Exploiting localized surface binding effects to enhance the catalytic reactivity of peptide-capped nanoparticles

Article abstract of DOI:10.1021/ja402215t

Peptide-based methods represent new approaches to selectively produce nanostructures with potentially important functionality. Unfortunately, biocombinatorial methods can only select peptides with target affinity and not for the properties of the final material. In this work, we present evidence to demonstrate that materials-directing peptides can be controllably modified to substantially enhance particle functionality without significantly altering nanostructural morphology. To this end, modification of selected residues to vary the site-specific binding strength and biological recognition can be employed to increase the catalytic efficiency of peptide-capped Pd nanoparticles. These results represent a step toward the de novo design of materials-directing peptides that control nanoparticle structure/function relationships.

Full text of DOI:10.1021/ja402215t

Article
pubs.acs.org/JACS
Exploiting Localized Surface Binding Effects to Enhance the Catalytic
Reactivity of Peptide-Capped Nanoparticles
Ryan Coppage,1 Joseph M. Slocik,2 Hadi Ramezani-Dakhel,3 Nicholas M. Bedford,2,1 Hendrik Heinz,3
Rajesh R. Naik,2,* and Marc R. Knecht1,*
1Department of Chemistry, University of Miami, Coral Gables, Florida 33146, United States
2Materials and Manufacturing Directorate, Air Force Research Laboratory, Wright-Patterson Air Force Base, Ohio 45433-7702,
United States
3Department of Polymer Engineering, University of Akron, Akron, Ohio, 44325, United States
S
* Supporting Information
ABSTRACT: Peptide-based methods represent new ap-
proaches to selectively produce nanostructures with potentially
important functionality. Unfortunately, biocombinatorial
methods can only select peptides with target affinity and not
for the properties of the final material. In this work, we present
evidence to demonstrate that materials-directing peptides can
be controllably modified to substantially enhance particle
functionality without significantly altering nanostructural
morphology. To this end, modification of selected residues
to vary the site-specific binding strength and biological
recognition can be employed to increase the catalytic efficiency of peptide-capped Pd nanoparticles. These results represent a
step toward the de novo design of materials-directing peptides that control nanoparticle structure/function relationships.
however, the residue-specific binding effects are anticipated to
be more important in controlling the surface bound peptide
structure and the biotic/abiotic interface of the material, which
likely participates in the reactivity.
INTRODUCTION
■
Biology provides exquisite inspiration for the production of
functional inorganic nanomaterials.1 Most of these materials are
prepared using peptides and proteins to direct structural growth
and engender the system with desired properties.1−3 Mimicking
these approaches for the generation of technologically useful
nanomaterials represents an avenue to achieve important
structures for sustainable activities. Many groups have turned
to biocombinatorial approaches to isolate peptides with affinity
for a range of nonbiological materials;1,4−7 however, the effect
of the biotic/abiotic interface over the structure/function
relationship remains poorly understood. To this end, the ability
to fine-tune desirable optical, catalytic, magnetic, and other
material properties could be achieved at the peptide sequence
level.
We have employed the Pd-specific Pd4 peptide
(TSNAVHPTLRHL) for the fabrication of peptide-capped
Pd nanoparticles with the ability to catalyze C−C coupling
reactions under nontraditional conditions of a water-based
solvent at room temperature.8,9 The peptide is anticipated to
anchor to the metallic surface through the histidine residues at
the six and eleven position,10 where substitutions at these sites
results in changes to the particle size and catalytic activity.11,12
In this regard, we hypothesize that weaker binding at the six
position, and stronger binding at the eleven position could
result in peptides that produce nanoparticles with optimized
catalytic functionality. Note that binding changes at these
specific positions may alter the global peptide binding strength;
In this work, we demonstrate that we can rationally modify
materials binding peptide sequences to enhance the structure/
function relationship of Pd nanoparticles for catalytic
functionality; however, such approaches could be adapted for
nanomaterials with other properties. To this end, the Pd4
peptide was employed as the parent sequence wherein the six
and eleven positions were substituted with cysteine, histidine,
and alanine residues. These amino acids were selected to
modulate the binding strength at these specific positions with
cysteine being the strongest binder, alanine the weakest, and
the native histidine being of intermediate binding strength.13
From this, six different peptides were generated, listed in Table
1, which were used to prepare peptide-capped Pd nanoparticles.
Complete characterization of the peptide binding effects were
studied using both experimental and modeling approaches,
followed by their catalytic analysis using Stille C−C coupling.
From these studies, the particles demonstrated reactivities that
varied based solely upon the binding strength of the residues at
the predetermined positions. Interestingly, no correlation
between the global peptide binding strength and catalytic
functionality was noted, supporting the hypothesis that
Received: March 2, 2013
Published: July 18, 2013
© 2013 American Chemical Society
11048
dx.doi.org/10.1021/ja402215t | J. Am. Chem. Soc. 2013, 135, 11048−11054

Journal of the American Chemical Society
Article
of at least 100 nanoparticles among multiple images was
employed to determine particle size distributions. For Quartz
Crystal Microbalance with dissipation energy (QCM-D)
measurements, Pd surfaces were prepared by sputter-coating
Pd onto Au QCM sensors for 180 s with a Cressington 108-
Auto sputter coater. The metallic surfaces were then cleaned via
UV/ozone exposure for 10.0 min, followed by a deionized
water rinse, and an additional 10.0 min UV/ozone exposure.
