Variation of formal hydrogen-bonding networks within electronically delocalized π-conjugated oligopeptide nanostructures

Article abstract of DOI:10.1021/la501999g

This photophysical study characterizes the generality of intermolecular electronic interactions present within nanomaterials derived from self-assembling oligopeptides with embedded π-conjugated oligophenylenevinylene (OPV) subunits stilbene and distyrylbenzene that in principle present two distinct β-sheet motifs. Two different synthetic approaches led to oligopeptides that upon self-assembly are expected to self-assemble into multimeric aggregates stabilized by β-sheet-like secondary structures. The target molecules express either two C-termini linked to the central OPV core (symmetric peptides) or the more common N-termini to C-termini polarity typical of natural oligopeptides (nonsymmetric peptides). Both peptide secondary structures were shown to form extended 1-D peptide aggregates with intimate intermolecular π-electron interactions. Differences in length of the π-conjugated OPV segments resulted in differing extents of intermolecular interactions and the resulting photophysics. The peptides containing the shorter stilbene (OPV2) units showed little ground state interactions and resulted in excimeric emission, while the longer distyrylbenzene (OPV3) peptides had different ground state interactions between adjacent π-conjugated subunits resulting in either perturbed electronic properties arising from exciton coupling or excimer-like excited states. Molecular dynamics simulations of nascent aggregate formation predict peptide dimerization to be a spontaneous process, possessing thermodynamic driving potentials in the range 2-6 kcal/mol for the four molecules considered. Antiparallel stacking of the peptides containing an OPV3 subunit is thermodynamically favored over the parallel orientation, whereas both arrangements are equally favored for the peptides containing an OPV2 subunit. This study validates the generality of peptide-π-peptide self-assembly to provide electronically delocalized supramolecular structures and suggests flexibility in peptide sequence design as a way to tune the material properties of π-conjugated supramolecular polymers.

Full text of DOI:10.1021/la501999g

Article
pubs.acs.org/Langmuir
Variation of Formal Hydrogen-Bonding Networks within
Electronically Delocalized π‑Conjugated Oligopeptide
Nanostructures
Brian D. Wall,† Yuecheng Zhou,∥ Shao Mei,∥ Herdeline Ann M. Ardona,†,‡ Andrew L. Ferguson,∥
̃
and John D. Tovar*,†,‡,§
‡
§
†Department of Chemistry, Krieger School of Arts and Sciences, Institute of NanoBioTechnology, and Department of Materials
Science and Engineering, Whiting School of Engineering, Johns Hopkins University, 3400 N. Charles St., Baltimore, Maryland 21218,
United States
∥Department of Materials Science and Engineering, University of Illinois at Urbana−Champaign, Urbana, Illinois 61801, United
States
S
* Supporting Information
ABSTRACT: This photophysical study characterizes the
generality of intermolecular electronic interactions present
within nanomaterials derived from self-assembling oligopep-
tides with embedded π-conjugated oligophenylenevinylene
(OPV) subunits stilbene and distyrylbenzene that in principle
present two distinct β-sheet motifs. Two different synthetic
approaches led to oligopeptides that upon self-assembly are
expected to self-assemble into multimeric aggregates stabilized
by β-sheet-like secondary structures. The target molecules
express either two C-termini linked to the central OPV core
(symmetric peptides) or the more common N-termini to C-termini polarity typical of natural oligopeptides (nonsymmetric
peptides). Both peptide secondary structures were shown to form extended 1-D peptide aggregates with intimate intermolecular
π-electron interactions. Differences in length of the π-conjugated OPV segments resulted in differing extents of intermolecular
interactions and the resulting photophysics. The peptides containing the shorter stilbene (OPV2) units showed little ground
state interactions and resulted in excimeric emission, while the longer distyrylbenzene (OPV3) peptides had different ground
state interactions between adjacent π-conjugated subunits resulting in either perturbed electronic properties arising from exciton
coupling or excimer-like excited states. Molecular dynamics simulations of nascent aggregate formation predict peptide
dimerization to be a spontaneous process, possessing thermodynamic driving potentials in the range 2−6 kcal/mol for the four
molecules considered. Antiparallel stacking of the peptides containing an OPV3 subunit is thermodynamically favored over the
parallel orientation, whereas both arrangements are equally favored for the peptides containing an OPV2 subunit. This study
validates the generality of peptide-π-peptide self-assembly to provide electronically delocalized supramolecular structures and
suggests flexibility in peptide sequence design as a way to tune the material properties of π-conjugated supramolecular polymers.
the molecular level.3,4 The overall properties of a peptidic
material can be tuned through variation in primary amino acid
INTRODUCTION
■
The folding and assembly of natural biological proteins
underpin many biological processes essential to life. Synthetic
oligopeptides have demonstrated great potential to self-
sequence and biological functionality can even be directly
perturbed by the peptidic with which it associates.5
The intermolecular associations among oligopeptides in their
assembled states can also be used in conjunction with π-
assemble unique macromolecular architectures with desirable
structural and functional properties.1 The rich structural
conjugated systems to drastically alter their (opto)electronic
possibilities when considering differing levels of peptide
organization, from primary amino acid sequence to secondary
properties.6−10 Such modifications in principle should lead to
an aggressive perturbation of protein-based assembly especially
and tertiary structure, showcase the power of peptides as
when non-natural aromatic quadrupoles of π-conjugated
versatile scaffolds for materials development. Beyond their
subunits are incorporated directly into the peptide backbones.
However, recent work from several laboratories has validated
structural diversity, the ability for peptides to assemble under
biologically relevant conditions and have specific interactions
with biological systems create a niche for powerful biomaterials
and tissue engineering schemes.2 The ability to control the
macromolecular architecture of peptide-based materials opens
up possibilities to rationally engineer the resulting materials at
Received: May 22, 2014
Revised: July 21, 2014
Published: September 2, 2014
© 2014 American Chemical Society
11375
dx.doi.org/10.1021/la501999g | Langmuir 2014, 30, 11375−11385

Langmuir
Article
Figure 1. DFAG peptides with different N-to-C backbone directionalities and different OPV lengths.
the self-assembly paradigm for peptide-π-peptide triblock
molecules where the π-conjugated internal units provide
spectroscopic handles for investigating peptide self-assem-
bly.11−14 The abilities to influence peptide association in a
predictable way and tune the intended macromolecular
structures are crucial. In this report we describe how the
nature of the peptide-based hydrogen-bonding network
influences the self-assembly and the nanostructural electronic
properties. Non-natural oligopeptides with the ability to form
1-D supramolecular aggregates with differing hydrogen bonding
motifs offer new design opportunities for unique electronic
biomaterials. The goal of this study is to provide new electronic
information about how the differences of peptide aggregation as
a primary function of hydrogen bonding motif can influence
optoelectronic properties of peptide aggregates. In a more
general sense, these studies also provide governing principles to
rationally design specific types of supramolecular aggregates
that may be useful in the development of biologically relevant
optoelectronic biomaterials.
side chains. We chose the oligophenylenevinylene chromo-
phores (OPVs) due to their established photophysical
properties, including stilbene (OPV2) and distyrylbenzene
(OPV3). In one presentation, the same sense of the peptide N-
to-C linkages (or “directionality”) was maintained along the
length of the molecule, thus leading to a final molecule with
one distinct N- and C-terminus (referred to here as
“nonsymmetric”) while in another presentation, the N-to-C
directionality emanates in opposing directions from the central
π-conjugated core in a mirror image sense, thus leading to a
final molecule presenting two C-termini (referred to here as
“symmetric”). We synthesized peptide-based OPV materials
with these two differing peptide sequence presentations using
methods developed in our lab as described further below. We
previously demonstrated how the variation of the peptide
sequence peripheral to the central chromophore unit impacts
the photophysical properties of the resulting supramolecular
nanostructures,15 and in this case we are in essence maintaining
the same structural units albeit with different backbone
directionality. In this report, we examine how these peptide
N-to-C polarities influence nanostructure photophysics, using
well-studied OPVs as prototypical examples.
