Triple recognition of B-DNA by a neomycin - Hoechst 33258-pyrene conjugate

Article abstract of DOI:10.1021/bi9016796

Recent developments have indicated that aminoglycoside binding is not limited to RNA, but to nucleic acids that, like RNA, adopt conformations similar to its A-form. Wefurther sought to expand the utility of aminoglycoside binding to B-DNA structures by conjugating neomycin, an aminoglycoside antibiotic, with the B-DNA minor groove binding ligand Hoechst 33258. Envisioning a dual groove binding mode, we have extended the potential recognition process to include a third, intercalative moiety. Similar conjugates, which vary in the number of binding moieties but maintain identical linkages to allow direct comparisons to be made, have also been prepared.We report herein novel neomycin- and Hoechst 33258-based conjugates developed in our laboratories for exploring the recognition potential with B-DNA. Spectroscopic studies such as UV melting, differential scanning calorimetry, isothermal fluorescence titrations, and circular dichroism together illustrate the triple recognition of the novel conjugate containing neomycin, Hoechst 33258, and pyrene. This study represents the first example of DNA molecular recognition capable ofminor versus major groove recognition in conjunction with intercalation. 2009 American Chemical Society.

Full text of DOI:10.1021/bi9016796

452 Biochemistry 2010, 49, 452–469
DOI: 10.1021/bi9016796
Triple Recognition of B-DNA by a Neomycin-Hoechst 33258-Pyrene Conjugate†
Bert Willis and Dev P. Arya*
Laboratories of Medicinal Chemistry, Clemson University, Clemson, South Carolina 29634
Received September 25, 2009; Revised Manuscript Received November 21, 2009
ABSTRACT: Recent developments have indicated that aminoglycoside binding is not limited to RNA, but to
nucleic acids that, like RNA, adopt conformations similar to its A-form. We further sought to expand the utility
of aminoglycoside binding to B-DNA structures by conjugating neomycin, an aminoglycoside antibiotic, with
the B-DNA minor groove binding ligand Hoechst 33258. Envisioning a dual groove binding mode, we have
extended the potential recognition process to include a third, intercalative moiety. Similar conjugates, which
vary in the number of binding moieties but maintain identical linkages to allow direct comparisons to be made,
have also been prepared. We report herein novel neomycin- and Hoechst 33258-based conjugates developed in
our laboratories for exploring the recognition potential with B-DNA. Spectroscopic studies such as UV melting,
differential scanning calorimetry, isothermal fluorescence titrations, and circular dichroism together illustrate
the triple recognition of the novel conjugate containing neomycin, Hoechst 33258, and pyrene. This study
represents the first example of DNA molecular recognition capable of minor versus major groove recognition in
conjunction with intercalation.
Targeting genes with small molecules has gradually become a
reality in medicine, because of a deeper understanding of the
structure and sequence of DNA. For example, the action of drugs
such as anthracyclines and acridines (both DNA intercalators)
involves the binding and stabilization of ternary complexes of
DNA with DNA topoisomerases, a phenomenon unknown until
recently (1-4). The understanding of recognition in structure-
specific and DNA sequence-specific binding modes is crucial to
defining the principles for consistent development of gene
regulatory drugs.
The minor groove is perhaps the most studied of the potential
sites of recognition of DNA by small molecules over the past
decade (5). Exploiting solid phase synthesis techniques for
preparation of polyamides based on distamycin and netropsin,
Dervan (6, 7) has set forth rules for base pairing recognition
centralized on tandem interactions between imidazole and/or
pyrrole pairs formed via polyamide dimers or hairpin structures.
A promising feature of such polyamides, because of their high
affinities for B-DNA, is their ability to compete with specific
protein-DNA binding (8-13). The inhibition is not limited to
minor groove binding transcription factors; polyamide binding
may also succeed at “locking” DNA into its B-form such that
major groove binding proteins may fail in their attempt to distort
the B-DNA conformation upon binding (8).
pyrrole/imidazole or hydroxypyrrole/pyrrole (15) pairs to dis-
criminate base pair recognition, Dervan has shown that
50-TGGCATACCA-30, 50-TGGCATTACCA-30, and 50-TGGCA-
TATACCA-30 can bind hairpin dimers at subnanomolar con-
centrations (14).
Another well-studied minor groove binding structural motif is
bis(benzimidazole), originating from the early discoveries of
compounds such as Hoechst 33258 binding to DNA (16, 17).
These benzimidazoles are known for their potential to poison
DNA topoisomerases or to stabilize complexes of DNA topo-
isomerases that ultimately result in strand cleavage (18). Numerous
reports on benzimidazoles exist, ranging from tris(benzimid-
azoles) (19, 20) to furamidine-benzimidazole (21-23) structures,
as well as conjugates with distamycin (among other polyamides),
polyamines (24) (including our work with neomycin-conjugated
Hoechst 33258) (25, 26), and intercalators. Minor groove binder-
polyamine conjugates have prospects for dual binding of DNA, by
binding both the minor groove and the major groove (polyamine)
in tandem (24, 27-30). We recently showed that neomycin can be
driven to bind the B-DNA major groove when conjugated with the
minor groove binding Hoechst 33258 (25, 26).
A number of natural products have been shown to possess
structures characteristic of both groove binding and intercala-
tion. Selected examples (Figure 1) include rebeccamycin (indolo-
carbazole family), nogalamycin (anthracycline family), altro-
mycin (pluramycin family), and mithramycin (aureolic acid
family). Some include an additional elecrophilic moiety capable
of alkylating DNA. Furthermore, the majority have been shown
to display multiple DNA binding modes. For example, nogala-
mycin has recently been reported to thread DNA, exhibiting
intercalation by the central chromophore, with saccharide re-
gions extending from both ends to bind both major and minor
grooves.
Gene regulation by small molecules is driven by the develop-
ment of ligands that bind unique stretches of DNA. Extending
beyond the conventional polyamide hairpins and bis(benzimid-
azoles) for recognition of up to 6 bp, current research has
uncovered a few remarkable conjugates that have been shown
to recognize longer stretches of DNA. For example, head-to-
head hairpin polyamide dimers have been shown to recognize
DNA from 10 to 12 bp in length (14). Utilizing the ability of
†We are grateful to the National Science Foundation (CHE 0134792)
for financial support.
*To whom correspondence should be addressed. E-mail: dparya@
clemson.edu. Phone: (864) 656-1106. Fax: (864) 656-6617.
In addition to their well-known potential as RNA-targeted
drugs, we have utilized aminoglycosides for probing nucleic acid
recognition in a variety of pathways (31-42). By utilizing
r
pubs.acs.org/Biochemistry
Published on Web 12/09/2009
2009 American Chemical Society

Article
Biochemistry, Vol. 49, No. 3, 2010 453
F
IGURE 1: Structures of naturally occurring intercalator-groove binder hybrids that bind DNA.
synthetic neomycin intermediates for their covalent attachment
with intercalators [pyrene (43, 44), BQQ (44), and methidium
(45, 46)] and groove binders [Hoechst 33258 (25, 26) and
DNA (47, 48)], we have shown an increased binding affinity
for both A-form (or A-like) and B-form nucleic acids (25, 26). We
extend the recognition potential by combining major groove
binding of neomycin with both intercalating and minor groove
binding moieties in one molecule (49). Outlined in the forth-
coming article are advancements toward multi-recognition of
DNA by the successful design, synthesis, and binding studies of a
novel neomycin-Hoechst 33258-pyrene conjugate, termed
“NHP”. For the sake of comparison, control compounds lacking
one of the binding moieties were also prepared. A combination of
spectroscopic, calorimetric, and viscometric techniques were
utilized in probing the binding properties of these novel ligands.
Results show that NHP is the first ligand designed to simulta-
neously recognize DNA via all three recognition motifs: major
groove, minor groove, and intercalation.
4117860021] was purchased from Amersham Pharmacia
(Piscataway, NJ). All concentrations were determined spectro-
photometrically using extinction coefficients provided by the
supplier.
(ii) Chemicals. Neomycin B (sulfate salt) was purchased from
ICN pharmaceuticals and used without further purification (both
synthesis and binding experiments); DIAD (diisopropylazodi-
carboxylate) and 1,10-thiocarbonyldi-2(1H)-pyridone were pur-
chased from Aldrich Chemical Co. All other reagents and solvents
were purchased from Acros Organics and used without further
purification. Reaction solvents were distilled accordingly; methyl-
ene chloride and pyridine were distilled over calcium hydride, and
diethyl ether and 1,4-dioxane were distilled over sodium metal.
Conjugate NH was synthesized as reported previously (26).
Quantitation of Hoechst 33258 (ε338 = 42000 M-1 cm-1), NH
(ε342 = 39241 M-1 cm-1), NH (ε342 = 21085 M-1 cm-1), and
NHP (ε345 = 53674 M-1 cm-1) in aqueous solutions was
conducted using UV absorbance. To ensure stability and mini-
mum adsorption of Hoechst compounds to container walls, all
solutions were stored in nontransparent, polystyrene tubes.
Methods. (i) UV Melting. Samples of DNA (1 μM in
duplex for oligomers, 15 μM for polymers) were mixed with
ligand in buffer before being heated at 95 °C for 5 min and slowly
annealed to 20 °C. UV analysis was then conducted at 260 nm
EXPERIMENTAL PROCEDURES
Materials. (i) Nucleic Acids. Oligonucleotides were pur-
chased from Integrated DNA Technologies and used without
further purification. Polymeric DNA [poly(dA) poly(dT), lot
3

