Molecular classification and drug resistance analysis of Escherichia coli isolated from poultry in China

Abstract

We aimed to understand the distribution of Escherichia coli in poultry and to reveal the virulence factors, the drug resistance and molecular epidemic regularity and characteristics of isolate strains from 6 provinces in China and to complete the characteristics of E. coli for the risk assessment. A total of 87 E. coli isolates were analyzed with 7 virulence genes by PCR drug sensitivity test in 13 kinds of antimicrobial agents and analyzed with PFGE and MLST genotyping. The PFGE genotyping of 87 isolates yielded 75 PFGE type. MLST analysis of isolates identified the 39 STs, the 7 housekeeping genes had the different variation. The most prevalent virulence genes were iucD (74.7%), followed by iss (55.2%), Irp2 (43.7%), tsh (28.7%), cva (19.5%), papC (9.2%) and vat (8.1%). All isolates were resistant to two or three antimicrobial agents highly resistant to SXT, TE (85.06%), SF (83.91%), AM (66. 67%), to fluoroquinolones (ENR, 63.22%, NOR, 50.57%) and to GM (57.47%). E. coli strains resistant spectrum was wide gene was polymorphism the distribution had a certain timeliness and regional in part region of China. These were a solid foundation for the epidemiological investigation and traceability laid.

Introduction

Escherichia coli is one of the normal microfloras in humans and animals intestine. Some of them is a well-known pathogen that cause diarrhea. hemorrhagic colitis and hemolytic uremic syndrome [1]. Most cases have been attributed to O157:H7, but the importance of non-O157 STEC is also increasingly recognized [2].

E. coli possesses a number of virulence factors. Some of which related to E. coli pathogenicity. The Yersinia high-pathogenicity island (HPI) carrying Irp2 (encoding the siderophore yersiniabactin) is also present in certain non-O157 STEC lineages which was previously reported only in stx2e carrying human isolates [3]. papC gene encode the adherence factor intimin and icuD is lysine monoxygenase encoding the aerocin [4]. Temperature sensitive hemagglutinin (tsh) and cavitation automatic transport toxins (vat) are automatically transfer protein family. Iss genes (encoding the outer membrane protein) is an important pathogenic factor of E. coli, it can enhance the resistance of Escherichia coli in serum or be associated with bacterial resistance to complement effect [5-7]. Some Iss genes containing ColV plasmid can produce colicin V (cva), which have close relationship with chicken Escherichia coli disease.

Pulsed-field gel electrophoresis (PFGE) is a DNA fingerprinting method that can distinguish different strains within a species by comparing genotypic characteristics. It is an analytical technique in the field of molecular epidemiology [8,9], where this method has been traditionally used for identifying the route and source of infection for specifying bacteria that are responsible for food illness [10]. PFGE has extremely high sensitivity reproducibility and discrimination ability compared with other MLST methods [11].

Multilocus sequence typing (MLST) has become the method of choice for typing epidemiologically important strains [12]. This method is based on determining short nucleotide sequences (450-500 bp) of several (five to seven) housekeeping genes that had undergone some evolutionary diversification leading to polymorphism [13]. MLST is highly discriminatory detecting even few nucleotide substitutions and enables easy automation of polymerase chain reaction (PCR) and sequence determination. Furthermore, it offers better standardization and lab-to-lab portability than its predecessor multilocus enzyme electrophoresis since it relies on sequence data, which are easily accessible in computer databases [14].

This study was designed to understand the distribution of Escherichia coli in large scale poultry population from 6 provinces (Shandong, Shanxi, Neimeng, Anhui, Henan and Chongqing) in China from 2000 to 2012 as well as reveal their virulence factors drug resistance molecular epidemic regularity and characteristics.

