Thymus transcriptome reveals novel pathways in response to avian pathogenic Escherichia coli infection.

Avian pathogenic Escherichia coli (APEC) can cause significant morbidity in chickens. The thymus provides the essential environment for T cell development; however, the thymus transcriptome has not been examined for gene expression in response to APEC infection. An improved understanding of the host genomic response to APEC infection could inform future breeding programs for disease resistance and APEC control. We therefore analyzed the transcriptome of the thymus of birds challenged with APEC, contrasting susceptible and resistant phenotypes. Thousands of genes were differentially expressed in birds of the 5-day post infection (dpi) challenged-susceptible group vs. 5 dpi non-challenged, in 5 dpi challenged-susceptible vs. 5 dpi challenged-resistant birds, as well as in 5 dpi vs. one dpi challenged-susceptible birds. The Toll-like receptor signaling pathway was the major innate immune response for birds to respond to APEC infection. Moreover, lysosome and cell adhesion molecules pathways were common mechanisms for chicken response to APEC infection. The T-cell receptor signaling pathway, cell cycle, and p53 signaling pathways were significantly activated in resistant birds to resist APEC infection. These results provide a comprehensive assessment of global gene networks and biological functionalities of differentially expressed genes in the thymus under APEC infection. These findings provide novel insights into key molecular genetic mechanisms that differentiate host resistance from susceptibility in this primary lymphoid tissue, the thymus.

Detection of Iss and Bor on the surface of Escherichia coli.

AIMS: To confirm the presence of Iss and Bor on the outer membrane of Escherichia coli using Western blots of outer membrane protein (OMP) preparations and fluorescence microscopy, and explore the use of fluorescence microscopy for the detection of avian pathogenic E. coli (APEC) and diagnosis of avian colibacillosis. METHODS AND RESULTS: Knockout mutants of iss and bor were created using a one-step recombination of target genes with PCR-generated antibiotic resistance cassettes. Anti-Iss monoclonal antibodies (Mabs) that cross-react with Bor protein were used to study the mutants relative to the wild-type organism. These Mabs were used as reagents to study OMP preparations of the mutants with Western blotting and intact E. coli cells with fluorescence microscopy. Iss and Bor were detected in Western blots of OMP preparations of the wild type. Also, Iss was detected on Deltabor mutants, and Bor was detected on Deltaiss mutants. Iss and Bor were also detected on the surface of the intact, wild-type cells and mutants using fluorescence microscopy. CONCLUSIONS: These results demonstrate that Bor and Iss are exposed on E. coli's outer membrane where they may be recognized by the host's immune system. SIGNIFICANCE AND IMPACT OF THE STUDY: To our knowledge, this is the first report confirming Iss' location in the outer membrane of an E. coli isolate. Such surface exposure has implications for the use of these Mabs for APEC detection and colibacillosis control.

Biofilm formation by avian Escherichia coli in relation to media, source and phylogeny.

AIMS: To assess the abilities of 105 avian pathogenic Escherichia coli (APEC) and 103 avian faecal commensal E. coli (AFEC) to form biofilms on a plastic surface and to investigate the possible association of biofilm formation with the phylotype of these isolates. METHODS AND RESULTS: Biofilm production was assessed in 96-well microtitre plates using three different media, namely, M63 minimal medium supplemented with glucose and casamino acids, brain-heart infusion broth, and diluted tryptic soy broth. Avian E. coli are highly variable in their ability to form biofilms. In fact, no strain produced a strong biofilm in all three types of media; however, most (75.7% AFEC and 55.2% APEC) were able to form a moderate or strong biofilm in at least one medium. Biofilm formation in APEC seems to be mostly limited to nutrient deplete media; whereas, AFEC are able to form biofilms in both nutrient deplete and replete media. Also, biofilm formation in E. coli from phylogenetic groups B2, D and B1 was induced by nutrient deplete conditions; whereas, biofilm formation by members of phylogenetic group A was strongest in a rich medium. CONCLUSIONS: Biofilm formation by APEC and phylotypes B2, D and B1 is induced by nutrient deplete conditions, while AFEC are able to form biofilms in both nutrient rich and deplete media. SIGNIFICANCE AND IMPACT OF THE STUDY: This is the first study to investigate biofilm formation by a large sample of avian E. coli isolates, and it provides insight into the conditions that induce biofilm formation in relation to the source (APEC or AFEC) and phylogenetic group (A, B1, B2 and D) of an isolate.

