Infl uence of a subinhibitory concentration of vancomycin on the in vitro expression of virulence-related genes in the vancomycin-resistant Enterococcus faecalis

Introduction: Exposure to subinhibitory concentrations (SICs) of antimicrobials may alter the bacterial transcriptome. Methods: Here, we evaluated the expression of nine virulence-related genes in vancomycin-resistant enterococci (VRE) urinary tract infection isolates grown at SICs of vancomycin. Results: A Subinhibitory concentrations of vancomycin interferes with gene modulation, but does not affect the phenotype of a VRE strain in vitro. Conclusions: Subinhibitory concentrations of vancomycin may regulate the expression of virulence factors in vivo or contribute to the selection of vancomycin-resistant strains.

Enterococcus faecalis, an important microorganism that causes healthcare-associated infections, is part of the normal human microbiota.Enterococcus faecalis is the fi fth most common pathogen that causes catheter-associated urinary tract infections (UTIs) (1) .Several different virulence factors have been reported in clinical strains of E. faecalis, including biofi lm formation and the expression of surface adhesion components.The ability of E. faecalis to adhere to medical devices such as ureteral stents and catheters and to develop biofi lms is likely associated with its pathogenicity (2) .
The biofi lm on plastic (BOP) operon is essential for biofi lm formation (3) .Adhesion of collagen of enterococci (ACE) and aggregation substance (ASC10) are proteins that are also crucial for biofi lm production (4) (5) .The sulfur assimilation system (SUF) is responsible for iron-sulfur cluster biogenesis and is essential for cellular survival, particularly during aerobic growth or oxidative stress.The SUF operon, which is associated with virulence, is controlled by the regulatory protein sensor for oxidative stress (OxyR) and the ferric-uptake regulator (FUR) (6) .
Vancomycin-resistant enterococci (VRE) have emerged as important nosocomial pathogens worldwide.The VanA phenotype (associated with vanA) is responsible for high levels of resistance to vancomycin (7) .The clinical use of vancomycin is a subject of major concern, because continuous monitoring is required to maintain a concentration of approximately 15mg/L of the drug in patients' sera.Occasionally, because of the complex pharmacokinetics of vancomycin, this value is not achieved and patients are subjected to subinhibitory concentrations (SICs) of the drug.Consequently, frequent incidences of vancomycin failure and poor outcomes have been observed recently.Under these conditions, there is a high probability of selecting resistant or heteroresistant isolates (8) .
Exposure to SICs of antimicrobials may alter the transcriptomic and phenotypic responses of pathogenic bacteria (9) (10) ; however, the effect of SICs of antimicrobials on virulence remains unclear.The aim of this study was to evaluate the expression of virulence-related genes in an E. faecalis VRE strain isolated from a urinary tract infection and grown under SICs of vancomycin.The bacterial strains used in this study are listed in Table 1.The growth capacities of four E. faecalis strains were examined, and the strain with the highest optical density (OD) value of biofi lm formation was selected for transcriptional profi ling.All the E. faecalis strains were streaked onto Tryptic Soy Broth (TSB) agar and incubated at 37°C overnight (ON).Four individual colonies were used to inoculate the same tube of 5mL of 2× YT broth (1% yeast extract, 1.6% tryptone, and 1% NaCl) and grown ON at 37°C.Urine from six healthy men and women volunteers with no history of UTIs or antibiotic use in the last 6 months was collected, pooled, centrifuged, sterilized twice by fi ltration, and added to 2× YT broth to make nine different sample concentrations: 50%, 25%, 12%, 10%, 8%, 6%, 4%, and 2% (2×YT-U50-2×YT-U2) and pure urine (100%).The positive control contained only 2× YT broth.The concentration of vancomycin chosen for the assays (64µg/mL) was based on range 64-1,000µg/mL, which has been shown to produce high levels of resistance (7) .
All the strains were used to inoculate plates containing 2× YT agar and incubated at 35°C ON.All the isolates were then evaluated for their ability to form biofi lms in ten different media: 2× YT broth (control), eight different concentrations of urine (2-50%) and 100% urine under two conditions -drug-free condition (DFC) and vancomycin condition (VC) -by using the crystal violet assay (11) .Four to eight colonies were diluted in 0.9% sterile saline (w/v) until the turbidity matched that of a 0.5 McFarland standard [approximately 1 × 10 8 colonyforming units per milliliter (CFU/ml)].Eight wells of a 96-well fl at-bottomed microplate were fi lled with 180μL of each media type and 20µL of bacterial suspension.The optical density (OD) was measured at 492nm in a spectrophotometer, and the OD of each strain was determined by comparing the arithmetic mean of the absorbance of the wells with the mean absorbance of the negative controls.The strains were categorized on the basis of their ODs into the following groups: non-adherent, weak biofi lm producers, moderate biofi lm producers, and strong biofi lm producers (11) .Non-inoculated 2× YT wells were used as negative controls, and Staphylococcus epidermidis ATCC 35984 was used as a positive control.The OD 492 of each well was measured; all tests were carried out in triplicate.The effects of the different media on bacterial growth and biofi lm formation were evaluated using the paired t-test (level of signifi cance = 0.05).
Eleven genes (vanA, bopA, bopB, bopC, bopD, ace, asc10, fur, oxyR, tuf, and 23S) were analyzed by quantitative polymerase chain reaction (qPCR).The oligonucleotides used in this study are listed in Table 2.Only one strain, exhibiting the maximum OD value, was selected for transcriptional profi ling.The qPCR analyses were performed at least three times.Fivehundred microliters of ON culture were inoculated in 49.5mL of 2× YT-U10 in the absence (DFC) and presence (VC) of SICs of vancomycin.Aliquots were collected and extracted at the early-exponential phase (OD 650 0.3), total ribonucleic acid (RNA) was purifi ed and quantifi ed, and cDNAs were synthesized (12) .Quantitative PCRs were performed as uniplex reactions in fi nal volumes of 20µL with 100pg of cDNA using the Eco TM Real Time PCR System (Illumina ® San Diego, USA) and universal cycling protocols.
