Diversity of Virulence Genes in Multidrug Resistant Escherichia coli from a Hospital in Western China.

BACKGROUND: Escherichia coli strains are the most commonly isolated bacteria in hospitals. The normally harmless commensal E. coli can become a highly adapted pathogen, capable of causing various diseases both in healthy and immunocompromised individuals, by acquiring a combination of mobile genetic elements. Our aim was to characterize E. coli strains from a hospital in western China to determine their virulence and antimicrobial resistance potential.
METHODS: A total of 97 E. coli clinical isolates were collected from the First Affiliated Hospital of Chengdu Medical College from 2015 to 2016. Microbiological methods, PCR, and antimicrobial susceptibility tests were used in this study.
RESULTS: The frequency of occurrence of the virulence genes fimC, irp2, fimH, fyuA, lpfA, hlyA, sat, and cnf1 in the E. coli isolates was 93.81, 92.78, 91.75, 84.54, 41.24, 32.99, 28.86, and 7.22%, respectively. Ninety-five (97.9%) isolates carried two or more different virulence genes. Of these, 44 (45.4%) isolates simultaneously harbored five virulence genes, 24 (24.7%) isolates harbored four virulence genes, and 17 (17.5%) isolates harbored six virulence genes. In addition, all E. coli isolates were multidrug resistant and had a high degree of antimicrobial resistance.
CONCLUSION: These results indicate a high frequency of occurrence and heterogeneity of virulence gene profiles among clinical multidrug resistant E. coli isolates. Therefore, appropriate surveillance and control measures are essential to prevent the further spread of these isolates in hospitals.

Introduction

Most Escherichia coli strains that colonize the human intestines rarely cause illness in healthy individuals. However, a number of pathogenic strains can cause intestinal or other diseases in healthy, as well as immunocompromised individuals.1 Commensal E. coli strains can evolve into highly adapted pathogens capable of inducing diseases following the acquisition of a combination of mobile genetic elements, including virulence genes.1–3

The occurrence of multidrug resistant (MDR) E. coli strains has increased in recent years, leading to a severe problem in healthcare settings, especially in developing countries.4–6 MDR E. coli strains complicate treatment, as they require prolonged hospitalization and antibiotic treatment and increase the need of surgery, which eventually increase mortality.7,8

E. coli strains have been well documented in healthcare settings in western China; however, their characterization has often been limited to phenotypic tests and the identification of resistance genes,9–13 with limited information regarding their virulence factors. Previously, we examined the virulence gene profiles of 13 diarrheagenic E. coli (DEC) strains isolated from a hospital in western China, as well as the molecular characteristics of their genes.14 Here, we characterized E. coli strains from a hospital in western China and determined their virulence and antimicrobial resistance potential, to better understand the prevalence of virulence genes and antimicrobial resistance in clinical E.coli isolates. This study emphasizes the importance of preventing the spread of E.coli isolates that harbor both antimicrobial resistance and virulence genes in the clinical setting.

Methods

Bacterial Isolates

A total of 97 non-duplicated clinical E. coli isolates were collected from 97 different patients in various departments (gastroenterology, urology, endocrinology, neurosurgery, and other wards) of the First Affiliated Hospital of Chengdu Medical College, Chengdu, Sichuan, China from 2015 to 2016. The isolates were identified using standard laboratory methods and the ATB New system (bioMérieux, Lyons, France). Patients who satisfied the following three criteria were included in the analysis: 1) age >18 years; 2) suspected of having an infection, based on their clinical symptoms (e.g. fever, abdominal pain, nausea, vomiting, dehydration and tenesmus); and 3) their bacterial culture yielded E. coli. The E. coli isolates were collected from biofluid samples including blood, urine, sputum, wound exudates and abscesses. Each isolate was further verified by PCR amplification of a 369-bp internal control region from the E. coli marker gene, alr.15 All bacterial strains were stored at −80 °C and were grown on MacConkey Agar (Oxoid, Hampshire, UK).

The study protocol was approved by the Ethics Committee of Chengdu Medical College, in accordance with the Helsinki declaration. In all cases, the patients or their family members were informed and their written consents was obtained.

