Sonorensin: an Antimicrobial Peptide, Belonging to the Heterocycloanthracin Subfamily of Bacteriocins, from a New Marine Isolate, Bacillus sonorensis MT93

Abstract

Marine environments are the greatest fronts of biodiversity, representing a resource of unexploited or unknown microorganisms and new substances having potential applications. Among microbial products, antimicrobial peptides (AMPs) have received great attention recently due to their applications as food preservatives and therapeutic agents. A new marine soil isolate producing an AMP was identified as Bacillus sonorensis based on 16S rRNA gene sequence analysis. It produced an AMP that showed a broad spectrum of activity against both Gram-positive and Gram-negative bacteria. The peptide, named sonorensin, was purified to homogeneity using a combination of chromatographic techniques. The intact molecular mass of the purified peptide, 6,274 Da, as revealed by matrix-assisted laser desorption ionization–time of flight (MALDI-TOF), was in agreement with Tricine-SDS-PAGE analysis. A PCR array of primers was used to identify AMP structural genes, which allowed the successful amplification of the related genes from strain MT93. The putative open reading frame of sonorensin was amplified, cloned into the pET-32a(+) vector, expressed as a thioredoxin (Trx) fusion protein in Escherichia coli, and then purified. Sequence alignment analysis revealed that the bacteriocin being reported could belong to new subfamily of bacteriocins, heterocycloanthracin. The peptide indicated its potential as a biocontrol agent or food antimicrobial agent, due to its antimicrobial activity against bacteria such as Listeria monocytogenes and Staphylococcus aureus. This is the first report of the production, purification, and characterization of wild-type and recombinant bacteriocin by B. sonorensis and the first bacteriocin of the heterocycloanthracin subfamily to be characterized.

INTRODUCTION

Antimicrobial peptides (AMPs) are evolutionarily conserved components of the innate immune response, which is the principal defense system for the majority of living organisms, and are found among all classes of life, ranging from prokaryotes to humans (1, 2). Production of AMPs is widespread among diverse bacteria (3). Bacteriocins are ribosomally synthesized AMPs produced by prokaryotes that are bactericidal and/or bacteriostatic against organisms often, but not always, related to the producer strain (4). These metabolites are heterogeneous compounds produced by various Gram-positive and Gram-negative microorganisms (3, 5–7). Bacteriocins are classified into different classes (8). Class I bacteriocins (lantibiotics) are small peptides that undergo extensive posttranslational modification to produce active peptide. Nisin, the most studied bacteriocin, belongs to the class I bacteriocins, which are active against a broad spectrum of food spoilage and pathogenic bacteria, including Listeria monocytogenes (9). Class II bacteriocins are heat-stable, low-molecular-weight, membrane-active peptides. Members of class III are large, heat-labile proteins that may be bacteriolytic (subclass IIIa) or nonlytic (subclass IIIb), and class IV encompasses cyclic peptides (8).

Lee et al. described the gene clusters that produce a subfamily of bacteriocins with thiazole or oxazole heterocyclic rings (10). Haft reported a new subfamily of putative thiazole-containing heterocyclic bacteriocins of the genus Bacillus and named this category of bacteriocins heterocycloanthracin (11). These bacteriocins are synthesized as protoxins with N-terminal leader peptides that show homology from one precursor to another. The C-terminal region contains a low-complexity sequence, often repetitive and rich in Cys, Ser, or Thr (11). Bacteriocins are reported to be synthesized as biologically inactive peptides (precursors) containing an N-terminal leader peptide that is cleaved off during maturation and exported by a dedicated ABC transporter (4, 12).

Bacteriocins produced by Gram-positive bacteria, especially Bacillus spp., have attracted worldwide attention because of their great potential as food preservatives, therapeutic agents, and biosurfactants (in the case of lipopeptides) (13, 14, 15). Though many bacteriocins or bacteriocin-like inhibitory substances (BLIS) in the genus Bacillus have been reported (e.g., cerein, produced by Bacillus cereus Gn105 [16], and cerein 7, produced by B. cereus Bc7 [17]), screening of naturally occurring habitats in different parts of the world could lead to isolation of peptide-producing microorganisms having potential antimicrobial activity.

One of the strategies to determine the identities of newly found bacteriocins has been random testing by PCR with primers targeting one or more suspect bacteriocins (18–20). The use of PCR techniques can readily identify bacteriocin genes in bacteriocinogenic bacteria. Yi et al. used a variation of colony PCR to facilitate the screening of colonies for class IIa bacteriocin-producing lactic acid bacteria (LAB) (21). They used degenerate primers based on the conserved N-terminal regions found in many class IIa bacteriocins and specific downstream primers to design PCRs, resulting in large amplimers specific for pediocin, enterocin, and plantaricin. Wieckowicz et al. used a PCR assay with a custom panel of bacteriocin-related primers to screen metagenomic DNA preparations obtained from the microflora of Polish artisanal cheeses and were able to identify class IIa bacteriocin sequences (22).

