Presence of virulence factors in Enterococcus faecalis and Enterococcus faecium susceptible and resistant to vancomycin.

Despite the increasing importance of Enterococcus as opportunistic pathogens, their virulence factors are still poorly understood. This study determines the frequency of virulence factors in clinical and commensal Enterococcus isolates from inpatients in Porto Alegre, Brazil. Fifty Enterococcus isolates were analysed and the presence of the gelE, asa1 and esp genes was determined. Gelatinase activity and biofilm formation were also tested. The clonal relationships among the isolates were evaluated using pulsed-field gel electrophoresis. The asa1, gelE and esp genes were identified in 38%, 60% and 76% of all isolates, respectively. The first two genes were more prevalent in Enterococcus faecalis than in Enterococcus faecium, as was biofilm formation, which was associated with gelE and asa1 genes, but not with the esp gene. The presence of gelE and the activity of gelatinase were not fully concordant. No relationship was observed among any virulence factors and specific subclones of E. faecalis or E. faecium resistant to vancomycin. In conclusion, E. faecalis and E. faecium isolates showed significantly different patterns of virulence determinants. Neither the source of isolation nor the clonal relationship or vancomycin resistance influenced their distribution.

In recent decades, Enterococcus have emerged as important nosocomial pathogens. Currently, they are recognised as the second-most common cause of urinary tract infections and the third-most common cause of nosocomial bacteraemia ( Lindenstrau et al. 2011). Although many species can be recovered from human infections, Enterococcus faecalis is the most frequently recovered, representing more than 90% of clinical isolates, followed by Enterococcus faecium (Dahlén et al. 2012), which has shown an increase in prevalence in recent years (Fisher & Phillips 2009). Classically, E. faecalis has been linked to increased virulence, whereas E. faecium commonly exhibits multiresistance characteristics ( Sharifi et al. 2012).

Recent studies have associated enterococcal virulence with different factors, such as gelatinase production, Enterococcus surface protein (Esp), aggregation substance (AS) and biofilm formation (Chuang et al. 2009). Gelatinase is a zinc metalloprotease, encoded by gelE, with hydrolytic capacity (Lindenstrau et al. 2011). AS, encoded by a plasmid gene, mediates binding to the host epithelium and it appears to mediate bacterial aggregation during conjugation, facilitating plasmid exchange (Schlievert et al. 2010). The Esp protein, encoded by the esp gene, seems to contribute to the colonisation and persistence of E. faecalis strains in ascending infections of the urinary tract. In addition, Esp may mediate the interaction with primary surfaces and participate in biofilm formation (Chuang-Smith et al. 2010), which substantially enhances bacterial survival in biopolymers and may also be involved in antimicrobial resistance (Ballering et al. 2009).

Knowledge of the virulence characteristics of circulating Enterococcus strains may help to understand the complex pathogenic process of these opportunistic microorganisms (Sharifi et al. 2012). Data about virulence of Enterococcus circulating strains in Brazil are still scarce (Ruzon et al. 2010). Therefore, the objective of this study was to evaluate the presence of virulence factors in E. faecalis and E. faecium from clinical and surveillance samples.

SUBJECTS, MATERIALS AND METHODS

Enterococcus were isolated from inpatients in Porto Alegre, southern Brazil, between 2000-2011. The bacteria were recovered from two distinct situations: an E. faecalis outbreak that occurred mainly during 2004, with some related strains isolated from 2000 until 2009, and an E. faecium outbreak that occurred during 2010-2011. To potentially increase the genetic diversity of the strains, non-epidemiologically related Enterococcus strains recovered during the same period were examined. The isolates were from clinically relevant sites and from surveillance cultures (rectal swabs), which were performed by infection control staff during the first outbreak. The samples were maintained in 10% (v/v) Skim Milk solution (Molico, Nestlé(r)) with 10% glycerol (v/v) and stored at -20°C.

