The Dimensionality of Hydrogen Bond Networks Induces Diverse Physical Properties of Peptide Crystals

Short peptides are attractive building blocks for the fabrication of self-assembled materials with significant biological, chemical, and physical properties. The microscopic and macroscopic properties of assemblies are usually closely related to the dimensionality of formed hydrogen bond networks. Here, two completely different supramolecular architectures connected by distinct hydrogen bond networks were obtained by simply adding a hydroxyl group to switch from cyclo-tryptophan-alanine (cyclo-WA) to cyclo-tryptophan-serine (cyclo-WS). While hydroxyl-bearing cyclo-WS molecules provided an additional hydrogen bond donor that links to adjacent molecules, forming a rigid three-dimensional network, cyclo-WA arranged into a water-mediated zipper-like structure with a softer two-dimensional layer template. This subtle alteration resulted in a 14-fold enhancement of Young’s modulus values in cyclo-WS compared to cyclo-WA. Both cyclo-dipeptides exhibit biocompatibility, high fluorescence, and piezoelectricity. The demonstrated role of dimensionality of hydrogen bond networks opens new avenues for rational design of materials with precise morphologies and customizable properties for bioelectronic applications.

The Rigaku CrysAlis Pro software was used to process the diffraction data.The crystal structures were solved and refined using Bruker SHELXTL software.Except for hydrogen atoms, all atoms were placed in calculated positions and refined in a riding mode.Parameters for data collection and refinement are provided in Table S1, and the final CIF files are given in the Supporting Information.The crystallographic data has been deposited in the CCDC with numbers 2304390 and 2304391.
Young's modulus and point stiffness measurements: Atomic force microscopy (AFM) nanoindentation was used to measure the Young's modulus of the crystals.All the experiments were performed using a commercial AFM (JPK, Nanowizard IV, Berlin, Germany).The crystals were spread on mica substrates and the cantilever was moved to the surface of the crystals.Nanoindentation was performed on the crystal surface (scan area: 5 μm × 5 μm) in QI mode (conditions: pixels: 60 × 60; Z length: 0.1 μm; extend and retract speed: 30 μm s -1 ; Z resolution: 80000 Hz; maximum loading force: 1000 nN) and RTESPA-525 cantilevers (Bruker Company, half-open angle of the pyramidal face of θ: < 10°, tip radius: ~10 nm, spring constant: ~200 N m −1 ) were used in all the experiments.Typically, the cantilever was extended to the surface of the crystal and retracted while indentation depths pressed by the cantilever tip were less than 8 nm.The force and displacement during the process were recorded.The Young's modulus of the crystals was calculated by fitting the retraction curve with the Hertz model (1): (1) F corresponds to the force,  corresponds to the depth of the crystal pressed by the cantilever tip,  is the radius of the tip,  is the Young's modulus of the crystals, and ν is the Poisson ratio (ν = 0.3).The point stiffness was determined as the normal force divided by the deformation of the sample and calculated from the force-displacement curves after deducting the deformation of the cantilever.For each sample, at least six regions were randomly selected to perform the nanoindentation, and at least three cantilevers were used in the experiments to exclude a tip dependency of the results.All the data was analysed and the two-dimensional diagrams were reconstructed using the JPK data processing 7.0.46software (JPK company).
DFT calculations: All modelling was performed using the CP2K package 1 for periodic density functional theory (DFT) calculations. 2The Orbital Transformation 3 (OT) SCF algorithm was used with Goedecker, Teter and Hutter (GTH) type pseudopotentials and a molecular optimized double-zeta gaussian basis set (akin to 6-31G**). 4The cut-off for the plane waves and gaussians was 900 Ry and 60 Ry, respectively, and energy was converged to 10 −8 hartree.
Exchange-correlation effects were treated using the Perdew, Burke, and Ernzerhof (PBE) 5 implementation of the Generalised Gradient Approximation (GGA). 6Grimme D3 dispersion corrections were used to capture van der Waals interactions. 7,8All crystal structures were optimized using conjugate gradient minimization in a supercell model of 3 × 3 × 2 for cyclo-WA and 3 × 2 × 3 for cyclo-WS. 9The piezoelectric parameters were calculated using a finite difference method, with the supercell strained by ± 0.015 in each of the Voigt directions.The piezoelectric tensor is then the response of polarization (considered here as a periodically corrected Berry phase) to the applied strain.The matrix product between the piezoelectric charge tensor, e, provides the piezoelectric strain tensor, d.The DOS distribution was obtained from the optimized supercell structure using a standard diagonalization and -point sampling.
The band structure, which requires large memory and computing time, was performed on a single optimized crystal unit cell using a standard diagonalization and multiple k-points based on the Monkhorst-Pack scheme: 3 × 3 × 2 for cyclo-WA, and 3 × 2 × 3 for cyclo-WS.Special k-points for the band structure path were generated with SeeK-path tools. 10A sampling of 25 data points between consecutive special k-points was used.
Cell viability measurement: A total of 1x10 6 HeLa cells mL -1 were cultured in 96-well tissue microplates (100 µl per well) and allowed to adhere overnight at 37°C for the analysis of cell viability.A fine powder of cyclo-WA or cyclo-WS crystals suspension was added to the cell growth medium at a concentration of 0.5, 1, 2, 5, and 10 mg ml -1 .The cells were seeded in one half of each plate, while the other half served as a blank control.A medium without cyclo-WA or cyclo-WS crystals was used as a negative control.To assess cell viability, 3-(4,5dimethylthiazolyl-2)-2, 5-diphenyltetrazolium bromide was used according to the manufacturer's instructions after a 24-hour incubation at 37°C.Then 10 µl of the 5 mg/mL MTT reagent dissolved in PBS was added to each of the 96 wells, followed by another 3 h of incubation at 37°C.The wells were then filled with 100 µl of extraction buffer (100% DMSO) and incubated for 30 minutes at 37 °C in the dark.Finally, absorbance intensity was measured using a multi-plate reader at 570 nm, with background subtraction at 680 nm.
Live cell imaging: Confocal microscopy was used to obtain images of live HeLa cells grown in the presence of cyclo-WA or cyclo-WS crystals.Briefly, the cells were grown in glass bottom dishes to a confluence of 75%.Afterward, the cells were cultured in media containing cyclo-WA or cyclo-WS crystals at a concentration of 2 mg/mL for varying periods of time.Then, the cells were washed twice with PBS.Images were acquired using a Leica     Figure S6 The unit cell of cyclo-WA crystals.
Figure S7 The "edge-to-face" interaction of cyclo-WA crystals.S3 Calculated piezoelectric charge tensor components e ij (in units of C m -2 ), strain tensor components d ik (pC V -1 ) and piezoelectric Voltage Tensor (mV mN -1 ) of cyclo-WS crystals.

