Identification of a Bacteria-Specific Binding Protein from the Sequenced Bacterial Genome

Abstract

Novel and specific recognition elements are of central importance in the development of a pathogen detection method. Here, we describe a simple method for identifying the cell-wall binding domain (CBD) from a sequenced bacterial genome employing homology search for phage lysin genes. A putative CBD (CPF369_CBD) was identified from a genome of Clostridium perfringens type strain ATCC 13124, and its function was studied with the CBD-GFP fusion protein recombinantly expressed in Escherichia coli. Fluorescence microscopy showed the specific binding of the fusion protein to C. perfringens cells, which demonstrates the potential of this method for the identification of novel bioprobes for specific detection of pathogenic bacteria.

Introduction

Antibodies are the most commonly used biorecognition elements for the detection of pathogen [11]. The sensitivity and specificity of techniques for bacterial detection are primarily dependent on the intrinsic properties of the antibodies. Most antibodies, however, suffer from cross-reactivity and low stability against changes of pH or temperature [22]. In addition, because antibodies are produced by injecting an antigen into a mammal, high production costs and ethical issues are inevitable. Therefore, the demand for alternative bioprobe molecules that fulfill both high specificity and affinity is increasing.

Phages and phage-derived proteins have received growing attention as alternatives to antibodies owing to their host specificity, high affinity, and relatively lower cost of production [18]. In particular, the cell-wall binding domains (CBDs) of phage-encoded peptidoglycan hydrolases (endolysins) showed great potential as promising bioprobes for pathogen detection [9,24]. Several studies have revealed that CBDs show high affinity to their target cells, with a genus- or species-level of specificity [5,19].

Although there has been an increasing interest in the use of CBDs to detect bacterial pathogens as alternatives to currently used antibodies, identifying a novel CBD from a phage endolysin is time-consuming and labor-intensive owing to the following reasons. First, a novel phage has to be isolated and its genome sequenced. Second, the full genome annotation requires skilled personnel and considerable time. Therefore, a better method to find a CBD is needed for the wide use of CBDs as bioprobes. Here, we present a facile and efficient method for identifying a novel CBD from a sequenced bacterial genome. This method relies on the idea that most bacteria harbor phage genes within their genome as prophages or short phage remnants [2]. For example, 16% of the Escherichia coli O157:H7 strain Sakai genome was reported to be prophage elements [7], and in Streptococcus pyogenes, phage-related sequences account for more than 10% of the total genome [1]. These phage genes are often involved in lateral gene transfer and they have a major impact on bacterial genome evolution [3]. Based on this idea, we hypothesized that we could exploit the genome database as a potential reservoir for extracting genes encoding CBDs of various target pathogens. We chose Clostridium perfringens as a model organism because the anaerobic nature of this bacterium makes it relatively difficult to grow compared with aerobic bacteria, which in turn leads to difficulties in isolation of C. perfringens phages. We successfully identified a CBD from a genome (NCBI Gene ID: 4201291) of C. perfringens ATCC 13124 and confirmed its specific binding to C. perfringens.

 

Materials and Methods

Bacterial Culture Conditions

The bacterial strains used in this study are listed in Table 1. The C. perfringens ATCC 13124 strain was used for the isolation of the CBD. All of the Clostridium strains were grown in re-inforced clostridial medium broth at 37℃ under anaerobic condition. Bacillus, Listeria, and Staphylococcus strains were grown in brain-heart infusion broth at 37℃ under shaking. All of the media used in this study were purchased from Difco and used according to the manufacturer’s instructions.

Table 1.ATCC, American Type Culture Collection; NCTC, National Collection of Type Cultures; NCCP, National Culture Collection for Pathogens.

Computational Analysis

The BLAST search program (http://www.ncbi.nlm.nih.gov/BLAST) was used to survey lytic enzymes in the genome of C. perfringens ATCC 13124. We exploited the amino acid sequences from catalytic domains of Psm (endolysin of C. perfringens episomal phage phiSM101; NCBI Accession ID: YP_699978.1) [16,23] and Ply3626 (endolysin of C. perfringens phage phi3626; NCBI Accession ID: NP_612849.1) [27] as query sequences because those two endolysins are the well-characterized N-acetylmuramidase and N-acetylmuramoyl-L-alanine amidase, respectively. The NCBI non-redundant protein sequences were utilized as a reference database, with default parameters. The positions of the selected lytic enzymes within the genome were identified using CLS Genomics Workbench (ver. 3. 61). A conserved domain search within the selected lytic enzymes was conducted using InterProScan [26] and the NCBI Conserved Domain Database [14]. Amino acid sequences of the putative lytic enzymes were aligned with one another or the reported endolysins using ClustalX2 [10]. Three-dimensional structural analysis of proteins was conducted using the PyMol program [4] and the Phyre server [8].

