Virulence-Associated Genes and Genetic Diversity of Avian Pathogenic (APEC) and Fecal (AFEC) E. Coli Isolates From Chickens

Avian pathogenic E. coli (APEC) is the etiologic agent of serious colibacillosis and causes extensive economic losses. To examine the genetic background of APEC, we characterized the serotypes, virulence genes, phylogenetic classication and MLST of 392 APEC and 586 AFEC strains isolated from infected chickens. The results showed that the most predominant serotypes were O78 (13.47%), O2 (9.16%), O18 (5.39%), O20 (4.42%) and O25 (4.09%). The major serotypes O78 (13.47%) and O2 (9.16%) were signicantly higher in the APEC isolates than in the AFEC isolates. Among the 16 analyzed virulence-associated genes (VAGs), iroN (100%), ompT (100%), mC (92.46%), iss (77.91%) and irp2 (71.98%) were the most frequently identied. Over half (54.85%) of the strains possessed > 8–13 VAGs, and 85.23% of the strains carried iroN-ompT-mC-iss/irp2 VAG patterns. According to the phylogenetic analysis, phylogroups A (32.11%) and B2 (31.36%) proved to be the most prevalent phylogenetic groups in the AFEC and APEC isolates, respectively. The strains that belonged to phylogroup B2 were associated with more VAGs. Based on MLST, 46 STs belonging to 15 different clonal complexes were identied, and 4 were novel. ST88 (10.67%) was found to be the most dominant ST, and it possessed at least 9 VAGs and belonged to phylogroups B2 or D. Furthermore, the isolates belonging to B2-O78/O2-ST88 were the most likely APEC isolates to be associated with epidemics, and they carried more VAGs than the other strains.


Conclusions
Our ndings have enriched our knowledge of the molecular characteristics of APEC isolates from chickens, which will be important for the prevention and control of avian colibacillosis.

Background
Avian pathogenic Escherichia coli (APEC) has been reported as an etiologic agent of colibacillosis in poultry worldwide [1][2][3]. APEC isolates can cause severe respiratory and systemic infections that lead to a signi cant economic burden in the poultry industry due to increased mortality, medication costs, and condemnation of carcasses in the prevention and control of the disease [4]. Avian colibacillosis is characterized by yolk sac infection, swollen-head syndrome, septicemia, and in ammation, such as pericarditis, perihepatitis, airsacculitis, salpingitis, arthritis, and peritonitis, of different organs or a combination of these syndromes [5].
APEC infects all ages of commercial poultry and is considered a major or minor pathogen that accounts for 3-4% of the mortality of birds on a farm; however, the mechanisms underlying infection and systemic displacement by APEC are ill de ned. Elucidation of the underlying molecular mechanisms of APEC pathogenicity is essential for controlling avian colibacillosis.
Recently, multiple virulence-associated genes (VAGs) have been identi ed that contribute to APEC pathogenesis [6]. These VAGs are involved in different steps of infection and/or adaptation, and their functions can be classi ed as adhesins, invasins, toxins, iron acquisition systems, and protectins/serum resistance [7][8], thereby protecting the pathogen from the host immune response and enabling its extraintestinal existence [7]. The nature and the combination of the VAGs of APEC could determine the degree of virulence and their potential to cause speci c diseases in speci c hosts. However, the importance and interaction of speci c VAGs that determine the pathogenesis of APEC infections are still poorly understood [9]. Furthermore, some avian fecal E. coli (AFEC) isolated from healthy poultry may also carry certain VAGs re ecting their virulence potential.
Other techniques, including phylogroup and multilocus sequence typing (MLST), are established tools to type APEC. E. coli isolates may be classi ed into four main phylogroups: A, B1, B2, and D. Among them, strains from the B2 and D phylogroups are most frequently found in APEC, while commensal intestinal strains commonly belong to the A and B1 groups [13]. MLST is a top-level genetic tool for differentiating E. coli that allows the assignment of closely related strains in clonal groups or complexes as a sequence type (ST) [14]. These standardized classi cations have facilitated the identi cation and monitoring of pandemic strains that cause nosocomial and community outbreaks [15][16].
Several studies have described APEC strains in the literature; however, little information is available on the characteristics of avian E. coli, especially their serotype level, VAGs and molecular characterization. Therefore, the objective of this study was to determine the genetic background of highly pathogenic E. coli isolates of avian colibacillosis outbreaks and compare this information with that of AFEC strains obtained from healthy poultry.

