Correspondence: Sergio H. Marshall, Laboratorio de Genética e Inmunología Molecular, Instituto de Biología, Pontificia Universidad Católica de Valparaíso, PO Box 4059, Valparaíso, Chile. Tel.: +56 322 274 866; fax: +56 322 274 835; e-mail: firstname.lastname@example.org
This is the first report of a functional toxin–antitoxin (TA) locus in Piscirickettsia salmonis. The P. salmonis TA operon (ps-Tox-Antox) is an autonomous genetic unit containing two genes, a regulatory promoter site and an overlapping putative operator region. The ORFs consist of a toxic ps-Tox gene (P. salmonis toxin) and its upstream partner ps-Antox (P. salmonis antitoxin). The regulatory promoter site contains two inverted repeat motifs between the −10 and −35 regions, which may represent an overlapping operator site, known to mediate transcriptional auto-repression in most TA complexes. The Ps-Tox protein contains a PIN domain, normally found in prokaryote TA operons, especially those of the VapBC and ChpK families. The expression in Escherichia coli of the ps-Tox gene results in growth inhibition of the bacterial host confirming its toxicity, which is neutralized by coexpression of the ps-Antox gene. Additionally, ps-Tox is an endoribonuclease whose activity is inhibited by the antitoxin. The bioinformatic modelling of the two putative novel proteins from P. salmonis matches with their predicted functional activity and confirms that the active site of the Ps-Tox PIN domain is conserved.
Eubacteria and archaea are known to contain numerous toxin–antitoxin (TA) loci, with many species possessing tens of TA cassettes that can be grouped into distinct evolutionary families (Ramage et al., 2009). Initially known as plasmid addiction or poison–antidote systems (Deane & Rawlings, 2004), TAs have been consistently characterized as plasmid stabilization agents (Boyd et al., 2003; Hayes, 2003; Budde et al., 2007) in which a plasmid-encoded TA functions as a postsegregational mechanism increasing the plasmid prevalence by selectively eliminating daughter cells that did not inherit a plasmid copy at cell division (Van Melderen & Saavedra de Bast, 2009). Nevertheless, in recent years they have also been detected in chromosomes of numerous free-living bacteria (Pandey & Gerdes, 2005). In contrast to the TA loci localized in plasmids, there is no general consensus on the functions of the chromosomal TA systems. A hypothesis was suggested that at least some of these systems (e.g. Escherichia coli mazEF loci) induced programmed cell death (PCD), acting as apoptotic tools (Engelberg-Kulka et al., 2006; Prozorov & Danilenko, 2010). Several researchers have determined that chromosome-borne TA systems are activated by various extreme conditions, including antibiotics (Robertson et al., 1989; Sat et al., 2001) infective phages (Hazan & Engelberg-Kulka, 2004), thymine starvation or other DNA damage (Sat et al., 2003), high temperatures, and oxidative stress (Hazan et al., 2004). Their involvement in the response to amino acid starvation (Hayes, 2003) also raises high interest; indeed, TA modules are believed to provide a backup system to the stringent response by controlling superfluous macromolecular biosynthesis during stasis (Sevin & Barloy-Hubler, 2007). The characteristic features of TA loci are that they comprise a TA gene pair in a bicistronic operon, consisting of an upstream antitoxin and a downstream toxin gene. Normally, two small proteins, a stable toxin and a labile antitoxin, associate tightly so as to keep the toxin component inert (Kwong et al., 2010). The putative role of the antitoxin gene product has been widely discussed, suggesting that there are at least two types of antitoxin. In Type I systems, the antitoxin is an RNA molecule that neutralized the toxin translation and in Type II systems the antitoxin is a small labile protein that binds avidly to the toxin, inhibiting its activity or by downregulating its expression (Hayes, 2003). On the other hand, the toxins of Type I systems are small, hydrophobic proteins that confer their toxicity by damaging cell membranes, while Type II toxins damage particularly either DNA or RNA molecules (Van Melderen & Saavedra de Bast, 2009). In short, whatever their real function is, TA modules can attack cells from within (Engelberg-Kulka et al., 2005) and a number of different intracellular targets have already been identified (Hayes, 2003).