454 Biochemistry, Vol. 49, No. 3, 2010
Willis and Arya
from 20 to 95 °C at a heating rate of 0.2 °C/min. Tm values were
determined by first-derivative analysis (where clearly discerned)
using instrument software or via the midpoint of the transtion
where broad transitions were observed.
(ii) Fluorescence Titrations. Equilibrium binding experi-
ments were conducted using a Photon Technology International
(Lawrenceville, NJ) instrument at ambient temperature (22 °C).
A solution of ligand (serially diluted to working concentrations)
was excited at its respective λmax (slits of 3 nm), and resulting
emission curves (from 390 to 600 nm) were recorded after serial
docked closest to its 50 position for the Hoechst linker, was
conducted prior to attachment to Hoechst using a Monte Carlo
routine in MacroModel, using the AMBER* force field and
water as the solvent. A similar protocol was used for the Hoechst
33258 and pyrene moieties. Five of the six amines in neomycin
were protonated, in agreement with NMR studies of neo-
mycin (52, 53), as well as the terminal amine in the piperazine
ring of the Hoechst moiety. Three MC runs (for all single bond
torsional angles and 1000 steps each) starting from different
initial conformations were conducted, yielding the same global
minimum conformation. All the flexible bonds were selected. The
ligands were built and optimized in Macromodel, and the RESP
charges were derived using ab initio calculations with a 6-31 G*
basis set in Jaguar. The minimized complexes were then remini-
mized with the distance constraints removed to a root-mean-
additions of a concentrated DNA solution [poly(dA) poly(dT)
3
was 200 μM base-1]. After each addition, the solution was mixed
by pipetting up and down with a Pasteur pipet treated with
silanizing agent (Sigmacote) to avoid adsorption of ligand to the
glass. Sample equilibrium was monitored by continually exciting
and scanning the sample at different times and was usually
reached within 2 min. All data were normalized to account for the
(small) dilution of sample upon addition of substrate. For the
self-complementary 12mer experiments, individual samples of
varying ligand:DNA ratios were prepared, all with a constant
ligand concentration of 100 nM. A silanized (SigmaCote) or
polystyrene cuvette was used in all experiments. Corresponding
fluorescence data were fit using Kaleidagraph (see the Supporting
Information for further details of the curve fitting).
-2
square gradient of 0.08 kcal mol-1
A
to allow unfavorable
˚
contacts to be removed. In the next step, all the restrictions were
removed except the movement of ring atoms of the terminal bases
and energy minimization reached a convergence threshold of
0.05 kJ/mol in all cases. After structural minimization, the data
file was imported into WebLabViewer and the surface potential
of the complex was rendered using WebLabViewer.
Synthesis. For synthesis of all intermediates, see the Support-
ing Information.
(iii) CD Spectropolarimetry Titrations. Circular dichro-
ism (CD) experiments were conducted at 20 °C using a Jasco
J-810 spectropolarimeter. A concentrated solution of ligand
(i) HPA (3). Compound 10 was converted to the correspond-
ing imidate ethyl ester by vigorously bubbling anhydrous HCl into
a solution of 10 in dry EtOH for 30 min before being stored at 4 °C
overnight. After acid removal (bubbling nitrogen through the
solution into a saturated bicarbonate solution) and EtOH eva-
poration, the imidate was rinsed well with dry ether before being
dried in vacuo and used immediately without characterization.
To the corresponding imidate of 10 were added diamine 11 and a
3:1 dry EtOH/glacial acetic acid solution (2 mL) before the mixture
was refluxed for 8 h under argon. The acetic acid was removed, and
the crude mixture was loaded onto a column of silica before
purification using a gradient of MeOH (0 to 15%) in CH2Cl2
containing a trace of NEt3. The product N-trifluoroacetamide of 3 is
represented by an intense yellow-green fluorescence under long UV.
Rf = 0.6 (85:15:0.1 CH2Cl2/MeOH/NH4OH). Unknown benzi-
midazole byproducts, fluorescent purple under long UV, were
difficult to separate; therefore, the crude product was deprotected
as indicated below, with full characterization. MALDI-MS con-
firmed the product: [M]þ 1053.18, found 1053.00.
Deprotection of the corresponding trifluoroacetamide (14 mg,
13.3 μmol) was eliminated when the solution was stirred with a
2:1 mixture of MeOH and water (1.5 mL) containing K2CO3
(25 mg, 0.18 mmol) for 24 h before solvent removal, dissolution in
MeOH and filtration, and further coevaporation of MeOH with
toluene (2 ꢀ 3 mL) and purified over a short column of silica with
a gradient from 20 to 50% MeOH in CH2Cl2 until less polar
impurities were flushed off (which fluoresced purple under long
UV) before elution of a 1:1 CH2Cl2/MeOH mixture containing
2% NEt3. Product fluoresces yellow-green under long UV. After
evaporation of solvent and a rinse with ether, 12 mg (94%) of the
pure amine 3 was obtained: 1H NMR (500 MHz, CDCl3) δ 8.44
(d, 1H, J = 8.7 Hz), 8.25-7.90 (m, 12H), 7.48-7.38 (m, 1H),
7.30-6.93 (m, 5H), 4.13 (m, 2H), 3.80-3.10 (m, 22H, peaks
masked by solvent signal), 2.93-2.89 (m, 4H), 2.05-1.95 (m,
2H); MS (MALDI-TOF) m/z for C57H64N8O6 [M]þ 957.17,
found 957.77.
(200 μM) was added to a 30 μM solution of poly(dA) poly(dT)
3
and the mixture stirred constantly before being scanned from
450 to 210 nm. As with fluorescence experiments, equilibrium
was determined by periodically scanning the sample over a period
of time (up to 10 min) for the first few additions of ligand and was
reached within 3 min. An average of three scans were gathered for
each titration point. All data were normalized to account for the
(small) dilution of sample upon addition of ligand.
(iv) Viscometry. Viscosity measurements were conducted
using a Cannon-Ubbelohde 75 capillary viscometer submerged in
a water bath at 27 ( 0.05 °C. Flow times of buffer only followed
by DNA (1030 μL of a 100 μM solution in base pairs) were
recorded in triplicate before titrations of concentrated ligand
solutions (500 μM) with mixing by bubbling of air (using a pipet
bulb) through the solution. Flow times after each titration were
recorded in triplicate. In all cases, standard deviations were less
than 0.1 s. All solutions were in 10 mM sodium cacodylate buffer
containing 150 mM KCl and 1 mM MgCl2 (pH 7.2). Flow times
for buffer alone were in the range of 106 s, whereas the flow time
for DNA alone was approximately 112 s. Flow times for
titrations of ligand into DNA ranged from 110 to 108 s. The
viscosity for each titration was determined using a protocol
similar to that reported previously (50). For further explanation
of the analysis, please see the Supporting Information.
(v) Differential Scanning Calorimetry. DSC experiments
were conducted on a MicroCal VP-DSC instrument. Samples of
DNA (512 μL of a 100 μM solution in base pairs) in the absence
and presence of ligand (8 μM) were heated from 20 to 110 °C at a
scan rate of 30 °C/h. After buffer subtraction (usually negligible),
baseline adjustments, Tm assignments, and peak integrations
were conducted using Origin version 5.0. For further explanation
of the analysis, see the Supporting Information.
(vi) Computer Modeling. The DNA duplex d(CGCA-
AATTTGCG)2 was extracted from Protein Data Bank (PDB)
entry 296d (51). Conformational optimization of neomycin,
(ii) NHP 4. To the Boc-protected conjugate 13 (5 mg,
2.2 μmol) were added dry CH2Cl2 (1 mL) and trifluoroacetic

Article
Biochemistry, Vol. 49, No. 3, 2010 455
F
IGURE 2: Structures of novel conjugates used in this study.
acid (1 mL) and 5 μL of ethanedithiol as a scavenger before the
mixture was stirred at room temperature for 3 h. Evaporation of
liquids and addition of ether gave a light brown solid which, after
subsequent rinses with ether and being dried under vacuum, was
dissolved in deionized water and lyophilized to afford 6 mg of
product (quant.). The product was judged to be sufficiently pure
by analytical HPLC (Supelcosil LC18S column; eluent, 0.1%
aqueous TFA with a gradient from 0 to 100% of a 95:5
acetonitrile/water mixture over a period of 20 min; flow rate,
1.2 mL/min; ambient temperature; retention time of product,
8.9 min): 1H NMR (500 MHz, D2O with CF3CO2D) δ 8.66 (br s,
1H), 8.51 (br s, 1H), 7.96 (br s, 1H), 7.52-6.30 (d, 14H), 5.96-5.80
(m, 3H), 5.30 (s, 1H), 4.10-2.97 (m, 52H), 2.64-2.17 (m, 5H),
1.98-1.92 (m, 2H), 1.85-1.70 (m, 2H); MS (MALDI-TOF) m/z
for C83H113N15NaO18S2 [M þ Na]þ 1696.0, found 1696.6; UV
(water) λmax = 345 nm (ε = 53674 M-1 cm-1).
(iii) Compound 2 (NP). A solution of 15 (5 mg, 2.88 μmol)
was stirred in a 1:1 mixture of dry CH2Cl2 and TFA containing a
trace of ethanedithiol at room temperature, under N2 for 3 h. The
liquids were evaporated, and the solid was taken up in deionized
water before lyophilization to afford 5 mg of product (96%). The
product was judged to be sufficiently pure by analytical HPLC
(Supelcosil LC18S column; eluent, 0.1% aqueous TFA with a
gradient from 5 to 100% acetonitrile over a period of 20 min; flow
rate, 1.5 mL/min; ambient temperature; retention time of product,
14.6 min): 1H NMR (500 MHz, D2O) δ 8.38 (d, 1H, J = 9.2 Hz),
8.28 (t, 2H, J = 6.9 Hz), 8.23 (d, 2H, J = 8.3 Hz), 8.13 (s, 2H),
8.09 (t, 1H, J = 7.8 Hz), 7.99 (t, 1H, J = 7.8 Hz), 7.69 (t, 1H, J =
4.4 Hz), 7.54 (d, 1H, J = 6.4 Hz), 5.96 (d, 1H, J = 4.1 Hz), 5.32 (d,
1H, J = 4.2 Hz), 5.00 (s, 1H), 4.22-3.08 (m, 32H), 2.81 (d, 1H),
2.55-2.36 (m, 6H), 2.21-2.17 (m, 2H), 1.82 (q, 1H); MS (MALDI-
TOF) m/z for C52H91N9O15S2 [M þ 2H]þ 1136.37, found 1136.75;
UV (water) λmax = 343 nm (ε = 19625 M-1 cm-1).
RESULTS AND DISCUSSION
Design and Synthesis of a Triple Recognition Agent.
(i) Modeling. As imaginable when viewing the DNA structure,
placing the key binding moieties in separate areas would require a
linker with significant length and flexibility. Regarding this, we
opted for designing the conjugate based on a novel neomycin-
Hoechst 33258 conjugate, NH, which maintains a hexaethylene
glycol linkage between the two binding moieties. Though NH
binding to DNA was shown to be weaker than that of an
earlier Neomycin-Hoechst 33258 conjugate (26), we chose to
maintain the hexaethylene glycol spacer for NHP for a number of
reasons. (a) The linker is long enough to accommodate bind-
ing from all three moieties when spaced evenly along the hexa-
ethylene glycol chain (see Figure 2). (b) Direct comparisons
can be made with NH to isolate pyrene’s effect on DNA binding.