Each peptide was dissolved in water at concentrations of 2.5,
5.0, 7.5, 10.0, and 15.0 μg/mL. QCM measurements were
obtained with a Q-Sense E4 QCM-D system, in which standard
flow modules were employed. A flow rate of 0.15 mL/min was
used for each analysis. The third overtone frequency was fit
using a Langmuir isotherm from which the binding constants
can be determined via known methods.16 CD spectra were
obtained on a Jasco J-815 CD spectrometer using a 750 μL
quartz cuvette with a path length of 0.5 cm. Peptides were
dissolved in water where a concentration of 16 μM was used for
all measurements. Nanoparticles were prepared at equivalent
ratios and likewise diluted for CD. The CDPro software
package was used for analysis of CD spectra to deconvolute
secondary structure contributions.
Molecular Simulation. Models of Pd {111} surfaces and
the six peptides (Pd4, C6, C11, C6,11, C6A11, and A6C11) in
aqueous solution were prepared on the scale of ∼3 × 3 × 6
nm3. The models were subjected to molecular dynamics
simulation using the INTERFACE-CHARMM force field
(same as CHARMM-METAL force field), including extensions
for partially covalent binding of Cys residues, as well as
advanced equilibration and sampling techniques. Binding
conformations, adsorption energies, and the role of specific
residues to surface binding were analyzed as an average over
more than 20 ns simulation time for each peptide (see details in
the Supporting Information).
Table 1. Pd4 and Cysteine Analogue Peptides
localized, residue-specific binding effects dominate property
control, which was reinforced by theoretical modeling of the
binding effects of the peptide on the Pd surface. Taken
together, these results are important as they demonstrate that
peptide design could be employed to generate sequences that
optimize material properties, which cannot be selected for
through biocombinatorial approaches.
EXPERIMENTAL SECTION
■
Chemicals. FMOC-protected Wang resins, functionalized
with a protected leucine residue, FMOC-protected amino acids,
diisopropyl ethylamine (DIPEA), piperidine, N-hydroxybenzo-
triazole (HoBT), and O-benzotriazole-N,N,N′,N’-tetramethyl-
uronium hexafluorophosphate (HBTU) were purchased from
Advanced Chemtech. Triisopropylsilane, 4-iodobenzoic acid,
and K2PdCl4 were purchased from Sigma-Aldrich. NaBH4,
CDCl3, phenyltin trichloride (PhSnCl3), trifluoroacetic acid
(TFA), anhydrous Na2SO4, NaCl, and KOH were acquired
from VWR. N,N-dimethylformamide (DMF), acetonitrile,
diethyl ether, and methanol were purchased from VWR. All
chemicals were used as received. Synergy Ultrapure UV water
(18mΩ•cm; Millipore) was used for all experiments.
Nanoparticle Synthesis. Peptides were synthesized via
standard FMOC synthetic protocols,14 purified by reverse-
phase HPLC, and confirmed by MALDI-TOF mass spectrom-
etry. Once the sequence was confirmed, the peptides were used
to fabricate peptide-capped Pd nanoparticles at a 3.3:1
Pd:peptide ratio, as previously described.9 Briefly, for each
peptide, a 4.90 mL solution of ∼150 μM of each peptide and
500 μM K2PdCl4 were prepared as Pd2+-peptide complex
reactions. After the complexes were allowed to incubate for
30.0 min to ensure completion of the binding event, 100 μL of
a freshly prepared 100 mM NaBH4 solution was added, which
resulted in nanoparticle formation. To complete the reduction
process, the reaction was allowed to stand for 1.00 h.
Catalytic Analysis. For the catalytic analyses, the Stille C−
C coupling reaction was employed.9,12,15 For this reaction, 1.25
mmol of 4-iodobenzoic acid and 1.56 mmol of PhSnCl3 were
codissolved in sufficient 2.25 M KOH. After dissolution, 1.25
mL of the stock Pd nanoparticles was added, resulting in a 0.05
mol % Pd concentration. The final volume of the reaction was
20.0 mL, which was accounted for based upon the volume of
KOH added. After reaction initiation, 2.0 mL aliquots were
extracted and quenched with 12.5 mL of 5.0% HCl at selected
time points over a 1.00 h reaction time. The final product,
biphenylcarboxylic acid, was extracted and quantitated using
described methods.9
RESULTS AND DISCUSSION
■
Based on modeling studies, which predicted metal surface
binding at the six and eleven positions of the native Pd4
peptide, a set of six different sequences was prepared (Table
1).10 Pd4 possesses histidine residues at both locations. To site
specifically increase the binding strength at these locations, two
single cysteine modified peptides, C6 and C11, were prepared
that substituted the histidine residue for cysteine at the six and
eleven positions, respectively, while a double modification
peptide, C6,11, was also generated that modified both
histidines to cysteines. Two additional peptides were also
designed that included both cysteine and alanine modifications:
the A6C11 that incorporated an alanine at the sixth position
and a cysteine at the eleventh position and the C6A11 peptide
with a cysteine and alanine modification at the six and eleven
positions, respectively. These two sequences were generated to
substantially change the binding strength at the two positions,
where minimal to no binding at the alanine sites was
anticipated, with the strongest binding from the cysteine
residues.
