controlling supramolecular chirality in peptide-π-peptide
TRANSCRIPT
1
Controlling supramolecular chirality in peptide-π-peptide networks by
variation of alkyl spacer length
Sayak Subhra Panda,†,‡ Kirill Shmilovich¶, Andrew L. Ferguson*,¶, and John D. Tovar*,†,‡,┴
† Department of Chemistry, Johns Hopkins University, 3400 North Charles Street, Baltimore,
Maryland 21218, United States.
‡ Institute of NanoBioTechnology, Johns Hopkins University, 3400 North Charles Street,
Baltimore, Maryland, 21218, United States.
┴ Department of Materials Science and Engineering, Johns Hopkins University, 3400 North
Charles Street, Baltimore, Maryland 21218, United States.
¶ Pritzker School of Molecular Engineering, University of Chicago, 5640 South Ellis Avenue,
Chicago, Illinois, 60637, United States.
*Corresponding author emails: [email protected], [email protected]
2
Abstract:
Self-assembled supramolecular organic materials with π-functionalities are of great interest due to
their applications as biocompatible nanoelectronics. Detailed understanding of molecular
parameters to modulate the formation of hierarchical structures can inform design principles for
materials with engineered optical and electronic properties. In this work, we combine molecular-
level characterization techniques with all-atom molecular simulations to investigate the subtle
relationship between the chemical structure of peptide-π-peptide molecules and the emergent
supramolecular chirality of their spontaneously self-assembled nanoaggregates. We demonstrate
through circular dichroism measurements that we can modulate the chirality by incorporating alkyl
spacers of various lengths in between the peptides and thienylene-phenylene π-system
chromophores: even numbers of alkyl carbons in the spacer units (0, 2) induce M-type helical
character whereas odd numbers (1, 3) induce P-type. Corroborating molecular dynamics
simulations and explicating machine learning analysis techniques identify hydrogen-bonding and
hydrophobic packing to be the principal discriminants of the observed chirality switches. Our
results present a molecular-level design rule to engineer chirality into optically and electronically
active nanoaggregates of these peptidic building blocks by exploiting systematic variations in alkyl
spacer length.
3
Introduction:
The self-assembly of small molecules into well-defined supramolecular aggregates has fascinated
synthetic chemists over the past few decades due to the immense potential for control of
bionanomaterials at the molecular scale.1-7 In this regard, peptide-based supramolecular aggregates
received significant attention due to their ability to form well-defined hierarchical structures based
on collective intermolecular interactions (hydrogen-bonding, electrostatics, van der Waals forces
etc.).3,8-11 Peptides have a rich structural diversity with multifarious secondary structures such as
α-helices, β-sheets and β-turns giving rise to diverse nanostructures that are difficult to achieve
through synthetic chemistry or lithographic techniques. Peptides functionalized with π-electron
conjugates can take advantage of this programmed assembly to form biologically-inspired π-
stacked conduits relevant for electronic applications such as organic solar cells, field-effect
transistors and chemical sensors.12-23 In general, a delicate balance of π-π interaction and other
non-covalent interactions determines the charge percolation pathways through the networks. For
all but the simplest molecular building blocks, a complete understanding of bottom-up principles
by which individual molecules may be designed to produce supramolecular aggregates with
engineered structure and function remains lacking, precluding the rational design of molecular
aggregation and associated electron couplings.
Chirality is a pervasive subject in chemical science, from the molecular stereocenters
whose inversions can impact drug efficacy to the polymeric tacticity considerations that influence
chain crystallization. Supramolecularly derived materials also express chirality as mediated
through the presence of molecular stereocenters or through regular patterns of intermolecular
associations that impart a local chiral environment (such as helicity).24-26 In the case of
chromophore-bearing molecules, the local chirality can influence chiroptical properties as well.
4
Although the control of supramolecular chirality has implications in applications such as optics
(spin-selected transport), catalysis (enhancement of enantioselectivity of asymmetric reactions),
biomedicine (increased cell proliferation of due to stereospecific interaction) and purification
(separation of enantiomers),27-34 the specific supramolecular building blocks with respect to
electronic function tend to be designed in isolation based on empirical guidance due to the
difficulty in predicting the chiroptical properties of supramolecular architectures via molecular
dynamics simulations.35-40 Henze et al. observed odd-even effects of chiral aggregation using CD
spectroscopy in a series of oligothiophene molecules substituted with oligo(ethylene oxide) chains
going from quinqueothiophene to septithiophene.35 Chirality driven self-sorting processes were
also investigated in naphthalenedimide and perylenediimide based systems using CD
spectroscopy.41-43 Chiral dopants have also been shown to influence helicity from achiral starting
materials.44, 45 Many groups investigated the formation of one-dimensional nanostructures from
chromophores bearing oligopeptides,46-57 which present obvious points of chirality arising from
the constituent amino acids.58-69 Indeed, supramolecular chirality of oligothiophene networks
could also be tuned with the change in chirality of a single amino acid attached to it.59 Chiral self-
assembly of amyloid-like oligopeptides was investigated by Marty et al. demonstrating a ‘two-
fold’ odd-even effect as observed in circular dichroism (CD) spectroscopy when oligopeptide-
polymer conjugates coupled to perylenediimide chromophores varied with the number of L-
alanine units and the spacer length between the chromophore and the oligopeptide.58 Despite these
advances over the past decade, a detailed understanding and predictive control of chirality within
peptide-π-peptide nanostructures has yet to be established.
In this report, we describe an expansion of our established synthesis methodology to
incorporate π-systems into peptidic segments where aliphatic spacers are placed between the
5
peptidic segments and the conjugated π-electron regions in order to decouple the electronic effect
of the attaching electron-withdrawing carboxamide peptidic segments to the central π-conjugated
chromophores. Specifically, Pd-mediated cross-coupling reactions were used to install the
necessary π-electron functionality within peptide conjugates by promoting aryl-aryl bond
formation.50, 70, 71 In principle, the carboxamide groups directly attached to the π-system should
decrease the efficacy of hole transport by destabilizing the formal cationic structure formed during
field-effect gating. We envisioned that introducing an “insulating” spacer would better facilitate
semiconductive properties. As representative π-core examples, we included different thienylene-
phenylene mixed oligomers and a thienoacene linkage as a “high performance” organic
semiconductor72 that could facilitate aggregation due to stronger π-π stacking and eventually
facilitate energy transport.
We present a comprehensive investigation into how alkyl spacers between π-cores and
flanking peptides influence the internal arrangement of the π-stacks within self-assembled
nanostructures. Incorporation of the spacer groups imparts different conformational preferences
around the linkages thus different molecular geometry starting points from which to form
supramolecular aggregates with different electronic couplings. In general, the dependence of
carrier transport properties as a function of the intermolecular geometric interaction of
wavefunctions of molecular organic semiconductors can be treated computationally,73, 74 but
waging predictive control is difficult to realize experimentally. The peptide self-assembly
processes were monitored using standard spectroscopic techniques (UV-Vis and PL spectroscopy)
and the resulting assemblies were visualized using electron microscopy. We show a remarkable
serendipity of systematically tuning the supramolecular chirality as observed by CD based on the
peptide-chromophore spacer length, regardless of chromophore composition. This shows how a
6
subtle modification in the molecular structure imparts a substantial change in the supramolecular
aggregates formed from these otherwise well-understood systems. Accompanying molecular
dynamics simulations precisely recapitulate the experimentally-observed trends in chirality, and
interrogative "white box" machine learning exposes the particular atoms and inter-molecular
interactions most discriminative of these trends.
Experimental:
General information
THF was acquired from an Innovative Technologies Pure Solv solvent purification system and
dried over 4Å molecular sieves. DMF and triethylamine were purchased from Sigma-Aldrich and
dried over 4Å molecular sieves. N-methyl-2-pyrrolidone (NMP) was obtained from Advanced
ChemTech and dried over 4Å molecular sieves. DCM and n-hexane were freshly distilled and
dried over 4Å molecular sieves. All solvents were degassed by sparging with nitrogen gas at least
30 min prior to use. O-(Benzotriazol-1-yl)-N,N,N′,N′-tetramethyluronium hexafluorophosphate
(HBTU), trimethylsilylacetylene, 3-(4-bromophenyl)propionic acid, 4-bromophenylacetic acid, 4-
(4-bromophenyl)butanoic acid, trimethylsilylacetylene and phenylboronic acid were purchased
from Oakwood Products Inc. Tetrakis(triphenylphosphine)palladium and trans-
dichlorobis(triphenylphosphine)palladium(II) were obtained from Strem Chemicals. 1,4-Diiodo-
2,5-dibromobenzene was purchased from TCI America. Wang resin (pre-loaded with amino acid)
and Fmoc-protected amino acids were obtained from Advanced Chem Tech. Ethanol was
purchased from Pharmco-AAPER. 5-Bromo-2-thiophenecarboxylic acid was obtained from
Accela ChemBio Co. Ltd. 4-iodobenzoic acid was purchased from Alfa Aesar. ((2,5-dibromo-1,4-
phenylene)bis(ethyne-2,1-diyl))bis(trimethylsilane),75 Benzo[1,2-b:4,5-b’]dithiophene76 and 2,6-
bis(trimethylstannyl)benzo[1,2-b:4,5-b’]dithiophene76 were prepared following the literature
protocol and 1H NMR matched literature data. Biotech grade cellulose ester dialysis tubing
7
(MWCO 500-1000) was purchased from Spectrum Labs. All other reagents and starting materials
were obtained from Sigma-Aldrich and were used as received.
NMR Spectroscopy: 1H-NMR spectra were obtained using a Bruker Advance 400 MHz FT-NMR
spectrometer and processed with Bruker Topspin 1.3. Peptide 1H NMR spectra were acquired
using a presaturation pulse to suppress water. Chemical shifts are reported in parts per million
relative to the residual protio solvent [HOD δ: 4.79, CHCl3 δ: 7.26].
Electrospray Ionization Mass Spectrometry (ESI-MS): ESI samples were collected using a
Thermo Finnigan LCQ Deca Ion Trap Mass Spectrometer in negative mode. Samples were
prepared in a 1:1 MeOH:water solution with 1% ammonium hydroxide.
Reverse-Phase HPLC: HPLC purification was performed on an Agilent 1100 series (semi-
preparative/analytical) and a Varian PrepStar SD-1 (preparative) instruments using Luna 5 μm
particle diameter C8 with TMS endcapping columns with silica solid support. An ammonium
formate aqueous buffer (pH 8) and acetonitrile was used as the mobile phase.
