Effects of a pulsed light-induced stress on Enterococcus faecalis

Abstract

Aims

Pulsed light (PL) technology is a surface decontamination process that can be used on food, packaging or water. PL efficiency may be limited by its low degree of penetration or because of a shadow effect. In these cases, surviving bacteria will be able to perceive PL as a stress. Such a stress was mimicked using low transmitted energy conditions, and its effects were investigated on the highly environmental adaptable bacterium Enterococcus faecalis V583.

Methods and Results

In these laboratory conditions, a complete decontamination of the artificially inoculated medium was performed using energy doses as low as 1·8 J cm⁻², while a treatment of 0·5, 1 and 1·2 J cm⁻² led to a 2·2, 6 and 7-log₁₀ CFU ml⁻¹ reduction in the initial bacterial population, respectively. Application of a 0·5 J cm⁻² pretreatment allowed the bacteria to resist more efficiently a 1·2 J cm⁻² subsequent PL dose. This 0·5 J cm⁻² treatment increased the bacterial mutation frequency and affected the abundance of 19 proteins as revealed by a global proteome analysis.

Conclusions

Enterococcus faecalis is able to adapt to a PL treatment, providing a molecular response to low-energy PL dose, leading to enhanced resistance to a subsequent treatment and increasing the mutation frequency.

Significance and Impact of the Study

This study gives further insights on Ent. faecalis capacities to adapt and to resist to stress.

Introduction

Enterococcus faecalis is a natural commensal of the gastro-intestinal tract of humans and animals and can be used as cultures starter in food fermentations, as animal health supplement and/or as probiotic. This bacterium is able to contaminate and stay viable in diverse environments such as soil, sand, water, plants and food, owing to its robustness (Mundt 1986). However, it is also capable of causing opportunistic infections, including bacteraemia, endocarditis, meningitis, and wound, urinary tract and nosocomial bloodstream infections (Murray and Weinstock 1999; Wisplinghoff et al. 2004). It is responsible for up to 80 % of enterococcal-associated nosocomial infections, because of its natural abundance and its ability to produce virulence factors (Hancock and Gilmore 2000). Although harmless in healthy individuals, enterococcal strains may become pathogenic in patients in intensive care units and in hospitalized patients with impaired immune systems (Ogier and Serror 2008).

Pulsed light (PL) technology can be used for decontaminating food, packaging and water (Elmnasser et al. 2007; Gomez-Lopez et al. 2007; Garvey et al. 2010), using ultrashort duration pulses of an intense broadband emission spectrum (200–1100 nm) that is rich in UV-C germicidal light (Wang et al. 2005; Gomez-Lopez et al. 2007). This latter is important because an alteration in the UV portion of the light spectrum can negatively impact the LP treatment effectiveness (Takeshita et al. 2003). PL efficiency may also be limited by its low degree of penetration or because of shadow effects (Marquenie et al. 2003; Wuytack et al. 2003; Gomez-Lopez et al. 2007), which could have significant consequences on the physiology of bacteria. In this case, the resulting PL treatment could be perceived as a stress, thus triggering a specific bacterial stress response. In a previous study, we have shown that the Gram-negative Pseudomonas aeruginosa is able to perceive and to adapt to such stress (Massier et al. 2012). We performed here a study on the highly robust Gram-positive Ent. faecalis in terms of survival capacity, growth parameters, mutagenic ability and global proteome pattern and discussed the similarities and differences of comportment between these two bacteria.

Materials and methods

Bacterial strains and growth conditions

Enterococcus faecalis V583 (Sahm et al. 1989) was grown at 37°C for 24 h without shaking in a MCDE Trp CAA chemically defined medium (Reffuveille et al. 2011; Massier et al. 2012). For bacterial numeration, cells were diluted in 0·9% NaCl and plated on GM17 agar supplemented with 0·5% glucose (Terzaghi and Sandine 1975). Colony forming units (CFU) were counted after 24 or 48 h of growth at 37°C. For bacterial growth recovery experiments, Ent. faecalis was grown in MCDE Trp CAA, a chemically defined medium, to a final OD₅₈₀ of 0·8 ± 0·05, and subjected to a PL treatment with fluences ranging from 2 × 10⁻⁵ to 3 J cm⁻². Treated and control (untreated) cultures were then diluted in fresh medium to an OD₅₈₀ of 0·08 in a 96-well microplate, and growth was monitored each 30 min for 24 h in a multimodal microplate reader (SAFAS Xenius XML; SAFASmonaco, Monaco).

