Important Virulence Factors and Related Genes in Uropathogenic E. coli and their Relation to Fluoroquinolone Resistance

Escherichia coli (E. coli) is the predominant pathogen causing urinary tract infection. E. coli expresses a wide spectrum of virulence factors that enable it to cause the disease. The present study was conducted on 50 E. coli isolates to assess some important phenotypic virulence factors of E. coli: haemolytic activity, cytotoxicity and haemagglutination and their relations to the following virulence genes: hlyA, cnf-1 and papC. Fluoroquinolones susceptibility was also assessed. Fifty E. coli isolates were isolated from non-hospitalized patients with UTIs and identified by colony morphology and standard biochemical reactions. The isolates were examined phenotypically for haemolytic activity using 5% sheep blood agar, cytotoxicity by their cultivation on HEp-2 cell line and haemagglutination of group O human RBCs. Virulence genes were detected by multiplex PCR. Fluoroquinolones susceptibility was done using disk diffusion method. Phenotypic methods showed that 30% of the isolates were bhaemolytic, 18% of the isolates showed haemagglutination with human RBCs, with higher prevalence for MRHA (12%) than MSHA (6%), and 92% of the isolates showed morphological changes, detachment and lysis in HEp-2 cells. Multiplex PCR showed that the most prevalent gene was cnf-1 (30%), while 26% of isolates were positive for hlyA and 26% of isolates were positive for papC. Only one UPEC isolate showed the simultaneous presence of papC, cnf-1 and hlyA, while 10 isolates were harbouring hlyA and cnf-1, 2 isolates were harbouring cnf-1 and papC genes and 2 isolates were harbouring hlyA and papC simultaneously. Fluoroquinolones susceptibility showed that 60% of the isolates were resistant to the used fluoroquinolones.

Urinary tract infection (UTI) is among the most common bacterial infectious diseases encountered at all ages. Escherichia coli is being the etiologic agent in 50-80%. (Farell et al., 2003;Matute et al., 2004;Zhanel et al., 2006). Certain strains of E. coli are consistently associated with uropathogenicity and are designated as uropathogenic E. coli (UPEC) (Connell et al., 1996). UPEC strains encode widespread virulence factors closely related to colonization, persistence, and pathogenesis of bacteria in the urinary tract; the most important of these factors include adhesins or fimbriae, biofilm formation, siderophores and toxins such as haemolysins (Hly) (Dhakal et al., 2008) and cytotoxic necrotizing factor-1 (CNF-1) (Wiles et al., 2008 a ). The virulence factors among UTI agents are really different and clinical isolates could not be detected as UPEC by using traditional methods. Therefore, using molecular detection methods is critical for identification of UPEC strains (Saraylu et al., 2012). In clinical management, it is usually accepted that resistance, especially multidrug resistance, equates to the virulence of the strain. However, molecular studies have indicated an inverse relationship between the distribution of virulence factors and antimicrobial resistance, particularly to quinolones as well as fluoroquinolones (Vila et al., 2002;Johnson et al., 2003;Moreno et al., 2006;Piatti et al., 2008).

Aim of Work
The present study was conducted to detect some important virulence factors of E. coli phenotypically: haemolytic activity, cytotoxic activity and haemagglutination and related virulence genes: hlyA, cnf-1 and papC, by using multiplex PCR. Study the relation between the bacterial resistance to fluoroquinolones and the virulence potential of E. coli.

MAteriAls And Methods
The study has been approved by the Research and Ethical Committee of Medical Microbiology and Immunology Department, Faculty of Medicine, Cairo University.

isolation and identification of E. coli
Urine samples from non-hospitalized cases of urinary tract infections were collected from clinical laboratories of Kasr El Ainy hospital. Microscopic examination of a wet film of urine was carried out. Samples with pus cells ≥ 10/ HPF were quantitatively inoculated on MacConkey's agar (Oxoid, UK) and then incubated at 37 ºC for 24 hrs. Culture plates with bacterial count ≥ 10 5 CFU/ ml and growth of single morphotype of lactose fermenting colonies were selected for identification of E. coli by standard biochemical methods.

Phenotypic detection of UPeC Virulence Factors
The E. coli isolates were tested for expression of haemolytic, cytotoxic activity and haemagglutination as described below:

Assessment of haemolytic activity
The cytolytic toxin secreted by most haemolytic E. coli isolates is known as alpha haemolysin. E. coli isolates were subcultured onto 5% sheep blood agar (Oxoid, UK) and incubated overnight at 37ºC. Haemolysin production was detected by presence of a zone of complete haemolysis around the inoculation site (Raksha et al., 2003;EL-Mosallamy et al., 2015;Tabasi et al., 2015).