Synthesis of OPV Amidation Partners and Peptides.
We reported several synthetic approaches to peptide-based
organic electronic materials, and we employ two of these
approaches here. The first method (leading to nonsymmetric
peptides 1 and 3) utilized π-conjugated OPV “amino acids” as
functional mimics of standard α-amino acids compatible with
standard solid-phase peptide synthesis (SPPS),11 where the
activated acid moiety of the amino acid N-acylated the amine
terminus of the growing oligopeptide while at the same time
presented a new (but protected) amine functionality for
continued amide bond formation. This synthesis approach
RESULTS AND DISCUSSION
■
We designed and synthesized four oligopeptides with
embedded phenylenevinylene chromophores as shown in
Figure 1 and studied their photophysical and morphological
properties. The oligopeptides studied here all consisted of the
same primary amino acid sequence, DFAG, with respect to
residue placement relative to the central π-conjugated core,
although the peptide backbone directionalities differ for each of
the two chromophore examples (Figure 1). The tetrapeptide
DFAG was chosen to provide adequate aqueous solubility at
neutral or basic pH and to undergo self-assembly upon
lowering the pH due to formal protonation of the carboxylate
11376
dx.doi.org/10.1021/la501999g | Langmuir 2014, 30, 11375−11385

Langmuir
Article
maintained the same sense of the peptide N-to-C linkages along
the length of the molecule. The second method (leading to
symmetric peptides 2 and 4) used π-conjugated diacids in an
on-resin site−site dimerization procedure that links two peptide
amine termini via amidation/N-acylation onto a common π-
electron core.16 Using the amino acid and diacid π-systems for
oligopeptide synthesis, we can now synthesize self-assembling
oligopeptide-π OPV hybrid models with different peptide
polarity presentation to probe how this important macro-
molecular assembly variable specifically impacts nanostructural
and electronic properties. We discuss briefly the synthesis of the
individual amino acid or diacid amidation partners used for the
peptide synthesis. More specific details regarding the peptide
syntheses can be found in the Supporting Information.
benzylphosphonate18 and methyl 4-formylbenzoate to give
the stilbene dimethyl ester20 which was then saponified to
afford the diacid coupling partner21 (Figure 2b) for the on-resin
dimerization among the N-deprotected terminal amines
presented on the synthesis support to provide the symmetric
OPV2 (2).
The synthesis of the OPV3 π-conjugated subunits revealed
solubility issues that initially motivated the development of the
on-resin dimerization procedure. The OPV3 amino acid proved
to be difficult to synthesize, with a successful route coming
from the two-step sequential Heck coupling onto p-
divinylbenzene (Figure 3a): the first with 4-iodoaniline and
The OPV2 amino acid was synthesized through a Horner−
Wadsworth−Emmons olefination with diethyl 4-nitrobenzyl-
phosphonate17 and methyl 4-formylbenzoate (Figure 2a).18
Figure 3. Synthesis of OPV3 amino acid (a) and diacid (b) for
incorporation into nonsymmetric and symmetric peptide backbones
present in 3 and 4, respectively.
the second with the NHS-activated ester of 4-iodobenzoic acid.
The sequential Heck reactions gave the isolable and directly
acid-activated OPV3 amino acid that could be directly coupled
onto the primary amine of the deprotected Wang resin without
the need for aryl amine protection. This reduced nucleophilicity
required that the subsequent amidation/N-acylation of the
terminal aryl amine with Fmoc-glycine to continue the SPPS be
conducted under the same electrophilic coupling conditions
(triphosgene and 2,6-lutidine) as necessary for SPPS with the
OPV2 amino acid, thus yielding the nonsymmetric OPV3
peptide (3). The symmetric OPV3 diacid was synthesized
previously16 through a double Horner−Wadsworth−Emmons
reaction between terephthaldehyde and a benzylic phosphonate
followed by subsequent saponification of the methyl esters22
(Figure 3b). This diacid was used for the on-resin dimerization
similarly to the OPV2 diacid, leading to the previously reported
symmetric OPV3 peptide (4).
Model of Peptide Aggregation. The standard model
invoked to understand the aggregation of these molecules is
their self-assembly into extended β-sheet-like structures (Figure
4).11−13 This picture is motivated by simplistic energy
minimization calculations, which predict these ordered one-
dimensional aggregates to be the most energetically favorable
conformations at 0 K. Our more recent work employing
molecular dynamics simulations at finite temperature has
shown that entropic considerations are also important for a
complete understanding of the thermodynamics and structure
of these aggregates15 and predict the perfect periodic order to
Figure 2. Synthesis of OPV2 amino acid (a) and diacid (b) for
incorporation into nonsymmetric and symmetric peptide backbones
present in 1 and 2, respectively.
Reduction of the nitro group using stannous chloride followed
by ester saponification and Fmoc protection of the amine gave
the OPV2 amino acid that could be used directly with SPPS.
Amidation between the OPV2 amino acid and the N-
deprotected peptide resin took place under normal HBTU
activation conditions. After Fmoc deprotection of the resin
bound OPV2, the weaker nucleophilicity of the resulting aryl
amine required the use of more electrophilic coupling reagents
for the next step in the SPPS sequence. The coupling of the
next amino acid was completed with the use of triphosgene and
2,6-lutidine,19 and all couplings of the remaining amino acids
could be completed using standard activation procedures
ultimately yielding nonsymmetric OPV2 (1). The OPV2 diacid
was synthesized through a Horner−Wadsworth−Emmons
olefination between diethyl 4-(methoxycarbonyl)-
11377
dx.doi.org/10.1021/la501999g | Langmuir 2014, 30, 11375−11385

Langmuir
Article
π-conjugated diyne subunits) to encourage electronic inter-
actions and even covalent chemistry in a topochemical sense.
Nanostructure Morphology. The acidic residues effec-
tively render all four peptides molecularly dissolved and not
significantly interacting at basic pH (ca. pH 10). Acidification of
the aqueous solutions (ca. pH 2) formally protonates the
carboxylate groups and screens the Coulombic repulsions
among the molecules, thereby allowing for intermolecular
association and subsequent nanostructure formation. The
OPV3 nanomaterials were found to bind Congo Red, thus
revealing their amyloidogenic nature. Transmission electron
microscopy (TEM) was employed to visualize the nanomateri-
als resulting from peptide assembly. It is well established that
variations in peptide sequence can have dramatic morphological
outcomes on molecular assembly, with structures spanning 0-D
vesicles and micelles, 1-D tapes, rods, and ribbons, and 2-D
sheet-like assemblies. When considering the photophysical
properties inherent to such geometrically disparate objects, it is
possible that local differences in packing densities, curvatures,
etc., could influence the spectral outcomes regardless of specific
structural variables. It was therefore important to find that the
nanomaterials afforded by the assembly of symmetric and
nonsymmetric peptides were comparable in their morpholo-
gies: all peptides examined yielded 1-D nanostructures with
consistent widths (10−12 nm) as determined by TEM (Figure
5). Given that the molecular lengths of these peptides are well
Figure 4. Energy-minimized models for the assembly of nonsymmetric
(a) and symmetric (b) OPV2 oligopeptides, illustrating idealized
antiparallel and parallel β-sheet structures, respectively. For simplicity,
the aspartic acid and phenylalanine residues have been replaced with
alanine.
break down, while still (and importantly) retaining one-
dimensionality. While our experimental and computational
results presented herein and elsewhere15 paint a significantly
more complex picture of assembly, the β-sheet-like structures
presented in Figure 4 provide a useful pedagogical idealization
for beginning to understand assembly and are broadly
consistent with the observed morphological outcomes,
electronic responses, and molecular simulations.
