Fifteen E. coli strains isolated between 2012 and 2014 from broilers presenting vertebral osteomyelitis or arthritis at commercial poultry farms in the state of Minas Gerais, Southeast of Brazil, were studied. The broilers were from eight different flocks, which represent seven different farms. They had variable ages and gender. All experimental procedures were approved by the Universidade Federal de Minas Gerais (UFMG), Committee for Ethics in Animal Experimentation (CETEA) under protocol 205/2011.
Clinical signs and pathology
For clinical examination, broilers presenting locomotor disorders were placed in station and encouraged to move for change in gait and posture assessment. Then, broilers were euthanized by cervical dislocation for necropsy and gross evaluation. The locomotor system was analyzed for size, shape, color, flexibility and breaking strength. The vertebral column of all broilers was sectioned along the longitudinal midline for vertebral body and spinal cord analysis. The free thoracic vertebra was considered as T4. Articulations were analyzed for size and aspects of the synovial fluid in the articular space. Samples for bacterial isolation were collected aseptically from broilers presenting osteomyelitis or arthritis. Tissue sections were fixed in 10 % neutral buffered formalin for 48 to 56 hours. Then, formalin-fixed-vertebral column, intertarsal and femorotibial articulations with lesions were decalcified in 24 % formic acid. For slide preparation, tissues were dehydrated in increasing ethanol concentrations, diaphoanised in xylene, embedded in paraffin to obtain 4-μm thick serial sections and then stained with hematoxylin-eosin (HE) and Goodpasture for further analysis under a light microscope.
Bacterial isolation and identification
Swabs of the lesions were inoculated onto two blood agar (BA) plates and one MacConkey agar (MCK) plate. One BA plate was incubated in microaerophilic conditions at 37 °C for 24 to 72 h, while the others were incubated at the same temperature and time under aerobic conditions. After the initial growth, morphology of isolated colonies was characterized and these same colonies were subcultured, Gram stained and submitted to catalase and oxidase tests. Bacterial isolates were subjected to automatic bacterial identification through VITEK 2 system (bioMérieux, Inc. Hazelwood, MO, USA) using commercially available identification cards for Gram-negative bacteria in accordance to the manufacturer’s recommendations. After bacterial identification, the colonies were inoculated into microtubes containing Brain-Heart Infusion (BHI) broth with 30 % glycerol and stored at - 80 °C until subsequent molecular and phenotypic tests described below.
APEC diagnosis tests
The diagnosis of APEC strains was performed by different methods previously described. The ability of E. coli strains to induce lethality in 1-day-old specific-pathogen-free (SPF) chicks (detailed on section Lethality test) was considered gold standard test to assess strain pathogenicity. In addition, two molecular methods based on genetic profiles were used: 1) detection of minimal predictors described by Johnson et al. , which classify an E. coli strain as pathogenic based on the minimum detection of four out of five virulence genes (iroN, ompT, hlyF, iss and iutA); and 2) genotyping method developed by Schouler et al. , which is based on the identification of different associations of virulence genes (iutA, sitA, aec26, P (F11) fimbriae, O78, frz
) that allow the APEC strains classification in four genetic patterns of virulence (A, B, C and D).
Determination of O antigens was carried out by agglutination using antisera O1, O2, O5, O8, O15, O18, O25, O45, O78, O88, O111 and O120, according to the method described by Blanco et al. . The O antisera were produced in the Laboratorio de Referencia de Escherichia coli (Lugo, Spain). Furthermore, PCR was performed to detect O1, O2, O4, O6, O7, O8, O12, O16, O18, O25a, O45a, O45b, O75, O78, O88 and O104 antigens, as previously described (Additional file 1: Table S1).
The strains were submitted to PCR to determine flagella type H4, H7, H8, H21 and H25 (Additional file 1: Table S1). Those strains negative for all flagellar types tested by PCR were submitted to motility test. Briefly, bacteria were grown on LB broth overnight. Then, the strains were deeply inoculated in LB plates 0.3 % agar using a Pasteur pipette and then incubated at 37 °C overnight for motility evaluation the following day .