Detection of Adherence and Virulence Genes

All E. coli isolates were subjected to PCR to detect 12 adherence (bfp, daaD, daaE, fimC, fimH, aggA, aafA, agg3A, agg4A, lpfA, sfa, and pap) and 27 virulence (aggR, pic, astA, stx1, stx2, eae, ipaH, est, elt, irp2, fyuA, escJ, escN, escV, espP, nleB, nleE, ent/espL2, cnf1, cnf2, cdt-I, cdt-II, invE, hlyA, pet, sat, and subAB) genes. The primers used to amplify these genes are listed in Table 1.

Table 1

Gene Primers Used in This Study

Gene	Primer Sequence (5´-3´)	PCR Product (bp)	Reference
alr	F: CTGGAAGAGGCTAGCCTGGACGAGR: AAAATCGCCACCGGTGGAGCGATC	369	15
bfp	F: GACACCTCATTGCTGAAGTCGR: CCAGAACACCTCCGTTATGC	324	55
daaD	F: TGAACGGGAGTATAAGGAAGATGR: GTCCGCCATCACATCAAAA	444	56
daaE	F: GAACGTTGGTTAATGTGGGGTAAR: TATTCACCGGTCGGTTATCAGT	542	57
fimC	F: GGGTAGAAAATGCCGATGGTGR: CGTCATTTTGGGGGTAAGTG	477	58
fimH	F: CGAGTTATTACCCTGTTTGCTGR: ACGCCAATAATCGATTGCAC	878	59
aggA	F: GCTAACGCTGCGTTAGAAAGACCR: GGAGTATCATTCTATATTCGCC	421	59
aafA	F: ATGTATTTTTAGAGGTTGACR: TATTATATTGTCACAAGCTC	518	60
agg3A	F: GTATCATTGCGAGTCTGGTATTCAGR: GGGCTGTTATAGAGTAACTTCCAG	462	59
agg4A	F: TGAGTTGTGGGGCTAYCTGGACACCR: ATAAGCCGCCAAATAAGC	169	41
lpfA	F: AGGCGGTGCATTCACTCTGGCATCTR: CCGCGTCGATAGCGGTATAGGCAGA	446	61
sfa	F: CTCCGGAGAACTGGGTGCATCTTACR: CGGAGGAGTAATTACAAACCTGGCA	408	59
pap	F: GACGGCTGTACTGCAGGGTGTGGCGR: ATATCCTTTCTGCAGGGATGCAATA	328	59
aggR	F: ACGCAGAGTTGCCTGATAAAGR: AATACAGAATCGTCAGCATCAGC	400	55
pic	F: GGGTATTGTCCGTTCCGATR: ACAACGATACCGTCTCCCG	1176	62
astA	F: CCATCAACACAGTATATCCGAR: GGTCGCGAGTGACGGCTTTGT	111	59
stx1	F: CGATGTTACGGTTTGTTACTGTGACAGCR: AATGCCACGCTTCCCAGAATTG	244	55
stx2	F: GTTTTGACCATCTTCGTCTGATTATTGAGR: AGCGTAAGGCTTCTGCTGTGAC	324	55
eae	F: TGAGCGGCTGGCATGAGTCATACR: TCGATCCCCATCGTCACCAGAGG	241	63
ipaH	F: GTTCCTTGACCGCCTTTCCGATACCGTCR: AAAATCGCCACCGGTGGAGCGATC	619	64
est	F: ATTTTTCTTTCTGTATTGTCTTR: CACCCGGTACAGGCAGGATT	190	65
elt	F: GGCGACAGATTATACCGTGCR: CGGTCTCTATATTCCCTGTT	450	65
irp2	F: AAGGATTCGCTGTTACCGGACR: TCGTCGGGCAGCGTTTCTTCT	264	66
fyuA	F: TGATTAACCCCGCGACGGGAAR: CGCAGTAGGCACGATGTTGTA	785	27
escJ	F: CACTAAGCTCGATATATAGAACCCR: GTCAATGTTGATGTCGTATCTAAG	824	40
escN	F: CGCCTTTTACAAGATAGAACR: CATCAAGAATAGAGCGGAC	854	67
escV	F: GATGACATCATGAATAAACTCR: GCCTTCATATCTGGTAGAC	2128	40
espP	F: AAACAGCAGGCACTTGAACGR: GGAGTCGTCAGTCAGTAGAT	1830	62
nleB	F: GGAAGTTTGTTTACAGAGACGR: AAAATGCCGCTTGATACC	297	68
nleE	F: GTATAACCAGAGGAGTAGCR: GATCTTACAACAAATGTCC	260	68
ent/espL2	F: GAATAACAATCACTCCTCACCR: TTACAGTGCCCGATTACG	233	68
cnf1	F: GGCGACAAATGCAGTATTGCTTGGF: GACGTTGGTTGCGGTAATTTTGGG	552	62
cnf2	F: GTGAGGCTCAACGAGATTATGCACTGR: CCACGCTTCTTCTTCAGTTGTTCCTC	839	62
cdt-Ⅰ	F: CAATAGTCGCCCACAGGAR: ATAATCAAGAACACCACCAC	412	69
cdt-Ⅱ	F: GAAAGTAAATGGAATATAAATGTCCGR: TTTGTGTTGCCGCCGCTGGTGAAA	556	69
invE	F:CGATCAAGAATCCCTAACAGAAGAATCACR: CGATAGATGGCGAGAAATTATATCCCG	766	55
hlyA	F: GCATCATCAAGCGTACGTTCCR: AATGAGCCAAGCTGGTTAAGCT	533	66
pet	F: TTTCCAGCACTTCCTGTTCCR: ATTTCCAACGTCTACGCCAT	297	70
sat	F: GCAGCAAATATTGATATATCAR: GTTGTTGACCTCAGCAAGGAA	2913	40
subAB	F: TATGGCTTCCCTCATTGCCR: TATAGCTGTTGCTTCTGACG	556	71