PL treatment

PL was delivered by a laboratory-scale PL pilot equipment (Claranor SA, Avignon, France) used in nonindustrial conditions because the aim of the study was not to decontaminate the medium but to allow some bacteria to survive and to adapt to the treatment (Massier et al. 2012). Briefly, short-time pulses of 250 μs and of broad spectrum (200–1100 nm) were produced by four xenon flash lamps. The four lamps were placed at the maximal distance (12 cm) from the quartz tank containing the bacterial culture, leading to lowering the energy dose emitted. In this configuration, the fluence generated by one pulse at 500–3000 V or two pulses at 2500 V or 2–5 pulses at 3000 V, ranged from 2 × 10⁻⁵ to 3 J cm⁻² (Massier et al. 2012). Enterococcus faecalis cells were grown to an OD₅₈₀ of 0·8 ± 0·05, treated by PL in UV-transparent quartz containers and plated onto GM17 agar plates for numeration. After 24 h at 37°C, CFU were counted, and the PL effect was expressed as a reduction rate, that is, the log₁₀ of the ratio between the number of CFU counted after (N) and before treatment (N₀). Unflashed samples and medium were used as negative controls. The experiments were performed at least three times.

Adaptation assays

Enterococcus faecalis V583 was grown to an OD₅₈₀ of 0·4 ± 0·05 and exposed or not (control) to low-energy PL doses ranging from 2 &times 10⁻⁵ to 0·2 J cm⁻². Cultures were then incubated at 37°C to reach a final OD₅₈₀ of 0·8 ± 0·05 and then submitted to a second PL treatment of 0·5, 1 or 1·2 J cm⁻² and plated onto GM17 agar plates for numeration. After 24 h at 37°C, CFU were counted, and the PL effect was expressed as a reduction rate, that is, the log₁₀ of the ratio between the number of CFU counted after (N) and before treatment (N₀).

Rifampicin resistance assay

The frequency at which rifampicin-resistant mutants appear in vitro, a method often used as an indicator of the overall mutation frequency in bacteria, including Enterococci (Gustafsson et al. 2003), was used for the comparison of PL with continuous UV treatment. One hundred microlitres of bacterial suspensions (treated or not) was diluted and plated onto GM17 agar supplemented or not with rifampicin (20 μg ml⁻¹). The bacterial growth on GM17 agar supplemented with rifampicin characterized the rifampicin-resistant cells after PL or continuous UV treatment. Otherwise, the number of CFU corresponded to the total number of bacteria surviving after the PL or the continuous UV stress. The control consisted of cells that were not treated by PL or continuous UV, to evaluate the ratio rifampicin-resistant cells/initial population, representative of a ‘natural’ mutagen ability of Ent. faecalis (Massier et al. 2012). This experiment was performed at least three times.

Continuous UV treatment

Continuous UV treatment was carried out at 254 nm (UV lamp: 90 W, VL-315-G; Fisher Bioblock Scientific, Illkirch, France), in the following conditions leading to the same bacterial reduction as PL, that is about 2·2 log₁₀. To perform this, cells were grown to an OD₅₈₀ of 0·8 ± 0·05 and then treated by continuous UV at a fluence of 0·091 ± 0·05 J cm⁻² for 11·5 s, before being subjected to the rifampicin resistance assay.

Statistical analyses

The mean values and standard deviations were calculated from data obtained from at least three independent experiments performed on different days with different bacterial cultures. Normality test and one-way analysis of variance (anova) were used for statistical analysis of the data, using Past ver. 1.84 for Windows (http://folk.uio.no/ohammer/past/).

