Virulence Genes among Enterococcus faecalis and Enterococcus faecium Isolated from Coastal Beaches and Human and Nonhuman Sources in Southern California and Puerto Rico

Most Enterococcus faecalis and E. faecium are harmless to humans; however, strains harboring virulence genes, including esp, gelE, cylA, asa1, and hyl, have been associated with human infections. E. faecalis and E. faecium are present in beach waters worldwide, yet little is known about their virulence potential. Here, multiplex PCR was used to compare the distribution of virulence genes among E. faecalis and E. faecium isolated from beaches in Southern California and Puerto Rico to isolates from potential sources including humans, animals, birds, and plants. All five virulence genes were found in E. faecalis and E. faecium from beach water, mostly among E. faecalis. gelE was the most common among isolates from all source types. There was a lower incidence of asa1, esp, cylA, and hyl genes among isolates from beach water, sewage, septage, urban runoff, sea wrack, and eelgrass as compared to human isolates, indicating that virulent strains of E. faecalis and E. faecium may not be widely disseminated at beaches. A higher frequency of asa1 and esp among E. faecalis from dogs and of asa1 among birds (mostly seagull) suggests that further studies on the distribution and virulence potential of strains carrying these genes may be warranted.

E. faecalis and E. faecium are among the most common species of enterococci found in the beach environment [18][19][20][21]. Enterococci found in the beach environment can include naturalized populations existing in soil and vegetation as well as strains from humans, sewage, animals, birds, reptiles, and insects [1]. Presumably, potentially pathogenic E. faecalis and E. faecium in human fecal waste would harbor higher numbers of virulence genes as compared to strains from animal and environmental sources. In Puerto Rico, beaches receive storm flows containing contaminated septage and agricultural runoff potentially carrying enterococci derived from human and animal fecal waste. In Southern California, urban runoff, beach sand, and sea wrack (macroalgae on beach sand) have been identified as important sources of enterococci to beach water [22][23][24][25].
Previous studies showed that E. faecalis and E. faecium from the beach water and sand harbor antibiotic resistant genes suggesting a potential health risk for beach goers [26][27][28][29][30]; however, the frequency of other virulence factors was not determined. Here, we compared the frequency of putative enterococcal virulence genes (esp, gelE, cylA, asa1, and hyl) among E. faecalis and E. faecium from beaches in Southern California and Puerto Rico impacted by different enterococci source inputs to assess beaches as an environmental reservoir of potentially virulent enterococci.  Table 1). The environmental isolates were randomly selected from a collection of strains obtained from previous studies [20,31] including beach water, eelgrass (Zostera marina), wrack (mainly Macrocystis pyrifera), sand, creek, or storm drain runoff upstream of beaches and sewage influent (untreated waste) and effluent. Isolates from animals were obtained from bird (mostly seagull) stools on beach and dog stools. Human (nonclinical) strains of E. faecalis and E. faecium were isolated from urine and fecal samples from 18 healthy (nonhospitalized) individuals residing in Southern California. Human fecal and urine specimens obtained from healthy individuals were considered representative of strains that could be found in beach water due to human shedding or contamination from sewage and/or septage. Ten E. faecium isolates identified as vancomycin resistant enterococci (VRE), 5 clinical strains of E. faecium (non-VRE), and 10 E. faecalis isolated from rectal swabs (3), urine (1), blood (1), abscess (1), ascites (2), vagina (1), and joint (1) were provided by Orange County Public Health Laboratory (OCPHL). Clinical isolates were included for comparison to strains with enhanced virulence potential.

Puerto Rico.
A total of 247 Enterococcus (174 E. faecalis and 73 E. faecium) isolates from Puerto Rico were analyzed ( Table 2). Enterococcal isolates from beach water were obtained from two beaches in Puerto Rico. Human (nonclinical) enterococcal strains were isolated from fresh fecal samples from nine healthy individuals from Mayaguez, Puerto Rico. Clinical enterococcal strains were isolated from urine specimens and identified to species level by a local hospital in Mayaguez. Six septage samples were obtained from individual houses or from septic tank trucks after emptying individual family tanks.