456 Biochemistry, Vol. 49, No. 3, 2010
Willis and Arya
(c) Due to both pyrene and Hoechst binding, the degree of
freedom of neomycin for outside solvent interactions is dimin-
ished when compared with the neomycin moiety in NH. (d) The
commercial availability of (terminal) diamines spaced with di-
and triethylene glycols fits conveniently with the synthetic design,
not only with NHP but also with NP (see Figure 2). (e) Molecular
modeling of NHP indicated a virtually ideal fit with such
distances between the binding moieties.
Pyrene is a popular fluorescence tag for studying or identifying
biochemical processes in vitro. The intense fluorescence upon
exposure to UV allows for sensitive measurements at nanomolar
concentrations. Various studies have indicated that pyrene
intercalates DNA, yet the strength and selectivity of binding
are significantly weaker than those of many well-known DNA
intercalators. The ease of its synthetic incorporation into bio-
molecules largely accounts for its popularity in biochemistry and
is the primary reason for its choice in these studies.
In modeling the triple recognition agent, we first docked the
three binding moieties in their respective areas: Hoechst in its
routine minor groove A/T stretch (as gathered from PDB entry
296d) (51), neomycin in the nearby major groove, and amino-
pyrene 2 bp above the terminal hydroxyl group of Hoechst 33258.
In docking neomycin, the 50-OH group of the ribose ring was
directed toward the minor groove, as would be required if
covalently linked to a minor groove binding ligand. Furthermore,
the amino position of aminopyrene was directed out of the minor
groove, as it would be when conjugated to the Hoechst moiety.
After structural minimization, the linker, based on hexaethylene
glycol, was constructed to connect the three binding moieties.
The structure was further minimized to provide the renderings in
Figure 3.
Inspection of the model of NHP with d(CGCAAATTTGCG)2
clearly indicates a snug fit of Hoechst within the DNA minor
groove and significant H-binding sites between neomycin and
specific bases, as well as with the strands that line the major
groove. A clear increase in the base pair distance is also noticeable
in the pyrene intercalation site. The resulting picture is somewhat
of a “clamp” on the DNA structure.
(ii) Synthesis of NHP and Control Ligands. The reaction
scheme for the preparation of NHP is depicted in Scheme 1.
Monotosylation of triethylene glycol using published procedures
to provide 5 and subsequent coupling with p-cyanophenol using
Mitsunobu conditions afforded 6 in good yields. Tosylate sub-
stitution with a monoprotected diamine 7 under basic conditions
provided intermediate 8. Reaction of the secondary amine of 8
with pyrene succinimide ester 9 followed by standard coupling
with 11 in an ethanol/acetic acid mixture provided the desired
Hoechst-pyrene-amine conjugate (HPA, 3) after subsequent
trifluoroacetyl removal (K2CO3, MeOH/H2O) at the terminal
amine (Scheme 1). Reaction of 3 with neomycin isothiocyanate
12 followed by NH-Boc deprotection of coupled product 13 gives
the desired triple recognition agent 4 (Scheme 2) as a trifluoro-
acetate salt.
Traditionally, modifications at the phenol hydroxy group of
Hoechst 33258 have been conducted prior to benzimidazole
formation. Therefore, functional groups within the linker (at the
phenol OH position) must be stable to anhydrous HCl, which is
used to convert the p-cyano position to a reactive imidate ester for
coupling with 1,2-diaminobenzenes to provide the benzimidazole.
Conveniently, the pyrene-amide intermediate, formed before
benzimidazole formation, is stable to such acidic conditions, as
F
IGURE 3: Computer model of NHP binding to d(CGCAAATTTG-
CG)2. The Hoechst 33258-DNA complex was extracted from PDB
entry 296d (51), and further model building was conducted in
MacroModel. Energy minimization was conducted before and after
the three binding moieties had been linked. The surface potential
rendering was created by exporting the PDB file of the minimized
structure into WebLabViewer Pro.
well as in a later step where base is used to deprotect the terminal
trifluoroacetyl group on the amine.
A novel neomycin-pyrene conjugate (NP) was also prepared for
the sake of comparison. NP maintains the identical linkage between
neomycin and Hoechst 33258. The synthetic route is outlined in
Scheme 3. For its preparation, monoprotected diamine 7 was first
reacted with pyrene succinimide ester 9, which was generated in situ,
before trifluoroacetyl deprotection and reaction with neomycin
isothiocyanate 12. N-Boc deprotection of 15 in a CH2Cl2/TFA
mixture afforded compound 2 (NP) as the trifluoroacetate salt.
(iii) UV Melting and DSC of NHP with the DNA Duplex
Suggest Significant Stabilization of the DNA Duplex. The
thermal stability of poly(dA) poly(dT) in the presence of various
3
ligands was first studied by UV melting analysis (Figure 4). In
addition to NH, NHP, and Hoechst 33258, additional compar-
isons were made with a novel neomycin-pyrene conjugate (NP)
and Hoechst-pyrene-amine conjugate (HPA), which is a pre-
cursor to NHP synthesis. A comparison including HPA allows a
direct analysis of neomycin’s (in NHP) effect on DNA stability.
A more detailed study of the binding effect by NHP was also
achieved with melting experiments consisting of a combination of
individual ligands (e.g., NH with aminopyrene in the same
solution, and so forth). When compared with those of NH and
Hoechst 33258, a noticeable shift in the Tm is lacking for NHP.
As indicated in Table 1, ΔTm values for NH and Hoechst 33258,
and combinations containing these ligands (see Figure 4), remain
in a similar range (ΔTm = 15 °C), while NHP melts exhibited a
less significant shift in Tm (ΔTm = 10 °C). This unanticipated
result prompted studies of the NHP-DNA complex by other
melting methods, most notably differential scanning calori-
metry (DSC). DSC is often used to monitor the dissociation of
DNA strands, providing valuable Tm and ΔH values for the
association-dissociation process.

Article
Biochemistry, Vol. 49, No. 3, 2010 457
Scheme 1: Preparation of the Hoechst 33258-Pyrene-Amine Conjugatea
aReagents and conditions: (i) p-cyanophenol, PPh3, DIAD, dioxane, 78%; (ii) 7, DMF, NaI, 61%; (iii) 9, DMAP, DMF, 65%;
(iv) (a) 11, HOAc, EtOH, 41%; (b) K2CO3, MeOH, H2O.
A previous differential scanning calorimetry (DSC) study of a
It is well-known that pyrene favors the alternating purine-
pyrimidine sequence over the nonalternating purine-pyrimidine
sequence polymers (54). It was therefore of interest to explore the
effect of NHP on the polymer duplex poly(dAdT)2. UV melting
analysis with poly(dAdT)2 was conducted in a fashion similar to
first-generation neomycin-Hoechst 33258 conjugate in a com-
plex with poly(dA) poly(dT) exhibited an excellent agreement
3
with Tm values determined in UV melting experiments (26). DSC
data gathered for the NHP-poly(dA) poly(dT) complex gave a
3
much different result (Figure 5). The relatively sharp peak
corresponding to duplex melting, otherwise present in DNA
alone, becomes virtually absent even up to temperatures as high
as 110 °C. Intermediate concentrations (less than saturating
conditions) indicate melting of unbound DNA (at 72 °C) with
one important difference from that of DNA melting with no
ligand present: ΔH values at the Tm were significantly lower. It
is possible that the Tm for NHP-bound DNA is above that
observable in DSC (and UV) experiments. These results, gath-
ered after melting experiments conducted with shorter DNA
oligomers (discussed below), struck a chord with what we initially
observed, namely, Tm values lower than those anticipated for
NHP in UV melting experiments.
that for poly(dA) poly(dT)2. NHP was found to exhibit a
3
significantly larger ΔTm (16 °C) under the conditions employed
versus that found for other ligands (Figure 6), including NP and
HPA, further illustrating the contributions of both Hoechst and
neomycin toward pyrene binding, a phenomenon not observed for
NP and less significant for HPA. Lastly, the direct comparison of
NHP with control samples containing a combination of lower-
order conjugates and individual ligand (Figure 6 and Table 2),
essentially providing all three binders in the DNA sample
(e.g., NP-Hoechst 33258), further illustrates the importance of
covalently linking the three moieties to achieve optimal recogni-
tion. The increase in the ΔTm of NHP with poly(dAdT)2 when
compared with poly(dA) poly(dT) can simply be a consequence
3

458 Biochemistry, Vol. 49, No. 3, 2010
Willis and Arya
Scheme 2: Preparation of Triple Recognition Agent NHPa
aReagents and conditions: (i) pyridine, DMAP, 52%; (ii) 1:1 TFA/CH2Cl2, quant.
of pyrene’s preferential binding to alternating purine-pyrimidine
sequences (54).
To test the effect of the conjugates on oligomeric DNA duplex
and HPA (Figure 9). These findings highlight the importance
of an extended stretch of A-T base pairs. Furthermore, fluore-
scence titrations have indicated that the neomycin conjugates
reported here prefer A/T stretches of 9 bp. Therefore, the central
A3T3 region in d(CGCAAATTTGCG)2 may be less favorable
stability, UV melting profiles of dA22 dT22 in the absence and
3
presence of various ligands were gathered in a fashion similar to
that used with polymeric DNA (Figure 7). A much more
than DNA oligomers such as dA22 dT22. Melting data (Figures 7
3
significant and noticeable shift in the Tm for dA22 dT22 with
and 8) with such DNA support this. As such, experiments with
control ligands as a combination of all three binders, as done with
3
NHP was apparent. When compared with other ligands, such as
Hoechst 33258, NH, NP, and HPA, the triple-binding agent
NHP was hailed as the strongest stabilizer of this DNA oligomer,
with a ΔTm of 24 °C. A concentration-dependent melting study
(Figure 8) illustrates the gradual shift in Tm values as ligand
concentrations are increased, with ΔTm values of up to 32 °C at
ligand concentrations of 3 μM. These results indicate a clear
distinction in binding effect for NHP, as solutions containing a
combination of individual ligands (NH with aminoypyrene,
neomycin with HPA, and NP with Hoechst 33258) exhibited
significantly lower Tm values (see Figure 7 and Table 3).
dA22 dT22 and polymeric DNA, exhibited melting profiles
3
similar to those of the individual ligands themselves (Figure 9
and Table 4).
(iv) Fluorescence Titrations Indicate Significant Binding
to Duplex DNA: Poly(dA) Poly(dT). Upon binding to
3
DNA, Hoechst 33258 exhibits a significant fluorescence enhance-
ment (17), due presumably to the loss of solvent exposure upon
snugly binding (van der Waals and H-bonding forces) within the
DNA groove (55). An established DNA minor groove binder
with A/T base pair specificity, Hoechst 33258 was disovered in
the early 1990s to display multiple modes of binding to polymeric
DNA (56). Binding constants were found to be on the order of
UV melting profiles with d(CGCAAATTTGCG)2 with NHP
exhibited results similar to those seen with Hoechst 33258, NH,