Protein extraction and two-dimensional gel electrophoresis (2-DE)

Enterococcus faecalis cells, treated or not with a sublethal PL dose of 0·5 J cm⁻², were harvested, and proteins were extracted as described by Kilic et al. (2010). Protein concentrations were determined with the Bio-Rad protein assay (Bio-Rad Laboratories, Marnes la Coquette, France) using bovine serum albumin as standard. Immobilized linear pH gradients (11 cm, pH 4–7; Bio-Rad Laboratories) were used for the first dimension. Two-dimensional gel electrophoresis gels were achieved according to the study by Massier et al. (2012). Image analyses were performed using PDQuest 2-DE analysis software (Bio-Rad Laboratories). For each condition, three independent samples were analysed on six gels (each sample was loaded onto two gels). Spots that showed reproducible variations in intensities (with a twofold change or greater) were selected for further analyses by trypsin digestion and peptide mass fingerprinting.

In-gel trypsin digestion and peptide mass fingerprinting

The procedure developed by Shevchenko et al. (2006) was slightly modified for in-gel digestion, as previously described by Massier et al. (2012). All data (MS and MS/MS) were analysed using mascot software (ver. 2.1) for interrogating several databases (Swiss Prot, NCBI) to identify the proteins present in each spot. The criteria used to accept protein identification based on MS/MS data included the extent of sequence coverage (minimum of 13%), the numbers of matching peptides (minimum of 3) and the score of probability (minimum of 120 for the Mascot score).

Results

In this study, the laboratory-scale PL device was used under nonindustrial conditions to stress Ent. faecalis. However, under these laboratory conditions, a complete decontamination of the artificially inoculated medium (9 log₁₀ CFU ml⁻¹) was performed using energy doses as low as 1·8 J cm⁻² (Fig. 1). A treatment at 0·5, 1 or 1·2 J cm⁻² led to about 2·2 log₁₀, 6 log₁₀ or 7 log₁₀ reduction in the initial Ent. faecalis population, respectively (Fig. 1). When exposed to very low PL energies, ranging from 2 × 10⁻⁵ to 0·2× J cm⁻², the reduction in the initial population was <1 log₁₀ (Fig. 1).

Figure 1. Effect of the pulsed light energy dose (fluence) on the population reduction rate of Enterococcus faecalis. Data are means of at least three independent experiments, and standards errors are represented by vertical bars.

Effect of a PL stress on Enterococcus faecalis growth parameters

We next studied the effect of a PL-induced stress on the bacterial growth parameters. Enterococcus faecalis was thus treated with increasing PL doses and then diluted to an OD₅₈₀ of 0·08 in fresh chemically defined medium and allowed to grow for 24 h. In these conditions, the lag time increased with regard to the induced PL stress. This phenotype could be correlated with stress intensity, resulting in bacterial reduction (Fig. 2). The doubling time increased gradually after a treatment ranging from 0·06 to 0·6 J cm⁻². It is noticeable that for these PL doses, the maximal OD₅₈₀ reached were similar to the control. In contrast, no significant growth recovery could be observed for treatments with fluences above 0·6 J cm⁻² (Fig. 2).

Figure 2. Pulsed light (PL) effect on Enterococcus faecalis growth recovery. Bacteria were not treated (■) or treated by PL doses of 0·00002 (△), 0·06 (○), 0·2 (♢), 0·3 (×), 0·5 (*), 0·6 (□), 1·2 (+), 1·8 (···) and 3·0 (—) J cm⁻². The average standard deviation of the points is 0·03.

