Phenotypic and Molecular Characterization of Plasmid- Mediated Virulence and Antimicrobial Resistance Traits among Multidrug Resistant Enterococcus Spp. in Egypt

Enterococcus spp. are remarkable multidrug resistant (MDR) bacteria that are causing serious healthcare-associated infections. The current study investigated the frequency of Enterococcus spp., antimicrobial susceptibility, biofilm formation and the presence of some plasmid-mediated virulence characters and antimicrobial resistance determinants in enterococcal isolates from Egyptian hospitals in Cairo. Enterococcus bacterial isolates were recovered from different clinical specimens and identified using biochemical testing and KB005A HiStrepTM identification kit. Kirby-Bauer disc diffusion method and/or broth microdilution method were used to determine the antimicrobial susceptibility patterns. Phenotypic assays were performed to study biofilm formation and cytolysin and gelatinase production. PCR assays targeting the plasmid-carried genes aac(6’)-aph(2’), aph(3)-IIIa, vanA, agg and cylA were performed. In this study, 50 isolates of diverse Enterococcus spp. were identified with E. faecium was the most frequently isolated one. High resistance profiles were determined against tested antimicrobials and all isolates were MDR. Moderate biofilm formation was detected in 20% of isolates, 18% showed complete blood hemolysis and 12% produced gelatinase. All isolates carried the tested aminoglycosides resistance genes, while vanA was found only in 4 isolates (8%). The virulence genes agg and cylA were detected in 4% and 32% of isolates, respectively. In conclusion, E. faecium was the most prevalent species. The entire isolates set were MDR and the plasmid-carried aminoglycoside resistance genes were extensively disseminated among MDR isolates. Thus, regular surveillance studies, from the area of study or other geographical regions in Egypt, and strict infection control measures are required to monitor the emerging MDR enterococci.


METHODOlOGy Enterococci isolates and species identification
The clinical isolates of Enterococcus spp., in this study, were recovered from 240 various clinical samples including 92 urine samples, 91 pus samples and 57 blood samples. These samples were routinely collected and processed by the dedicated team through the medical care of patients in four Egyptian hospitals in Cairo including Homiat Elabasia Hospital, Elsayed Galal University Hospital, El Hussien University Hospital and Helwan General Hospital, over the period from November 2015 to March 2017. The study was approved by the ethical committee of the Faculty of Pharmacy (Girls), Al-Azhar University. The Enterococcus isolates were identified by Gram stain, cultural characteristics on different microbiological media such as cysteine, lactose, and electrolyte-deficient (CLED) agar (CONDA, Madrid, Spain), Bile Esculin agar (CONDA, Madrid, Spain), HiCrom UTI agar (HiMedia, Mumbai, India), in addition to biochemical testing. The different species of Enterococcus were identified using the KB005A HiStrep™ Identification Kit (HiMedia, Mumbai, India) based on the biochemical activities of enterococci isolates 12 .

Phenotypic assays for determination of virulence traits Cytolysin production
Detection of blood haemolysis capability due to cytolysin production was performed according to Chow et al. 14 .

Gelatinase production assay
Gelatinase activity was investigated according to the study of Pollack et al. 15 .

Biofilm formation assay
The ability of biofilm formation was assayed by the microtiter plate method according to Mohamed et al. 16 Optical density (OD) of stained adherent biofilm was determined using ELISA auto reader (Model 680, Biorad, UK) at a wavelength of 570 nm. This assay was performed in triplicate wells and three independent repeats. The result of biofilm formation assay was interpreted as follows: if mean of the three repeats OD readings ≤ ODc (3 × SD + mean OD of the negative control), it was considered as a non-biofilm producer, and if ODc ˂ OD ≤ 2 × ODc, it was considered as weak biofilm producer, if 2 × ODc < OD ≤ 4×ODc, it was considered as moderate biofilm producer and if 4×ODc < OD, it was considered as strong biofilm producer. Staphylococcus aureus ATTCC 29213 was the positive control in this assay and the negative control was sterile tryptic soy broth (TSB) 17 .