Production and Purification of Recombinant Protein

A gene fragment encoding 140 amino acids from the C-terminal of YP_694826 was predicted as a putative CBD, and the coding sequence was PCR (polymerase chain reaction) amplified from a colony of C. perfringens ATCC 13124 with the following primers: GCGGGATCCGTAAAAAATAATTTTAAATTGTATAATGCAACC ACTAAG (fwd) and GCGAAGCTTCTAAATCTTTTTAACAAA GTCAGCCTTAACAAAA (rev). To confirm and visualize the cell binding activity of the CPF369_CBD, we fused EGFP (enhanced green fluorescent protein) at the N-terminal of CPF369_CBD. The EGFP encoding gene was PCR amplified from pEGFP (Clontech, Palo Alto, CA, USA), and its native stop codon was omitted for translational fusions. The amplified DNA products were cloned into the NdeI/BamHI sites of a modified 10His-pET28a vector (Novagen, Madison, WI, USA), which intentionally inserts an additional four histidine residues to make an N-terminal decahistidine tag. The CBD fragment was inserted into the BamHI/HindIII sites of EGFP-containing 10His-pET28a, creating an in-frame fusion of N-terminal deca-His-tagged EGFP-CBD. The constructed plasmid was transformed into E. coli BL21 (DE3) for protein expression. The clone was grown in LB broth and induced with 0.5 mM isopropyl-β-D-thiogalactoside once an optical density at 600 nm of 0.8 was reached. The induced culture was shaken for 21 h at 18℃. The cells were pelleted, resuspended in a buffer A (200 mM NaCl and 50 mM Tris-Cl (pH 8.0)), and disrupted on ice with the Sonifier 250 (Branson, Danbury, CT, USA) at a duty cycle of 20% and output control of 5. The suspension was centrifuged (21,000 ×g, 1 h, 4℃) and the supernatant was filtered (0.2-μm-pore-size syringe filter; Sartorious, Göttingen, Germany). The soluble protein was purified by immobilized metal affinity chromatography using a Poly-Prep Chromatography column (Bio-Rad, Hercules, CA, USA) packed with Ni-NTA superflow resin (Qiagen, Chatsworth, CA, USA). Buffer B (200 mM Tris-Cl (pH 8.0), 200 mM NaCl, and 250 mM imidazole) was used for protein elution. Eluates were dialyzed against buffer A and stored at -20℃ after addition of 50% (v/v) glycerol.

Bacterial Cell Binding Assay with Fluorescence Microscopy

Binding assays of EGFP-CBD fusion proteins were performed as described before [12]. Briefly, exponentially growing bacterial cells in Dulbecco’s phosphate-buffered saline (DPBS; GenDepot, Barker, TX, USA) were incubated with 0.4 μM EGFP-CBD fusion proteins for 5 min at room temperature. The cells were washed twice with DPBS buffer and observed by epifluorescence microscopy using a DE/Axio Imager A1 microscope (Carl Zeiss, Oberkochen, Germany) and a filter set (excitation 470/40; emission 525/50) for EGFP. The effect of NaCl on the binding of the CBDs to C. perfringens cells was investigated in 20 mM Tris-Cl (pH 8.0) buffer supplemented with 0, 0.05, 0.1, 0.2, 0.5, 1.0, or 2.0 M NaCl. Cells were mixed with the EGFP-CBD proteins, incubated (5 min at room temperature), and centrifuged (16,000 ×g, 1 min). After the supernatant was discarded, cells were resuspended in 200 μl of PBS, and the cell-associated fluorescence was measured by using a SpectraMax i3 multimode microplate reader (Molecular Devices, Sunnyvale, CA, USA) with excitation at 485 nm and emission at 535 nm. The cell density at 600 nm was measured, and the normalization was calculated as whole-cell fluorescence per OD600. The relative binding capacity of the CBD was given in percentage by comparing with the optimal binding condition (the highest measured value). The influence of pH on the CBD binding was assessed using cells resuspended in a universal pH buffer (10 mM KH2PO4, 10 mM Na-citrate, and 10 mM H3BO4) adjusted to different pH values from 4 to 10 [17]. The cell binding assay was carried out as described above.