The symmetric (2 and 4) and nonsymmetric (1 and 3)
peptide architectures synthesized in this work present an ideal
starting point to understand how formal variations in the
requisite hydrogen-bonding patterns possible within idealized
nanostructures impact their photophysical properties. Our
simple model would predict that the nonsymmetric peptides 1
and 3 would maximize favorable energetic interactions by
assembling into 1-D aggregates consisting of antiparallel β-sheet
motifs, whereas the architecture of the symmetric peptides 2
and 4 would constrain them to form less favorable parallel β-
sheets. These assembled ribbons are presumed to aggregate
into higher order structures (stacked fibrils, etc.) typical of
oligopeptide and amyloid-like materials.23 This simple model-
ing scenario places the π-conjugated units into distances close
enough to expect electronic interactions and specific spectral
signatures, and the computational and experimental studies we
describe below provide strong support for the presence of these
interactions. The designed symmetries of these molecules
necessarily require phenylalanine steric clashing in both the
parallel and antiparallel configurations. We have previously
characterized these interactions for peptide 4 (symmetric
OPV3)15 and also observed polydiacetylene formation from
assembled peptide−diacetylene−peptide precursors.24 These
findings are all consistent with sufficient local intermolecular
ordering among the individual oligopeptides (and the reactive
Figure 5. TEM micrographs of peptide nanostructures prepared from
0.1 wt % of the respective DFAG peptides in acidic solution along with
their calculated widths: (a) nonsymmetric OPV3 (3: 9.6 1.4 nm);
(b) nonsymmetric OPV2 (1: 10.4 1.6 nm); (c) symmetric OPV3
(4: 10.5 1.3 nm); (d) symmetric OPV2 (2: 11.6 1.8 nm).
under 10 nm in their fully extended conformations, the present
structures reflect the inherent hierarchical bundling known to
be operative in peptide-based self-assembling materials but
nevertheless all remain comparable in size and shape. In all
cases, nanostructures with high aspect ratios were observed,
typically extending to micron lengths. These nanostructures can
be viewed as experimental manifestations of the idealized
assemblies shown in Figure 4 once extended along the π-
stacking axis.
Photophysical Characterization. The assembly of these
triblock molecules offers several possible intermolecular
electronic outcomes that have distinct spectral observables.
For example, exciton coupling via a classic H-like aggregate
11378
dx.doi.org/10.1021/la501999g | Langmuir 2014, 30, 11375−11385

Langmuir
Article
would lead to a blue-shift in the absorption maxima relative to
the isolated molecules, a quenched photoluminescence, and,
given the chirality expected for natural peptide aggregates, an
induced CD signal and bisignate Cotton effect corresponding
to the π−π* transition the OPV subunit. These general trends
have been observed in similar peptide-π-peptide motifs.11,13,16
Other possibilities include aggregates that may place the π-
electron units into J-like aggregates that would lead to red-
shifted absorption maxima and oftentimes enhanced photo-
luminescence or aggregated states with differing degrees of
chromophore planarity/rigidity that otherwise do not exhibit
any noticeable intermolecular electronic interactions. Given the
differences in hydrogen-bonding networks expected to exist
within nanostructures formed from the symmetric and
nonsymmetric peptide frameworks, a careful spectroscopic
explication of these electronic properties is warranted. The
polydispersities in nanomaterial morphologies indicate that the
spectral measurements should be treated as ensemble averages
over all local chromophore orientations present rather than
corresponding to a single defined orientation (such as within a
crystalline lattice).
OPV2. Aggregated stilbene systems have typically shown two
distinct spectral behaviors: one exhibiting broad featureless red-
shifted emission typically accompanied by subtle changes in the
absorption spectra and the other exhibiting structured emission
profiles and perturbed absorption spectra compared to the
isolated stilbene molecules. These two cases have been
attributed to structural differences of the aggregate stilbenes
with the first class corresponding to excimer-like emission
appearing to have neighboring face-to-face interactions of the
stilbene subunits and the second class corresponding to
excitonic-like emission with edge-to-face interactions between
the subunits.25−27
The nonsymmetric and symmetric OPV2 peptides 1 and 2
showed typical absorption and emission profiles in basic
solution. Both peptides had identical absorption properties
under both molecular and assembled conditions with λmax
values around 330 nm (Figure 6 and Table 1). The lack of
apparent blue-shifts upon OPV2 assembly can be attributed to
the shorter expected transition dipole (or the weaker oscillator
strength of the π−π* transition) of these small chromophores
that therefore do not experience as dramatic of an exciton
coupling as in longer chromophores. The differences in PL
between the two molecules can be rationalized by the different
chemical compositions of the chromophores themselves (an N-
acylcarboxamide and a dicarboxamide, respectively). Although
the N-acyl nitrogen is likely participating in amide bond
resonance, it is possible that it could serve as an electron donor
for the accepting carboxamide on the other end of the
chromophore unit. Once the pH of the spectroscopic solution
was lowered (thus triggering the assembly), the absorption
profiles did not change to any appreciable extent, but the
photoluminescence was predominately excimeric in nature with
featureless and broad peaks appearing at lower energies versus
the respective monomers (λmax differences of ca. 64 and 79 nm,
Figure 6a,b). The PL red-shifts noted for the nonsymmetric
peptides under both molecular and assembled conditions when
compared to the symmetric could reflect a stronger “donor−
acceptor” character that is better stabilized in the excited state.
Remarkably, although the absolute values were low, there were
no dramatic decreases in quantum yield upon assembly of
either OPV2 examples; in fact, there was a notable increase in
PL quantum yield for 1 upon assembly. The trends in excited
Figure 6. UV−vis and PL spectra (a, b) and CD spectra (c, d) of
symmetric OPV2 (2: a, c) and nonsymmetric OPV2 (1: b, d) peptides
in basic solution (- - -) and assembled in acidic solution ().
state properties were comparable among the two examples:
The single-exponential photoluminescence lifetimes recorded
for the symmetric and nonsymmetric OPV 2s in basic solution
were 0.31 and 0.23 ns, respectively, while extending to 0.74 and
0.99 ns, respectively, in acidic solution (Table 1). We observed
that, under continual irradiation, basic molecular solutions and
acidic aggregate solutions showed evidence for trans−cis
isomerism, but we were unable to conclusively observe or
otherwise characterize any significant contributions from
intermolecular [2 + 2] stilbene photodimerizations as being
facilitated by close contact within the assembled nanostructures
(Figures S24 and S25).
Circular dichroism was used to interrogate the associations of
the π-conjugated OPV chromophores before and after
nanostructure assembly. Both OPV2 peptides showed no
appreciable low-energy features in CD spectra for the basic
molecularly dissolved samples, but upon lowing the pH to
trigger the assemblies, bisignate spectral features crossing over
around 310 nm were apparent (Figure 6c,d). The induction of
such pronounced Cotton effects indicates that the stilbene
moieties are held in chiral environments and, along with the
excimeric-like emission profiles observed, that the stilbene
subunits are in close contact as suggested by our modeling
studies. Circular dichroism and infrared spectroscopies were
also employed to ascertain information about the hydrogen-
bonding networks. Both measurements indicated β-sheet-like
character: CD revealed high-energy negative peaks at 205−210
nm along with classic IR amide I bands appearing at 1641 and
1632 cm−1 and amide II bands appearing at 1531 and 1537
cm−1 for the nonsymmetric and symmetric OPV2s, respec-
tively. We emphasize that this data, recorded on lyophilized
solid samples, does not allow us to specifically categorize the
nature of the β-sheet character (specifically being parallel or
antiparallel as idealized in Figure 4), nor does it rule out other
more disordered peptide conformations. The intermolecular
electronic interactions evident in the OPV2 signatures indicate
that the π-electron portions are able to interact comparably
11379
dx.doi.org/10.1021/la501999g | Langmuir 2014, 30, 11375−11385

Langmuir
Article
Table 1. Steady-State and Time-Resolved Spectroscopic Data for the Peptide Variants
a
b
c
Abs λmax/nm
PL λmax/nm
QY
τ/ns
peptide sequence
basic
acidic
basic
acidic
basic
acidic
basic
acidic
nsOPV2 (1)
sOPV2 (2)
nsOPV3 (3)
sOPV3 (4)
330
330
367
368
330
334
332
344
451
383
500
447
530
447
490
456
0.02
0.06
0.12
0.38
0.04
0.06
0.06
0.12
0.23 (100)
0.31 (100)
0.68 (100)
1.09 (100)
0.99 (100)
0.74 (100)
6.96 (28); 0.74 (72)
7.53 (100)
a
b
Excitation at the absorption λmax. Recorded on solutions with absorbance below 0.1 au and relative to quinine sulfate in 0.5 N H2SO4 (0.55).
c
Fitted percent contributions for the decay profiles in parentheses.
within the nanostructures regardless of the specific hydrogen-
bonding geometries being adopted within them.