UV-Vis and Photoluminescence: UV-Vis spectra were obtained using a Varian Cary 50 Bio UV-
Vis spectrophotometer. Photoluminescence spectra were obtained using a PTi Photon Technology
International Fluorometer (QuantaMaster 40) with a 75-W Ushio Xenon short arc lamp and
operated with Felix32 Version 1.2 software. Spectroscopic samples were prepared by diluting the
peptide solution to the appropriate concentration (exact concentrations given in spectra captions)
in Millipore water to achieve an optical density near 0.1. The pH was then adjusted by adding 30
μL of 2M KOH (basic) followed by addition of 80 μL of 2M HCl (acidic).
Circular Dichroism (CD): CD spectra were obtained using a Jasco J-810 spectropolarimeter.
Spectroscopic samples were prepared by diluting the peptide solution to the appropriate
8
concentration (exact concentrations given in spectra captions) in Millipore water. The pH was then
adjusted by adding 30 μL of 2M KOH (basic) followed by addition of 80 μL of 2M HCl (acidic).
Dynamic Light Scattering (DLS): CD spectra were obtained using a Zetasizer Nano-ZS90
(Malvern Instruments). Spectroscopic samples were prepared by diluting the peptide solution to
the appropriate concentration (exact concentrations given in spectra captions) in Millipore water.
The pH was then adjusted by adding 30 μL of 2M KOH (basic) followed by addition of 80 μL of
2M HCl (acidic).
Attenuated total reflection Fourier transform infrared (ATR-FTIR): ATR-FTIR spectra were
obtained on lyophilized acidic peptide solutions using a Thermo Scientific Nicolet iD5 ATR-IR.
Transmission Electron Microscopy (TEM): Imaging was performed on a Philips EM 420
transmission electron microscope equipped with an SIS Megaview III CCD digital camera at an
accelerating voltage of 100 kV. Acidic solutions (0.1%) of each peptide were prepared by placing
the samples in a closed container with a vial of conc. HCl opened within, which initiated the
assembly process by diffusion of HCl vapor to the sample. The assembled peptides in water were
pipetted (drop of 1 mg/mL solution) onto 200 mesh copper grids coated with carbon and incubated
for 5 minutes at 25°C. Excess solution was wicked off by touching the side of the grid to filter
paper. The samples were then stained with a 2% uranyl acetate solution and excess moisture was
wicked off. The grid was allowed to dry in air before imaging.
General solid phase peptide synthesis (SPPS)
All peptides were synthesized using the standard Fmoc solid phase technique with Wang resin pre-
loaded with the terminal amino acid (Wang-Val, 0.700 mmol/g). To the resin in a peptide chamber,
Fmoc-deprotection was accomplished by adding a (1:4) piperidine/DMF solution twice
9
(successive 5-and 10-minute treatments) followed by washing with NMPx3, methanolx3 and
DCMx3. For the amino acid couplings, 3.0 eq. of the Fmoc-protected amino acid was added along
with 2.9 eq. of HBTU and 10 eq. diisopropylethylamine (DIPEA). The reaction mixture was
allowed to mix for 45–90 minutes, after which was rinsed with NMP, methanol and DCM (3 times
each). The completion of all couplings was monitored using a Kaiser test on a few dry resin beads,
repeating same amino acid coupling as needed. The general procedure for amino acid coupling
was repeated until the desired peptide sequence was obtained.
General N-acylation procedure for peptides
Following our previous procedure,70 a solution containing 3 eq. of aryl halide carboxylic acid,
HBTU (2.9 eq.) and DIPEA (10 eq.) was mixed for 3 h with the resin forming an N-acylated
peptide capped with desired halide. The completion of the couplings was assessed using a Kaiser
test on a few dry resin beads. The resin was washed with NMP, methanol and DCM (3 times each).
General On-Resin Stille Coupling Procedure
The solid supported peptide capped with an aryl halide was made following the general SPPS and
N-acylation procedures. The resin (1 eq.) was transferred to a Schlenk flask equipped with a reflux
condenser. The resin was dried under vacuum. Pd(PPh3)4 (4 mol%, relative to resin loading) was
added to the reaction vessel. An approximately 15 mM solution of the bis-stannylated aryl reagent
(0.50 eq) was prepared in DMF. The solution was added to the reaction flask via syringe. The
mixture was heated to 80°C for 16-21 h and was agitated constantly by bubbling nitrogen through
the solution. The mixture was allowed to cool to room temperature. The peptide was subjected to
the general cleavage and work-up procedure to yield the crude product, then further purified by
HPLC.
10
General On-Resin Suzuki Coupling Procedure
The solid supported peptide capped with an aryl halide was made following the general SPPS and
N-acylation procedures. The resin (1 eq.) was transferred to a Schlenk flask equipped with a reflux
condenser. The resin was dried under vacuum. Pd(PPh3)4 (4 mol% relative to resin loading) and
benzene-1,4-diboronic acid (0.55 eq.) were added to the reaction vessel. K2CO3 (8 eq.) was
dissolved in 0.50 mL of water and was added to the reaction flask along with 5-10 mL DMF via
syringe. The mixture was heated to 80°C for 20-27 h and was agitated constantly by bubbling
nitrogen through the solution. The mixture was allowed to cool to room temperature. The resin
was washed with water and then subjected to the general cleavage and work-up procedure to yield
the crude product, then further purified by HPLC.
General cleavage, work-up procedure of peptides
Following solid-phase cross-coupling, the resin was returned to the peptide chamber and again
subjected to a standard wash cycle: 3x NMP, 3x DMF, 3x methanol and 3x DCM. The resin was
treated with 9.50 mL of trifluoroacetic acid, 250 μL water, and 250 μL of triisopropylsilane for 3
h. The peptide solution was filtered from the resin beads, washed 3x with DCM, and was
concentrated by evaporation under reduced pressure. The crude peptide was then precipitated from
solution with 60-80 mL of diethyl ether and isolated through centrifugation. The resulting pellet
was triturated with diethyl ether to yield crude product, which was dissolved in approximately 10-
15 mL of water and 50 μL potassium hydroxide (1M) and lyophilized. The solution was placed
inside dialysis tubing of the appropriate length. The tubing was stirred in 1L of water for 1 h, the
water was exchanged, and the tubing was allowed to stir for another 1 h. This process was repeated
twice and then the tubing stirred overnight (approx. 15 h). The tubing was removed from water,
and the peptide solution transferred to a separate container and lyophilized.
11
Molecular simulations
All simulations were conducted using the Gromacs 2018.6 simulation suite77 patched with the
PLUMED version 2.2 enhanced sampling libraries78. The Automated Topology Building (ATB)
server was used to generate topologies and initial configurations of oligopeptides79 modeled using
the GROMOS 54a7 force field80. Dimers of peptides 1a, 1b, 1c, and 1d (Figure 1) containing no
(n = 0), methylene (n = 1), ethylene (n=2), and propylene (n = 3) spacers were initialized with a
center of mass separation |𝒅| = 0.45 nm and a relative twist angle of 𝜃 = 10°, placed in a cubic
simulation box with 1.5 nm spacing from the box edges to the molecule and employing three-
dimensional periodic boundary conditions, and solvated in water with simple point charge (SPC)
model81. Simulations were conducted in the NPT ensemble at 1 bar and 300 K employing a
Parrinello-Rahman barostat82 and velocity-rescaling thermostat83. Initial dimer configurations
were equilibrated using steepest descent energy minimization to remove forces larger than 50
kJ/mol.nm followed by 100 ps NVT and NPT equilibration runs. Initial molecular velocities were
drawn from a Maxwell−Boltzmann distribution at 300 K. Equations of motion were numerically
integrated using the leap-frog algorithm with a 2 fs time step84. Lennard-Jones interactions were
smoothly shifted to zero at 1.0 nm. Particle mesh Ewald was used to treat electrostatic interactions
with a 1.0 nm real-space cutoff and 0.16 nm Fourier grid spacing, both of which were optimized
for performance during runtime85. Covalent bonds involving hydrogen were held fixed using the
LINCS algorithm86. Well-tempered metadynamics simulations87 employing the signed twist angle
𝜃 between the dimers as a collective variable were performed for each spacer length with a
Gaussian deposition pace of 1 ps, width of 0.2 radians, initial height of 1.2 kJ/mol, and bias factor
of 𝛾 = 5.0. Harmonic restraining potentials with a force constant of 500 kJ/mol.rad are applied if
the twist angle exits the window 𝜃 = [-1.25,1.25] radians. Harmonic restraining potentials with a
12
force constant of 500 kJ/mol.nm are applied at center of mass separations of |𝒅| > 0.6 nm.
Estimates of the unbiased potential of mean force (PMF) as a function of the twist angle F(θ) were
extracted from the converged well-tempered metadynamics runs by reweighing contributions to
negate the effects from all sources of imposed bias88. By tracking the height of the deposited
Gaussians, the frequency of transitions in 𝜃, and stablization of the PMF profiles over the course
of the simulation, 2 μs were found sufficient to produce well-converged PMF estimates for all
spacer lengths considered with data from the terminal 1 μs employed for analysis.
