Chem Biol Drug Des 2006; 67: 297304
Journal compilation 2006 Blackwell Munksgaard
Research Article
DNA-binding by Functionalized Gold
Nanoparticles: Mechanism and Structural
Requirements
Catherine M. Goodman, Nandini S. Chari,
Gang Han, Rui Hong, Partha Ghosh and
Vincent M. Rotello*
Department of Chemistry, University of Massachusetts, Amherst
01003, MA, USA
*Corresponding author: Vincent M. Rotello, rotello@chem.umass.edu
Present addresses: Department of Biochemistry and Biophysics,
University of Pennsylvania, Philadelphia 19104, PA, USA
Department of Chemistry and Biochemistry, University of Texas,
Austin 78712, TX, USA
A family of nanoparticles featuring surfaces of
varying hydrophobicity was synthesized. The
efficiency of DNA-binding was determined, dem-
onstrating in a fivefold modulation in binding a
37-mer DNA strand. Nanoparticle-binding causes a
reversible conformational change in the DNA
structure, as demonstrated by circular dichroism
and fluorescence experiments. Furthermore, the
affinity of the nanoparticle for the DNA can be
regulated by external agents, though stability of
the complex is observed at relatively high ionic
strengths.
Key words: DNA, intercalator, hydrophobicity,
LC50, nanoparticle
Received 25 February 2006, revised and accepted
for publication 14 March 2006
Gene regulation as a means of controlling disease states or altering
cellular activity has become a realistic goal within medicinal chem-
istry. Exploration of synthetic molecules capable of DNA-binding
has resulted in the development of several classes of designed
scaffolds. Notable success in the creation of DNA transcription reg-
ulators can be traced to the use of substituted polyamides that are
highly sequence-selective (15). Alternate approaches in the search
for molecules appropriate for interaction with DNA sequences have
resulted in the identification of systems capable of covalent modifi-
cation of the target DNA strand (6,7), as well as peptide and sac-
charide scaffolds that utilize known DNA-binding components to
confer activity to unique sequences (813). Significant progress has
also been made toward understanding DNA reactivity via the
detailed analysis of less functionally diverse systems (1416), inclu-
ding polyamines (17) and cationic dendrimers (1820). Consideration
2006 The Authors
doi: 10.1111/j.1747-0285.2006.00372.x
of the structure of the DNA strands, as well as its regulation
in vivo; however, suggests one additional structure for DNA-binding
has not been sufficiently utilized. Histones are a necessary pack-
aging element for DNA within the nucleus. With their large diam-
eter (approximately 6.5 nm) and correspondingly low surface
curvature, they are well suited for interacting with and condensing
extensive stretches of DNA (21). However, with the exception of
dendrimers, which present synthetic challenges (22), none of the
major systems under investigation employs a similar shape. Our
research focuses on the creation of nanoparticles as scaffolds for
DNA-binding.
Mixed monolayer-functionalized gold nanoparticles present a prom-
ising structure for the development of DNA-regulating molecules
(Figure 1A) (23), and have already been shown to be highly effect-
ive transfection vectors (24). In a previous report, we demonstrated
that quaternary ammonium-functionalized nanoparticle binds 37-mer
DNA in a non-aggregated, stoichiometric fashion with high affinity
(25). Centrifugation of the DNA:nanoparticle complexes provided a
molar ratio of 34 nanoparticles per 37-mer DNA strand, while
dynamic light scattering (DLS) of the bound species confirmed that
the complexes were not aggregated, but had a specific and small
radius in solution of approximately 10 nm. The affinity of the nano-
particles for the DNA segment was demonstrated by their ability to
interrupt transcription by T7 RNA polymerase in vitro. As this
polymerase has an affinity of approximately 5 nM, one possible con-
clusion is that the affinity of the nanoparticles for the DNA exceeds
that of the polymerase, and thus blocks access to the DNA duplex
via tight binding or steric hindrance, preventing transcription (26). A
second explanation for our results is that, upon binding to the
nanoparticle, the DNA helix becomes unrecognizable to the polym-
erase as an appropriate substrate. Our initial focus, therefore, was
to elucidate the mechanism of the nanoparticleDNA interaction.
The question of availability of nanoparticle-bound DNA as a sub-
strate for T7 RNA polymerase arises from the fact that this protein
recognizes DNA in its helical form (27,28). As the anionic phosphate
backbone of DNA is charge neutralized, the conformational rigidity
of the helix is relaxed, as is observed for peptide nucleic acid
(29,30). Decreases in anion density and a corresponding unwinding
of the strands can be observed as salt concentrations are increased
(31), pH is decreased (32,33), or during binding by cationic species,
such as proteins (34,35) or polyamines (3638). With a precedent of
conformational change for tightly bound DNA sequences set, we
surmised that one explanation for the lack of DNA transcription in
our in vitro assay would be that the 37-mer DNA would not display
the regular features of a duplex strand when it binds to the
297
Goodman et al.
cationic nanoparticles,. Alteration of the topology of the DNA would
then have a significant impact on the ability of the polymerase to
successfully bind the duplex. To provide insight into this aspect of
the nanoparticle:DNA interaction, we have examined the DNA con-
formation during the course of binding. By developing a more elab-
orate model of the mode of binding, we will be able to further
optimize the functionalized particles for interaction with DNA.