The interpretation that the stilbene moieties undergo
intermolecular electronic interactions is further corroborated
by the excimer formation found in Letsinger’s work with
stilbene dicarboxamides placed into the backbones of
complementary oligonucleotides.28 These studies revealed
only a minor (6 nm blue-shift) ground state perturbation
upon hybridization (thus bringing the stilbene units sufficiently
close for intermolecular electronic interactions) as well as a new
fluorescent emission signature centered at 450 nm. Comparable
results were found in Collard’s work with cofacial stilbenes
linked through a common [2.2]paracyclophane core that
provides a more geometrically well-defined yet pre-existing
ground-state through-space interaction.29 Lewis’ studies of
arenedicarboxamide single crystals also provide insight into the
effects of how torsional angles impact intermolecular
interactions through specific perturbations arising from
increased interplanar distances, lateral displacement, or twist
angle with respect to adjacent monomers.30 The peptide−OPV
hybrids allow us to explore these effects with similar stilbene
cores after the peptides assemble into nanostructures with
comparably well-defined geometrical constraints.
OPV3. The distyrylbenzene chromophore is another well-
defined OPV oligomer that has been subject to extensive
photophysical study, similarly revealing a spectrum of electronic
outcomes that varies as a function of intermolecular orientation.
We previously interrogated a series of symmetric OPV3
peptides with systematic and site-specific variation of amino
acid sequence and demonstrated the accessibility of excitonic
and excimeric excited states.15 In the present work, we
challenge the central OPV3 chromophore with different
peptide N-to-C directionalities of an otherwise constant
amino acid sequence.
Figure 7. UV−vis (a, b), PL (c, d), and CD spectra (e, f) of symmetric
OPV3 (4: a, c, e) and nonsymmetric OPV3 (3: b, d, f) peptides in
basic solution (- - -) and assembled in acidic solution ().
Both OPV3s have similar molecularly dissolved UV−vis
absorption profiles (λmax 367−368 nm) regardless of the
chromophore connectivity (N-acylcarboxamide in 3 vs
dicarboxamide in 4) although the emission λmax values were
separated by ca. 50 nm (Figure 7 and Table 1). The OPV3 PL
lifetimes for the molecularly dissolved solutions were 1.09 and
0.68 ns for the symmetric and nonsymmetric cases,
respectively. Upon assembly, both OPV3 peptides showed
the absorption trends typically associated with H-like
aggregates: the UV−vis profiles showed an overall blue-shift
in the absorption maxima with respect to the molecularly
dissolved molecules (λmax values of 332 and 343 nm for the
nonsymmetric and symmetric OPV3s 3 and 4, respectively)
along with the appearance of weak lower energy shoulders at ca.
425 nm (Figure 7a,b). Unlike the OPV2 examples, the
photoluminescence for both OPV3 aggregates was substantially
quenched with respect to the isolated molecules (Figure 7c,d),
and the lifetimes of the excited states were significantly
increased. This shows that the radiative and nonradiative
rates of excited state decay are fundamentally different for these
particular OPV2 and OPV3 chromophores under the
conditions studied. Interestingly, the PL spectral footprints
for the two OPV3 systems were markedly different. The PL of
the symmetric OPV3 peptide 4 previously was shown to exhibit
an excimeric structure with a red-shifted emission maxima
centered at 515 nm with a single-component lifetime of 7.53 ns.
In contrast, the nonsymmetric OPV3 peptide 3 had a spectral
profile (evolution of vibronic fine structure but with no
dramatic red-shift) and an excited-state lifetime (two
components: 0.74 ns (72%) and 6.96 (28%)) more suggestive
of the excitonic extreme. Polymodal PL signatures are
characteristic of donor−acceptor types of OPV3 chromophores
in exciton-coupled environments.31,32 Although the assembled
nonsymmetric OPV3 does not present clear vibronic features, it
is much more structured than the broad and featureless
11380
dx.doi.org/10.1021/la501999g | Langmuir 2014, 30, 11375−11385

Langmuir
Article
spectrum recorded under molecular dissolution at basic pH.
The absolute differences in PL maxima, as for the OPV2 cases,
can be attributed to the donor−acceptor character present in
the nonsymmetric OPV3 molecule. Although the two OPV3
chromophores are different, the greater contributions of the
long-lived excited state component for symmetric OPV3s were
previously correlated to the excimeric states, while the
assemblies with greater contributions of shorter lived
components were attributed to exciton-coupled states.
simulated the formation of both the parallel (i.e., molecules
possess the same N-terminal to C-terminal directionality;
parallel β-sheet motif) and antiparallel (i.e., peptides possess
opposing N-terminal to C-terminal directionality; antiparallel β-
sheet motif) dimers. In Figure 8, we present our computed
The circular dichroism spectra of the two OPV3 peptides
were similar, with a negative Cotton effect showing a crossover
point at ca. 330 nm. Furthermore, both OPV3 peptides showed
negative signals centered at ca. 210 nm accompanied by IR
signatures characteristic of amide I stretches (1635 and 1640
cm−1 for the symmetric and nonsymmetric, respectively) and
amide II stretches (1535 and 1517/1531 cm−1, respectively). As
was the case for the OPV2 comparison, we are unable to extract
definitive peptide hydrogen-bonding content out of these data.
The split couplets in the low-energy portion of the CD spectra
coupled with the UV−vis and PL data support the presence of
exciton coupling among the longer OPV3 chromophores
regardless of peptide sequence directionality or inherent
conformational preferences.
Molecular Dynamics Simulations. The origin of the
differences in excited state outcomes for the OPV2 peptides
(both excimeric in nature) and the OPV3 peptides (one
excimeric and one excitonic) were not immediately clear after
considering intermolecular associations in energy-minimized
models. We therefore turned to molecular dynamics
simulations which present a means to furnish microscopic
understanding of the interactions and thermodynamic driving
forces driving assembly at a resolution inaccessible by
experimentation. Because the first step in the self-assembly of
nascent 1-D peptide ribbons is the formation of noncovalent
peptide homodimers, we performed atomistic molecular
dynamics simulations to model dimerization for each of the
four DFAG peptides in Figure 1 (rather than model extended
oligomeric stacks) to compute the free energy of dimerization,
identify stabilizing interactions within the dimer, and test our
hypothesis that the nonsymmetric peptides should preferen-
tially adopt an antiparallel, rather than parallel, β-sheet motif.
This dimeric restriction offers a more quantitative local picture
about dimer geometries and energetics that should reasonably
propagate within an extended supramolecular stack but with
less computational demand. Details of our simulation protocols
are provided in the Experimental Section.
Figure 8. PMF profiles for the formation of peptide homodimers as a
function of peptide center of mass distance, Δr, for symmetric and
nonsymmetric (a) OPV2 and (b) OPV3 peptides.