Cloning and sequencing of cnf1 from Escherichia coli incriminated in mink and bovine colibacillosis.

Colibacillosis is responsible for significant losses to the mink and cattle industries. Previous work in our laboratory and by others has suggested that possession of cnf1, the gene encoding cytotoxic necrotizing factor (CNF1), may contribute to the virulence of isolates of E. coli from mink and cattle. The cnf1 gene from E. coli isolated from a mink with colisepticaemia and a bovid with scours was amplified and cloned as a 3.5 kb fragment, and the fragment was sequenced. The cnf1 sequences from the mink and bovine isolates of E. coli were compared to each other and to cnf1 sequences of E. coli from urinary tract and diarrhoea-associated infections of humans. The difference was only 7 nucleotides between the cnf1 sequences of the mink and bovine isolates of E. coli, which translated into 7 differences in amino acids. The cnf1 sequence of the mink isolate of E. coli had 15 nucleotide differences from the cnf1 sequences of the human isolate of E. coli (GenBank X70670), which translated into 11 differences in amino acids between these proteins. The cnf1 sequence of the bovine isolate of E. coli had 14 nucleotide differences from the cnf1 sequence of the human isolate of E. coli (GenBank X70670), which translated into 10 differences in amino acids between these proteins. The highly conserved sequences of the amino acids of CNF1 proteins make them a promising target for detection and control of the CNF1-producing E. coli involved in disease among various host species.

Characterization of Escherichia coli isolates incriminated in colisepticaemia in mink.

Colisepticaemia is a major health and economic concern for the mink industry, yet little information is available about the Escherichia coli that cause this disease. In this study, 40 E. coli, isolated from mink clinically diagnosed with colisepticaemia that had been submitted to the North Dakota State University Veterinary Diagnostic Laboratory, were randomly selected for characterization. These isolates were serotyped and screened for resistance to 18 antimicrobials, possession of transmissible R plasmids, and the presence of several virulence traits or genes using bioassays or the polymerase chain reaction. Several serotypes were identified that have previously been associated with septicaemia in other animal species. The majority of the isolates exhibited multiple antimicrobial resistance phenotypes. Common resistance phenotypes observed included those to tetracycline, sulfamethoxazole, streptomycin, ampicillin and kanamycin. Several of the isolates that could be studied by conjugation contained transmissible R plasmids coding for multiple antimicrobial resistance phenotypes. About half of the isolates produced colicin; all produced enterobactin: and all but one-quarter produced aerobactin. None of the isolates tested produced enterohaemolysin, and one-fifth were considered to be beta haemolytic. About half appeared to contain the gene encoding cytotoxic necrotizing factor-1; three contained the gene encoding EAE, but none appeared to contain the genes coding for LT, Sta/b, SLT-I/II or CNF-II toxins or K99 antigen. Approximately one-third of the isolates elaborated capsule. The results show that the E. coli isolates implicated in mink colisepticaemia possess similar virulence traits and antimicrobial resistance phenotypes to those associated with diarrhoeal diseases in food animals.

Resistance to serum complement, iss, and virulence of avian Escherichia coli.

Control of avian colibacillosis is hampered by lack of easily identifiable markers for virulent Escherichia coli. Resistance to serum complement appears to be a widespread trait of virulent avian E. coil, suggesting that bacterial factors promoting survival in serum may be useful in discriminating between virulent and avirulent isolates. Such distinguishing factors may prove useful in diagnostic protocols or as targets in future colibacillosis control protocols. Interestingly, the factors responsible for resistance to complement differ in the E. coli isolated from mammalian and avian hosts, which may reflect differences in the nature of avian and mammalian colibacillosis. In some cases, genetic determinants for serum complement resistance in avian E. coli are found on aerobactin- or Colicin V-encoding plasmids. One such gene, iss, first described for its role in the serum resistance associated with a ColV plasmid from a human E. coli isolate, occurs much more frequently in isolates from birds with colibacillosis than in faecal isolates from healthy birds. Efforts to identify the genomic location of iss in a single, virulent avian E. coli isolate have revealed that it occurs in association with several purported virulence genes, all linked to a large conjugative R plasmid. At this time, it is not known whether iss merely marks the presence of a larger pathogenicity unit or is itself a contributor to virulence. Nevertheless, the presence of the complement-resistance determinant, iss, may be a marker of virulent avian E. coli exploitable in controlling avian colibacillosis.

