Hydrogen peroxide-mediated killing of Caenorhabditis elegans by Enterococcus italicus and Lactococcus garvieae isolated from food

Abstract

In this study, we used the nematode Caenorhabditis elegans as a model to assess the pathogenic potential of two species isolated from food, Enterococcus italicus and Lactococcus garvieae, for which few indications on pathogenicity are available. We identified the conditions under which E. italicus and L. garvieae are able to kill the nematode and suggest that the production of hydrogen peroxide (H₂O₂) by these two bacteria was involved in the death of C. elegans in our model system. The efficacy of E. italicus and L. garvieae to kill C. elegans differed, most likely related to each species’ distinct ability to accumulate H₂O₂ (4.9 mM and 0.9–1.1 mM, respectively). Genome analysis of both species revealed that the genome of E. italicus contains a gene encoding a NADH oxidase which shows high amino acidic similarity with H₂O₂ -forming NOX-1 enzymes, while that of L. garvieae contains a gene codifying for a water-forming NADH-oxidase (NOX-2). Reverse transcriptase-PCR experiments carried out in presence of flavin adenine dinucleotide (50 mM) confirmed the presence of the two different genes and likely explains the different toxicity of E. italicus and L. garvieae against C. elegans in our study. The results obtained show for the first time the production of H₂O₂ in E. italicus and L. garvieae and indicate its toxic effect in the nematode C. elegans.

Introduction

In the last two decades, foodborne diseases have emerged as an important and growing public health concern (EFSA 2008). This has led to the development and use of methods to monitor and detect undesirable microorganisms in food products. Today, together with conventional microbiological analyses, more sensitive methods, such as immunology-based methods involving antigen–antibody interactions and PCR analysis for DNA characterization, are available. However, little is known of the presence and function of virulence factors in several emerging foodborne bacteria or of their impact bacteria on specific ecosystems, and their mechanism of pathogenesis is not fully understood (Hatcher et al. 2012). Innovative strategies and methods are thus needed for surveying foodborne disease and food contamination (Tauxe 2002).

The nematode Caenorhabditis elegans can be used as a simple and well-established model host in studies on emerging foodborne pathogens. Comprehensive studies have reported that this worm is a versatile model metazoan in which to assess the virulence of many human pathogens (Ewbank and Zugasti 2011). This free-living nematode offers numerous practical and economic advantages, including its small size (adults are approx. 1 mm long) and short life cycle (only 3 days from the fertilized egg to a sexually active adult) (Blaxter 2011). Moreover, the bacteriovorous C. elegans can be easily challenged with pathogens, and its life span is readily measured using simple assays (Tan et al. 1999; Ballestriero et al. 2010; Penesyan et al. 2013). Laboratory experiments have shown that C. elegans responds differently to pathogenic species, such as Salmonella enterica and Enterococcus faecalis (Aballay et al. 2000; Garsin et al. 2001), and to probiotic bacteria, such as Lactobacillus paracasei and Bifidobacterium spp. (Lee et al. 2011). Caenorhabditis elegans also recognizes toxic solutes, such as alcohol, reactive oxygen species, heavy metals and sulphydryl reactive compounds, which induce the production of cellular stress proteins by inhibiting feeding (Sifri et al. 2002). Consequently, worm-based assays have been widely employed to study two different types of mechanisms of pathogenesis. Bacterial pathogens can kill the nematode either through an infection-like process (Tan et al. 1999) or by diffusible toxins (Jansen et al. 2002; Moy et al. 2004). In the infection model system, C. elegans ingests the bacteria and dies over the course of several days, depending on the pathogen. In contrast, killing by diffusible toxins occurs more quickly and does not require contact with viable bacterial cells.