We analyzed the effect of different concentrations of urine on biofi lm formation by phenotypic characterization and on the basis of the OD values.Biofi lm formation in the VRE strains was negatively infl uenced by the presence of pure urine at 35 °C (no growth was observed).With urine concentrations of 2-50%, all the strains showed growth (OD, 0.016 ± 0.004); however, they were unable to form biofi lms under both conditions.The UTI-2389 isolate produced the highest OD value (0.021 ± 0.004).No signifi cant association was observed between biofi lm formation and urine concentration.These fi ndings might be a consequence of the in vitro environment, which lacked essential components of the in vivo urinary system such as urothelial cells, glucose, mineral salts, and albumin.Moreover, it has been suggested that biological signals in human urine play an important role in modulating virulence at enterococci infection sites (13) .On the basis of our biofi lm assay, all the strains could be classifi ed as non-adherent according to both DFC and VC.Although maximal biofi lm induction has been observed at a concentration of 3/4 th the minimal inhibitory concentration (MIC) (10) , in other studies, vancomycin at concentrations ≤ 1/2 the MIC (9) had little or no effect on S. epidermidis biofi lm formation.Thus, we concluded that the low concentration of vancomycin used in our study (1/6 th the MIC) produced no effect on biofi lm formation, which is consistent with the OD values observed in our study.
On t he basis of the results of our biofilm assays, we selected the UTI-2389 strain and 10% urine (2×YT-U10) for subsequent qPCR analysis.The relative quantifi cation of the messenger ribonucleic acid (mRNA) concentrations is  shown in Table 3. Signifi cant upregulation of vanA, which is vancomycin-inducible (7) , was detected under the VC condition.We used this result to investigate the up/downregulation of the other selected genes.Genes of the bop operon were partially regulated in the presence of SICs of vancomycin.Treatment with SICs of vancomycin signifi cantly upregulated bopA and bopC and slightly downregulated bopB and bopD.Similar partial regulation of the bop operon in a VRE strain has been reported previously (14) under growth conditions lacking antimicrobial agents, which suggests that this operon behaves erratically, even under normal growth conditions.Under the conditions tested, the expression of ace was signifi cantly downregulated.However, in a previous study, ace expression was scarcely detected (15) in cells at the mid-and late exponential phases and was not detectable in cells during the stationary phase, suggesting that ace transcription is very low under standard in vitro growth conditions.The expression of acs10 was induced slightly by SICs of vancomycin.ASC10 is expressed in the presence of an intercellular signal cCF10, which is a peptide pheromone (4) .Moreover, antibiotics can act as signaling molecules; therefore, vancomycin could have induced asc10 expression.We also detected minor downregulation of oxyR.During intracellular oxidative stress, the E. faecalis SUF system is induced by OxyR and requires the expression of the integration host factor (6) .
In this in vitro study, no hydrogen peroxide production was observed, which may explain the lack of stimulation of oxyR.
In contrast to oxyR, the expression of the ferric uptake regulator fur was signifi cantly upregulated in response to the stress caused by vancomycin, indicating that this gene is not regulated in the same manner as oxyR.For example, fur may operate in response to a cofactor in cell homeostasis, because the SUF system is the housekeeping machinery for E. faecalis (6) .
In summary, we present here, for the first time, data on the behavior of VRE from a UTI strain in the presence of SICs of vancomycin.The nine genes evaluated in this study are related to virulence factors that are often present in Enterococcus faecalis.Although there was an increase in the expression of several VRE genes in the presence of SICs of vancomycin, no differences in their phenotypes were observed in vitro.Although the data presented here are from in vitro assays, SICs of vancomycin may contribute to a similar increase in the expression of virulence factors in vivo, leading to adverse clinical outcomes and to the selection of vancomycin-resistant strains.In-vitro experiments may not effectively represent the complex in vivo conditions during enterococcal urinary tract infection.Therefore, additional in vitro and in vivo studies should be carried out to corroborate our fi ndings.The current study is limited because the strains tested were isolated at local hospitals before the study commenced, meaning that we have no patient records detailing their treatments or clinical outcomes.
Our main interest was to investigate the behavioral response of a clinical VRE strain against SICs of vancomycin.Therefore, we did not perform tests with other antimicrobials or compare our results to a reference strain.

TABLE 3 -Delta cycle threshold (ΔCt) and ratio values of the expression of virulence related genes in a vancomycin-resistant enterococcus strain grown in 2× YT + 10% urine in absence (DFC) and presence (VC) of subinhibitory concentrations of vancomycin.
ΔCt: cycle threshold variation; YT: yeast extract, tryptone and broth; drug-free condition; VC: vancomycin condition; E: effi ciency of the reaction; R: the relative expression ratio is the average value of the normalized based on the reference genes; vanA: vancomycin resistance gene A-type; bop: biofi lm on plastic; ace: adhesion of collagen of enterococci; asc10: aggregation substance; fur: ferric-uptake regulator; oxyR: oxidative stress regulator; tuf: elongation factor Tu; 23S ribosomal RNA; na: not applicable; *Level of signifi cance = 0.05.