Materials and methods

E. coli isolates

87 isolates used in this study were randomly selected from the collection of Microbiology laboratory of Quality & Safety Risk Assessment for Animal Products of Ministry of Agriculture (CAHEC) in Qingdao, China. All these strains were isolated from 32 intensive poultry farms ranging from 10, 000 to 100, 000 poultry per farm in the above 6 provinces from 2000 to 2012: 2000 (13), 2008 (31), 2009 (5), 2010 (19), 2011 (16) and 2012 (3). All isolates were confirmed to be E. coli using API 20E biochemical test strips (bioMérieux, France). Sorbitol fermentation characteristic was examined using sorbitol-MacConkey agar (SMAC) (Oxoid, UK). The determination of O antigens was firstly carried out by testing for specific E. coli O groups of interest targeting group specific genes within the O-antigen gene cluster described by Deb Roy [15]. The entire coding sequence of the fliC gene was amplified by PCR with the primers fliC-F (5’-ATGGCACAAGTCATTAATACCCAAC-3’) and fliC-R (5’-CTAACCCTGCAGCAGAGACA-3’) reported by Qing M [16]. Serotypes of each isolate were determined by agglutination tests with anti-Escherichia coli sera (SSI Denmark).

Identification of virulence genes

For E. coli isolates 7 virulence-associated genes (iucD, iss, vat, Irp2, papC, tsh and cva) were sought as previously described by PCR Primers specific for these virulence-associated genes were shown in Table 1. PCR products were sequenced on ABI Prism 3100 automated sequencer (Applied Biosystems, USA) and were analyzed using NCBI BLAST program (http://www ncbi nlm nih gov/).

Table 1.

PCR primers used for E. coli virulence genes amplification

Targets	Primer	Oligonucleotide sequence (5’-3’)	Amplicon size (bp)	Reference
iucD	iucD-F	ACAAAAAGTTCTATCGCTTCC	714	[17]
	iucD-R	CCTGATCCAGATGATGCTC
irp2	irp2-F	AAGGATTCGCTGTTACCGGAC	280	[18]
	irp2-R	TCGTCGGGCAGCGTTTCTTCT
papC	papC-F	TGATATCACGCAGTCAGTAGC	501	[19]
	papC-R	CCGGCCATATTCACATAA
iss	iss-F	ATCACATAGGATTCTGCCG	309	[17]
	iss-R	CAGCGGAGTATAGATGCCA
tsh	tsh-F	ACTATTCTCTGCAGGAAGTC	824	[17]
	tsh-R	CTTCCGATGTTCTGAACGT
vat	vat-F	TCCTGGGACATAATGGTCAG	1000	[17]
	vat-R	GTGTCAGAACGGAATTGT
cva	cva-F	TGGTAGAATGTGCCAGAGCAAG	1200	[17]
	cva-R	GAGCTGTTTGTAGCGAAGCC

Antimicrobial susceptibility testing

Susceptibility of the isolates to antimicrobial agents was evaluated according to Clinical Laboratory Standards Institute (CLSI) guidelines [20] using the disc diffusion methodon Mueller-Hinton agar (Becton Dickinson, USA). Following 13 antimicrobial agents were tested, including chloramphenicol (florfenicol), penicillins (ampicillin and auge door-keeper), sulfonamides (sulfisoxazole and trimethoprim-sulfamethoxazole), cephems (cefotaxime), aminoglycosides (gentamicin and spectinomycin), tetracyclines (tetracycline and doxycycline), fluoroquinolones (enrofloxacin and ofloxacin), and other (polymyxin). Results were interpreted using the Clinical and Laboratory Standards Institute (CLSI, 2012) breakpoints when available E. coli ATCCR 25922 was used as quality control.

Pulsed-field gel electrophoresis

PFGE of the strains were performed using the non-O157 STEC subtyping protocol (www pulsenetinternational org) with some modifications. The bacteria genomic DNA was digested with 50 U of Xba I (Takara, China) at 37°C for 3 h. XbaI-digested Salmonella enterica serovar Braenderup H9812 was used as the DNA size marker, PFGE was repeated twice to determine reproducibility. For untypeable isolates, 50 μM thiourea (Sigma USA) was added to the 0 5× TBE buffer prior to PFGE run as described by Römling and Tümmler [21]. A contour-clamped homogenous electric field apparatus CHEF-Mapper (Bio-Rad, USA) was used. The pulse time was ramped from 2.16 s to 54.17 s over 19 h at 6.0 V/cm Gel images were captured with a Gel Documentation 2000 software (Bio-Rad, USA) and converted to Tiff files, then analyzed using BioNumerics software (Applied Maths Belgium).