PL pretreatment favours the survival of Enterococcus faecalis to a subsequent PL stress

We next investigated whether an initial PL stress could induce a resistance of Ent. faecalis to a subsequent PL challenge. Bacteria were thus grown until an OD₅₈₀ of 0·4 and first exposed to low PL energies below or equal to 0·5 J cm⁻². In these conditions, the reduction rate was more than 2·2 log₁₀ (Fig. 1). The stressed bacteria were then allowed to grow to reach a final OD₅₈₀ of 0·8 before being treated with fluences of 0·5, 1 or 1·2 J cm⁻². The pretreatment had no effect on bacterial survival when the second PL stress was of 0·5 or 1 J cm⁻² (Fig. 3). However, when bacteria were submitted to a second PL stress of 1·2 J cm⁻², the bacterial reduction was lower for pretreated bacteria comparatively to the bacteria that were not exposed to a low-energy PL dose prior challenge. The survival gain was dependent of the adaptation dose and reached about 2 log₁₀ when Ent. faecalis was previously subjected to a 0·5 J cm⁻² dose (Fig. 3). Taken together, these results suggest that a first PL treatment can induce a better resistance of Ent. faecalis to a second subsequent PL stress.

Figure 3. Enterococcus faecalis is able to adapt to a pulsed light (PL) treatment. Bacteria were treated with a fluence of 0·5, 1 or 1·2 J cm⁻² after being exposed to low-energy PL doses or not then plated onto GM17 agar and numerated. The effect of PL was expressed as reduction rate defined by doing the log₁₀ of the ratio of the number of colony forming units, after (N) and before the second PL exposition (N₀). The dashed line represents the initial bacterial population. Data are means ± standard errors and represent three independent experiments. 0 (■); 0·00002 (□); 0·06 (); 0·2 (); 0·5 ().

The 0·5 J cm⁻² PL treatment increases the mutagen ability of Enterococcus faecalis

The effectiveness of the PL treatment in terms of microbial decontamination has been at least partly attributed to its content in UV-C (Wang et al. 2005; Gomez-Lopez et al. 2007), which is well known as a powerful mutagen agent (Bintsis et al. 2000). The application of a low-energy dose of 0·5 J cm⁻² led to altered growth parameters (Fig. 2) and to a better survival rate after a subsequent PL stress (Fig. 3). We then investigated the effect of this energy dose on the Ent. faecalis mutation frequency using a rifampicin resistance assay. Enterococcus faecalis is sensitive to this antibiotic, but is able to undergo spontaneous mutations yielding subclones resistant to this antibiotic at a frequency of 10⁻⁷ (Fig. 4, control). By treating the bacteria with the 0·5 J cm⁻² PL dose, the number of rifampicin-resistant cells was significantly increased compared with that of untreated cells (Fig. 4). These results indicate that the 0·5 J cm⁻² dose increased the mutagen ability of Ent. faecalis, suggesting that the UV content of such a PL dose may be sufficient to induce DNA alterations. However, when Ent. faecalis was treated with 1 J cm⁻², the frequency of rifampicin-resistant bacteria became null (data not shown). We finally compared the PL treatment to continuous UV in terms of mutant frequency. The experiments were performed to allow the determination of UV treatment conditions that led to the same rate of decontamination, that is 2·2 log₁₀ CFU ml⁻¹. This was performed by delivering 0·091 ± 0·05 J cm⁻² for 11·5 s (data not shown). By treating bacteria in these conditions, a similar mutagen effect was observed compared with PL (Fig. 4), showing that PL is not more mutagen than continuous UV in our conditions.

Figure 4. A 0·5 J cm⁻² treatment increases Enterococcus faecalis mutation frequency. Results are the ratio of the CFU counted on GM17 agar plates containing rifampicin (N) vs the one counted without antibiotic (N₀), before (control: white bar) or after a 0·5 J cm⁻² pulsed light (PL) treatment (black bar) or a 0·091 ± 0·05 J cm⁻² continuous UV irradiation during 11·5 s (hatched bar). Bars represent the mean of at least three measurements with the corresponding standard deviation. The absence of statistical significance difference is indicated (NS; P > 0·05). The dashed line represents the initial bacterial population.

Proteomic analysis of the bacterial response to a 0·5 J cm⁻² PL treatment

We finally investigated the effects of the 0·5 J cm⁻² PL treatment at a molecular level by performing a global proteome analysis of Ent. faecalis. Figure 5 shows two representative 2-DE gels performed on Ent. faecalis cells that have been subjected to a 0·5 J cm⁻² PL stress or not (control). Both were highly similar but nevertheless allowed the detection of 20 spots, showing reproducible intensity variations. Spots of interest were excised for further identification by mass spectrometry analysis and corresponded to 19 proteins that were classified into six groups on the basis of their nature or function: energy metabolism, adaptation to atypical conditions, cell division, transcription and translation processes, phage-related proteins and carbohydrate transport protein (Table 1).

