Responses of Enterococcus faecalis resistance and cytolysin up-regulation to nutrients in constructed mesocosms

2.1 Experimental design of model ecosystem

Mesocosms used in this study were located at the Fujian Agriculture and Forestry University (Fuzhou, China, latitude: 26°05′24.9″ N, longitude: 119°13′53.2″ E). The average precipitation during the study period was ranges from 74 to 288 mm. These mesocosms were identical to those described previously (Ali et al., 2016). Briefly, the mesocosms consisted of twelve sterile plastic pails (w × d × h = 20 × 20 × 20 cm) in approximately 27 m³ polyethylene enclosure established in May 2010. Prior to filling with five liters of tap water to each pail, uncontaminated soil from the Fuzhou National Forest Park was placed symmetrically to a depth of 5 cm at the bottom of each pail. The average annual temperature of the water was 24 ± 4.45 °C during the entire experiment. The pH, dissolved oxygen, Secchi Depth (SD) and primary conductivity in mesocosms were inspected regularly in situ, using Water Checker Field Monitor (Horiba Instruments) to ensure that the environmental parameters of all the mesocosms were equivalent at the start of the study. During heavy rain, the pails were temporarily covered with lids. To avoid the evaporation effect, tap water was sprayed daily to nearly re-establish the original water levels.

The mesocosms consisted of four treatments (T1; control without N and P supplementation), T2 (low-dose of N and P), T3 (mid-dose of N and P), T4 (high-dose of N and P) and were arranged in parallel triplicates: (i) T1: unenriched control treatment (ii) T2: P was supplemented with KH₂PO₄ at a final concentration of 0.15 mg/L P and N (1.5 mg/L) supplemented with NH₄NO₃ (iii) T3: the concentration of P (0.45 mg/L) supplemented with KH₂PO₄ and N (4.5 mg/L) supplemented with NH₄NO₃ (iv) T4 was enriched by 1.25 mg/L P and 12.5 mg/L N. These nutrient-enrichment concentrations were selected to modulate the bacterioplankton community structure under different spatial aquatic environments and mesocosm experiments (Ali et al., 2016). Furthermore, these nutrients were chosen since E. faecalis is likely adapted to the aquatic environments containing certain nutrients including N and P. All treatments were inoculated with E. faecalis, resulting in a final concentration of ≈4 × 10⁴ CFU/mL on day 0. Before starting the experiment, trophic assessment standards were established and evaluated by conducting a pre-experiment, to make sure that the different treatment groups have different trophic levels.

2.2 Sample collection

The experiments reported here were conducted from August to November 2015. The nutrients (N and P) were added to the treatments (T2, T3, and T4) at a day 0. Ten samples from each treatment pail were collected on different sampling points to avoid overlapping. These samples (100 µL) were taken from the top of the mesocosms sediment in sterile Eppendorf tubes on day 0, 1, 7, 14, 28, 40, 60 and 96. At the same days, water samples (20 mL) from the pails were also collected in glass jars for the determination of Total Nitrogen (TN), Total Phosphorus (TP), Chemical Oxygen Demand (COD_Mn), pH, and Conductivity (Cond) as described previously (Hou et al., 2013).

2.3 Isolation and identification of enterococci

Ten sediment samples each at 8 sampling events of the mesocosms were taken. Each sample was inoculated in Mueller-Hinton (MH) broth at 25 °C for 24 h with gentle shaking. Then, the samples were spread on Slanetz-Bartley agar medium (CM0377, Oxoid Ltd, Hampshire, UK) at 37 °C for 48 h. Deep red colored colonies were selected and further spread on bile-aesculin agar plates (CM0888, Oxoid Ltd, Hampshire, UK) at 37 °C for 18–24 h and small, brown colonies were selected for subsequent analyses. A total of n = 382 isolates were obtained from day 0 to day 96. DNA of the E. faecalis isolates was extracted using MiniBEST Bacteria Genomic DNA Extraction Kit (Takara, Dalian, China) according to the to the manufacturer's instructions. E. faecalis 16S rRNA gene was identified using primers 72-F (5′-CCG AGT GCT TGC ACT CAA TTG G-3′) and 210-R (5′-CTC TTA TGC CAT GCG GCA TAA AC-3′). E. faecalis ATCC 29,212 was used as a reference strain. For each sampling of the selected days, only one strain was identified and preceded for quantitative real-time PCR (qPCR) analysis.