A second focus in our continued investigation is to observe the
impact of changes to the nanoparticle surface on DNA-binding. The
functional groups displayed on the monolayer surface can have a
significant impact on the behavior and properties of the nanoparti-
cle. The addition of charged or polar terminal substituents will ren-
der an otherwise hydrophobic nanoparticle soluble in water and
other polar solvents (39,40). Reactions can be performed on the sur-
face of the particle, demonstrating the ready access of these intro-
duced components to the surrounding environment and thus their
relevance in focusing on variation of these moieties for optimization
of binding to a given target (41). Additionally, single substitute can
direct binding of the nanoparticle host to a targeted guest mole-
cule: carboxylic acid-functionalized nanoparticles do not interact
with DNA due to the net negative charge of both molecules (25),
yet when a single DNA intercalator is introduced to the monolayer,
binding is observed (42). This behavior indicates that the selection
of thiol components for a DNA-binding nanoparticle can have a
298
Figure 1: (A) To scale depic-
tions of the histone octamer and
the cationic nanoparticle, both
with DNA bound on the surface.
The similar curvature and particle radius
suggests similar modes of binding may
be important. Histone tails have been
removed for clarity. (B) Structure of the
functionalized monolayer components.
The coverage of these quaternary
ammonium salt chains in the monolayer
was determined to be 60 10% by
nuclear magnetic resonance (NMR).
significant impact on the resultant activity, as even a single func-
tional group can determine particle behavior. Our study encompas-
ses a range of quaternary amines, from the smallest (trimethyl) to
the very bulky methyldihexyl head group (Figure 1B). Comparison of
groups with increasing hydrophobic bulk has allowed us to clarify
the relative contributions of non-covalent forces to DNA-binding
affinity and provides insight into the optimal balance of electro-
static charge and hydrophobic volume.
The importance of the hydrophobic components becomes apparent
upon consideration of the DNA duplex. Though dominated by the
phosphate groups within the backbone, this structure also displays
several non-polar regions. One strategy in the design of DNA-bind-
ing molecules is to occupy these regions with complementary
hydrophobic bulk, excluding water from this interface. The resultant
favorable binding to these regions, including the 5-methyl group of
the thymine base as well as several atoms from the sugar ring
found in the groove walls (43), has a significant impact on overall
affinity (44). The size of the grooves can also play an important role
in discriminating between potential drugs, as the steric bulk that
can be accommodated within this flexible cavity is limited (45,46).
We anticipated that analysis of our series of increasingly bulky qua-
ternary amines would result in an increased affinity for the DNA
target with increasing head group size up to the point of the struc-
tural limits of groove binding. The variation of the amine head
Chem Biol Drug Des 2006; 67: 297304

groups allows determination of not only the optimum volume of
hydrophobes for interaction with the DNA grooves, but also the ori-
entation and placement of these groups.
The studies reported herein examine the nanoparticleDNA interac-
tion to elucidate these issues. Our investigations, utilizing a series
of functionalized monolayers, clearly indicate that the DNA structure
is responsive to its host, providing insight into the mechanism of
binding and of the observed inhibition of T7 RNA polymerase. This
characterization of the nanoparticle properties will allow optimiza-
tion for further applications of these versatile molecules in DNA
regulation.
Methods and Materials
Materials
Reagents were purchased from Sigma and used as received. Solvents purchased from
Fisher Scientific (UMASS, Amherst, MA, USA) and used without further purification,
except for dichloromethane and tetrahydrofuran (THF), which were distilled over cal-
cium hydride and sodium, respectively.
Nanoparticle fabrication
Fabrication of the C8-functionalized nanoparticle was completed via a previously repor-
ted procedure (47). In a typical preparation of the mixed monolayer-protected gold clus-
ters (MMPCs), 50 mg of each functionalized thiol was added to 50 mg of
C8-functionalized nanoparticles in 10 mL THF and degassed for approximately 1 h (see
Supplementary material section for details of thiol syntheses). After stirring under
argon for 2 days at room temperature, the black precipitate of the trialkyl-functional-
ized nanoparticle was purified by removal of solvent in vacuo, and repeated washing
with dichloromethane, followed by dialysis in ddH2O. The monodispersity of the parti-
cles as well as confirmation of the particle size was determined by transmission elec-
tron microscopy (TEM) prior to incorporation of the substituted thiols. Nuclear magnetic
resonance (NMR) end-group analysis was used to determine the percentage of ammo-
nium side chain functionality in each thiol monolayer as illustrated in Figure 1B.
DNA synthesis
37-residue oligonucleotides were synthesized (1.0 lmol scale) on a Perseptive Biosys-
tems Expedite 8909 Nucleic Acids Synthesis System (Perspective Biosystems, Framing-
ham, MA, USA) using standard procedures. Cleavage from the support and base
deprotection was accomplished by treatment with concentrated NH4OH (16 h, 37
C).
Strands were purified via elution with 20% acetonitrile from an Amberchrom CG-161C
Synthetic Adsorbant/Chromatographic Resin column (Toso-Haas). The oligomers were
then concentrated to dryness and redissolved in TE buffer (pH 8.0). Complementary
strands were annealed by combining equivalent molar amounts of the individual
sequences (100 lM final concentration), heating to 90
C, and slowly cooling to approxi-
mately 30
C. DNA was stored at )20
C.