ECOR phylogenetic grouping
E. coli strains were classified into the four main ECOR phylogenetic groups by triplex PCR as described by Clermont et al. . Strains were assigned to phylogenetic groups A, B1, B2, or D according to the amplification of the chuA and yjaA genes and the TspE4C2 fragment. Strains MG1655, ECOR26, ECOR62, and ECOR50 were used as controls for phylogenetic groups A, B1, B2, and D, respectively.
Total DNA extracts were prepared by a rapid boiling method . The presence of genes encoding virulence factors were determined using primers and PCR amplification programs previously described, together with positive control strains (Additional file 1: Table S2).
Single PCR assays were used to detect sfaS, focG, tsh, ibeA, aatA, neuC, irp2, ireA, sat, vat, astA, fyuA, hlyA, traT, cva/cvi, iucD, hra, iha, pic, csgA, tia, malX (=rpai), KpsMTII, cnf 1 and cnf 2. Furthermore, some multiplex assays were performed to detect simultaneously clbB and clbN, and fimA, fim
and fimH. DNA fragments were amplified in a 25-μL PCR mix containing 1 U of GoTaq®G2 Flexi DNA polymerase (Promega), 12.5 pmol of the forward and reverse primers, and 5 nmol of deoxynucleotide triphosphate mix (Eurogentec) in 1x GoTaq®G2 Flexi buffer. The PCR conditions were as follows: initial denaturation at 94 °C for 4 to 5 min, followed by 30 cycles of 94 °C for 30 s, annealing temperature according to GC-content of primers for at least 30 s, 72 °C for 30 s to 45 s according to the size of the amplified fragment (1 min/kbp), and then a final extension at 72 °C for 7 min.
Pulsed-field gel eletrophoresis (PFGE)
Pulsed-field gel electrophoresis was conducted as previously described . Bacterial cells (equivalent to an OD600 of 1.0) grown in BHI broth were harvested by centrifugation. The cellular pellet was resuspended in 500 μL of buffer TE 100 (10 mM Tris/HCl, pH 9, 100 mM EDTA) and incubated for 30 min at 37 °C. The bacterial suspension was mixed with an equal volume of 2.0 % low-melting-point agarose and dispensed into plug molds (Biorad). Agarose plugs were incubated in a lysozyme solution (10 mM Tris/HCl, pH 9, 100 mM EDTA, 5 mg lysozyme ml−1, 0.05 % sarkosyl) for 2 h at 37 °C, and then incubated overnight at 55 °C in a lysis solution (10 mM Tris/HCl, pH 9, 100 mM EDTA, 1 mg proteinase K ml−1, 1 % SDS). After lysis, agarose plugs were washed three times in a TE buffer (10 mM Tris/HCl, pH8, 1 mM EDTA) for 1 h at room temperature, where the first washing buffer was supplemented with 100 mM PMSF (Phenylmethylsulfonyl fluoride). For digestion, plugs were equilibrated in incubation buffer containing XbaI restriction enzyme (Takara) overnight. Pulsed-field gel electrophoresis was conducted in a CHEF-DRIII apparatus (Bio-Rad). Gels (1 % agarose) were run at 14 °C for 24 h in TBE buffer (4 mM Tris, 4 mM borate, 1 mM EDTA, pH 8.3) at 6 V cm−1. Pulse times were increased from 10 to 30 s. XbaI restriction fragments of Salmonella enterica serovar Braenderup H9812 were used as size markers. Cluster analysis using Dice similarity indices was done in BioNumerics 6.6 software (at 0.5 % tolerance and 0.5 % optimization) (Applied Maths, Ghent, Belgium) to generate a dendrogram describing the relationships among PFGE profiles.