In recent years, the TA system has been consistently associated with a crucial regulatory process in living organisms better known as PCD. PCD is an active process that results in cell suicide and is an essential mechanism in multicellular organisms, required for the elimination of superfluous or potentially harmful cells (Engelberg-Kulka & Glaser, 1999). PCD is currently used to refer to any form of cell death mediated by an intracellular program, no matter what triggers it and whether or not it displays the characteristics of apoptosis (Hengartner & Bryant, 2000). The recent discovery of TA modules in many bacteria suggests that PCD may be a general phenomenon in bacteria (Picardeau et al., 2001).
In this study, we report the presence of a TA locus in the genome of Piscirickettsia salmonis, a Gram-negative fish bacterial pathogen that has affected the salmonid industry since 1989 (Bravo & Campos, 1989). Piscirickettsia salmonis, the aetiological agent of the Salmonid Rickettsial Septicaemia (SRS) or Piscirickettsiosis, belongs to the Gammaproteobacteria group (Fryer & Hedrick, 2003) and was recently reclassified as a facultative intracellular organism (Mauel et al., 2008; Mikalsen et al., 2008; Gómez et al., 2009). Piscirickettsiosis was first reported in coho salmon (Oncorhynchus kisutch) (Bravo & Campos, 1989), but infectivity has also been demonstrated in cultured salmonid species such as the Atlantic salmon (Salmo salar), Chinook salmon (Oncorhynchus tshawytscha), and rainbow trout (Oncorhynchus mykiss) from the south of Chile to the northern hemisphere (Rojas et al., 2009). Recently, the presence of this bacterium has been detected in specimens of white sea bass (Atractoscion nobilis) on the Southern California coast (Arkush et al., 2005). European sea bass (Dicentrarchus labrax) in Greece have been affected by a pathogen similar to P. salmonis (Athanassopoulou et al., 2004); also in Hawaii, tilapia populations (Oreochromis mossambicus and Sarotherodon melanotheron), both free-living as well as farmed fish, have suffered a Piscirickettsiosis-type disease (Mauel et al., 2003), suggesting the expansion of this agent to other fish of commercial importance (Marshall et al., 2007). Although the disease affects several fish species of commercial importance, to date the biology, genetics and epidemiology of P. salmonis have been poorly studied, and so details of relevant aspects of the life cycle of the pathogen are still unknown. The P. salmonis TA locus, named Ps-Tox-Antox, includes its respective regulatory sequences. By in silico comparative genomics of the ps-Tox-Antox locus, we determined that it is homologous to the VapBC TA system of Rickettsia felis and other chromosomal TA operons (Ogata et al., 2005). When the P. salmonis TA genes were cloned and expressed in E. coli for functional analysis, we observed that the characteristics of these genes and their products were similar to other TA systems.
Materials and methods
Growth conditions of P. salmonis and DNA procedures
Piscirickettsia salmonis strain LF-89 (ATCC VR 1361) was grown on Blood Cysteine Glucose (BCG) agar plates at 23 °C (modified from Mauel et al., 2008). A single colony was used to inoculate 25 mL of MC5 broth, and was incubated at 23 °C with agitation of 100 r.p.m. Two-day-old bacterial cultures were processed using the AxyPrepTM Multisource Genomic DNA Miniprep Kit (AxyGen Bioscience) according to the manufacturer's instructions.
Purified P. salmonis DNA was used to construct a genomic DNA library in the plasmid pBluescript SK (+) (Fermentas) and has been described previously (Marshall et al., 2011).
The DNA sequenced data were analysed with the softberry server software (http://linux1.softberry.com/berry.phtml) using the algorithms, FgenesB (to find possible ORFs in the sequences), and Bprom (to search for putative bacterial promoters).
The products of the ORFs predicted by FgenesB were used in blastp analysis, with the search limited to bacterial sequences (http://blast.ncbi.nlm.nih.gov) to determine their possible identities.
Additionally, the primary structure analysis of the new proteins was made by the protparam tool available on the Expasy Proteomic Server (http://www.expasy.org). Thus, the amino acid composition, the hypothetical molecular weight, and the isoelectric point (pI) were all calculated.