Circular dichroism
Samples were prepared in TE buffer [10 mM Tris, 1 mM ethylenediaminetetraacetic acid
(EDTA), pH 7.4]. The DNA 37-mer was diluted from a stock solution of 500.25 lM
(2 mL final volume). Nanoparticles were added in aliquots to the DNA solution from
0.1, 1, 10, or 100 lM stocks in ddH2O. Sodium dodecyl sulfate (SDS) was prepared
from a dry powder to 100 mM final concentration. Three stocks of 10-fold dilutions
were also prepared (10, 1 and 0.1 mM). Circular dichroism (CD) data were obtained
using either a Jasco 715 or 720 spectropolarimeter (Jasco, Easton, MD, USA), using a
quartz cuvette with a path length of 1 cm. Scans were taken from 240300 nm at a
rate of 10 nm/min, with a 0.1 nm step resolution and a 16 seconds response. For each
spectrum, 13 runs were averaged at a constant temperature of 20
C, with a 5 min
equilibration before each scan. Data obtained with high-tension voltage (HT) over 600
were not included in the analysis.
Chem Biol Drug Des 2006; 67: 297304
Mechanism of DNANanoparticle Binding
Fluorescent intercalator displacement assay
This assay was modified from a reported procedure (48). Samples were prepared in TE
buffer except as noted. The DNA 37-mer was diluted from a stock solution of 50 lM to
0.1 or 0.25 lM (0.5 mL final volume) or from 100 to 0.11 lM (2.0 mL final volume). In
experiments performed in the presence of salt, the sodium chloride was mixed into the
TE buffer prior to addition of DNA. Ethidium bromide (EB) was diluted from a stock solu-
tion of 1 mM to a final concentration of 20 times the DNA concentration in each case.
Nanoparticles were added in aliquots to the DNA solution from 0.1, 1, 10, or 100 lM
stocks in ddH2O. The SDS solution was prepared from a dry powder to 100 mM final
concentration, and an additional three stocks of 10-fold dilutions (10, 1 and 0.1 mM)
were prepared from the initial stock. SDS was introduced to the solution using which-
ever stock concentration resulted in the addition of the least volume, above a minimum
0.5 lL volume. Spectra were obtained on a Shimadzu RF-5301 PC spectrofluorophotome-
ter (Shimadzu, Columbia, MD, USA), using a 2 10 mm quartz cuvette (kem 595 nm,
kex 545 nm). The excitation slit width was maintained at 1.5 or 3 nm, and the emis-
sion slit width open to 10 or 20 nm, depending on initial conditions.
Results and Discussions
Information regarding the DNAnanoparticle interaction was first
obtained using the fluorescent intercalator displacement assay (FID)
as described by Boger et al. (48). In this assay, added guests com-
pete with the hydrophobic intercalator, EB, for DNA-binding sites
either directly (via intercalative insertion) or indirectly, by promoting
a DNA conformation incompatible with EB-binding. Removal of the
intercalator in a concentration-dependent manner provided a binding
curve for the Me3 nanoparticle (Figure 2A). The EB signal was fully
quenched at a Me3 concentration of approximately 0.35 lM, or at
a approximately 3.5:1 nanoparticle:DNA ratio, in good agreement
with previous results (25). Control experiments with carboxylate-
functionalized nanoparticles demonstrated that the decreasing fluor-
escence is not due to light absorption by or interference due to the
particles. Similarly, addition of monolayer periphery substituents at
concentrations similar to the density of these groups on the particle
surface resulted in no change in the fluorescence spectra, insuring
that the multivalency of the particle is responsible for complex for-
mation.
To provide evidence for conformational change of the DNA guest,
we examined the nanoparticle:DNA interaction using CD (Figure 2B).
The CD spectrum of DNA alone shows a maximum at approximately
280 nm, indicative of the B-form duplex. Upon addition of the Me3
nanoparticle at the same concentrations used in the FID assay,
there is an immediate and significant change in the DNA conforma-
tion. The decrease in intensity at 280 nm coupled with a shift in
the maximum wavelength is indicative of conversion to a denatured
double strand.
We further evaluated the nanoparticle:DNA interaction by examin-
ing the DNA-binding capacity of the Me3 in the presence of differ-
ing environmental conditions. The initial design of the nanoparticles
focused on the electrostatic interaction of the cationic monolayer
components with the anionic DNA backbone. To determine whether
this charge pairing was the only driving force for binding, we
probed the DNA conformation during binding in buffers containing
increasing concentrations of sodium chloride (Figure 3). The signifi-
cance of any electrostatic interactions will be diminished by the
high levels of electrolytes present in solution, providing insight into
the role of charge pairing in the nanoparticle:DNA association.
299
Goodman et al.
Figure 2: Effect of cationic nanoparticles on 37-mer double
stranded (ds)DNA. (A) Fluorescence of ethidium bromide (EB) intercalated
within the helix is quenched as the Me3 particles bind the DNA duplex.
DNA concentration is 0.11 lM. Curve is meant to lead the eye. (B) Circular
dichroism of DNA with increasing aliquots of Me3 demonstrates a substan-
tial rearrangement of the DNA duplex. Each addition increases the nanopar-
ticle concentration by 0.1 lM, for a final concentration of 1.2 lM. Arrows
indicate that ellipticity decreases at approximately 280 nm and increases at
approximately 245 nm as the nanoparticles are added to the solution. DNA
concentration is 0.25 lM.