Multi-locus sequence typing

Multi-locus sequence typing (MLST) was performed according to the recommendations of the E. coli MLST website (http://mlst ucc ie/mlst/dbs/Ecoli) using 7 housekeeping genes (adk, fumC, gyrB, icd, mdh, purA and recA). Alleles and sequence types (STs) were determined following the web site instructions [22]. A minimum spanning tree based on these STs was generated with BioNumerics software.

Results

Virulence gene

The results of the distribution of virulence determinants in E. coli isolates in relation were reported in Figure 2. All the 7 virulence factor genes sought were identified in at least 7 isolates. The most prevalent virulence genes were iucD (74.7%), followed by iss (55.2%), Irp2 (43.7%), tsh (28.7%), cva (19.5%), papC (9.2%) and vat (8.1%). The ST23 isolates didn’t exhibit the same virulence profiles. Only five different virulence genes were uniformly present in 7 ST23 isolates, including iucD, Irp2, iss, cva, tsh genes and other 3 ST23 isolates exhibit less one. The virulence profiles correspond inconsistently with PFGE type, but they gathered in a cluster, suggesting similar evolution of virulence genotypes. Meanwhile, the ST131 isolates exhibit the same virulence profiles and PFGE type.

Figure 2.

Dendrogram of PFGE profiles of 87 E. coli isolates from poutry frams PFGE patterns and the corresponding dendrogram for 87 isolates obtained in the present study are depicted. The 6 PFGE clusters were marked on the node as A to F. The different clusters observed are designated on the left side of the figure. Displayed on the right hand side are key (strain name), province, year, PFGE-Pattern, virulence gene, sequence type (ST) and antibiotic resistance. Abbreviations for antibiotics are: AM, Ampicillin, AC, Auge door-keeper, PME, Polymyxin, EFT, Cefotaxime, GM, Gentamicin, SPT, Spectinomycin, TE, Tetracycline, DOX, Doxycycline, SF, Sulfisoxazole, SXT, Trimethoprim-sulfamethoxazole, NOR, Ofloxacin, ENR, Enrofloxacin, FFC, Florfenicol.

Antibiotic resistance

All E. coli isolates were highly resistant to trimethoprim-sulfamethoxazole (SXT), tetracycline (TE) (85.06%), sulfisoxazole (SF) (83.91%), ampicillin (AM) (66.67%), fluoroquinolones [enrofloxacin (ENR), 63.22%, ofloxacin (NOR), 50.57%] and gentamicin (GM) (57 47%) (Figure 1). All isolates were multi-drug resistant as they were resistant to at least 2 groups of antimicrobials. Advantage of resistant performance is NOR-TE-DOX-SF-SXT-AM-GM (Figure 2).

Figure 1.

The dug resistance rate of all isolates for 13 kind of antimicrobial agents. Statistical test was only performed in all isolated. Antibiotics abbreviations are: AM, Ampicillin, AC, Auge door-keeper, PME, Polymyxin, EFT, Cefotaxime, GM, Gentamicin, SPT, Spectinomycin, TE, Tetracycline, DOX, Doxycycline, SF, Sulfisoxazole, SXT, Trimethoprim-sulfamethoxazole, NOR, Ofloxacin, ENR, Enrofloxacin, FFC, Florfenicol. PFGE typing.

All 87 E. coli isolates were analyzed by PFGE using enzymes XbaI resulted in 75 distinguishable patterns demonstrating a high level of genetic diversity among the isolates. An UPGMA dendrogram was constructed (Figure 2). Fifteen E. coli isolates were untypeable by PFGE. After the addition of thiourea to the running buffer, all isolates remained typeable. The 87 isolates could be divided into six clusters A to F at a similarity of 60% or greater. Cluster A contains AH08-8 (ST10) and NM10-11 (ST93). Most of isolates belongs to cluster B, in which including all ST23, ST77, ST2309 and ST2505 isolated, There were a variety of PFGE types for E. coli strains, but these were not very similar.