Primer design, cloning and expression in E. coli
PCR primers for P. salmonis ps-Tox, ps-Antox, and ps-Tox-Antox genes were designed the Oligo Calc tool (http://www.basic.northwestern.edu/bio-tools/oligocalc.html). The forward primer of ps-Tox (Tox-For: 5′-GATCATATGCTTTATATGCTGGATACGAATATTT-3′) included a NdeI restriction site and the reverse primer (Tox-Rev: 5′- GACGGATCCTTAACTTGCCCAGTCTTCAAACTCTA-3′) had a BamHI restriction site. The Ps-Antox forward primer (Antox-For: 5′-GATGGATCCATGGCAAAATCACGAATTTTTAAA-3′) and the Ps-Antox reverse (Antox-Rev: 5′-GATGGATCCCTAAAACCAGTCACGTTCTTGTGCT-3′) primer have a BamHI restriction site. The primers Tox-Rev and Antox-Rev were designed to be immediately behind the terminator codon to ensure involvement of the terminator in the PCR product. All genes were amplified using P. salmonis genomic DNA as templates, which were purified as described above. Ps-Tox, ps-Antox, and ps-Tox-Antox were amplified by PCR in a 40 μL reaction, using the primers described above. To amplify the ps-Tox-Antox gene, we used the Antox-For and Tox-Rev primers.
The PCR products were purified with the kit MSB Spin PCRapace (Invitek) according to the manufacturer's instructions. Five micrograms of ps-Antox and ps-Tox-Antox PCR products were digested with BamHI endonuclease (New England Biolabs) for 6 h at 37 °C. Five micrograms of purified ps-Tox PCR product was digested with BamHI and NdeI endonucleases (New England Biolabs) for 12 h at 37 °C. The vector pET27b+ (Novagen) (5 μg) was digested with NdeI and BamHI endonucleases (New England Biolabs) for 14 h at 37 °C in order to eliminate the PelB signal for secretion. All digested DNA was purified by the kit MSB Spin PCRapace and the DNA concentration was measured. Two micrograms of digested ps-Antox, ps-Tox-Antox and 2 μg of digested pET27b+ vector was incubated with Klenow DNA Polymerase I (Promega) according to the manufacturer's instructions, and the vector was then dephosphorylated with alkaline phosphatase (Promega). The genes and vector were purified with the MSB Spin PCRapace. Finally, 300 ng of ps-Tox, ps-Antox, and ps-Tox-Antox were ligated with 100 ng of digested pET27b+ in the presence of T4 DNA ligase (Promega) for 16 h at 16 °C. The ps-Tox gene was ligated in pET27b+ vector that was not treated with Klenow.
The ps-Tox, ps-Antox, and ps-Tox-Antox genes ligated on pET27b+ were chemically transformed on competent cells of E. coli BL21 DE3 (Novagen) as described previously (Sambrook et al., 1989). The transformants were plated on Luria–Bertani (LB) agar plates supplemented with kanamycin 50 μg mL−1 and incubated at 37 °C overnight. The colonies were analysed by PCR using the forward primer of each gene and the primer T7 terminator (plasmid reverse primer). Positives colonies were grown overnight and stored at 15% glycerol at −80 °C.
In order to express the recombinant proteins, a duplicate experiment was performed according as follows: 2 mL of LB broth was inoculated with 50 μL of transformant cells and incubated overnight at 37 °C and 250 r.p.m. Fifty microlitres of the overnight cultures was used to reinoculate 2 mL of fresh LB broth and was then incubated at 37 °C and 250 r.p.m. for 2 h. After this, the cultures were induced by the addiction of 1 mM of IPTG (Winkler Ltd.) and incubated for a further 3 h at the same culture conditions as before. An aliquot of 200 μL was taken at the end of every hour and centrifuged at 15 500 g for 10 min, the resultant pellets were resuspended in 100 μL of Laemmli sample buffer. The expression of the recombinant Ps-Tox, Ps-Antox, and Ps-Tox-Antox proteins was visualized on an 18% Tris-tricine urea sodium dodecyl sulphate polyacrylamide gel electrophoresis stained with Coomassie Blue R-250 (Winkler Ltd.).