Surprisingly, there was no change in binding even the salt concen-
tration reached 0.2 M. The inability of the sodium chloride to
decrease binding indicates that the tight binding of the nanoparti-
cles to the DNA is not mediated solely by electrostatic pairing. This
high affinity interaction may be stabilized by the van der Waals
interactions between either the exposed methyl groups or the inter-
ior alkyl chains of the nanoparticle with the thymine methyl substit-
uent or other hydrophobic groups of the DNA duplex, as previously
mentioned. The electrostatic interaction between the DNA and na-
noparticle was disrupted by adding 0.5 M NaCl. Interestingly, CD
shows the release of the DNA in its native form from the particle
surface at the salt concentration of 1 M (Figure 3B).
300
Figure 3: Introduction of sodium chloride attenuates nanopar-
ticle:DNA interaction. DNA concentration is 0.25 lM. (A) Fluorescent
intercalator displacement (FID) assay performed in the presence of increasing
concentrations of sodium chloride. (B) The helical circular dichroism (CD)
spectrum of DNA alone (solid line) is disrupted upon addition of saturating
levels of Me3 (dashed line), but this interaction is completely screened out
by 1 M NaCl at the same nanoparticle concentration (dotted line).
The addition of salt was performed to investigate the driving force
due to charge pairing between the nanoparticles and DNA. These
individual ions were not sufficient at low concentrations to disrupt
the DNAnanoparticle complex. In contrast, addition of long chain
anions was anticipated to affect the equilibrium of the system, as
the alkyl tails of these surfactants can be inserted into the non-
polar monolayer of the particle, simultaneously burying hydrophobic
groups and attenuating the positively charged nanoparticle surface.
This DNA-release mechanism is unique to the nanoparticle scaffold
due to the surface curvature of the gold core: the subsequent out-
ward radiation of surface-bound thiols leaves exposed hydrophobic
surfaces in the monolayer interior, a feature not observed with two-
dimensional interfaces. While it is possible that the surfactant
could also compete for the DNA-binding sites on the surface of the
nanoparticle host due to its like charge, we did not anticipate that
the anionic functionality on its own would have any effect, as the
chloride ions in the previous experiments were not observed to
cause any change at low concentrations. Sodium dodecyl sulfate
Chem Biol Drug Des 2006; 67: 297304

Figure 4: Addition of sodium dodecyl sulfate (SDS) to the
unwound, fluorescence-quenched samples results in regener-
ation of the DNA signal due to duplex formation. Points plotted
are the value at 280 nm within the circular dichroism (CD) spectrum in refer-
ence to the free and fully bound duplexes. Inset shows full CD scans. DNA
concentration is 0.25 lM. Me3 concentration is 1.2 lM.
was indeed able to release the DNA from the particles, resulting in
complete regeneration of the double helical structure (Figure 4). The
concentration of SDS needed (approximately 0.1 mM) to effect this
change is well below any reported critical micelle concentration
(CMC) values (110 mM; 49,50), clearly demonstrating that the
observed release is not related to interactions with micellar or
other higher-order structures. The incorporation of substituted alkyl
chains into a nanoparticle monolayer has previously been examined,
resulting in the release of proteins from carboxylate-functionalized
nanoparticles using long chain cations (51). The in situ modification
of the nanoparticle surface to alter the behavior of the cationic host
demonstrates the dynamic nature of the gold particles, a feature,
which can be, optimized for future in vivo applications. This
dynamic behavior is again reminiscent of the mechanism of DNA
regulation associated with histone-binding; although in the case of
the histone octamer, the cationic lysines are covalently modified to
alter the charge distribution on the protein surface, the requirement
for an external agent (histone acetylase versus SDS) as well as the
outcome of DNA release is the same.
Understanding of the importance of external cues and the behavior
of the nanoparticleDNA system in response to those stimuli pro-
vides insight into mechanisms by which the nanoparticles may inter-
act with the DNA in vivo. Further design of the nanoparticle itself,
however, can also be used to elucidate the significance of the inter-
molecular forces at play in this complex system. Our evaluation of
the nanoparticle functionality focused on a series of particles with
increasing hydrophobic bulk in the cationic head group. Examination
of the change in hydrophobicity versus the strength of the binding
affinity should provide details as to the specific balance of charge
density versus van der Waals packing and hydrophobicity for opti-
Chem Biol Drug Des 2006; 67: 297304
Mechanism of DNANanoparticle Binding
Figure 5: DNA-binding by six functionalized nanoparticles, as
determined by using fluorescent intercalator displacement
(FID) assay. DNA concentration is 0.1 lM. Head group volume calculated
as the volume of the nitrogen atom and unsubstituted alkyl chains of each
incorporated thiol using MACROMODEL 7.0 with the AMBER forcefield.
mal interaction with DNA. Indeed, analysis of the FID assay for
each of our six particles indicates several transitions as the head
group volume increases in the substituted nanoparticles (Figure 5).