Machine learning
We consider intermolecular distances between the 𝑚 = 46 peptide wing atoms in each molecule
to furnish 2𝑚. pairwise distances – 𝑚. for each of the two sets of proximate peptide wings – and
define a featurization 𝑥0 ∈ ℝ34 = ℝ.,667 for each snapshot 𝑖 of the simulation trajectory. The
snapshot is additionally assigned a label 𝑦0 ∈ {+1,−1} defining the instantaneous supramolecular
chirality: (+1) if it is M-type (left-handed) with 𝜃 = [0,π] and (-1) if it is P-type (right-handed) with
𝜃 = [-π,0]. Given training data 𝑋 = {(𝑥6, 𝑦6), (𝑥., 𝑦.), … , (𝑥C, 𝑦C)} a linear SVM attempts to learn
a maximal margin separating hyperplane between the two classes within the feature space under
the decision rule,
𝑓(𝑥) = 𝑠𝑔𝑛(𝑤 ⋅ 𝑥 + 𝑏), Eqn. 1
where 𝑤 and 𝑏 are parameters to be learned89. The vector 𝑤 corresponds to weights given to
different features in 𝑥 discriminating between the two classes and can be interpreted as the
orientation of the SVM hyperplane in the feature space. The scalar 𝑏 is a bias term that can be
interpreted as the offset of the hyperplane from the origin. In practice, the data is typically not
13
linearly separable, and we instead adopt a soft margin SVM corresponding to the following convex
optimization for the parameters 𝑤 and 𝑏89,
min𝑤, 𝑏, 𝜉
6.M|𝑤|MN + 𝐶∑ 𝜉0C
0Q6 Eqn. 2
subject to: 𝑦0(𝑤 ⋅ 𝑥0 + 𝑏) ≥ 1 −𝜉0; 𝜉0 ≥ 0∀𝑖 = 1,… ,𝑁,
where the slack variables 𝜉0 measure the distance from the corresponding classification boundary
of misclassified samples, the penalty parameter 𝐶 controls the relative weighting within the
objective function between misclassification of points to maximization of the margin, and 𝑝
defines the norm used for the weight decay term. To train this soft margin SVM for chirality
classification each feature over the extracted 𝑁 samples is standardized to zero mean and unit
variance with 80% partitioned as training data and the remaining 20% held-out to evaluate
classification accuracy. Conventionally, SVMs use the L2 norm (𝑝 = 2) where all features in 𝑥
contribute to the decision through non-zero weights in 𝑤. Fitting an L2 norm SVM achieved
reasonable classification accuracies of 65-75% (Table S1).
If instead we employ the L1 norm (𝑝 = 1), this has the property of generating trained SVMs with
sparse solutions wherein some of the weights 𝑤 are set exactly to zero. Varying 𝐶 under the L1
norm may be conceived as a means of performing feature selection to identify those features that
are most important in discriminating supramolecular chirality. At 𝐶=0 the objective function in
Eqn. 2 contains no contribution from classification accuracy and the L1 norm returns the null
model (𝑤 = 0). As 𝐶 increases from zero the trained model accumulates progressively more non-
zero elements within 𝑤 corresponding to the participation of more features (i.e., inter-atomic
pairwise distances) into the classification model. As more features are included in the model the
classification accuracy improves, but we observe a knee in the accuracy at a particular number of
14
features. Beyond this point there are diminishing marginal improvements in classification accuracy
upon incorporating more features (Figure S1). The L1 norm SVM performs automated feature
selection by picking out those intermolecular pairwise distances that are most discriminatory of
the dimer chirality. To identify those atoms in the peptide wings most important in distinguishing
chirality we analyze the SVMs trained up to the knee and assign each atom a rank. The rank is
defined as the sparsity (i.e., number of non-zero coefficients) in the L1 norm SVM model within
which it first appeared. For example, in the case of methylene spacer (n = 1) the C18 atom first
appeared within the C18-C14 pairwise distance in the 5th non-zero coefficient model, meaning it
was assigned a rank of 5. We average the ranks of each atom over the four molecules (n = 0, 1, 2,
3). Atoms with high ranks (i.e., low numerical values) are those that play important roles in
discriminating chirality across the four molecules.
Results:
Scheme 1: Peptide synthesis protocol using Pd-catalyzed on-resin coupling method. X = Br, I; Y = SnMe3, B(OH)2.
Design consideration and synthetic strategy: Syntheses of the peptide-spacer-π conjugates were
achieved via Pd-mediated cross coupling. We previously showed that pre-formed di-acids with
alkyl spacers separating the π-system were difficult to synthesize and were poorly soluble.90 We
were also worried that the increased acidity of the methylene protons adjacent to the π-system
might also lead to undesired reactivity in longer conjugated structures. Herein, Fmoc-based solid-
15
phase peptide synthesis was used to synthesize the peptidic regions followed by N-acylation of the
oligopeptides to install a portion of the chromophore (Ar1) onto the peptidic backbone. These
acylation partners were functionalized with bromine/iodine making them suitable for palladium
catalyzed on-resin couplings with another arene (Ar2) to synthesize peptide-π-peptide compounds
(Scheme 1). The generality of the method was validated by using either distannylated (Y = SnMe3)
cores for Stille coupling and diboronated (Y = B(OH)2) cores for Suzuki coupling. In contrast to
our prior synthetic strategies, we show here how N-acylating agents with varied alkyl spacers are
suitable for this chemistry.
Using this method, all peptides were synthesized in moderate yield (Figure 1). Following
our prior molecular design,70 4-iodobenzoic acid was used as zero-carbon spacer N-acylating agent
which was then coupled with stannyl or boronic acid reagents leading to 1a, 2a, 3a or 4a
respectively (n = 0, Figure 1). Peptides with methylene spacers (1b, 3b, 4b) were synthesized
using 4-bromophenylacetic acid as the N-acylating Ar1 group with different Ar2 groups under Stille
conditions, whereas 2b was synthesized using Suzuki coupling conditions. Ethylene spacer
peptides (1c, 2c, 3c, 4c) were synthesized using 3-(4-bromophenyl)propionic acid as the Ar1 group,
and propylene spacer peptides (1d, 2d, 3d, 4d) were synthesized similarly with 4-(4-
bromophenyl)butanoic acid as the Ar1 group. Attempts to use a thiophene based Ar1 group alkyl
spacer (e.g. 2-(5-bromothiophen-2-yl)acetic acid) were unsuccessful: a significant color change
was observed during activation, and the corresponding peptides were isolated in extremely low
yield. It is possible that during activation of the carboxylic acid to promote solid-phase coupling,
the acidic a-carbon protons (between the thiophene and the carboxylic acid) might be deprotonated
competitively. The peptide sequences were mainly chosen to maintain molecular solubility in
aqueous conditions. We found previously that VEVAG-sequences led to soluble peptide-π-peptide
16
conjugates at basic pH and hence they were used again here. At basic pH, the carboxylic acid
groups (C-terminal and glutamic acid side chains) remain deprotonated and impart intramolecular
repulsion thus preventing aggregation. The pentapeptide sequence was kept constant for all the
peptides whereas spacer length and chromophore structure were varied to identify the effect of
spacer and chromophores in the self-assembly processes which were visualized using electron
microscopy and monitored through solution phase spectroscopic techniques.
Figure 1: Peptides synthesized with different π-chromophores and with varying spacer length.
Electron microscopy: The nanostructures were prepared by acidifying the aqueous ‘dissolved’
peptide solutions with HCl vapor in a closed chamber (assembled peptides were pipetted onto
copper grids coated with carbon followed by staining with 2% uranyl acetate solution) and were
observed by TEM. TEM imaging revealed that the introduction of a spacer group did not have a
dramatic or predictable influence on the adopted supramolecular morphology. All the
nanostructures form 1-D tape like structures with high aspect ratios that are on the order of
micrometers in length which are comparable to other peptide-π-peptide conjugates reported
previously.70, 71 The nanostructures showed different widths and persistence lengths as expected
from these highly disperse supramolecular networks. For example, peptide 1a formed
nanostructures that were generally over microns in length whereas 1b-1d all formed relatively
shorter nanostructures that were only few hundreds of nanometers in length (Figure 2).
Nanostructures formed from 1a and 1d were around 10.2 ± 0.6 nm and 10.4 ± 0.6 nm respectively
in diameter while those from 1b and 1c had relatively smaller widths (ca. 5.2 ± 0.5 nm and 6.7 ±
17
0.7 nm respectively). Other TEM images may be found in the Supporting Information (Figures
S50-S53). Although the sizes of the nanostructures do not follow any specific trends, these
observed bundling characteristics indicate different extents of lateral interactions of one or more
filaments are impacted by the spacer length and the associated changes in electrostatic interactions
and solvation of the molecules. We did not observe any finer superstructure details of these
nanostructures using AFM or TEM (e.g. twisting or coiling) that could be related in any way to
the chirality of molecular precursors or the overall handedness of the resulting supramolecular
assemblies.
Figure 2: Representative TEM images of peptide 1a (1st row, left), 1b (1st row, right), 1c (2nd row, right), 1d (2nd row, left).
UV-Vis absorption and photoluminescence: Photophysical responses of different peptides in
acidic (pH 2) and basic (pH 8) conditions provide information about the extent of electronic
coupling within peptide-π-peptide nanostructures. In these types of triblock molecules, the acidic
glutamic acid side-chains are deprotonated at basic pH which renders the peptides essentially
“dissolved” while acidic pH screens the carboxylate charges via protonation and allows for self-
assembly. The photophysical responses recorded here should be treated as ensemble averages of
different local chromophore environments because the materials tend to form highly disperse
18
networks with varying widths and persistence lengths as evident from TEM images. Four different
chromophore segments were studied, where each segment was subject to different spacer lengths.
In line with established studies,54, 55, 70, 91 exciton coupling (i.e. interaction of the transition dipoles
associated with interacting chromophores that leads to new electronic states) in these tri-block
molecules is expected to result from the hydrogen-bonding between amino acids in the peptidic
regions that forces cofacial interactions among the chromophores. Absorption spectra of all the
peptides in the assembled state showed significant blue-shifts with respect to the spectra recorded
in molecularly dissolved solutions, which indicates classic H-type aggregation (i.e. transition
dipoles of the chromophores are oriented in parallel-like fashion shifting the absorption maxima
to higher energies side).92 Generally, stronger co-facial interaction should be responsible for larger
blue-shifts in the absorption maxima, when moving from the molecularly dissolved to the
assembled states. However, quadrupole repulsion between the chromophores might compete with
cofacial stacking, potentially moving to more of a twisted slip-stack type configuration.
Incorporation of a spacer should enable conformational freedom which was not accessible in
peptides with direct conjugation of carboxamide groups and internal π-systems.
The absorption profiles of the chromophores in molecularly dissolved (basic) conditions
show that carboxamide groups are decoupled from the chromophore regions when the spacer
groups are inserted. As a representative example, the direct carboxamide attachment of 3a in basic
solution shows a λmax of 397 nm vs a substantially blue-shifted 372 nm in methylene-spaced 3b
(Figure 3a). Furthermore, the methylene, ethylene and propylene spaced compounds (3b, 3c and
3d, respectively) had near-identical λmax values in basic conditions which indicates a lack of π-
character in the intervening alkyl spacers regardless of the length (Figure S62c). These trends
were also apparent in the molecularly dissolved spectra of 1a-d and 2a-d (Figure S62a-b). Upon
19
assembly in acidic media, the classic signatures associated with H-like aggregation were observed:
3a showed a blue shift of about 40 nm in assembled state whereas 3b, 3c and 3d showed blue shifts
of about 25-30 nm (Figure 3a, S62c). More conformational freedom around the peptide-π
chromophore connection region for spacer-peptides may lead to geometric changes in the excited
states resulting in low oscillator strengths which are also possible due to breaks in conjugation
between carboxamide group and the π-cores. Similar results were found for assemblies derived
from 1a-d and 2a-d (Figures S62a-b).