Isolate collections
In this study, a total of 2947 E. coli strains were isolated, of which 426 strains were collected from tissue and 2531 strains were collected from feces. After PCR detection, 928 E. coli strains were identi ed as ExPEC. The 928 ExPEC strains were further used to study the VAGs and genetic diversity among APEC and AFEC isolates. Among them, 392 APEC isolates were collected from tissues of freshly dead chickens with suspected colibacillosis, and 536 AFEC isolates with VAGs were collected from feces.

Prevalence of virulence-associated genes
The frequencies and combinations of 16 VAGs in 928 E. coli strains (392 APEC and 536 AFEC) were assessed by PCR.
The prevalence of each gene in the APEC and AFEC isolates is shown in Table 2. The iron acquisition system gene iroN and protectin/serum resistance gene ompT were carried by all of the detected strains. More than 70% of the isolates also carried mC (92.46%), iss (77.91%) and irp2 (71.98%). Gene ibeA was not detected in any strain, and papC and neuC were detected less frequently, with 1.94% and 4.74%, respectively. Compared with AFEC, all the VAGs, except iroN, ompT and ibeA, were signi cantly more prevalent in the APEC strains (P < 0.01). Based on the different combinations of VAGs, the E. coli isolates were divided into 27 virulence gene pro le types (VTs) ( Table 3). Each strain possessed at least 3 different VAGs, and over half (54.85%) of the strains possessed > 8- 13 VAGs.

Phylogenetic Classi cation analysis
The phylogenetic analysis showed that phylogroups A and B2 proved to be the most prevalent phylogenetic groups, with 32.11% and 31.36%, respectively, followed by phylotypes D (18.64%) and B1 (17.89%). Within the APEC strains, phylogroup B2 was the largest and contained over half (59.95%) of the strains, while for the AFEC strains, 47.95% belonged to phylotype A and showed the most common groups (Table 4). The strains that belonged to phylogroups B2 and D were associated with more virulence genes, presenting an average of 9.64 and 9.13 virulence genes for each strain, respectively, which was signi cantly higher than that for the B1 (5.68) and A (5.18) groups (Table S1).

MLST-based genotype analysis
MLST was performed to analyze the genotypic diversity of E. coli isolates based on 7 housekeeping genes. According to MLST, 928 strains were divided into 46 STs that belonged to 15 different clonal complexes, and 4 were novel and found in this study (Fig. 3, Table S2). ST88 (10.67%) was found to be the most dominant ST, followed by ST243 (8.94%), ST461 (7.76%), and ST142 (7.22%), while ve STs, ST65, ST118, ST122, ST141, and ST298, were less prevalent and were represented in only one AFEC strain. ST142 was the most common ST in the AFEC strains but was absent in the APEC strains. ST461, ST88 and ST243 were the three most common ST types among the APEC strains, accounting for 11.48%, 10.97% and 9.95%, respectively. Furthermore, all of these strains possessed at least 9 VAGs and belonged to phylogroups B2 or D.
The serotype and STs of the strains exhibited epidemic preferences (Fig. 4). ST88 was the most predominant ST among the O78 and O2 serotypes, and all the APEC strains with ST88 belonged to these two serotypes. Eleven different STs of O78 strains were found, and the top ve STs were ST88 (29.60%), ST243 (20.00%), ST461 (15.20%), ST131 (10.40%), and ST85 (8.80%), which were also the most dominant STs of the APEC strains. Eighty-ve O2 strains were divided into 17 STs, and ST254 (18.82%) and ST855 (12.94%) were the most common STs, except ST88. ST254 was the most common ST in the APEC strains but was missing in the AFEC strains among the O2 serotypes.