Plasmid DNA extraction and detection of purified plasmids
Overnight

Statistical analysis
Results were statistically analysed using SPSS software (version 14.0, Chicago, IL, USA).
The P-value of less than 0.05 was considered as statistically significant. The correlation between different virulence factors and biofilm formation was evaluated by the Chi-Square (χ2) test.
Study of the antimicrobial resistance profiles of the tested isolates showed that all isolates showed resistance to ampicillin, ciprofloxacin, tetracycline, rifampicin and tobramycin, while 68%, 48%, 80%, 56% and 8% of the entire isolates collection were resistant to high-level streptomycin, high-level gentamicin, chloramphenicol, nitrofurantoin and linezolid, respectively. The 50 isolates were sensitive to teicoplanin by either disc diffusion method or MIC determination. Antimicrobial susceptibility results are listed in Table 2. Concerning resistance to vancomycin, although disc diffusion method indicated resistance to vancomycin, MIC determination, using the broth microdilution assay, indicated the sensitivity of all isolates as MIC values were ≤ 4 µg/ml except one E. faecium isolate which showed intermediate resistance to vancomycin (MIC = 8 µg/ml). Resistance to teicoplanin was also studied by determining MIC using the broth microdilution assay. These assays indicated that all isolates were sensitive to teicoplanin with MIC ≤ 1 µg/ml.

Plasmid analysis and the Prevalence of some plasmid -carried virulence and vancomycin and aminoglycosides resistance genes
Agarose gel electrophoresis-based analysis of the extracted plasmid DNA from 50 isolates indicated intact isolated plasmids in  Fig. 2).

Correlation between biofilm-forming capacity and antimicrobial and virulence factors
The correlations between biofilm-forming ability and incidence of different antimicrobial or virulence factors encoding genes were insignificant as P > 0.05 (Table 4).

