Mechanistic Insights and Rational Design of a Versatile Surface with Cells/Bacteria Recognition Capability via Orientated Fusion Peptides

Abstract Hospital‐acquired infection causes many deaths worldwide and calls for the urgent need for antibacterial biomaterials used in clinic that can selectively kill harmful bacteria. The present study rationally designs fusion peptides capable of undergoing 2D self‐assembly on the poly(methyl methacrylate) surface to form a smart surface, which can maintain a desirable orientation via electrostatic interactions. The in vitro assay shows that the smart surface can recognize bacteria to exert antibacterial activity and is nontoxic toward mouse bone mesenchymal stem cells. Excitingly, the smart surface can distinguish different bacterial strains. This selective feature, from being broad‐spectrum to being highly selective against S. aureus, can be altered by varying the number of amino acids in the recognition sequences. By all‐atom molecular dynamics simulations, it is also found that the recognition sequence in the peptide is critical for the selectivity toward specific bacterial strains, in which a less accessible surface area for the bacteria in the antimicrobial peptide sequence is responsible for such selectivity. Finally, the smart surface can inhibit S. aureus infection in vivo with much more rapid tissue‐healing compared to the control.


Preparations of substrate
The titanium wafers (5 mm × 5 mm × 1 mm) were purchased from Zhongjingkeyi Technology Co., Ltd. (Beijing, China). The polymethyl methacrylate (PMMA) surface was prepared by spinning PMMA (Beiaolu, Shanghai, M w = 8 × 10 4 ) solution onto the titanium wafer. Briefly, 50 μL of PMMA solution in methylbenzene (1 wt%) was dropped onto the titanium wafer and was spun at a speed of 4000 rpm for 30 s.
For the QCM-D assay, the PMMA electrodes were prepared by spinning as above on Ti electrodes, which were purchased from Biolin Scientific (Goteborg, Sweden).

Preparation of the surface
The PMMA surfaces were washed with distilled water following preparation and were dried by nitrogen. After cleaning and drying, the surfaces were treated by oxygen plasma for 0, 2.5 or 5 min, which was abbreviated as PMMA-0min (pristine PMMA), PMMA-2.5min and PMMA-5min, respectively. Then, the surfaces were immersed into 100 μL of the HHC36, CtrlP, KD14 or KD17 solution at a concentration of 500 μM in water for 30 min for the self-assembly.
Afterwards, the surfaces were washed with distilled water for 3 times and dried by nitrogen.

Quartz crystal microbalance with dissipation (QCM-D) assay
The QCM electrodes were treated by oxygen plasma for 0, 2.5 and 5 min, and placed into sensor chambers of the Q-Sense E4 QCM-D system (Q-Sense AB, Sweden). Distilled water was introduced to the electrode until the frequency was balanced. After that, 100 μM of different peptides (KD14 or KD17) solution in water was introduced onto the electrode at a speed of 30 μL/min at 25 °C. After being balanced, the electrodes were rinsed again with distilled water to reach balance again.

Characterization of the surfaces
The water contact angles of the surfaces were recorded on the OCA15 contact angle goniometer (Dataphysics, Filderstadt, Germany) at room temperature with 1 μL of distilled water as the reference liquid. Atomic Force Microscope (AFM) images were obtained by a MutiMode Nanoscope IIIa AFM (Digital Instruments Inc., Santa Barbara, CA). Zeta potentials of different peptides and surfaces were measured on a SZetasizer Nano ZS (Malvern Instruments Inc., Malvern, UK).

In vitro cell assay
Mouse bone mesenchymal stem cells (mBMSCs) were cultured in H-DMEM medium with 10% FBS under 5% CO 2 at 37 °C. The medium was replaced every three days. Cells were passaged after the coverage reached 80% confluence and 3-5 passaged cells were used for the experiments.
All the samples used for the cell assay were sterilized with 75% ethanol for 2 h. After sterilization, the samples were removed to a new 24-well plate and washed twice with PBS.
Then, the mBMSCs were added directly onto each surface (10000 cells per well for the CCK-8 assay and 2 × 10 4 cells for the confocal microscopy assay) and cultured under 5% CO 2 at 37 °C.
The biocompatibility of the sample was evaluated using a CCK-8 kit.  After incubation at 37 °C for 15 h, the number of bacteria on each agar plate was counted.