Simple variation of the alkyl chain length causes up to a fivefold
difference in the LC50 values. The mid-sized head groups (MeBu2,
Bu3, and Me2Hex) display the smaller values, and the values
increase with the bulky hydrophobe (MeHex2) or small hydropho-
bes (Me3, Me2Bu). The high LC50 value observed for the largest
head group (MeHex2) suggests that this amine substituent is too
bulky to be accommodated in the DNA groove, and that the binding
observed is based solely on electrostatic pairing of the cationic par-
ticles and anionic DNA. Interestingly, the correlation of LC50 and
head group volume for the smallest and largest nanoparticles falls
on a straight line. The binding of the DNA with the intermediately
sized particles is tighter than expected based on this correlation,
emphasizing the importance of optimal volume of head groups to fit
into the DNA groove.
As the grooves are more fully packed with the growing alkyl chains,
in MeBu2, Bu3, and Me2Hex, the LC50 values drop approximately
twofold, representative of an enhanced DNA-binding compound.
Notably, the LC50 values for these three nanoparticles are identical
within error even though the structure of the head groups is quite
distinct. However, consideration of the likely conformations of these
flexible chains indicates a greater similarity than their structure
would suggest: Me2Hex, for example, can approximate MeBu2 by
simply incorporating a hairpin-type conformation within the hexyl
chain, which would result in a similar shape to that of the two bu-
tyl chains of MeBu2. The low binding affinity (high LC50) observed
for MeHex2 further confirms that the full-length hexyl chain
is unsuitable for interaction with the DNA grooves. A similar
301
Goodman et al.
Figure 6: Circular dichroism (CD) data of the three selected nanoparticles binding to the DNA. Arrows indicate direction of change during
nanoparticle titration. DNA concentration is 0.25 lM. (A) Addition of MeBu2 particles. Two of the last four overlapping scans have been removed for clarity.
(B) Titration with MeHex2 nanoparticles. (C) Comparison of the CD traces for the three nanoparticles. The long dashed line demonstrates the effects of two
equivalents of nanoparticles per DNA strand, and the short dashed line is after the introduction of four equivalents.
consideration of the conformation of MeHex2 suggests the reason
for this disparity: the folding of the two hexyl chains would approxi-
mate a head group displaying four butyl groups rather than the
three that are displayed on Bu3. We can therefore define the limit
of aliphatic components that can efficiently bind with DNA as three
intermediate length chains. To further quantify the space available
in the DNA groove for binding, head groups should be composed of
chains that are more constrained. Using this methodology, the use
of varying topologies in in situ optimization of binding, as observed
in the nanoparticles under investigation, would be severely limited.
Further investigation of these three groupings was carried out to
determine if the differences in LC50 values reported were related
to any changes in the mechanism of the nanoparticle:DNA interac-
tion. CD titrations were performed with MeBu2 (low LC50) and
MeHex2 (highest LC50) to compare their effect on conformational
change with that observed for Me3 (Figure 6). In each case, con-
tinuous change is seen as additional equivalents of particles are
introduced to the solution. For MeBu2, no isosbestic point can be
identified during addition of the first few aliquots of nanoparti-
cles, but the latter half of the titration elucidates an isosbestic
point at approximately 262 nm (Figure 6A). For the MeHex2 na-
noparticles, a clear isosbestic point is seen in the transition from
duplex to unwound DNA at approximately 255 nm (Figure 6B).
Interestingly, both of these isosbestic points were observed in the
initial titration with Me3 particles. This result indicates similarities
between both Me3 and MeBu2 as well as Me3 and MeHex2.
We hypothesize that the transition with an isosbestic point of
approximately 255 nm is related to a simple charge pairing of the
cationic head groups of Me3 and MeHex2 with the DNA back-
bone, whereas the transition centered at approximately 262 nm
has a more direct correlation to interaction with the DNA grooves
with the hydrophobic portion of either Me3 or MeBu2. A direct
comparison of spectra with two and four equivalents of nanoparti-
cles added to the DNA provides further insight into the three sys-
tems (Figure 6C). Introduction of two equivalents MeBu2 particle
causes the most immediate change. This lower molecular ratio
suggests that less charge neutralization is required for DNA
302
unwinding, emphasizing the optimal interaction of the nanoparticle
chains with the DNA grooves. Me3 shows an intermediate
change after addition of two particles, with fully unwound DNA
at the saturating concentration of four equivalents. This is in good
agreement with the intermediate LC50 value illustrated in Figure 5.
Addition of MeHex2 causes a smaller change in conformation
after two particles, and even after addition of four equivalents of
nanoparticles, we do not see a full conformational switch to the
unwound strand. This lack of structural change supports our previ-
ous hypothesis that MeHex2 particles bind the DNA strand only
via electrostatic pairing of the cationic quaternary amines and
phosphate backbone.
One of the central questions driving these experiments was to
clarify potential reasons that T7 RNA polymerase is incapable of
using a nanoparticle-bound DNA duplex as a template for RNA
synthesis. Two mechanisms, which could account for this inhibition
are that the polymerase does not recognize the DNA strand as its
target or that the binding of the nanoparticle is too tight to allow
access to the double helix. The CD titration of DNA with the
Me3 nanoparticles clearly demonstrates that the helical conforma-
tion is not maintained upon binding. As mentioned, T7 RNA po-
lymerase requires the helical structure for recognition of its
promoter sequence. This conformational change, then, presents a
significant barrier in further binding by the protein, and could
explain the lack of RNA transcript synthesis. While the lack of
structural cues to initiate polymerase-binding can explain our pre-
vious results, the alternate mechanism of tight binding cannot be
discounted. The transcription assay is performed by adding the po-
lymerase to the already bound nanoparticle:DNA system (25). In
the studies reported herein, we observe that even in the presence
of 0.2 M NaCl, the nanoparticles bind efficiently to the DNA 37-
mer, and we anticipate that a higher concentration of salt would
be needed to strip the particles from the duplex if added after
complexation had occurred. A similar phenomenon is likely occur-
ring in the case of the polymerase: the multivalent binding of
Me3 prevents the polymerase (in much lower than 0.2 M concen-
trations) from accessing the DNA.