The aim of our study was to employ the C. elegans system to assay the pathogenic potential of two emerging foodborne bacteria, Enterococcus italicus and Lactococcus garvieae. The genus Enterococcus encompasses a considerable number of different species, some of which are well characterized in regards to their pathogenicity (Eaton and Gasson 2001; Semedo et al. 2003). Enterococcus italicus, which is dairy in origin (Fortina et al. 2004), has been found to have a low virulence profile (Fortina et al. 2008; Borgo et al. 2009), but the identification of a clinical isolate belonging to E. italicus (Carvalho et al. 2008) and genome analysis of the type strain of the species (Borgo et al. 2013) indicate the possible presence of several virulence factors for which further studies are needed. Lactococcus garvieae is a pathogen that causes septicaemia in fish and serious economic damage to fish aquaculture worldwide (Vendrell et al. 2006). In addition, this bacterium has been found in different food matrices (Fortina et al. 2003; Fernández et al. 2010; Ferrario et al. 2012), and several cases of L. garvieae infection in humans have been reported (Fihman et al. 2006; Aubin et al. 2011; Russo et al. 2012). However, despite the growing importance of L. garvieae in both human and veterinary medicine, little data directly related to pathogenicity factors are available. In our study, we show that E. italicus and L. garvieae are able to produce hydrogen peroxide (H₂O₂) and that the production of H₂O₂ is likely related to their potential pathogenicity.

Materials and methods

Bacterial and nematode strains

Bacterial strains were grown on Brain Heart Infusion (BHI medium (Difco, Necton Dickinson, Sparks, MD) at 37 °C. Liquid cultures were typically grown without shaking. We used Enterococcus italicus strain TP1.5^T (DSM 15952^T) isolated from Italian cheese and Lactococcus garvieae strains TB25 and LG9, originating from Italian cheese and diseased fish, respectively. Caenorhabditis elegans strain Bristol N2 (Brenner 1974) was maintained by using standard practices (Sulston and Hodgkin 1988).

Nematode killing assay

Bacterial cultures were grown in BHI medium at 37 °C and diluted to an optical density at 600 nm (OD₆₀₀) of 0.1–0.2 in fresh medium; 10-μl aliquots were then spread onto 35-mm BHI agar plates and the plates incubated overnight at 37 °C. AnaeroGen (Oxoid, Basingstoke, UK) envelopes and jar containers were used to generate an anaerobic environment. Before the nematodes were transferred to the anaerobically grown lawns of E. italicus and L. garvieae, the plates were cooled aerobically at room temperature for 30–60 min. Nematodes at the L4 developmental stage were suspended and washed twice in M9 buffer (3 g KH₂PO₄, 6 g Na₂HPO₄, 5 g NaCl, 1 ml 1 M MgSO₄, H₂O to 1 l), then transferred to the bacterial lawns in drops (5–10 ml; 30–40 per plate) and incubated at 25 °C. Worms were scored for survival every 24 h. Nematodes that did not respond to touch with a platinum wire pick were considered to be dead. For the experiments with catalase (Sigma, St. Louis, MO), catalase was mixed into liquid BHI agar at a concentration of 1,000 U ml⁻¹. Assays for each bacterial strain and condition were carried out at least twice, and each experiment was done in triplicate.

H₂O₂ sensitivity assays

Bacteria were grown in BHI medium at 37 °C until the mid-log phase (OD ₆₀₀ = 0.4–0.6) was reached, centrifuged for 2 min at 10,000 g and 4 °C, washed in ice-cold saline solution (0.9 % NaCl, w/v) and resuspended in fresh BHI medium. Resuspended bacteria were added to microtiter plate wells in duplicate containing twofold dilutions of H₂O₂ (Sigma,) in BHI medium and incubated for 30 min. Aliquots from each well were applied to BHI agar plates containing 200 U of catalase (Sigma) per milliliter for viable counts. The concentration of H₂O₂ required to cause a 99.9 % decrease in the number of colonies compared to the negative control without H₂O₂ was recorded as the minimum bactericidal concentration (MBC).