Toxicity evaluation in E. coli
In order to determine the potential toxic effect of the Ps-Tox protein of P. salmonis, we evaluated the growth rate of E. coli cells. The E. coli strains that contain the ps-Tox, ps-Antox, and ps-Tox-Antox genes were grown on LB broth, in 96-well plates, supplemented with 50 μg mL−1 kanamycin and 1 mM IPTG and incubated at 37 °C for 8 h in constant shaking (200 r.p.m.).
Absorbance (OD600 nm) was measured every hour to determine the growth level of the cells. As an experimental control, we used the same E. coli with the P. salmonis TA genes, which were grown on LB without IPTG in the same conditions described above. Additionally, the E. coli transformant cells were streaked out on agar plates supplemented with 50 μg mL−1 of kanamycin and 1 mM of IPTG. The plates were incubated at 37 °C overnight and the growth level was evaluated.
Structural model of Ps-Tox and Ps-Antox proteins
Based upon the recently determined structure of the VapBC complex of Mycobacterium tuberculosis (Miallau et al., 2008) (PDB ID: 3DBO), we performed a homology model of the Ps-Tox protein. We used the Swiss Model server (Schwede et al., 2003; Arnold et al., 2006), and constructed the model with an alignment of the Mycobacterium VapC-5 toxin and the P. salmonis Ps-Tox toxin. The antitoxin sequence has a 20% identity (%ID) and the toxin sequence has 24% ID. The alignment between Ps-Tox and VapC-5 was made with jalview (Clamp et al., 2004) and the figures were made with the vmd software (Humphrey et al., 1996).
Determine the putative target of Ps-Tox
In order to determine the putative target of the Ps-Tox protein, it was tested for RNase activity based on the presence of a PIN domain. Piscirickettsia salmonis was grown on 5 mL of MC5 medium under the same conditions described above. Two-day-old cultures were centrifuged at 6000 g for 20 min at 4 °C. The RNA was extracted from the bacterial pellet with Trizol® LS reagent (Invitrogen), according to the manufacturer's instructions. The RNA concentration was measured by spectrophotometry. The RNA was kept at −80 °C until use.
The recombinant proteins Ps-Tox, Ps-Antox, and Ps-Tox-Antox were expressed on E. coli. Frozen vials of E. coli BL21 (DE3) bearing the Ps-Tox, Ps-Antox, and Ps-Tox-Antox containing plasmid were used to inoculate 5 mL of LB broth supplemented with 50 μg mL−1 of kanamycin. The culture was grown overnight at 37 °C and 250 r.p.m. Then, 2 mL of these cultures was added to 50 mL of LB broth supplemented with 50 μg mL−1 of kanamycin and the cultures were incubated 250 r.p.m. and at 37 °C until the OD600 nm reached 0.4. When the cultures reached the expected OD they were supplemented with 1 mM of IPTG and incubated for 2 h under the same conditions described above. The cultures were centrifuged at 5000 g for 20 min at 4 °C. The resultant pellet was resuspended in 3 mL of lysis buffer (Tris-HCl 50 mM, NaCl 100 mM, 50 μg mL−1 lysozyme, pH 8), and incubated at 37 °C for 30 min. The samples were sonicated at 11 r.m.s. (three pulses of 20 s) and centrifuged at 16 000 g for 45 min at 4 °C. The protein concentration of supernatants was determined by BCA Protein Assay Kit (Thermo Corporation) according to the manufacturer's instructions.
Finally, 2 μg of P. salmonis RNA was incubated with 100 μg of the E. coli protein extract for 1.5 h at 37 °C. As a positive control, 2 μg of RNA was treated with commercial RNase A (E.Z.N.A Omega-Biotek) and as negative control 2 μg of P. salmonis RNA alone was incubated under the same conditions described above. The digested RNA was visualized on 1% agarose gel stained with GelRed™.
Nucleotide sequence accession number
The GenBank accession number for the P. salmonis ps-Tox-Antox locus is HQ008719.