Antimicrobial Susceptibility Testing

The minimum inhibitory concentrations (MICs) of 24 antimicrobial agents against the E. coli isolates were determined by agar dilution methods, according to the 2019 Clinical and Laboratory Standards Institute guidelines.16 The following 24 antimicrobial agents were tested: sulfonamide, doxycycline, tetracycline, cefotaxime, ampicillin, ticarcillin, nalidixic acid, cefoperazone, piperacillin, gentamicin, ciprofloxacin, levofloxacin, ofloxacin, tobramycin, cefoxitin, ceftazidime, minocycline, aztreonam, kanamycin, amikacin, chloramphenicol, meropenem, imipenem, and ertapenem. The results were used to classify the isolates as resistant or susceptible to a particular antibiotic using standard reference values.16

Results

Detection of E. coli Adherence and Virulence Genes

The presence of 12 adherence genes and 27 toxin-encoding genes was examined in all E. coli strains by PCR. As shown in Figure 1, the detection rate of fimC, irp2, fimH, fyuA, lpfA, hlyA, sat, and cnf1 in the isolated E. coli strains was 93.81, 92.78, 91.75, 84.54, 41.24, 32.99, 28.86, and 7.22%, respectively. All isolates were negative for the other genes tested (bfp, daaD, daaE, aggA, aafA, agg3A, agg4A, sfa, pap, aggR, pic, astA, stx1, stx2, eae, ipaH, est, elt, escJ, escN, escV, espP, nleB, nleE, ent/espL2, cnf2, cdt-I, cdt-II, invE, pet, and subAB).

Figure 1

Frequency of virulence genes among E.coli isolates.

Different combinations of multiple virulence genes were detected in the E. coli isolates. The number of virulence genes in each isolate and the specific virulence gene combinations are shown in Table 2. Two or more different virulence genes were identified in ninety-five (97.94%) isolates. Of these, 44 (45.37%) isolates simultaneously harbored five virulence genes, 24 (24.74%) isolates harbored four virulence genes, 17 (17.53%) isolates harbored six virulence genes, five (5.15%) isolates harbored three virulence genes, two (2.06%) isolates harbored two virulence genes, two (2.06%) isolates harbored seven virulence genes, and only one (1.03%) isolate harbored eight virulence genes.