Assessment of cytotoxic activity
The cytotoxic effect of E. coli isolates was detected according to Ghadir et al. (2010), however, HEp-2 cells was used instead of Vero cells according to Mills et al. (2000).

Preparation of bacterial isolates
E. coli isolates were obtained from frozen stock and subcultured on blood agar (Oxoid, UK) then on MacConkey's agar (Oxoid, UK). Colonies were then inoculated in test tubes containing 10 ml of tryptic soy broth (TSB) (Oxoid, UK) and incubated for 18-20 h at 37ºC and kept at 4ºC and used for inoculum preparation. Bacteria were harvested by centrifugation at 3000 rpm (MSE Centaur 2, UK). The supernatant was discarded and the bacterial pellets were washed once with phosphate buffered saline (PBS), then resuspended in Eagle's minimum essential medium with Earl's balanced salts (EMEM) (Lonza, Belgium) and adjusted to a count1×10 8 CFU/ ml.

Cell line and growth conditions
HEp-2 cells were maintained in EMEM (Lonza, Belgium) supplemented with 2% fetal bovine serum (FBS) (Lonza, Belgium), 100 U/ml penicillin and 100 µg/ml streptomycin at 37°C in 5 % CO 2 and subcultured every 3-4 days. HEp-2 cells suspended in MEM Earl's and supplemented with 10 % FBS were seeded in 24-well tissue culture plates and kept at 37°C with 5 % CO 2 for 24 hours to form a confluent monolayer. 100 µl of the bacterial cell suspension prepared above was added to each well and incubated at 37ºC in 5 % CO 2 for 3 hours. HEp-2 cell morphology was assessed by examination using an inverted microscope (Olympus bx 50, Japan) 1, 2 and 3 hours post inoculation.

Assessment of haemagglutination
E. coli isolates were inoculated into nutrient broth and incubated at 37 °C for 48 hours for production of fimbriae. Five milliliter of group O venous blood was collected then washed three times in normal saline and a suspension of 3% RBCs in phosphate buffered saline (pH 7.4) was prepared (Lonza, Belgium). They were used immediately or within a week when stored at 3-5 °C. The slide haemagglutination test was carried out on a multiple-concavity slide by adding 50 µl of the RBC suspension to 50 µl of the broth culture and the slide was rocked at room temperature for 5 minutes. Presence of clumping was taken as positive for haemagglutination. At the same time, 50 µl of 3% D-mannose (Fluka AG, Switzerland) was added to the red cells and broth culture in another set and rocked as above. Haemagglutination inhibited in presence of D-mannose was labeled as mannose sensitive haemagglutination (MSHA); indicating type1 fimbriae, if agglutination occurred even in presence of D-mannose, it was called mannose resistant haemagglutination (MRHA); indicating presence of P fimbriae (Uma et al., 2013).

Molecular detection of UPeC Virulence Genes
E. coli isolates were tested for the presence of virulence genes including: hlyA, cnf-1and papC using multiplex PCR.

dnA extraction
DNA extraction was performed for all the clinical isolates using CinnaPure DNA extraction kit (Tehran, Iran) according to manufacturer's instructions.

Amplification
Amplification was done according to Bashir et al. (2012)  The thermal cycler (Biometra T 3000) conditions were as follows: an initial denaturation step at 94°C for 5 min followed by 30 cycles of amplification (denaturation at 94°C for 1 min, annealing at 56°C for 1 min and extension at 72°C for 1 min) and final extension of 5 min at 72°C.
The E. coli colonies were suspended in sterile normal saline and adjusted to 0.5 MacFarland standard, then inoculated on Muller-Hinton agar plates and incubated with the previously mentioned antimicrobial disks at 35ºC ±2 for 18 hours. The interpretation of the diameters of the inhibition zones were done according to CLSI, (2014) ( Table  3). statistical Analysis Data were statistically described in terms of frequencies (number of cases) and percentages and compared using Chi-square (c 2 ) test. Exact test was used instead when the expected frequency is less than 5, P values less than 0.05 was considered statistically significant. All statistical calculations were done using computer program SPSS (Statistical Package for the Social Science;

resUlts
The present study was performed on 50 E. coli isolates isolated from urine samples of patients with UTI. E. coli isolates were identified by their colony morphology on MacConkey's agar and standard biochemical reactions.