Virulence factors of Escherichia cofi from cellulitis or colisepticemia lesions in chickens.

This study was designed to compare virulence factors of cellulitis-derived Escherichia coli to colisepticemic E. coli in order to clarify whether E. coli associated with cellulitis comprise a unique subset of pathogenic E. coli. Isolates were tested for serotype, capsule, aerobactin production, colicin production, the presence of the iss gene, and serum resistance. Untypable isolates made up the greatest percentage of each group. Serotypes O2 and O78 were the most commonly identified among both groups of isolates. No statistical differences in the distribution of aerobactin or colicin production, capsule, or iss gene were observed between groups. Cluster analysis showed that 90% of the E. coli isolates had greater than 42% livability in serum-resistance tests. No separation of colisepticemic vs. cellulitis E. coli isolates was observed on the basis of SR. Colicin production by E. coli was highly correlated with serum resistance (P = 0.0029). These data suggest that cellulitis E. coli have virulence traits similar to those of colisepticemic E. coli.

Complement resistance-related traits among Escherichia coli isolates from apparently healthy birds and birds with colibacillosis.

In this study, 294 Escherichia coli isolates from birds with colibacillosis were collected from disease outbreaks throughout the United States and were compared with 75 fecal E. coli isolates of apparently healthy chickens by their possession of several purported virulence genes, resistance to rough-lipopolysaccharide-specific bacteriophages (rLPSr), and elaboration of capsule. Traits were selected for study on the basis of their association with complement resistance. The genes targeted in this study included those encoding colicin V (cvaC) and the outer membrane proteins TraT (traT), OmpA (ompA), and Iss (iss). No significant differences were found between the two groups of isolates in the occurrence of cvaC-, traT-, or ompA-homologous sequences or in rLPSr. Only a few isolates were encapsulated, and the isolates of healthy birds were significantly more likely to be encapsulated than were the isolates of sick birds. However, iss, whether detected through hybridization or amplification, was found in more of the disease-associated isolates than in those of healthy birds. This difference was highly significant. Further, iss sequences were widely distributed among isolates of different serotypes from various avian host species and sites within these hosts. Such results suggest that possession of the iss sequence by an avian E. coli isolate may be a good indicator of that isolate's potential to cause disease. This association warrants further study because iss and the protein it encodes may be useful targets of future colibacillosis control efforts.

Cloning and sequencing of the iss gene from a virulent avian Escherichia coli.

Control of colibacillosis is important to the poultry industry. We have found that the presence of a gene for increased serum survival, iss, is strongly correlated with Escherichia coli isolated from birds with colibacillosis. Therefore, the iss gene and its protein product, Iss, are potential targets for detection and control of avian colibacillosis. The iss gene was amplified from a virulent avian E. coli isolate and sequenced. The sequences of the gene and the predicted protein product were compared with those of iss from a human E. coli isolate and lambda bor. The iss gene from the avian E. coli isolate has 96.8% identity with the iss gene from the human E. coli isolate and 89.4% identity with lambda bor. The Iss protein from the avian isolate has 87% identity with Iss from the human isolate and 90% identity with Bor. The low identity between the two Iss proteins is because of a frame-shift in their respective coding sequences. In sum, iss from this avian E. coli isolate is very similar to iss from a human E. coli isolate, but because of a frameshift mutation in the coding sequence of iss from the human E. coli isolate, Iss proteins from avian and human E. coli isolates have only 87% identity. The strong association of iss with E. coli isolated from birds with colibacillosis, suggests that this sequence be studied for its value as a marker or target to be used in colibacillosis control.

Iss from a virulent avian Escherichia coli.