Discussion
Colibacillosis is caused by APEC and is considered one of the serious threats to the poultry industry and public health.
APEC infections in birds cause many different kinds of clinical manifestations, ranging from respiratory tract infections to swollen head syndrome, which leads to death [12]. Although previous studies have reported the epidemic characteristics and pathogenic mechanism of APEC strains, detailed data on serotype, VAGs and molecular characteristics are often unavailable in many regions of China. Given that the zoonotic potential of APEC strains is still questionable, a considerable number of strains were isolated from chickens affected with colibacillosis, and several characteristics were compared between APEC and AFEC isolates in the present study.
APEC isolates often have diverse serotypes, and some overlapping serogroups are commonly detected in AFEC isolates.
Although different O serogroups have been associated with colibacillosis, certain speci c serogroups (O78, O2, and O1) are more frequently reported than others [11,17]. Here, O78 was the most predominant serogroup, followed by O2, which is similar to the results from previous studies. O1, however, was the sixth most frequently observed serotype among the Methods Sample collection and bacterial isolation APEC isolates in this investigation, and it was signi cantly higher than that in the AFEC isolates. O18 and O8 were also major serogroups, although the results have been controversial in different studies [12,18].
APEC strains usually carry a large number of VAGs with different functions, and more VAGs in the same strain are often detected among E. coli from lesions [19][20]. Similar results were presented in our ndings, and signi cantly more APEC isolates than AFEC isolates were collected from lesions. O-antigen is well proven to be the virulence factor of E. coli and can protect the bacteria from clearance by the neutrophils and macrophages of the host [21]. It is clear from the results of this study and previous evidence that serogroups O78 and O2 often possess more virulence and are considered virulent. However, different strains from the same serotype may vary in their virulence [18].
APEC strains are characterized by the possession of several VAGs, which enable these bacteria to survive an extraintestinal life and to cause colibacillosis [22][23]. In recent years, the mechanisms behind the pathogenesis and epidemic characteristics of VAGs of APEC strains have been extensively studied [12,18]. Some essential virulence genes, iroN (siderophore), ompT (outer membrane protease), iss (serum survival) and hlyF (hemolysin), carried by plasmids are considered preferential molecular markers for APEC [12,24]. APEC isolates from poultry clinically diagnosed with colibacillosis were positive for at least one of these VAGs, and the frequencies vary greatly in different studies [11].
Sixteen VAGs were detected in the current study, and the frequencies of most VAGs were similar to those in these previous reports [11]. Among those VAGs, the frequencies of iroN and ompT were higher, while the frequencies of iucD, fyuA and vat were lower. Subedi et al. [25] reported an identical result: the virulence genes iroN and ompT were harbored by each APEC isolate, but iss and hlyF were higher than those in the present study. We also found that all the AFEC isolates (carrying VAGs) were positive for both iroN and ompT.
It is speculated that the association of several different VAGs could increase the pathogenicity of bacteria [20]. The functions of virulence genes tested in the present study are well documented, and accumulation of these genes may be a potential risk factor for APEC infection. The epidemic characteristics of VAGs constitute unique VTs and evolution steps. Therefore, monitoring VTs with multiple VAGs in different hosts and understanding the evolution steps may be helpful for reducing economic losses in the poultry industry and the potential zoonotic risks of APEC strains[26].
Phylogenetic classi cation and MLST have several important advantages over PFGE, including shorter assay times, better standardization, and repeatability of data among laboratories. Epidemiological surveys in most previous studies have classi ed APEC strains as predominantly phylogroup B2, followed by phylogroup D, which are mainly responsible for extraintestinal infections and possess more VAGs [27]. The B2 group is closely related to pathogenicity and is frequently found among serogroups O78 and O2 isolates [28]. In the present study, the APEC isolates belonging to B2-O78/O2-ST88 were the most associated with epidemics and carried more VAGs than the other strains. Our studies have suggested that there is a relationship among different E. coli phylogenetic groups, STs, serogroups and the virulence capabilities of the strains.
In conclusion, this study demonstrated the prevalence of common VAGs in APEC and AFEC strains recovered from colibacillosis tissue and fresh tissue. Our studies suggested that different VAGs have accumulated in APEC strains.
However, the presence of these VAGs in AFEC isolates poses a potential risk of causing colibacillosis. Furthermore, the identi cation of predominant serotypes and molecular characteristics, which are closely related to pathogenicity, may be particularly useful in the diagnostic approach. Thus, regular screening and monitoring of APEC strains is essential for implementing intervention programs to reduce the risk of colibacillosis.
Animal-based active surveillance was conducted for 1568 tissue swab samples (liver, heart, lung and spleen), and 2799 fresh samples from infected chickens with typical lesions of an E. coli infection were collected from 26 different farms in Hebei Province from 2018 to 2020. All the samples were cultured on MacConkey agar overnight at 37°C, and pinkcolored suspected bacterial colonies were further isolated on LB agar. For each sample, only one colony was isolated and used for subsequent examination.
The pure and presumptive positive strains were then con rmed as E. coli via biochemical analysis using the API20E system (bioMerieux, Marcy-l'Etoile, France) according to the manufacturer's recommendations. The isolates were stored at − 80°C in 25% glycerol.