Characteristic of the ps-Tox-Antox TA operon
The resultant sequences were analysed by FgeneB tool, finding that a sequence of 905 bp contains two putative ORFs. The ORF1 encodes a putative protein of 75 amino acid residues and the ORF2 encodes a putative protein with 135 amino acid residues. Both amino acid sequences were submitted to blastp analysis to determine protein identities. The blastp analysis shows that the protein encoded by the ORF1 has a high level of similarity to antitoxin proteins of bacterial TA modules, specifically to VapB and VagC antitoxins (Table 1). The product of the ORF1, named Ps-Antox, contains an SpoVT/AbrB domain, which is a DNA-binding domain, and, as such, belongs to the super family of transcriptional regulators of the same name. The protein encoded by ORF2, named Ps-Tox, seems to be strikingly similar to toxin proteins of bacterial TA modules, specifically the VapC toxin (Table 1). Additionally, the protein encoded by ORF2 shows the presence of a PIN domain (a homologous domain to the N-terminal domain of the pili biogenesis protein PilT), which is highly conserved in the VapC homologues. The sequence alignment of the Ps-Tox, with other homologues VapC proteins of bacterial TA modules shows a high degree of conservation between them (see Supporting Information, Fig. S1). These results indicate that we have found a typical TA locus in the genome of P. salmonis, named Ps-Tox-Antox.
Table 1. blastp results for the Ps-Antox (ORF1) and Ps-Tox (ORF2) protein sequences
The table shows proteins obtained with query coverage above 98% and e value of e−30.
The P. salmonis ps-Tox-Antox locus consists of a bicistronic operon conformed by an upstream 228-bp gene (ps-Antox) and a downstream 408-bp gene (ps-Tox) separated by an 8-bp intergenic spacer (Fig. 1). By analysis with bprom, we have found a putative promoter region and a Shine–Dalgarno sequence upstream of the ps-Antox gene (Fig. 1). This putative promoter contains a pair of 7-bp inverted repeat sequences (IRs) between the −10 and −35 regions, which is characteristic of other TA operons.
Ps-Tox and Ps-Antox characterization
In order to determine the chemical characteristics of these new P. salmonis proteins, the amino acid sequences of Ps-Tox and PS-Antox were analysed using the protparam tool of the Expasy Proteomic Server. The molecular weight of the Ps-Antox protein was 8.9 kDa and its theoretical pI is 4.79. The predicted molecular weight of Ps-Tox was 15.9 kDa with a pI of 6.7.
In order to characterize this pair of predicted proteins, we cloned the ps-Antox and ps-Tox genes into the pET27b+ expression vector, either individually or together and attempted to express the genes in E. coli BL21 DE3. After the IPTG induction, the expression of the recombinants proteins was checked every 1 h for a 3-h period, finding that the greatest amount of the two proteins is obtained 2-h post-IPTG induction. The expression level of Ps-Antox was much lower than that of its partner Ps-Tox when expressed alone (Fig. 2). When the two proteins were expressed simultaneously in a bicistronic operon the expression level of both proteins was similar, not showing a significant polar effect (Fig. 2).
Toxic effect of Ps-Tox in E. coli
To determine the toxic proprieties of the P. salmonis toxin, we made cultures of the E. coli transformant cells in the presence of IPTG. The growth of the E. coli carrying the pET27b+ vector that contains the ps-Tox gene was minimal in the presence of IPTG during 8 h of growth kinetics (Fig. 3a). In contrast, the growth of E. coli strains that contained the ps-Antox and ps-Tox-Antox in the pET27b+ was normal compared with the host that had the vector without insertion (Fig. 3a). All the transformants strains grew normally in the absence of IPTG, including the strain with the ps-Tox gene in the pET27b+ vector (Fig. 3b). When the strains were streaked out on LB agar plates supplemented with IPTG, the results were the same as those obtained in LB broth (data not shown).