DISCuSSION
T h e M D R e nte ro c o c c i b a c te r i a are emerging as a leading cause of healthcare opportunistic infections in hospitalized patients, mostly the immunocompromised 22-25 . Consequently, it is important to control and prevent the outbreak of Enterococcus infections by regular identification of the prevalence and antimicrobial resistance patterns of these bacteria 26,27 . The current study explored the prevalence and antimicrobial susceptibility of Enterococcus spp. in Cairo, Egypt. Furthermore, the study aimed to find some plasmid-mediated antimicrobial resistance and virulence factors encoding genes among isolated Enterococcus species and study the correlation between the antimicrobial and virulence factors phenotypes and/or the presence of their corresponding plasmid-carried genes.
In the current study, 50 Enterococcus isolates were recovered from 240 clinical samples representing 20.8% of bacterial isolates during the period of study. This frequency was almost like the frequencies reported in other geographical regions in Egypt as well as other countries such as Iran and Pakistan [28][29][30] .
Five different Enterococcus species were identified in the study. The most predominant species was E. faecium followed by E. faecalis representing 46% and 30% of total Enterococcus isolates, respectively. The other Enterococcus species identified in lower percentages were E. pseudoavium (12%), E. avium (6%) and E. raffinosus (6%). These findings were consistent with other previous studies 3, 31 . However, earlier studies in the same geographical region of the current study reported a high prevalence of E. faecalis over E. faecium 21,32,33 . Besides, regarding the type of clinical samples, the highest number of MDR Enterococcus isolates was recovered from urine samples. This finding is comparable to previous studies from Egypt as well as other countries 10,28,32,33 .
This high distribution of MDR Enterococcus spp., particularly in developing countries, is partly due to the widespread and misuse of antimicrobial drugs which consequently results in limitation in the therapeutic options 35, 36 . In addition to the disc diffusion method, resistance to glycopeptides was determined by broth microdilution method as per CLSI recommendation 13 , since the disc test does not differentiate vancomycin-susceptible isolates from vancomycin-intermediate strains 37 . The 50 enterococcal isolates showed high sensitivity to teicoplanin by both disc diffusion and broth microdilution assay. These records were relatively comparable to other studies 21,35, [38][39][40] .
Contrary to the disc diffusion method, which showed full vancomycin-resistant enterococci (VRE) profile; broth microdilution assay revealed only one E. faecium isolate had intermediate resistance pattern to vancomycin showing a VanB phenotype since susceptible to teicoplanin by PCR we detected a plasmidic vanA gene in this isolate, in other 3 strains still completely susceptible to vancomycin, these incongruent Enterococcus genotypes-phenotypes should be related to mutations or deletions in the VanA operons as previously described [41][42][43] .
Results suggested that both MIC and PCR should be taken into consideration in evaluating the susceptibility to vancomycin and identification of Vancomycin incongruence genotypes-phenotypes isolates, respectively and recommended the exclusion of vancomycin as therapy in patients were these isolates are detected.
The Plasmid carrying-vancomycin resistance gene vanA is common among Enterococcus isolates in Europe and the United States 44 . The low VRE occurrence here found (8%) agreed with several studies in Cairo, Egypt that reported no or the same low frequency of VRE 21,35,44 . Strains like these capable of expressing resistance to vancomycin, if used is worrying, because these isolates are not revealed biochemically but only by molecular method.
Concerning aminoglycosides resistance in this study, HLGR and HLSR were observed in 48% and 68% of Enterococcus isolates, respectively, while 38% of isolates showed both HLGR and HLSR. The higher frequency of HLSR over HLGR in this study is inconsistent with previous studies where HLGR was more prevalent 2,3,45,46 while it is agreed with another study 47 . That could be explained by the extensive use of gentamicin over streptomycin in the hospitals involved in the current study. Although, both findings designate that the resistance to streptomycin does not always correlate with resistance to gentamicin 48 .
Linezolid is the drug of choice for treatment of drug-resistant Enterococcus infections; however, the emergence of linezolid resistance in enterococci 39 . In the present study, fortunately, 4 (8 %) isolates only were resistant to linezolid, which was consistent with previous studies 39,46 .
Plasmid profile is main epidemiological tool to study genetic diversity and the dissemination of antimicrobial resistance among bacteria 49 . In this study, different patterns were detected among enterococcal isolates indicating diversity of isolates due to the higher exposure of people to Enterococcus spp. in this geographical area and/ or high rate of plasmid transfer amongst bacteria. The high-level of resistance to aminoglycosides among Enterococcus species is attributed to acquiring the extrinsic resistance genes encoding the aminoglycoside-modifying enzymes. HLGR and HLSR in enterococci are mostly mediated by the plasmid-carried aac(6')-Ie-aph(2'')-Ia andaph(3')-IIIa, respectively 47 . The PCR investigation of the plasmid-carried aminoglycoside resistance genes aac(6')-aph (2')and aph(3)-IIIa in our study revealed the presence of both genes in all tested isolates (Table 3, Fig. 1). While the inconsistency between the phenotypic results and carrying of these genes is similar