All-atom molecular dynamics (MD) simulation
All the MD simulations were performed by the Gromacs 4.5.4 package [1] with the general Amber force field. [2] The partial charge was derived by RESP fitting [3] to the electrostatic potential computed by HF/6-31G * . The initial structures of the syndiotactic PMMA molecule with 10 units and the peptide were optimized at a HF/6-31G * level by the Gaussian package. [4] Then, the substrate composed of 135 PMMA molecules with the size of 8 × 8 × 3 nm 3 and density of 1.18 g/cm 3 was built by the "genbox" program in Gromacs. This substrate was energy minimized by the steepest descent method followed by a 200-ns NVT simulation. Next 5, 15 or 25 negative charges were added on the surface of the substrate by demethylation of the methyl ester group in the PMMA molecule randomly, and the substrates were abbreviated as 5_PMMA, 15_PMMA and 25_PMMA, respectively. The partial charge of the PMMA molecule after demethylation was re-derived by RESP fitting [3] to a HF/6-31G * electrostatic potential.
To obtain initial conformations of KD14 and KD17, we first performed MD simulations of these two peptides in bulk solution. In particular, we employed Modeller [5,6] to prepare a helical conformation for both peptides. We then dissolved each of them in a simulation box with the dimension of 5 × 5 × 5 nm 3 and containing 3557 and 4675 TIP3P water molecules [7] for KD14 and KD17 respectively. For both systems, 5 ions were added to neutralize the system. We performed the energy minimization by the steepest descent method followed by a 200-ps NPT simulation with the positions of all heavy atoms restrained. We then performed a 30-ns NPT simulation. In these NPT simulations, the temperature was maintained at 300 K by the V-rescale thermostat, [8] while the pressure was set at 1 bar using the Parrinello-Rahman method.
[9] The LINCS algorithm [10] was used to restrain all the bonds, and the PME method [11] was applied to calculate the long-range electrostatic interactions. The cutoffs of the short-range electrostatic and van der Waals interactions were set to 1.2 nm and 1.1 nm, respectively. We choose the last conformation of this simulation as the initial conformation for further modeling of the binding of these two peptides to the surface as described below.
To model the binding of these two peptides to the surface, we first placed them above the surface by randomly choosing an orientation (the distance between the center of mass of the peptide and the surface was 2 to 3 nm) with the conformation discussed above. We then dissolved each of the peptides in a simulation box of 8 × 8 × 10 nm 3 containing TIP3P water molecules.
[7] For KD14, the system contains 14665, 14666, and 14665 water molecules for the 5_PMMA, 15_PMMA and 25_PMMA surface, respectively. For the KD17, the system contains 14658, 14659, and 14658 water molecules for the 5_PMMA, 15_PMMA and 25_PMMA surface, respectively. For all the systems, we added ions make the simulation box neutral.
We then performed energy minimization using the steepest descent method followed by a 500-ps position restrained NVT simulation, and then another 1-ns position restrained NPT simulation. For each system, we performed four 100-ns production MD simulations in the NVT ensemble with different initial orientations of the peptide relative to the surface with an aggregated simulation time of 2,400 ns. The other set-up of MD simulations were identical as those discussed in the previous paragraph.

