Screening of bacteria-binding peptides and one-pot ZnO surface modification for bacterial cell entrapment

Short functional peptides are promising materials for use as targeting recognition probes. Toll-like receptor 4 (TLR4) plays an essential role in pathogen recognition and in activation of innate immunity. Here, the TLR4 amino acid sequence was used to screen for bacterial cell binding peptides using a peptide array. Several octamer peptides, including GRHIFWRR, demonstrated binding to Escherichia coli as well as lipopolysaccharides. Linking this peptide with the ZnO-binding peptide HKVAPR, creates a bi-functional peptide capable of one-step ZnO surface modification for bacterial cell entrapment. Ten-fold increase in entrapment of E. coli was observed using the bi-functional peptide. The screened peptides and the simple strategy for nanomaterial surface functionalization can be employed for various biotechnological applications including bacterial cell entrapment onto ZnO surfaces.


Introduction
The introduction of desired properties onto material surfaces is one of the most important techniques for realizing various bionanotechnological applications. Short peptides, derived easily by combination of the 20 natural amino acids, are promising molecules for material surface modication as they exhibit specic recognition of material surfaces. For example, amino acids with carboxylate groups bind preferentially to titanium and silicon owing to their strong electrostatic interactions with the charged surface. 1,2 The strong interaction between gold surfaces and tryptophan is considered to be derived from the planar structure of the amino acid with p-electron bonds characterizing its aromatic ring. [3][4][5] Various peptides with these unique characteristics enabling them to bind to specic materials are used as non-covalently bound linkers for material surface modication. 6,7 In addition to their usage for material surface modications, peptides can bind to various targets including bacterial cells. Because of increasing worldwide concerns about pathogenic bacteria, particularly drug-resistant bacteria, bacterial recognition probes are gaining attention as tools for microbial detection. [8][9][10][11][12] Peptides are promising probes as they are relatively stable and function in ligand-receptor interaction. The SPOT-synthesis technique, a method for customized peptide synthesis on a cellulose membrane established by Frank,13 has been adopted for the screening of functional peptides. 14,15 In particular, the technique is useful for identifying the active site in functional proteins, owing to self-dened peptide library construction. 16,17 Toll-like receptors (TLRs) are a class of proteins with an indispensable role in the immune system. This group of membrane bound-receptors is responsible for recognizing pathogen-associated molecular patterns 18 and activation of innate immunity. 19 Specically, TLR4 recognizes lipopolysaccharides (LPS, endotoxin), eliciting signaling and inducing immune responses against invading bacteria. 20 ZnO is receiving increasing attention because of its antimicrobial activity and its applicability for the synthesis of various nanostructures (e.g., particulates and wires). [21][22][23] For example, continuous mechanical cell lysis and biosensing were developed using bare ZnO nanowires. 24,25 In this study, we conducted array-based screening of bacteria-binding peptides designed around the amino acid sequence of TLR4. We also demonstrated bacterial capturing on the ZnO surface by linking a bacteria-binding peptide to a modular peptide with affinity to ZnO. 26 Surface functionalization is achieved via a one-step process, namely incubation with the peptide solution. The simple ZnO surface functionalization technique demonstrated herein could be utilized for future ZnO-based biotechnological applications, including bioseparation and biosensing.
TLR4 was synthesized. 14,15 The array was used for the binding assay with bacterial cells (Escherichia coli, 2 Â 10 9 cell per mL) and various peptide candidates were identied (ESI Fig. 1 †). To re-evaluate the binding ability of these peptide candidates, peptide arrays comprising the 60 highest binding peptide sequences from the rst screening were constructed and used for the next binding assay. Table 1 lists the top 10 bacteriabinding peptides; they exhibited a high proportion of arginine (R) and lysine (K) residues, high pI, positive charge, and low GRAVY value (hydrophilic). These results are in agreement with previous research because bacterial binding to TLR4 has been suggested to occur via electrostatic interactions between positively charged regions in TLR4 and negatively charged regions in the LPS. 27 Interestingly, the top 10 peptide sequences were present in three regions of TLR4 : GRHIFWRRLRKALL (aa 802-816), KLKSLKRLTFTS (aa 348-360), and RLHKLTLRNN (aa 226-236). KLKSLKRLTFTS and RLHKLTLRNN are located in the leucine-rich repeat domain, which is thought to be responsible for TLR molecule recognition. 14,15 Interestingly, GRHIFWRRLRKALL was also capable of binding to the bacterial cell, even though the peptide was in the cytoplasmic region. This might be ascribed to the presence of the LPS-binding motif (BBXB; B is basic amino acid and X is hydrophobic amino acid) 28 within its sequence, even if the region seems inconsistent with the LPS-binding site of TLR4. Based on bacteriabinding results, the three linear peptides (p402:GRHIFWRR, p175:KLKSLKRL, and p114:RLHKLTLR) showing the highest uorescent intensity in each region were subjected to further investigation.