Article
Biochemistry, Vol. 49, No. 3, 2010 459
Scheme 3: Synthesis of NPa
aReagents and conditions: (i) (a) pyrenebutyric acid, TSTU, DMF, NEt3, 64%; (b) TFA/CH2Cl2, 96%; (ii) (a) neomycin isothiocyanate
12, pyridine, 62%; (b) 1:1 TFA/CH2Cl2, ethanedithiol, 96%.
Table 1: Tm Data for Poly(dA) Poly(dT) Melting in the Presence of the
3
Various Indicated Ligandsa
compound
Tm2f1 (°C)
ΔTm2f1 (°C)
none
72
72
86
73
78
86
78
82
87
81
85
0
0
neomycin
Hoechst 33258
aminopyrene
NP
14
1
6
NH
14
6
HPA
NHP
10
15
9
NH-aminopyrene
HPA-neomycin
NP-Hoechst 33258
13
aSamples of DNA (15 μM) were mixed with ligand (2 μM) in buffer
[10 mM sodium cacodylate, 0.5 mM EDTA, and 150 mM KCl (pH 7.2)]
before being heated at 95 °C for 5 min and slowly annealed to 20 °C before
UV analysis at 260 nm from 20 to 95 °C at a heating rate of 0.2 o/min.
Tm values were determined by first-derivative analysis (where clearly
discerned) using instrument software or via the midpoint of the transtion
where broad transitions were observed.
tions (56), fluorescence techniques have been primarily utilized
for probing Hoechst 33258 binding. Furthermore, because of its
tight binding to DNA, the requirement for experimental sub-
strate concentrations on the order of 1/Kb for gathering appro-
priate binding isotherms eliminates the utility of other spectro-
scopic techniques due to weak signal detection.
Fluorescence titrations of the novel conjugates, as well as
Hoechst 33258, with poly(dA) poly(dT) were conducted in a
3
F
IGURE 4: UV melting profiles of poly(dA) poly(dT) with various
3
fashion similar to that reported previously (56-58). The experi-
ments typically consisted of titrations of a fixed ligand con-
centration (10 nM) with a concentrated DNA sample, with
fluorescence emission scans gathered at each ligand:DNA ratio
after the appropriate equilibration time. The resulting binding
isotherm could then be fit to a theoretical model, providing
binding constants (Kb) for the respective ligands. Because of
multiple modes of binding of Hoechst 33258 with polymeric
DNA, with ligand:phosphate stoichiometries of up to 2:1 (56),
titrations of low (nanomolar) ligand concentrations with DNA to
ligands. Samples of DNA (15 μM) were mixed with ligand (2 μM) in
buffer [10 mM sodium cacodylate, 0.5 mM EDTA, and 150 mM KCl
(pH 7.2)] before being heated at 95 °C for 5 min and slowly annealed
to 20 °C before UV analysis at 260 nm from 20 to 95 °C at a heating
rate of 0.2 °C/min. Tm values were determined by first-deri-
vative analysis (where clearly discerned) using instrument software
or via the midpoint of the transtion where broad transitions were
observed.
108 (M-1) for the sequence-specific interaction (A/T stretch of
4-5 bp). Because of its aggregation at micromolar concentra-

460 Biochemistry, Vol. 49, No. 3, 2010
Willis and Arya
F
IGURE 6: UV melting profiles of poly(dAdT)2 with various ligands.
Samples of DNA (20 μM) were mixed with ligand (3 μM) in buffer
[10 mM sodium cacodylate, 0.5 mM EDTA, and 150 mM KCl
(pH 7.2)] before being heated at 95 °C for 5 min and slowly annealed
to 20 °C before UV analysis at 260 nm from 20 to 95 °C at a heating
rate of 0.2 °C/min. Tm values were determined by first-derivative
analysis (where clearly discerned) using instrument software or via
the midpoint of the transtion where broad transitions were observed.
Table 2: Tm Data for poly(dAdT)2 Melting in the Presence of Various
Indicated Ligandsa
compound
none
Tm2f1
ΔTm2f1
66
66
75
65
66
76
74
82
77
72
74
0
0
neomycin
Hoechst 33258
aminopyrene
NP
9
-1
0
NH
10
8
HPA
NHP
16
11
6
NH-aminopyrene
HPA-neomycin
NP-Hoechst 33258
8
aSamples of DNA (20 μM) were mixed with ligand (3 μM) in buffer
[10 mM sodium cacodylate, 0.5 mM EDTA, and 150 mM KCl (pH 7.2)]
before being heated at 95 °C for 5 min and slowly annealed to 20 °C before
UV analysis at 260 nm from 20 to 95 °C at a heating rate of 0.2 °C/min.
Tm values were determined by first-derivative analysis (where clearly
discerned) using instrument software or via the midpoint of the transtion
where broad transitions were observed.
F
IGURE 5: DSC profiles of 100 μM poly(dA) poly(dT) with increas-
3
ing concentrations of NHP: (a) 1 μM NHP, (b) 4 μM NHP, and
(c) 8 μM NHP. Buffer conditions were identical to those in UV
melting experiments. Samples were heated at a rate of 0.5 °C/min.
Values for Tm and ΔH were determined using Origin version 7.0
provided with the instrument.
yield low ligand:DNA ratios were conducted to simulate nearly
independent site binding characteristics. In this approach, the low
ligand:DNA ratio exhibits a fluorescence signal corresponding to
a sequence-specific binding event. Thus, the resulting binding
curve can be fit using an independent site model for tight binding,
provided the binding site size, N, is known. Once N is known, the
DNA (base pairs) concentrations at each titration point can be
divided by N to provide the binding site size that is necessary for
the determination of binding constant data.

Article
Biochemistry, Vol. 49, No. 3, 2010 461
Table 3: Tm Data for dA22 dT22 Melting in the Presence of Various
3
Indicated Ligandsa
ligand
Tm2f1
ΔTm2f1
none
48
48
0
0
neomycin
Hoechst 33258
aminopyrene
NP
70
22
2
50
53
5
NH
72b
64b
83
24b
16b
35
25
13b
21
HPA
NHP
NH-aminopyrene
HPA-neomycin
NP-Hoechst 33258
73
61b
69
aSamples of DNA (1 μM in duplex) were mixed with ligand (3 μM) in
buffer [10 mM sodium cacodylate, 0.5 mM EDTA, and 150 mM KCl (pH
6.8)] before being heated at 95 °C for 5 min and slowly annealed to 20 °C
before UV analysis at 260 nm from 20 to 95 °C at a heating rate of 0.2 °C/
min. Tm values were determined by first-derivative analysis (where clearly
discerned) using instrument software or via the midpoint of the transtion
where broad transitions were observed. bFor compounds 1 and 3, biphasic
transitions were observed (at an unbound Tm of 51 and the indicated Tm).
F
IGURE 7: UV melting profiles of dA22 dT22 with various ligands.
3
Samples of DNA (1 μM in duplex) were mixed with ligand (3 μM) in
buffer [10 mM sodium cacodylate, 0.5 mM EDTA, and 150 mM KCl
(pH 6.8)] before being heated at 95 °C for 5 min and slowly annealed
to 20 °C before UV analysis at 260 nm from 20 to 95 °C at a heating
rate of 0.2 °C/min. Tm values were determined by first-derivative
analysis (where clearly discerned) using instrument software or via
the midpoint of the transtion where broad transitions were observed.
F
IGURE 8: UV melting profiles of dA22 dT22 with NHP. Samples of
F
IGURE 9: UV melting profiles of d(CGCAAATTTGCG)2 with
3
DNA (1 μMinduplex) weremixed withligand (0, 1, 2, and 3 μMfrom
left to right) in buffer [10 mM sodium cacodylate, 0.5 mM EDTA,
and 150 mM KCl (pH 6.8)] before being heated at 95 °C for 5 min and
slowly annealed to 20 °C before UV analysis at 260 nm from 20 to
95 °C at a heating rate of 0.2 °C/min. Tm values were determined by
first-derivative analysis (where clearly discerned) using instrument
software or via the midpoint of the transtion where broad transitions
were observed.
various ligands. Samples of DNA (1 μM in duplex) were mixed with
ligand (1 μM) in BPES buffer [6 mM Na2HPO4, 2 mM NaHPO4,
1 mM EDTA, and 185 mM NaCl (pH 7.0)] before being heated at
95 °C for 5 min and slowly annealed to 20 °C before UV analysis at
260 nm from 20 to 95 °C at a heating rate of 0.2 °C/min. Tm values
were determined by first-derivative analysis (where clearly discerned)
using instrument software or via the midpoint of the transition where
broad transitions were observed.
In our experiments with Hoechst 33258, an N value of 10 was
used, as established in the literature (59). Separate experiments, in
which a constant DNA concentration was titrated with concen-
trated ligand to provide values of n (number of base pairs per