UV-Vis spectroscopy of molecularly-dissolved peptides 4a-4d showed similar features at
ca. 240 nm and 290 nm. The lower-energy features revealed the spacer-dependent differences,
whereby 4a had a peak absorption at 354 nm with a shoulder at 402 nm, while 4b had a blue shifted
peak at 338 nm with a shoulder at 387 nm (Figure 3b). Ethylene and propylene spaced variants
4c and 4d were essentially identical to the methylene spacer system (Figure S62d). Upon going
from molecularly dissolved (basic pH) to aggregated (acidic pH) state, the low-energy absorption
maxima were blue shifted for all the chromophores (4a-4d) indicating H-type aggregation that
would suggest more co-facial interaction of the chromophores. 4a showed blue shifts of ̴10 nm
whereas 4b-d showed blue shifts of ̴5 nm.
Figure 3: UV-Vis spectroscopy of (a) 3a,3b and (b) 4a,4b at basic (solid lines) and acidic (dashed lines) condition acquired at a concentration of 20 µM. Basic solutions were made by adding 30 µL of KOH (2M) and acidified by adding 50 µL of HCl (2M) to the basic solution.
20
Monomeric peptides with no inserted spacers show relatively unstructured emission
profiles whereas those with alkyl spacer units show more structured peaks with clear vibronic
progressions (Figures 4a, 4b, S63, S64, S65). The spectral signatures also vary significantly
between the unassembled (basic pH) and aggregated (acidic pH) state, again suggestive of the
formation of H-type aggregates as evident from the pronounced quenching of the signal intensity.
Subtle variations in the component peptide residues can tune excimer-like (weakly coupled
chromophores) to exciton-like (tightly coupled chromophores) structure formation which was
primarily justified based on the hydrophobicity of the component residues and associated aqueous
solvent influences.91 In general, no-spacer molecules show broad, featureless emission peaks
resembling excimer type-structures in their aggregated states whereas more vibronic features were
apparent for peptides with alkyl spacer units. Representative spectra are shown in Figure 4 for the
phenyl-benzodithiophene-phenyl series (4a-4d). This shows that peptides with spacer units (4b,
4c, 4d) show structured spectra with pronounced higher-energy vibronic features whereas the
emission from the direct peptide-attached 4a shows a broad featureless spectral signature.
Figure 4: PL spectroscopy of (a) 4a-4d in basic (ca. pH 10) solution (excitation λ: 355 nm (4a), 344 nm (4b-4d)) and (b) 4a-4d in acidic (ca. pH 2) solution (excitation λ: 345 nm (4a), 338 nm (4b-4d)) acquired at a concentration of 5 µM. Basic solutions were made by adding 30 µL of KOH (2M) and acidified by adding 50 µL of HCl (2M) to the basic solution.
Circular dichroism: Circular dichroism (CD) was also employed to examine the locally chiral
chromophore environments imposed during molecular self-assembly. Peptide hydrogen bonding
21
networks are routinely analyzed by CD spectroscopy based on the collective response of the amide
n-π* transitions.93 The assembled peptides here typically showed a clear minimum around 220-
230 nm region which is indicative of a twisted β-sheet formation that is red-shifted from the ideal
β-sheet signature of 216 nm (Figure S66). FT-IR spectra from lyophilized samples also showed
amide I bands (ca. 1640-1620 cm-1) associated with typical β-sheet type signature (Figure S54-
S57). More importantly, all the peptides showed exciton-coupled bisignate Cotton effects (i.e. the
sign change of the CD signal around absorption maxima due to the chiral interaction of transition
dipoles of adjacent chromophores)94 for the π-π* transitions associated with the π-conjugated
chromophores, where the CD signatures undergo a change in sign near the chromophore’s
absorption maximum. The influences of subtle changes in the molecular structure of the local
arrangement of the chromophores were clearly visible from the CD spectra. Longer chromophores
usually show more intense signatures, as expected from their greater UV-Vis extinction
coefficients. A pronounced “odd-even” effect depending on the spacer length in between the π-
core and the peptidic region was visible which resulted in variation of the sign of the CD signal.
We found negative Cotton effects for peptides with no spacer (n = 0) and ethylene spacer (n = 2)
that correspond to M-type helical supramolecular aggregates (akin to the left-handed
supramolecular twist in natural β-sheet peptides) whereas peptides with methylene (n =1) and
propylene (n = 3) spacers show positive Cotton effects corresponding to P-type helical bias
(Figure 5a and 5b). Idealized illustrations are shown in Scheme S1.
Another obvious CD feature was the “odd-even” change in intensity. The intensity of the
Cotton effect varied from one spacer length to another which gives an idea about the extent of π-
π interaction and the strength of the subsequent exciton coupling. For example, 3a (no-spacer) and
3c (2-carbon spacers) showed higher molar ellipticities whereas 3b (1-carbon spacer) and 3d (3-
22
carbon spacers) showed relatively weaker signatures (Figure 5c). These trends were also apparent
in the CD signature of 1a-d and 4a-d (Figure 5a, 5b). In principle, a more organized nanostructure
with a stronger helical response should result in higher molar ellipticity, however, the possibility
of having both left and right-handed helical orientations and highly dynamic helical segments
within a more disordered assembly can also contribute to relatively lower molar ellipticities.
Peptides with embedded benzodithiophene units (4a-4d) showed interestingly different
behavior as compared to the other series (Figure 5d). 4a without a spacer displayed a small degree
of exciton coupling whereas peptides (4b, 4c and 4d) with spacers did not show the split Cotton
band but rather a non-zero response with peaks that mirrored those in the absorption spectrum.
This would indicate that these chromophores are not undergoing exciton coupling but rather are
nevertheless held in the chiral environments dictated by the peptide helical assemblies. We
speculate that the larger quadrupolar surfaces of the phenyl-appended benzodithiophene cores are
frustrated from significant exciton coupling in the assembled state due to the additional freedom
imparted by the intervening alkyl spacers.
Odd-even effects are commonly observed for n-alkanes and their derivatives95, impacting
for example their melting point96 and glass-transition temperature97. Pronounced odd-even effects
are also observed for assembly of n-alkanes on monolayer surfaces resulting from different
conformation of chains98, 99 and also impacting their charge conduction properties100. In liquid
crystals, odd-even effects lead to notable changes in thermal properties, and the movement of an
asymmetric center along the terminal alkyl spacer led to alternating ferroelectric
properties101. Field-effect mobilities were also tuned periodically with odd and even number of
carbons in alkyl side chains of a diketopyrrolopyrrole-vinylene thiophene conjugated polymer102,
23
103. To the best of our knowledge, chiroptical impacts of odd-even alternation and their impacts in
peptide-π conjugated systems have not been widely investigated58.
Figure 5: CD spectroscopy of peptides 1a-1d (a), 2a-2d (b), 3a-3d (c) and 4a-4d (d) in their assembled states (acidic pH). Spectroscopic data were collected at 20 µM concentration. Solutions were prepared adding 30 µL of KOH (2M) and then made acidic using 50 µL of HCl (2M).
Computational Studies:
To complement the experimental studies and resolve the molecular origin of the observed
preferences for supramolecular chirality as a function of spacer length, we conducted all-atom
molecular dynamics simulations of dimerized pairs of peptides 1a-1d. We previously employed
molecular dynamics simulation and free energy calculations to quantify the thermodynamic
driving forces underpinning the assembly into nanoaggregates of a closely related family of
24
oligopeptides containing oligophenylvinylene π-systems104. These calculations revealed (i) a
favorable free energy for the addition of each subsequent peptide monomer to the growing
nanoaggregate of ~25 kBT , (ii) a favorable peptide-peptide driving free energy of ~100 kBT that is
counterbalanced by a large unfavorable solvent contribution of ~80 kBT and a small unfavorable
intrapeptide contribution of ~10 kBT, and (iii) the interaction between the π-cores to be responsible
for ~20% of the peptide-peptide interaction with the remainder made up by interactions between
the π-cores and the peptide wings and between the peptide wings. In the present work we conduct
molecular simulations to quantify the chiral preferences of dimers as a function of spacer length.
The chiral preferences resolved by our molecular simulations of the dimer are in excellent
agreement with those observed in experiment, suggesting that our reductionist focus on the dimer
as the fundamental unit of a self-assembled nanoaggregate stack is well founded. We employ well-
tempered metadynamics87 as an enhanced sampling technique to estimate the potential of mean
force (PMF) as a function of the twist angle 𝜃 between the dimer pair. A similar approach was
previously employed by Kang et al. to identify metastable configurations of camptothecin
dimers.67 The twist angle is defined as the signed angle between the cores of two dimerized
oligopeptides arbitrarily indexed as molecule 1 and 2,
𝜃 = cosZ6(𝒎𝟏 ⋅ 𝒎𝟐) ⋅ 𝑠𝑔𝑛((𝒎𝟐 ×𝒎𝟏) ⋅ 𝒅) Eqn. 3
where 𝒎𝟏 and 𝒎𝟐 are unit vectors connecting the terminal aromatic core carbon atoms in
molecules 1 and 2, respectively, and 𝒅 is the unit vector directed from the center of mass of
molecule 1 to molecule 2. The sense of the vectors 𝒎𝟏 and 𝒎𝟐 and identity of molecules 1 and 2
is immaterial provided that they are consistently defined. The definition of 𝜃 in Eqn. 3 can be
understood as follows. The first term, cosZ6(𝒎𝟏 ⋅ 𝒎𝟐), computes the unsigned angle between the
two vectors 𝒎𝟏 and 𝒎𝟐 and is restricted to the range [0-π] radians. The second term,
25
𝑠𝑔𝑛((𝒎𝟐 ×𝒎𝟏) ⋅ 𝒅), endows 𝜃 with a sign by computing the handedness of the twist through the
projection of cross product 𝒎𝟐 ×𝒎𝟏 onto 𝒅 and taking the sign. So defined, the angle lies in the
range 𝜃 = [-π, π], where positive values of 𝜃 correspond to left-handed twist angles (M-type
supramolecular stacking) and negative values to right-handed (P-type). Inverting the identity of
the molecules leaves the twist angle unchanged. Inverting the sense of one or other of 𝒎𝟏 and 𝒎𝟐
replaces the twist angle with its negative supplement 𝜃 → -(π-𝜃) = (𝜃- π). By the symmetry of the
oligopeptides, twist angles of 𝜃 and (𝜃- π) define equivalent configurations of the dimer and the
PMF need only be estimated over the range 𝜃 = [-π/2, π/2]. In fact, we resolve all the interesting
features of the PMF and markedly improve our sampling efficiency by restricting sampling to the
range 𝜃 = [-1.25,1.25] wherein we apply harmonic restraining potentials if the twist angle moves
outside of this window. We elected not to perform sampling over the full range 𝜃 = [-π/2, π/2] as
tests showed that this frequently led to the formation of interlocked configurations in the vicinity
of ±π/2 where each peptide would curl around the other in a configuration resembling a pair of
interlocking “U” characters or a link in a chain. These configurations are artifacts associated with
the dimer at large twist angles and are not representative of configurations observed in multimeric
peptide stacks. These kinetically trapped configurations led to hindered sampling in 𝜃 and required
prohibitively long simulations to reach convergence. We also applied harmonic restraining
potentials to the center of mass separation between the oligopeptides for |𝒅| > 0.6 nm in order to
maintain the oligopeptides within a contact dimer and better mimic the conditions within a
multimeric peptide stack. Full details of our molecular simulation and enhanced sampling
protocols are provided in the Experimental Methods.