In this report, we describe the use of an AMP PCR array to detect the presence of AMP-related genes in isolates showing antimicrobial activity. The genomic DNA from these bacteria was subjected to an AMP-specific PCR array in individual reactions with primers representing different structural genes of class II bacteriocins and other AMPs. Sequencing of the amplimers, followed by sequence analysis, helped to determine if the sequences had identity with others currently in GenBank or were unique sequences. The gene encoding the putative bacteriocin was identified, cloned, and expressed. The bacteriocin was purified and characterized.

MATERIALS AND METHODS

Materials.

Six indicator strains, Bacillus subtilis MTCC 121, Staphylococcus aureus MTCC 1430, Escherichia coli MTCC 1610, Pseudomonas aeruginosa MTCC 1934, Vibrio vulnificus MTCC 1146, and L. monocytogenes MTCC 839, were procured from the Microbial Type Culture Collection (MTCC), Chandigarh, India. E. coli Rosetta 2 was used as the expression host. E. coli DH5α was used as the host for subcloning and plasmid amplification; pET-32a(+) was used as the expression plasmid. Restriction enzymes NcoI and XhoI and T4 DNA ligase were purchased from New England BioLabs Inc. (Ipswich, MA, USA). The enzymes SDS, tricine, and Diaion HP-20 (Supelco) resin were purchased from Sigma-Aldrich Inc. (St. Louis, MO, USA). Microtiter plates were procured from Nunc (Nalge Nunc International, Roskilde, Denmark). All other reagents and medium components used were either analytical grade or of the highest purity grade available in India. The sequence alignment studies were carried out using the Clustal W program of MEGA 5.2.

Isolation and growth media.

Marine soil samples were collected from Parangipettai, India (11°30′N, 79°47′E), at the Vellar riverbank. The soil samples were serially diluted with sterile water and plated onto sterile tryptone soya broth (TSB) (Hi-Media, Mumbai, India) agar medium containing 1.5% agar at pH 7.2. Following incubation of the plates at 30°C for 4 days, colonies that inhibited the growth of surrounding colonies, as evidenced by clear zones, were isolated, purified by subculturing, and preserved at −70°C for further study. A total of 470 isolates were selected for further screening for antimicrobial activity against six indicator strains using the agar disk diffusion method as described by Kimura et al. (23). Inhibitory activity against at least one indicator strain was detected for 170 isolates. Among these, a strain designated MT93 exhibited broad-spectrum antibacterial activity by inhibiting the growth of all six indicator strains.

Bacterial identification.

Isolate identities were determined by sequencing 16S rRNA genes. The genomic DNA of the isolate was extracted using a commercial DNA extraction kit (AxyPrep Bacterial Genomic DNA Miniprep kit; Axygen Biosciences, USA). PCR amplification of the 16S rRNA gene was done using the universal primers 8-27f (5′-AGAGTTTGATCCTGGCTCAG-3′) and 1492r (5′-TACGGYTACCTTGTTACGACTT-3′). The targeted DNA sequence was amplified in a thermocycler using a standard protocol. The PCR products were examined by 1% agarose gel electrophoresis and then recovered, purified using a gel extraction kit (HiYield gel/PCR DNA minikit; Real Biotech Corporation [RBC], Taiwan), and sequenced with an ABI 310 genetic analyzer (Applied Biosystems). Identification of phylogenetic neighbors and the calculation of pairwise 16S rRNA gene sequence similarities were achieved using the EzTaxon server (http://www.eztaxon.org/) (24).

Determination of antimicrobial activity of MT93.

Strain MT93 was grown for 34 h in sterile TSB (Hi-Media, Mumbai, India) in an incubator shaker at 30°C and 200 rpm. Subsequently, the cells were removed by centrifugation (10,000 × g for 10 min at 4°C). The antimicrobial activity of the cell culture supernatant was detected and tested against all six indicator strains. A growth curve up to 60 h was plotted for strain MT93 to examine the bacteriocin production at different stages of growth. Samples were withdrawn at regular intervals, and an activity assay, in triplicate, was performed with the cell culture supernatant. The bacteriocin titer was determined by the serial 2-fold dilution method described by Mayr-Harting et al. (25). The antimicrobial activity was defined as the reciprocal of the highest dilution yielding a clear zone of growth inhibition of the indicator strain and expressed as activity units (AU) per milliliter.

Bacteriocin production and purification.