Model of three-dimensional structure
The model constructed is presented in Fig. 4b and c. In general, the secondary structure is conserved compared with that of the M. tuberculosis VapC-5 toxin. Some amino acids implicated in the toxin function are conserved, in particular, three of the four acidic amino acids present in the PIN domains that are related with an exonuclease activity (Miallau et al., 2008), VapC-5: D26, E57, D115, D135, and Ps-Tox: D6, E44, D100, and E121, as can be seen Fig. 4a (see Table S1).
RNAse activity of Ps-Tox protein
In order to determine whether the newly described toxin behaves in the same way as most described toxins, we tested Ps-Tox for putative RNAse activity. When P. salmonis RNA was treated with a crude extract of E. coli containing the recombinant Ps-Tox protein, a significant degradation was observed compared with that of the untreated sample (Fig. 5, lanes 2 and 6, respectively). The same effect was observed in the corresponding extract containing Ps-Antox and Ps-Tox-Antox proteins (Fig. 5, lanes 1 and 3, respectively). The RNA degradation produced by the protein extract that contains Ps-Antox and Ps-Tox-Antox could also have been produced by E. coli RNases present in the sample, as inferred by looking at the E. coli control (Fig. 5, lane 4).
Twenty-five years after its characterization as an obligate intracellular Alphaproteobacteria (Fryer et al., 1992), it has only recently been demonstrated that P. salmonis is truly a free-living bacterial pathogen, belonging to the Gammaproteobacteria group (Fryer & Hedrick, 2003). The bacteria is known to survive in either fresh (Graggero et al., 1995) or marine waters (Olivares & Marshall, 2010) and moreover it is also known to be highly adaptable when exposed to limiting and/or stressing conditions, which mimics its natural situation in the oceans (Rojas et al., 2008). Additionally, the presence of insertion sequences and putatively other mobile genetic elements in P. salmonis represents a solid evidence that the adaptability potential of the bacteria resides in its versatile genome (Marshall et al., 2011).
In this context, the description of a TA locus in P. salmonis appears to be a natural consequence of this versatility. Indeed, TA loci are conserved (often in multiple copies) in the genomes of many organisms that can cause persistent infections and/or persist in the environment: M. tuberculosis, Helicobacter pylori, Coxiella burnetii, Leptospira interrogans, Vibrio cholerae, and Salmonella enterica serovars Typhi and Typhimurium, as well as Haemophilus influenzae, are good examples of this fact (Daines et al., 2007). Additionally, it is important to consider that TA loci are highly abundant in free-living bacteria, but lost from host-associated microorganisms (Pandey & Gerdes, 2005). To date, nine TA families have been reported in the literature: VapBC, RelE, ParE, MAzF, Doc, HipA, HigB, CcdB, and ω-ɛ-ζ (Van Melderen & Saavedra De Bast, 2009). The VapBC is the largest family of bacterial TA modules, representing close to 40% of all the TA loci known, and grouped together by virtue of their toxin components, in most cases belong to the PilT N-terminal domain family of proteins, which in turn function as ribonucleases (Cooper et al., 2009; Robson et al., 2009). Thus, it appears logical and important to identify TA loci in emerging prokaryotic organisms in order to improve our understanding of these systems, and more broadly, in attempting to understand the cellular mechanisms behind bacterial adaptation (Sevin & Barloy-Hubler, 2007).
We have characterized a new and functional bicistronic operon that encodes the two genes of a Type II TA module in P. salmonis. The organization of the P. salmonis TA locus shows many characteristics of other bacterial TA modules. The presence of IRs in the promoter region (Fig. 1) is a feature that is present in various Type II TA systems, such as the vapBC and ChpK operons of L. interrogans (Picardeau et al., 2001; Zhang et al., 2004). The localization of the antitoxin gene upstream of the toxin ORF is a distinctive feature shared by all Type II TA loci homologous to the P. salmonis system. The P. salmonis toxin and antitoxin genes (ps-TA and ps-Antox) are separated by an 8-bp intergenic spacer (Fig. 1), but in many TA operons the antitoxin and toxin genes overlap, indicative of translational coupling between the two cistrons (Gerdes et al., 2005). The sequence of the Ps-Antox protein also shares high identity values with other reported antitoxins, specifically with the well-described VapB and VagC antitoxins (Table 1). Additionally, the Ps-Antox contains a putative SpoVT/AbrB domain, which is present in toxins of the VagC family. Bacillus subtilis SpoVT/AbrB domain proteins are transcriptional regulators, which are expressed during the transition state between vegetative growth and the onset of stationary phase and sporulation (Robertson et al., 1989). The presence of a SpoVT/AbrB domain in the Ps-Antox protein could be explained by the fact that all the Type II TA operons are autoregulated at the level of transcription by the antitoxins, which bind to the TA locus promoters (Gerdes et al., 2005). The best reported example of this issue is the E. coli YefM–YoeB system, which is transcriptionally autoregulated (Kedzierska et al., 2007). We have not explored this possibility in this report.