Analysis of the MD simulation trajectories
Each peptide were observed to adsorb on the PMMA surface in the 100ns MD simulations, which can be monitored by the distance of the center of mass (COM) of peptide and COM of PMMA along the Z-axis. The coordinates at the Z-axis between the COM of two peptide (KD14 and KD17) and the COM of the three types of substrates (5_PMMA, 15_PMMA and 25_PMMA) were calculated by "g_traj" tool that implemented in Gromacs 4.5.4, respectively. [1] The mean and the standard deviation of the distance was calculated from the 4 independent MD trajectories at indicated time points.
The contact area between the peptide and the substrate was the difference of the solvent accessible surface area between the peptide without PMMA surface and the peptide adsorbed on the PMMA surface, which was calculated by g_sas program implemented in Gromacs 4.5.4. [1] The mean and the standard deviation of the contact surface area was calculated from the 4 independent MD trajectories at indicated time points.
The secondary structures of the KD14 and KD17 were calculated by DSSP program.
[12] The radius of gyration of the peptide was calculated by the "g_gyrate" program implemented in Gromacs 4.5.4. [1] For each of the four MD simulations of each peptide adsorbing on a particular surface, we extracted 4,000 conformations from its last 80 ns segment to compute the secondary structure and radius of gyration.
The binding free energy between the peptide and the surface was calculated by the MM/PBSA method with the "g_mmpbsa" program [13] implemented in Gromacs 4.5.4. In order to properly estimate long-range electrostatic interactions between two groups (i.e. AMP sequence in the peptide and the substrate) in PME, we computed electrostatic interactions for three systems: one containing both AMP sequence and substrate ( ), one containing only AMP sequence ( ), and one containing only the substrate ( ). We can then compute the electrostatic interactions between AMP sequence and substrate by: . The dielectric constant was set as 1 (ε=1). [14,15] For each peptide adsorbing on a particular substrate, the MM/PBSA calculations were performed on a total of 320 conformations extracted from the last 80 ns of four MD simulations. By using the same method, we obtained the binding free energy between the RGD sequence in the peptide and the substrate. We note that it has been suggested recently that choosing a larger value of dielectric constant (e.g. ε=2) may yield more consistent results with experiment especially for systems with extended binding interface.
To compute the bacterial accessible surface area (BASA) of the AMP sequence, we adopted a similar approach for computing the solvent accessible surface area [17] by rolling a probe sphere with the radius of 0.14 nm, 0.6 nm, 0.9 nm or 1.8 nm over the surface of the peptide. To achieve this, we applied the "g_sas" program implemented in Gromacs 4.5.4 on a total of 1,600 conformations extracted from the last 80 ns of four MD simulation for each system, where the peptide was found to be bound to the surface.

In vivo assay
All SD rat's experiments were approved by Guangdong Medical Laboratory Animal Center in After surgery, the SD rats were cultured in cages in a temperature-controllable facility with access to antibiotic-free food and water. After 7 days of the implantation, the SD rats were euthanized and the sample was extracted from the incision site. were acquired by microtome, and the slices were baked at 60 °C for 1 h.
H&E staining was employed to evaluate the histological morphology of in vivo infection. The slices were subsequently immersed in dimethylbenzene I for 15 min (50 °C), dimethylbenzene II for 5 min, dimethylbenzene III for 5 min, 100% ethanol for 5 min, 95% ethanol for 5 min, 80% ethanol for 5 min, 70% ethanol for 5 min, and distilled water for 5 min. After that, the slices were immersed in hematoxylin for 5 min, rinsed with distilled water for 1 min, stained with eosin solution for 5 min, and rinsed with distilled water for 1 min. Finally, the slices were dehydrated by 80% ethanol for 5 min, 95% ethanol for 5 min, 100% ethanol for 5 min twice, dimethylbenzene I for 10 min, dimethylbenzene II for 10 min.     and Sec.10 for details of g_sas calculations.

Figure S24
Projection of the conformations of peptide onto the radius of gyration and the distances of the C-N terminal of KD17 and KD14 on the surfaces with different charge densities. See Sec.9 and Sec.10 for details of these calculations.

Figure S25
The antibacterial activity of the indicated surface in vivo. See Sec.11 for details of the in vivo assay. Zeta Potential (mV) +14.1 -9.9 +5.8 +5.1

Table S2
The abbreviations of the samples.