An isolated single colony was used to inoculate a 250-ml flask containing 50 ml of TSB medium and incubated in a rotary shaker at 30°C and 200 rpm for 16 h. This culture was used to inoculate 2-liter Erlenmeyer flasks containing 500 ml of TSB, and the flasks were incubated in a rotary shaker at 30°C and 200 rpm for 34 h. Subsequently, the cells were separated by centrifugation at 8,000 × g and 4°C for 20 min, and the cell supernatant was used as crude bacteriocin. Initial purification of the bacteriocin was carried out by mixing the cell supernatant with 1% (wt/vol) Diaion HP-20 (Supelco) resin, followed by shaking at 30°C and 200 rpm for 2 h. The resin was washed with 10% methanol to remove nonspecific and loosely bound proteins. The bacteriocin was eluted with 100% methanol. The methanol was evaporated using a Rotavapor R-200 (Buchi) device, and the dried peptide content was dissolved in Milli-Q water. This extract was applied on a manually packed SP Sepharose column preequilibrated with 10 mM sodium phosphate buffer (pH 7.2) and connected to an ÄKTA Explorer chromatography system (GE Healthcare, USA), followed by elution with a 0- to 1.0-M linear gradient of sodium chloride. Active fractions, concentrated by lyophilization, were loaded onto a manually packed Sephadex G-25 column, connected to an ÄKTA Explorer, that was previously equilibrated with 20 mM sodium phosphate buffer (pH 7.2) containing 20 mM NaCl. The elution was done in the same buffer at a flow rate of 0.3 ml/min and monitored through a UV detector (220 nm). The active fractions were freeze-dried, redissolved in Milli-Q water, and injected onto a reversed-phase (RP) column (Waters Spherisorb S100DS2 20- by 250-mm semipreparative column; part no. PSS832595; Waters Corporation, USA) equilibrated with 5% acetonitrile (ACN)–0.1% trifluoroacetic acid (TFA)–water, and the concentration of the ACN in the eluted solvent was increased from 10% to 60% (vol/vol) over 30 min using a linear gradient at a flow rate of 5 ml/min. The absorbance was monitored at 220 nm with a Waters 2489 UV-visible light (Vis) detector, and the fraction corresponding to the peak of the high-performance liquid chromatography (HPLC) chromatogram (purified bacteriocin) was collected. Following freeze-drying, the purified bacteriocin was redissolved in water to a concentration of 0.5 mg/ml for its characterization and determination of the antimicrobial activity. The purified bacteriocin was run on 16.5% Tricine-SDS-PAGE gels (26), and the in-gel activity of the bacteriocin was determined as described by Bouksaim et al. (27). Matrix-assisted laser desorption ionization (MALDI) was also used to characterize the purified AMP, as described by Mandal et al. (28).

PCR array for AMP structural genes.

PCR was used to detect the presence of genes encoding known class II bacteriocins and other AMPs. Total DNA of the isolate was used as a template for PCR amplification. The primers used for the PCR amplification of genes related to AMP production and the annealing temperature of each primer pair are listed in Table 1. The PCR products were recovered and purified using a gel extraction kit (HiYield gel/PCR DNA minikit; RBC) and then sequenced. The nucleotide sequence information was analyzed by using the BLAST algorithm at the National Center for Biotechnology Information (http://www.ncbi.nlm.nih.gov/blast).

TABLE 1.

Oligonucleotide primers used to detect bacteriocin genes in strain MT93

Primer	Expression product	Sequence (5′→3′)a	Gene(s)	Expected product size (bp)	Annealing temp (°C)	Reference
oligo 48	Bacteriocin	TAYGGIAAYGGIGTITAYTG	Class IIa bacteriocin genes	3,000	51	30
oligo 105		CYTCDATNGCRTTRTC
Bli F	Lichenicidin	GGAAATGATTCTTTCATGG	Bli04127	215	60	50
Bli R		TTAGTTACAGCTTGGCATG
Thu F	Thuricin	GTAGGTCAAATGGAAACAC	tucA1, tucA2, tucA3	589	52	50
Thu R		TTAACTTGCAGTACTAGCTC
ITUC F	Iturin	GGCTGCTGCAGATGCTTTAT	ituC	423	60.1	51
ITUC R		TCGCAGATAATCGCAGTGAG
SRFA F	Surfactin	TCGGGACAGGAAGACATCAT	srfAA	201	60.4	51
SRFA R		CCTCTCAAACGGATAATCCTGA
Coa F	Coagulin	GGTGGTAAATACTACGGTAATGGGGT	coaA	600	66	52
Coa R		GTGTCTAAATTACTGGTTGATTCGT
Sbo F	Subtilosin	GGTTGTGCAACATGCTCGAT	sbo	300	58	52
Sbo R		CTCAGGAAGCTGGTGAACTC
Son F	Sonorensin precursor	ATGAAGGGGGATATTAAAAT	Sonorensin precursor	330	58	This study
Son R		CTACATGCCTCCGTGCCA
Son Ap F	Sonorensin active peptide	CATGCCATGGATCCGTGTTGGAGCTGTATGGGTCAC	Sonorensin bacteriocin	171	58	This study
Son Ap R		CCGCTCGAGCTACATGCCTCCGTGCCA

^aThe underlined sequences in the Son Ap F and Son Ap R primers are the restriction sites of the NcoI and XhoI enzymes, respectively.

Cloning of the bacteriocin-encoding gene.