462 Biochemistry, Vol. 49, No. 3, 2010
Willis and Arya
Table 4: Tm Data for d(CGCAAATTTGCG)2 Melting in the Presence of
the Indicated Ligandsa
compound
Tm2f1
ΔTm2f1
none
54
54
68
54
72
62
71
69
56
68
0
0
neomycin
Hoechst 33258
aminopyrene
NH
14
0
18
8
HPA
NHP
17
15
2
NH-aminopyrene
HPA-neomycin
NP-Hoechst 33258
14
aSamples containing DNA (1 μM in duplex) and ligand (1 μM) in buffer
[6 mM Na2HPO4, 2 mM NaHPO4, 1 mM EDTA, and 185 mM NaCl
(pH 7.0)] were heated at 95 °C for 5 min before being slowly annealed to
room temperature before UV melting analysis (monitored at 260 nm from
20 to 95 °C at a heating rate of 0.2 °C/min.). Tm values were determined by
first-derivative analysis (where clearly discerned) using instrument software
or via the midpoint of the transtion where broad transitions were observed.
F
IGURE 11: Fluorescence titration of NH into poly(dA) poly(dT).
3
Small aliquots of concentrated ligand were added to DNA (2 μM)
before fluorescence analysis (excitation at 345 nm, emission scanning
from 370 to 600 at 1 nm/s; 4 nm slit width; T = 25 °C). Buffer: 10 mM
sodium cacodylate, 0.5 mM EDTA, and 150 mM KCl (pH 7.2).
F
IGURE 10: Fluorescence titration of NHP into poly(dA) poly(dT).
3
Small aliquots of concentrated ligand were added to DNA (2 μM)
before fluorescence analysis (excitation at 345 nm, emission scanning
from 370 to 600 at 1 nm/s; 4 nm slit width; T = 25 °C). Buffer: 10 mM
sodium cacodylate, 0.5 mM EDTA, and 150 mM KCl (pH 7.2).
ligand), were conducted and used for determining N. It has been
shown that N = 2(n - 1) (60). As indicated by a break in the
fluorescence signal when plotted against varying DNA (base
pairs):ligand ratios, n values of 9 for both conjugates NH and
NHP were observed (Figures 10 and 11). This value is reasonable
due to the presence of the neomycin moiety, which likely extends
beyond the Hoechst binding site and/or perturbs the DNA
environment to exclude binding further down the DNA lattice.
In the case for NHP, one would anticipate the binding site to be
larger than NH, due to pyrene intercalation. However, the free-
dom of the hexaethylene glycol-neomycin arm of NH potentially
extends further than that in NHP, due to a “locking in” of the
Hoechst and pyrene moieties to their preferred binding areas,
restricting the reach of the neomycin arm relative to that in NH.
Binding curves depicting the ligand-DNA interaction are
represented in Figures 12-14. For all ligands, as with NHP,
fluorescence enhancement with an increasing DNA concentra-
tion was observed. All curves, as suggested above, could be
successfully curve fit to provide binding constants for each
ligand. Resulting binding data are listed in Table 5. Similar to
that reported previously (57), Hoechst 33258 binding is approxi-
mately 1 ꢀ 108 M-1. Binding of the neomycin-Hoechst 33258
F
IGURE 12: Fluorescence-detected binding of NHP to poly-
(dA) poly(dT). Small aliquots (1.5-20 μL) of DNA (200 μM bp-1
)
3
wereadded toa solutionofligand (2.5mLofa10nMsolution) before
fluorescence analysis (excitation at 345 nm, emission scanning from
370 to 600 at 1 nm/s; 4 nm slit width; T = 25 °C). Buffer: 10 mM
sodium cacodylate, 0.5 mM EDTA, and 150 mM KCl (pH 7.2).
conjugate possessing the hexaethylene glycol linker (NH) exhib-
ited markedly less binding affinity (Kb = 0.3 ꢀ 108 M-1) than
the other conjugates, including Hoechst 33258. The triple recog-
nition agent, NHP, displayed the highest binding affinity (2.2 ꢀ
108 M-1) of all the ligands, including NH (1.6 ꢀ 108 M-1), which
has been shown to significantly stabilize poly(dA) poly(dT) over
3
Hoechst 33258. We therefore believe that the affinity of NHP for
poly(dA) poly(dT) can be partially attributed to pyrene inter-
calation.
3

Article
Biochemistry, Vol. 49, No. 3, 2010 463
F
IGURE 14: Fluorescence-detected binding of poly(dA) poly(dT)
F
IGURE 13: Fluorescence-detected binding of poly(dA) poly(dT)
3
3
and NH. Small aliquots (1-20 μL) of DNA (200 μM bp-1) were
added to a solution of ligand (2 mL of a 10 nM solution) before
fluorescence analysis (excitation at 338 nm, emission scanning from
390 to 600 at 1 nm/s; 4 nm slit width; T = 25 °C). Buffer: 10 mM
sodium cacodylate, 0.5 mM EDTA, and 150 mM KCl (pH 7.2).
with Hoechst 33258. Small aliquots (1-20 μL) of DNA (200 μM
bp-1) were added to a solution of ligand (2 mL of a 10 nM solution)
before fluorescence analysis (excitation at 338 nm, emission scanning
from 390 to 600 at 1 nm/s; 4 nm slit width; T = 25 °C). Buffer: 10 mM
sodium cacodylate, 0.5 mM EDTA, and 150 mM KCl (pH 7.2).
Table 5: Binding Data for Studied Ligandsa
d(CGCAAATTTGCG)2. Fluorescence titrations of the
novel ligands were also conducted with d(CGCAAATTTGCG)2,
an oligomer possessing a single site for Hoechst 33258 binding.
Like reported observations (56), the binding of Hoechst 33258
and similar ligands (61) was found to exhibit a 1:1 stoichiometry
by way of fluorescence mixing curves [Job plots (Supporting
Information)]. With a 1:1 stoichiometry, the fluoresecence-
detected binding isotherms, the resulting titrations could be fit
to an independent, single-site model for tight binding to provide
binding constants for each ligand.
conjugate
poly(dA) poly(dT) K (ꢀ107 M-1
)
A3T3 K (ꢀ107 M-1
)
3
NH
3.4 ( 0.9
22 ( 0.6
0.5 ( 0.3
3.5 ( 1.2
NHP
aValues for Kb were determined by nonlinear curve fitting of fluore-
scence titrations according to an independent site model using conditions
similar to those reported. Values for N (binding site size) were either used as
reported (Ht33258) or determined experimentally (see Methods).
Figures 15-17 depict the binding isotherms recorded for the
conjugates with d(CGCAAATTTGCG)2. Surprisingly, the bind-
ing of all the neomycin conjugates was found to be significantly
the affinity of Hoechst 33258, and the neomycin conjugates for
sequences in which a GC junction interrupts the AnTn stretch.
Hoechst 33258 has been shown to bind A/T stretches from 4 to
6 bp in length. Introduction of G/C base pairs has also been
shown to significantly weaken the binding of Hoechst 33258.
Therefore, studying a sequence with a GC junction can provide
deeper insight into neomycin’s role in binding. For example, does
binding enhancement by neomycin perturb the sequence specifi-
city of Hoechst? Furthermore, what effects do pyrene and neo-
mycin conjugation have on Hoechst binding a G-C base pair
breaking up the A-tract? Thus, we chose to study DNA duplex
d(CGCAAGCTTGCG)2.
Fluorescence titration experiments with d(CGCAAGCTT-
GCG)2 with Hoechst 33258 and NHP were conducted like those
experiments with the central A3T3 stretch (Figure 18). In the case
of Hoechst 33258, the observed Kb was significantly lower than
what was observed for d(CGCAAATTTGCG)2. The binding is
nearly 1000-fold weaker (3.7 ꢀ 105 M-1), validating previous
weaker than that found with Hoechst 33258 (1.8 ꢀ 108 M-1
;
comparable with reported values of 3 ꢀ 108 M-1) (62). Clearly,
neomycin has little preference for this A-tract duplex. Neomycin
dimers on the other hand have recently been shown to prefer A/T
rich duplexes (63). The larger binding site size of the ligand, absent
in this duplex, also contributes to the lower affinity. Enhanced
binding (vs that of Hoechst 33258) is observed in extended A/T
stretches [poly(dA) poly(dT)], suggesting neomycin’s role in the
3
binding event. What is absolutely clear, however, is the effect
of pyrene intercalation when comparing NH to NHP. There is a
7-fold increase in affinity observed upon going from NH to NHP
for the polynucleotide as well as the oligomer (Table 5).
(i) Breaking up the A/T Stretch: Specificity of Hoechst
Binding in the NHP Conjugate. To improve our knowledge of
the design of dual binding ligands, we found it necessary to probe

464 Biochemistry, Vol. 49, No. 3, 2010
Willis and Arya
F
IGURE 15: Fluorescence-detected binding of NHP with d(CGCA-
F
IGURE 16: Fluorescence-detected binding of Hoechst 33258 with
AATTTGCG)2. Samples of ligand (100 nM in 2 mL) mixed with
varying concentrations of DNA were analyzed for fluorescence
emission over a 390-600 nm range. Conditions: λexc = 345 nm;
6 nm slits; T = 20 °C, Buffer: BPES [6 mM Na2HPO4, 2 mM
NaH2PO4, 180 mM NaCl, and 2 mM Na2EDTA (pH 7.0)].
d(CGCAAATTTGCG)2. Samples of ligand (100 nM in 2 mL) mixed
with varying concentrations of DNA were analyzed for fluorescence
emission over a 390-600 nm range. Conditions: λexc = 338 nm; 4 nm
slits; T = 20 °C. Buffer: BPES [6 mM Na2HPO4, 2 mM NaH2PO4,
180 mM NaCl, and 2 mM Na2EDTA (pH 7.0)].
accounts which indicated a strong binding preference of Hoechst
ligands for AnTn regions of DNA. Comparison with HPA was
difficult due to the solubility of the ligand under the conditions
used. However, experiments with NHP indicated a more favor-
able binding interaction with this duplex, stabilizing it nearly
10-fold more than Hoechst 33258. Therefore, pyrene and neo-
mycin both play a clear role in binding to d(CGCAAGCTT-
GCG)2 and force Hoechst to bind within the groove (as indicated
by the observed fluorescence enhancement in these experiments)
with a greater impact than Hoechst 33258 alone.
characteristic of pyrene absorption appear to overlap (270-
400 nm, likely due to a π-π* absorption of the pyrene moiety)
with the Hoechst region. Since asymmetry is induced in both
chromophores upon binding to DNA, these data support the
hypothesis that both Hoechst 33258 and pyrene regions of NHP
are interacting with the DNA.
Pyrene binding to poly(dA) poly(dT) has been shown to occur
3
in an “external”, nonintercalative fashion that is characterized by
little change in Tm, viscosity, or circular dichroism in pyrene’s
absorption region (54, 66). Our observations with NHP in the
(ii) Circular Dichroism of Poly(dA) Poly(dT) in the
3
Presence of NHP Indicates Considerable Conformational
Changes in both DNA and the Hoechst-Pyrene Regions.
We utilized circular dichroism (CD) spectroscopy to study the
presence of poly(dA) poly(dT) suggest an intercalative binding
3
mode similar to that of poly(dAdT)2 (66, 67). Therefore it is likely
that a combination of conformational restrictions in NHP and
the resulting unwinding of the duplex by dual groove binding of
neomycin and Hoechst 33258 designates pyrene to an intercala-
tive binding mode. Virtually no change in the CD spectrum of the
DNA region (200-260 nm) suggests that DNA conformation is
relatively unperturbed upon ligand binding.
changes, if any, in the poly(dA) poly(dT) structure as well as in
3
the Hoechst moiety of the conjugate. Numerous reports have
explained binding-induced chirality of Hoechst 33258 by DNA,
typically indicated by a change in the CD signal at the λmax of
Hoechst 33258 UV absorbance (64, 65). In our studies, we
conducted titrations of ligand into a solution of DNA, with
CD scans of the solution taken after appropriate equilibration
times between ligand additions. The resulting scans were overlaid
to compare the CD spectra at different ligand:DNA ratios
(Figures 19-22). We found that there is a significant change in
(iii) Viscometric Titrations Suggest Intercalation by the
Pyrene Moiety. The experimental results at this point led us to
propose a multivalent recognition process for NHP. Equilibrium
binding data indicated an enhanced binding over both NH (1)
(which contains a linker nearly identical to NHP) and NH0
(a conjugate with a shorter linker than NH), previously shown to
significantly bind DNA better than Hoechst 33258 (26). Circular
dichroism results (Figure 22) suggested Hoechst and pyrene
interaction. We further chose to confirm pyrene intercalation
using viscometric analysis.
the CD signal in the poly(dA) poly(dT) region, which is repre-
3
sented by a negative band at 248 nm and a positive band at
260 nm. Also, in the region of 360 nm, there is an increasingly
positive CD signal as ligand:DNA ratios are increased, indicative
of complexation of Hoechst with the DNA. Additionally, peaks