No single characteristic of virulent avian Escherichia coli has been identified that can be exploited in colibacillosis detection protocols. Research in our lab suggests a strong association between the presence of an iss DNA sequence with an isolate's disease-causing ability. The study presented here focuses on the techniques used in the expression, purification, and characterization of avian E. coli Iss protein. In brief, iss was cloned into an expression vector, the construct was transformed into a protease-deficient E. coli, and expression was induced. The protein was expressed as a glutathione-S-transferase (GST) fusion and purified by affinity chromatography. The GST portion was cleaved from Iss, Iss was harvested by affinity chromatography, and the identity of Iss was confirmed by N-terminal sequencing. Currently, purified Iss is being used to prepare hybridomas for production of monoclonal antibodies with the goal of evaluating anti-Iss as a reagent for the detection of virulent avian E. coli.

Characterization of Salmonella dublin and Salmonella typhimurium (Copenhagen) isolates from cattle.

Eight Salmonella typhimurium (Copenhagen) and eight Salmonella dublin isolates from cattle were compared by their antibiotic resistance patterns, by their production of colicin, aerobactin, haemolysin and capsule, by their possession of transmissible R plasmids and the spvC gene, and by their ability to invade and replicate within cultured epithelial cells. The two groups differed in their antibiotic resistance profiles, with more of the host-adapted S. dublin isolates resistant to tetracycline than were the non-host-adapted S. typhimurium (Copenhagen) group, but more of the S. typhimurium (Copenhagen) isolates resistant to the other antibiotics tested. None of the isolates produced colicin, but all produced aerobactin. One isolate in each group was encapsulated. All of the S. typhimurium (Copenhagen) and S. dublin isolates contained plasmids, and all of them contained the spvC-homologous sequences. Four of the S. typhimurium (Copenhagen) isolates were able to transfer an R plasmid to a recipient organism by conjugation. One of the five S. dublin isolates, which showed resistance to some of the antibiotics tested, was able to transfer an R plasmid by conjugation. Both groups of isolates invaded cultured epithelial cells to a similar degree after 1 h, but the S. dublin isolates reached significantly higher levels within the cells than did S. typhimurium (Copenhagen) after 9 h. This ability may, in part, explain the association of S. dublin with more severe forms of salmonellosis and prolonged carrier states. Further study of the intracellular growth of these isolates seems warranted.

Decreased intracellular survival of an fkpA mutant of Salmonella typhimurium Copenhagen.

The fkpA gene of Salmonella typhimurium encodes a protein similar to the macrophage infectivity potentiator (Mip) proteins of Legionella pneumophila and Chlamydia trachomatis. Because Mip proteins enhance the ability of these intracellular pathogens to survive within macrophages and epithelial cells, we tested whether the product of the fkpA gene would have the same effect on the intracellular growth of a virulent strain of S. typhimurium. By a series of P22 transductions, the fkpA gene of S. typhimurium Copenhagen was replaced with the inactive fkpA1::omega-Cm gene from Escherichia coli, creating the mutant S. typhimurium KY32H1. The Copenhagen and KY32H1 strains were equally able to enter Caco-2 cells (an epithelial cell line) and J774.A1 cells (a macrophage-like cell line). However, compared to the parent, the fkpA mutant survived less well in both types of cells during the first 6 h after infection. The number of viable intracellular S. typhimurium Copenhagen bacteria remained constant 6 h after infection of Caco-2 cells, but the viability of S. typhimurium KY32H1 decreased significantly by 4 h postinfection. The fkpA mutant also exhibited a reduced ability to survive intracellularly in J774.A1 cells as little as 2 h postinfection. Complementation of the fkpA mutation by a plasmid-borne wild-type fkpA gene from E. coli restored the ability of S. typhimurium KY32H1 to grow normally in J774.A1 cells. Thus, expression of the mip-like fkpA gene confers on S. typhimurium Copenhagen properties analogous to those mediated by the Mip proteins in other intracellular pathogens, suggesting that this mechanism may play a role in the virulence and/or intracellular growth of numerous bacteria.

Further characterization of a complement-sensitive mutant of a virulent avian Escherichia coli isolate.