The unbiased PMFs as a function of the twist angle F(𝜃) estimated from well-tempered
metadynamics simulations at each of the four spacer lengths are presented in Figure 6. In excellent
26
agreement with experimental observations, the global minimum of the PMF for no and ethylene
spacers (n = 0, 2) lies at positive values of 𝜃 corresponding to a thermodynamic preference for a
M-type (left-handed) supramolecular stacking, whereas for ethylene and propylene spacers (n = 1,
3) it lies at negative values of 𝜃 corresponding to P-type (right-handed) stacking. This agreement
between experiment and computation lends strong support to our hypothesis that pairwise
intermolecular interactions govern supramolecular chirality. We now proceed to interrogate our
simulation data to identify the molecular-level root of this effect.
Figure 6. Potentials of mean force (PMF) as a function of the signed twist angle θ within the oligopeptide dimer for each of the four spacer lengths. Positive twist angles θ = [0, π] correspond to M-type (left-handed) supramolecular stacking whereas negative twist angles θ = [-π, 0] correspond to P-type (right-handed) stacking. Harmonic restraining potentials with a force constant of 500 kJ/mol.rad is applied if the twist angle exits the window θ = [-1.25,1.25] radians and of 500 kJ/mol.nm for center of mass separations |𝒅| > 0.6 nm. The PMFs F(θ) are computed over the terminal 1 μs of the metadynamics simulation runs. Uncertainties are estimated from block averages over PMFs computed over five 200 ns contiguous blocks where each PMF is shifted to be mean-free and error bars calculated pointwise. In agreement with experimental observations the oligopeptides with no and ethylene spacers (n = 0, 2) show a thermodynamic preference for M-type supramolecular chirality (θ>0) whereas for methylene and propylene spacers (n = 1, 3) impose a preference for P-type chirality (θ<0). Insets show representative snapshots of the dimer system near the respective global minima where water is removed for clarity. Molecular renderings are constructed using VMD105.
27
A number of prior experimental and computational studies have explored chirality transfer
from molecular chiral centers to supramolecular assemblies24, 39, 67-69, with a subset of these
focused on chirality inversion mediated by molecular spacers43, 58, 106, 107. Proposed molecular
mechanisms underpinning supramolecular chirality transfer include hydrogen bonding,
hydrophobicity, electrostatics, host-guest interaction, metal-ligand coordination, and solvent
mediation24. To date, however, it has remained a challenge to identify from molecular simulations
the molecular mechanism underpinning the observed chirality inversion. This failure can be largely
attributed to the high-dimensional nature of the molecular simulation trajectories and challenges
in comprehensively sampling the thermally-accessible phase space. In this work, we surmount
both of these challenges by performing metadynamics enhanced sampling to comprehensively
explore the relevant oligopeptide configurations and by training simple "white box" machine
learning models to identify the molecular variables that are most responsible in discriminating the
supramolecular chirality. Specifically, we take each frame of our metadynamics molecular
simulations where |𝒅| < 0.6 nm and featurize it with the intermolecular pairwise distances
between all atoms in the peptide wings (i.e., excluding atoms in the core or spacer regions).
Considering the 𝑚 = 46 peptide wing atoms in each molecule furnishes 2𝑚. pairwise distances
– 𝑚. for each of the two sets of proximate peptide wings – defining a featurization 𝒙0 ∈ ℝ34 =
ℝ.,667 for snapshot 𝑖 of the simulation trajectory. The simulation snapshot is assigned a label 𝑦0 ∈
{+1,−1} defining the instantaneous supramolecular chirality: (+1) if it is M-type (left-handed)
with 𝜃 = [0,π] and (-1) if it is P-type (right-handed) with 𝜃 = [-π,0]. It is our hypothesis that training
L1-regularized linear support vector machine (SVM) classifiers to predict supramolecular chirality
as a function of intermolecular pairwise distances through a linear relation 𝑦a0 = 𝑓(𝒙0) we can
distill from our model those atoms that are the most important discriminants of chirality. This
28
approach can be conceived of as performing simultaneous model training and feature selection to
identify a sparse model – one that uses a small number of intermolecular pairwise distances – to
make a predictive classification of chirality. Those intermolecular distances that are retained by
the model identify those atoms with the greatest discriminatory power in correctly identifying
whether the molecules have a thermodynamic preference for M-type or P-type stacking. We
present full mathematical details of our approach in the Experimental Methods.
We present in Figure 7A a rank ordered list of the atoms identified by our SVMs to possess
the highest discriminatory power in distinguishing molecular chirality. A chemical structure
providing a key to the atom labels is provided in Figure 7B. The highest-ranked atoms are those
involved in hydrogen bonding (H11, O9, H6, O8) and the hydrophobic alanine and valine side-
chain atoms (C18, C14, C15). This indicates that the formation of hydrogen bonds and packing of
hydrophobic side-chain atoms are critical discriminants of supramolecular chirality. The high
ranks of hydrophobic side-chain atoms suggest that unfavorable handedness may lead to
conformational changes along the backbone which discourage good packing of these sidechains
into hydrophobic cavities created by the neighboring molecule (Figure 7C). The consistently
highly ranked backbone carbonyl oxygens (O9, O8, O7, O6) and amide hydrogens (H11, H6, H10)
further reveal that hydrogen bonding along the backbone plays an important role with hydrogen-
bonding participants nearer to the core generally selected as more impactful in discriminating
chirality than those farther along the peptide wing.
29
Figure 7. Atoms identified by the L1 norm SVM as those most discriminatory of supramolecular chirality. (A) Atomic ranks averaged over all four molecules with spacers n = 0, 1, 2, and 3. The rank is defined as the sparsity of the L1 norm SVM model in which the atom first appears. High ranks (i.e., 1, 2, 3, …) imply that the atom is an important discriminant of chirality. (B) Chemical structure indicating the locations of the atoms within the peptide wings. The color coding of the chemical structure matches that in (A). (C) Selected snapshot from the molecular simulation of the methylene spacer (n = 1) dimer. The thermodynamically preferred right-handed configuration exhibits better hydrophobic packing of the Ala sidechain united-atom carbon (C18, circled) against the other chain, whereas in the thermodynamically disfavored left-handed configuration it is solvent exposed.
Finally, we performed a decomposition of the intermolecular interaction energy into the
Coulombic, π-π interactions, and hydrogen bonding interactions as a function of spacer length
averaged over our molecular simulation trajectories (Table 1). We observe that interactions
between the π-cores constitutes ~15% of the total intermolecular interaction energy and is
30
approximately constant as a function of spacer length. The magnitude of this contribution is similar
to the value of ~20% we previously determined for oligopeptides containing oligophenylvinylene
π-cores.104 The hydrogen bonding interactions are responsible for ~20% of the interaction energy
for no spacer (n=0), but this falls to ~10% for n=1,2,3. A similar trend is observed in the Coulombic
contribution, which decreases monotonically from ~25% at n=0 to ~15% at n=3. These trends
reflect our observations from simulation wherein hydrogen bonds – particularly those nearest to
the aromatic cores – are less frequently made for the molecules containing spacers than those
without, likely due to the larger configurational flexibility provided to the peptide wings by the
spacers. Our results also show the Coulombic contribution to the intermolecular interaction energy
is dominated 3:1 or more by the Lennard-Jones contribution at all spacer lengths.
Table 1. Decomposition of the intermolecular interaction energy within the peptide dimer as a function of spacer length. The total interaction energy is the sum of the Columbic and Lennard-Jones interactions between all atoms in one molecule and all of those in the other. The π-core – π-core interaction comprises the intermolecular Coulomb and Lennard-Jones interactions restricted to those atoms within the phenyl-thiophene-phenyl core. The hydrogen bonding interaction comprises the intermolecular Coulomb and Lennard-Jones interactions between the C=O and N-H groups within the peptide wings.
no spacer
(n=0)
methylene spacer
(n=1)
ethylene spacer
(n=2)
propylene spacer
(n=3)
Coulombic (25.5 ± 5.1) % (19.7 ± 0.8) % (18.7 ± 4.7) % (15.9 ± 1.0) %
π-core – π-core (13.0 ± 0.5) % (14.8 ± 0.3) % (15.3 ± 0.6) % (14.1 ± 0.5) %
hydrogen bonding
(19.7 ± 1.7) % (11.8 ± 0.3) % (10.6 ± 1.4) % (11.3 ± 0.6) %
31
Conclusions:
We have developed a synthetic methodology to incorporate alkyl spacers in between peptides and
π-electron groups as a way to mitigate the deactivating/electron withdrawing influence of the
directly conjugated peptide carboxamide groups. UV-Vis and CD spectroscopy of the
supramolecular aggregates indicated that subtle changes in molecular structure correspond to
different arrangements of π-π stacking in the supramolecular aggregates. This was especially
pronounced in the CD spectra as the changes in spacer length resulted in changes of
supramolecular chirality of the nanostructures. TEM showed a variety of nanostructure bundling
characteristics. Accompanying all-atom molecular dynamics simulations and interrogative
machine learning revealed a thermodynamic origin for the experimentally-observed chiral
preferences and revealed the formation of hydrogen bonds and packing of hydrophobic side-chain
atoms to be primary discriminants of the supramolecular chirality The observed formation of well-
defined nanostructures with substantially different chiroptical properties provides an essential
guideline for the synthetic chemist to design hierarchical structures from different molecular
starting points. It is thus necessary to not only understand the molecular origins of a given self-
assembly but to also develop reliable and predictive design rules for self-assembled nanowires
with hierarchical order that can be optimized for energy migration as we have done here for
supramolecular helicity and chromophore coupling. Future studies will probe the chiroptical
implications of these nanostructures in the context of energy transport and biological recognition.