 

Results and Discussion

Analysis of Lytic Enzymes in C. perfringens ATCC 13124

C. perfringens ATCC 13124, originally isolated from a gas gangrene patient [15], was used as a source for finding a putative CBD within a bacterial genome, as its complete genome sequence is available at the NCBI database (reference sequence: NC_008261.1). We used amino acid sequences of catalytic domains of the selected endolysins for BLAST search because the catalytic domains are conserved relatively higher than the full-length endolysins, which generally contain host-specific CBDs. The BLAST results showed four Psm homologs and one Ply3626 homolog in the genome of C. perfringens ATCC 13124 (Table 2). Among these five putative lytic enzymes, YP_695420 is identical to the well-characterized muramidase, PlyCM [20], and is encoded within the prophage region of the genome containing genes encoding phage terminase, holin, and capsid protein. YP_696011 is almost the same as Psm (96% identity) and resides in another prophage region. The absence of phage-associated genes nearby genes encoding either YP_695964 or YP_696189 suggests that these two lytic enzymes may be bacterial cell wall autolysins. In particular, YP_696189 (putative amidase) showed a 64% amino acid sequence identity with the phage endolysin Ply3626, indicating a close relationship between phage endolysins and bacterial autolysins [13]. YP_694826 seemed to reside in a small prophage remnant region because no phage-related genes, except a holin (YP_694827), existed nearby.

Table 2.Homologs of Psm and Ply3626 in C. perfringens ATCC 13124.

Identification of CBD from CPF369

Among the five lytic enzymes, we analyzed YP_694826 (Gene Symbol: CPF_0369; hereafter referred to as CPF369) further for two reasons. First, the CPF369 protein contains an N-terminal catalytic N-acetylmuramidase (glucoside hydrolase 25 family) domain and two C-terminal tandem repeated SH3_3 domains (Fig. 1A), which are presumed to function in bacterial cell wall binding [25]. Second, we could predict the protein structure of CPF369 more accurately (Fig. 1B) because the crystal structure of Psm endolysin, which shares significant amino acid homology (71%) with the CPF369 (Fig. 1C), is available at the Protein Data Bank (PDB ID: 4KRT) [23]. Although both YP_696011 and YP_695420 (PlyCM) also fulfill these criteria, we preferred to study CPF369 because it has low similarity to the previously reported endolysins [6,16,20,21,27]. As shown in Fig. 1B, we could identify a putative linker region between the N-terminal enzymatic active domain and the C-terminal CBD consisting of two SH3_3 domains. These results made it easier for us to predict the CBD of CPF369 (CPF369_CBD), starting from Val 203 to the stop codon.

Fig. 1.Schematic description of CPF369.

C. perfringens-Specific Binding Activity of CPF369_CBD

The EGFP-CPF369_CBD fusion protein containing the N-terminal decahistidine tag was expressed in E. coli and easily purified to homogeneity by single Ni-NTA affinity chromatography (Fig. 2A). EGFP-CPF369_CBD could bind all C. perfringens strains tested, whereas other genera or species bacteria could not be labeled by the fusion protein (Fig. 2B and Table 1). EGFP alone did not label the C. perfringens cells (Fig. 2C). These results confirmed that CPF369_CBD specifically binds to C. perfringens cells. The effects of NaCl and pH on the binding activity of EGFP-CPF369_CBD fusion protein were investigated (Fig. 3). The optimum NaCl concentration was 50 mM, but the CPF369_CBD retained about 43% of its binding activity even at 2 M of NaCl. The optimum pH condition for binding activity of CPF369_CBD was between pH 7 and 8. The binding activity decreased rapidly at pH 10, resulting in only 17% of binding compared with the value at pH 7.

Fig. 2.C. perfringens-specific binding activity of EGFPCPF369_CBD.

Fig. 3.Influence of NaCl and pH on the CBD binding capacity.

In conclusion, we have identified a C. perfringens-specific CBD from the C. perfringens ATCC 13124 genome through simple homology analysis without phage isolation. We think that it could be a general method for CBD identification from sequenced bacterial genomes. With the rapid increase of bacterial genome data, our method will be useful for discovering novel bioprobes for the detection of pathogenic bacteria.