Southern California.
Enterococcal isolates from all samples (except for clinical specimens) were obtained using mEI agar and identified to species level using the Vitek II (bioMérieux) plus additional biochemical tests and pigment and motility as per Ferguson et al. [32]. Clinical strains were isolated by OCPHL using TSA with 5% sheep's blood; presumptive enterococcal colonies were gram-stained and identified using MicroScan (Siemens Healthcare) and/or API Strep 20 (bioMérieux). Up to 3 isolates per sample identified as E. faecalis and E. faecium were randomly selected for virulence gene analysis. Species identification of 8 different isolates obtained using biochemical methods was also confirmed by 16S rRNA sequencing conducted at GenoSeq, University of California, Los Angeles.

Puerto Rico.
Enterococcal isolates from all samples (except for clinical specimens) were obtained using mE agar. All isolates were divided into four groups based on pigmentation and motility. The isolates were identified to the genus level based on growth in BHI with 6.5% NaCl, growth at 45 ∘ C, esculin hydrolysis, catalase, and PCR amplifying of the Tuf gene [33]. Species level identification was done by a double digestion of the PCR product of the ATP synthase subunit gene in combination with a restriction fragment length polymorphism (RFLP) assay (paper in preparation). Clinical strains obtained from a local hospital were identified using MicroScan system (Siemens Health Care). Up to 12 isolates each of E. faecalis or E. faecium were randomly selected per sample for virulence gene analysis.

Southern California and Puerto
Rico. E. faecalis and E. faecium strains were grown in BHI broth, incubated overnight at 37 ∘ C, and harvested by centrifugation (13,000 RPM for 5 min). The cells were washed three times in TE buffer and resuspended in 200 L 1x TE (10 mM Tris-HCl; 1 mM EDTA, pH 8.0) and lysed by heating at 95 ∘ C for 10 min. The lysed cells were transferred to tubes with glass beads, subjected to bead beating for five minutes, and centrifuged as before.

Multiplex PCR for the Detection of Enterococcal Virulence
Genes. Total DNA extracted from all isolates obtained from California and Puerto Rico was screened for enterococcal virulence genes (gelE, asa1, esp, cylA, and hyl) using PCR primers and multiplex method developed by Vankerckhoven et al. [34] with the following modifications: we used Promega Flexi Taq DNA polymerase instead of Hot-StarTaq DNA polymerase in the master mix; the initial activation step was done at 95 ∘ C for 2 min, followed by 35 cycles of denaturation (95 ∘ C for 30 sec), annealing (49.5 ∘ C for 30 sec), and extension (72 ∘ C for 2 min) and 1 cycle of elongation at 72 ∘ C for 10 min. Each set of primers has a characteristic product size to differentiate within the five virulence genes (asa1 at 375 bp, gelE at 213 bp, cylA at 688 bp, esp at 510 bp, and hyl at 276 bp). PCR products obtained by the Puerto Rico laboratory were confirmed by 1.8% agarose-gel electrophoresis (90 v, 2.5 hrs), stained with ethidium bromide, and visualized by UV transillumination (VersaDoc MP 4000). In Southern California, PCR products were visualized using the FlashGel5 (Lonza) system. 2 L of extracted DNA was diluted in 2 L FlashGel loading dye and inserted into 12 + 1-cassette wells. A 50 bp-1.5 kb DNA ladder (Lonza) was used as a molecular size marker. FlashGels were run at 150 V for up to 13 minutes. Each PCR run included a no-template control; the positive control strain used for gelE, esp, asa1, and cylA was E. faecalis MMH594 kindly donated by N. Shankar, Department of Medicinal Chemistry and Pharmaceutics, University of Oklahoma Health Sciences Center, Oklahoma City [14].