Chem Biol Drug Des 2006; 67: 297304

Previous studies of the cationic nanoparticles indicated their utility
as transfection vectors in mammalian cells (24). Examination of the
Me3:DNA system in the presence of external agents allows us to
speculate on potential mechanisms of intracellular trafficking and
release. The salt concentration in vivo (approximately 170 mM; 52)
is not sufficient to release the DNA from its cationic host, as dem-
onstrated by the sodium chloride results. The DNA, then, is carried
into the cell by the electrostatic complementarity of the highly cati-
onic nanoparticles with the lipid bilayer. Once internalized into
endosomal vesicles1, the DNA must be released into the cytosol.
If the nanoparticles are capable of disrupting the vesicle wall at
even a single point, some lipids would be likely to insert their
hydrophobic chains into the nanoparticle monolayer, as observed
with the introduction of surfactant to the bound system. This inter-
action with the lipids serves two purposes: it would release the
DNA from the bound complex as observed with addition of SDS, as
well as further disrupt the endosomal membrane, potentially releas-
ing the DNA into the cytosol for targeting to the nucleus. While
the in vitro experiments described herein cannot definitively demon-
strate the accuracy of this potential mechanism, they serve as use-
ful tools toward gaining an improved understanding of the
mechanisms by which nanoparticles interact with cellular machinery.
Increasing understanding of these unique scaffolds will allow fur-
ther optimization of these functionalized gold particles for specific
applications.
Comparison of variations in the substituents on the nanoparticle
surface shed light on the preferred balance between charge den-
sity and hydrophobic bulk. As previously mentioned (43), the non-
polar groove of the helix can accommodate some steric bulk from
an interacting molecule to promote binding, but the capacity is
limited. The sharp transitions in LC50 values as the size of the
head groups increases clearly defines the optimal packing for
nanoparticle DNA-binding as being between 158
3 and 221
3.
The immediate unwinding of the helix observed by CD also dem-
onstrates that these intermediate head groups are well suited to
their DNA target. Smaller substituents cannot displace the water
from the DNA surface effectively, resulting in decreased binding
(LC50 values increased by approximately twofold). Similarly, once
the gaps have been filled to capacity, the introduction of addi-
tional hydrophobic bulk causes an abrupt and unfavorable change
in the LC50 values: comparison of binding of Me2Hex (221
3)
and MeHex2 (242
3) shows a fivefold increase. This result is
mirrored by the CD titration of MeHex2, in which the decreased
affinity is observed as an incomplete conformational change. Fur-
ther development of the cationic nanoparticles for DNA-binding,
then, should focus on the substituents of intermediate size for
optimal non-covalent interactions.
Conclusions
In conclusion, functionalized gold nanoparticles are a viable scaffold
for DNA-binding. The DNA readily undergoes a conformational
change upon binding, perhaps attributable to the very similar
The suggested mechanism developed in the previous report (24), based on
several controls.
Chem Biol Drug Des 2006; 67: 297304
Mechanism of DNANanoparticle Binding
structure of the nanoparticles and a biologically relevant partner,
the histone octamer. The responsiveness of the nanoparticles to
change in the environment highlights the utility of these scaffolds
in vivo. The variation in DNA-binding with changes in the cationic
head group indicates that the binding affinity may be tuned accord-
ing to the functionality incorporated into the monolayer. Examina-
tion of additional thiols with alternative substituents will lead to
further understanding and optimization of intermolecular forces
involved in these complex systems.
Acknowledgment
This research was funded by grant EB004503. CMG acknowledges NIH Chemistry-Bio-
logy Interface training grant GM 08515.
References
1. Nguyen-Hackley D.H., Ramm E., Taylor C.M., Joung J.K., Dervan P.B., Pabo C.O.
(2004) Allosteric inhibition of zinc-finger binding in the major groove of DNA by
minor-groove binding ligands. Biochemistry;43:38803890.
2. Gopal Y.N.V., Van Dyke M.W. (2003) Combinatorial determination of sequence spe-
cificity for nanomolar DNA-binding hairpin polyamides. Biochemistry;42:68916903.
3. Fechter E.J., Dervan P.B. (2003) Allosteric inhibition of protein-DNA complexes by
polyamide-intercalator conjugates. J Am Chem Soc;125:84768485.
4. Ellervik U., Wang C.C.C., Dervan P.B. (2000) Hydroxybenzamide/pyrrole pair distin-
guishes T center dot A from A center dot T base pairs in the minor groove of
DNA. J Am Chem Soc;122:93549360.