An attempt was made to characterize the mechanism of complement resistance operating in a virulent avian Escherichia coli isolate. Using flow cytometry to detect antibody to C3, we found that there was significantly more antibody bound to a complement-sensitive mutant of this wild type than to the parent organism, suggesting that more C3 subunits were bound to the wild type. Neither the wild type nor the mutant degraded C3. Further, the mutant was phagocytosed to a significantly greater degree than the wild type by cultured phagocytes in the presence of C5-deficient serum. These data suggest that the wild type is resistant to complement, at least in part, because of its ability to restrict C3 deposition on its surface. Therefore, the decrease in virulence seen in the mutant may be related to its increased sensitivity to complement-mediated bacteriolysis or its enhanced susceptibility to complement-opsonized phagocytosis or both.

Large plasmids of avian Escherichia coli isolates.

The plasmid DNA of 30 Escherichia coli isolates from chickens was extracted and examined using techniques designed to isolate large plasmids. This plasmid DNA was examined for the presence of certain known virulence-related genes including cvaC, traT, and some aerobactin-related sequences. Seventeen of the 30 isolates contained from one to four plasmids greater than 50 kb in size. Eleven of these 17 strains possessed plasmids greater than 100 kb in size. Therefore, E. coli isolates of chickens frequently contain large plasmids, and many of these plasmids are likely to contain virulence-related sequences.

Analysis of plasmids cloned from a virulent avian Escherichia coli and transformed into Escherichia coli DH5 alpha.

Three of four plasmids from a virulent wild-type avian Escherichia coli were cloned or transformed into an avirulent laboratory recipient E. coli DH5 alpha and tested for the ability to confer a virulence phenotype. The three plasmids transformed into E. coli DH5 alpha were 5, 6, and 56 kb. A fourth plasmid of 64 kb was not successfully transformed. Parameters used to measure virulence included presence of type 1 pili and a smooth lipopolysaccharide (LPS) layer, motility, production of Colicin V, resistance to host complement, and embryo lethality. The 5-kb plasmid encoded for ampicillin resistance, whereas the 6-kb plasmid encoded for tetracycline resistance. The 56-kb plasmid encoded for streptomycin, sulfisoxazole, and tetracycline resistance. Twelve-day embryos inoculated with 467 colony-forming units of E. coli DH5 alpha containing the 56-kb plasmid had increased death rates (45%) in the embryo lethality assay and a decreased weight of surviving embryos with cranial hemorrhages as compared with embryos inoculated with similar amounts of E. coli DH5 alpha (0%) and phosphate-buffered saline (0%). Embryos inoculated with the wild-type virulent E. coli had 90% deaths. The 56-kb plasmid also had homology with a probe for Colicin V production (cvaC). No differences in LPS layer, complement resistance, motility, Colicin V activity, type 1 pili, cell-free supernatant proteins, or outer membrane proteins were observed in the transformants when compared with nontransformed E. coli DH5 alpha.

Invasion of Caco-2 cells by Salmonella typhimurium (Copenhagen) isolates from healthy and sick chickens.

In a previous study, Salmonella isolates of sick birds were distinguished from those of apparently healthy birds by their high degree of invasion of tissue culture cells. In this study, a single pair of Salmonella isolates was examined to determine the source of this observed difference in invasion. When isolates were allowed to invade Caco-2 cells for 8 hours, the isolate from the sick bird (S) appeared to invade in greater numbers than did the isolate from the healthy bird (H). However, when invasion was distinguished from intracellular growth/survival, it was found that H invaded in greater numbers than S, but once inside the cell, H declined in number, and S increased. Inhibition of RNA, protein, and DNA syntheses lessened the degree to which both invaded. The presence of mannose inhibited invasion by S but did not appear to inhibit invasion by H. Trypsin treatment of monolayers affected invasion of S and H, whereas neuraminidase treatment did not. There was no significant difference noted between S and H in ability to adhere to fixed monolayers. Therefore, the two isolates tested differ in their mechanisms of entry into Caco-2 cells, the efficiency with which they invade, and their ability to survive within Caco-2 cells.

The distribution of invA, pagC and spvC genes among Salmonella isolates from animals.