2.4 Enterococcus faecalis sensitivity testing

Antibiotic resistances of all isolated E. faecalis strains (3 8 2) were grown on MH broth overnight. Broth microdilution method was used according to the (Wikler et al., 2007) breakpoints (Table 1) to determine the resistance of the following antibiotics: ampicillin (AMP), oxytetracycline (OXY), ciprofloxacin (CIP), vancomycin (VAN), chloramphenicol (CHL), and erythromycin (ERY). These antibiotics were selected primarily because they are most frequently found on surface water and sediment in China due to anthropogenic activities.

2.5 Biofilm assay

Biofilm assay of the E. faecalis isolates was performed as described previously (O’Toole, 2010) with minor modifications. Briefly, overnight cultures of E. faecalis isolates were diluted 100-fold in MH broth. 150 μL of diluted cell suspensions were added to 96-well polystyrene plates and incubated at 37 °C for 24 h. After incubation, media from the wells were gently removed, washed with phosphate-buffered saline, left to dry and then stained with 150 μL of 1% crystal violet for 10 min at room temperature. The wells were then washed twice with de-ionized water and left to dry. To destain the biofilm, 150 μL of (33%) glacial acetic acid was added to each well. The optical densities of the wells were measured at 630 nm (OD₆₃₀) in triplicate by using a microplate reader (Tecan, Austria).

2.6 Gelatinase assay

For the qualitative detection of gelatinase activity, cultures of E. faecalis were transferred into gelatin broth (Guangdong Huankai Microbial Sci. & Tech. Co. Ltd. China) having 12% gelatin in 0.8% nutrient broth. The tubes were incubated at 37 °C for 24–48 h in a shaking incubator (190 rpm) and then held at 4 °C for 1 h. The solidified tubes were phenotypically considered gelatinase negative. We used E. faecalis ATCC 29,212 strain because it possesses gelE gene and cyl operon, however, this strain does not contain fsr operon.

2.7 Physiochemical analyses

Standard stock solution (250 mM NaOH, 55 mM K₂S₂O₈, and 242 mM H₃BO₃) was prepared and stored at room temperature. The standard calibration curves of TN and TP were constructed as previously described (Lei Yanzhi, 2007). Water samples were collected to measure TN, TP using a persulfate oxidation method (Valderrama, 1981). For the simultaneous analysis of TN and TP, 10 mL of a water sample from each pail was added to 10 mL standard stock solution and autoclaved at 124 °C for 30 min and further processed as described previously (Valderrama, 1981). The absorbance of the samples were analyzed for TN at 220 nm and 275 nm and for TP at 700 nm using Beckman DU-800 spectrophotometer (Fullerton, USA) and Milli-Q water was used in the reference cuvette. COD_Mn was measured by acid titration method (Ma et al., 2016).

2.8 Quantitative real-time PCR (qPCR)

Total RNA was extracted from 28 sediment samples (5 g) (4 treatments × 7 different time point) using TRIzol method (Shanghai Biological Engineering Co. LTD) as previously described (Ran et al., 2013). The purity of mRNA was determined using Nano-200 Nucleic Acid Analyzer at 260 and 280 nm: A260/A280 optical density ratios range from 1.8 to 2.0 was accepted for further analysis. cDNA was synthesized from 2 µg of total RNA using Prime Script RT Reagent Kit with gDNA Eraser. In qPCR, using SYBR Premix Ex Taq II (Takara, Japan) for the subsequent genes expression of cytolysin large subunit gene (cylL_L), cytolysin small subunit (cylL_S), cytolysin synthetase (cylM), ATP-binding transporter (cylB), component A (cylA), cytolysin regulatory genes (cylR1 and cylR2), cytolysin immunity (cylI), and gelatinase (gelE) as described previously (Ran et al., 2013). The results of these genes were based upon three independent experiments and the relative expression values were internally normalized against reference gene (23S rRNA) expression. Primers used for PCR and qPCR, along with the corresponding size of the amplified fragments are listed in Table S1.