5. Kikuta E., Murata M., Katsube N., Koike T., Kimura E. (1999) Novel recognition of
thymine base in double-stranded DNA by zinc(II)-macrocyclic tetraamine complexes
appended with aromatic groups. J Am Chem Soc;121:54265436.
6. Oyoshi T., Kawakami W., Narita A., Bando T., Sugiyama H. (2003) Inhibition of tran-
scription at a coding sequence by alkylating polyamide. J Am Chem Soc;125:4752
4754.
7. Wada T., Minaminoto N., Inaki Y., Inoue Y. (2000) Peptide ribonucleic acids (PRNA):
2. A novel strategy for active control of DNA recognition through berate ester for-
mation. J Am Chem Soc;122:69006910.
8. Jantz D., Amann B.T., Gatto G.J., Berg J.M. (2004) The design of functional DNA-
binding proteins based on zinc finger domains. Chem Rev;104:789799.
9. Vazquez M.E., Caamano A.M., Mascarenas J.L. (2003) From transcription factors to
designed sequence-specific DNA-binding peptides. Chem Soc Rev;32:338349.
10. Kovacic R.T., Welch J.T., Franklin S.J. (2003) Sequence-selective DNA cleavage by
a chimeric metallopeptide. J Am Chem Soc;125:66566662.
11. Lajmi A.R., Lovrencic M.E., Wallace T.R., Thomlinson R.R., Shin J.A. (2000) Minimal-
ist, alanine-based, helical protein dimers bind to specific DNA sites. J Am Chem
Soc;122:56385639.
12. Xuereb H., Maletic M., Gildersleeve J., Pelczer I., Kahne D. (2000) Design of an
oligosaccharide scaffold that binds in the minor groove of DNA. J Am Chem
Soc;122:18831890.
13. Zondlo N.J., Schepartz A. (1999) Highly specific DNA recognition by a designed
miniature protein. J Am Chem Soc;121:69386939.
14. Joubert A., Sun X.W., Johansson E., Bailly C., Mann J., Neidle S. (2003)
Sequence-selective targeting of long stretches of the DNA minor groove by a novel
dimeric bis-benzimidazole. Biochemistry;42:59845992.
15. Bentin T., Nielsen P.E. (2003) Superior duplex DNA strand invasion by acridine con-
jugated peptide nucleic acids. J Am Chem Soc;125:63786379.
16. Proudfoot E.M., Mackay J.P., Karuso P. (2001) Probing site specificity of DNA bind-
ing metallointercalators by NMR spectroscopy and molecular modeling. Biochemis-
try;40:48674878.
303
Goodman et al.
17. Esposito D., DelVecchio P., Barone G. (1997) Interactions with natural polyamines
and thermal stability of DNA. A DSC study and a theoretical reconsideration. J Am
Chem Soc;119:26062613.
18. Bielinska A.U., Chen C.L., Johnson J., Baker J.R. (1999) DNA complexing with pol-
yamidoamine dendrimers: implications for transfection. Bioconjug Chem;10:843850.
19. Tang M.X., Szoka F.C. (1997) The influence of polymer structure on the interactions
of cationic polymers with DNA and morphology of the resulting complexes. Gene
Ther;4:823832.
20. Bielinska A.U., KukowskaLatallo J.F., Baker J.R. (1997) The interaction of plasmid
DNA with polyamidoamine dendrimers: mechanism of complex formation and ana-
lysis of alterations induced in nuclease sensitivity and transcriptional activity of
the complexed DNA. Biochim Biophys Acta-Gene Struct Expr;1353:180190.
21. Voet D., Voet J.G. (1995) Biochemistry, 2nd edn. New York, USA: John Wiley &
Sons, Inc.;p. 11241131.
22. Grayson S.K., Frechet J.M.J. (2001) Convergent dendrons and dendrimers: from
synthesis to applications. Chem Rev;101:38193867.
23. Goodman C.M., Rotello V.M. (2004) Biomacromolecule surface recognition using
nanoparticles. Mini Rev Org Chem;1:103114.
24. Sandhu K.K., McIntosh C.M., Simard J.M., Smith S.W., Rotello V.M. (2002) Gold na-
noparticle-mediated transfection of mammalian cells. Bioconjug Chem;13:36.
25. McIntosh C.M., Esposito E.A., Boal A.K., Simard J.M., Martin C.T., Rotello V.M.
(2001) Inhibition of DNA transcription using cationic mixed monolayer protected
gold clusters. J Am Chem Soc;123:76267629.
26. Ujvari A., Martin C.T. (1996) Thermodynamic and kinetic measurements of promoter
binding by T7 RNA polymerase. Biochemistry;35:1457414582.
27. Steitz T.A. (2004) The structural basis of the transition from initiation to elongation
phases of transcription, as well as translocation and strand separation, by T7 RNA
polymerase. Curr Opin Struct Biol;14:49.
28. Kukarin A., Rong M.Q., McAllister W.T. (2003) Exposure of T7 RNA polymerase to
the isolated binding region of the promoter allows transcription from a single-
stranded template. J Biol Chem;278:24192424.
29. Rasmussen H., Kastrup J.S., Nielsen J.N., Nielsen J.M., Nielsen P.E. (1997) Crystal
structure of a peptide nucleic acid (PNA) duplex at 1.7
resolution. Nat Struct
Biol;4:98101.