QCM analysis was initially employed to determine changes
in the global Pd affinity of the peptides based upon the
sequence changes. Figure 1a presents the observed frequency
changes for the C6A11 peptide binding to a Pd sensor at five
selected concentrations. As anticipated for all peptide
concentrations, as time increases, the change in frequency
decreases until saturation, consistent with binding of the target
surface. Note that an inverted plot is presented to enhance the
Characterization. A 5.00 μL sample of each nanoparticle
solution was deposited and allowed to dry on a 400-mesh Cu
grid coated with an ultrathin carbon layer. TEM images were
obtained with a JEOL 2010F or Phillips CM 200, operating at
200 kV, or an FEI Titan TEM operating at 300 kV. The analysis
11049
dx.doi.org/10.1021/ja402215t | J. Am. Chem. Soc. 2013, 135, 11048−11054

Journal of the American Chemical Society
Article
the binding strength regardless of the position; however,
changes to the peptide sequence/conformation may occur to
increase the affinity. Indeed, based upon computational
modeling discussed later, when bound to a Pd surface,
increased peptide flexibility was noted for the A6C11 peptide
as compared to the C6A11 sequence, consistent with their
degrees of surface affinity. Furthermore, altered biotic/abiotic
interfaces are likely, which were confirmed via CD analysis of
the peptides before and after nanoparticle binding (discussed
below).
Upon confirmation of the Pd affinity, the peptides were
employed as passivating ligands to fabricate Pd nanoparticles
using standard approaches.8,9 The materials were then analyzed
via UV−vis spectroscopy and transmission electron microscopy
(TEM). UV−vis analysis of each peptide solution displayed a
relatively featureless spectrum as anticipated (Supporting
Information, Figure S2). Upon addition of K2PdCl4, the
spectra of the Pd2+ complexed peptides demonstrated the
appearance of a ligand-to-metal charge transfer (LMCT) band
at ∼215 nm. After addition of NaBH4, the reduced materials
displayed a broad absorbance that increased toward lower
wavelengths and the lack of an LMCT peak, all of which
suggested nanoparticle formation.8,9
Figure 1. QCM analysis of peptide binding: (a) observed frequency
changes based upon peptide binding as a function of C6A11
concentration, (b) plot of the calculated kobs values as a function of
C6A11 concentration, and (c) Pd adsorption analysis for the parent
Pd4 peptide and cysteine analogues.
TEM imaging of the materials capped with the designed
peptides is presented in Figure 2. For the Pd4-capped particles,
clarity of data presentation. From the raw QCM data, a series of
five kobs values were determined employing previously described
methods,16 which were then plotted against the peptide
concentration (Figure 1b). Linear fitting of the data was then
used to determine the association (ka) and dissociation (kd)
constants of peptide binding from the slope and y-intercept
values of the best-fit line, respectively.16 Similar studies were
conducted for all of the peptides of Table 1 (Supporting
Information, Figure S1), where the binding constants for all
sequences are listed in Figure 1c. From this analysis, the parent
Pd4 peptide was observed to possess a ka of 1693 169 M−1
s−1 and a kd of 0.0028 0.0006 s−1, which results in a Keq value
of 6.1 × 105 1.3 × 105 M−1 and a ΔG value of −33.0 0.5
kJ/mol. Interestingly, when cysteine-only modifications were
employed (C6, C11, and C6,11), while slightly lower ka values
were noted compared to the Pd4, substantially diminished kd
values were observed. As a result, significantly increased Keq
values of 1.99 × 106 7.3 × 105 M−1, 1.47 × 106 7.4 × 105
M−1, and 1.72 × 106 8.5 × 105 M−1 were determined for the
C6, C11, and C6,11 peptides, respectively. This in turn leads to
more negative ΔG values of −35.9 0.9 kJ/mol for the C6,
−35.1 1.2 kJ/mol for the C11, and −35.5 1.2 kJ/mol for
the C6,11 peptides, indicating a greater affinity for Pd.
Figure 2. TEM analysis of the Pd nanoparticles capped with (a) Pd4,
(b) C6, (c) C11, (d) C6,11, (e) A6C11, and (f) C6A11. Main scale
bars are 10 nm, while inset bars are 2 nm.