The 330-bp gene encoding the putative bacteriocin was identified in the PCR array. The primers corresponding to the 5′ and 3′ ends of the putative active bacteriocin gene sequence were designed and synthesized (Table 1). NcoI and XhoI restriction sites were contained upstream of the forward primer and downstream of the reverse primer, respectively. To facilitate chemical cleavage, the site for formic acid cleavage was contained in the forward primer. The bacteriocin gene was introduced downstream of the formic acid cleavage sequence. In addition, between the thioredoxin (Trx) tag and the bacteriocin-coding sequence there was a His tag, which was beneficial for the fusion protein separation by Ni²⁺-chelating affinity chromatography and recognition of a formic acid site to facilitate cleaving of the fusion protein and release of the mature bacteriocin peptide. The bacteriocin PCR-amplified product and pET-32a(+) vector were digested with NcoI and XhoI and gel purified. After that, the PCR product was ligated into the pET-32a(+) vector with the Trx gene to obtain an in-frame fusion gene. The resulting plasmid, pET-32a-bacteriocin, was transformed into E. coli DH5α, and recombinant E. coli DH5α cells were selected on ampicillin-containing Luria-Bertani (LB) plates and screened by the colony PCR method. The recombinant plasmid was isolated from the positive clones and sequenced to ensure that the coding sequence of pET-32a-bacteriocin was correct and in frame with the Trx gene.

Expression and purification of recombinant fusion protein.

The recombinant plasmid was transformed into E. coli strain Rosetta 2 for expression. A single colony containing the recombinant plasmid was inoculated into 20 ml LB medium supplemented with ampicillin (100 μg/ml) and chloramphenicol (34 μg/ml) and incubated for 16 h at 30°C and 200 rpm. This culture was used to inoculate 2-liter Erlenmeyer flasks containing 500 ml of LB medium, followed by incubation in a rotary shaker at 37°C and 200 rpm until the cell optical density at 600 nm (OD₆₀₀) reached 0.6 to 0.7. Induction was initiated by 0.2 mM IPTG (isopropyl-β-d-thiogalactopyranoside), and the flasks were further incubated at 16°C and 200 rpm for 16 h. Following centrifugation (10,000 × g for 10 min at 4°C), the cells were separated from the supernatant, resuspended in 10 ml of buffer (20 mM Tris-HCl, 300 mM NaCl, 10% glycerol, and 10 mM imidazole, pH 7.9), and lysed by sonication (ultrasonic cell crusher) at 300 W for 20 cycles (30 s on, 30 s off) and kept in an ice-water bath. The supernatant of the cell lysate resulting from centrifugation at 14,000 × g for 60 min was applied to an Ni²⁺-chelating column packed with 3 ml of Ni²⁺-chelating resin that had been previously equilibrated with the same buffer. The column was washed with 20 mM imidazole at a flow rate of 1 ml/min, and elution was done with 200 mM imidazole. The fractions were collected and applied to 15% SDS-PAGE, and a bicinchoninic acid (BCA) protein assay was used for quantitative analysis of proteins. To remove the imidazole, which was necessary for further purification, buffer exchange with 20 mM Tris-HCl, 300 mM NaCl, pH 7.9, was carried out using Amicon centrifugal devices (3,000 molecular weight cutoff [MWCO]).

Peptide release and purification of recombinant sonorensin.

The bacteriocin-containing fusion protein was cleaved using 50% (vol/vol) formic acid at 50°C for 24 h, and the mixture was then lyophilized to remove the formic acid. The His-tagged carrier and residual undigested fusion proteins were rapidly removed by using Ni²⁺ affinity chromatography. The unbound material was collected and desalted by dialyzing the sample against 20 mM Tris HCl buffer, pH 7.9. It was then subjected to reversed-phase HPLC using the same protocol as for native bacteriocin; 16% Tricine-SDS-PAGE of recombinant bacteriocin (r-bacteriocin) and antimicrobial activity assays were carried out as discussed previously.

Effects of temperature, pH, hydrolytic enzymes, organic solvents, surfactants, and reducing agents on bacteriocin activity.

The sensitivity of the purified bacteriocin to temperature, pH, hydrolytic enzymes, organic solvents, surfactants, and reducing agents was evaluated. The purified bacteriocin samples, in triplicate, were incubated at pH 2.0, 3.0, 5.0, 7.0, 8.0, 10.0, 11.0, and 12.0 and temperatures of 45°C, 60°C, 75°C, and 100°C for 30 min and at 121°C for 20 min to determine their pH and temperature stability. In another set of experiments, the purified bacteriocin was treated with various enzymes (including lysozyme, proteinase K, α-amylase, pepsin, catalase, lipase, and RNase A at final concentrations of 2 mg/ml). Samples were boiled for 5 min to inactivate the added enzymes before the analysis of bacteriocin activity. To study the effects of organic solvents, the purified bacteriocin was incubated with various organic solvents (ethanol, methanol, isopropanol, acetone, ethyl acetate, chloroform, acetonitrile, and butanol) at a final concentration of 10% (vol/vol), along with appropriate controls, at room temperature for 1 h. After incubation, the solvents were removed by vacuum concentration at 45°C for 1 h, and then residual antimicrobial activity was determined. The effects of surfactants, such as SDS (1% [wt/vol]), Tween 80 (1% [wt/vol]), and Triton X-100 (1% [wt/vol]), and reducing agents, such as dithiothreitol (DTT) (10 mmol/liter) and β-mercaptoethanol (10% [vol/vol]), were studied by addition of the reagents to the purified bacteriocin, followed by incubation at 25°C for 2 h. The samples were then assayed for antimicrobial activity.