Multilocus Sequence Typing (MLST)
The phylogenetic relationships between strains were studied using MLST method initially described by Maiden et al.  and E. coli Achtman’s scheme (http://mlst.warwick.ac.uk/mlst/dbs/Ecoli/). E. coli MLST scheme used internal fragments of seven house-keeping genes: adk (adenylate kinase), fumC (fumarate hydratase), gyrB (DNA gyrase), icd (isocitrate/isopropylmalate dehydrogenase), mdh (malate dehydrogenase), purA (adenylosuccinate dehydrogenase) and recA (ATP/GTP binding motif). They were amplified in a total volume of 20 μL containing 4 μL of DNA crude extract as a template, 2.5 U of GoTaq®G2 Flexi DNA polymerase (Promega), 10 pmol of each primer, 5 nmol of deoxynucleoside triphosphate 30 mM MgCl2 in 1x buffer. PCR conditions were as follows: 94 °C for 5 min; 30 cycles of 94 °C for 40 s, variable annealing temperature (54 °C, 60 °C, 64 °C, 58 °C, 62 °C, 62 °C or 58 °C, respectively) for 45 s, and 72 °C for 45 s; and a final extension at 72 °C for 5 min. The amplicons were sequenced on both strands and sequence type (ST) of each allele was attributed according to Achtman’s scheme. Novel STs described in this work were submitted to the E. coli MLST database and identified as ST5856 and ST5766.
Strain virulence was evaluated by a lethality test using 1-day-old chicks as previously described . Lethality test was carried out in the experimental infection unit PFIE (Plateforme d’Infectiologie Expérimentale, INRA Val de Loire). For each strain, groups of five 1-day-old SPF chicks were inoculated subcutaneously with 0.5 mL of an overnight culture in LB-Miller broth without agitation (inoculum in stationary phase was ~108 CFU). Mortality was recorded at 4 days post inoculation and the strains were classified as pathogenic when at least one chick died . Avian E. coli strains BEN2908 and BEN2269 (a non-pathogenic avian E. coli isolate of serogroup O2) were used as positive (5 chicks died) and negative control (no chicks died), respectively. The housing, husbandry and slaughtering conditions conformed to European Guidelines for care and use of laboratory animals. French regional ethics committee number 19 (Comité d’Ethique en Expérimentation Animale Val de Loire) approved this protocol under the reference 2012-11-5.
Serum bactericidal test
The serum bactericidal assay was performed as previously described by Dozois et al.  with some modifications. Briefly, bacteria were grown overnight in LB broth at 41 °C with agitation (180 rpm). Then, bacterial cultures were resuspended in fresh medium (OD600 = 0.02), incubated at 41 °C with agitation (180 rpm), and harvested during the logarithmic growth phase (DO600 = 0.35). Bacteria were washed at room temperature with dulbeco’s phosphate-buffered saline (pH 7.0 ~ 7.3) and then resuspended to a concentration of 2x106 CFU/mL. A volume of 500 μL of bacterial suspension was added to 500 μL of complement or inactivated (56 °C, 30 min) SPF chicken serum, which were incubated at 41 °C without agitation. Viable cell counts were counted at 0 h and 3 h by plating 10-fold dilutions in sterile saline solution on LB agar plates. Compared to the bacterial count in inactivated serum, a strain was considered resistant when the bacterial count increased or did not change, intermediate when the bacterial count decreased up to one order of magnitude, and sensitive when bacterial count decreased more than one order of magnitude. Serum resistant (BEN2908) and serum sensitive (BEN4134) E. coli strains were used as positive and negative controls.
Antibiotic susceptibility testing
Susceptibility testing was performed by the disk diffusion method according to the guidelines of the Antibiogram Committee of the French Society of Microbiology (http://www.sfm-microbiologie.org). The antibiotics tested belong to seven different classes: aminoglycosides (gentamicin, Gen; neomycin, Neo; apramycin, Apr), beta-lactams (amoxicillin, Amx; amoxicillin + clavulanic acid, Amc), cephalosporins (cephalotin, Cef; cefoxitin, Fox; ceftiofur, Xnl), phenicols (florfenicol, Ffc), polypeptides (colistin, Cst), quinolones (nalidixic acid, Nal; flumequine, UB; enrofloxacin, Enr), sulfonamides (trimethoprim, Tmp; Tmp + sulfamethoxazole, TmpStx), and tetracyclines (tetracycline, Tet). The presence of extended spectrum β-lactamases (ESBL) was detected by double-disk synergy method . E. coli ATCC 25922 strain was used as quality control.