MLST typing

Thirty-nine discrete STs were identified among the 87 E. coli isolates, indicating a high degree of genotypic diversity. Of these 39 STs, 20 were represented by single isolates, 19 were represented by more than one isolate (n=2 to 10). The predominant STs were ST23 and ST354 containing 10 (256%) and 6 (154%) isolates respectively (Table 2). Isolates characterized as the same ST did not necessarily have the same PFGE pattern. For example, the 10 isolates characterized as ST23 had 8 distinct PFGE patterns (Figure 2). All of the isolates that shared a PFGE pattern had the same ST. Meanwhile, isolates of the same STs generally showed the same or similar drug resistance patterns (Figure 2). ST23, ST117, ST113 and ST2732 isolates showed the same or similar multi-drug resistance to 13 antimicrobial agents respectively.

Table 2.

ST and allele profile of each isolate

ST	adk	fumC	gyrB	icd	mdh	purA	recA	No of isolates
10	10	11	4	8	8	8	2	4
23	6	4	12	1	20	13	7	10
48	6	11	4	8	8	8	2	1
88	6	4	12	1	20	12	7	1
93	6	11	4	10	20	8	6	4
101	43	41	15	18	11	7	6	1
115	4	26	39	25	5	31	19	5
117	20	45	41	43	5	32	2	5
131	53	40	47	13	36	28	29	3
155	6	4	14	16	24	8	14	2
156	6	29	32	16	11	8	44	1
162	9	65	5	1	9	13	6	1
165	10	27	5	10	12	8	2	1
189	10	27	5	10	12	8	49	1
354	85	88	78	29	59	58	62	6
362	62	100	17	31	5	5	4	1
453	99	6	33	33	24	8	7	2
533	6	4	5	18	11	8	14	1
539	6	19	57	18	9	13	6	1
602	6	19	33	26	11	8	6	2
711	9	6	15	131	24	7	7	3
746	10	7	4	8	12	8	2	1
770	52	116	55	101	113	40	38	2
871	64	7	1	8	8	8	6	2
1079	6	19	14	16	11	12	2	2
1101	9	8	5	1	9	8	7	1
1125	6	4	15	18	24	26	7	1
1140	83	23	164	181	80	1	42	1
1158	18	3	17	6	5	5	4	1
1431	6	65	3	1	11	13	6	1
1551	6	250	83	28	1	1	2	1
1724	9	29	12	26	11	8	7	1
2165	6	23	3	16	9	7	7	2
2176	9	65	5	1	9	13	58	1
2309	271	26	39	25	5	31	19	4
2505	6	41	12	1	20	13	7	4
2732	46	26	208	6	5	16	4	3
3285	6	6	15	10	20	23	6	2
3714	6	4	14	402	24	8	14	1

A minimum spanning tree was constructed (Figure 3). Most STs differed from each other by 2 or more alleles while four pairs of STs (ST155 and ST3714, ST162 and ST2176, ST10 and ST48, ST165 and ST189 and ST115 and ST2309) and one set of 3 STs (ST23, ST88 and ST2505) differed from each other by only 1 allele.

Figure 3.

Genetic relationships of E. coli isolates based on MLST. A. The separation of area to build the genetic relationship of all isolates STs. The colors for the slices of the pie represent places of isolates: Shandong province in green, Neimeng province in red, Shanxi province in purple, Anhui province in yellow, Henan province in wathet and Chongqing city in reseda. The numbers on connecting lines show the number of allelic difference between two STs. The number in a circle is the ST number; B. The separation of time to build the genetic relationship of all isolates STs. Each circle represents a given ST with size proportional to the number of isolates. The number in a circle is the ST number. The colors for the slices of the pie represent places of isolates: Red for 2010, green for 2008, purple for 2011, 2000 in yellow, 2009 in bule and 2012 in reseda. The numbers on connecting lines show the number of allelic difference between two STs.