Determination of MICs.

The MICs of bacteriocin for different indicator strains were evaluated by using a microtiter broth dilution method as described by Stalons and Thornsberry (29). The lowest concentration that inhibited the growth of test strains and did not show any increase in absorption after 48 h was considered the MIC.

UV spectroscopy.

The UV absorbance spectrum from 190 to 600 nm at 0.5-nm resolution was recorded using a Shimadzu UV 2550 double-beam spectrophotometer (Shimadzu Corporation, Tokyo, Japan).

Nucleotide sequence accession numbers.

The 16S rRNA gene sequence of MT93 has been submitted to the EMBL nucleotide sequence database under the accession number HF944961.1. The gene sequence of sonorensin has been submitted to GenBank under the accession number KF493891.

RESULTS

Isolation and identification of microorganisms producing antimicrobial activity.

A total of 470 bacterial strains were obtained from marine soil collected from Parangipettai, India (11°30′N, 79°47′E). Screening of the antimicrobial activity of these 470 isolates was carried out against six indicator strains: B. subtilis MTCC 121, S. aureus MTCC 1430, E. coli MTCC 1610, P. aeruginosa MTCC 1934, V. vulnificus MTCC 1146, and L. monocytogenes MTCC 839. Inhibitory activity against at least one indicator strain was detected for 170 isolates. Among these, a strain designated MT93 exhibited broad-spectrum antibacterial activity by inhibiting the growth of all six indicator strains. BLAST analysis of the 16S rRNA gene sequence of MT93 revealed significant identity (98.9%) with Bacillus sonorensis. Further, the neighbor-joining phylogenetic tree constructed with 16S rRNA gene sequences of other members of the genus Bacillus confirmed that strain MT93 belonged to the genus Bacillus (data not shown), and hence, strain MT93 was designated B. sonorensis strain MT93 and the antimicrobial agent, bacteriocin, produced by the organism was named sonorensin.

Cell culture supernatants collected at different time intervals during the growth of strain MT93 were used to perform antimicrobial activity assays. The growth curve of MT93 was plotted and revealed initiation of sonorensin production after 18 h of growth (data not shown). However, there was a significant increase in sonorensin production between 20 and 34 h, and the antimicrobial activity remained constant thereafter, as measured from the zone of inhibition of the indicator strain.

Purification of native sonorensin.

The sonorensin present in the cell culture supernatant, obtained after 34 h of culture growth, was extracted by affinity chromatography using Diaion HP-20 resins and purified by sequential SP-Sepharose and Sephadex G-25 chromatography. The sample fractions exhibiting antimicrobial activity against all six indicator strains were subjected to semipreparative reversed-phase HPLC (Fig. 1), and the purified sonorensin was found to be positive for antimicrobial activity (Fig. 1, inset). The intact mass of purified sonorensin determined by MALDI-time of flight (TOF) spectroscopy revealed its size to be 6,274 Da (Fig. 2), which was in good agreement with Tricine-SDS-PAGE data for the peptide (Fig. 2, inset B). Zymographic analysis of bacteriocin activity resulted in inhibition of the indicator strain corresponding to the same band present in the gel stained with Coomassie blue, with a molecular mass of around 6.5 kDa (Fig. 2, inset A).

FIG 1.

RP-HPLC elution profile of native MT93 bacteriocin purification. Elution was performed with a 10 to 60% acetonitrile-H₂O linear gradient over a 30-min period. The active fraction was eluted at 9 min. (Inset) Inhibition activity of the HPLC-purified peptide.

FIG 2.

Mass spectrum of the pure native bacteriocin fraction. (Inset) Tricine-SDS-PAGE of purified bacteriocin and its in-gel activity. (A) Direct overlay of SDS-PAGE gel demonstrating a clear inhibition zone against a test strain of B. subtilis (indicated by an arrow). (B) Left lane, purified peptide fraction collected from an RP-HPLC run; right lane, polypeptide SDS-PAGE molecular weight standard.

Amplification of AMP genes by PCR.

