociation constants of peptides were determined
by gel mobility-shift assay as described in the litera-
ture.11 A series of concentrations of the investigated
peptides were prepared. These solutions were incubated
with dsDNA solutions and the mixtures were subjected
to gel shift analysis. KD values were simply determined
as the concentration at which 50% of the target dsDNA
was mobility shifted. At least 3 independent gel shift
assays were performed for each compound under study
to determine the apparent Kd. Concentration-response
curves, obtained by analysis of the gel shifts, were fitted
with the use of the equation Y=Emax/[1+(Kd/C)nH])
with Y standing for the response (% shift), Emax for the
maximal response and nH for the Hill coefficient using
Graphpad Prism (Graphpad Software Inc.).
Ethidium bromide displacement test13
Wells of Costar black 96-well plates were loaded with 2
mL of a 50 mM dsDNA solution, 2 mL of a 0.35 mM

Z. Zhang et al. / Bioorg. Med. Chem. 10 (2002) 3401–3413
3413
DNase I footprinting
9. Hyrup, B.; Nielsen, P. E. Bioorg. Med. Chem. 1996, 4, 1 5.
10. Lescrinier, T.; Hendrix, C.; Kerremans, L.; Rozenski, J.;
Link, A.; Samyn, B.; Van Aerschot, A.; Lescrinier, E.; Eritja,
R.; Van Beeumen, J.; Herdewijn, P. Chem. Eur. J. 1998, 4, 3
425.
11. Chaltin, P.; Lescrinier, E.; Lescrinier, T.; Rozenski, J.;
Hendrix, C.; Rosemeyer, H.; Busson, R.; Van Aerschot, A.;
Herdewijn, P. Helv. Chim. Acta: In press.
12. Isshiki, H.; Akira, S.; Tanabe, O.; Nakajima, T.; Shima-
moto, T.; Hirano, T.; Kishimoto, T. Mol. Cell .Biol. 1990, 10,
2757.
13. Boger, D. L.; Fink, B. E.; Brunette, S. R.; Winston, C. T.;
Hedrick, M. P. J. Am. Chem. Soc. 2001, 123, 5878.
14. Zhang, Z.; Herdewijn, P. Curr. Med. Chem. 2000, 8, 517.
15. Lowe, G.; Vilaivan, T. J. Chem. Soc., Perkin Trans. 1
1997, 539.
16. Zhang, Z.; Van Aerschot, A.; Hendrix, C.; Busson, R.;
David, F.; Sandra, P.; Herdewijn, P. Tetrahedron 2000, 56, 2513.
17. Nagao, Y.; Miyasaka, T.; Seno, K.; Fujita, E.; Shibata,
D.; Doi, E. J. Chem. Soc., Perkin Trans. 1 1984, 2439.
18. Borch, R. F.; Hassid, A. I. J. Org. Chem. 1972, 37, 10
1673.
19. Carpino, L. A.; Han, G. Y. J. Org. Chem. 1972, 37, 22
3404.
20. Chang, C.-D.; Felix, A. M.; Jimenez, M. H.; Meienhofer,
J. Int. J. Peptide Protein Res. 1980, 15, 485.
Cleavage reactions by DNase I were performed essen-
tially according to the protocol outlined previously.11
The 271 dsDNA fragment used in these experiments
was derived from the vector pUC 19, which contained
the 14-mer target sequence for the screening assays. The
preparation of the vector and the 271 fragment was
performed as previously described.11 Reactions were
conducted in a total volume of 10 mL containing 2 mL
labeled DNA fragment, 5 mL of the buffered solution
with the investigated molecules in the appropriate con-
centrations. After a 30 min equilibration period, clea-
vage was initiated via addition of 3 mL of a DNase I
solution of which the concentration was adjusted to
obtain an enzyme attack of ꢄ30% of the starting
material (final enzyme concentrations of 0.0005 to 1 U/
10 mL reaction mixture). After 3 min, the reaction was
stopped by freeze drying. The DNA cleavage products
were resolved by polyacrylamide gel electrophoresis
under denaturing conditions. Following electrophoresis
(ꢄ2.5 h. at 65 W in TBE pH 8.0 buffered solution), gels
were soaked in 10% acetic acid for 5 min, transferred to
Whatmann 3 MM paper, dried under vacuum at 80 ꢃC
and then exposed to a storage phosphor screen (Phar-
macia). G-tracks were obtained by treatment with
dimethylsulphate (DMS), followed by heating in piper-
idine.