an average size of 2.0
0.4 nm was noted, consistent with
previous studies;8 however, for the particles prepared with the
cysteine-substituted peptides, a slight size increase to 2.2 0.3
nm (C6), 2.4 0.4 nm (C11), and 2.3 0.4 nm (C6,11) was
observed. Furthermore, for the materials capped with either the
A6C11 or C6A11 peptide, particles of an identical size and
When both alanine and cysteine-based peptide modifications
were employed, unique changes in the Pd binding constants
were observed. For instance, analysis of the A6C11 peptide
demonstrated ka and kd values of 1563
210 M−1 s−1 and
0.0028 0.0009 s−1, respectively, thus resulting in a Keq of 5.6
× 105 1.9 × 105 M−1 and a ΔG of −32.7 0.8 kJ/mol. Such
values are quite similar to those observed for the parent Pd4
peptide, suggesting that the increased binding effect of the
cysteine residue may be offset by the weakened binding of the
alanine. Analysis of the C6A11, however, demonstrated binding
constants of 1827 113 M−1 s−1, 0.0007 0.0004 s−1, 2.61 ×
106 1.42 × 106 M−1, and −36.5 1.3 kJ/mol for the ka, kd,
Keq, and ΔG values, respectively. From this analysis, the
strongest binding was observed from the C6A11 peptide,
within the error of the analysis, while the weakest binding was
observed from the A6C11. Such results were somewhat
surprising as alanine modifications were anticipated to diminish
distribution of 2.4
0.4 nm were noted. High-resolution
analysis of selected cysteine-containing materials, shown as
inserts in Figure 2, demonstrated the formation of single crystal
particles; Pd4-based systems have been shown to produce
single crystals particles.8,11 Such similar sizes for the particles
prepared with the A6C11 and C6A11 peptides were surprising
as they have different binding affinities, based upon their ΔG
values; however, such results may be related to the interaction
mechanism of individual residues with the surface rather than
the aggregate metallic affinity of the peptide.12,17 This suggests
that the resultant particle size is based upon factors such as
peptide binding motifs and surface recognition elements, rather
than on the absolute degree of affinity for the target material. It
11050
dx.doi.org/10.1021/ja402215t | J. Am. Chem. Soc. 2013, 135, 11048−11054

Journal of the American Chemical Society
Article
is also interesting to note that particles of nearly the same size
were generated with the designed sequences (2.2 − 2.4 nm).
Such results are different than previous work with alanine-only
substituted peptides that displayed controllable changes in
particle size based upon sequence modifications.11 This
indicates that the presence of cysteine residues may modify
the biorecognition capabilities of the peptides through metal−
sulfur interactions; however, the overall structure of the
different bound peptides is anticipated to vary based upon
the altered sequences and the binding of the other residues to
the particle surface.
at 205 nm and slight positive shoulder at 228 nm).18 The Pd-
bound C6 peptide possessed similar CD peaks to C6A11
(Figure 3c), although the C6 bound peptide appears to be less
structured and more disordered (59.3%) based on differences
in ellipticity. For the set of peptides with cysteine only in the 11
position, the CD spectra of the Pd-bound C11 and A6C11
peptides exhibited similarities in terms of peak shapes and
wavelengths, where both represented unordered structures on
Pd (Figure 3d). Again, the Pd-bound C11 peptide has a
decreased ellipticity relative to A6C11 on Pd and is more
unfolded. Lastly, the cysteine double substituted C6,11 adopts a
hybrid surface structure on Pd similar to C6 and C11 based on
equal ellipticity values and characteristic spectral components of
both Pd-bound C6 and C11 peptides. In total, this suggests that
the cysteine in the peptide six position likely affects the overall
surface structure, while the presence of histidine or alanine at
the eleven position attenuates the magnitude of ellipticity,
resulting in a less native and more disordered structure. When a
cysteine residue was present in the six position, followed by a
binding residue at the eleven position, which is observed for the
C6 and C6,11 sequences, the peptides adopt a more disordered
structure (59.3% and 55.2% on the Pd surface) due to peptide
pinning at the central position and additional strong binding at
the end of the sequence. It is interesting to note that peptides
with cysteine modifications at the same site (C6 and C6A11 vs.
C11 and A6C11, Figures 3c and d, respectively) have
remarkably similar CD spectra when bound to Pd. More
importantly, the particles capped with the peptides from the
first group possessed similar catalytic activities, while those
capped with peptides from the second group had increased,
similar reactivities. Furthermore, the peptide structural differ-
ences can also be resolved using computational modeling,
where it appears that these subtle conformational differences
observed by CD for the peptide set contribute to the large
differences in catalytic performance.
To investigate the trends in peptide binding noted using CD
studies, the adsorption of the six different sequences on Pd
{111} surfaces was investigated using the INTERFACE-
CHARMM force field,17,19,20 including the specifically stronger
bonds of the cysteine thiol sulfur to Pd to reflect covalent
bonding contributions (see Supporting Information for de-
tails).21 The equilibrium binding conformation of single
peptides was flat-on in all cases and surface-bound water was
almost fully replaced (Figure 4 and Supporting Information,
Figure S4). Attractive sites on the metal surface for the peptides
were not the atoms in the top metal layer, but the atoms in the
two layers underneath. This epitaxial interaction mechanism,
which is well supported by other simulation and experimental
studies,8,17,22−24 leads to the meandering patterns of the
peptides that roughly follow the hexagonal symmetry of the
{111} surface. The thiol group also showed a preference toward
epitaxial sites in agreement with experimental observations on
Au surfaces.25,26 Moreover, the relative trend in computed
binding energies of individual peptides to the Pd surface
(Supporting Information, Figure S5) agrees with the reported
binding constants of Figure 1c. Accordingly, the C6A11 peptide
is more strongly bound than the A6C11 peptide even though
particles of the same size are produced. The stronger attraction
of C6A11 is related to tighter surface contact at the sixth
residue next to a sterically demanding proline in the seventh
position as opposed to loosening the surface contact by
positioning an alanine residue in the sixth position as in A6C11.