Free Energy of Dimerization and Dimer Structure. To
evaluate the free energy of dimerization, we computed from our
simulations the Gibbs free energycommonly referred to as
the potential of mean force (PMF)as a function of the
distance between the molecular centers of mass, G(Δr) = −kBT
ln P(Δr) + C, where kB is Boltzmann’s constant, T is the
temperature, P(Δr) is the equilibrium probability of observing
the molecules at intermolecular center of mass distance Δr, and
C is an additive constant due to the unknown normalization
factor for the probability distribution (i.e., the partition
function). Because of this additive constant, only differences
in G(Δr) are meaningful. The negative gradient of the PMF
may be identified as an effective driving force along Δr averaged
over the ensemble of system configurations at the correspond-
ing value of Δr,33 and the global minimum of the PMF
corresponds to the most probable intermolecular separation at
equilibrium. For the nonsymmetric peptide variants, we
PMF profiles for dimerization of the OPV2 and OPV3
peptides, and Figure S23 presents representative snapshots of
dimer configurations generated using VMD.34
At large intermolecular distances, interactions between the
molecules are screened by the intervening solvent, and the
molecules are effectively noninteracting, reflected in the plateau
in the PMF profiles at large separations, where the free energy
ceases to be a function of separation. Consistent with the
expectation that dimerization should proceed spontaneously,
the PMF decreases as the intermolecular separation is reduced
until it reaches a global minimum corresponding to the most
probable intermolecular separation of the dimer contact pair.
For all six systems considered, the PMF minimum lies in the
range 0.3−0.7 nm, consistent with the expected model of
aggregate formation in which the molecules mimic the
assembly of natural peptides into β-sheet motifs that possess
interstrand distances of about 0.5 nm.26,35
11381
dx.doi.org/10.1021/la501999g | Langmuir 2014, 30, 11375−11385

Langmuir
Article
The free energy change upon dimerization of the OPV2
peptides is about −2.2 kcal/mol for the symmetric architecture
(2) and −3.7 kcal/mol for the nonsymmetric architecture (1).
The PMF profile for the symmetric dimer possesses a
twist angle between the molecules in the OPV2 dimers
averaged over the configurational ensemble observed in our
simulations is not statistically significantly different from zero,
suggesting that the moleculesat least in the dimershow
little preference to deviate from a near parallel stacking
configuration. In contrast, the OPV3 dimers all possess nonzero
twist angles within the standard deviation of the ensemble,
suggesting an intrinsic preference for a skewed stacking
configuration (cf. Figure S23).
Energetics of Dimerization. Consistent with our molecular
model of self-assembly of peptides into 1-D aggregates
consisting of β-sheet-like motifs, it was our anticipation that
the stability differences between the various homodimers may
be explained by the efficiency of hydrogen bonding within the
dimer. We report in Table 3 the number of intermolecular
hydrogen bonds formed between the peptides in each of the six
homodimers. The similar stability of the parallel and antiparallel
nonsymmetric OPV2 dimers is consistent with the formation in
each case of around two interpeptide hydrogen bonds. The
elevated stability of the antiparallel OPV3 dimer over the
parallel configuration is consistent with the formation of
approximately one extra hydrogen bond in the antiparallel
orientation. Because of the polydisperse nature of the
assembled structures that form, we are unable to definitively
correlate these energetics to specific assembly morphologies
(e.g., those shown in Figure 5). Future work will seek to further
understand the energetics of the longer range assembly
processes operative in these nanomaterials.
minimum at 0.53
0.06 nm, whereas those for the parallel
and antiparallel nonsymmetric homodimers exhibit minima at
intermolecular separations of 0.44 0.06 and 0.31 0.06 nm,
respectively. There is no statistically significant difference in the
dimerization free energy change for the parallel and antiparallel
nonsymmetric dimer pairs. The absence of a strong
thermodynamic orientational preference suggests that multi-
meric 1D aggregates of nonsymmetric OPV2 peptides may
contain a statistical mixture of idealized parallel and antiparallel
β-sheet motifs. Thus, there is a strong statistical overlap in the
intermolecular spacings for both OPV2 molecules that can
serve to rationalize the similarities in excited state excimeric
outcomes regardless of peptide directionality or hydrogen-
bonding networks present within the resulting nanostructures.
For the OPV3 peptides, the free energy change upon
dimerization for the symmetric architecture (4) is about −5.8
kcal/mol, and the intermolecular separation at the PMF
minimum is 0.48
0.06 nm. This separation is statistically
consistent with the spacing found for this peptide as part of a
35-membered stack reported previously (DFAG).15 For the
nonsymmetric OPV3 peptides (3), the free energy change for
formation of the antiparallel homodimer is about −3.7 kcal/
mol, compared to only −0.6 kcal/mol for the parallel
orientation. The intermolecular separation in the antiparallel
homodimer is 0.36 0.06 nm, compared to 0.67 0.06 nm for
the less stable parallel orientation. This suggests that 1D
aggregates of nonsymmetric OPV3 peptides possess an intrinsic
bias toward the more stable antiparallel peptide orientations
that resemble antiparallel β-sheet motifs. The essentially
complete lack of intermolecular stacking overlap among the
likely configurations of the two OPV3 configurations could help
to rationalize the different excited state outcomes observed.
Consistent with our prior sequence variation analysis,15
aggregates with smaller intermolecular spacings tended toward
excitonic states whereas larger spacings coincided with
excimeric states.
The relative stabilities of the symmetric and nonsymmetric
dimers cannot be rationalized by hydrogen bonding alone, so to
more quantitatively probe the energetics of dimerization, we
decomposed the total system interaction energy to compute:
(i) ΔUpeptide as the portion of the energy change resulting from
bonded interactions (bonds, angles, dihedrals) and intra-
molecular nonbonded contributions (van der Waals, Coulomb)
vdW
within both peptides upon dimerization and (ii) ΔU
peptide−peptide
Coul
and ΔU
as the changes in the intermolecular van der
peptide−peptide
Waals and Coulomb interactions between the molecules upon
formation of the dimer. We summarize these energy changes
upon dimerization in Table 3. Negative terms correspond to
energetic contributions that favor dimerization. For all systems
studied, the change in intramolecular peptide energies upon
dimerization cannot be distinguished from zero with the
calculated standard deviations. This is consistent with our
observation that peptide conformations change very little over
the course of dimerization.
We summarize in Table 2 the intermolecular separation,
Δrdimer, and relative twist angle between the peptides, Δθdimer
,
in the dimer and the free energy change for dimerization,
ΔGdimerization, for each of the six configurations. The relative
Table 2. Free Energy Change, Intermolecular Separation,
and Relative Twist Angle between Molecules in the Dimer
The aggregated van der Waals and Coulombic contributions
to the peptide−peptide dimerization energy change,
b
d
vdW
peptide−peptide
Coul
peptide−peptide
ΔGdimerization
Δθdimer
ΔUpeptide−peptide = ΔU
+ ΔU
, is signifi-
a
c
dimer
orientation
(kcal/mol)
Δrdimer (nm)
(deg)
cantly smaller for the formation of the parallel nonsymmetric
OPV3 dimer (−34 kcal/mol) relative to both the antiparallel
nonsymmetric (−46 kcal/mol) and symmetric (−45 kcal/mol)
dimers. The more favorable Coulombic interactions in the
antiparallel nonsymmetric dimer relative to the parallel
orientation is consistent with a strong planar hydrogen bonding
pattern in the idealized antiparallel β-sheet motif, compared to
the weaker nonplanar pattern in the parallel arrangement. The
values of ΔUpeptide−peptide for the antiparallel nonsymmetric and
symmetric dimers are very similar, indicating that the ∼2 kcal/
mol higher stability of the symmetric dimer cannot be explained
by intermolecular interactions alone (cf. Table 2). In the case of
OPV2, both the van der Waals and Coulombic contributions to
ΔUpeptide−peptide are substantially larger for the antiparallel
s OPV2 (2)
ns OPV2 (1)
ns OPV2 (1)
s OPV3 (4)
ns OPV3 (3)
ns OPV3 (3)
−2.2 0.4
−3.6 0.4
−3.8 0.4
−5.8 0.4
−0.6 0.5
−3.7 0.4
0.53 0.06
0.44 0.06
0.31 0.06
0.48 0.06
0.67 0.06
0.36 0.06
4
6
14
15
8
parallel
antiparallel
12
8
6
parallel
11
22
6
antiparallel
9
a
b
s = symmetric, ns = nonsymmetric. Uncertainties computed by 50
rounds of bootstrap resampling and propagation of uncertainties.
c
d
Uncertainties from bin resolution of calculated PMF curve. Standard
deviation computed over all umbrella sampling configurations
0.06 nm. Since the peptides in the
satisfying Δr = Δrdimer
homodimer are identical, we have adopted a convention in which the
relative angle is taken to be non-negative.