21. Lam, K. S.; Salmon, S. E.; Hersch, E. M.; Hruby, V. J.;
Kazmierski, W. M.; Knapp, R. Nature 1991, 82, 354.
22. Zhang, Z.; Van Aerschot, A.; Chaltin, P.; Busson, R.;
Herdewijn, P. Collect. Czech. Chem. Commun. 2001, 66, 923.
23. Rozenski, J.; Chaltin, P.; Van Aerschot, A.; Herdewijn, P.
Rapid Commun. Mass Spectrom. 2002, 16, 10 982.
24. Bruice, T. W.; Lima, W. F. Biochemistry 1997, 36, 5004.
25. Chaltin, P.; Borgions, F.; Zhenyu, Z.; Rozenski, J.; Van
Aerschot, A.; Herdewijn, P. Bioconj. Chem.: Submitted.
Acknowledgements
26. Bailly, C.; He
27. Bailly, C.; Pommery, N.; Houssin, R.; He
Pharm. Sci. 1989, 78, 910.
28. Bailly, C.; Helbecque, N.; He
´
nichart, J. P. Bioconjugate Chem. 1991, 2, 6.
The work was supported by a grant from K.U.Leuven
(GOA 02/13) and from the Flemish Science Foundation
(G.0269.98 and G.0089.02). We thank Chantal Bier-
naux for editorial help.
´
nichart, J. P. J.
´
nichart, J. P.; Colson, P.;
Houssier, C.; Rao, K. E.; Shea, R. G.; Lown, J. W. J. Mol.
Recognit. 1990, 3, 26.
29. Eliadis, A.; Phillips, D. R.; Reiss, J. A.; Skorobogaty, A.
J. Chem. Soc. Chem. Commun 1988, 1049.
References and Notes
30. Denny, W. A. Anti-Cancer Drug Res. 1990, 4, 241.
31. Baguley, B. C. Anti-Cancer Drug Res. 1991, 6, 1.
32. Wilson, W. D.; Wang, Y.; Kusuma, S.; Chandrasekaran,
S.; Yang, N. C.; Boykin, D. W. J. Am. Chem. Soc. 1985, 107,
4989.
1. Gottesfeld, J. M.; Neely, L.; Trauger, J. W.; Baird, E. E.;
Dervan, P. B. Nature 1997, 387, 202.
2. Kielkopf, C. L.; White, S.; Szewczyk, J. W.; Turner, J. M.;
Baird, E. E.; Dervan, P. B.; Rees, D. C. Science 1998, 282, 111.
3. White, S.; Turner, J. M.; Szewczyk, J. W.; Baird, E. E.;
Dervan, P. B. J. Am. Chem. Soc. 1999, 121, 260.
4. Dervan, P. B. Science 1986, 232, 464.
33. Bailly, C.; Pommery, N.; Houssin, R.; He
Pharm. Sci. 1989, 78, 910.
34. Bailly, C.; Helbecque, N.; He
´
nichart, J. P. J.
´
nichart, J. P.; Colson, P.;
5. Goodsell, D.; Dickerson, R. E. J. Med. Chem. 1986, 29, 727.
6. Harshman, K. D.; Dervan, P. B. Nucleic Acids Res. 1985,
13, 4825.
7. Lane, M. J.; Dabrowiak, J. C.; Vournakis, J. N. Proc. Natl.
Acad. Sci. U.S.A. 1983, 80, 3260.
Houssier, C.; Rao, K. E.; Shea, R. G.; Lown, J. W. J. Mol.
Recognit. 1990, 3, 26.
35. Bailly, C.; Denny, W. A.; Mellor, L. E.; Wakelin, L. P. G.;
Waring, M. J. Biochemistry 1992, 31, 13.
36. Augustyns, K.; Vandendriessche, F.; Van Aerschot, A.;
Busson, R.; Urbanke, C.; Herdewijn, P. Nucleic Acids Res.
1992, 20, 4711.
8. Portugal, J.; Waring, M. J. Biochim. Biophys. Acta 1988,
949, 158.