Table 2

Distribution of Virulence Genes Among E. coli Isolates

No. of Virulence Genes	Virulence Genes Profile	No. (%) of Bacterial Strain	Total No. (%)
0 genes		2(2.06)	2(2.06)
2 genes	fimC, fimH	1(1.03)	2(2.06)
irp2, fyuA	1(1.03)
3 genes	fimC, fimH, lpfA,	2(2.06)	5(5.15)
	fimC, lpfA, sat	1(1.03)
	fimC, irp2, fyuA	1(1.03)
irp2, fyuA, lpfA	1(1.03)
4 genes	fimC, fimH,irp2, fyuA	14(14.44)	24(24.74)
	fimC, fimH, irp2, lpfA	4(4.12)
	fimC,irp2, fyuA, lpfA,	2(2.06)
	fimH, irp2, fyuA, hlyA	2(2.06)
	fimC, fimH, lpfA, hlyA	1(1.03)
fimC, fimH, hlyA, irp2	1(1.03)
5 genes	fimC, fimH, irp2, fyuA, lpfA	19(19.59)	44 (45.37)
	fimC, fimH, irp2, fyuA, sat	12(12.38)
	fimC, fimH, irp2, fyuA, hlyA	10(10.31)
	fimC, fimH, irp2, lpfA, hlyA	2(2.06)
fimC, fimH, irp2, fyuA,cnf1	1(1.03)
6 genes	fimC, fimH, irp2, fyuA, hlyA, sat	7(7.23)	17(17.53)
	fimC, fimH,irp2,fyuA,lpfA, sat	4(4.12)
	fimC, fimH, irp2, fyuA,hlyA,cnf1	3(3.09)
	fimC, fimH, irp2, fyuA, lpfA,hlyA	2(2.06)
fimC,fimH,irp2,lpfA,hlyA,sat	1(1.03)
7 genes	fimC,fimH,irp2,fyuA,hlyA,sat,cnf1	2(2.06)	2(2.06)
8 genes	fimC, fimH,irp2, fyuA, lpfA,hlyA sat, cnf1	1(1.03)	1(1.03)

Resistance to Antimicrobial Agents

The 24 most commonly used antimicrobials in Chinese practice clinical were used in this study to test the antibiotic resistance of the 97 E. coli isolates,14,18–20 including penicillin (ampicillin, ticarcillin, piperacillin), cephems (cefoxitin, cefoperazone, cefotaxime, ceftazidime), monobactams (aztreonam), carbapenems (meropenem, imipenem, ertapenem), aminoglycosides (tobramycin, kanamycin, gentamicin, amikacin, chloramphenicol), tetracyclines (deoxycycline, minocycline, tetracycline), quinolones (levofloxacin, ofloxacin, nalidixic acid, ciprofloxacin).16 The resistance profiles of the E. coli isolates against these 24 antibiotics are detailed in Table 3. The isolates exhibited a high degree of resistance, especially against sulfonamide (97.94%), ampicillin (94.85%), ticarcillin (90.72%), nalidixic acid (90.72%), tetracycline (81.44%), doxycycline (78.49%), ciprofloxacin (70.10%), ofloxacin (68.04%), cefotaxime (68.04%), and levofloxacin (60.82%). Furthermore, all E. coli isolates were susceptible to meropenem and imipenem. The sensitivity rate of the E. coli strains to ertapenem, amikacin, cefoxitin, ceftazidime, aztreonam, and chloramphenicol was 92.79, 88.66, 74.22, 67.01, 67.01, and 64.95%, respectively.

Table 3

Antimicrobial Susceptibility of E. coli Clinical Isolates

Antimicrobial Agent	Resistant n (%)	Intermediate n (%)	Susceptible n (%)
Sulfonamide	95 (97.94)	—	2 (2.06)
Ampicillin	92 (94.85)	0(0)	5 (5.15)
Ticarcillin	88 (90.72)	2 (2.06)	7 (7.23)
Nalidixic acid	88 (90.72)	—	9 (9.28)
Tetracycline	79 (81.44)	1 (1.03)	17 (17.53)
Doxycycline	73 (78.49)	4 (4.12)	20 (20.62)
Ciprofloxacin	68 (70.10)	2 (2.06)	27 (27.84)
Ofloxacin	66 (68.04)	2 (2.06)	29 (29.90)
Cefotaxime	66 (68.04)	4 (4.12)	37 (38.14)
Levofloxacin	59 (60.82)	10 (10.31)	28 (28.87)
Piperacillin	58 (59.79)	19 (19.59)	20 (20.62)
Cefoperazone	51 (52.58)	18 (18.55)	28 (28.87)
Gentamicin	51 (52.58)	4 (4.12)	42 (43.30)
Kanamycin	39 (40.21)	1 (1.03)	57 (58.76)
Tobramycin	39 (40.21)	17 (17.53)	41 (42.26)
Chloramphenicol	33 (34.02)	1 (1.03)	63 (64.95)
Minocycline	33 (34.02)	13 (13.40)	51 (52.58)
Aztreonam	28 (28.87)	4 (4.12)	65 (67.01)
Ceftazidime	21 (21.65)	11 (11.34)	65 (67.01)
Cefoxitin	17 (17.53)	8 (8.25)	72 (74.22)
Amikacin	8 (8.25)	3 (3.09)	86 (88.66)
Ertapenem	3 (3.09)	4 (4.12)	90 (92.79)
Meropenem	0(0)	0(0)	97 (100)
Imipenem	0(0)	0(0)	97 (100)

