Detection of phenotypes , virulence genes and phylotypes of avian pathogenic and human diarrheagenic Escherichia coli in Egypt

Introduction: The purpose from this study was to determine phenotypes, intestinal virulence-associated genes, and phylotypic profiling of human diarrheagenic E. coli (DEC) and avian pathogenic E. coli (APEC). Methodology: A total of 108 chicken visceral organs (liver, spleen, heart) from 36 diseased birds (three organs per each bird) and 78 human stool samples (50 diarrheic patients and 28 healthy persons) were randomly collected during the first half of 2015 in the district of Mansoura city, Egypt. Conventional culturing, serotyping, and molecular characterization of virulence genes and phylogroups were performed. Results: Sixty-five (35%) biochemically identified E. coli isolates were detected from chicken visceral (29/108; 26.9%) and human stool samples (36/78; 46.2%). Serotypes O78, O2, and O1 were the most prevalent serotypes (62%) distinguished from APEC isolates, and only two similar serotypes (O119:H4 and O26:H11) were identified from both APEC and DEC isolates. By polymerase chain reaction (PCR), the respective percentages of 100 and 35 with eae and Shiga toxin genes were detected from APEC isolates while 50%, 27.8%, and 19.4% of human DEC isolates harbored eae, stx1, and stx2 genes, respectively. Phylogrouping revealed a significantly higher occurrence of pathogenic phylogroups (D and B2) in APEC (19/29; 65.5%) than in human DEC isolates (8/36; 22.2%). Conclusions: APEC isolates shared serotypes, virulence genes, and phylotypes with human DEC isolates, which is a subsequent potential public health concern. To the best of our knowledge, this is the first report in Egypt that determines virulence gene and phylogroup coexistence between APEC and DEC isolates.


Introduction
Escherichia coli (E.coli) is a naturally occurring organism that forms part of the animal and human gut microbiota; however, some strains have the ability to produce pathogenicity in both [1].Pathogenic strains have been divided into intestinal (diarrheagenic; DEC) and extra-intestinal (ExPEC) pathogenic E. coli.Avian pathogenic E. coli (APEC), a subdivision of ExPEC that produces a systemic disease in poultry, could serve as a potential zoonotic hazard to humans [2,3].
APEC colonizes the intestinal tract of the chicken and has the ability to disseminate systemically either through intestinal or respiratory mucosa in the presence of stressors such as inappropriate husbandry measures [4].The conventional methods of slaughtering and evisceration that done manually, especially in developing countries, lead to overstate the incidence of carcass contamination with different bacterial pathogens [5] and accordingly a great value of the identification of these pathogens from chicken viscera.
The investigations of the zoonotic burden of chicken isolates depend on phenotypic characterization and assessment of the common serotypes that have been isolated from infected birds.However, the overlap between different serotypes makes virulence genotyping the valued method to categorize different E. coli pathotypes [6,7].Nonetheless, the role of serotyping in distinguishing APEC and other E. coli cannot be absolutely ignored, as some serotypes such as O78, O1, O2, and O18 have been more commonly associated with APEC than other pathotypes [7][8][9].
The implementation of PCR methods to screen and identify the common virulence genes and phylogroups between different isolates is crucial [10].The main phylogenetic groups to which E. coli predominately belong are A, B1, B2, and D; the virulence genes have been mostly associated with phylogroups D and B2 rather than the other phylogroups.The simple and rapid triplex PCR method that determines chuA and yjaA and a DNA fragment, TspE4.C2, also provides a link between virulence genotyping and phylotyping by identifying potential pathogenic strains from commensal ones [11].
There is a paucity of information in Egypt about the possibility of diarrheagenic virulence genes sharing between APEC strains isolated from chicken viscera and human intestinal E. coli strains.Thus, the overall objective of this study was to investigate the presence of pathogenic E. coli in chicken viscera and human stool (diarrheic and healthy samples) by conventional isolation and serotyping, followed by the determination of the virulence genes eae, stx1, and stx2 and the distribution of phylogroups in these isolates.

Sample collection
A total of 108 chicken visceral organs (liver, spleen, and heart) from 36 diseased chickens (three organs per each bird) were collected randomly from different poultry farms (each farm contained 10,000-20,000 birds) located in the district of Mansoura city (latitude 31° north and longitude 31° east), Egypt, during the first quarter of 2015.Upon necropsy, the common lesions detected in these diseased birds were pericarditis, air sacculitis, perihepatitis, ascites, splenitis, and peritonitis.The visceral organs from each bird were individually packed in a polyethylene bag and immediately transferred in an ice tank to the laboratory of the Microbiology, Immunology and Mycology Department, Faculty of Veterinary Medicine, Mansoura University, for bacteriological analysis under sterile conditions.
Seventy-eight human stool samples (50 from diarrheic patients and 28 from healthy persons) were included in this study.All the stool samples (1 sample per individual) were taken from a clinical pathology laboratory of a small charity hospital in the district of Meniet Sandoub, Mansoura, Egypt.When informed consent was received from those enrolled in this experiment, a detailed questionnaire was administered, which included information about age, health status, eating habits, and previous exposure to diarrheic episodes.Diarrheic patients admitted to the hospital presented symptoms of abdominal disturbances with foul-smelling diarrhea.All human stool samples were taken in sterile sample collection vials during the second quarter of 2015, transferred as fast as possible to the laboratory of Hygiene and Zoonoses Department, Faculty of Veterinary Medicine, Mansoura University, and tested for the presence of pathogenic E. coli.