Specific PCR primers were designed to obtain the bacteriocin structural gene (Table 1). Our preliminary examination indicated that the sonorensin produced by MT93 could belong to class II, because it was shown to be active against Listeria. Class IIa bacteriocins have a conserved amino acid sequence, YGNGV, in the amino-terminal region and LDNAIE in histidine kinases (30). Based on these amino acid sequences, two primers, oligo 48 and oligo 105 (Table 1), were synthesized, and amplification PCR was carried out with total genomic DNA of MT93 as the template. An approximately 3-kb DNA fragment was amplified and sequenced (data not shown), and a 330-bp sequence (Fig. 3) was identified that showed significant similarity to a putative thiazole-containing heterocyclic bacteriocin of Bacillus licheniformis ATCC 14580 (accession no. YP_006712426.1). The amino acid sequence of sonorensin was aligned with the sequences of the putative bacteriocins of the other Bacillus strains using the Clustal W program of MEGA 5.2 (Fig. 4), which revealed homology with the leader sequences of protoxins from various Bacillus strains. A 159-bp sequence was identified as the gene encoding the leader sequence of the putative bacteriocin, which was predicted to be cleaved off during the posttranslational maturation of the bacteriocin (Fig. 3). The remaining 171-bp gene sequence (6,225 Da, assuming all cysteines are reduced) was predicted to encode the putative active bacteriocin. The amino acid sequence showed that the bacteriocin contained a repeated Cys-Xaa-Xaa motif (where Xaa is any amino acid) at the C terminus. Previously, Haft had reported a strain-variable bacteriocin with a repeated Cys-Xaa-Xaa motif-containing N-terminal leader sequence and C-terminal bacteriocin sequence (11).

FIG 3.

Nucleotide sequence of the open reading frame (ORF). Shown is the nucleotide sequence of the protoxin sonorensin gene encoding the indicated amino acids. The arrow shows the site of cleavage during the maturation of the active peptide. The nucleotide and amino acid sequences of the active peptide, Sonorensin, are underlined. The stop codon is designated by an asterisk.

FIG 4.

Sequence alignment of protoxin sequences. (A) Alignment of N-terminal leader sequences of putative bacteriocins of some Bacillus spp. (B) Alignment of C-terminal putative active bacteriocins of Bacillus spp.: B. cereus E33L (sequence 1), B. cereus ATCC 14579 (sequence 2), B. thuringiensis serovar Konkukian strain 97-27 (sequence 3), B. cereus AH187 (sequence 4), B. weihenstephanensis KBAB4 (sequence 5), B. cereus G9842 (sequence 6), B. cereus B4264 (sequence 7), B. cytotoxicus NVH 391-98 (sequence 8), B. licheniformis ATCC 14580 (sequence 9), and B. sonorensis MT93 (sequence 10). The sequence used in this study (sequence 10) was aligned with the other protoxin sequences using Clustal W. The N-terminal leader sequence showed significant similarity to the other leader sequences of this new group. However, the C-terminal active sonorensin sequence revealed that it is a different bacteriocin. The asterisks denote amino acids that are present at the same positions in all sequences under study. The color codes are the default settings for amino acids in Clustal W.

The primers Son F and Son R (Table 1), specific for this 330-bp sequence, were designed, and the bacteriocin gene was successfully amplified. Also, PCR analysis was performed using primer pairs specific for known Bacillus AMPs (the primers used in this study are listed in Table 1). The primers specific for the srfAA gene for detecting the presence of the synthetase gene cluster of nonribosomally synthesized lipopeptides yielded an amplicon that revealed 75% homology to the fenD gene (fengycin nonribosomal peptide synthetase) of Bacillus amyloliquefaciens subsp. plantarum (HE617159.1). This suggested that MT93 could have harbored the synthetase gene cluster of nonribosomally produced peptides. Recently, Adimpong et al. have predicted the presence of nonribosomal lipopeptide-biosynthesizing gene clusters in B. sonorensis strain L12 (31). The primers for the Bli04127, ituC, sbo, coaA, and thu genes failed to generate appropriate-size amplicons. This indicated the absence of the associated gene clusters from strain MT93.

Cloning and expression of sonorensin.

The primers corresponding to the 171-bp sonorensin-encoding gene were synthesized. The gene was amplified and cloned into the pET-32a(+) vector to be expressed as a Trx-fused protein. All sonorensin gene expression experiments were done using E. coli Rosetta 2 as a host in order to overcome the problem of rare codons. The whole-cell lysates were analyzed by 15% SDS-PAGE (Fig. 5A) and compared to the negative-control E. coli Rosetta 2 pET-32a(+): recombinant expression products revealed an approximately 24-kDa Trx-sonorensin target band at the expected position after staining with Coomassie brilliant blue G-250.

FIG 5.

SDS-PAGE analysis of fusion protein expression. (A) 15% SDS-PAGE. Left lane, induced empty pET-32a(+) control; right lane, 0.2 mM induced recombinant pET-32a(+). (B) 15% SDS-PAGE. Lane 1, uninduced control; lane 2, 0.2 mM induced fusion protein; lane 3, unbound pass through of Ni-NTA; lanes 4 and 5, wash of Ni-NTA with 20 mM imidazole; lanes 6, 7, 8, and 10, elutions with 200 mM imidazole; lane 9, low-molecular-weight marker. (C) 15% SDS-PAGE. Lane 1, fusion protein after formic acid cleavage; lane 2, fusion protein before formic acid cleavage. (D) 16.5% Tricine-SDS-PAGE. Lane 1, polypeptide SDS-PAGE molecular weight standard; lane 2, purified recombinant sonorensin.