Cytotoxic assay
Upon visual assessment of the HEp-2 cell line one hour after infection with 1 x 10 8 CFU/ ml of the tested isolates, no effect was seen on HEp-2 cells. Two hours post infection; it was found that out of 50 UPEC isolates, 46 (92%) isolates showed cytotoxic effect with mild detachment and lysis of HEp-2 cell line and almost complete detachment of the cell line 3 hours post infection, while only 4 (8%) isolates showed no effect on the cell line (Figures 1, 2 and 3).

haemagglutination assay
Out of the 50 E. coli isolates, 9 (18%) isolates showed haemagglutination (HA) of group O RBCs, while 41 (82%) were nonhaemagglutinating. Out of the 9 isolates positive   Twenty five isolates were positive for one or more of the targeted 3 genes, while the remaining isolates were negative for the targeted genes. It was found that one isolate was positive for the three targeted virulence genes, 10 isolates showed the presence of both cnf-1 and hlyA, 2 isolates showed the presence of both cnf-1 and papC, 2 isolates showed the presence of both hlyA and papC, while 8 isolates showed the presence of papC and 2 isolates showed the presence of cnf-1 gene (Table 4).
It was found that the most prevalent gene was cnf-1 which was detected in 15 (30%) isolates, while 13 (26%) isolates were positive for hlyA and 13 (26%) isolates were positive for papC. On the other hand, 10 (24.4%) out of the 41 non-haemagglutinating isolates were positive for papC gene. Although, haemagglutination positive isolates showed a higher percentage of papC gene than the non-haemagglutinating isolates there was no statistically significant difference between them (P-value = 0.679).

results of Fluoroquinolones sensitivity of the E. coli isolates
In vitro susceptibility pattern of E. coli isolates showed that 30 (60%) isolates of E. coli were resistant to all fluoroquinolones, while 20 (40%) isolates were sensitive ( Figure 6).

Association of Virulence Factors expression with resistance to Fluoroquinolones among E. coli isolates h a e m o l y t i c a c t i v i t y i n r e l a t i o n t o fluoroquinolones resistance
Phenotypic analysis showed that out of the 15 isolates with bhaemolysis; 10 (33.3%) were fluoroquinolone resistant and 5 (25%) were fluoroquinolones sensitive isolates with no statistically significant difference between the two groups (P-value = 0.380).   Virulence Factors in relation to Virulence Genes haemolytic activity in relation to hlyA gene It was found that hlyA gene was detected in higher percentage in bhaemolytic isolates (61.5%) than non-haemolytic isolates (38.5%) with statistically significant difference (P-value = 0 .011).

Cytotoxicity in relation to cnf-1 and hlyA
Out of 46 isolates showing cytotoxicity by phenotypic methods, 17 (37%) isolates were positive for the cytotoxic genes; 11 isolates were

Cytotoxicity in relation to fluoroquinolones resistance
Out of the 46 isolates showing cytotoxicity to HEp-2 cells; 27 (90%) isolates were fluoroquinolone resistant and 19 (95%) isolates were fluoroquinolones sensitive with no statistically significant difference between the two groups (P-value = 0.641).

Association of Virulence Genes with resistance to Fluoroquinolones among E. coli isolates
Out of the 30 fluoroquinolones resistant isolates; 16 (53.3%) isolates harboured the virulence genes with variable distribution. It was found that hlyA, cnf-1 and papC genes were detected in the following percentages: 27%, 30% and 30%, respectively. On the other hand, in the susceptible group 8 (40%) isolates harboured the virulence genes with variable distribution. It was found that hlyA, cnf-1 and papC genes were detected in the following percentages respectively 25%, 30% and 20%. There was no statistically significant difference between the resistant and susceptible groups regarding the 3 virulence genes.