11382
dx.doi.org/10.1021/la501999g | Langmuir 2014, 30, 11375−11385

Langmuir
Article
Table 3. Hydrogen Bonds Formed and Decomposition of the Energy Change upon Dimerization
a
b
b
b
b
b
vdW
peptide−peptide
Coul
peptide−peptide
dimer
orientation
no. H-bonds
ΔUpeptide (kcal/mol) ΔU
(kcal/mol) ΔU
(kcal/mol) ΔUpeptide−peptide (kcal/mol)
s OPV2
2.0 1.2
2.0 1.3
1.8 1.0
2.3 0.7
0.9 0.9
1
29
−23
−26
−30
−32
−27
−31
4
4
−14
−13
−17
−13
−7
8
6
5
4
5
−37
−39
−47
−45
−34
9
7
7
5
6
ns OPV2
ns OPV2
s OPV3
parallel
−22 34
−2 31
antiparallel
5
3
29
10 33
39
3
ns OPV3
ns OPV3
parallel
4
antiparallel
2.3 1.7
3
5
−15 12
−46 13
a
b
s = symmetric, ns = nonsymmetric. Standard deviations computed over all umbrella sampling configurations satisfying Δr = Δrdimer 0.06 Å and
propagation of uncertainties.
nonsymmetric OPV2 dimer relative to the symmetric
orientation, again consistent with the elevated stability of the
antiparallel β-sheet motif. Nevertheless, the free energy
calculations (cf. Table 2) reveal the antiparallel dimer to be
only ∼0.2 kcal/mol more stable than the parallel, despite
possessing an approximately 8 kcal/mol more favorable
peptide−peptide interaction energy.
dichotomy in excited state outcome that depends on the
nature of the hydrogen-bonding networks. This can be
attributed in part to magnified differences in intermolecular
spacing for the OPV3 peptides relative to the OPV2s.
Molecular dynamics simulations for the formation of peptide
homodimers as the first stage of peptide self-assembly reveal
free energy changes driving dimerization of 2−6 kcal/mol.
Intermolecular interactions between nonsymmetric OPV3
peptide pairs favor the formation of antiparallel dimers in a
motif akin to an antiparallel β-sheet, contributing to the ∼3
kcal/mol enhanced stability of the antiparallel relative to the
parallel orientation. In the case of the nonsymmetric OPV2
peptides, the antiparallel and parallel dimers are equally stable
despite a ∼8 kcal/mol larger intermolecular interaction energy
in the antiparallel orientation. These findings suggest an
important role for solvent-mediated interactions and entropic
effects in determining dimer stability. The geometric relation-
ships within the dimer pairs provide a useful starting point for
electronic structure calculations of the expected photophysical
outcomes that will be the subject of future investigations.
Potential results that come as a result of the hydrogen-bonding
motif could help shed information about naturally occurring
peptide aggregation as well as potential supramolecular
materials, and differences that arise from this and the differing
π-conjugated subunit can aid in understanding interchain
interactions of extended π-conjugated systems that may more
useful for exploring the bioelectronics potential of these
materials.
The results of our energy analysis support our hypothesis
that the antiparallel arrangement of nonsymmetric dimers
should be energetically favored over the parallel arrangement
due to more efficient intermolecular interactions. In all cases
the van der Waals contribution to the peptide−peptide
interaction outweighs the Coulomb, with dispersion inter-
actions responsible for 62−79% of the total interaction energy,
indicating an important role for nonspecific dispersion
interactions in addition to specific hydrogen bonds. The free
energy of dimerization comprises enthalpic and entropic
contributions, ΔGdimerization = ΔH − TΔS, which may be
further decomposed as ΔGdimerization = ΔUpeptide
ΔUpeptide−peptide + ΔUpeptide−solvent + ΔUsolvent + pΔV − TΔS.
We have quantified in the present study ΔGdimerization, ΔUpeptide
+
,
and ΔUpeptide−peptide. That the observed trends in ΔGdimerization
cannot be explained by the trends in ΔUpeptide and
ΔUpeptide−peptide alone indicates that the peptide−solvent
interaction, solvent energetics, and volumetric and entropic
effects also make important contributions to the free energy of
dimerization. While in principle this decomposition might be
experimentally accessible from calorimetry,36 these experiments
would likely be exceedingly difficult to conduct in practice
because controlling the formation of dimers without forming
higher order aggregates would be extremely challenging. We are
currently undertaking additional simulation studies to quantify
the entropic and solvent-mediated enthalpic components of the
calculated free energy of dimerization using the methodology
detailed elsewhere.37 Such calculations will furnish insight into
the relative enthalpic and entropic contributions to the
dimerization process and help guide the design of peptides
engineered to stabilize and control aggregation.
EXPERIMENTAL SECTION
■
Specific details about the synthesis of π-conjugated amino acids,
diacids, and peptide target molecules along with relevant character-
ization data for all synthetic intermediates and final peptide molecules
may be found in the Supporting Information.
General Synthesis of Peptides. All peptides were synthesized
using standard solid phase 9-fluorenylmethoxycarbonyl (Fmoc)
chemistry on a Wang resin preloaded (0.8 mmol/g) with the Fmoc-
protected Asp(OtBu). Fmoc deprotection was performed by mixing
the resin in a piperidine/DMF (2:8) solution for 10 min (2×),
followed by rinsing with DMF, MeOH, and CH2Cl2. For all standard
amino acid couplings, 3.0 equiv (relative to the resin substitution) of
Fmoc protected amino acid was activated externally with 2.9 equiv of
HBTU and 10 equiv of diisopropylethylamine (DIPEA) dissolved in
NMP. The activated Fmoc-protected amino acid was then added to a
peptide chamber containing the deprotected Wang resin and mixed for
3 h. The resin was then drained, rinsed with NMP, MeOH, and
CH2Cl2, and then allowed to dry. All coupling and deprotection steps
were monitored by performing a Kaiser test.
CONCLUSIONS
■
We have reported a series of OPV-containing oligopeptides
possessing similar primary amino acid sequences but differing
in the presentation of overall peptide directionality. We
explored the photophysical properties of their self-assembled
nanostructures to probe how hydrogen-bonding networks
would perturb the aggregate structure. Although the hydro-
gen-bonding motifs differ in their relative stability, the ability
for each of these peptide constructs to self-assemble does not
appear to be dramatically influenced. The OPV2 peptides seem
to show nearly quantitative excimer emission regardless of
peptide directionality, while the OPV3 peptides show a
Installation of π-Conjugated “Amino Acids” (for Non-
symmetric Peptides 1 and 3). The Fmoc-protected OPV amino
acids were coupled to immobilized and deprotected DFAG
tetrapeptide sequences prepared as above but were activated with
HATU rather then HBTU and a solution of 2:1 NMP/DCM was used
11383
dx.doi.org/10.1021/la501999g | Langmuir 2014, 30, 11375−11385

Langmuir
Article
for the couplings. After successful coupling and Fmoc deprotection as
described above, the next Fmoc-protected α-amino acid residue
(glycine) was activated with triphosgene and 2,4,6-trimethylpyridine
using a published procedure.6 At this point, standard SPPS was carried
out as described above to install the remainder of the peptide sequence
(A, F, and D amino acids).
free energy profile computed by solving the WHAM equations.42 Full
details of our simulation protocols are provided in the Supporting
Information.
ASSOCIATED CONTENT
* Supporting Information
■
S
Installation of π-Conjugated Diacids (for Symmetric Pep-
tides 2 and 4). The general on-resin dimerization procedure of
Vadehra et al. was followed for OPV2 diacid installation,11 while the
symmetric OPV3 was prepared similarly and as reported previously.
