Peptide-Controlled Assembly of Macroscopic Calcium Oxalate Nanosheets

The fabrication of two-dimensional (2D) biomineral nanosheets is of high interest owing to their promise for applications in electronics, filtration, catalysis, and chemical sensing. Using a facile approach inspired by biomineralization in nature, we fabricate laterally macroscopic calcium oxalate nanosheets using β-folded peptides. The template peptides are composed of repetitive glutamic acid and leucine amino acids, self-organized at the air–water interface. Surface-specific sum frequency generation spectroscopy and molecular dynamics simulations reveal that the formation of oxalate nanosheets relies on the peptide–Ca2+ ion interaction at the interface, which not only restructures the peptides but also templates Ca2+ ions into a calcium oxalate dihydrate lattice. Combined, this enables the formation of a critical structural intermediate in the assembly pathway toward the oxalate sheet formation. These insights into peptide–ion interfacial interaction are important for designing novel inorganic 2D materials.


Peptide Synthesis and Purification
Microwave-Controlled Solid-Phase Peptide Synthesis and Peptide Purification.
LE10 oligopeptides were prepared on a CEM Liberty peptide synthesizer. DMF and NMP were used as solvents and the standard CEM Liberty coupling protocols were followed. To achieve 1 mmol of oligomer, preloaded Wang resins containing glutamic acid were used. The amino acid coupling was facilitated by use of Oxyma Pure/HBTU (0.5 M in DMF) and DIPEA (2 M in NMP).
Deprotection and release of the oligopeptides from the Wang resins was accomplished by 24 h treatment with a cleavage solution containing 95% TFA, 2.5% TIS (as a scavenger), and 2.5% H2O. The products were filtered and then precipitated in cold diethyl ether and centrifuged before lyophilization. The chemical purity of the oligopeptides was characterized by HPLC. 1.56 mL CaCl2 (0.1M/L) was injected into the sub-phase of the solution, allowing the Calcium ions to be coordinated with the deprotonated glutamic acid residues of LE10 peptides; after 2 hours peptide−Ca 2+ ion interaction, the CaC2O4 mineralization was initiated by injecting 1.56 mL (0.1m/L) Na2C2O4 into the sub-phase of solution. After 20 minutes, the fabricated peptide-CaC2O4 films were lifted off using Langmuir-Schaefer technique by either (50 nm) Au coated silicon wafers or (2000 meshes) copper transmission electron microscopy (TEM) grids without supporting carbon film.

Sample Preparation
The CaC2O4 mineralized peptide films on Au were characterized by X-ray diffraction (XRD) and X-ray photoelectron spectroscopy (XPS); the films on the TEM grids were examined by scanning electron microscopy (SEM) and TEM.

XPS
XPS was conducted using a Kratos Axis Ultra DLD spectrometer (Kratos, Manchester, England) using an Al K excitation source with a photon energy of 1487 eV. The data was acquired in the hybrid mode using a 0° take-off angle, defined as the angle between the surface normal and the axis of the analyzer S3 lens. Detailed and high resolution XP spectra were collected with setting analyzer pass energy at 80 eV and 20 eV, respectively. Neutralizer was always used during spectra collection. Binding energy scales were calibrated to Au 4f7/2 emission at 84 eV. 1 A linear background was subtracted for all peak quantifications. The peak areas were normalized by the manufacturer supplied sensitivity factors and surface concentrations were calculated using CasaXPS software. The thickness of LE10 peptide film after 20 minutes mineralization (dsample) was calculated by evaluating the intensity ratios of the C 1s and Au 4f emissions, 2 according to a standard expression: after calcium oxalate mineralization. The C 1s spectra in (d) was self-fitted with individual peaks assigned for the carbon atoms from leucine side chains (dark yellow), peptide backbone (navy), and glutamic acid side chains and oxalate mineral (purple).

SEM and TEM
SEM measurement was carried out on a Zeiss 1530 LEO Gemini microscope. TEM was carried out with JEM 1400, JEOL Ltd., Japan, and operating voltage of 120 kV was applied during measurement.  with crystal lattice of calcium oxalate dihydrate, however, the diffuse appearance of the diffraction rings suggest that the oxalate crystals in the sheet are arranged into ordered nuclei in short range.