Conventional identification of E. coli
From each human stool and chicken visceral organ, 2 g was directly enriched in tryptic soya broth (TSB; Becton Dickinson, Sparks, USA) and incubated at 37°C for 18 hours.Then, a loopful from the overnight enriched culture was streaked onto MaCconkey agar (Becton Dickinson, Sparks, USA) and incubated at 37°C aerobically for 18 to 24 hours.A single colony from each plate with the typical morphological pattern of E. coli on MacConkey was picked, streaked onto eosin methyelene blue (EMB; Becton Dickinson, Sparks, USA) and incubated overnight at 37°C.The identification of E. coli isolates depends upon the colony morphological criteria and biochemical testing [12].E. coli isolates were stored in 25% glycerolsupplemented TSB at -80°C until used.

Serotyping
The biochemically identified E. coli isolates were subjected to serotyping as described by Kok et al. [13] at the Department of Food Hygiene Control, University of Benha, Egypt, by using rapid diagnostic E. coli antisera sets (Difco Laboratories, Detroit, USA).

DNA extraction
Three representative colonies of the same morphological type were picked from the slants of the previously isolated bacteria, transferred into a tube containing 3 mL of TSB, and incubated at 37C for 18 hours.One milliliter of the overnight bacterial culture was centrifuged at 8,000 × g for 2 minutes and then sediment was washed with nuclease-free water, homogenized, and heated at 95C for 15 minutes.The supernatants from boiled lysates were used as DNA template and transferred to the Central Diagnostic and Research Laboratory, Faculty of Veterinary Medicine, Kafrelsheikh University, for the virulence gene and phylogroup identification.

Molecular identification of virulence genes
The primer pairs used (sequence, target gene, and PCR product) are summarized in Table 1 [14,15,11].PCR was performed in a volume of 15 µL consisting of 7.5 µL of 2X PCR Master Mix (Promega, Madison, USA), 0.15 µL of each primer (100 µM each), and 2 µL DNA template.PCR program for both stx1 and stx2 genes was similar to that done by Paton and Paton [15], with slight modifications, and started with an initial denaturation for 5 minutes at 95°C, 35 cycles (95°C for 30 seconds, 62°C for 30 seconds and 72°C for 30 seconds), and a final extension step at 72°C for 7 minutes.The PCR condition for the eae gene was similar to that described by Fagan et al. [16], with the annealing temperature at 64°C.DNA extracted from E. coli O157:H7 and Salmonella Typhimurium (personal unpublished data; isolated from cloacal swab of infected chicken, identified biochemically and serotyped) were used as positive and negative control in each PCR run, respectively.

Phylogenetic grouping of APEC and DEC isolates
The main E. coli phylogenetic lineages (A, B1, B2, and D) were determined by a triplex PCR for the amplification of gene specific markers chuA, yjaA, and the DNA fragment TspE4.C2 that yielded three characteristic amplicons (Table 2).PCR reaction was done using a Bio-Rad (Munich, Germany) thermal cycler with a condition similar to that done by Clermont et al. [11].

Statistical analysis
The association of phylogroup distribution between chicken and human isolates was determined by the Chisquare (χ 2 ) test performed with SPSS version 16.0 (SPSS Inc., Chicago, USA) software at a probability value p < 0.05.

Results
In this study, a total of 186 samples were tested using the conventional cultural methods mentioned to determine the occurrence of pathogenic E. coli from chicken visceral organs (n = 108) and human stool samples (n = 78; 50 diarrheic and 28 healthy subjects).Sixty-five isolates (35%) were biochemically identified to be E. coli from chicken visceral (29/108; 26.9%) and human stool samples (36/78; 46.2%).All the 29 biochemically identified E. coli isolates from chicken viscera were serogrouped and confined to 9 serotypes (Table 3).On the other hand, only 36.1% (13/36) of human E. coli isolates were identified and grouped into 8 serotypes (Table 4).
belonged to commensal phylogroups that harbored eae and stx genes.