2.9 Statistical analysis

E. faecalis biofilm development was determined by one-way ANOVA and Tukey's multiple pair-wise comparisons among different treatments, and quorum-sensing-related genes expression analyses were performed using Student’s t-test. These analyses were analyzed in MS Excel and Minitab v.18 (State College, PA) at a 5% significance level. A canonical discriminant analysis was conducted on the environmental variables in R version 3.4..

3.1 Mesocosms appearance and survival of E. faecalis

After enrichment of the nutrients, no changes were seen in the physiological appearance of the mesocosms across treatments. The nutrient contents have impacted by several factors, such as the type of organic matter, sediment type, water SD rate, evaporation, and redox conditions. However, the average SD was decreased from the beginning (day 0) to end date (day 96) of the experiment. It was observed that TN concentrations of the different treatments generally decreased over time. The decrease in the relative humidity and longer-term (14 weeks) of the experiment offset the evaporation effect (Table S2). Consequently, the concentration of TN in T3 (2.78 ± 0.35 mg/L N) and T4 (6.64 ± 1.16 mg/L N) treatments were significantly decreased on day 96. In addition, TP concentration also significantly decreased (p < 0.05) on day 96 in the enriched treatments (Table S2). However, control, T2, and T3 was significantly similar on day 1. In the aquatic environments, TN and TP both are the key factors to measure Trophic Level Index (TLI), which indicate a gradual decreased in mesocosm TLI. In addition, the approximate ratio of N:P (14:1) in T3 and T4 treatments was closed to redfield value (16:1), however, this ratio was either high or low in different mesocosms on different days. The multivariate linear model showed that the canonical scores of two dimensions account for 99.2% in the environmental variables (Fig. 1). The first dimension (horizontal) was significantly correlated with conductivity, pH, COD_Mn, TN, and TP while the second dimension (vertical) was significantly correlated with SD and temperature. T4 treatment was more separated from control treatment compared to the T2 and T3 treatments.

Close

Fig. 1

Multivariate linear model (canonical heplot), visualizing the two canonical dimensions of the environmental variables, with alpha = 0.05. T1; control (without nutrient-enrichment), T2; (low-dose of N and P), T3; (mid-dose of N and P), and T4; (addition of nitrogen and phosphorus with high concentration) are four treatments. The lengths of environmental variables indicate their relative contribution to discrimination among the group means. The close angles between variables show high correlations between them. From right to left, the first canonical dimension (88.8%) essentially corresponds to the treatments. The variables such as conductivity (Cond), pH, chemical oxygen demand (COD_Mn), and Total Nitrogen (TN) are largely associated with this dimension (Can1), indicates that they mostly explain by T4. Whereas, Secchi Depth (SD), and Temperature (Temp) is point towards the second canonical dimension (Can2) and negatively correlated with the other four variables.

Export to PPT

Electric conductivity of the mesocosms on day 1 in nutrient amended treatments were 10.43 ± 0.75, 11.67 ± 0.96, and 12.27 ± 0.45 mS/cm in T2, T3 and T4, respectively, significantly different from control 7.43 ± 0.76. The conductivity of mesocosms was gradually increased from day 1 to day 96 in all treatments (Table S2).

The average pH of all the treatments was slightly decreased, however, no significant difference was found among the treatments or days. The lowest pH was 6.8 ± 0.10 on day 14 in control mesocosm and highest was 8.43 ± 0.05 on day 96 in T2 treatment (Table S2). The COD_Mn ranged from 1.56 ± 0.17 to 11.28 ± 8.61 mg/L in all treatments. The overall COD_Mn of T4 was significantly high (p < 0.05) than T2, T3 and control treatments. The number of E. faecalis isolates declined over time in all treatments (Fig. S1). On day 0, the average isolation rates among all four treatments were 100% (10 isolates). However, on day 96, the number of isolates decreased to 5%, 5%, 0% and 15% in T1, T2, T3 and T4, respectively (Fig. S1).

3.2 Effect of nitrogen and phosphorus on Enterococcus faecalis antibiotic resistance

The prevalence of antibiotic (AMP, OXY, CIP, and ERY) resistant E. faecalis strains was detected in four different treatments (Table 1). In a total of 382 isolates, ERY (MIC, 8–256 μg/mL), was the most common of all the treatments. In the nutrient-enriched treatments, 188 (49.2%), 83 (21.7%), and 21 (5.5%) E. faecalis isolates exhibited resistance to OXY, AMP, and CIP, respectively (Table 1). Resistance to VAN and CHL was the least prevalent in the mesocosm experiment. Overall, statistical analysis revealed a significant increase of antibiotic resistance in E. faecalis after exposure to the water environment.