to the findings of another study in India 50 where 26.02% of isolates were phenotypically resistant to high-level gentamicin, and 15% were resistant to high-level streptomycin, although 94.75% of isolates were carrier for aac (6′)-Ie-aph (2″)-Ia which encodes a very powerful bi-functional aminoglycosidemodifying enzymes resulted from the fusion of two genes. Other studies reported a high prevalence of aminoglycoside resistance that expressed phenotypically equivalent to the presence of the encoding genes aac(6')-Ie-aph(2'')-Ia and aph(3')-IIIa 46,47 . These findings might indicate that the phenotypic expression of HLGR or HLSR may require the presence of multiple aminoglycosides resistance mechanisms and/or genes other than targeted ones in our study. Also, it can be explained by the interaction among multiple genes involved in this trait and/or the environmental factors that can impact phenotype, besides some alleles may be silenced by other genes that minimize their expression.
HLGR and HLSR encoding genes are also found in Enterococcus spp. other than the most common species faecalis and faecium. Moreover, significantly, the coexistence of both genes was detected in all tested isolates which is the first time to be reported. This can be explained by the high possibility and dissemination of plasmid-mediated horizontal gene transfer among enterococci in the geographical area of this study, which leads to the rapid dispersion of resistance to aminoglycosides 51 . In the current study, 32% of Enterococcus isolates were positive for cylA gene although cytolysin production indicated by complete (β) haemolysis was phenotypically observed in 18% of total isolates. All isolates showed β-haemolysis were carrier for cylA gene. These findings were consistent with a previous study where the percentage of phenotypic expression of cytolysin was 16.7% 52 and with another study where the phenotypic and genotypic expression of cytolysin was 8% and 16%, respectively 21 . The phenotypic expression of cytolysin was in lower frequency than that of the detected encoding gene. This could be attributed to the fact that the expression of cytolysin requires other proteins or factors encoded by genes on cylL locus 24 . Each the phenotypic and genotypic profile of cytolysin was high among E. faecalis than in E. faecium isolates which agreed with previous studies 11,21,53 .
Screening of biofilm formation capability among MDR isolates, 20% of isolates showed moderate biofilm production and 24% of isolates were weak biofilm producers while the remaining isolates (56%) were non-biofilm producers. Out of 10(20%) moderate biofilm producers isolates; 4 were E. faecium. These records are less comparable to previous studies where a higher percentage of strong, medium and weak biofilm producers were documented 32,54,55 . This could be explained that their isolates were mostly recovered from urine which considered as one of the main sources in which biofilm can develop a severe problem. Biofilm can be formed in the urothelium, prostate stones and implanted foreign bodies 56 . Aggregation substances (agg) mediate the transfer of mobile genetic elements and the formation of aggregates during conjugation 57,58 .
Only 2 (4%) Enterococcus isolates (one E. faecalis isolate and one E. avium isolate) were carriers for agg gene in the current study; one of them was categorized phenotypically as moderate biofilm producer and the other was non-biofilm producer. This record is like the study of Hashem et al. 21 from the same geographical region Cairo, Egypt, which recorded the frequency of 3% of agg gene. Conversely, other studies reported much higher frequencies where agg gene was carried by moderate and strong biofilm forming isolates 3,54,59 . Gelatinase action has been known as one of the initial steps in the biofilm formation 57 . In the current study, 24% of isolates were capable to hydrolyse gelatine indicated by gelatine liquefaction. This record was comparable with previous studies 32, 55,56 .
There was no significant correlation between virulence factors investigated phenotypically and/or distribution of their encoding genes to biofilm formation, and between biofilm-forming capacity and incidence of other virulence factors was found to be statistically not significant (P < 0.05) calculated by Chi-Square (χ2) test, although moderate and weak biofilm production was found to be frequently associated with gelatine liquefaction capability 32,56 .

CONCluSION
Regular recent data on antimicrobial resistance profile is important for treatment guidelines for infections caused by MDR Enterococcus spp. The resistance of Enterococcus spp. to aminoglycosides is of great concern, since it limits the therapeutic options of serious Enterococcus infections even by the synergy with beta-lactam antibiotics. Thus, strict infection control measures in hospitals, as well as a public health strategy on the appropriate prescribing and the rational use of antimicrobial drugs must be followed to limit the spread of MDR enterococci. Besides, more reliable screening of Enterococcus isolates for vancomycin resistance would be considered including the recommended MIC tests over the disc diffusion method. Consequently, it should be taken into consideration in evaluating the susceptibility to vancomycin that depending only on the disc diffusion test may result in the exclusion of this effective antibiotic from the empirical antibiotic therapy. while it is recommended that MIC tests should be performed to determine the susceptibility to vancomycin. Furthermore, molecular detection should be used to detected vancomycin susceptible Enterococci carrying low or unexpressed van genes.

DATA AVAIlABIlITy
All datasets generated or analysed during this study are included in the manuscript.