Dielectric Constants
Figure S1 Microscopy images of cyclo-WA crystals obtained in a water and methanol mixed solution at room temperature.

Figure
Figure S2 SEM images of cyclo-WA crystals obtained in water at -5 ℃.

Figure
Figure S3 SEM images of cyclo-WS crystals obtained in water at -5 ℃.The cyclo-WS assembles into polycrystal structures where many crystals are stacked together.This may be induced by heterogeneous nucleation and an uneven distribution of temperature.

Figure
FigureS4XRD patterns of cyclo-WA and cyclo-WS crystals.The cyclo-WA-1 crystal, obtained in water-methanol mixtures at room temperature, and the cyclo-WA-2 crystal, obtained in water alone at -5 ℃, both exhibit the same structure.

Figure
Figure S5 ORTEP diagram of the cyclo-WA crystal with ellipsoid probability of 50%.

Figure
Figure S8 Overlay of space-filling spheres showing the weak van der Waals interactions mediating the layer-layer contacts.Color code: cyan, C; blue, N; and red, O.

Figure
Figure S9 ORTEP diagram of the cyclo-WS crystal with ellipsoid probability of 50%.

Figure
Figure S10The unit cell of cyclo-WS crystals.

Figure
Figure S12 AFM topography images of (a) cyclo-WA and (b) cyclo-WS crystals.The significant heterogeneity or roughness of the crystal surface shown in Figure S12b results from larger ripples on the sample, which may be induced by small crystals or aggregations on its surface.

Figure
Figure S13 Statistical point stiffness of (a) cyclo-WA and (b) cyclo-WS crystals.The point stiffness data appear homogeneous in Figure S13b, indicating that the same materials exhibit similar point stiffness.

Figure
Figure S14 Typical force-displacement traces on (a) cyclo-WA and (b) cyclo-WS crystals.

Figure
Figure S16 (a, b) MTT cell viability analysis of HeLa cells grown in the presence of (a) cyclo-WA and (b) cyclo-WS crystals at different concentrations, as indicated.

Figure
Figure S17 (a,b) Confocal fluorescence images of HeLa cells incubated with (a) cyclo-WA and (b) cyclo-WS crystals.

Figure
Figure S18 (a) Calculated band structure showing the showing the bandgap for cyclo-WS crystals.(b) The corresponding density of states of cyclo-WS crystal, with the Fermi energy level set to zero.

Figure
Figure S19 UV-visible absorption spectra of (a) cyclo-WA and (b) cyclo-WS.The inset shows corresponding band gaps.

Figure
Figure S20 Fluorescence emission spectra of (a) cyclo-WA and (b) cyclo-WS at different excitations.

Table S1
Data collection and refinement statistics of cyclo-WA and cyclo-WS crystals.

Table S2
Calculated piezoelectric charge tensor components e ij (in units of C m -2 ), strain tensor components d ik (pC V -1 ) and piezoelectric Voltage Tensor (mV mN -1 ) of cyclo-WA crystals.