Local tightening of surface contact appears to be a main
Structurally, CD spectroscopy showed significant changes in
peptide conformation before and after Pd binding that is strictly
dependent on the location of cysteine within the sequence
(Figure 3 and Supporting Information, Figure S3). Quantita-
Figure 3. CD analysis of the peptides (a) before and (b) after binding
to the Pd nanoparticles. Part (c) presents the CD spectra of C6 and
C6A11 peptides bound to Pd, while (d) displays the CD spectra of
C11 and A6C11 peptides bound to Pd.
tively, the largest change in peptide structure was observed with
the C6 and C6,11 peptides after Pd binding. Both Pd bound
peptides showed a large decrease in ellipticity at ∼203 nm
compared to the free peptides, indicative of highly disordered
secondary structures (59.3% for C6 and 55.2% for C6,11). This
is a result of strong binding to the Pd surface through histidine/
cysteine and cysteine/cysteine binding pairs. For the C6A11
and A6C11 peptides on Pd, the peptide structures were
substantially less disordered, 35.3% and 26.4%, respectively,
because of the presence of only a single binding residue at the
six or eleven position. The low extent of unfolding reflects the
increased conformational flexibility of the peptides on Pd when
anchored at either cysteine residue. For C11 and the original
Pd4 peptide, these structures unfolded by 41.0% and 41.4%,
respectively, as compared to the analysis of the free peptides in
solution.
The CD spectra of Pd-bound C6A11 (Figure 3d) showed the
presence of a shoulder at ∼240 nm (slightly positive peak) and
a wavelength shift from 201 to 205 nm for the dominant
negative CD peak. This most likely is indicative of a compact
peptide structure on the Pd surface and exhibits partial
characteristics of a polyproline type II (PP2) helix (∼18%
contribution). This structure assignment is consistent with the
spectral features of a 9-mer polyproline peptide (negative peak
11051
dx.doi.org/10.1021/ja402215t | J. Am. Chem. Soc. 2013, 135, 11048−11054

Journal of the American Chemical Society
Article
As is evident from the CD studies and molecular modeling,
different peptide conformations are present on the metallic
surface. This effect arises from the position of the cysteine
within the overall sequence. To this end, when cysteine is
present in the sixth position, the near covalent interaction of
the thiol with Pd pins the sequence directly at the center of the
biomolecule to the metallic surface. This binding is enhanced
by the sterically demanding proline residue at the seven
position that has significant effects on peptide secondary
structure, thus greatly minimizing the degree of sequence
flexibility once bound. This is in contrast with the peptides that
position the cysteine residue at the eleven position, which
demonstrated increased flexibility when bound to Pd. In this
regard, peptide pinning is present at the C-terminal region,
allowing the rest of the sequence additional freedom to bind to
the metallic surface. Such pinning and flexibility effects were
predicted by computational modeling and observed via CD
studies that displayed similar peptide structures when bound to
the particle surface for the C6 and C6A11 peptides, as well as
for the C11 and A6C11 sequences. This degree of flexibility
may play a dramatic role over the catalytic activity of the
particles, which is dependent upon the exposure of the metallic
surface to reagents in solution.
Figure 4. Top view onto equilibrium conformations of single peptides
adsorbed on extended Pd {111} surfaces in aqueous solution from MD
simulation for (a) Pd4, (b) C6, (c) C11, (d) C6,11, (e) C6A11, and
(f) A6C11. The position of partly covalently bound SH groups in
cysteine is highlighted. Metal atoms are shown in decreasing size from
the top to sublayers.
Upon characterization of the particle/peptide surface
structure, Stille C−C coupling was used to probe the catalytic
activity of the particles (Figure 5). Such studies can be
contributor to stronger adsorption of cysteine-containing
peptides versus cysteine-free peptides (Supporting Information,
Figure S5 and Figure 1c). The localized attraction of cysteine to
one epitaxial site on a facet is also in contrast to stabilization of
an area of multiple epitaxial sites by histidine, especially the
conformationally flexible histidine in the eleventh position
(Figures 4a and b). In agreement with TEM data (Table 1 and
Figure 2), histidine appears to be a better growth-limiting
residue than cysteine and leads to smaller particle sizes.
It is important to realize that the nanoparticles produced
consist of multiple facets27 and binding measurements by QCM
rely on polycrystalline substrates that contain {111}, as well as
{100} and {110} surfaces. Moreover, adsorption of peptides
may not create perfect monolayers of 0.4 nm thickness.28,29
Earlier QCM and AFM studies on Au {111} surfaces showed
irregular peptide agglomerates that were related to interactions
between multiple peptides and potentially biomolecule folding
on the surface.16 The current QCM data similarly show the
doubling of surface coverage from lower to higher concen-
tration (Figure 1a and Supporting Information, Figure S1),
suggesting increased peptide adsorption at higher concen-
trations. The Langmuir adsorption parameters including
binding constants are averaged over {hkl} facets and multiple
peptides, thus potentially reflecting peptide-surface as well as
peptide−peptide interactions. It has recently been suggested
that the influence of such interactions can diminish adsorption
energies from −50 kcal/mol for single peptides to less than
−10 kcal/mol for peptides in multilayers.30 Binding free
energies of −8 to −9 kcal/mol from the Langmuir fit reflect
possible multilayer formation and correspond to the strength of
about 3 hydrogen bonds (Figure 1). The adsorption of the first
peptide, including a partly covalent thiol bond, is expected to be
higher, possibly −50 to −85 kcal/mol as seen in the simulation
(Supporting Information, Figure S5). Furthermore, the long
time to reach binding saturation likely arises from peptide
surface rearrangement.