Table 1. Putative functions of the identified proteins whose expression is altered in response to a low-energy pulsed light (PL) treatment

Spot numbera	EF number	Protein name	Number of matched peptides	Coverage (%)	Mascot Scoreb	PL effect on proteins
+: More protein on the treated Enterococcus faecalis sample; −: less protein on the treated Ent. faecalis sample, compared with the control. A minimum of twofold change expression was considered as significant. ^a Numbering is according to Fig. 5. ^b Mascot score is 10 × log (P) where P is the probability that the observed match is a random event. The last column represents fold changes expressed as the ratio adapted/control.
Energy metabolism
1	EF1961	Enolase	3	29	318	−
2	EF1045	6-phosphofructokinase	6	27	149	−
3	EF1964	Glyceraldehyde-3-phosphate dehydrogenase	6	24	403	−
4	EF0105	Ornithine carbamoyltransferase	3	18	186	−
5	EF1962	Triosephosphate isomerase	5	33	460	−
Adaptation to atypical conditions/Stress related proteins
6	EF1367	Cold shock domain contain protein	4	87	383	+
7	EF2925	Cold shock domain contain protein	4	78	362	+
8	EF3233	Dps family protein	4	19	369	−
9	EF0080	Gls24 protein	3	18	163	+
Cell division
10	EF1151	Cell division protein DivIVA, putative	4	46	126	+
Transcription and translation processes
11	EF0201	Elongation factor Tu	5	18	316	+
12	EF2395	Ribosome-recycling factor	4	34	241	+
13	EF2397	Elongation factor Ts	6	21	333	+
13′	EF2397	Elongation factor Ts	4	17	205	+
14	EF2398	30S ribosomal protein S2	3	13	123	+
15	EF2914	Elongation factor GreA	4	41	409	+
Phage-related proteins
16	EF0318	Hypothetical protein	3	26	197	+
17	EF0322	Hypothetical protein	3	13	153	+
18	EF1285	Major tail protein	3	22	252	+
Carbohydrate transport protein
19	EF0709	Phosphocarrier protein HPr	5	42	449	+

Figure 5. Effect of the 0·5 J cm⁻² treatment on Enterococcus faecalis protein expression. Proteins of Ent. faecalis treated (b) or not (a) with a low-energy pulsed light (PL) dose were separated by 2-DE. The identified proteins were numbered according to Table 1. The gels are representative of three independent protein extractions.

Energy metabolism

This group consisted of five proteins, involved in energy metabolism, among which the enolase EF1961, the triosephosphate isomerase EF1962, the glyceraldehyde 3 phosphate dehydrogenase EF1964, the 6-phosphofructokinase EF1045 and the ornithine carbamoyltransferase EF0105, which expression was downregulated in response to the PL-induced stress (spots 1–5, Fig. 5 and Table 1).

Adaptation to atypical conditions

Four proteins belong to this group, among which two cold shock domain contain proteins (EF1367 and EF2925) and the stress response Gls24 protein EF0080 (Choudhury et al. 2011), which expression was induced in response to the PL-induced stress, while that of the Dps family protein EF3233 involved in redox homoeostasis (Calhoun and Kwon 2011) was reduced (Fig. 5 and Table 1).

Cell division

This group was represented by the EF1151 DivIVA, an essential protein required for cell division (Ramirez-Arcos et al. 2005), and its expression was increased in response to PL stress.

Transcription and translation processes

This group consisted of five identified proteins, Tu, Ts and GreA elongation factors (Ef-Tu: EF0201, Ef-Ts: EF2397, GreA: EF2914), the 30S ribosomal protein S2 (EF2398) and the ribosome-recycling factor (EF2395), which were overproduced (Fig. 5 and Table 1). Two isoforms of Ef-Ts were detected (Spots 13 and 13′ on Fig. 5), both being overproduced.