3.3 Biofilm development

To determine whether nutrients and different isolation time point (day 0 to day 96) effect E. faecalis attachment to surfaces, we evaluated biofilm development with the microtiter dish assay. The biofilm development of E. faecalis isolates on day 60 and day 96 were significantly increased (F_7,24 = 62.26, p < 0.01; Fig. 2). Notably, on day 0 no significant difference (F_3,8 = 2.86, p > 0.05) was found among treatments, however, biofilm development was increased significantly in different treatments over time (Fig. 2; Tukey's test (p < 0.05).

Close

Fig. 2

E. faecalis biofilm formation with different time points. E. faecalis biofilm formation shows in T1 (without nutrient-enrichment), T2 (low-dose of N and P), T3 (addition of nitrogen and phosphorus with mid concentration) and T4 (addition of nitrogen and phosphorus with high concentration) treatments. Different letters indicate a significant difference between treatments on the same day (One-way ANOVA and Tukey's test). Error bars represent standard deviation (n = 3). Asterisks indicate statistically significant difference of isolates on different days (* p < 0.05; ** p < 0.01).

Export to PPT

3.4 Analysis of Quorum-sensing gene expression

The expressions of E. faecalis cyl operon genes in response to different concentrations of nutrients in mesocosms were examined by qPCR. The relative expression of quorum-sensing-related genes was made between experimental mesocosms and with respect to those obtained from day 0 (calibrator), regarding strain-dependent expression. The level of cylL_L expression for E. faecalis isolates was determined for each of the different treatments relative to control for seven different time points (Fig. 3a). A significant increase in cylL_L expression was observed on day 7 in T4 treatment (18.93 ± 4.31) compared to control (5.43 ± 1.36), T2 (5.90 ± 0.46), and T3 (6.43 ± 0.83) treatments. Although, cylL_L abundance was significantly similar in all treatments on day 14 and day 28. On day 40, down-regulation was observed in T2 treatment (10.07 ± 1.46) compared to control (15.17 ± 4.30). Fig. 3b illustrates the relative abundance (mRNA relative level in fold difference) of cylL_S at different sampling points of the four different treatments. On day 0, the values (3.5 ± 0.3), (2.53 ± 0.40), (1.83 ± 0.21), and (1.47 ± 0.25) of E. faecalis cylL_S were obtained for control, T2, T3, and T4 treatments, respectively. The cylL_S abundances on day 1 were slightly increased in control (3.87 ± 0.11), T2 (3.63 ± 0.21), T3 (5.77 ± 0.15), and T4 (4.57 ± 0.15) treatments. A significant level of up-regulation (14.7 ± 0.95) was observed in T4 treatment on day 7 compared to control (4.13 ± 0.21). Similar trends were observed for the abundance of cylL_S at day 28, day 40, and day 60. Relative levels of cylR1 expression in the initial isolation (day 0) were less than twofold in all treatments (Fig. 3c). However, the expression of cylR1 in T3 treatment was approximately tenfold (10.73 ± 1.80) increased on day 1. The gradual up-regulation of was observed in the four treatments on day 7. The expression of cylR1 was significantly different (p < 0.05) in T2 (15.83 ± 4.37), T3 (15.8 ± 0.5), and T4 (20.16 ± 6.36) as compared to control (8.5 ± 4.43) on day 7 and 14 (Fig. 3c). In contrast, cylR gene expression was stimulated much more in control than other treatments on day 28 and day 40. The transcription of cylR2 gene was also showed similar trends like cylR1 on first two isolation time points, and the patterns for the day 7 and day 14 were parallel those observed for cylL_S and cylR1 (Fig. 3b; Fig. 3c). The nutrient supplement treatments were significantly higher than control on day 7 and 14. The highest abundance of cylR2 transcription was observed in T4 treatment on day 14 (18.47 ± 2.53), while the lowest expression was observed in T2 (0.83 ± 0.15) treatment on day 0 (Fig. 3d). The level of cylM expression for E. faecalis isolates was determined for each of the different treatments relative to MH broth for seven different time points (Fig. 3e). A significant increase in cylM expression was observed on day 1 in T3 treatment (15.9 ± 1.18) compared to control (3.27 ± 1.55), T2 (2.23 ± 0.32), and T4 (3.13 ± 0.15) treatments. Although overall cylM abundance was significantly higher in T4 treatment on day 7, day 14 and day 28. On day 40, down-regulation was observed in T4 treatment (8.73 ± 0.85) compared to T2 (16.6 ± 0.92). The level of cylB gene was also showed similar trends like cylM on first two isolation time points. However, the patterns for the day 7 in control and T2 were down-regulated (Fig. 3f). On the other hand, the abundances of the cylB gene were up-regulated on other time points. On day 14 the expression of cylB in the T3 and T4 treatments were significantly different from control, nevertheless, by day 28, 40, and 60 a student t-test of this gene abundance in all treatments indicated that both N and P addition were not significantly different from control mesocosm. On the other hand, neither gelE expression by qPCR nor gelatinase positive phenotype by gelatinase assay was detected throughout the experiment (Fig. S 2).