Importantly, all isolates were resistant to at least three different classes of antimicrobial agents and were considered as multidrug resistant.17 Of the 97 MDR E. coli isolates, five (5.16%), one (1.03%), one (1.03%), three (3.09%), three (3.09%), six (6.19%), nine (9.28%), six (6.19%), nine (9.28%), twelve (12.37%), nine (9.28%), eight (8.25%), eight (8.25%), four (4.12%), three (3.09%), two (2.06%), three (3.09%), two (2.06%), and three (3.09%) isolates exhibited resistance to 3–21 types of antibiotics, respectively, as shown in Table 4 and Figure 2.

Table 4

Number of E. coli Isolates Resistant to Different Classes of Antibiotics

Different classes of antibiotics		3	4	5	6	7	8	9	10	11	12	13	14	15	16	17	18	19	20	21
Isolates	n	5	1	1	3	3	6	9	6	9	12	9	8	8	4	3	2	3	2	3
%	5.16	1.03	1.03	3.09	3.09	6.19	9.28	6.19	9.28	12.37	9.28	8.25	8.25	4.12	3.09	2.06	3.09	2.06	3.09

Figure 2

Number of E. coli isolates resistant to different classes of antimicrobial agents.

Frequency of Virulence Gene Occurrence in Isolated E. coli Strains Exhibiting Antimicrobial Resistance

The frequencies of virulence gene occurrence in isolated E. coli strains exhibiting antimicrobial resistance are detailed in Table 5. The frequencies for fimC, irp2, and fimH among the resistant E. coli isolates were nearly > 90%, whereas that of fyuA was > 80%. Moreover, the frequencies of lpfA, hlyA, sat, and cnf1 in the resistant isolates were higher than 40, 30, 20, and 5%, respectively.