Isolates were identified by conventional biochemical tests (Teixeira et al. 2011) and susceptibility to vancomycin was determined following the Clinical Laboratory Standards Institute guidelines (CLSI 2011) using broth microdilution to determine the minimal inhibitory concentration (MIC).

Gelatinase activity was characterised as described elsewhere (Marra et al. 2007). Briefly, samples were inoculated into tubes containing 4 mL of brain heart infusion broth with 4% gelatin. After incubation at 35-37°C for 24 h, the tubes were cooled at 4°C for 30 min and the liquefaction of the medium was observed.

Biofilm formation was determined using the method described proposed by Stepanovic et al. (2000) . Briefly, microorganisms were grown in polystyrene microplates and the wells were washed, stained with crystal violet and washed again. The dye was solubilised using an alcohol solution. Then, the absorbance of the solution was measured by determining the optical density (OD) of the wells. The cut-off was defined as the mean OD of eight negative control wells. Based on the OD, the isolates were classified as follows: non-adherent, weakly adherent, moderately adherent and strongly adherent.

The presence of the gelE, esp and asa1 genes was verified by polymerase chain reaction, as described by Eaton and Gasson (2001) . The sequences of primers and annealing temperatures used are shown in Table I . Statistical analysis was performed with SPSS version 19.0, using the chi-square test. The clonal relationships among the isolates were determined by pulsed-field gel electrophoresis (PFGE) and by comparing the fragment patterns after restriction of chromosomal DNA with SmaI, as described elsewhere (Teixeira et al. 1997). Electrophoresis was performed in a CHEF DR III (Bio-Rad) apparatus.

TABLE I

Primers and annealing temperatures (AT) used to detect the

virulence genes

Gene	Primers	Sequence (5'-3')	Product (bp)	AT (°C)
gelE	TE9	ACCCCGTATCATTGGTTT	419	53
	TE10	ACGCATTGCTTTTCCATC
esp	TE34	TTGCTAATGCTAGTCCACGACC	933	52
	TE36	GCGTCAACACTTGCATTGCCGAA
agg	TE3	AAGAAAAAGAAGTAGACCAAC	1,553	52
	TE4	AAACGGCAAGACAAGTAAATA

bp: base pairs.

RESULTS

This study evaluated 50 Enterococcus isolates: 30 E. faecalis and 20 E. faecium. Twenty E. faecalis were vancomycin-resistant enterococci (VRE), with a MIC ≥ 256 µg/mL, as were all E. faecium . Table II describes the characteristics of the Enterococcus included in the study. Most of the isolates were recovered from urine (30%). Other specimens included blood, stool (surveillance cultures), abdominal secretions, catheters, wounds, soft tissue and nephrostomy secretions. Seven isolates had an unidentified source. No relationship was found between the source of isolation and the presence or activity of any virulence factor.

TABLE II

Characteristics of Enterococcus isolates included in the study

Species	Vancomycin MIC (μg/mL)	Source	Clonal type	Year
E. faecalis	> 256	Stool	FB	2000
E. faecalis	> 256	Urine (n = 2)	FA2	2001/2004
E. faecalis	> 256	Stool	FA2	2004
E. faecalis	> 256	NI	FA1	2004
E. faecalis	> 256	Urine	FA1	2004
E. faecalis	> 256	Blood (n = 4)	FA1	2004/2004/2005/2008
E. faecalis	> 256	Nefrostomy secretion	FA1	2004
E. faecalis	> 256	Stool (n = 4)	FA1	2004
E. faecalis	> 256	Catheter	FA1	2004
E. faecalis	> 256	Wound	FA1	2004
E. faecalis	> 256	NI	FA1	2006
E. faecalis	> 256	Sputum	FA1	2008
E. faecalis	> 256	Abdominal secretion	FA1	2008
E. faecalis	1	Soft tissue	SA	2008
E. faecalis	0.5	Blood	SB	2008
E. faecalis	0.5	Soft tissue	SC	2009
E. faecalis	1	NI	SD	2009
E. faecalis	2	Abdominal secretion	SE	2009
E. faecalis	1	NI	SF	2009
E. faecalis	0.5	NI	SG	2010
E. faecalis	0.25	NI	SH	2010
E. faecalis	2	NI	SI	2010
E. faecalis	0.5	Abdominal secretion	SJ	2010
E. faecium	> 256	Urine	EB	2010
E. faecium	> 256	Blood (n = 2)	EA2	2011
E. faecium	> 256	Blood (n = 5)	EA1	2011
E. faecium	> 256	Urine (n = 2)	EA2	2011
E. faecium	> 256	Urine (n = 8)	EA1	2011
E. faecium	> 256	Abdominal secretion	EA1	2011