The recombinant fusion protein consisted of the following regions sequentially from its N terminus: the Trx domain (consisting of 109 residues), the linker sequence of pET-32a(+) with a His tag, and formic acid recognition sites. The His-tagged recombinant protein was purified from lysate using nickel-nitrilotriacetic acid (NTA) chromatography, and the Trx-sonorensin fusion protein was eluted from the column with 200 mM imidazole (Fig. 5B). Following buffer exchange, the desalted fusion protein was subsequently cleaved by formic acid, which resulted in two bands on 15% SDS-PAGE corresponding to approximately 6.3 kDa of sonorensin and approximately 17.5 kDa of the fusion partner (Fig. 5C). After the cleavage, the sample was reapplied to the nickel-NTA resin, and the unbound material containing the free sonorensin was collected and run on Tricine-SDS-PAGE gels (Fig. 5D). On further purification of sonorensin using semipreparative RP-HPLC, a pattern of elution similar to that in the case of native sonorensin was observed (data not shown).

Characterization of sonorensin.

The temperature stability studies of sonorensin revealed that it was stable at low temperature and could stand treatments up to 100°C for 30 min without losing its antimicrobial activity, although the activity was totally lost by autoclaving sonorensin preparations (Table 2). The treatment of sonorensin at pH values in the range of 2 to 12 did not affect its antimicrobial activity. The proteinaceous nature of sonorensin was established by its sensitivity to proteinase K and pepsin at the tested concentrations, whereas lysozyme, lipase, catalase, and RNase A did not affect its activity (Table 2). Investigations of the effects of several chemicals on the antimicrobial activity of sonorensin indicated its activity was unaffected by treatment with various organic solvents at 10%; different surfactants, like SDS, at concentrations of 1% (wt/vol); and the reducing agents tested. However, an increase in the antimicrobial activity was observed when sonorensin was treated with 1% Triton X-100 and Tween 80 (Table 2). The MIC assay for Gram-positive and Gram-negative bacteria with purified sonorensin revealed both L. monocytogenes and V. vulnificus to be highly sensitive to the peptide, as they were found to be inhibited at lower MIC values—1 ± 0.1 μg/ml and 25 ± 1.5 μg/ml, respectively—than B. subtilis, S. aureus, P. aeruginosa, and E. coli, whose MIC values were 50 ± 1 μg/ml, 45 ± 1.5 μg/ml, 43 ± 1.15 μg/ml, and 96 ± 0.75 μg/ml, respectively (Fig. 6). On examining the UV absorption spectrum of the purified sonorensin between 190 and 600 nm, the purified peptide was observed to have absorbance maxima at 210 nm and 280 nm, and there was no appreciable absorbance above 300 nm (data not shown). A shoulder at 210 nm to 220 nm corresponded to characteristic absorption of peptide bonds (32).

TABLE 2.

Effects of temperature, organic solvents, enzymatic treatment, and environmental conditions on MT93 bacteriocin activity

Parameter	Type or value	Activitya
Control		15
Enzymatic treatment	Proteinase K	0
	Pepsin	0
	RNase	15
	Lysozyme	15
	Lipase	15
	α-Amylase	15
	Catalase	15
Organic solvents (10% [vol/vol])	Acetone	15
	Chloroform	15
	Acetonitrile	15
	Ethanol	15
	Isopropanol	15
	Butanol	15
	Methanol	15
Surfactants (1% [wt/vol])	Triton X-100	18
	Tween 80	18
	SDS	15
Reducing agents	DTT (10 mmol liter−1)	15
	β-Mercaptoethanol (10% [vol/vol])	15
Temp (°C)	45	15
	60	15
	75	15
	100	15
Autoclaving	121°C, 20 min	0
Storage	4°C (30 days)	15
	−20°C (30 days)	15
pH	2.0	15
	3.0	15
	5.0	15
	7.0	15
	8.0	15
	10.0	15
	11.0	15
	12.0	15

^aActivity was determined as the diameter of the zone of growth inhibition, in millimeters (average of triplicates).

FIG 6.

Antimicrobial activity (MIC) of sonorensin. The MIC assay for Gram-positive and Gram-negative bacteria with purified sonorensin using microtiter plates revealed that L. monocytogenes and V. vulnificus are highly sensitive. The error bars indicate standard deviations.

DISCUSSION

Screening of naturally occurring habitats in different parts of the world is anticipated to result in isolation of microorganisms producing potential antimicrobial agents. Here, a potential antimicrobial-activity-producing isolate identified as B. sonorensis by its genotypic characteristics (designated MT93) was selected for further characterization, as it exhibited broad-spectrum antibacterial activity. An interesting finding with this isolate was the ability of its bacteriocin to inhibit the growth of a Gram-negative bacterium, E. coli, as bacteriocins produced by Gram-positive strains are mostly inhibitory to Gram-positive strains and are less effective against Gram-negative strains (33, 34).