Phage-related proteins

The proteins of spots 14–16 (Fig. 5), corresponding to EF0318, EF0322 and EF1285, were identified as phage-related proteins and were overproduced in response to the PL treatment (Table 1).

Carbohydrate transport protein

The phosphocarrier protein HPr (EF0709), which is required for carbohydrate entering into the bacterial cell through the sugar phosphotransferase system (PTS, Maurer et al. 2001), was overproduced after a PL stress (Fig. 5 and Table 1).

Discussion

In this study, the PL pilot was used under nonindustrial conditions to stress Ent. faecalis to test its adaptation potential to PL. The four xenon lamps were placed at 12 cm from the bacterial sample, as previously described (Massier et al. 2012), and a limited number of flash was applied, resulting in lowering the energy dose delivered (Farrell et al. 2010), because it has been shown that these parameters had a significant effect on the decontaminating efficiency of the PL treatment (Farrell et al. 2010; Massier et al. 2012). Under these laboratory conditions, a suspension containing 10⁹ CFU ml⁻¹ was decontaminated by applying a fluence of 1·8 J cm⁻², comforting PL as an efficient decontamination process. This was also the case for the Gram-negative Ps. aeruginosa treated in similar conditions (Massier et al. 2012). Comparing the decontamination curves of Ent. faecalis and Ps. aeruginosa led to the conclusion that Ent. faecalis was more resistant to the PL-induced stress than Ps. aeruginosa, which is consistent with previous reports, indicating that Gram-positive bacteria are less sensitive to a PL treatment than Gram-negative bacteria in industrial conditions (Farrell et al. 2010). Enterococcus faecalis was able to recover growth, reaching the same final density of untreated bacteria, after receiving energy doses lower than or equivalent to 0·6 J cm⁻². We show furthermore that (i) the 0·5 J cm⁻² PL dose resulted in 2·2 log₁₀ reduction in a total initial population of 9 log₁₀ CFU ml⁻¹, (ii) Ent. faecalis was less susceptible to a second PL treatment when previously treated with a low-energy PL dose, and (iii) a low-energy dose (0·5 J cm⁻²) increased the mutation frequency of Ent. faecalis. Taken together, these results show that Ent. faecalis was able to perceive a low-energy PL dose and to respond to this stress leading to a better survival rate. In a previous study, we applied a similar protocol on Ps. aeruginosa cells (Massier et al. 2012) and showed that a pretreatment of Ps. aeruginosa with a 0·2 J cm⁻² energy dose (which was responsible for <1 log₁₀ of bacterial reduction when taken alone) led to a survival increase of 3 log₁₀ when bacteria where further treated with a 1 J cm⁻² energy dose. When Ent. faecalis was treated exactly in these conditions (first treatment of 0·2 J cm⁻² followed by a second dose of 1 J cm⁻²), no gain in terms of bacterial survival could be observed, except when increasing the second PL treatment to 1·2 J cm⁻² (survival gain lower than 1 log₁₀). Furthermore, it was noticeable that only a slight reduction (<0·5 log₁₀ CFU ml⁻¹) was obtained when treating Ent. faecalis with a 0·2 J cm⁻² energy dose (Fig. 1), consistently with the higher resistance to PL of Enterococci compared with Gram-negative bacteria (this study and Farrell et al. 2010). In our experimental conditions, the best survival gain (2 log₁₀) was observed when pretreating Ent. faecalis with a 0·5 J cm⁻² energy dose followed by a second treatment of 1·2 J cm⁻². Taken together, these results suggest that despite its better resistance to a PL stress, Ent. faecalis seems to have a more limited capacity to adapt to PL-induced stress compared with Ps. aeruginosa.