Notes:
The authors declare no financial conflicts.
Acknowledgements:
32
This research is based upon work supported by the National Science Foundation’s Designing
Materials to Revolutionize and Engineer our Future (DMREF) program (Grant Nos. DMR-
1728947 and DMR-1841807). K.S. is supported by the National Science Foundation’s Graduate
Research Fellowship (Grant No. DGE-1746045). This work was completed in part with resources
provided by the University of Chicago Research Computing Center. We gratefully acknowledge
compute time on the University of Chicago high-performance GPU-based cyberinfrastructure
(Grant No. DMR-1828629). We also acknowledge support from Johns Hopkins University. We
thank Mr. Nicholas Marin for his contributions with peptide synthesis, the Center for Molecular
Biophysics (JHU) for the use of a circular dichroism spectrometer, and Prof. Hai-Quan Mao for
the use of a Zetasizer for the dynamic light scattering experiments.
References:
(1) Hoeben, F. J. M.; Jonkheijm, P.; Meijer, E. W.; Schenning, A. P. H. J. About Supramolecular Assemblies of π-Conjugated Systems. Chem. Rev. 2005, 105 (4), 1491–1546.
(2) Aida, T.; Meijer, E. W.; Stupp, S. I. Functional Supramolecular Polymers. Science 2012, 335 (6070), 813–817.
(3) Tovar, J. D. Supramolecular Construction of Optoelectronic Biomaterials. Acc. Chem. Res. 2013, 46 (7), 1527–1537.
(4) Webber, M. J.; Appel, E. A.; Meijer, E. W.; Langer, R. Supramolecular Biomaterials. Nat. Mater. 2016, 15, 13–26.
(5) Leclère, P.; Surin, M.; Viville, P.; Lazzaroni, R.; Kilbinger, A. F. M.; Henze, O.; Feast, W. J.; Cavallini, M.; Biscarini, F.; Schenning, A. P. H. J.; Meijer, E. W. About Oligothiophene Self-Assembly: From Aggregation in Solution to Solid-State Nanostructures. Chem. Mater. 2004, 16 (23), 4452–4466.
(6) Hasegawa, M.; Iyoda, M. Conducting Supramolecular Nanofibers and Nanorods. Chem. Soc. Rev. 2010, 39, 2420–2427.
(7) Ajayaghosh, A.; George, S. J.; Schenning, A. P. H. J. Hydrogen-Bonded Assemblies of Dyes and Extended π-Conjugated Systems. In: Würthner F. (eds) Supermolecular Dye Chemistry. Topics in current chemistry, Berlin, Heidelberg, Springer, 2005: 258, 83−118.
(8) Gazit, E. Self-Assembled Peptide Nanostructures: The Design of Molecular Building Blocks and Their Technological Utilization. Chem. Soc. Rev. 2007, 36, 1263-1269.
(9) Amit, M.; Yuran, S.; Gazit, E.; Reches, M.; Ashkenasy, N. Tailor-Made Functional Peptide Self-Assembling Nanostructures. Adv. Mater. 2018, 30, 1707083.
(10) Raymond, D. M.; Nilsson, B. L. Multicomponent Peptide Assemblies. Chem. Soc. Rev. 2018, 47, 3659–3720.
33
(11) Ulijn, R.V.; Smith, A. M. Designing peptide based nanomaterials. Chem. Soc. Rev. 2008, 37, 664-675.
(12) Panda, S. S.; Katz, H. E.; Tovar, J. D. Solid-State Electrical Applications of Protein and Peptide Based Nanomaterials. Chem. Soc. Rev. 2018, 47, 3640-3658.
(13) Kumar, R. J.; MacDonald, J. M.; Singh, T. B.; Waddington, L. J.; Holmes, A. B. Hierarchical Self-Assembly of Semiconductor Functionalized Peptide α-Helices and Optoelectronic Properties. J. Am. Chem. Soc. 2011, 133 (22), 8564–8573.
(14) Ardona, H. A. M.; Tovar, J. D. Peptide π-Electron Conjugates: Organic Electronics for Biology? Bioconjug. Chem. 2015, 26 (12), 2290–2302.
(15) Khalily, M. A.; Bakan, G.; Kucukoz, B.; Topal, A. E.; Karatay, A.; Yaglioglu, H. G.; Dana, A.; Guler, M. O. Fabrication of Supramolecular n/p-Nanowires via Coassembly of Oppositely Charged Peptide-Chromophore Systems in Aqueous Media. ACS Nano 2017, 11 (7), 6881–6892.
(16) Stone, D. A.; Tayi, A. S.; Goldberger, J. E.; Palmer, L. C.; Stupp, S. I. Self-Assembly and Conductivity of Hydrogen-Bonded Oligothiophene Nanofiber Networks. Chem. Commun. 2011, 47, 5702-5704.
(17) Shao, H.; Seifert, J.; Romano, N. C.; Gao, M.; Helmus, J. J.; Jaroniec, C. P.; Modarelli, D. A.; Parquette, J. R. Amphiphilic Self-Assembly of an n-Type Nanotube. Angew. Chemie Int. Ed. 2010, 49, 7688–7691.
(18) Tsai, W. W.; Tevis, I. D.; Tayi, A. S.; Cui, H.; Stupp, S. I. Semiconducting Nanowires from Hairpin-Shaped Self-Assembling Sexithiophenes. J. Phys. Chem. B 2010, 114 (45), 14778–14786.
(19) Kitamura, T.; Nakaso, S.; Mizoshita, N.; Tochigi, Y.; Shimomura, T.; Moriyama, M.; Ito, K.; Kato, T. Electroactive Supramolecular Self-Assembled Fibers Comprised of Doped Tetrathiafulvalene-Based Gelators. J. Am. Chem. Soc. 2005, 127 (42), 14769–14775.
(20) Saberi Moghaddam, R.; Draper, E. R.; Wilson, C.; Heidari, H.; Adams, D. J. Effect of Electric Field on the Electrical Properties of a Self-Assembled Perylene Bisimide. RSC Adv. 2018, 8, 34121–34125.
(21) Eakins, G. L.; Pandey, R.; Wojciechowski, J. P.; Zheng, H. Y.; Webb, J. E. A.; Valéry, C.; Thordarson, P.; Plank, N. O. V; Gerrard, J. A.; Hodgkiss, J. M. Functional Organic Semiconductors Assembled via Natural Aggregating Peptides. Adv. Funct. Mater. 2015, 25, 5640–5649.
(22) Moulin, E.; Cid, J. J.; Giuseppone, N. Advances in Supramolecular Electronics - From Randomly Self-Assembled Nanostructures to Addressable Self-Organized Interconnects. Adv. Mater. 2013, 25, 477–487.
(23) Besar, K.; Ardona, H. A. M.; Tovar, J.D.; Katz, H. E. Demonstration of Hole Transport and Voltage Equilibriation in Self-Assembled π-Conjugated Peptide Nanostructures Using Field-Effect Transistor Architectures. ACS Nano 2015, 9 (12), 12402-12409.
(24) Liu, M.; Zhang, L.; Wang, T. Supramolecular Chirality in Self-Assembled Systems. Chem. Rev. 2015, 115 (15), 7304–7397.
(25) Xing, P.; Zhao, Y. Controlling Supramolecular Chirality in Multicomponent Self-Assembled Systems. Acc. Chem. Res. 2018, 51 (9), 2324–2334.
(26) Smith, D. K. Lost in translation? Chirality effects in the self-assembly of nanostructured gel-phase materials. Chem. Soc. Rev. 2009, 38, 684-694.
(27) Yashima, E.; Maeda, K.; Iida, H.; Furusho, Y.; Nagai, K. Helical Polymers: Synthesis, Structures, and Functions. Chem. Rev. 2009, 109 (11), 6102–6211.
34
(28) Yashima, E.; Ousaka, N.; Taura, D.; Shimomura, K.; Ikai, T.; Maeda, K. Supramolecular Helical Systems: Helical Assemblies of Small Molecules, Foldamers, and Polymers with Chiral Amplification and Their Functions. Chem. Rev. 2016, 116 (22), 13752–13990.
(29) Palmans, A. R. A.; Meijer, E. W. Amplification of Chirality in Dynamic Supramolecular Aggregates. Angew. Chemie - Int. Ed. 2007, 46, 8948–8968.
(30) Jiang, J.; Meng, Y.; Zhang, L.; Liu, M. Self-Assembled Single-Walled Metal-Helical Nanotube (M-HN): Creation of Efficient Supramolecular Catalysts for Asymmetric Reaction. J. Am. Chem. Soc. 2016, 138 (48), 15629-15635.
(31) Kumar, M.; Brocorens, P.; Tonnele, C.; Beljonne, D.; Surin, M.; George, S. J. A dynamic supramolecular polymer with stimuli-responsive handedness for in situ probing of enzymatic ATP hydrolysis. Nat. Commun. 2014, 5, 5793.
(32) Liu, G.-F.; Zhang, D.; Feng, C.-L. Control of Three-Dimensional Cell Adhesion by the Chirality of Nanofibers in Hydrogels. Angew. Chem., Int. Ed. 2014, 53, 7789-7793.
(33) Lee, C. C.; Grenier, C.; Meijer, E. W.; Schenning, A. P. H. J. Preparation and characterization of helical self-assembled nanofibers. Chem. Soc. Rev. 2009, 38, 671-683.