MIC: minimal inhibitory concentration; NI: not identified.

The distribution of virulence factors among the isolates is shown in Table III . Overall, 60% (30/50) of the enterococci carried gelE . Of these, 66.6% (20/30) presented gelatinase activity. The esp gene was detected in 76% (38/50) of isolates and 38% (19/50) of all Enterococcus carried the asa1 gene. Furthermore, 74% (37/50) of the isolates were biofilm formers. Most E. faecalis were moderate or strong formers, whereas most E. faecium were not able to produce biofilm, or if so, the biofilm was weakly adherent.

TABLE III

Distribution of virulence genes, the gelatinase activity and biofilm formation among isolates

Group (n)	Clonal type (n)	agg n (%)	esp n (%)	gelE n (%)	Gelatinase n (%)	Biofilm n (%)	Virulence genes
VREfm (20)	EA1 (15)	1 (5)	13 (65)	-	1 (5)	7 (35)	agg , esp , gelatinase, biofilm
	EA2 (3)	-	2 (10)	1 (5)	-	1 (5)	esp , gelE , biofilm
	EB (1)	-	1 (5)	-	-	1 (5)	esp , biofilm
	Total	1 (5) a	16 (80)	1 (5) a	1 (5) a	9 (45) a	-
VREf (20)	FA1	11 (55)	14 (70)	18 (90)	13 (65)	17 (85)	agg , esp , gelE , gelatinase, biofilm
	FA2	2 (10)	1 (5)	-	-	1 (5)	agg , esp , biofilm
	FB	-	1 (5)	1 (5)	-	1 (5)	esp , gelE , biofilm
	Total	13 (65) a	16 (80)	19 (95) a	13 (65) a	19 (95) a	-
VSEf (10)	NRI	5 (50)	6 (60)	10 (100)	6 (60)	9 (90)	agg , esp , gelE , gelatinase, biofilm
Total	Total	19 (38)	38 (76)	30 (60)	20 (40)	37 (74)	-

a : p < 0.05; NRI: non-related isolates (unique profiles); VREf: Enterococcus faecalis resistant to vancomycin; VREfm: E. faecium resistant to vancomycin; VSEf: E. faecalis susceptible to vancomycin.

Considering vancomycin resistance as a variable, we did not find any significant difference in the presence or activity of virulence factors between resistant and susceptible enterococci. On the other hand, the presence and activity of virulence factors were significantly different between the species. Only VRE isolates were selected to compare virulence characteristics between E. faecalis and E. faecium .

E. faecalis carried gelE at a significantly higher frequency and presented gelatinase activity more frequently than E. faecium (p < 0.001). The presence of the asa1 gene was also significantly different between the two species (p < 0.001), being more commonly found in E. faecalis than in E. faecium . On the other hand, there was no difference in the frequency of the esp gene between the two species. Significantly more E. faecalis isolates formed biofilms than E. faecium isolates (p = 0.003).

We also evaluated the association between biofilm formation and the presence of the esp, asa1 and gelE genes in E. faecalis and E. faecium . Biofilm formation was significantly associated with the presence of the gelE (p= 0.001) and asa1 genes (p = 0.03), although it was not associated with the esp gene (p = 0.06). Nevertheless, all strongly adherent isolates carried, along with gelE, the esp gene and more than 70% of the strongly adherent isolates carried all three virulence genes.