Table 5

Frequency of Virulence Genes Among Antibiotic Resistant E. coli Isolates

Antibiotic (n)	Virulence Genes, n (%)
fimC	irp2	fimH	fyuA	lpfA	hlyA	sat	cnf1
Sulfonamide (95)	89 (93.68)	88 (92.63)	88 (92.63)	80 (84.21)	39 (41.05)	32 (33.68)	27 (28.42)	7 (7.37)
Ampicillin (92)	87 (94.56)	85 (92.39)	85 (92.39)	77 (83.69)	38 (41.30)	32 (34.78)	25 (27.17)	7 (7.61)
Ticarcillin (88)	84 (95.45)	82 (93.18)	83 (94.32)	74 (84.09)	37 (42.05)	32 (36.36)	25 (28.41)	7 (7.95)
Nalidixic acid (88)	84 (95.45)	82 (93.18)	84 (95.45)	74 (84.09)	38 (43.18)	32 (36.36)	27 (30.68)	7 (7.95)
Tetracycline (79)	75 (94.94)	75 (94.94)	74 (93.67)	68 (86.07)	33 (41.77)	25 (31.65)	22 (27.85)	5 (6.32)
Deoxycycline (73)	68 (93.15)	68 (93.15)	67 (91.78)	63 (86.30)	30 (41.10)	23 (31.51)	20 (27.40)	5 (6.85)
Ciprofloxacin (68)	67 (98.53)	65 (95.59)	67 (98.53)	59 (86.76)	28 (38.36)	23 (33.82)	22 (32.35)	6 (8.82)
Ofloxacin (66)	65 (98.48)	63 (95.45)	65 (98.48)	57 (86.36)	26 (39.39)	24 (36.36)	21 (31.82)	6 (9.09)
Cefotaxime (66)	66 (100)	63 (95.45)	63 (95.45)	58 (87.88)	32 (48.48)	21 (31.82)	17 (25.76)	7 (10.61)
Levofloxacin (59)	58 (98.31)	56 (94.92)	58 (98.31)	50 (84.75)	24 (40.68)	21 (35.59)	15 (25.42)	6 (10.17)
Piperacillin (58)	56 (96.55)	55 (94.82)	55 (94.82)	53 (91.37)	26 (44.82)	17 (29.31)	16 (27.50)	5 (8.62)
Cefoperazone (51)	51 (100)	50 (98.04)	49 (96.07)	45 (88.24)	25 (49.02)	14 (27.45)	12 (23.53)	3 (5.88)
Gentamicin (51)	49 (96.08)	47 (92.16)	48 (94.12)	44 (86.27)	21 (41.18)	16 (31.37)	12 (23.53)	3 (5.88)
Kanamycin (39)	37 (94.87)	36 (92.31)	35 (89.74)	32 (82.05)	19 (48.72)	9 (23.08)	8 (20.51)	3 (7.69)
Tobramycin (39)	38 (97.44)	36 (92.31)	35 (89.74)	36 (92.31)	19 (48.72)	5 (12.82)	8 (20.50)	3 (7.69)
Chloramphenicol (33)	32 (96.97)	30 (90.91)	29 (87.88)	27 (81.82)	13 (39.9)	8 (24.24)	10 (30.30)	2 (6.06)
Minocycline (33)	31 (93.94)	32 (96.97)	30 (90.91)	31 (93.94)	17 (51.52)	8 (24.24)	11 (33.33)	2 (6.06)
Aztreonam (28)	28 (100)	28 (100)	27 (96.43)	26 (92.86)	12 (42.86)	7 (25.00)	7 (25.00)	2 (7.14)
Ceftazidime (21)	21 (100)	21 (100)	21 (100)	19 (90.48)	9 (42.86)	6 (28.57)	7 (33.33)	1 (4.76)
Cefoxitin (17)	15 (88.23)	15 (88.23)	12 (70.59)	13 (76.47)	9 (52.94)	6 (35.29)	2 (11.76)	2 (11.76)
Amikacin (8)	7 (87.50)	6 (75.00)	7 (87.50)	4 (50.00)	5 (62.50)	1 (12.50)	1 (12.50)	0 (0)
Ertapenem (3)	3 (100)	3 (100)	3 (100)	2 (66.67)	1 (33.33)	1 (33.33)	2 (66.67)	1 (33.33)

Discussion

E. coli strains are the most commonly isolated bacteria in hospitals.18–20 Although these strains have been frequently reported in hospitals in western China, data regarding the virulence genes present in these strains are limited.9–13 Thus, in this study, we investigated the presence of virulence genes and antimicrobial resistance in E. coli strains at a hospital in the western region of China in order to further expand our knowledge of the characteristics of E. coli strains prevalent in China.

We first detected 12 adherence and 27 virulence genes in 97 clinical E. coli isolates. Our results showed that most of the E. coli isolates contained multiple and heterogeneous virulence genes (Table 2). Type 1 fimbriae is an E. coli adhesion factor encoded by the fimC and fimH genes. It enables E. coli to bind to intestinal epithelial cells by attaching on mannose-containing receptors. In our study, fimC and fimH were identified in 93.81 and 91.75% of the strains, respectively. Nuesch-Inderbinen et al21 detected the presence of fimC and fimH in all human E. coli strains isolated in Switzerland, while Malekzadegan and Khashei22 found fimH in all isolates from Iranian patients. These reports are in agreement with our findings; the high frequency of occurrence of fimC and fimH among E. coli strains points to their importance in E. coli adhesion.

Some E. coli strains, contain another type of fimbria, long polar fimbriae (LPF), encoded by the conserved gene lpfA.23,24 We found that 41.24% of the E. coli isolates carried lpfA, which is similar to the frequency (50%) reported in Mexico.25 Initial studies conducted on human biopsy samples have suggested that adherence and the attaching and effacing lesion caused by E. coli do not require LPF.24 Therefore, it is possible that LPF are not necessary for E. coli pathogenicity.

The High-Pathogenicity Island (HPI) marker genes, irp2 and fyuA, were detected in 92.78 and 84.54%, respectively, of E. coli isolates in this study. The irp2 and fyuA genes have been detected in a number of studies examining pathogenic E. coli isolated from humans,26–28 similar to the results of the present study. The iron-uptake system of highly pathogenic strains is mediated via yersiniabactin, which is encoded by irp2 and fyuA and is associated with strain virulence.29,30 A considerable number of bacteria isolated from food harbor irp2 and fyuA (involved in iron capture systems).31,32 This could be the reason for the frequent detection of irp2 and fyuA in pathogenic E. coli isolated from humans.