(34) Wang, Y.; Xu, J.; Wang, Y.; Chen, H. Emerging chirality in nanoscience. Chem. Soc. Rev. 2013, 42, 2930-2962.
(35) Henze, O.; Feast, W. J.; Gardebien, F.; Jonkheijm, P.; Lazzaroni, R.; Leclère, P.; Meijer, E. W.; Scheming, A. P. H. J. Chiral Amphiphilic Self-Assembled α,α′-Linked Quinque-, Sexi-, and Septithiophenes: Synthesis, Stability and Odd-Even Effects. J. Am. Chem. Soc. 2006, 128 (17), 5923–5929.
(36) Bouman, M. M.; Meijer, E. W. Stereomutation in Optically Active Regioregular Polythiophenes. Adv. Mater. 1995, 7 (4), 385–387.
(37) Zhao, Y.; Chen, H.; Yin, L.; Cheng, X.; Zhang, W.; Zhu, X. Chirality Induction of Achiral Polydialkylfluorenes by Chiral Solvation: Odd-Even and Side Chain Length Dependence. Polym. Chem. 2018, 9, 2295–2301.
(38) Lavicoli, P.; Xu, H.; Feldborg, L. N.; Linares, M.; Paradinas, M.; Stafastrom, S.; Ocal, C. Nieto-Ortega, B.; Casado, J.; Navarette, J. T. L.; Lazzaroni, R.; Feyter, S. D.; Amabilino, D. B. Tuning the Supramolecular Chirality of One- and Two-Dimensional Aggregates with the Number of Stereogenic Centers in the Component Porphyrins. J. Am. Chem. Soc. 2010, 132 (27), 9350-9362.
(39) Kulkarni, C.; Korevaar, P. A.; Bejagam, K. K.; Palmans, A. R. A.; Meijer, E. W.; George, S. J. Solvent Clathrate Driven Dynamic Stereomutation of a Supramolecular Polymer with Molecular Pockets. J. Am. Chem. Soc. 2017, 139 (39), 13867–13875.
(40) Jin, W.; Fukushima, T.; Niki, M.; Kosaka, A.; Ishii, N.; Aida, T. Self-Assembled Graphitic Nanotubes with One-Handed Helical Arrays of a Chiral Amphiphilic Molecular Graphene. Proc. Natl. Acad. Sci., U.S.A, 2005, 102, 10801−10806.
(41) Sarkar, A.; Dhiman, S.; Chalishazar, A.; George, S. J. Visualization of Stereoselective Supramolecular Polymers by Chirality-Controlled Energy Transfer. Angew. Chemie - Int. Ed. 2017, 56, 13767–13771.
(42) Xie, Z.; Stepanenko, V.; Radacki, K.; Würthner, F. Chiral J-Aggregates of Atropo-Enantiomeric Perylene Bisimides and Their Self-Sorting Behavior. Chem. Eur. J. 2012, 18, 7060−7070.
(43) Wang, K.; Han, D.; Cao, G.; Li, X. Link spacer controlled supramolecular chirality of Perylene Bisimide-Carbohydrate Conjugate. RSC Advances 2015, 5, 47728–47731.
35
(44) George, S. J.; de Bruijn, R.; Tomović, Ž.; Van Averbeke, B.; Beljonne, D.; Lazzaroni, R.; Schenning, A. P. H. J.; Meijer, E. W. Asymmetric Noncovalent Synthesis of Self-Assembled One-Dimensional Stacks by a Chiral Supramolecular Auxiliary Approach. J. Am. Chem. Soc. 2012, 134 (42), 17789–17796.
(45) Aimi, J.; Oya, K.; Tsuda, A.; Aida, T. Chiroptical Sensing of Asymmetric Hydrocarbons Using a Homochiral Supramolecular Box from a Bismetalloporphyrin Rotamer. Angew. Chem., Int. Ed 2007, 46, 2031−2035.
(46) Stone, D. A.; Hsu, L.; Stupp, S. I. Self-Assembling Quinquethiophene–oligopeptide Hydrogelators. Soft Matter 2009, 5, 1990-1993.
(47) Wall, B. D.; Diegelmann, S. R.; Zhang, S.; Dawidczyk, T. J.; Wilson, W. L.; Katz, H. E.; Mao, H. Q.; Tovar, J. D. Aligned Macroscopic Domains of Optoelectronic Nanostructures Prepared via Shear-Flow Assembly of Peptide Hydrogels. Adv. Mater. 2011, 23 (43), 5009–5014.
(48) Tsai, W.-W.; Li, L.-S.; Cui, H.; Jiang, H.; Stupp, S. I. Self-Assembly of amphiphiles with terthiophene and tripeptide segments into helical nanostructures. Tetrahedron, 2008, 64, 8504−8514.
(49) 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 (42), 13840–13841.
(50) Sanders, A. M.; Kale, T. S.; Katz, H. E.; Tovar, J. D. Solid-Phase Synthesis of Self-Assembling Multivalent π-Conjugated Peptides. ACS Omega 2017, 2 (2), 409–419.
(51) Schillinger, E. K.; Mena-Osteritz, E.; Hentschel, J.; Börner, H. G.; Bäuerle, P. Oligothiophene versus β-Sheet Peptide: Synthesis and Self-Assembly of an Organic Semiconductor-Peptide Hybrid. Adv. Mater. 2009, 21, 1562–1567.
(52) Nelli, S. R.; Lin, J-H.; Nguyen, T. N. A.; Tseng, D. T-H.; Talloj, S. K.; Lin, H-C. Influence of amino acid side chains on the formation of two component self-assembling nanofibrous hydrogels. New J. Chem., 2017, 41, 1229—1234.
(53) Zhan, F-K.; Hsu, S-M.; Cheng, H.; Lin, H-C. Remarkable influence of alkyl chain lengths on supramolecular hydrogelation of naphthalene diimide-capped dipeptides. RSC Adv., 2015, 5, 48961–48964.
(54) Draper, E. R.; Walsh, J. J.; McDonald, T. O.; Zwijnenburg, M. A.; Cameron, P. J.; Cowan, A. J.; Adams, D. J. Air-stable Photoconductive Films Formed from Perylene Bisimide Gelators. J. Mat. Chem. C. 2014, 2, 5570-5575.
(55) Draper, E. R.; Archibald, L. J.; Nolan, M. C.; Schweins, R.; Zwijnenburg, M. A.; Sproules, S.; Adams, D.J. Controlling Photoconductivity in PBI Fims by Supramolecular Assembly. Chem. Eur. J. 2018, 24, 4006-4010.
(56) Nalluri, S. K. M.; Berdugo, C.; Javid, N.; Frederix, P. W. J. M.; Ulijn, R. V. Biocatalytic Self-Assembly of Supramolecular Charge-Transfer Nanostructures Based on n-type Semiconductor-Appended Peptides. Angew. Chem. Int. Ed. 2014, 53, 5882-5887.
(57) Mba, M.; Moretto, A.; Armelao, L.; Crisma, M.; Toniolo, C.; Maggini, M. Synthesis and Self-Assembly of Oligo(p-phenylenevinylene) Peptide Conjugates in Water. Chem. Eur. J. 2011, 17, 2044-2047.
(58) Marty, R.; Nigon, R.; Leite, D.; Frauenrath, H. Two-Fold Odd-Even Effect in Self-Assembled Nanowires from Oligopeptide-Polymer-Substituted Perylene Bisimides. J. Am. Chem. Soc. 2014, 136 (10), 3919–3927.
36
(59) Schillinger, E. K.; Kümin, M.; Digennaro, A.; Mena-Osteritz, E.; Schmid, S.; Wennemers, H.; Bäuerle, P. Guiding Suprastructure Chirality of an Oligothiophene by a Single Amino Acid. Chem. Mater. 2013, 25 (22), 4511–4521.
(60) Cui, H.; Cheetham, A. G.; Pashuck, E. T.; Stupp, S. I. Amino Acid Sequence in Constitutionally Isomeric Tetrapeptide Amphiphiles Dictates Architecture of One-Dimensional Nanostructures. J. Am. Chem. Soc. 2014, 136 (35), 12461−12468.
(61) Digennaro, A.; Wennemers, H.; Joshi, G.; Schmid, S.; Mena-Osteritz, E.; Bäuerle, P. Chiral Suprastructures of Asymmetric Oligothiophene-Hybrids Induced by a Single Proline. Chem. Commun. 2013, 49, 10929–10931.
(62) Selegård, R.; Rouhbakhsh, Z.; Shirani, H.; Johansson, L. B. G.; Norman, P.; Linares, M.; Aili, D.; Nilsson, K. P. R.; Distinct Electrostatic Interactions Govern the Chiro-Optical Properties and Architectural Arrangement of Peptide–Oligothiophene Hybrid Materials. Macromolecules 2017, 50 (18), 7102–7110.
(63) López-Andarias, A.; López-Andarias, J.; Atienza, C.; Chichón, F. J.; Carrascosa, J. L.; Martín, N.; Tuning Optoelectronic and Chiroptic Properties of Peptide-Based Materials by Controlling the Pathway Complexity. Chem. Eur. J. 2018, 24, 7755–7760.
(64) Wang, Y.; Qi, W.; Huang, R.; Yang, X.; Wang, M.; Su, R.; He, Z. Rational Design of Chiral Nanostructures from Self-Assembly of a Ferrocene-Modified Dipeptide. J. Am. Chem. Soc. 2015, 137 (24), 7869–7880.
(65) Percec, V.; Dulcey, A. E.; Peterca, M.; Ilies, M.; Nummelin, S.; Sienkowska, M. J.; Heiney, P. A. Principles of self-assembly of helical pores from dendritic dipeptides. Proc. Natl. Acad. Sci. U. S. A. 2006, 103, 2518−2523.
(66) Garifullin, R.; Guler, M. O. Supramolecular Chirality in Self-Assembled Peptide Amphiphile Nanostructures. Chem. Commun. 2015, 51, 12470–12473.
(67) Kang, M.; Zhang, P.; Cui, H.; Loverde, S. M. π-π Stacking Mediated Chirality in Functional Supramolecular Filaments. Macromolecules 2016, 49 (3), 994–1001.
(68) Wang, M.; Zhou, P.; Wang, J.; Zhao, Y.; Ma, H.; Lu, J. R.; Xu, H. Left or Right: How Does Amino Acid Chirality Affect the Handedness of Nanostructures Self-Assembled from Short Amphiphilic Peptides? J. Am. Chem. Soc. 2017, 139 (11), 4185–4194.