PFGE analysis was performed in three groups of isolates: E. faecalis resistant to vancomycin (VREf) - representative isolates of the first outbreak period), E. faecium resistant to vancomycin (VREfm) (representative isolates of the second outbreak period) and E. faecalis susceptible to vancomycin (VSEf), which was included to possibly increase the genetic variability. Although VREf and VREfm presented a clonal relationship, with a major clone each, the distribution of virulence factors was not homogeneous among the isolates (Table III ). The major clones of VREf and VREfm represented 19 of each 20 isolates, with two subclones that were highly related. On the other hand, VSE demonstrated a highly heterogeneous PFGE profile (no clonal relationship among the isolates was observed), with most isolates presenting all evaluated virulence determinants.

DISCUSSION

Studies focused on enterococcal virulence are complex because the essential factors for pathogenicity have not yet been described. Indeed, virulence in this genus has been typically considered a multifactorial process, with the participation of several genes and their products.

Production of gelatinase is a characteristic mediated by the gelE gene and phenotypically expressed in vitro by liquefaction of a culture medium containing the substrate (Tsikrikonis et al. 2012). We observed Enterococcus isolates that did not produce the enzyme, even though gelE was detected. Indeed, according to Marra et al. (2007), the presence of gelE is not necessarily correlated with gelatinase activity. Some studies suggest that other genes may be associated with gelE expression control. Indeed, Lindenstrau et al. (2011) indicated that mutated genes affect gelE gene expression and possibly regulate gelatinase production. These findings reinforce the complexity of the processes involved in Enterococcus virulence.

There are diverging opinions concerning the role of Esp in biofilm production (Garth et al. 2008). Dworniczek et al. (2012) concluded that there is no clear relationship between the expression of esp or gelE and biofilm formation. Indeed, an analysis by Sillanpää et al. (2010) showed efficient biofilm production in the absence of Esp in E. faecium isolates. On the other hand, Chuang-Smith et al. (2010) speculated that Esp may mediate the interaction with primary surfaces and participate in the formation of this phenotype. In addition, Heinkens et al. (2007) showed that Esp is involved in biofilm formation in E. faecium isolates. As our results did not show an association between the presence of esp and biofilm production, we assume that other factors are associated with this phenotype.

We observed a considerable number of E. faecium carrying esp compared with asa1 and gelE genes, which were detected at very low frequencies among this species. Some clones of vancomycin-resistant E. faecium that are well established in hospitals show high positivity for esp, suggesting that this gene may play an important role in the virulence process ( Willems & Bonten 2007). Besides esp, Willems et al. (2005) detected some other genes and combinations of mutations and recombinations that give nosocomial E. faecium clones adaptive advantages that contribute to their dissemination.

As with esp, gelE gene participation in biofilm formation is also controversial. Some authors have observed that, in fact, the presence of gelatinase enzyme can affect the virulence and the process of biofilm formation in Enterococcus (Dworniczek et al. 2012). However, other authors (Ballering et al. 2009) have demonstrated that gelE is not associated with such phenotypes. Nevertheless, our study demonstrated that gelatinase expression may have influenced biofilm formation.

AS is one of several Enterococcus adhesins that can be targeted for therapy against infectious endocarditis. The expression of asa1 has an important effect on biofilm formation because this genotype promotes the accumulation of adherent microorganisms to a surface (Schlievert et al. 2010). One study identified several functional subdomains in the Enterococcus asa1 gene and mutations in these loci drastically reduced the ability to form biofilm (Chuang-Smith et al. 2010). Our study corroborates these findings, since asa1 was significantly more frequent in biofilm-producing strains.

However, based on our statistical analysis, biofilm formation could not be linked to any specific gene. In fact, this phenotype is multifactorial and depends on a number of genes working together along with extrinsic factors. So far, several other genes or gene sets have been reported as auxiliaries in biofilm formation in Enterococcus, which highlights the complexity and the multifactorial nature of this trait (Dworniczek et al. 2012).