Figure 5. TOF analysis of the peptide-capped Pd nanoparticles for
Stille coupling. Part (a) presents the model catalytic reaction, while
part (b) displays the TOF values achieved using the nanoparticles
capped with the indicated peptide. Pd4 and A6,11 data acquired
previously.12
employed to ascertain the structural effects of the peptides on
the overall material properties. Note that particles of nearly
identical sizes were used for this reaction, thus differences in the
reactivity are likely attributable to changes in the peptide/
particle surface structure. For this reaction, the coupling of 4-
iodobenzoic acid with PhSnCl3 to prepare biphenylcarboxylic
acid in water at room temperature was used (Figure 5a). From
these studies, a clear trend in the catalytic activity was noted
that directly depended upon the cysteine position within the
11052
dx.doi.org/10.1021/ja402215t | J. Am. Chem. Soc. 2013, 135, 11048−11054

Journal of the American Chemical Society
Article
peptide sequence; however, the reactivity was independent
from the global peptide binding strength. Such results were
quite surprising as thiols are well-known poisons for noble
metal-based nanocatalysts.31,32 The opposite increase in
reactivity could be related to an active role of the thiol groups
to support atom leaching, a proposed step in nanoparticle-
driven C−C coupling, by stabilizing local defects. Further
poisoning effects are likely to be attenuated by the binding of
the additional sequence residues to the particle surface, as
discussed below.
and even enhanced through individual residue changes at the
peptide sequence level. Note that the Stille C−C coupling
reaction has been suggested to occur via an atom leaching
mechanism.5,33 In this approach, during oxidative addition at
the particle surface via the aryl halide, abstraction of Pd is
possible. From this, the reaction then occurs in solution, where
a free Pd species is released after completion of the reaction.
This species can then be recycled through the reaction,
providing sufficient starting material is present, or be quenched
by the remaining nanoparticles. In the present study, greater
TOF values were observed from the particles capped with
cysteine-containing peptides over the cysteine-free parent Pd4
sequence. This suggests that such modifications alter the
structure/binding of the peptide on the particle surface,
consistent with the CD and modeling data, to give rise to
this level of reactivity increase. This is surprising in light of
known thiol-based poisoning of catalytic particles.31,32 In
general, strong thiol binding at active sites in dispersed Pd
catalysts is observed, along with steric crowding at the surface,
both of which contribute to diminished reactivity. Such
capabilities are likely to be attenuated by the additional
functional groups of the peptide sequence binding to the
particle. For the peptides, it is likely that the sequences strongly
bind through the thiol moiety of the cysteine residue, which
binds at the 3-fold interstitial site on the metallic surface to pin
the biomolecule to the inorganic material. From this, the
additional residues of the peptide are then arranged to provide
optimal surface binding, which changes based upon the position
of the cysteine within the sequence. This binding mechanism
would provide steric bulk to the ligands to prevent additional
thiol-based surface poisoning, while allowing substrates access
to the catalytic surface to drive Pd leaching during oxidative
addition. To this end, these results indicate that thiol-based
ligands can be designed as nanoparticle stabilizers for catalytic
applications should sufficient offsetting binding moieties be
present.
Furthermore, when comparing the modified sequences, an
interesting trend was observed; for those peptides with stronger
binding at the eleven position and weaker binding at the six
position, increased reactivity with greater TOF values was
demonstrated (Scheme 1). Such TOFs were larger than those
observed for the materials prepared with the sequences that
positioned amino acids with stronger and/or equal binding
strengths at the six position over the eleven position. In
addition, the Pd4-capped Pd nanoparticles, which were the
smallest of the set, thus having the greatest surface area,
demonstrated the lowest degree of reactivity. Together, this
evidence indicates that the peptide sequence and highly
localized residue/surface binding inherently affects the particle
functionality, controlled via surface binding that can be used as
a design point to modify particle properties for applications. To
this end, highly localized biomolecular binding, based upon the
identity and position of functional groups within the peptide
was directly proportional to the catalytic reactivity of the
materials. Furthermore, no observable trends were present
between the global peptide Pd affinity (ΔG values), the particle
size, and the observed reactivity of the materials as nearly the
same size nanoparticles were prepared by all five modified
sequences. As such, the observed changes in TOF can be
attributed to the surface-bound peptide structure, directed via
thiol surface pinning based upon the location of cysteine
residues within the sequence. These changes in the particle
surface morphology, as confirmed by CD and computational
For comparison, as previously reported, the Pd4-capped
particles demonstrated a TOF value of 2234 99 mol product
(mol Pd × hr)−1.12 For the particles capped with C6, C6,11,
and C6A11 peptides, an approximate 2-fold increase in the
reaction efficiency was observed, displaying TOFs of 3963 28
mol product (mol Pd × hr)−1, 3974 280 mol product (mol
Pd × hr)−1, and 4147 340 mol product (mol Pd × hr)−1,
respectively. Note that for all of these peptides, the strongest
residue binding was incorporated at the six position of the se-
quence. Remarkably, a nearly 3-fold TOF increase over the
Pd4-capped particles for the model reaction was observed for
the materials prepared using the C11 and A6C11 with TOF
values of 6138 55 mol product (mol Pd × hr)−1 and 6097
65 mol product (mol Pd × hr)−1, respectively. For these
sequences, weaker binding was incorporated at the sixth
position as compared to the stronger binding of the cysteine
residue at the eleventh position. Such high TOF values
indicated a direct correlation between the peptide sequence and
localized surface binding, particle surface structure (based upon
CD and theoretical modeling), and catalytic reactivity. For
comparison, the A6,11 peptide, with both histidines replaced
with alanines, was previously employed to prepare peptide-
capped Pd nanoparticles that demonstrated a TOF value of 361
21 mol product (mol Pd × hr)−1.12 These materials, whose
peptide removed thiol-binding residues completely, demon-
strated the lowest degree of reactivity, following the anticipated
trend, which likely was an effect of the individual residues at
specific peptide sequence positions.