Dot blot analysis
The binding activity of the screened peptide candidates (p114, p175, and p402) was investigated by LPS dot blot analysis ( Fig. 1). Whereas the negative control and other two candidate peptides (p114 and p175) showed negligible binding activity, peptide GRHIFWRR (p402), showed considerable binding to LPS. Hence, as peptides of the same length revealed different binding to LPS, we concluded that function did not derive simply from the number of amino acids. Accordingly, the unique binding property of each amino acid appears to result from a combination of their physicochemical properties and sequences. As LPS is the major component of the outer membrane of Gram-negative bacteria, peptide candidate p402 could bind to LPS as well as Gram-negative bacteria. On the contrary, the other two peptides could bind to bacterial cells but not to LPS, implying that they might target other cell membrane components such as membrane proteins.

Evaluation of the binding efficiency of a bi-functional peptide on ZnO NPs
To shed further light on the potential of the screened peptide, we investigated the functionalization of nanomaterial using a simple process. ZnO-based nanomaterials have recently received signicant attention for the development of biosensors and photocatalytic anti-microbial materials, 21-23 and introducing bacteria-binding to a material surface would be highly valuable for ZnO-based biotechnological applications. Here, we demonstrated the functionalization of a ZnO surface using a bifunctional peptide (GRHIFWRRGGGHKVAPR) consisting of the screened peptide bound to bacteria and LPS (p402:GRHIFWRR), a ZnO binding peptide (HKVAPR), and a linker peptide (GGG). 26 First, the optimization of peptide concentration was investigated. As shown in Fig. 2; at peptide concentrations exceeding 50 mM, modication efficiency seemed to be saturated as 20 mmol g À1 . Modication using the bi-functional peptide was observed also under the microscope. As shown in Fig. 2ii, uorescence from the uorescein isothiocyanate (FITC)-labeled  bi-functional peptide coincided with sites of ZnO NPs visualized by phase-contrast. At the same time, FITC-labeled bacteriabinding peptide revealed negligible binding to ZnO NPs. It should be noted that the intensity of the uorescently labeled peptide seemed to be consistently stronger for large particles and saturated for ZnO particles at this size and condition. Therefore, the investigated peptides seemed to be homogeneously distributed on the ZnO surface. These data clearly conrmed the binding property of the ZnO-binding peptide HKVAPR when fused with the bacteria-binding peptide (GRHIFWRR).

E. coli entrapment on the ZnO surface using a bi-functional peptide with roles in bacteria and ZnO binding
Once the ZnO-binding property of the bi-functional peptide (GRHIFWRRGGGHKVAPR) was conrmed, bacterial cell capture on the peptide-modied ZnO surface was investigated (Fig. 3i). Fig. 3ii clearly shows the uorescent signals derived from the stained cells bound to the peptide-modied ZnO, whereas unmodied ZnO rarely exhibited any uorescence. In addition, quantitative evaluation based on the uorescent images, revealed that binding efficiency increased approximately 10-fold following peptide modication (Fig. 3iii). These data indicate that the screened peptide possesses bacteriabinding function even when the molecule is fused to the ZnO-binding peptide and modied onto a nanomaterial surface. Furthermore, it should be noted that the ZnO surface modication was conducted using a single step; i.e., incubation with the peptide solution. In this study, peptide functionalization on the ZnO surface and bacterial entrapment were conducted. However, the present conditions would benet from further optimization of peptide sequence (e.g., length of linker sequence) and peptide molecule density based on a detailed analysis of material surface properties; this would allow for the development of more effective bacterial cell entrapment tools.

Conclusions
In conclusion, a peptide array technique was applied for the screening of bacteria-binding peptides based on the sequence of TLR4, a receptor that recognizes LPS within Gram-negative bacterial cells. Among several candidates, the GRHIFWRR peptide showed relatively high binding to LPS. The fusion peptide (GRHIFWRRGGGHKVAPR) consisting of the above peptide, linker (GGG), and ZnO-binding peptide (HKVAPR), revealed ZnO surface modication potential using a singlestep method for bacterial entrapment. This simple strategy for nanomaterial surface functionalization using peptide bifunctionality can be expanded to other nanomaterials for various biotechnological applications, including biosensors.  This journal is © The Royal Society of Chemistry 2018