H₂O₂ measurement

To measure H₂O₂ accumulation upon aeration of cultures grown under anaerobic conditions, we cultivated 100-ml BHI medium cultures of E. italicus and L. garvieae strains in 500-ml flasks at 37 °C without agitation in jar containers. Cultures were removed from the anaerobic conditions and aerated on an orbital shaker. To measure bacterial growth and H₂O₂ accumulation in aerobic cultures, overnight cultures were grown without shaking, diluted into 100 ml of prewarmed BHI medium aerated at 37 °C and centrifuged at 10,000 g. The H₂O₂ levels of bacterium-free culture eluates were measured with horseradish peroxidase type VI-A (Sigma) at 505 nm. The H₂O₂ was quantified using standard curve determined with H₂O₂ (Sigma) solutions at concentrations ranging from 1.5 to 100 mmol l⁻¹.

Catalase and NADH oxidase activity assay

Aerobic cultures grown under anaerobic conditions were harvested at an OD_600nm of 0.4–0.6 and the cells washed with ice-cold phosphate buffer (PB) (50 mmol l⁻¹ KH₂PO_4, 0.5 mmol l⁻¹ EDTA, pH 7.0). Cells were disrupted using glass beads (diameter <106 μm; Sigma) and a Bead Beater (Precellys®24; Bertin Technologies, Montigny le Bretonneux, France) and were then centrifuged for 15 min at 4 °C at 10,000 g. The protein concentration was determined by the Bradford (1976) method using serum albumin as a standard.

Catalase activity was measured by adding 25 ml of cell extract to a cuvette containing 0.1 % H₂O₂ in 1 ml of 50 mmol l⁻¹ PB. The rate of H₂O₂ decomposition was recorded as the change in adsorption at 240 nm. The extinction coefficient for H₂O₂ (ε₂₄₀ = 0.0436 cm² mmol⁻¹) was used to calculate catalase activity units. One unit decomposes 1 mmol H₂O₂ min⁻¹.

NADH oxidase activity was measured by adding 25 ml of cell extract to 1 ml reaction mixture containing 150 mmol l⁻¹ NADH (Sigma) with or without 50 mmol l⁻¹ flavin adenine dinucleotide (FAD; Sigma) in the PB, and absorbance at 340 nm was monitored for 2 or 3 min.

nox detection

Total DNA was extracted as described in Borgo et al. (2013). The NADH oxidase gene (nox) primers were designed on the basis of comparisons of the nucleotide sequences of the E. italicus DSM 15952^T and L. garvieae TB25 and LG9 whole genome shotgun sequence (Accession nos: PRJNA61487, AGQX01000000, AGQY01000000, respectively). Before use, the sequence of the forward primer was checked with BLAST (Basic Local Alignment Search Tool) at the National Center for Biotechnology Information (NCBI) to ensure that no matches with other genes were present (Altschul et al. 1990). The primers were synthesized by PrimmBiotech (Milan, Italy), and the sequences were 5′-GCATTATATGTCGGCGGTGT-3′ and 5′-CGCAATAATTGGCCATGAGC-3′ (forward and reverse, respectively). Each 25-μl reaction mixture contained 100 ng of bacterial DNA, 2.5 μl of 10× reaction buffer, 200 mM of a deoxynucleoside triphosphate mixture, 2.5 mmol l⁻¹ MgCl₂, 0.5 mmol l⁻¹ of each primer and 0.5 U of Taq polymerase (Dream Taq DNA-polymerase; MBI-Fermentas, Vilnius, Lithuania). Amplification was performed in a Gene Amp PCR System 2400 Thermal Cycler (Perkin-Elmer, Norwalk, CT) and consisted of 30 cycles of 94 °C for 1 min (denaturation), 58 °C for 1 min (annealing) and 72 °C for 2 min (extension), followed by an additional 7-min elongation period at 72 °C. PCR products were separated on 1.5 % agarose gel, stained with ethidium bromide in 1× Tris–acetate–EDTA buffer (40 mmol l⁻¹ Tris-acetate, 1 mmol l⁻¹ EDTA, pH 8.0) and photographed under UV light.