Discussion
The overall occurrence of E. coli isolates (26.9%; 29/108) from chicken viscera in this study as well as the detected serotypes (predominantly O78:H -, O2:H6 and O1:H7) was not far from those previously reported by other researchers [2,[7][8][9].It is difficult to compare between the incidences of pathogenic E. coli along with the distribution of its serotypes among different countries owing to the contribution of many conditions such as geographic area, seasonal variation, sampling techniques, and the conventional methods used for its isolation.However, it might be helpful in the prediction of certain outbreaks caused by these serotypes, especially in countries of the same geographical and climatic conditions, with a subsequent application of a suitable control regime.
Concerning human isolates, the occurrence of pathogenic E. coli was higher among diarrheic patients than healthy persons [17,18].Different serotypes determined from the diarrheic isolates were mostly the classical non-O157 enterohemorrhagic E. coli (EHEC) serotypes such as O26:H11, O111:H2, O113:H7, O103:H2, and O145:H28 [19].From healthy persons, only two serotypes were identified (O26:H11 and O111:H2); this finding is in agreement with many previous studies which verified that O26 is considered one of the most clinically serotypes that could be isolated from both diarrheic and healthy persons [20,21].
From the above findings, the predominant APEC serotypes O78, O1, and O2 were not recovered from human stool samples.This was in agreement with many previous studies that reported the genetic relatedness between APEC and human ExPEC isolates with an explanation of the origin similarity of both isolates along with the possession of common virulence and antimicrobial resistance genes [22][23][24][25].Meanwhile, some other studies previously isolated O78 serotypes from stools of diarrheic patients [26][27][28].Interestingly, the presence of the same serotypes in this study (O26:H11 and O119:H4) isolated from diseased chicken viscera and human stool increases the awareness about the presence of substantial overlap at the level of serogroups between chicken and human isolates.
The PCR assays in our study screened a subset of chicken-and human-derived E. coli isolates with serotypes that have been associated with diarrhea in human patients.Approximately 35% of APEC isolates possessed stx genes, similar to the findings of many investigations [29][30][31] that detected stx genes from APEC isolates.However, many previous reports [32][33][34] failed to genetically identify stx genes from chicken viscera and they asserted that chickens could not act as potential sources and reservoirs of Shiga toxinproducing E. coli.
Our findings identified that two serotypes (O113:H7 and O26:H11), which belong to the classical EHEC, isolated from diarrheic patients, were eae gene positive and stx genes negative.This was similar to the findings of Kozub-Witkowski et al. [35], who molecularly identified eight strains that belonged to classical EHEC serotypes and were stx genes negative.This is not surprising, as there is still a possibility that these strains might acquire stx genes by horizontal transfer with a subsequent disease burden to humans [36], or they may have originated as EHEC strains and then lost phage-encoded stx genes [37].
Concerning phylogrouping, higher frequencies of virulent phylogroups of D and B2 were found among APEC isolates (particularly serotypes O78, O2, and O1), confirming the previous results about the distribution of extra-intestinal pathogenic strains in these phylotypes [38,39].Nonetheless, the presence of phylogroup A in approximately 25% of APEC isolates in our study, which is predominantly associated with commensal E. coli, could be ascribed to the origin of these strains as commensals that acquired virulencerelated genes [40].
The phylotype distribution of human E. coli isolates was influenced by many factors such as host genetic *A significant association of phylogroups (B2 and D) distribution between chicken and human isolates; #Non-significant association of phylogroups (B2 and D) distribution between human diarrheic and healthy isolates.
factors, dietary factors, use of medications, and geographical conditions [41,42].In this study, there was an over-representation of commensal phylogroups A and B1 among human isolates, which is consistent with the reports of previous studies that DEC isolates were included in phylogroups A, B1, and D [43,44].The molecular identification of eae and stx genes in 50% of the human commensal phylotypes in this study is in agreement with the explanation of Escobar-Paramo et al. [42], who reported that human DEC strains (which belong to the commensal phylogroups) might have the genetic precursors and virulence genes necessary for disease emergence.

Conclusions
This study concluded that there is an overlap between APEC and DEC serotypes, besides the association of intestinal virulence genes and phylogroups among these isolates.A PCR-based method for the identification of virulence genes and phylotypes with conventional phenotyping would be a valuable tool in the epidemiological surveillance that could identify the zoonotic potential sources and possible risks to humans.Nevertheless, our results recommend further investigations into these isolates using the sequence typing method to delineate their genetic diversity and clonal circulation between animal food sources and humans.

Figure 1 .
Figure 1.Phylogenetic typing of representative E. coli isolates by triplex PCR.

Table 1 .
List of primers used for virulence genes identification and phylotyping of E. coli isolates.

Table 2 .
The key of E. coli phylotyping using triplex polymerase chain reaction of chuA, yjaA, and TspE4.C2 genes.
+ Presence of the target amplicon; -Absence of the target amplicon.
. All E. coli isolates from chicken viscera in this study carried the eae gene, and only 1 (3.4%) and 10 (34.5%) isolates harbored stx1 and stx2, respectively.However, only half of the human E.

Table 3 .
The distribution of virulence genes and among chicken E. coli phenotypes.

Table 4 .
The distribution of virulence genes and phylogroups among human E. coli phenotypes.

Table 5 .
Phylogroups of E. coli isolates from chicken carcasses and human stool samples.