Article
Biochemistry, Vol. 49, No. 3, 2010 465
F
IGURE 17: Fluorescence-detected binding of d(CGCAAATTTG-
F
IGURE 18: Fluorescence-detected binding curves of d(CGCAAG-
CG)2 with NH. Individual samples of ligand (100 nM in 2 mL) mixed
with varying concentrations of DNA were prepared immediately
after quantitation of both stock ligand and DNA solutions. After
sufficient mixing of solutions and equilibration time (30 min), each
sample was analyzed for fluorescence emission over the 390-600 nm
range. Conditions: λexc = 338 nm; 4 nm slits; T = 20 °C. Buffer:
BPES [6 mM Na2HPO4, 2 mM NaH2PO4, 180 mM NaCl, and 2 mM
Na2EDTA (pH 7.0)].
CTTGCG)2 binding to Hoechst 33258 (left) and NHP (right).
Individual samples of ligand (100 nM in 2 mL) mixed with varying
concentrations of DNA were prepared immediately after quantita-
tion of both stock ligand and DNA solutions. After sufficient mixing
of solutions and equilibration time (30 min), each sample was
analyzed for fluorescence emission over the 390-600 nm range.
Conditions: λexc = 338 nm; 4 nm slits; T = 20 °C. Buffer: BPES
[6 mM Na2HPO4, 2 mM NaH2PO4, 180 mM NaCl, and 2 mM
Na2EDTA (pH 7.0)].
Viscometry has long been utilized for investigating ligand-
substrate binding modes, particularly to confirm intercalation
events. Intercalation of nucleic acids results in an increase in the
helical length, because of space displacement between the base
pairs by the ligand. Nucleic acids of appropriate length are
rodlike, so an increase in helical length results in an increase in
solution viscosity (68). Thus, we conducted viscometric titrations
of neomycin, Hoechst 33258, NH, and NHP with poly(dA)
3
poly(dT) (Figure 23). The results were, at first, unanticipated
since groove binding molecules such as Hoechst 33258 with DNA
display little change in intrinsic solution viscosity. In all cases
except Hoechst 33258, the DNA solution viscosity decreased
when ligand was added. The most marked decrease was with NH,
whereas that with NHP was clearly higher. Similar to that
observed before (69, 70), the decrease in viscosity can be
attributed to groove binding, which, in contrast to intercalation,
can shorten the helix by compaction. This compacting of DNA is
clearly a result of neomycin interaction, further corroborating
neomycin’s role in the multirecognition process. However, the
binding degree of NHP can be extended further.
By subtracting the results of NH-DNA binding from
NHP-DNA binding in the viscosity experiments, we can ascer-
tain the effect of the pyrene moiety in NHP on the DNA. A closer
look, by this consideration, clears the view somewhat. By over-
laying the subtracted data [NHP minus NH (Figure 24)], we
observe a clear resemblance to theoretical intercalation for the
F
IGURE 19: CD scans of poly(dA) poly(dT) with increasing concen-
3
trations of Hoechst 33258. Samples of DNA (40 mM) were scanned
from 400 to 210 nm after serial additions of concentrated ligand with
stirring. Peaks around 360 nm correspond to ligand-DNA com-
plexation. Buffer: 10 mM sodium cacodylate, 0.5 mM EDTA, and
150 mM KCl (pH 7.2). T = 20 °C.
pyrene moiety in NHP. Therefore, there is a clear correlation
of NHP with intercalation of poly(dA) poly(dT). The slightly
lower value can be attributed to intercalation occurring less
periodically within the DNA lattice due to the large binding site
size (vs a simple intercalator that would intercalate every 3 bp).
3

466 Biochemistry, Vol. 49, No. 3, 2010
Willis and Arya
F
IGURE 23: Viscometric analysis of poly(dA) poly(dT) with various
3
ligands. DNA solutions (100 μM) were titrated with the respective
drug (200 μM), and corresponding flow times were recorded in
triplicate with a deviation of <0.1 s. Error bars are indicated for
each titration point. Buffer: 10 mM sodium cacodylate, 0.5 mM
EDTA, 150 mM KCl, and 1 mM MgCl2 (pH 7.2). T = 27 ( 0.05 °C.
F
IGURE 20: CD scans of poly(dA) poly(dT) with increasing concen-
3
trations of NH. Samples of DNA (40 mM) were scanned from 400 to
210 nm after serial additions of concentrated ligand with stirring.
Peaks around 360 nm correspond to ligand-DNA complexation.
Buffer: 10 mM sodium cacodylate, 0.5 mM EDTA, and 150 mM KCl
(pH 7.2). T = 20 °C.
F
IGURE 24: Subtracted viscosity data to convey the effect of pyrene
on solution viscosity. DNA solutions (100 μM) were titrated with the
respective drug (200 μM), and corresponding flow times were re-
corded in triplicate with a deviation of <0.1 s. Buffer: 10 mM sodium
cacodylate, 0.5 mM EDTA, 150 mM KCl, and 1 mM MgCl2 (pH
7.2). T = 27 ( 0.05 °C.
F
IGURE 21: CD scans of poly(dA) poly(dT) with increasing concen-
3
trations of NP. Samples of DNA (40 mM) were scanned from 400 to
210 nm after serial additions of concentrated ligand with stirring.
Buffer: 10 mM sodium cacodylate, 0.5 mM EDTA, and 150 mM KCl
(pH 7.2). T = 20 °C.
F
IGURE 25: Viscometric analysis of poly(dAdT)2 with various li-
gands. DNA solutions (100 uM) were titrated with the respective
drug (200 μM), and corresponding flow times were recorded in
triplicate. Error bars are indicated for each titration point. Buffer:
10 mM sodium cacodylate, 0.5 mM EDTA, 150 mM KCl, and 1 mM
MgCl2 (pH 7.2). T = 27 ( 0.05 °C.
F
IGURE 22: Circular dichroism of poly(dA) poly(dT) complexed
3
with NHP. A solution of DNA (30 μM) was titrated with small
aliquots of concentrated ligand (200 μM) before equilibration and
scanning from 450 to 210 nm. Peaks around 360 nm correspond to
ligand-DNA complexation. Solutions were in 10 mM sodium
cacodylate, 0.5 mM EDTA, and 150 mM KCl (pH 7.2). The
temperature was ambient.
presence in DNA binding. Thus, convincing evidence, from a
combination of analytical techniques, gives strong candidacy for
a DNA triple recognition agent in NHP.
Further viscometric analysis of alternating A/T base pair
intercalation was done using poly(dAdT)2. Unlike that seen with
Furthermore, the mere observation of viscosity decrease, as seen
in other accounts, confirms neomycin binding. Lastly, multiple
spectroscopic experiments have conveyed the Hoechst moiety’s
poly(dA) poly(dT), the viscosity of poly(dAdT)2 solutions was
3
found to increase upon titration of NHP (Figure 25), suggesting