New molecular diagnostic techniques often rely on hybridization or amplification of specific DNA regions to detect pathogenic bacteria. The choice of genes to be used as probes or as the targets of amplification techniques is critical to the success of these procedures. The genes so used might best be those associated with virulent isolates and having a wide distribution among such isolates. In this study three genes, invA, pagC and spvC, thought to be associated with the virulence of salmonellae, were labelled and used to probe the total DNA from 103 Salmonella isolates from animals in an attempt to determine whether these genes might be useful in diagnostic procedures. pagC was detected in 99% of the Salmonella tested, and invA was detected in 94.2% of the isolates. Both pagC and invA were detected with a significantly higher frequency than spvC in isolates from chickens and swine, but no significant difference in detection of these three genes occurred when bovine isolates were examined. Failure to detect any of these genes occurred in only one isolate. Isolates from apparently healthy or from clinically ill chickens and swine could not be distinguished by detecting these three genes. The genes were not detected in the non-Salmonella strains tested. These results suggest that, of these three genes, pagC may be the best choice for use as a probe or polymerase chain reaction target in future detection protocols.

Detection and characterization of Salmonella typhimurium from a dairy herd in North Dakota.

Nasal secretions, faecal samples and buffy coats were obtained from 102 cattle from a North Dakota dairy herd with a history of calf scours. Treated buffy coats, faecal samples and nasal secretions were inoculated into tetrathionate broth (TB), incubated at 37 degrees C overnight, and plated onto brilliant green agar medium with novobiocin (BGAN). The TB was left at room temperature for 5 days and then used to inoculate fresh TB. The fresh TB was incubated at 37 degrees C over night and plated onto BGAN medium. All the plates were incubated at 37 degrees C over night and observed for Salmonella-like growth. Suspect colonies were further tested and Salmonella isolates were serotyped by the National Veterinary Services laboratory. Twenty-two of the 36 calves sampled harboured S. typhimurium in their faeces, but no samples from cows were positive. No Salmonella were isolated from the buffy coats, but 4 calves were shown to have Salmonella in their nasal secretions. Extended enrichment of the faecal cultures in TB resulted in a significant increase in Salmonella isolations, although 2 samples were positive following the initial enrichment period and not after secondary enrichment. The typical Salmonella isolate detected from this herd contained a transmissible R-plasmid encoding resistance to tetracycline, kanamycin, sulphisoxazole and ampicillin. This study confirmed that delayed secondary enrichment in TB is superior to primary enrichment for detection of Salmonella from cattle.

Phenotypic expression of recombinant plasmids pKT107 and pHK11 in an avirulent avian Escherichia coli.

An avirulent wild-type avian Escherichia coli strain (Av) was electrotransformed with plasmids coding for complement resistance (pKT107) and Colicin V (ColV) production (pHK11) in order to study the effects of complement resistance and ColV production on virulence. Transformants were also compared with the wild type for embryo lethality, uptake by macrophages, motility, growth rate, plasmid content, and hemolysis. Growth rates and complement resistance patterns of strain Av and transformant Av+pHK11 were similar, but Av+pHK11 caused a significantly greater number of deaths in embryos and acquired motility. Transformant Av+pKT107 had a lower rate of phagocytosis, a slower growth rate, and a greater sensitivity to complement, and it changed from being non-hemolytic to expressing alpha-hemolytic action. The 35-kb plasmid present in the wild type was not present in the transformants. Although some of the results demonstrate the difficulties encountered in using wild-type organisms as recipients in virulence studies, the results with Av+pHK11 indicate that ColV production plus the acquisition of motility contributes to the virulence of avian E. coli.

Effect of normal intestinal flora of chickens on colonization by virulent colicin V-producing, avirulent, and mutant colicin V-producing avian Escherichia coli.

Colonization of the intestinal tracts of newly hatched chicks with Escherichia coli was attempted by swabbing test organisms onto the air-shell of 19-day-old embryos. Test organisms consisted of two virulent E. coli isolates, one avirulent isolate, and one laboratory-derived mutant of the avirulent isolate carrying a recombinant plasmid coding for Colicin V production. Chicks were cultured weekly for 3 weeks for total E. coli and for the test organisms using selective media. Control chicks were sampled on weeks 1 and 5, and the normal E. coli intestinal microflora were examined for the production of colicins. The two virulent E. coli isolates maintained colonization of the chicks for the 3-week test period, with titers decreasing from 10' to 10'- colony-forming units (CFU)/g of intestine. The avirulent isolate and laboratory mutant did not consistently colonize the intestinal tracts. The majority of intestinal samples taken from the control chicks at 1 and 5 weeks had colicin-producing E. coli that were inhibitory to the test organisms.