Quantitative RT-PCR

RNA was extracted with a NucleoSpin RNA II extraction kit (Macherey-Nagel, GmbH & Co. KG, Düren, Germany) according to the manufacturer’s instructions. The concentration and purity of the extracted mRNA were optically determined using a UV–Vis spectrophotometer (SmartSpecTM Plus; Bio-Rad, Hercules, CA), which measured the absorbance of the sample at wavelengths of 260 and 280 nm. Based upon the final concentrations, 100 ng of mRNA was rewritten into cDNA using a RevertAid First Strand cDNA Synthesis kit (MBI-Fermentas) and thermal cycling of 5 min at 25 °C, 30 min at 42 °C and 5 min at 85 °C. The mRNA expression levels of nox were analyzed with SYBR® Green technology (Life Technologies, Carlsbad, CA) in real-time quantitative PCR (qPCR) using the SsoFast EvaGreen Supermix (Bio-Rad) on a Bio-Rad CFX96 system according to the manufacturer’s instructions. In order to quantitatively measure the expression of the E. italicus and L. garvieae nox genes, we used the housekeeping gene 16S rRNA as internal control to which all other gene expression was normalized. The primers used were 5′-GAAGTCGTAACAAGG-3′ and 5′-GTGTCTCAGTCCAGTGTGG-3′ for 16S rRNA and the previously cited for nox genes. Samples were examined for differences in gene expression using relative quantification in which relevant gene expression is normalized to a housekeeping gene, according to Pfaffi (2001).

The specificity of the primers was tested with melting curves during amplification and in 1 % agarose gels. qPCR amplification was carried out as follows: an initial hold at 96 °C for 30 s, followed by 40 cycles at 96 °C for 2 s and at 60 °C for 5 s. The amount of template cDNA used for each sample was 1.25 ng.

Results

E. italicus- and L. garvieae-mediated killing of C. elegans

Preliminary experiments were carried out to verify their ability to grow in BHI medium under aerobic or anaerobic conditions. No significant differences in cell growth were found when the bacterial strains were grown under aerobic or anaerobic conditions (Table 1). The dry cell weight under aerobic or anaerobic conditions was also similar after the addition of FAD (50 mmol l⁻¹). Following growth under aerobic conditions E. italicus and L. garvieae did not have any apparent deleterious effect on the nematode C. elegans when the bacteria were provided to the worms as the sole source of food (Fig. 1a). However, when grown under anaerobic conditions on BHI medium and then exposed to aerobic conditions, the E. italicus and L. garvieae strains tested rapidly killed C. elegans (Fig. 1b). Furthermore, nematode mobility decreased within 45 min (data not shown). Lactococcus garvieae strains TB25 and LG9 killed the nematode within 5 days, while in the presence of E. italicus TP1.5 ^T 50 % of the nematodes died within 2–4 h, and nearly all of the nematodes had died by 8 h (expanded view ofFig. 1b). As shown in the expanded view of Fig. 1b, the addition of exogenous catalase to the solid BHI medium allowed the nematodes to remain mobile, with 100 % of survival. The same behavior was observed when L. garvieae strains were tested.

Table 1 Effects of different culture conditions on dry cell weight, NADH oxidase activity and hydrogen peroxide accumulation in BHI medium at 25 °C after a 24-h incubation with Enterococcus italicus strain TP1.5^T and Lactococcus garvieae strains TB25 and LG9

Fig. 1

Caenorhabditis elegans survival when fed Enterococcus italicus strain TP 1.5 and Lactococcus garvieae strains TB25 and LG9, respectively. The tested strains were grown on solid BHI medium under aerobic conditions (a) and under anaerobic followed by aerobic conditions (b). L4-stage C. elegans was transferred onto the lawns of tested strains, incubated at 25 °C and scored for survival. Symbols: Filled black circle Escherichia coli strain OP50 (positive control), filled square L. garvieae strain TB 25, X E. italicus strain TP 1.5, filled diamond L. garvieae strain LG9. Expanded view C. elegans survival, expressed in hours, when the worm was fed E. italicus grown under anaerobic/aerobic conditions in the presence (filled diamond) or absence (filled circle) of exogenous catalase