Results
A total of 170 E. faecalis and E. faecium isolates from Southern California (SC) and 247 isolates from Puerto Rico (PR) from beach water and potential sources of enterococci to beaches were analyzed for enterococcal virulence genes gelE, asa1, esp, cylA, and hyl.
esp was the most commonly found virulence gene detected among E. faecium isolates (0% to 47.9%), followed by gelE (0% to 18.8%), asa1 (0% to 12.5%), and hyl (0% to 1.3%) (Figure 2). At both study sites, human derived E. faecium isolates had the highest frequency of esp (36.4% to 47.9%). that differed in frequency depending on source. At both study locations, there was a higher prevalence of virulence genes among E. faecalis as compared to E. faecium. Among both species groups, virulence genes were less abundant among beach strains overall compared to human isolates, which was also consistent with a similar study conducted in Australia [35]. Enterococcal virulence genes asa1 (aggregation substance) and cylA (cytolysin activator) were found among E. faecalis isolates from beach water, humans, dogs, and birds, indicative of strains with enhanced virulence potential. asa1 and cylA were first identified in the genome of multidrug resistant E. faecalis strain MMH594 and have also been associated with E. faecalis pathogenicity islands [16,36]. Aggregation substance is encoded on a sex pheromone plasmid and mediates aggregation between bacteria, enabling the transfer of plasmids [37]. Cytolysins are toxins secreted by bacteria that damage cell membranes, facilitating the infection process. cylA can be carried on a plasmid or occur on the bacterial chromosome [38].

Discussion
The distribution of asa1 and cylA among E. faecalis from human clinical specimens was 90% and 70%, respectively, of E. faecalis from SC as compared to 50% and 19%, respectively, of isolates from PR. These differences likely reflect variability in the types of clinical specimens analyzed from each study location; clinical isolates of E. faecalis from SC were obtained from rectal swabs, urine, blood, abscess, ascites, vagina, and joints; those from PR were obtained primarily from urine specimens.
asa1 and esp were also found among E. faecalis strains in dogs and birds (mostly seagull), suggesting that they may be important reservoirs of strains that could potentially be transferred to humans. esp is thought to aid enterococci in evading the immune system and also form biofilm [36,39], which facilities colonization of E. faecalis in acute urinary tract infections [14]. Animals and birds have been suggested as potential sources of virulent strains to humans; gelE, asa1, esp, and cylA were detected in fecal E. faecalis isolated from dogs at veterinary hospitals [40,41], poultry [42], and ducks and wild geese [43]. The presence of these genes among E. faecalis strains from dogs and birds warrants further studies to assess potential human health risks. Among the virulence genes analyzed, gelE was the most frequently detected and widely distributed among E. faecalis strains from multiple sources, including the environment which is consistent with previous studies [9,44,45]. gelE is thought to enhance survivability of enterococci in extraintestinal environments [46].
In the beach environment, E. faecium was rare among enterococci identified from eelgrass, sewage influent, and dog samples, thus limiting the number of isolates that could be analyzed for virulence genes. E. faecium and E. faecalis were also rare or not detected in composite fecal samples from horses, goats, and pigs from PR, which is consistent with studies showing the low prevalence of these species in livestock [43,47,48]. Birds were rarely observed at the study beaches in PR, which precluded efforts to obtain enterococci isolates.
It is important to note that the presence of virulent strains among E. faecalis and E. faecium alone is not predictive of infection as there may be other mediators of pathogenicity that have yet to be elucidated [49]. It has been suggested that pathogenicity is also related to the ability of virulent strains to grow in high densities in the intestinal tract and spread to other sites in the body [50]. Host factors, such as predisposing medical conditions, immune status, and exposure to antibiotics, are also thought to play a role in the ability of enterococci to establish infection [51].

Conclusion
The low incidence of asa1, esp, and cylA among E. faecalis and E. faecium from the PR and SC beaches indicates that these virulence genes were not widely disseminated among strains found here, suggesting low potential health risks to humans. Still, the presence of E. faecalis and E. faecium harboring asa1, esp, and cylA suggests humans, birds, and dogs as potential sources of enterococci to beach water. Future surveys of enterococcal virulence genes at beaches should include those with different source inputs and populations of enterococci.