(69) Lin, S.; Li, Y.; Li, B.; Yang, Y. Control of the Handedness of Self-Assemblies of Dipeptides by the Chirality of Phenylalanine and Steric Hindrance of Phenylglycine. Langmuir 2016, 32 (29), 7420-7426.
(70) Sanders, A. M.; Dawidczyk, T. J.; Katz, H. E.; Tovar, J. D. Peptide-Based Supramolecular Semiconductor Nanomaterials via Pd-Catalyzed Solid-Phase “Dimerizations.” ACS Macro Lett. 2012, 1 (11), 1326–1329.
(71) Ardoña, H. A. M.; Besar, K.; Togninalli, M.; Katz, H. E.; Tovar, J. D. Sequence-Dependent Mechanical, Photophysical and Electrical Properties of Pi-Conjugated Peptide Hydrogelators. J. Mater. Chem. C 2015, 3, 6505–6514.
(72) Takimiya, K.; Shinamura, S.; Osaka, I.; Miyazaki, E. Thienoacene-Based Organic Semiconductors. Adv. Mater. 2011, 23 (38), 4347–4370.
(73) Cornil, J.; Dos Santos, D. A.; Crispin, X.; Silbey, R.; Bre´das, J. L. Influence of Interchain Interactions on the Absorption and Luminescence of Conjugated Oligomers and Polymers: A Quantum-Chemical Characterization. J. Am. Chem. Soc., 1998, 120, 1289-1299.
(74) Cornil, J.; Beljonne, D.; Calbert, J-P.; Bre´das, J. L. Interchain Interactions in Organicp-Conjugated Materials:Impact on Electronic Structure,Optical Response, and Charge Transport. Adv. Mater., 2001, 13, 1053-1067.
37
(75) Takimiya, K.; Konda, Y.; Ebata, H.; Niihara, N.; Otsubo, T. Facile Synthesis, Structure, and Properties of Benzo[1,2-b:4,5-b‘]dichalcogenophenes. J. Org. Chem., 2005, 70(25), 10569-10571.
(76) Kashiki, T.; Shinamura, S.; Kohara, M.; Miyazaki, E.; Takimiya, K.; Ikeda, M.; Kuwabara, H. One-pot Synthesis of Benzo[b]thiophenes and Benzo[b]selenophenes from o-Halo-Substituted Ethynylbenzenes: Convenient Approach to Mono-, Bis-, and Tris-Chalcogenophene-Annulated Benzenes. Org Lett., 2009, 11(11), 2473-2475.
(77) Abraham, M. J.; Murtola, T.; Schulz, R.; Smith, J. C.; Hess, B.; Lindahl, E. GROMACS: High Performance Molecular Simulations through Multi-Level Parallelism from Laptops to Supercomputers. SoftwareX 1-2, 2015, 2, 19–25.
(78) Tribello, G. A.; Bonomi, M.; Branduardi, D.; Camilloni, C.; Bussi, G. PLUMED 2: New Feathers for an Old Bird. Comput. Phys. Commun. 2014, 185 (2), 604–613.
(79) Koziara, K.B.; Stroet, M.; Malde, A. K.; Mark, A. E. Testing and Validation of the Automated Topology Builder (ATB) Version 2.0: Prediction of Hydration Free Enthalpies. J. Comput. Aided Mol. Des. 2014, 28, 221–233.
(80) Schmid, N.; Eichenberger, A. P.; Choutko, A.; Riniker, S.; Winger, M.; Mark, A. E.; Gunsteren, W. F. Van. Definition and Testing of the GROMOS Force-Field Versions 54A7 and 54B7. Eur. Biophys. J. 2011, 40, 843–856.
(81) Berendsen, H. J. C.; Postma, J. P. M.; Gunsteren, W. F. Van; Hermans, J. Interaction Models for Water in Relation to Protein Hydration. In: Pullman B. (eds) Intermolecular Forces. The Jerusalem Symposia on Quantum Chemistry and Biochemistry, Dordrecht, Springer, 1981: 331-342.
(82) Parrinello, M.; Rahman, A. Polymorphic transitions in single crystals: A new molecular dynamics method. J. Appl. Phys. 1981, 52(12), 7182-7190.
(83) Bussi, G.; Donadio, D.; Parrinello, M. Canonical sampling through velocity rescaling. J. Chem. Phys. 2007, 126, 014101.
(84) Hockney, R. W. Computer Simulation Using Particles. In: Boca Raton, CRC Press, 1988. (85) Essmann, U.; Perera, L.; Berkowitz, M. L. A smooth particle mesh Ewald method. J. Chem.
Phys.1995, 103(19), 8577-8593. (86) Hess, B.; Bekker, H.; Berendsen, H. J. C.; Fraaije, J. G. E. M. LINCS: A Linear Constraint
Solver for Molecular Simulations. J. Comput. Chem. 1997, 18 (12), 1463–1472. (87) Barducci, A.; Bussi, G.; Parrinello, M. Well-Tempered Metadynamics: A Smoothly
Converging and Tunable Free-Energy Method. Phys. Rev. Lett. 2008, 100, 020603. (88) Tiwary, P.; Parrinello, M. A Time-Independent Free Energy Estimator for Metadynamics.
J Phys. Chem. B 2015, 119, 736-742. (89) Cortes, C.; Vapnik, V. Support-Vector Networks. Machine Learning 1995, 297, 273–297. (90) Diegelmann, S. R. The synthesis and characterization of self-assembling peptides bearing
organic-electronic untis for the development of biologically active nanowires. Ph.D., Johns Hopkins University, 2011.
(91) Wall, B. D.; Zacca, A. E.; Sanders, A. M.; Wilson, W. L.; Ferguson, A. L.; Tovar, J. D. Supramolecular Polymorphism: Tunable Electronic Interactions within π-Conjugated Peptide Nanostructures Dictated by Primary Amino Acid Sequence. Langmuir 2014, 30 (20), 5946–5956.
(92) Kasha, M.; Rawls, H. R.; El-Bayoumi, M. A. The Exciton Model in Molecular Spectroscopy. Pure Appl. Chem. 1965, 11, 371-392.
38
(93) Kelly, S. M.; Price, N. C. The Use of Circular Dichroism in the Investigation of Protein Structure and Function. Curr. Protein Pept. Sci. 2000, 1(4), 349-384.
(94) Harada, N.; Nakanishi, K. Circular Dichroic Spectroscopy - Exciton Coupling in Organic Stereochemistry; University Science Books, Mill Valley, CA, 1983.
(95) Kishore, K.; Bharat, S.; Kannan, S. Correlation of Kauzmann Temperature with Odd-Even Effect in n-alkanes. J. Chem. Phys. 1996, 105, 11364-11365.
(96) Badea, E.; Gatta, G. D.; D’Angelo, D.; Brunetti, B.; Reckova ̌ , Z. Odd-Even Effect in Melting Properties of 12 alkane-α,ώ-diamides. J. Chem. Thermodyn. 2006, 38, 1546-1552.
(97) Yang, K.; Tyagi, M.; Moore, J. S.; Zhang, Y. Odd-Even Glass Transition Temperatures in Network-Forming Ionic Glass Homologue. J. Am. Chem. Soc. 2014, 136(4), 1268-1271.
(98) Tao, F.; Bernasek, S. L. Understanding Odd-Even Effects in Organic Self-Assembled Monolayers. Chem. Rev. 2007, 107(5), 1408-1453
(99) Akkerman, H. B.; Mannsfeld, S. C.; Kaushik, A. P.; Verploegen, E.; Burnier, L.; Zoombelt, A. P.; Saathoff, J. D.; Hong, S.; AtahanEvrenk, S.; Liu, X.; Aspuru-Guzik, A.; Toney, M. F.; Clancy, P.; Bao, Z. Effects of Odd−Even Side Chain Length of Alkyl-Substituted Diphenylbithiophenes on First Monolayer Thin Film Packing Structure. J. Am. Chem. Soc. 2013, 135(30), 11006−11014.
(100) Thuo, M. M.; Reus, W. F.; Nijhuis, C. A.; Barber, J. R.; Kim, C.; Schultz, M. D.; Whitesides, G. M. Odd-Even Effects in Charge Transport across Self-Assembled Monolayers. J. Am. Chem. Soc. 2011, 133, 2962−2975.
(101) Goodby, J. W.; Chin, E.; Leslie, T. M.; Geary, J. M.; Patel, J. S. Helical Twist Sense and Spontaneous Polarization Direction in Ferroelectric Smectic Liquid Crystals. J. Am. Chem. Soc. 1986, 108, 4729-4735.
(102) Yu, H.; Park, K. H.; Song, I.; Kim, M.-J.; Kim, Y.-H.; Oh, J. H. Effect of the alkyl spacer length on the electrical performance of diketopyrrolopyrrole-thiophene vinylene thiophene polymer semiconductors. J. Mater. Chem. C 2015, 3, 11697−11704.
(103) Wang, X.-Y.; Zhuang, F.-D.; Zhou, X.; Yang, D.-C.; Wang, J.-Y.; Pei, J. Influence of Alkyl Chain Length on the Solid-State Properties and Transistor Performance of BN-Substituted Tetrathienonaphthalenes. J. Mater. Chem. C 2014, 2, 8152−8161.
(104) Thurston, B. A.; Tovar, J. D.; Ferguson, A. L. Thermodynamics, morphology, and kinetics of early-stage self-assembly of π-conjugated oligopeptides. Molecular Simulations 2016, 42(12), 955-975.
(105) Humphrey, W.; Dalke, A.; Schulten, K. VMD: Visual Molecular Dynamics. J. Mol. Graph. 1996, 14, 33-38.
(106) Kumar, J.; Nakashima, T.; Tsumatori, H.; Mori, M.; Naito, M.; Kawai, T. Circularly Polarized Luminescence in Supramolecular Assemblies of Chiral Bichromophoric Perylene Bisimides. Chem. Eur. J. 2013, 19, 14090–14097.
(107) Kumar, J.; Nakashima, T.; Kawai, T. Inversion of Supramolecular Chirality in Bichromophoric Perylene Bisimides: Influence of Temperature and Ultrasound. Langmuir 2014, 30, 6030−6037.
39
For Table of Contents Only