The FgeneB analysis of the putative sequence of the Ps-Tox protein was submitted to blastp analysis, yielding high identity with members of the VapC family proteins (Table 1). The sequence alignment of Ps-Tox with VapC homologues showed a high level of conservation (see Table S1), indicating that it does correspond to a toxin coded in a bacterial TA module. The Ps-Tox protein contains a PIN domain, which is another distinctive feature of the toxins from the VapC and ChpK families (Arcus et al., 2005; Miallau et al., 2008). The PIN domains (homologues of the pilT N-terminal domain) are small protein domains of ∼140 amino acids (Arcus et al., 2005). In eukaryotes, PIN-domain proteins function as ribonucleases with activity linked to RNAi and nonsense-mediated RNA degradation (Clissold & Ponting, 2000). In prokaryotes, the majority of PIN-domain proteins are the toxic components (by virtue of their ribonuclease activity) of chromosomally encoded TA operons (Arcus et al., 2005). Because the Ps-Tox toxin does display endoribonuclease activity (Fig. 5) as do other VapC homologues and also contains a PIN domain, we could speculate a putative action similar to the Mycobacterium tuberculosis VapC-5 product, which specifically blocks protein translation via mRNA cleavage (Ramage et al., 2009). In fact, the structural model of Ps-Tox (Fig. 4b) shows that the secondary structure elements of the toxin are preserved in comparison with the M. tuberculosis VapC-5 toxin, with some helix and beta sheet shorter residues in Ps-Tox. Notably, the active site defined by VapC-5 from M. tuberculosis and shared by PIN domains (Miallau et al., 2008) is conserved in the Ps-Tox protein (Fig. 4c). Three of the four acidic amino acids are conserved with only a semiconservative change of E121 to D121 (see Table S1), suggesting that the acidic cavity responsible for the exonuclease activity is preserved.
ps-Tox and ps-Antox genes expressed in E. coli BL21 DE3, yielded products with molecular weights perfectly matching those predicted by the protparam device (15.9 and 8.9 kDa, respectively) (Fig. 2). Additionally, expression of the ps-Tox gene in E. coli cells in the presence of the inducer IPTG showed the expected toxic phenotype for at least the first 8 h of growth (Fig. 3a). The toxic effect of Ps-Tox is counteracted when it is coexpressed with the ps-Antox gene (Fig. 3a). Notwithstanding, and as expected, the bacterial growth is normal in the absence of the inducer (Fig. 3b). Our results also suggest that the molecular target of the Ps-Tox protein (RNA) is conserved between E. coli and P. salmonis, specifically by the presence of a PIN domain. Similarly, other studies have shown that a chromosome-encoded TA system, such as that from L. interrogans (the VapBC and ChpK modules), is able to inhibit the growth of E. coli cells and both the TA products do interact in the heterologous system (Picardeau et al., 2001; Zhang et al., 2004). Thus, we assume that the toxin gene is also functional in P. salmonis.
In conclusion, our data clearly demonstrate that the Ps-Tox-Antox system of P. salmonis corresponds to a fully active module belonging to the VapBC family. Considering that the expression of the ps-Tox gene has been demonstrated to be highly toxic to E. coli cells, the newly described module appears as a potential innovative tool for pathogen control via peptide interference (Lioy et al., 2010).
This work was supported by Innova Corfo grant 05CT6IPD-22 to S.M. and Conicyt Doctoral Scholarship to F.G.