Close

Fig. 3

Expression of E. faecalis cyl operon genes isolated at different time points, relative to expression levels in Mueller-Hinton broth. (a) Abundances of the cylL_L gene (b) cylL_S gene (c) cylR1 gene (d) cylR2 gene (e) cylM gene, and (f) cylB gene, during the time course of the mesocosm experiment. Error bars represent the standard deviation (n = 3).

Export to PPT

4.1 Correlation of nitrogen and phosphorus concentrations on Enterococcus faecalis survival

The persistence and survival of E. faecalis in aquatic environment have been largely influenced by environmental conditions. Hartke et al. (1998) evaluated the survival rate of E. faecalis in tap water over a period of 85 days in an oligotrophic model ecosystem. In addition, our data indicated that the survivals of E. faecalis strain in the mesocosm experiments were gradually decreased from day 0 to day 96 in all treatments (Fig. S1). The natural habitat of E. faecalis is a gastrointestinal tract of humans and animals, however, in aquatic (secondary) environments, E. faecalis adapting the survival strategy against environmental conditions including complex nutrients (Lebreton et al., 2014). On the contrary, mesocosm experiments retain oligotrophic conditions which may affected the survival of E. faecalis (Irankhah et al., 2016). The isolation rates in nutrient-enriched treatment were similar to, or higher than control (Fig. S1), endorses the hypothesis that oligotrophic conditions could somehow influence the survival of E. faecalis in a water environment. Nonetheless, other factors such as salinity, sunlight, and temperature may influence the survival of E. faecalis in environmental settings (Byappanahalli et al., 2012).

4.2 Correlation of nutrients with antibiotic resistance

High levels of N and P in the water environments can cause eutrophication. Eutrophication is highly increased the spreading of antibiotic-resistant bacteria in the aquatic environment. Thus, we hypothesized that N and P play a role in emergence of antibiotic resistance in the water. Consistent with this hypothesis, the supplementations of N and P to mesocosms were induced resistance in E. faecalis over time (Table 1). E. faecalis in the nutrient-enriched treatments were shown resistance to OXY, CHL, AMP, and CIP, indicating that this bacterium evolves resistance with increased nutrients availability. It is apparent that nutrients would affect E. faecalis, as this bacterium had developed resistance in enriched treatments compared to control (Table 1). This finding is consistent with the previous observation in which showing that E. faecalis develop resistance to antibiotics that inhibit protein synthesis (Ali et al., 2016). However, ERY resistant bacteria were found in all treatments including control. Besides N and P, it is likely that other environmental factors may influence antibiotic resistance in aquatic environments. However, additional research is needed to explore environmental factors related to wastewater and sediment which contribute to the development of resistant bacteria.