Comparing the reactivity of the particles as a function of the
cysteine position (Scheme 1) provides important evidence to
indicate that the materials functionality can be directly tuned
Scheme 1. Effects of peptide sequence on the catalytic
activity of peptide-capped Pd nanoparticles
11053
dx.doi.org/10.1021/ja402215t | J. Am. Chem. Soc. 2013, 135, 11048−11054

Journal of the American Chemical Society
Article
modeling, are likely to enhance the reaction mechanism, which
has been suggested to follow a unique atom leaching
process.5,27,33 For this, changes in the surface bound peptide
morphology may more efficiently display the reactive metallic
species to the reagents in solution to allow for more rapid Pd
leaching during oxidative addition. Should this process be
enhanced, additional Pd species would be present in solution to
drive the reaction, thus raising the TOF value. Furthermore,
increased peptide flexibility on the particle surface was noted
for the C11 and A6C11 peptides, as compared to the C6 and
C6A11 sequences. This flexibility may also directly modulate
the reactivity, where the biomolecules can more readily distort
their surface structure to accommodate the interactions of the
aryl halide with the metallic surface to facilitate the leaching
process. Such peptide structural arguments are supported by
the CD observations and modeling studies where similar
biomolecular structures were observed for the peptides with
cysteine at either the six or eleven position, thus directly
correlating with the catalytic observations. As such, localized
residue binding affects appear to play a more important role in
the activity of peptide-capped materials over the global peptide
interaction.
Computational resources from the Ohio Supercomputing
Center are also acknowledged.
REFERENCES
■
(1) Dickerson, M. B.; Sandhage, K. H.; Naik, R. R. Chem. Rev. 2008,
108, 4935.
(2) Bhandari, R.; Coppage, R.; Knecht, M. R. Catal. Sci. Technol.
2012, 2, 256.
(3) Briggs, B. D.; Knecht, M. R. J. Phys. Chem. Lett. 2012, 3, 405.
(4) Chiu, C. Y.; Li, Y. J.; Ruan, L. Y.; Ye, X. C.; Murray, C. B.; Huang,
Y. Nat. Chem. 2011, 3, 393.
(5) Pacardo, D. B.; Slocik, J. M.; Kirk, K. C.; Naik, R. R.; Knecht, M.
R. Nanoscale 2011, 3, 2194.
(6) Mao, C. B.; Solis, D. J.; Reiss, B. D.; Kottmann, S. T.; Sweeney, R.
Y.; Hayhurst, A.; Georgiou, G.; Iverson, B.; Belcher, A. M. Science
2004, 303, 213.
(7) Ahmad, G.; Dickerson, M. B.; Cai, Y.; Jones, S. E.; Ernst, E. M.;
Vernon, J. P.; Haluska, M. S.; Fang, Y.; Wang, J.; Subrarnanyarn, G.;
Naik, R. R.; Sandhage, K. H. J. Am. Chem. Soc. 2008, 130, 4.
(8) Coppage, R.; Slocik, J. M.; Briggs, B. D.; Frenkel, A. I.; Heinz, H.;
Naik, R. R.; Knecht, M. R. J. Am. Chem. Soc. 2011, 133, 12346.
(9) Pacardo, D. B.; Sethi, M.; Jones, S. E.; Naik, R. R.; Knecht, M. R.
ACS Nano 2009, 3, 1288.
(10) Pandey, R. B.; Heinz, H.; Feng, J.; Farmer, B. L.; Slocik, J. M.;
Drummy, L. F.; Naik, R. R. Phys. Chem. Chem. Phys. 2009, 11, 1989.
(11) Coppage, R.; Slocik, J. M.; Briggs, B. D.; Frenkel, A. I.; Naik, R.
R.; Knecht, M. R. ACS Nano 2012, 6, 1625.