The hlyA gene was detected in 32.99% of the E. coli isolates. In Iran, Malekzadegan and Khashei22 reported that 28.6% of the E. coli strains were positive for hlyA, whereas Dale et al33 found that 26% of E. coli strains in the UK carried hlyA, and Bozcal et al34 identified this gene in 15.4% of E. coli strains in Turkey. The percentage of E. coli harboring hlyA in our study was higher than detected in the above-mentioned studies. α-hemolysin (HlyA) belongs to a group of pore-forming leukotoxins containing RTX repeats, and is thus consider a virulence factor in E. coli.35–37 Depending on its concentration and the type of cell affected, HlyA either displays cytolytic activity or hijacks innate immune signaling pathways.37–39 The high percent of hlyA in this study suggests that HlyA is involved in the mechanisms underlying E. coli pathogenicity in 32 (32.99%) patients.

The sat gene was detected in 28.86% of E. coli isolates. Sat is frequently detected in pathogenic E. coli strains.5,40,41 As demonstrated by Guignot et al,42 Sat can cause tight junction lesions between epithelial cells, which may lead to an increase in their permeability. These findings indicate that Sat probably plays a role in E. coli pathogenesis in 28 (28.86%) of the patients.

The cnf1 gene was found in seven (7.22%) E. coli isolates, similar to the 7.2% reported in Turkey.34 Moreover, Bouzari et al43 reported that 29.4% of E. coli strains harbor cnf1 genes. Cytotoxic necrotizing factor type 1 (CNF1) is a monomeric protein previously shown to effect rabbit skin cell necrosis and multinucleation of various cultured eukaryotic cells.44–46 Our results are in agreement with the low occurrence of cnf1 in E. coli strains.

We next examined the antimicrobial resistance of the 97 E. coli strains. The E. coli isolates were insensitive to first-line antibiotics such as nalidixic acid, sulfonamide, ticarcillin, ampicillin, tetracycline, doxycycline, ofloxacin, cefotaxime, ciprofloxacin, and levofloxacin (Table 3). The antibiotic resistance rates of the E. coli isolates exceeded those reported in developing countries such as Brazil, Turkey, and Ghana.5,34,47 Moreover, the resistance rates observed in our study were higher than noted in the CHINET project.18–20 Unexpectedly, we found that all E. coli isolates were MDR and over half of them were resistant to > 12 classes of antibiotics (Table 4 and Figure 2). These results highlight the increasing severity of antibiotic misuse in clinical practice in western China.

Carbapenem-resistant Enterobacteriaceae (CRE) are highly prevalent in China, the United States, Italy, Israel, Colombia, Greece, the Indian subcontinent, North Africa, and Turkey.48,49 China (especially the regions of Beijing, Changsha, Chongqing, Fuzhou, Guangzhou, Hangzhou, Hebei, Hong Kong, and Zhengzhou) is thought to be one of main endemic regions of these bacteria around the world.50,51 In our study, we found that three (3.09%) CRE among the 97 E. coli isolates were resistant to ertapenem (Table 3). Carbapenem-resistant E. coli have been frequently reported in western China in recent years;52–54 most probably owing to the use of carbapenems as antimicrobial agents in this region.

Lastly, but most importantly, we found that the E. coli strains harbor a high rate of virulence genes in addition to high antimicrobial resistance (Table 5). These findings explain how the E. coli isolates are able to successfully invade the human body and evade antibiotic treatment. Our findings indicate that clinical MDR E. coli isolates harbor a high frequency of virulence genes and that their virulence gene profiles are highly heterogeneous. Therefore, surveillance and control measures need to be enhanced to prevent these isolates from spreading further in hospitals.

Conclusions

This study demonstrates a high frequency of occurrence and heterogeneity of virulence gene profiles among clinical multidrug resistant E. coli isolates. We conclude that appropriate surveillance and control measures are essential to prevent the further spread of these isolates in hospitals. However, further investigations are needed including additional hospitals in western China and a greater number of E.coli isolates to better understand the prevalence of virulence genes and antimicrobial resistance of the E.coli in western China.