Materials and methods
Screening of E. coli-binding peptides on the peptide array synthesized by spot technology The peptide array was synthesized using a peptide auto-spotter (ResPep SL; Intavis AG, Köln, Germany) as reported previously. 14,15,29 A cellulose membrane (grade 542; Whatman, Maidstone, UK) was activated using b-alanine as the Nterminal basal spacer. Activated Fmoc amino acids (0.5 M) were spotted on the membrane, using a peptide auto-spotter (MultiPep RSi; Intavis) according to the manufacturer's instructions and in all cases double couplings were applied with some modications. 15,30 Aer addition of the rst residue, Acp(6) were used as additional spacers between the peptides and the cellulose membrane and the remaining amino groups were blocked with 4% acetic anhydride. For each elongation step, the membrane was deprotected with 20% piperidine in N,N 0 -dimethylformamide and subsequently washed thoroughly with N,N 0 -dimethylformamide followed by methanol. Aer the nal deprotection, side-chain-protecting groups were removed with a solution of water, triisopropylsilane, and tri-uoroacetic acid in a volume ratio of 2 : 3 : 95 for 3 h. Finally, the membrane was washed thoroughly with diethyl ether, methanol, phosphate-buffered saline (PBS, pH 7.4) and applied for binding assay. For the rst screening, an 8-mer peptide library containing sequences derived from the 839 amino acid-long human TLR4 (gene ID: 7099) was synthesized. E. coli (BW25113) was cultured in Tryptic Soy Broth at 37 C. In a typical binding assay experiment, 40 mL of PBS containing E. coli at a concentration of 2 Â 10 9 cell per mL was used. The synthesized peptide array was soaked in the E. coli cell suspension and incubated at 25 C for 5 h. When screening for E. coli-binding peptides using the peptide array, a relatively longer incubation time was chosen, allowing for the isolation of a wide range of weak to strong binders. Subsequently, the peptide array was washed with PBS three times, stained by SYTO®9 (nal concentration 1 mM) for 30 min, and washed with PBS three times. The membranes were scanned using a 494 nm excitation and 520 nm emission lter pair (Typhoon FLA 9500; GE Healthcare; Little Chalfont, UK) and were quantied using ImageQuant (GE Healthcare).

Dot blot analysis for the evaluation of LPS binding
A polyvinylidene uoride (PVDF) membrane was soaked in methanol, and then washed with PBS two times. Aer drying, 2 mL of LPS solution (0.1, 0.5, 1, 5, and 10 mg) diluted in 65% ethanol was spotted on PVDF strips and allowed to dry. The strips were blocked using 1% bovine serum albumin in PBS at 25 C for 1 h and washed subsequently in PBS three times. The strips were incubated for 1 h 25 C in a FITC-labeled peptide solution (nal concentration 1 mM) diluted in PBS. All FITClabeled peptides (p114, p175, p402, and AAAA; purity > 90%) were purchased from Sigma-Aldrich (Tokyo, Japan). Aer incubation, strips were washed three times in PBS-Tween 20 (0.05%), scanned, and quantied as above.
ZnO surface modication using a bi-functional peptide A bi-functional fusion peptide (GRHIFWRRGGGHKVAPR) was synthesized, puried with high-performance liquid chromatography (purity > 85%), and conrmed by mass spectrometry. Using ZnO particles (400 mg mL À1 ), ZnO surface modication efficiency was evaluated in the presence of different FITClabeled peptide concentrations (from 0 to 100 mM). Aer 1 h of incubation at 25 C, the particles were washed with PBS ve times. Fluorescence intensity of all washed solutions was measured by a plate reader (PowerScan 4; DS Pharma Biomedical Co., Ltd., Osaka, Japan). Images of peptide-modied ZnO and E. coli capture were taken using a uorescent microscope (DM6B; Leica, Tokyo, Japan).

Entrapment of E. coli cells using a bi-functional peptide
Entrapment of E. coli cells on a peptide-modied ZnO surface was investigated. To fabricate the ZnO surface, ZnO particles (average diameter 25 nm) in milliQ (1 mg mL À1 ) were added to a silicon rubber-encircled hole (diameter 3 mm) attached on a glass slide. By drying up the aqueous solution, a white layer was observed indicating that a ZnO surface was fabricated onto the glass slide. Aer washing the surface by PBS two times, a bifunctional peptide (GRHIFWRRGGGHKVAPR, 100 mM) was added and incubated for 2 h at 25 C. Then, the surface was washed with PBS two times. Prior to the binding assay, E. coli cells were stained with SYTO®9 (150 nM) for 30 min and washed twice with PBS. The stained cells (1.0 Â 10 5 cells, 10 mL) were spotted onto the peptide-modied ZnO surface and incubated for 2 h. The incubation time was shorter during the screening process to eliminate non-specic binding to ZnO surface. Following ve washes with PBS, the binding of peptide-modied ZnO and E. coli was observed using a DM6B uorescent microscope (Leica). Bound cells were counted manually from ve randomly selected areas on the images. As a negative control, ZnO particles without peptide modication were evaluated for this assay.

Conflicts of interest
The authors declare no competing nancial interest.