Discussion

This study provided molecular-epidemiological data on E. coli strains isolated in the 32 intensive poultry farms in 6 provinces (Shandong, Shanxi, Neimeng, Anhui, Henan and Chongqing) in China from 2000 to 2012.

We analyzed multiple colonies from 39 samples to determine diversity within a sample (Figure 2). Two samples contained isolates with identical properties, suggesting they were the same strain, while the majority of the samples contained isolates belonging to the same sequence type but differing by one or more of the phenotypic or genetic properties tested, indicating that they were variants of the same clone. Most common variations were non-expression of the H antigen, variation of antibiotic resistance and/or variation in PFGE patterns.

Many studies have underlined the potential key role of the HIP subtypes in the severity of disease. Irp2 gene was the HPI core part, Zhu [23] proved that HPI already existed in poultry E. coli and the frequency was 17.1%. In this study, the prevalence of Irp2 was 43.7%. Fimbrial adhesins play an important role in colonization of the chicken intestine, papC gene encodes the adherence factor intimin, in the study papC gene the frequency is low at 9.2%. Other virulence factors may contribute to the pathogenicity of E. coli, iucD gene was the most prevalent virulent gene (74 7%) among the 7 virulence genes in our text.

Many non-O157 STEC isolated from humans and animals have shown resistance to multiple antimicrobials including resistance to trimethoprim-sulfamethoxazole [24-27]. In our study, we found that only 1 of the 8 categories of antimicrobial resistance types (sulfonamides) and 2 of the 13 antimicrobial agents (sulfisoxazole (SF) and trimethoprim-sulfamethoxazole (SXT)) were active against most the isolates. The high prevalence (>50%) of resistance to tetracycline trimethoprim-sulfamethoxazole is similar to that of other studies in China [25,27]. This suggests that the poultry farms in those countries may have used the similar antimicrobials for prophylactics as the poultry farms in China. PFGE, which is known for its discriminatory power as a molecular typing tool in epidemiologic studies, these isolates which separation in 2000 also shared identical ST and antimicrobial resistance profiles (Figure 2). All the isolates were obtained from different chicken raised in the same region, strongly suggesting that the transmission of the E. coli clone among the animals have occurred. In addition, the strains with the same PFGE type, the ST and resistant genes are the same, the opposite is not established, this results was consistent with Yu [28] studied, thus it can be seen that the same strains exist in the process of proliferation spread of resistant plasmids of gain or loss, lead to the change of the resistance, make have different resistant strains of the same type PFGE spectrum.

In this study, the variation of PFGE genotypes for E. coli is small and the overall similarity value is about 60%~100%, 87 E. coli isolates are divided into 75 PFGE typing, 39 subtypes, in which, 13 E. coli isolates from 2000 are divided into 8 PFGE typing, 3 subtypes, besides SD00-3, most of the strains are gathered in a cluster, although it is believed that these strains may be from the same clone, Dai [29] studied on 16 strains from different farms multi-resistant source of chicken E. coli PFGE classification analysis, 16 different belt type, which fully embodies the polymorphism distribution of E. coli.

All isolates of mapping results show PFGE and correlated with MLST, the strains have the same ST sequences, they are not necessarily the same PFGE typing, but the PFGE type with consistent. ST sequence must be the same, such as ST23 isolates, they come from the 2000 and 2008 different time and region, the PFGE typing are different. SD08-5. SD08-13 and SD08-14 they consistent with PFGE typing, the ST series are ST131. This relationship shows that E. coli between horizontal distribution and vertical transmission of isolates thus combining PFGE and MLST classification method is helpful to find popular advantage of molecular characteristics determine the region characteristics of the strain.

This study foud there was no obvious input or appearance of heterologous strains in the epidemic strains from 2000 to 2012 in China. The relationship between E. coli strains in China and E. coli strains in other regions around the country need further studies. Some of other countries and regions have established their own PulseNet [19]. The results of this study provide PFGE fingerprints of E. coli strains in China and establish a good foundation for the realization of data sharing, which will help to realize active surveillance of E. coli disease and tracing the source of infection in China.

Disclosure of conflict of interest

None.