(12) Coppage, R.; Slocik, J. M.; Sethi, M.; Pacardo, D. B.; Naik, R. R.;
Knecht, M. R. Angew. Chem., Int. Ed. 2010, 49, 3767.
(13) Hoefling, M.; Iori, F.; Corni, S.; Gottschalk, K. E. Langmuir
2010, 26, 8347.
(14) Chan, W. C.; White, P. D. Fmoc Solid Phase Peptide Synthesis: A
Practical Approach; Oxford University Press: New York, 2000.
(15) Garcia-Martinez, J. C.; Lezutekong, R.; Crooks, R. M. J. Am.
Chem. Soc. 2005, 127, 5097.
CONCLUSIONS
■
In summary, we have demonstrated that the properties of
peptide-capped nanomaterials can be modulated based upon
the position of amino acids within the peptide sequence. In the
present catalytic system, we observed an increase in reactivity as
a result of cysteine and alanine substitutions at selected
sequence positions. From empirical observations, localized
peptide modifications were exploited where the placement of
such residues within the sequence was determined based upon
individual amino acid affinities and computationally predicted
motifs. These results indicate that while biocombinatorial
techniques can isolate peptides with strong affinity, rational
design could be employed to optimize material functionality.
Through this, property enhancement was demonstrated with
no apparent changes to the metallic core structure. Such
capabilities could prove to be useful for a variety of materials
outside of catalysis, including plasmonic, magnetic, and
electronic structures.
(16) Tamerler, C.; Oren, E. E.; Duman, M.; Venkata-subramanian,
E.; Sarikaya, M. Langmuir 2006, 22, 7712.
(17) Feng, J.; Pandey, R. B.; Berry, R. J.; Farmer, B. L.; Naik, R. R.;
Heinz, H. Soft Matter 2011, 7, 2113.
(18) Rucker, A. L.; Creamer, T. P. Protein Sci. 2002, 11, 980.
(19) Heinz, H.; Vaia, R. A.; Farmer, B. L.; Naik, R. R. J. Phys. Chem. C
2008, 112, 17281.
(20) Heinz, H.; Lin, T. J.; Mishra, R. K.; Emami, F. S. Langmuir 2013,
29, 1754.
(21) Dubois, L. H.; Zegarski, B. R.; Nuzzo, R. G. J. Am. Chem. Soc.
1990, 112, 570.
ASSOCIATED CONTENT
* Supporting Information
Details of QCM, UV−vis analysis, CD deconvolution,
computational methods and results. This material is available
free of charge via the Internet at http://pubs.acs.org.
(22) Heinz, H.; Farmer, B. L.; Pandey, R. B.; Slocik, J. M.; Patnaik, S.
S.; Pachter, R.; Naik, R. R. J. Am. Chem. Soc. 2009, 131, 9704.
(23) Feng, J.; Slocik, J. M.; Sarikaya, M.; Naik, R. R.; Farmer, B. L.;
Heinz, H. Small 2012, 8, 1049.
(24) Ruan, L. Y.; Ramezani-Dakhel, H.; Chiu, C. Y.; Zhu, E. B.; Li, Y.
J.; Heinz, H.; Huang, Y. Nano Lett. 2013, 13, 840.
(25) Dubois, L. H.; Nuzzo, R. G. Annu. Rev. Phys. Chem. 1992, 43,
437.
■
S
AUTHOR INFORMATION
Corresponding Author
*Corresponding Author knecht@miami.edu; rajesh.naik@
wpafb.af.mil
■
(26) Ulman, A. Chem. Rev. 1996, 96, 1533.
(27) Ramezani-Dakhel, H.; Mirau, P. A.; Naik, R. R.; Knecht, M. R.;
Heinz, H. Phys. Chem. Chem. Phys. 2013, 15, 5488.
(28) Heinz, H.; Vaia, R. A.; Krishnamoorti, R.; Farmer, B. L. Chem.
Mater. 2007, 19, 59.
Notes
The authors declare no competing financial interest.
(29) Drummy, L. F.; Jones, S. E.; Pandey, R. B.; Farmer, B. L.; Vaia,
R. A.; Naik, R. R. ACS Appl. Mater. Interfaces 2010, 2, 1492.
(30) Ramezani-Dakhel, H.; Ruan, L. Y.; Huang, Y.; Heinz, H.
Submitted 2013.
(31) Ibanez, F. J.; Zamborini, F. P. Langmuir 2006, 22, 9789.
(32) Quinn, R.; Dahl, T. A.; Toseland, B. A. Appl. Catal., A 2004,
272, 61.
ACKNOWLEDGMENTS
■
This material is based upon work supported by the National
Science Foundation under Grant No. CBET-1157431 (M.K.),
DMR-0955071 (H.H.), and by AFRL/RX and AFOSR (H. H.
and R.N.). R.C. acknowledges fellowship support from the UM
CAS Dissertation Year Fellowship, while N.B. is supported by
the National Research Council Research Associateship award.
(33) Pacardo, D. B.; Knecht, M. R. Catal. Sci. Technol. 2013, 3, 745.
11054
dx.doi.org/10.1021/ja402215t | J. Am. Chem. Soc. 2013, 135, 11048−11054