Vancomycin resistance was not associated with more virulent strains in our study. Indeed, according to Giridhara et al. (2010), there is no significant difference in virulence factors, ability to cause infection or vancomycin susceptibility among Enterococcus isolates . Acquired resistance and virulence factors may compromise microorganism fitness. Thus, multiresistance characteristics in strains presenting restricted virulence arsenals are a common feature and are frequently seen in opportunistic pathogens (Foucault et al. 2010, Rigottier-Gois et al. 2011 ).

Molecular epidemiological studies have reported the spread, around the world, of a hospital-adapted complex of E. faecium designated epidemic clonal complex-17 (CC17), which is associated with the majority of hospital outbreaks and clinical infections on all continents. The global success of CC17 seems to have been facilitated by the cumulative acquisition of antibiotic resistance, putative virulence traits and the ability to acquire different mobile genetic elements (Deshpande et al. 2007). Although the esp gene seems to be a common characteristic of CC17 strains, its occurrence is not homogeneous (Brilliantova et al. 2010). Brilliantova et al. (2010) evaluated 129 vanA -positive E. faecium isolates, which included two major clones belonging to CC17. Virtually all isolates (91%) were positive for the gelE gene, in contrast to the 5% positive rate determined in our study. On the other hand, Brilliantova et al. (2010) did not observe any asa1 -positive isolates, which is in accordance with our findings.

Our VREfm group presented some characteristics of CC17, such as resistance to ampicillin and non-susceptibility to quinolones (data not shown). Moreover, the esp gene was highly prevalent, but not always present. Further studies must be done to determine the sequence type of our isolates.

VREf enterococci have been observed by virtually all authors to have a high prevalence of virulence determinants (Udo & Al-Sweih 2011, Hasani et al. 2012). Nevertheless, Sharifi et al. (2012 ) studied the distribution of virulence factors and they observed a high frequency of all virulence factors (including gelE and asa1) except esp, which reinforces the complexity of the virulence process in this genus.

Although one could expect that virulence traits change depending on the source of isolation, we did not find any such association. Dahlén et al. (2012) compared E. faecalis recovered from oral mucosal infections and deep infections and they did not find any discrepancies in the distribution of asa1, gelE or esp between both groups. Significant differences in the distribution of virulence factors are potentially achievable by comparing infecting and colonising isolates. However, Johansson and Rasmussen (2012) compared virulence factors (ace, asa1, gelE, esp) of E. faecalis from infective endocarditis and normal flora and observed that only biofilm formation had a significantly different occurrence (more pronounced in the normal flora group). Indeed, the source of isolation (including surveillance cultures) did not influence the virulence factor distribution among our enterococcal population.

In terms of molecular epidemiology, it was impossible to associate any virulence factor with any specific subclone of VREf or VREfm. On the other hand, among the highly heterogeneous population of VSEf, the distribution of virulence factors was homogeneous, with most isolates presenting almost all the evaluated virulence determinants. Thus, molecular relatedness does not affect virulence gene distribution (Willems et al. 2005).

Our study has some limitations. The small number of isolates could have compromised the statistical analysis, as could the clonal relationship among them. Moreover, our isolates of VREfm and VREf were closely related and potentially do not represent the real distribution of the putative virulence factors in the enterococcal population. Studies with a higher number of heterogeneous VREs should be performed to clarify each one's role in enterococcal pathogenesis.

In the results presented here, E. faecalis and E. faecium isolates showed significantly different patterns of virulence determinants, which reinforces the findings of other authors (Hasani et al. 2012, Sharifi et al. 2012). In addition, neither the source of isolation nor the clonal relationship or vancomycin resistance influenced the distribution of virulence determinants.

Given the increasing importance of Enterococcus as nosocomial pathogens, the identification of virulence factors associated with invasiveness and disease severity has become an important subject of concern. The development of other mechanisms to control infection, such as preventing Enterococcus biofilm formation or inhibiting the action of other virulence factors, may provide an alternative method of therapy, especially considering that antimicrobial treatment is challenging for this genus.