4.3 Nutrients and biofilm formation

In the aquatic environment, like other Gram-positive bacteria, E. faecalis is able to develop biofilms and increase their acquired and intrinsic resistance to antibiotics. However, the factors which are implicated in the biofilm development remain poorly understood. The nutrients such as nitrates and phosphates are crucial for bacterial growth and often found in oligotrophic environments. To some extent, these nutrients are considered as limiting factors of bacterial growth in drinking water (Fish et al., 2016). In this study, biofilm-forming E. faecalis on day 0 (under nutrient-limited conditions) were significantly similar (Fig. 2). This is in accordance with observations by (Villanueva et al., 2011), concerning the biofilm development of bacteria in a nutrient-limited mesocosm experiment. Nutrient addition was showed a gradual but significant biofilm production in E. faecalis (Fig. 2). It is likely that increased total biofilm formation was favored with increased nutrients availability in the aquatic environment (Miao et al., 2021). Furthermore, in the presence of N and P, bacteria can stimulate their growth rate and reproduction to take advantage of nutrients (Adams et al., 2015). Similarly, bacterial attachment and biofilm accretion to microplate wells were tended to increase over time. A plausible reason for this phenomenon could be the optimal adaptability of E. faecalis in the oligotrophic microcosm environments (Hartke et al., 1998). Based on our findings, it appears that N and P availability in the water might be associated with the biofilm development. In addition to nutrients, other factors, such as genetic and environmental (pH, oxygen, and temperature) may influence the biofilm development (Fish et al., 2016). In the future, it will be of major interest to explore those factors and their mechanisms which are potentially developed bacterial biofilm in the water.

4.4 Influence of nutrients on Quorum-sensing in Enterococcus faecalis

In the aquatic environment, quorum-sensing mechanisms in bacteria usually depend on growth physiology and environmental conditions. Bacteria also regulate the expression of autoinducers in response to N and P within sequencing batch reactors (Liu et al., 2006). To gain insight into how E. faecalis quorum-sensing response to the nutrients, qPCR was used to determine the abundance of structural and regulatory genes of cyl operon for two months on seven different occasions (Fig. 3). In this study, we noted a gradual increase in the expression of cylL_S within nutrient-enriched treatments compared to the unenriched control over time (Fig. 3b). The cylL_S encodes CylL_S protein, a small subunit of cytolysin toxin that is implicated in the autoinduction and control of cyl operon (Van Tyne et al., 2013). The expression of cylL_S in T4 treatment at day 7 was nearly tenfold higher compared to control, suggesting that water containing nitrate and phosphate ions may play a pivotal role in E. faecalis quorum-sensing. In contrast, quorum-sensing regulating genes were induced under P-limited condition in water; N and P with ratio of 5:0.05, respectively (Liu et al., 2006). CylL_S interacts with CylR1 (membrane protein) and further leads to the dissociation of CylR2 from the promoter region that triggers high-level activation of cyl operon (Ali et al., 2017). Since E. faecalis are capable of expressing a cyl operon, we expected to see that there would be a correlation between structural and regulatory genes of cyl operon in the nutrients enriched and control mesocosms. However, relative abundance of mRNA in the control mesocosms was either lower or uniform. This may indicate that the expression of quorum-sensing-related mRNA degradation or regulation might be controlled by environmental and nutritional signals. In response to environmental signals, mRNA decay or regulations lead to express a co-transcribed genes at distinct levels (Mohanty and Kushner, 2016). Furthermore, nutrient alterations in a variety of bacteria can lead to altering quorum-sensing regulation expression. Such studies are needed to explore the quorum-sensing regulation mechanism at high E. faecalis density in the nutritional environment.

E. faecalis is also used fsr quorum-sensing system that regulates social activity, virulence and biofilm development. The role of fsr quorum-sensing in the aquatic environment is thus far undocumented. A complete fsr operon is required for the expression of gelatinase phenotype in E. faecalis (Dundar et al., 2015). The partial deletions or complete absence of fsr operon in E. faecalis are eliminating the gelatinase ability to be expressed in standard assays (Dundar et al., 2015). Furthermore, it has been suggested that a partial deletion of fsr operon is more likely due to recombination and horizontal transfer (Galloway-Peña et al., 2011). Thus, we postulated that as developing of antibiotic resistance, E. faecalis might also be restored gelatinase phenotype in the aquatic environment. However, we did not detect gelatinase positive phenotypes which appear consistent with the previous results (Galloway-Peña et al., 2011). This suggests that complete fsr operon in E. faecalis is essential for the expression of gelE gene in the aquatic environment.