4,4′-((1E,1′E)-1,4-Phenylenebis(ethene-2,1-diyl))dibenzoic acid (0.3
equiv) and PyBOP (0.6 equiv) were dissolved in 2:1 NMP:CH2Cl2.
DIPEA (7.0 equiv) was added and mixed for approximately 1 min.
This solution was then added to the deprotected Wang resin with
immobilized DFAG tetrapeptide (1.0 equiv) in a peptide chamber, and
the reaction was mixed for 18 h. The chamber was drained, and the
resin was washed with CH2Cl2, MeOH, and NMP. A second round of
coupling was then performed: 4,4′-((1E,1′E)-1,4-phenylenebis-
(ethene-2,1-diyl))dibenzoic acid (0.2 equiv) and PyBOP (0.4 equiv)
were dissolved in 2:1 NMP:CH2Cl2. DIPEA (7.0 equiv) was added
and mixed for approximately 1 min. The solution was then added to
deprotected Wang resin (1.0 equiv) in a peptide chamber, and the
reaction was mixed for 18 h. The chamber was drained, and the resin
was washed with CH2Cl2, MeOH, and NMP.
General Procedure for Peptide Cleavage from the Resin. A
2:1 solution of 9.5:0.25:0.25 trifluoroacetic acid/H2O/triisopropylsi-
lane and CH2Cl2 were added to the peptide chamber and mixed with
the resin for 3 h. The resin was removed by filtration and washed 3×
with DCM, and the filtrate was then concentrated via solvent
evaporation under reduced pressure. Cold Et2O was added to the
solution, and the resulting solid material was collected via
centrifugation. After decanting the Et2O, the crude material was
dissolved in water, and ammonium hydroxide was added until the
material was completely dissolved. This solution was lyophilized, and
the obtained crude dry material was purified via reverse phase high
performance liquid chromatography (RP-HPLC).
UV−vis and Photoluminescence. UV−vis spectra were collected
using a Varian Cary 50 Bio UV−vis spectrophotometer. PL data were
collected using a PTI Photon Technology International QuantaMaster
Fluorometer with an Ushio xenon short arc lamp. Lifetime
measurements were collected on the QuantaMaster fluorometer with
a PTI TimeMaster LED TM-200 strobe. PTI FeliX32 V.1.2 (Build 56)
was used for data processing. Micromolar samples were prepared using
Millipore water, and the pH was adjusted by the addition of 1000 mL
(for 1 and 2) or 10 μL (for 3 and 4) of 1 M HCl or 1 M KOH.
Circular Dichroism. CD spectra were acquired using a Jasco J-810
spectropolarimeter. Acidic and basic samples were prepared by adding
10 μL of 1 M HCl or 1 M KOH, respectively, to micromolar stock
solution of the peptide in Millipore water.
General synthesis details, molecular characterization data, full
details of the molecular simulations, and Figures S1−S25. This
material is available free of charge via the Internet at http://
pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
*E-mail: tovar@jhu.edu (J.D.T.).
■
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We thank Johns Hopkins University and the U.S. Department
of Energy Office of Basic Energy Sciences (DE-SC0004857,
J.D.T. and A.L.F. peptide nanomaterial synthesis and
simulation) for financial support. We thank the Fairbrother
group at JHU for assistance with ATR-IR measurements.
B.D.W. was supported by the Harry and Cleio Greer
Fellowship (JHU).
REFERENCES
■
(1) Caplan, M. R.; Lauffenburger, D. A. Nature’s complex
copolymers: Engineering design of oligopeptide materials. Ind. Eng.
Chem. Res. 2002, 41, 403−412.
(2) Holmes, T. C. Novel peptide-based biomaterial scaffolds for
tissue engineering. Trends Biotechnol. 2002, 20, 16−21.
(3) Ulijn, R. V.; Smith, A. M. Designing peptide based nanomaterials.
Chem. Soc. Rev. 2008, 37, 664−675.
(4) Zelzer, M.; Ulijn, R. V. Next-generation peptide nanomaterials:
molecular networks, interfaces and supramolecular functionality.
Chem. Soc. Rev. 2010, 39, 3351−3357.
(5) Engler, A. J.; Sen, S.; Sweeney, H. L.; Discher, D. E. Matrix
elasticity directs stem cell lineage specification. Cell 2006, 126, 677−
689.
(6) Jatsch, A.; Schillinger, E. K.; Schmid, S.; Bauerle, P. Biomolecule
̈
assisted self-assembly of π-conjugated oligomers. J. Mater. Chem. 2010,
20, 3563−3578.
(7) Gonzalez-Rodriguez, D.; Schenning, A. P. H. J. Hydrogen-
bonded supramolecular π-functional materials. Chem. Mater. 2011, 23,
310−325.
Attenuated Total Reflection Infrared Spectroscopy. Data
were obtained on dry lyophilized peptides using a Thermo Nicolet
NEXUS 670 FTIR.
(8) Kim, S. H.; Parquette, J. R. A model for the controlled assembly
of semiconductor peptides. Nanoscale 2012, 4, 6940−6947.
(9) Marty, R.; Szilluweit, R.; Sanchez-Ferrer, A.; Bolisetty, S.;
Adamcik, J.; Mezzenga, R.; Spitzner, E.-C.; Feifer, M.; Steinmann, S.
N.; Corminboeuf, C.; Frauenrath, H. Hierarchically structured
microfibers of “single stack” perylene bisimide and quaterthiophene
nanowires. ACS Nano 2013, 7, 8498−8508.
Transmission Electron Microscopy. TEM images were acquired
on a Philips EM 420 transmission electron microscope equipped with
an SIS Megaview III CCD digital camera. 300 mesh Formvar carbon
coated copper grids were purchased from Electron Microscopy
Sciences. Grids were prepared as follows: A stock solution of a 0.1
mg/mL peptide in Millipore water was exposed to concentrated HCl
vapor for 1 min. 10 μL of this solution was pipetted onto a grid. The
grid was incubated for 8 min. The grid was then dipped sequentially
into water and then into a solution of 2% uranyl acetate stain and
allowed to dry in air. ImageJ 1.47 (National Institutes of Health,
Bethesda, MD) was used to approximate the widths and associated
standard deviations of the nanostructures (n = 50 per TEM image).
Molecular Dynamics Simulations. Molecular dynamics simu-
lations were conducted using the GROMACS 4.6 simulation suite.38
Peptides were modeled using the CHARM27 force field,39 and water
was modeled explicitly using the TIP3P model.40 Umbrella sampling41
was employed to simulate dimerization by restraining the
intermolecular distance between the peptide centers of mass and the
(10) Tovar, J. D. Supramolecular construction of optoelectronic
biomaterials. Acc. Chem. Res. 2013, 46, 1527−1537.
(11) Diegelmann, S. R.; Gorham, J. M.; Tovar, J. D. One-dimensional
optoelectronic nanostructures derived from the aqueous self-assembly
of π-conjugated oligopeptides. J. Am. Chem. Soc. 2008, 130, 13840−
13841.
(12) Schillinger, E.-K.; Mena-Osteritz, E.; Hentschel, J.; Borner, H.
̈
G.; Bauerle, P. Oligothiophene versus β-sheet peptide: Synthesis and
̈
self-assembly of an organic semiconductor-peptide hybrid. Adv. Mater.
2009, 21, 1562−1567.
(13) Stone, D. A.; Hsu, L.; Stupp, S. I. Self-assembling
quinquethiophene-oligopeptide hydrogelators. Soft Matter 2009, 5,
1990−1993.
11384
dx.doi.org/10.1021/la501999g | Langmuir 2014, 30, 11375−11385

Langmuir
Article
(14) Mba, M.; Moretto, A.; Armelao, L.; Crisma, M.; Toniolo, C.;
Maggini, M. Synthesis and self-assembly of oligo(p-phenyleneviny-
lene) peptide conjugates in water. Chem.Eur. J. 2011, 17, 2044−
2047.
(15) Wall, B. D.; Zacca, A. E.; Sanders, A. M.; Wilson, W. L.;
Ferguson, A. L.; Tovar, J. D. Supramolecular polymorphism: Tunable
electronic interactions within pi-conjugated peptide nanostructures
dictated by primary amino acid sequence. Langmuir 2014, 30, 5946−
5956.