Article
Biochemistry, Vol. 49, No. 3, 2010 467
an intercalating mode of binding by pyrene. Also, as illustrated in
Figure 25, NH titrations resulted in decreased solution viscosity.
At higher ligand concentrations, solution viscosity significantly
dropped in both NHP and NH experiments. Such a phenomenon
might be explained by an increase in the level of neomycin
binding along the DNA grooves, significantly compacting the
DNA as mentioned above. The difference in viscometry profiles
between the two polymeric DNAs can be a result of stronger
pyrene intercalation between the A-T steps in poly(dAdT)2,
our results illustrate the first steps in the development of multi-
valent (triple recognition) DNA binding molecules. Structural
studies and refinement of molecules such as NHP can now be
performed to better probe the molecular requirements for multi-
recognition of B-DNA.
SUPPORTING INFORMATION AVAILABLE
Synthetic procedures and characterization of all intermediates.
This material is available free of charge via the Internet at http://
pubs.acs.org.
stronger neomycin binding to poly(dA) poly(dT), or a combina-
3
tion of both. Altogether, these studies clearly illustrate neomycin
and pyrene’s role in the DNA binding event. Multiple spectro-
scopic experiments have conveyed the Hoechst moiety’s presence
in DNA binding. Thus, convincing evidence, from a combination
of analytical techniques, provides a strong candidacy for a first
step in DNA triple recognition agent in NHP.
REFERENCES
1. Chen, A. Y., Yu, C., Gatto, B., and Liu, L. F. (1993) DNA Minor
Groove-Binding Ligands: A Different Class of Mammalian DNA
Topoisomerase I Inhibitors. Proc. Natl. Acad. Sci. U.S.A. 90, 8131–
8135.
2. Redinbo, M. R., Champoux, J. J., and Hol, W. G. J. (1999) Structural
Insights into the Function of Type IB Topoisomerases. Curr. Opin.
Struct. Biol. 9, 29–36.
CONCLUSION
3. Redinbo, M. R., Stewart, L., Kuhn, P., Champoux, J. J., and Hol,
W. G. J. (1998) Science 279, 1504–1513.
We recently reported that the affinity of neomycin for nucleic
acids includes structures known to adopt A-like conformations.
We have since endeavored to explore neomycin’s utility in
binding to B-DNA to improve our understanding of the mole-
cular forces that dictate binding within the DNA major groove.
Specifically, a progression from dual recognition to triple recog-
nition has been explored by including an intercalating region
(pyrene) within the central spacer separating Hoechst 33258 and
neomycin. From the current study of the novel triple recognition
agent NHP, the following conclusions can be drawn. (1) NHP
significantly enhances DNA stability. UV melting experiments
with both polymeric and oligomeric DNA indicate significant
shifts in Tm when compared with those of samples in the absence
of ligand. The shifts were significantly greater than those with
control ligands NP and HPA, indicating contributions by all
three binding regions. DSC experiments indicate a significant
decrease in the ΔH of duplex melting at the unbound Tm,
indicating complexation Tm at a higher temperature (above the
limits of instrumental conditions). (2) NHP binds stronger to
B-DNA than Hoechst 33258 and other conjugates studied.
Fluorescence binding studies with NHP indicate a binding
affinity greater than those with Hoechst 33258, NH, and NH
4. Todd, A. K., Adams, A., Thorpe, J. H., Denny, W. A., Wakelin, L. P.
G., and Cardin, C. J. (1999) Major Groove Binding and ’DNA-
Induced’ Fit in the Intercalation of a Derivative of the Mixed
Topoisomerase I/II Poison N-(2-(Dimethylamino)ethyl)acridine-
4-carboxamide (DACA) into DNA: X-ray Structure Complexed
˚
to d(CG(5-BrU)ACG)2 at 1.3-A Resolution. J. Med. Chem. 42,
536–540.
5. Dervan, P. B. (2001) Molecular Recognition of DNA by Small
Molecules. Bioorg. Med. Chem. 9, 2215–2235.
6. Kielkopf, C. L., Baird, E. E., Dervan, P. B., and Rees, D. C. (1998)
Structural Basis for G C Recognition in the DNA Minor Groove.
Nat. Struct. Biol. 5, 104–109.
3
7. Kielkopf, C. L., White, S., Szewczyk, J. W., Turner, J. M., Baird,
E. E., Dervan, P. B., and Rees, D. C. (1998) A Structural Basis for
Recognition of A T and T A Base Pairs in the Minor Groove of
3
3
B-DNA. Science 282, 111–115.
8. Dickinson, L. A., Trauger, J. W., Baird, E. E., Dervan, P. B., Graves,
B. J., and Gottesfeld, J. M. (1999) Inhibition of Ets-1 DNA Binding
and Ternary Complex Formation between Ets-1, NF-κB, and DNA
by a Designed DNA-Binding Ligand. J. Biol. Chem. 274, 12765–
12773.
9. McBryant, S. J., Baird, E. E., Trauger, J. W., Dervan, P. B., and
Gottesfeld, J. M. (1999) Minor Groove DNA-Protein Contacts
Upstream of a tRNA Gene Detected with a Synthetic DNA Binding
Ligand. J. Mol. Biol. 286, 973–981.
10. Neely, L., Trauger, J. W., Baird, E. E., Dervan, P. B., and Gottesfeld,
J. M. (1997) Importance of Minor Groove Binding Zinc Fingers
within the Transcription Factor IIIA-DNA Complex. J. Mol. Biol.
274, 439–445.
11. Lenzmeier, B. A., Baird, E. E., Dervan, P. B., and Nyborg, J. K. (1999)
The Tax Protein-DNA Interaction is Essential for HTLV-I Transacti-
vation in Vitro. J. Mol. Biol. 291, 731–744.
with poly(dA) poly(dT). (3) The mode of binding of Hoechst to
3
DNA is the same in all ligands studied. Circular dichroism
experiments indicate similar spectral patterns for poly(dA) poly-
3
(dT) when titrated with either Hoechst 33258 or other Hoechst-
neomycin conjugates. (4) The optimal binding site for the
conjugates includes a contiguous stretch of nine A-T base
pairs. Fluorescence titrations indicated base pair:ligand ratios of
12. Winston, R. L., Ehley, J. A., Baird, E. E., Dervan, P. B., and
Gottesfeld, J. M. (2000) Asymmetric DNA Binding by a Homo-
dimeric bHLH Protein. Biochemistry 39, 9092–9098.
13. Wurtz, N. R., Pomerantz, J. L., Baltimore, D., and Dervan, P. B.
(2002) Inhibition of DNA Binding by NF-κB with Pyrrole-Imidazole
Polyamides. Biochemistry 41, 7604–7609.
9:1 for NHP. A higher binding constant for NH with poly(dA)
3
poly(dT) than with Hoechst 33258 was also observed. The importance
of the extended A/T site for strong binding was further illustrated
in studies with both shorter A/T stretches and interrupted A/T
stretches. NHP binding to d(CGCAAATTTGCG)2 exhibited a
10-fold lower binding affinity than Hoechst 33258, yet compar-
isons with d(CGCAAGCTTGCG)2 indicate a value for Kb
greater than that observed with NH or Hoechst 33258. Such
an increase in the level of binding can thus be attributed to
intercalation by the pyrene moiety. (5) Pyrene intercalation is
apparent using viscometric techniques. Comparison analysis of
NHP with NH, Hoechst 33258, and neomycin indicates an
increase in solution viscosity when pyrene (in NHP) is present,
a phenomenon consistent with DNA intercalation. Altogether,
14. Weyermann, P., and Dervan, P. B. (2002) Recognition of Ten Base
Pairs of DNA by Head-to-Head Hairpin Dimers. J. Am. Chem. Soc.
124, 6872–6878.
15. Renneberg, D., and Dervan, P. B. (2003) Imidazopyridine/Pyrrole
and hydroxybenzimidazole/pyrrole Pairs for DNA Minor Groove
Recognition. J. Am. Chem. Soc. 125, 5707–5716.
16. Bontemps, J., Houssier, C., and Fredericq, E. (1975) Physico-Chemi-
cal Study of the Complexes of “33258 Hoechst” with DNA and
Nucleohistone. Nucleic Acids Res. 2, 971–984.
17. Weisblum, B., and Haenssler, E. (1974) Fluorometric Properties of the
Bibenzimidazole Derivative Hoechst 33258, a Fluorescent Probe
Specific for AT Concentration in Chromosomal DNA. Chromosoma
46, 255–260.
18. Pilch, D. S., Xu, Z., Sun, Q., Lavoie, E. J., Liu, L. F., Geacintov, N. E.,
and Breslauer, K. J. (1996) Characterizing the DNA Binding Modes
of a Topoisomerase I-Poisoning Terbenzimidazole: Evidence for both