PCR analysis using primer pairs specific for known bacteriocins showed the presence of a 330-bp sequence having significant similarity to the putative thiazole-containing heterocyclic bacteriocin of B. licheniformis ATCC 14580 (accession no. YP_006712426.1). The Clustal W alignment of the N terminus of sonorensin revealed homology with the leader sequences of protoxins from various Bacillus strains. This indicated that sonorensin, produced initially as a protoxin, was posttransitionally modified to active bacteriocin (the cleavage site is indicated by an arrow in Fig. 3). This led to the suggestion that sonorensin could belong to a new subfamily of bacteriocin, heterocycloanthracin (named by Haft), a group of putative peptides containing oxazole and/or thiazole heterocycles (11). However, these bacteriocins have not been purified or demonstrated to have antibacterial activity so far (35). The C-terminal putative active bacteriocin of B. licheniformis ATCC 14580 also consists of a repetitive Cys-Xaa-Xaa sequence, but it contains only 33 amino acids, whereas the active sonorensin contains 57 amino acids, signifying that sonorensin is a different bacteriocin (Fig. 4). Moreover, the predicted molecular mass of the active sonorensin corresponds to the theoretical mass obtained by SDS-PAGE and MALDI-TOF analyses.

The PCR array indicated that, besides producing the bacteriocin, strain MT93 could have harbored the gene cluster for nonribosomally synthesized lipopeptides. However, this reported sonorensin is not likely to be a lipopeptide, as the peptide remained active even after treatment with lipase (Table 2). Strains possessing genes for more than one AMP have also been previously reported (36).

It is essential to obtain bacteriocin in a purified form for its further characterization and the use of various chromatographic techniques (37, 38), including gel permeation and affinity chromatography, after its ammonium sulfate precipitation was reported (39–43). Compared to ammonium sulfate precipitation, which precipitates most of the proteins, few studies have reported the utilization of hydrophobic interaction of resins to extract selective peptides (44, 45). In the present case, the extraction of native sonorensin from a 34-h culture supernatant using Diaion HP-20 resins was followed by cation-exchange chromatography, gel filtration chromatography, and RP-HPLC. Separation of sonorensin by Diaion HP-20 resins suggested that the sonorensin molecule is hydrophobic in nature and in this sense is similar to many other bacteriocins. The purified native sonorensin showed a single band at around 6.5 kDa on Tricine-SDS-PAGE; this was further confirmed by MALDI-TOF, which showed a molecular mass of 6,274 Da. In addition, sonorensin, cloned and expressed as a fusion protein with thioredoxin in E. coli, revealed an approximately 6.5-kDa molecular mass on cleavage from the fusion partner. Investigations of the MIC values of purified sonorensin showed L. monocytogenes and V. vulnificus to be highly sensitive to the AMP, and its UV spectrum suggested that it was a peptide antibiotic that produces absorbance maxima at 210 to 230 and 270 to 280 nm (46, 47). The purified sonorensin was also found to be heat stable, stable at low temperature, biologically active over a wide pH range, and not affected in the presence of organic solvents, surfactants, and reducing agents. However, the antimicrobial activity of sonoresin was lost upon exposure to proteases, indicating the proteinaceous nature of the bacterial peptide. Our results are consistent with those reported for other bacteriocins, such as bacillocin 490 from B. licheniformis 490/5 (48), which were inactivated by proteinase K. Furthermore, treatment of sonoresin with RNase, lysozyme, and α-amylase (Table 2) did not affect its antimicrobial activity, suggesting that a peptidic compound is only essential for the antibacterial activity of an AMP. Interestingly, the activity of sonorensin increased in the presence of a nonionic detergent, such as Triton X-100, which could be due to the hydrophobic nature of the peptide, which might have resulted in the formation of self-aggregates of peptide. Triton X-100 might have dissolved the aggregates, resulting in the increase of antimicrobial activity. Similar results were reported by Oscariz and Pisabarro, where the activity of cerein, produced by B. cereus, increased upon treatment with Triton X-100 (49).

In summary, the present study is the first report of the production, purification, and characterization of wild-type and recombinant bacteriocin from B. sonorensis. Also, to our knowledge, this is the first report of the purification and demonstration of the antimicrobial activity of a bacteriocin belonging to a new subfamily of bacteriocins, heterocycloanthracin. It was found to have a molecular mass of 6,274 Da and to be heat stable, stable under a wide pH range, and unaffected in the presence of various organic solvents, surfactants, and reducing agents. An important finding with this peptide was its ability to inhibit a Gram-negative bacterium, E. coli, as bacteriocins produced by Gram-positive strains are mostly inhibitory to Gram-positive strains and are less effective against Gram-negative strains. These properties of sonorensin demonstrate its potential application as a biocontrol agent or food preservative (antimicrobial agent) due to its antimicrobial activity against food-spoiling bacteria, such as L. monocytogenes and S. aureus.