Using a rifampicin resistance assay, we further showed that the 0·5 J cm⁻² energy dose led to an increase in the mutation frequency of Ent. faecalis. Considerable research has been performed to elucidate the different mechanisms underpinning microbial inactivation by PL, all of them emphasizing pivotal role of the UV component of the pulsed spectrum along with intimating other minor photochemical and photothermal effects (Anderson et al. 2000; Takeshita et al. 2003; Wuytack et al. 2003; Wang et al. 2005). The UV-C germicidal spectrum is required for the decontaminating efficiency of the PL process, and UVs are known to be powerful mutagen agents (Takeshita et al. 2003; Gomez-Lopez et al. 2005, 2007). This germicidal effect has been mainly associated with the formation of pyrimidine dimers, inhibiting the formation of new DNA that ultimately derails the vital process of cell replication (Bolton and Linden 2003). As the mutation frequency increased for treated Ent. faecalis cells, it may suggest that the UV content of the low-energy PL dose was sufficient to induce mutations in these bacteria, and consequently, that the DNA may be damaged, and/or that DNA repair mechanisms could be active, leading to these mutations (Takeshita et al. 2003; Gomez-Lopez et al. 2005, 2007). Furthermore, treating Ps. aeruginosa with the low-energy dose (0·2 J cm⁻²), which led to the best survival gain after a second subsequent PL treatment (Massier et al. 2012) or with a 0·5 J cm⁻² dose (our unpublished data), did not increase the mutation frequency of this bacterium. Taken together, these data suggest that the UV content of the low-energy dose was sufficient to provoke mutations in Ent. faecalis, but not in Ps. aeruginosa.

A proteomic study was then performed to get insights into the effects of a PL-induced stress. We found that the production of four stress response proteins was altered; among which Gls24, two cold shock domain containing proteins and a Dps family protein, suggesting that in our conditions, PL induced a stress response on Ent. faecalis. While the first three proteins were overproduced, Dps was underproduced in response to a PL-induced stress. The general stress response Gls24 protein (EF0080) is known to be increased in response to different stressors, including pig faecal extract, sausage medium (Hew et al. 2007), starvation, cadmium chloride (Giard et al. 2000) bile salts (Giard et al. 2000; Lenz et al. 2010; Choudhury et al. 2011), nitrite (Lenz et al. 2010) or copper (Stoyanov et al. 2010), leading to increased resistance of Ent. faecalis to such stressors. It has been furthermore shown that disruption of Gls24 in Ent. faecalis OG1RF resulted in significant attenuation in a mouse peritonitis (Teng et al. 2005) and a rat endocarditis models (Nannini et al. 2005), indicating that Gls24 is important for Ent. faecalis virulence (Choudhury et al. 2011). EF1367 and EF2925 are members of the cold shock proteins family and appear to function as DNA or RNA protective chaperones. Members of this family are involved in various cellular processes and play important roles in stress adaptation in wide variety of organisms (Chaikam and Karlson, 2010). These data indicate that the PL treatment was perceived as a stress by Ent. faecalis. Dps protein interacts with DNA and DNA-binding proteins in relation to the organization of the nucleoid and microbial survival (Chiancone and Ceci 2010; Calhoun and Kwon 2011). In Escherichia coli, Dps is involved in DNA binding, iron sequestration and has a ferroxidase activity, making Dps extremely important in iron and hydrogen peroxide detoxification (Calhoun and Kwon 2011). In our conditions, Dps expression was lower when Ent. faecalis grown to exponential phase was treated with a low-energy PL-induced stress. It is well known that PL, used as a decontamination process, is an advanced oxidation process generating in aqueous solutions very reactive oxidant components and free radicals (Moreau et al. 2008). It has been furthermore shown that Dps expression is downregulated at both the transcriptional and the post-transcriptional levels in E. coli grown to exponential phase, while it is overexpressed during the stationary growth phase. Downregulation of dps expression occurs when hydrogen peroxide levels are low during exponential phase and involves the nucleoid-associated proteins Fis and H-NS, blocking dps promoter and thus its transcription. Post-translational regulation of Dps through proteolysis during exponential growth involves the ClpAP and ClpXP proteases and occurs in the absence of oxidation stress (Calhoun and Kwon 2011). Taken together, our data suggest that the low-energy PL dose was not able to generate a sufficient amount of oxidative species to activate oxidative stress defence mechanisms in Ent. faecalis. Furthermore, a Dps-like protein was also underproduced in Ps. aeruginosa in response to a sublethal PL dose (Massier et al. 2012).