XRD
XRD patterns were recorded on a Phillips PW 1820 diffractometer with an X-ray source producing a Cu K wavelength of 1.5406 Å.  peptides in bulk solution.

Vibrational SFG Spectroscopy:
The vibrational SFG spectra were obtained by overlapping in time and space, visible and IR pulses of light. A Ti:Sapphire amplified system (Spitfire Ace, Spectra Physics Inc.) delivers 35 fs long pulses at a central wavelength of ~800 nm and 1 KHz repetition rate. The beam is split in two parts: one it is spectrally narrowed using a Fabry-Perot etalon to achieve spectral resolution (15cm -1 , lambda=800, E~25mJ/pulse).
The other part is used to generate tunable broadband IR pulses thanks to a parametric optical amplifier followed by a noncollinear difference frequency generation module (TOPAS Prime, LightConversion).
The average power at 6000 nm wavelength is 2J/pulse. Visible and IR beams are focused onto the sample trough using respectively a 20cm and 5cm focal length lenses. The polarization of both beams can be controlled (s or p) with a polarizer and a half waveplate. Beams are temporally and spatially overlapped at the sample position. The SFG signal is generated with visible and IR beam angles of 55° and 60° respective to the surface normal, and is collimated using a 20cm FL lens, further focused into a spectrograph using a 5cm FL achromatic lens, dispersed by a grating and collected by an intensified CCD camera.
The polarization of the SFG signal can also be controlled.
Spectra were recorded in the achiral SPS (sum, visible, and infrared) and chiral PSP polarization combinations. Each spectrum was acquired for 10 minutes, and the spectra are normalized by non-resonance reference spectra of z-cut quartz crystal after background correction, and energy calibration is performed according to the vibration bands of water. D2O solvents was used due to the spectra interference in amide I region by bending mode of H2O.
SFG spectra were fitted by Lorentzian peak shapes according to the following equation: 9 In equation (2) : (2) and (2) are the non-resonant and resonant contributions to the SF signal, respectively. A NR and  NR represent the amplitude and phase of non-resonant term, whereas An is the amplitude of n-th vibrational mode with resonant frequency n and linewidth Γn.  Figure 2A for LE10 peptides during Ca 2+ ion interaction and CaC2O4 mineralization. The parameters A, Γ, and  are derived by fitting SFG spectra according to equation (2).

Molecular Dynamics Simulations:
GROMACS 4.6 was used for all simulations. 10 Peptide parameters were taken from the AMBER99SBildn force field 11 , for calcium the CM parameter set was used. 12 To simulate a water slab, the TIP3P model was used. 13 The simulation box was 9 x 9 x 16 nm in size. Within the box, a 4 nm thick water slab with peptides at the interface was packed with Packmol. 14 The water slab was positioned at half height separating two vacuous spaces of 6 nm in height. Initial configurations of simulation boxes were prepared with twelve beta-sheet LE10 peptides at one of the two vacuum-water interfaces. Simulation boxes were neutralized with an appropriate amount of potassium ions or calcium ions. In order to generate the respective topology and coordinate files, we used the Tleap . The amber input files were subsequently converted to GROMACS input files with Acpype. [15][16] Constraining all bonds using the linear constraint solver (LINCS)allowed for a 2 fs time step. Periodic boundary conditions and the particle mesh Ewald (PME) method was used for electrostatic interactions. 17 Lennard-Jones potential cut-off was set to 1 nm. In order to maintain the temperature at 300 K, velocity rescaling with a stochastic term (v-rescale) was employed. 18 Radial distribution functions were calculated with VMD using in total 890 frames from 445 -500 ns simulation time. 19 Simulation snapshots were rendered with pymol and VMD. 19 found an optimal match with the experimental spectra for = 1000. The total linewidth (homogeneous linewidth + FWMH of the visible pulse) was found to be ~16 cm -1 , and the base frequency was 1622 cm -1 . To keep the spectral calculations as simple as possible we assumed that there is no non-resonant contribution. This might explain why we find rather large interfacial refractive index (ni) changes when optimally reproducing the experimental magnitude changes observed upon the addition of Ca 2+ (see