468 Biochemistry, Vol. 49, No. 3, 2010
Willis and Arya
Intercalative and Minor Groove Binding Properties. Drug Design
Discovery 13, 115–133.
19. Xu, Z., Li, T. K., Kim, J. S., LaVoie, E. J., Breslauer, K. J., Liu, L. F.,
and Pilch, D. S. (1998) DNA Minor Groove Binding-Directed
Poisoning of Human DNA Topoisomerase I by Terbenzimidazoles.
Biochemistry 37, 3558–3566.
20. Aymami, J., Nunn, C. M., and Neidle, S. (1999) DNA Minor Groove
Recognition of a Non-Self-Complementary AT-Rich Sequence by a
Tris-Benzimidazole Ligand. Nucleic Acids Res. 27, 2691–2698.
21. Wang, L., Bailly, C., Kumar, A., Ding, D., Bajic, M., Boykin, D. W., and
Wilson, W. D. (2000) Specific Molecular Recognition of Mixed Nucleic
Acid Sequences: An Aromatic Dication that Binds in the DNA Minor
Groove as a Dimer. Proc. Natl. Acad. Sci. U.S.A. 97, 12–16.
22. Wang, L., Carrasco, C., Kumar, A., Stephens, C. E., Bailly, C.,
Boykin, D. W., and Wilson, W. D. (2001) Evaluation of the Influence
of Compound Structure on Stacked-Dimer Formation in the DNA
Minor Groove. Biochemistry 40, 2511–2521.
23. Wang, L., Kumar, A., Boykin, D. W., Bailly, C., and Wilson, W. D.
(2002) Comparative Thermodynamics for Monomer and Dimer
Sequence-Dependent Binding of a Heterocyclic Dication in the
DNA Minor Groove. J. Mol. Biol. 317, 361–374.
24. Satz, A. L., and Bruice, T. C. (2000) Synthesis of Fluorescent
Microgonotropens (FMGTs) and their Interactions with dsDNA.
Bioorg. Med. Chem. Lett. 8, 1871–1880.
25. Arya, D. P., and Willis, B. (2003) Reaching into the Major Groove of
B-DNA: Synthesis and Nucleic Acid Binding of a Neomycin-Hoechst
33258 Conjugate. J. Am. Chem. Soc. 125, 12398–12399.
26. Willis, B., and Arya, D. P. (2006) Recognition of B-DNA by
Neomycin-Hoechst 33258 Conjugates. Biochemistry 45, 10217–10232.
27. Satz, A. L., and Bruice, T. C. (2001) Recognition of Nine Base Pairs in
the Minor Groove of DNA by a Tripyrrole Peptide-Hoechst Con-
jugate. J. Am. Chem. Soc. 123, 2469–2477.
28. Satz, A. L., and Bruice, T. C. (2002) Recognition of Nine Base
Pair Sequences in the Minor Groove of DNA at Subpicomolar
Concentrations by a Novel Microgonotropen. Bioorg. Med. Chem.
10, 241–252.
44. Arya, D. P., Xue, L., and Tennant, P. (2003) Combining the Best in
Triplex Recognition: Synthesis and Nucleic Acid Binding of a BQQ-
Neomycin Conjugate. J. Am. Chem. Soc. 125, 8070–8071.
45. Shaw, N. N., Xi, H., and Arya, D. P. (2008) Molecular Recognition of
a DNA:RNA Hybrid: Sub-Nanomolar Binding by a Neomycin-
Methidium Conjugate. Bioorg. Med. Chem. Lett. 18, 4142–4145.
46. Shaw, N. N., and Arya, D. P. (2008) Recognition of the Unique
Structure of DNA:RNA Hybrids. Biochimie 90, 1026–1039.
47. Charles, I., Xue, L., and Arya, D. P. (2002) Synthesis of Aminoglyco-
side-DNA Conjugates. Bioorg. Med. Chem. Lett. 12, 1259–1262.
48. Charles, I., Xi, H., and Arya, D. P. (2007) Sequence-Specific Target-
ing of RNA with an Oligonucleotide-Neomycin Conjugate. Biocon-
jugate Chem. 18, 160–169.
49. Willis, B., and Arya, D. P. (2009) Triple Recognition of B-DNA.
Bioorg. Med. Chem. Lett. 19, 4974–4979.
50. Bordelon, J. A., Feierabend, K. J., Siddiqui, S. A., Wright, L. L., and
Petty, J. T. (2002) Viscometry and Atomic Force Microscopy Studies
of the Interactions of a Dimeric Cyanine Dye with DNA. J. Phys.
Chem. B 106, 4838–4843.
51. Spink, N., Brown, D. G., Skelly, J. V., and Neidle, S. (1994) Sequence-
Dependent Effects in Drug-DNA Interaction: The Crystal Structure
of Hoechst 33258 Bound to the d(CGCAAATTTGCG)2 Duplex.
Nucleic Acids Res. 22, 1607–1612.
52. Kaul, M., Barbieri, C. M., Kerrigan, J. E., and Pilch, D. S. (2003)
Coupling of Drug Protonation to the Specific Binding of Aminoglyco-
sides to the A Site of 16 S rRNA: Elucidation of the Number of Drug
Amino Groups Involved and their Identities. J. Mol. Biol. 326, 1373–
1387.
53. Botto, R. E., and Coxon, B. (1983) Nitrogen-15 Nuclear Magnetic
Resonance Spectroscopy of Neomycin B and Related Aminoglyco-
sides. J. Am. Chem. Soc. 105, 1021–1028.
54. Geacintov, N. E., Shahbaz, M., Ibanez, V., Moussaoui, K., and
Harvey, R. G. (1988) Base-Sequence Dependence of Noncovalent
Complex Formation and Reactivity of Benzo[a]Pyrene Diol Epoxide
with Polynucleotides. Biochemistry 27, 8380–8387.
55. Pjura, P. E., Grzeskowiak, K., and Dickerson, R. E. (1987) Binding of
Hoechst 33258 to the Minor Groove of B-DNA. J. Mol. Biol. 197,
257–271.
29. Satz, A. L., and Bruice, T. C. (1999) Synthesis of a Fluorescent
Microgonotropen (FMGT-1) and its Interactions with the Dodeca-
mer d(CCGGAATTCCGG). Bioorg. Med. Chem. Lett. 9, 3261–3266.
30. Sengupta, D., Blasko, A., and Bruice, T. C. (1996) A Microgono-
tropen Pentaaza Pentabutylamine and its Interactions with DNA.
Bioorg. Med. Chem. 4, 803–813.
31. Xi, H., Gray, D., Kumar, S., and Arya, D. P. (2009) Molecular
Recognition of Single-Stranded RNA: Neomycin Binding to Poly(A).
FEBS Lett. 583, 2269–2275.
32. Willis, B., and Arya, D. P. (2006) Major Groove Recognition of DNA
by Carbohydrates. Curr. Org. Chem. 10, 663–673.
33. Willis, B., and Arya, D. P. (2006) An Expanding View of Aminoglyco-
side-Nucleic Acid Recognition. Adv. Carbohydr. Chem. Biochem. 60,
251–302.
56. Loontiens, F. G., Regenfuss, P., Zechel, A., Dumortier, L., and Clegg,
R. M. (1990) Binding Characteristics of Hoechst 33258 with Calf
Thymus DNA, Poly[d(A-T)], and d(CCGGAATTCCGG): Multiple
Stoichiometries and Determination of Tight Binding with a Wide
Spectrum of Site Affinities. Biochemistry 29, 9029–9039.
57. Breusegem, S. Y., Clegg, R. M., and Loontiens, F. G. (2002) Base-
Sequence Specificity of Hoechst 33258 and DAPI Binding to Five
(A/T)4 DNA Sites with Kinetic Evidence for More than One High-
Affinity Hoechst 33258-AATT Complex. J. Mol. Biol. 315, 1049–
1061.
58. Breusegem, S. Y., Loontiens, F. G., Regenfuss, P., and Clegg, R. M.
(2001) Kinetics of Binding of Hoechst Dyes to DNA Studied
by Stopped-Flow Fluorescence Techniques. Methods Enzymol. 340,
212–233.
34. Arya, D. P. (2005) Aminoglycoside-Nucleic Acid Interactions: The
Case for Neomycin. Top. Curr. Chem. 253, 149–178.
35. Charles, I., and Arya, D. P. (2005) Synthesis of Neomycin-DNA/
peptide Nucleic Acid Conjugates. J. Carbohydr. Chem. 24, 145–160.
36. Napoli, S., Carbone, G. M., Catapano, C. V., Shaw, N., and Arya,
D. P. (2005) Neomycin Improves Cationic Lipid-Mediated Transfec-
tion of DNA in Human Cells. Bioorg. Med. Chem. Lett. 15, 3467–
3469.
37. Arya, D. P., Xue, L., and Willis, B. (2003) Aminoglycoside
(Neomycin) Preference is for A-Form Nucleic Acids, Not just
RNA: Results from a Competition Dialysis Study. J. Am. Chem.
Soc. 125, 10148–10149.
38. Arya, D. P., Coffee, R. L., Jr., Willis, B., and Abramovitch, A. I.
(2001) Aminoglycoside-Nucleic Acid Interactions: Remarkable Sta-
bilization of DNA and RNA Triple Helices by Neomycin. J. Am.
Chem. Soc. 123, 5385–5395.
39. Arya, D. P., Coffee, R. L., Jr., and Charles, I. (2001) Neomycin-Induced
Hybrid Triplex Formation. J. Am. Chem. Soc. 123, 11093–11094.
40. Arya, D. P., and Coffee, R. L., Jr. (2000) DNA Triple Helix
Stabilization by Aminoglycoside Antibiotics. Bioorg. Med. Chem.
Lett. 10, 1897–1899.
41. Arya, D. P., Coffee, R. L., Jr., and Xue, L. (2004) From Triplex to
B-Form Duplex Stabilization: Reversal of Target Selectivity by
Aminoglycoside Dimers. Bioorg. Med. Chem. Lett. 14, 4643–4646.
42. Arya, D. P. (2007) Aminoglycoside Antibiotics: From Chemical
Biology to Drug Discovery, Wiley-Interscience, Hoboken, NJ.
43. Xue, L., Charles, I., and Arya, D. P. (2002) Neomycin-Pyrene
Conjugate: Dual Modes of Triplex Recognition. Chem. Commun.,
70–71.
59. Loontiens, F. G., McLaughlin, L. W., Diekmann, S., and Clegg, R.
M. (1991) Binding of Hoechst 33258 and 40,60-Diamidino-2-phenyl-
indole to Self-Complementary Decadeoxynucleotides with Modified
Exocyclic Base Substituents. Biochemistry 30, 182–189.
60. Macgregor, R. B., Jr., Clegg, R. M., and Jovin, T. M. (1985) Pressure-
Jump Study of the Kinetics of Ethidium Bromide Binding to DNA.
Biochemistry 24, 5503–5510.
61. Bostock-Smith, C. E., and Searle, M. S. (1999) DNA Minor Groove
Recognition by Bis-Benzimidazole Analogs of Hoechst 33258: Insights
into Structure-DNA Affinity Relationships Assessed by Fluorescence
Titration Measurements. Nucleic Acids Res. 27, 1619–1624.
62. Haq, I. (2002) Thermodynamics of Drug-DNA Interactions. Arch.
Biochem. Biophys. 403, 1–15.
63. Arya, D. P., Coffee, R. L., and Xue, L. (2004) From Triplex to B-
Form Duplex Stabilization: Reversal of Target Selectivity by Ami-
noglycoside Dimers. Bioorg. Med. Chem. Lett. 14, 4643–4646.
64. Rao, K. E., and Lown, J. W. (1991) Molecular Recognition between
Ligands and Nucleic Acids: DNA Binding Characteristics of Analogs
of Hoechst 33258 Designed to Exhibit Altered Base and Sequence
Recognition. Chem. Res. Toxicol. 4, 661–669.
65. Canzonetta, C., Caneva, R., Savino, M., Scipioni, A., Catalanotti, B.,
and Galeone, A. (2002) Circular Dichroism and Thermal Melting
Differentiation of Hoechst 33258 Binding to the Curved (A4T4)
and Straight (T4A4) DNA Sequences. Biochim. Biophys. Acta 1576,
136–142.
66. Chen, F. M. (1983) Binding of Pyrene to DNA, Base Sequence
Specificity and its Implication. Nucleic Acids Res. 11, 7231–7250.

Article
Biochemistry, Vol. 49, No. 3, 2010 469
67. Hartley, J. A., Webber, J., Wyatt, M. D., Bordenick, N., and Lee, M.
(1995) Novel Cytotoxic DNA Sequence and Minor Groove Targeted
Photosensitizers: Conjugates of Pyrene and Netropsin Analogues.
Bioorg. Med. Chem. 3, 623–629.
68. Cohen, G., and Eisenberg, H. (1969) Viscosity and Sedimenta-
tion Study of Sonicated DNA-Proflavine Complexes. Biopolymers
8, 45–55.
69. Arya, D. P., Micovic, L., Charles, I., Coffee, R. L., Jr., Willis, B., and
Xue, L. (2003) Neomycin Binding to Watson-Hoogsteen (W-H) DNA
Triplex Groove: A Model. J. Am. Chem. Soc. 125, 3733–3744.
70. Jin, E., Katritch, V., Olson, W. K., Kharatisvili, M., Abagyan, R., and
Pilch, D. S. (2000) Aminoglycoside Binding in the Major Groove of
Duplex RNA: The Thermodynamic and Electrostatic Forces that
Govern Recognition. J. Mol. Biol. 298, 95–110.