The production of five proteins involved in energy conversion was decreased. This is consistent with a lowering of the energetic metabolism in response to the PL treatment. The phosphocarrier protein HPr (EF0709), which is required for carbohydrate entering into the bacterial cell through the sugar PTS (Maurer et al. 2001), was overproduced after a PL stress as well as GreA transcriptional elongation factor, the Ef-Tu and Ef-Ts translational elongation factors, the 30S ribosomal protein S2 and the ribosome-recycling factor EF2395. Ef-Tu favours the entering of the aminoacyl-tRNA into the ribosome, and Ef-Ts is the guanine nucleotide exchange factor for elongation factor Tu (Dahl et al. 2006). These data indicate that transcriptional and translational processes were increased in Ent. faecalis, in response to the PL stress. Similar results were obtained for Ps. aeruginosa (Massier et al. 2012). It is noticeable that the protein DivIVA (EF1151) was overproduced in response to PL stress. As this protein is required for proper cell division, chromosome segregation and cell growth (Ramirez-Arcos et al. 2005), it suggests that the PL stress resulted in altered cell division of Ent. faecalis. Finally, we found that phage-related proteins were overproduced in treated Ent. faecalis, as it was the case for Ps. aeruginosa (Massier et al. 2012). UV irradiation is a condition well known to occasion lysogenic induction. The overproduction of proteins EF0318, EF0322 and EF1285 whose genes are located in prophage 01 for EF0318 and EF0322 and in prophage 02 for EF1285 (Lepage et al. 2006) may thus be a consequence of induction of these prophages. These data suggest also that the UV contain of the low-energy PL dose treatment may be sufficient to induce the production of these proteins, in accordance with the increase mutation frequency observed through the rifampicin resistance assay. Although the exact role of these phages in Ent. faecalis is yet unknown, they may have a function in bacterial fitness, giving competitive growth advantages between bacterial strains (Nakayama et al. 2000). Furthermore, phages potentially play an important role in the acknowledged spread of antibiotic resistance genes from enterococcal strains to pathogenic bacteria (Yasmin et al. 2010).

Our study shows that Ent. faecalis is able to adapt to a PL treatment because a proteomic response to a low-energy PL dose was observed. It appears that under these conditions, Ent. faecalis reduced the energy conversion systems and increased transcription and translation processes, which led to produce proteins involved in chaperone mechanisms. It is noticeable that the proteomic response of Ent. faecalis was closely similar to that observed in Ps. aeruginosa, suggesting that these highly adaptable Gram-negative and Gram-positive bacteria may perceive a low-energy PL dose as a stress leading to a stress response, which may be involved in the bacterial adaptation to a subsequent PL stress. However, Ent. faecalis response was also associated with an increased mutation frequency, which was not observed for Ps. aeruginosa. It is conceivable that the low GC content of Ent. faecalis genome (compared with the high GC content of that of Ps. aeruginosa), or perharps that the cell wall structure of the bacteria, could be involved in such differences. While this issue remains to be more deeply investigated, our study provides guidance for future experiments addressing this point.

Finally, while PL is a highly efficient decontamination process, we suggest that careful consideration of factors such as the shadow effect, linked either to the matrix to be treated or to the contamination density, will be required, because bacteria as different as Ent. faecalis and Ps. aeruginosa are able to adapt and to respond to a low-energy PL treatment.

Acknowledgements

We thank B. Lacour (Claranor) for helpful discussions. S. Massier was recipient of a doctoral fellowship from CLARANOR and ‘Association Nationale de la Recherche Technologique’ (CIFRE, ANRT). This work was supported by grants from the ‘Conseil Général de l'Eure’, the ‘Grand Evreux Agglomeration’, the Région Haute Normandie, the European Union (FEDER) and the ‘Plate Forme Technologique d'Evreux Normandie Sécurité Sanitaire’.