30. Wittung P., Nielsen P.E., Buchardt O., Egholm M., Norden B. (1994) DNA-like double
helix formed by peptide nucleic-acid. Nature;368:561563.
31. Cheng Y.K., Pettitt B.M. (1992) Stabilities of double-strand and triple-strand helical
nucleic-acids. Prog Biophys Mol Biol;58:225257.
32. Tajmirriahi H.A., Ahmad R., Naoui M., Diamantoglou S. (1995) The effect of Hcl on
the solution structure of calf thymus DNA a comparative-study of DNA denatura-
tion by proton and metal-cations using Fourier-transform IR difference spectroscopy.
Biopolymers;35:493501.
33. Oconnor T., Mansy S., Bina M., McMillin D.R., Bruck M.A., Tobias R.S. (1982) The
pH-dependent structure of calf thymus DNA studied by Raman-spectroscopy. Bio-
phys Chem;15:5364.
34. Manning G.S. (2003) Is a small number of charge neutralizations sufficient to bend
nucleosome core DNA onto its superhelical ramp? J Am Chem Soc;125:15087
15092.
35. Mahtab R., Rogers J.P., Singleton C.P., Murphy C.J. (1996) Preferential adsorption
of a 'kinked' DNA to a neutral curved surface: comparisons to and implications for
nonspecific DNA-protein interactions. J Am Chem Soc;118:70287032.
36. Delcros J.G., Sturkenboom M., Basu H.S., Shafer R.H., Szollosi J., Feuerstein B.G.,
Marton L.J. (1993) Differential-effects of spermine and its analogs on the structures
of polynucleotides complexed with ethidium-bromide. Biochem J;291:269274.
304
37. Geall A.J., Al-Hadithi D., Blagbrough I.S. (2002) Efficient calf thymus DNA conden-
sation upon binding with novel bile acid polyamine amides. Bioconjug
Chem;13:481490.
38. Minagawa K., Matsuzawa Y., Yoshikawa K., Matsumoto M., Doi M. (1991) Direct
observation of the biphasic conformational change of DNA induced by cationic pol-
ymers. FEBS Lett;295:6769.
39. Hong R., Fischer N.O., Verma A., Goodman C.M., Emrick T., Rotello V.M. (2004)
Control of protein structure and function through surface recognition by tailored
nanoparticle scaffolds. J Am Chem Soc;126:739743.
40. Simard J., Briggs C., Boal A.K., Rotello V.M. (2000) Formation and pH-controlled
assembly of amphiphilic gold nanoparticles. Chem Commun;19431944.
41. Templeton A.C., Hostetler M.J., Kraft C.T., Murray R.W. (1998) Reactivity of mono-
layer-protected gold cluster molecules: steric effects. J Am Chem Soc;120:1906
1911.
42. Wang G.L., Zhang J., Murray R.W. (2002) DNA binding of an ethidium intercalator
attached to a monolayer-protected gold cluster. Anal Chem;74:43204327.
43. Neidle S. (2002) Nucleic Acid Structure and Recognition. New York: Oxford Univer-
sity Press;p. 89138.
44. Chaires J.B. (1998) Drug-DNA interactions. Curr Opin Struct Biol;8:314320.
45. Biswas K., Pal S., Carbeck J.D., Kahne D. (2000) The molecular basis for pyrimid-
ine-selective DNA binding: analysis of calicheamicin oligosaccharide derivatives by
capillary electrophoresis. J Am Chem Soc;122:84138420.
46. Tien J., Terfort A., Whitesides G.M. (1997) Microfabrication through electrostatic
self-assembly. Langmuir;13:53495355.
47. Brust M., Walker M., Bethell D., Schiffrin D.J., Whyman R. (1994) Synthesis of thi-
ol-derivatized gold nanoparticles in a 2-phase liquid-liquid system. J Chem Soc
Chem Commun;801802.
48. Boger D.L., Fink B.E., Brunette S.R., Tse W.C., Hedrick M.P. (2001) A simple, high-
resolution method for establishing DNA binding affinity and sequence selectivity.
J Am Chem Soc;123:58785891.
49. Chatterjee A., Moulik S.P., Sanyal S.K., Mishra B.K., Puri P.M. (2001) Thermodynam-
ics of micelle formation of ionic surfactants: a critical assessment for sodium dode-
cyl sulfate, cetyl pyridinium chloride and dioctyl sulfosuccinate (Na salt) by
microcalorimetric, conductometric, and tensiometric measurements. J Phys Chem
B;105:1282312831.
50. Rubio D.A.R., Zanette D., Nome F., Bunton C.A. (1994) Effect of 1-butanol on micel-
lization of sodium dodecyl-sulfate and on fluorescence quenching by bromide ion.
Langmuir;10:11511154.
51. Fischer N.O., Verma A., Goodman C.M., Simard J.M., Rotello V.M. (2003) Reversible
'irreversible' inhibition of chymotrypsin using nanoparticle receptors. J Am Chem
Soc;125:1338713391.
52. Griffiths M. (1974) Introduction to Human Physiology. New York, USA: Macmillan
Publishing Co., Inc.;p. 121.
Supplementary Material
The following supplementary material is available for this article
online at http://www.blackwell-synergy.com
Appendix S1. Synthetic procedures and characterization for ammo-
nium-substituted thiols.
Chem Biol Drug Des 2006; 67: 297304