disCUssion
UPEC is the most frequent cause of community and hospital-acquired urinary tract infections (Calhau et al., 2015). These infections can vary from a simple cystitis to a serious parenchymatous attack. The degree of severity depends on the virulence of the responsible isolates (Srinivasan et al., 2003;Soto et al., 2011).
Several virulence factors such as fimbriae, toxins, and siderophores contribute to the colonization and pathogenicity of UPEC (Qin et al., 2013). Measuring a phenotype in vitro does not always correlate with in vivo expression and very often underestimates the presence of a virulence factor in vivo (Kaper et al., 2004). The distribution of virulence properties can vary depending upon host characters, type of infection and predisposing factors which determine the host parasite interaction in vivo which can culminate in an active infection (Vijayalakshmi et al., 2015). In addition there is always the possibility of mutation at the level of the corresponding gene, leading to the absence of its detection. Therefore, a positive PCR shows the presence of the virulence gene, but a negative PCR does not point to the absence of the corresponding operon (Tarchouna et al., 2013).
This study was designed to assess some important phenotypic virulence factors of E. coli: haemolytic activity, cytotoxic activity and haemagglutination and their relations to the following genes: hlyA, cnf-1 and papC. Regarding cytotoxic effects of E. coli in this study, it was found that no change was detected in the HEp-2 cell line during the first hour of its incubation with E. coli suspension. However, 2 hours post infection; 92% of the tested E. coli isolates showed morphological changes and almost complete detachment in HEp-2 cells was detected 3 hours post infection. Our results were in accordance with Guignot et al. (2000) and Ghadir et al. (2010).
Agglutination of human erythrocytes by E. coli strain is an indirect evidence of the presence of fimbriae on that strain. MRHA of human RBCs is the phenotypic expression of P fimbriae on E. coli, while MSHA is the phenotypic expression of type 1 fimbriae and both of them mediate adherence and they are important in bladder colonization (Vijayalakshmi et al., 2015).
In the present study, it was found that 18% of the isolates showed haemagglutination. In this study, higher prevalence for MRHA (12%) than MSHA (6%) was observed. Our results were in agreement with Fatima et al. (2012) and Desai et al. (2013). On the other hand, higher prevalence for MSHA than MRHA was recorded by Sai et al. (2013) and Vijayalakshmi et al. (2015).
The genes coding for alpha-haemolysin (hlyA), cytotoxic necrotizing factor (cnf-1) and P fimbrial adhesin (papC) were detected by PCR in the current study. It was observed that 13 isolates (26%) of E. coli isolates were positive for hlyA gene which was in agreement with Bingen-Bidois et al. Lower papC gene carriage rates of 23% and 24% were obtained by Arisoy et al. (2006) and Bashir et al. (2012), respectively. The diversity in frequency of pap gene among different studies can be attributed to the fact that UPEC strains can utilize a variety of adhesins to bind to the urinary epithelial cells, and start the infection. Hence, the strains lacking the pap operon may use other adhesins encoding operons such as afa, and sfa for binding (Neamati et al., 2015).
In addition, the differences in prevalence of UPEC virulence genes showed that the virulence properties of UPEC strains is closely depending on geographic region, weather climate of each regions, diets, the levels of public health, hospital's health and even methods of sampling (Karimian et al., 2012).
In the present study, the simultaneous presence of papC, cnf-1 and hlyA was observed in only one UPEC isolate. Bashir et al. (2012) found that the three genes were found collectively in 4 isolates (out of a total of 59 isolates).The cumulative presence of papC, cnf-1 and hlyA is the evidence of the presence of pathogenicity island II J96 , which is highly prevalent among UTI isolates (Piatti et al., 2008).
Regarding the relation between the phenotypic virulence factors and the virulence genes, it was found that hlyA gene was detected in a higher percentage in b-haemolytic E. coli 53% (8/13) than non-haemolytic E. coli 14% (5/13) and this difference was statistically significant (Pvalue = 0.011). The presence of the gene in some non-haemolytic strains may be explained by the possibility of modification of the gene leading to inactive haemolysin (Landraud et al., 2000). Also, since alpha haemolysin is a pore-forming toxin, at low concentra tion its binding does not always lead to lysis (Boehm et al., 1990;Ostolaza and Goni, 1995). On the other hand, 7 b-haemolytic E. coli isolates in this study were negative for hlyA gene which means that haemolysis on sheep blood agar might be due to haemolysins other than a-haemolysin (Fatima et al., 2012).
In our study, we found that all isolates positive for hlyA gene exhibited cytotoxic activity on HEp-2 cell line; 11 of which harboured cnf-1 gene in addition. Out of the 13 isolates positive for papC gene, only 2 isolates were positive for MRHA. This can be explained by Wullt (2002) who found that some papC positive isolates, especially those isolated form asymptomatic infections, do not express P fimbriae. In addition, the expression of P fimbriae is subjected to phase variation so in vitro growth conditions differing in glucose concentration, temperature and pH may switch the fimbriae on or off. We found that the remaining 4 isolates positive for MRHA were negative for papC gene. According to Sabitha et al. (2014), MRHA can be mediated by P fimbriae and also by X, F1C and Dr fimbriae.
Regarding sensitivity of E. coli to fluoroquinolones, it was observed that 30 isolates In the current study, although there was a higher percentage of both virulence factors and virulence genes in the fluoroquinolones resistant group than in the fluoroquinolones sensitive group, this difference was not statistically significant. This was in accordance with Oliveira et al. (2011). According to previous studies, Johnson et al. found a lower incidence of papC, hlyA, and cnf-1 among fluoroquinolones resistant group than among susceptible group, which seems to be due to a loss of the corresponding PAI, probably as a result of the mutation that causes resistance.

ConClUsion
Virulence determinants may help us to get insights into pathogenesis of UPEC isolates and management of UTIs. Phenotypic methods are simple but not very specific, while multiplex PCR is rapid, easy and more specific in detecting virulence genes. The tested E. coli isolates exhibited a high resistance to fluoroquinolones with no significant difference was found between fluoroquinolones resistant and sensitive E. coli isolates regarding the virulence genes as well as virulence factors.

ACKnoWledGeMents
This study was funded partly by Faculty of Medicine, Cairo University and partly by the authors.