Killing of C. elegans is related to H₂O₂ production

On the basis of previously published data (Moy et al. 2004), we examined the capacity of the E. italicus and L. garvieae strains tested to produce H₂O₂. E. italicus TP1.5 and L. garvieae TB25 and LG9 accumulated H₂O₂ in relation to cell growth, with the concentration of H₂O₂ in the culture media of these strains reaching 4.9, 0.9 and 1.1 mmol l⁻¹, respectively, within the first 8 h of incubation (Table 1). The ability to accumulate H₂O₂ was related to a high level of resistance to this toxic substance. When cells were exposed to 0.1 and 1 mmol l⁻¹ of H₂O₂ for 30 min, the survival of the tested strains was unaffected. Following exposure to 10.0 mmol l⁻¹ H₂O₂, E. italicus decreased to undetectable levels (0.5 log), whereas the number of L. garvieae cells decreased approximately by 2 log cycles. These results indicate that E. italicus TP1.5 has a higher ability to survive in the presence of exogenously added H₂O₂ than the L. garvieae strains tested.

Detection of NADH oxidase genes and expression studies

To study the biosynthetic enzymes involved in H₂O₂ production we investigated the genomes of all three strains (for accession numbers, see section nox detection) for the presence of genes related to this process. We did not detect any genes encoding catalase and NADH peroxidase enzymes, but in both species we did find a nucleotide sequence highly similar to that of nox genes which encoded NADH oxidase activity. Thus, E. italicus and L. garvieae would appear to use oxidases to reduce molecular oxygen to form H₂O₂, and the absence of NADH peroxidase, the major mechanism to scavenge H₂O₂, could result in increased accumulation of this reactive molecule. Two types of NADH oxidases are known, corresponding to H₂O₂-forming oxidase (NOX-1) and H₂O-forming oxidase (NOX-2). A comparison between the amino acid sequences of NOX in the tested strains and other reported amino acid sequences in other species and genera showed a high similarity (78 %) of the nox gene of E. italicus with NOX-1 enzymes and a lower level of similarity with the water-forming NADH oxidases (NOX-2). On the contrary, the NADH oxidase found in the L. garvieae strains showed a higher similarity (81 %) with NOX-2 enzymes.

Figure 2 shows the alignment based on the primary structure of the E. italicus 446-amino acid NOX protein and L. garvieae strains TB25 and LG9 462 amino acid NOX proteins. BLAST analysis of the amino acid sequence in NOX revealed conserved sequences, including the known FAD and NADH binding domain motifs and the LAXXAXXXG sequence, which appears to be highly conserved in flavoproteins. A cysteine residue in position 42 characterizing water-forming NADH oxidase (NOX-2) was found only in the L. garvieae NADH oxidase gene (Fig. 2): this amino acid allows the capture of H₂O₂, forming the sulfenic acid (cys42-SOH) intermediate. Electrons from a second molecule of NADH reduce sulfenic acid to water and thiolate (Jiang et al. 2005). Based on our results, E. italicus appears to possess a NADH oxidase H₂O₂-forming (NOX-1) in the presence of FAD, as also verified by measuring the enzymatic activity and the relative gene expression. Our results are similar to those previously published in which the addition of free FAD stimulated the oxidase activity of NOX-1 (Riebel et al. 2002), but not the activity of the NOX-2 enzyme (Jiang et al. 2005), thereby verifying the enzymatic activity in the presence/absence of FAD. As shown in Table 1, FAD added exogenously to the culture system at 50 mmol l⁻¹ increased the activity of NADH oxidase in E. italicus (by 6-fold), but did not significantly change the activity of this enzyme in the two L. garvieae strains tested.