(16) Vadehra, G. S.; Wall, B. D.; Diegelmann, S. R.; Tovar, J. D. On-
resin dimerization incorporates a diverse array of π-conjugated
functionality within aqueous self-assembling peptide backbones.
Chem. Commun. 2010, 46, 3947−3949.
(17) Anthony, N. G.; Breerl, D.; Clarke, J.; Donoghue, G.;
Drummond, A. J.; Ellis, E. M.; Gemmell, C. G.; Helesbeux, J.-J.;
Hunter, I. S.; Khalaft, A. I.; Mackay, S. P.; Parkinson, J. A.; Suckling, C.
J.; Waiah, R. D. Antimicrobial lexitropsins containing amide, amidine,
and alkene linking groups. J. Med. Chem. 2007, 50, 6116−6125.
(18) Durantini, E. N. Synthesis of functional olefins using the Wittig-
Horner reaction in different media. Synth. Commun. 1999, 29, 4201−
4222.
(19) Falb, E.; Yechezkel, T.; Salitra, Y.; Gilon, C. In situ generation of
Fmoc-amino acid chlorides using bis(trichloromethyl) carbonate and
its utilization for difficult couplings in solid-phase peptide synthesis. J.
Pept. Res. 1999, 53, 507−517.
(20) Eastoe, J.; Dominguez, M. S.; Wyatt, P.; Beeby, A.; Heenan, R.
K. Properties of a stilbene-containing gemini photosurfactant: Light-
triggered changes in surface tension and aggregation. Langmuir 2002,
18, 7837−7844.
(21) Gourianov, N.; Kluger, R. Conjoined hemoglobins. Loss of
cooperativity and protein-protein lnteractions. Biochemistry 2005, 44,
14989−14999.
(22) Sancho-Garcia, J. C.; Bredas, J. L.; Beljonne, D.; Cornil, J.;
Martinez-Alvarez, R.; Hanack, M.; Poulsen, L.; Gierschner, J.; Mack,
H. G.; Egelhaaf, H. J.; Oelkrug, D. Design of π-conjugated organic
materials for one-dimensional energy transport in nanochannels. J.
Phys. Chem. B 2005, 109, 4872−4880.
(32) Hoeben, F. J. M.; Herz, L. M.; Daniel, C.; Jonkheijm, P.;
Schenning, A. P. H. J.; Silva, C.; Meskers, S. C. J.; Beljonne, D.;
Phillips, R. T.; Friend, R. H.; Meijer, E. W. Efficient energy transfer in
mixed columnar stacks of hydrogen-bonded oligo(p-phenylene
vinylene)s in solution. Angew. Chem., Int. Ed. 2004, 43, 1976−1979.
(33) Yu, T.; Lee, O.-S.; Schatz, G. C. Steered molecular dynamics
studies of the potential of mean force for peptide amphiphile self-
assembly into cylindrical nanofibers. J. Phys. Chem. A 2013, 117,
7453−7460.
(34) Humphrey, W.; Dalke, A.; Schulten, K. VMD: Visual molecular
dynamics. J. Mol. Graphics 1996, 14, 33−38.
(35) Hughes, A. B. Amino Acids, Peptides and Proteins in Organic
Chemistry, Protection Reactions, Medicinal Chemistry, Combinatorial
Synthesis; Wiley-VCH: New York, 2011.
(36) Leavitt, S.; Freire, E. Direct measurement of protein binding
energetics by isothermal titration calorimetry. Curr. Opin. Struct. Biol.
2001, 11, 560−566.
(37) Choudhury, N.; Pettitt, B. M. Enthalpy-entropy contributions to
the potential of mean force of nanoscopic hydrophobic solutes. J. Phys.
Chem. B 2006, 110, 8459−8463.
(38) Hess, B.; Kutzner, C.; van der Spoel, D.; Lindahl, E. GROMACS
4: Algorithms for highly efficient, load-balanced, and scalable
molecular simulation. J. Chem. Theory Comput. 2008, 4, 435−447.
(39) MacKerell, A. D.; Bashford, D.; Bellott, M.; Dunbrack, R. L.;
Evanseck, J. D.; Field, M. J.; Fischer, S.; Gao, J.; Guo, H.; Ha, S.;
Joseph-McCarthy, D.; Kuchnir, L.; Kuczera, K.; Lau, F. T. K.; Mattos,
C.; Michnick, S.; Ngo, T.; Nguyen, D. T.; Prodhom, B.; Reiher, W. E.;
Roux, B.; Schlenkrich, M.; Smith, J. C.; Stote, R.; Straub, J.; Watanabe,
M.; Wiorkiewicz-Kuczera, J.; Yin, D.; Karplus, M. All-atom empirical
potential for molecular modeling and dynamics studies of proteins. J.
Phys. Chem. B 1998, 102, 3586−3616.
(40) Jorgensen, W. L.; Chandrasekhar, J.; Madura, J. D.; Impey, R.
W.; Klein, M. L. Comparison of simple potential functions for
simulating liquid water. J. Chem. Phys. 1983, 79, 926−935.
(41) Torrie, G. M.; Valleau, J. P. Non-physical sampling distributions
in Monte Carlo free-energy estimation - umbrella sampling. J. Comput.
Phys. 1977, 23, 187−199.
(42) Roux, B. The calculation of the potential of mean force using
computer-simulations. Comput. Phys. Commun. 1995, 91, 275−282.
(23) Davies, R. P. W.; Aggeli, A.; Beevers, A. J.; Boden, N.; Carrick, L.
M.; Fishwick, C. W. G.; McLeish, T. C. B.; Nyrkova, I.; Semenov, A.
N. Self-assembling β-sheet tape forming peptides. Supramol. Chem.
2006, 18, 435−443.
(24) Diegelmann, S. R.; Hartman, N.; Markovic, N.; Tovar, J. D.
Synthesis and alignment of discrete polydiacetylene-peptide nano-
structures. J. Am. Chem. Soc. 2012, 134, 2028−2031.
(25) Whitten, D. G. Photochemistry and photophysics of trans-
stilbene and related alkenes in surfactant assemblies. Acc. Chem. Res.
1993, 26, 502−509.
(26) Lewis, F. D.; Yang, J. S.; Stern, C. L. Ground and excited-state
aromatic-aromatic interactions with distance control by hydrogen-
bonding. J. Am. Chem. Soc. 1996, 118, 2772−2773.
́ ́
(27) Catalan, J.; Zimanyi, L.; Saltiel, J. Medium-controlled
aggregation of trans-stilbene. J. Am. Chem. Soc. 2000, 122, 2377−2378.
(28) Letsinger, R. L.; Wu, T. Control of excimer emission and
photochemistry of stilbene units by oligonucleotide hybridization. J.
Am. Chem. Soc. 1994, 116, 811−812.
́
(29) Mukhopadhyay, S.; Jagtap, S. P.; Coropceanu, V.; Bredas, J. L.;
Collard, D. M. π-Stacked oligo(phenylene vinylene)s based on
pseudo-geminal substituted [2.2]paracyclophanes: Impact of inter-
chain geometry and interactions on the electronic properties. Angew.
Chem., Int. Ed. 2012, 51, 11629−11632.
(30) Lewis, F. D.; Yang, J. S.; Stern, C. L. Crystal structures of
secondary arenedicarboxamides. An investigation of arene-hydrogen
bonding relationships in the solid state. J. Am. Chem. Soc. 1996, 118,
12029−12037.
(31) Bartholomew, G. P.; Bazan, G. C. Synthesis, characterization,
and spectroscopy of 4,7,12,15-[2.2]paracyclophane containing donor
and acceptor groups: Impact of substitution patterns on through-space
charge transfer. J. Am. Chem. Soc. 2002, 124, 5183−5196.
11385
dx.doi.org/10.1021/la501999g | Langmuir 2014, 30, 11375−11385