Preparation of DNA templates
The DNA templates for PCR (serotype, VAGs, phylogroup, MLST) were directly extracted from bacterial colonies using the boiled lysate method. Brie y, a single colony from an overnight culture at 37°C on LB agar was suspended in 30 µL sterile molecular grade water and boiled at 100°C for 10 min. The sample was immediately cooled on ice for 5 min and centrifuged at 13,000 g at 4°C for 10 min. The supernatant, containing DNA, was transferred to a fresh tube for use [29].

Serological characterization
Serotyping characterization was carried out using a multiplex PCR method for analyzing 162 different O antigens within 20 groups [30]. Multiplex PCR was performed as follows: each 30 µL reaction mixture contained 2 µL DNA template, 0.5 µL of each primer, 10 µL Taq MasterMix (Takara, Japan) and deionized water to a nal volume of 30 µL.

Phylogenetic Classi cation
The E. coli isolates were assigned to four phylogenetic groups (A, B1, B2, or D) based on three genes, ChuA and YjaA and an anonymous DNA fragment, TSPE4.C2, and a rapid and simple method as previously described [34], and the primers are shown in Table S4. The chuA and TspE4.C2-negative and TspE4.C2-positive E. coli strains were classi ed as groups A and B2, respectively, and chuA-negative and TspE4.C2-positive, chuA-positive and yjaA-negative E. coli strains were grouped into B1 and D (Fig S1) ExTaq DNA polymerase (Takara, Japan). The PCR products were bidirectionally sequenced, and the sequences of the 7 housekeeping genes were edited by using SeqMan 7.0. Each unique allele was assigned a different number, and the allelic pro le (string of seven allelic loci) was used to de ne each isolate's sequence type (ST) [36]. Clustering analysis was used to infer relationships among the isolates using the ngerprint analysis software BioNumerics (version 7.1).

Statistical Analysis
Page 12/18 The data for the APEC and AFEC isolates were analyzed using a chi-square test to nd any signi cant differences. These differences were considered statistically signi cant when P < 0.05.

Declarations
Ethics approval and consent to participate Ethical approval was granted for this study. Our study was conducted according to the Ethics Committee of Animal Experiments at the Institute of Husbandry and Pharmaceutical Sciences of Chinese Academy of Agricultural Sciences in Lanzhou, China. We gained consent from the owners of the animals for use in the study.

Consent for publication
All the authors agreed to the publication of the paper.
Availability of data and material The data supporting the ndings of this study are contained within the manuscript. Minimum spanning tree of the 938 E. coli isolates from chickens based on multilocus sequence typing (MLST). The minimum spanning tree was constructed using the 7 identi ed STs obtained from the 938 isolates using BioNumerics

Competing interests
Software. Each circle corresponds to a single ST. The shadow zones in different colors correspond to different E.coli (APEC or AFEC).

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.