Fig. 2

Structure-based amino acid alignment of NADH oxidase in L. garvieae TB25, L. garvieae LG9 and E. italicus TP 1.5. Shading Conserved sequence motifs. BD Binding domain, FAD flavin adenine dinucleotide

To confirm the nature of the NOX enzymes found in E. italicus and L. garvieae, we also performed qRT-PCR experiments, designing a primer set able to amplify a 222-bp internal region of the nox genes. Expression levels under standard culture conditions (BHI medium, 37 °C, 16 h of incubation, anaerobic conditions) and under aerobic conditions (BHI medium, 37 °C, 16 h of incubation) in the presence or absence of exogenously added 50 mmol l⁻¹ FAD. In the latter case, a ten-fold increase in the expression level of nox was observed only in the E. italicus strain, confirming that in this species the NADH oxidase gene encodes a putative H₂O₂-forming oxidase, whose expression is positively modulated in presence of additional FAD.

Discussion

In our study, we used C. elegans as an innovative tool for studying two emerging foodborne pathogens, E. italicus and L. garvieae, for which limited information on their potential pathogenicity is available. This study has not fully resolved the question of the nature of the killing of C. elegans by E. italicus and L. garvieae, but our results do support the hypothesis of a mechanism involving H₂O₂, which has been proposed for other enterococcal species (Garsin et al. 2001; Moy et al. 2004). The first step in our study was to grow the three strains to be tested under different culture conditions. This assay revealed that E. italicus and L. garvieae displayed a killing phenotype when grown under anaerobic conditions, leading us to hypothesize that the tested strains were producing a toxic compound during the anaerobic growth phase. However, further experiments revealed that the toxin H₂O₂ was not produced until the lawns of E. italicus and L. garvieae were exposed to oxygen. This led to a new hypothesis, namely, that following exposure to oxygen, these bacteria accumulate the H₂O₂ produced because the absence of NADH peroxidase in the cells (as verified in the genomic analysis of the strains) cannot adequately scavenge H₂O₂. The bacterial strains tested showed a different ability to kill C. elegans which was linked to a distinct accumulation of H₂O₂ (4.9 mmol l⁻¹ for E. italicus and 0.9–1.1 mmol l⁻¹ for L. garvieae). Furthermore, nematode killing was prevented by an addition of catalase to cultures of E. italicus and L. garvieae that had killed the worms.

This difference in the ability to kill C. elegans can be explained by the results of our analysis of the amino acid sequence of the NADH oxidases found in the two species. A cysteine residue in position 42, present only in the L. garvieae NADH oxidase gene (Fig. 2) characterizes H₂O-forming oxidase (nox-2); this amino acid allows the capture of H₂O₂, forming the sulfenic acid (cys42-SOH) intermediate. Electrons from a second molecule of NADH reduce sulfenic acid to water and thiolate (Jiang et al. 2005). Enterococcus italicus seems to possess a NADH oxidase that forms H₂O₂ (nox-1), as also verified by measuring the enzymatic activity and the relative gene expression in the presence of FAD. Riebel et al. (2002) reported that the addition of this cofactor induces an increase expression of nox1 gene, as found in our study of the E. italicus NADH oxidase.

Here, we report for the first time genotypic and phenotypic evidence of a possible mechanism of virulence in two potential emerging food pathogens, E. italicus and L. garvieae. H₂O₂ production was reported and its toxic effects studied in vivo. Caenorhabditis elegans, a free-living nematode, provided the model for investigating toxin-mediated killing. The presence along the genome of E. italicus NOX-1 and L. garvieae NOX-2 could explain the distinct killing phenotype of these two species against C. elegans. This genetic diversity may also explain the different accumulation of H₂O₂ and, consequently, the distinct killing phenotype of E. italicus and L. garvieae against C. elegans. As reported for other Gram-positive bacteria, including pathogens, such as Streptococcus pneumoniae and S. pyogenes (Spellerberg et al. 1996; Jansen et al. 2002), the H₂O₂ produced by the cells can contribute to virulence by damaging host tissue or surrounding cells. Further studies are necessary to understand if this molecule may play a role in E. italicus and L. garvieae colonization or infection and if additional virulence factors are responsible for nematode killing.

In conclusion, in this preliminary study we reported for the first time genotypic and phenotypic evidence of a possible mechanism of virulence in two potential emerging food pathogens. The results of our study also highlight the utility of the C. elegans pathogenicity model for identifying bacterial virulence factors.