The extraintestinal pathogen, avian pathogenic E. coli (APEC), known to cause systemic infections in chickens, is responsible for large economic losses in the poultry industry worldwide. In order to identify genes involved in the early essential stages of pathogenesis, namely adhesion and colonization, Signature-tagged mutagenesis (STM) was applied to a previously established lung colonization model of infection by generating and screening a total of 1,800 mutants of an APEC strain IMT5155 (O2:K1:H5; Sequence type complex 95). The study led to the identification of new genes of interest, including two adhesins, one of which coded for a novel APEC fimbrial adhesin (Yqi) not described for its role in APEC pathogenesis to date. Its gene product has been temporarily designated ExPEC Adhesin I (EA/I) until the adhesin-specific receptor is identified. Deletion of the ExPEC adhesin I gene resulted in reduced colonization ability by APEC strain IMT5155 both in vitro and in vivo. Furthermore, complementation of the adhesin gene restored its ability to colonize epithelial cells in vitro. The ExPEC adhesin I protein was successfully expressed in vitro. Electron microscopy of an afimbriate strain E. coli AAEC189 over-expressed with the putative EA/I gene cluster revealed short fimbrial-like appendages protruding out of the bacterial outer membrane. We observed that this adhesin coding gene yqi is prevalent among extraintestinal pathogenic E. coli (ExPEC) isolates, including APEC (54.4%), uropathogenic E. coli (UPEC) (65.9%) and newborn meningitic E. coli (NMEC) (60.0%), and absent in all of the 153 intestinal pathogenic E. coli strains tested, thereby validating the designation of the adhesin as ExPEC Adhesin I. In addition, prevalence of EA/I was most frequently associated with the B2 group of the EcoR classification and ST95 complex of the multi locus sequence typing (MLST) scheme, with evidence of a positive selection within this highly pathogenic complex. This is the first report of the newly identified and functionally characterized ExPEC adhesin I and its significant role during APEC infection in chickens.
Citation: Antão E-M, Ewers C, Gürlebeck D, Preisinger R, Homeier T, Li G, et al. (2009) Signature-Tagged Mutagenesis in a Chicken Infection Model Leads to the Identification of a Novel Avian Pathogenic Escherichia coli Fimbrial Adhesin. PLoS ONE 4(11): e7796. doi:10.1371/journal.pone.0007796
Editor: Leonardo A. Sechi, Universita di Sassari, Italy
Received: September 22, 2009; Accepted: October 7, 2009; Published: November 12, 2009
Copyright: © 2009 Antao et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This work was funded by the German Research Foundation (DFG) grants GRK-1121 (Project A2) and Wi1436/5-3, and supported by the National Genome Research Network within the framework of the “Funktionelle Genomanalyse Tierische Organismen (FUGATO) E. coli chick Project” funded by the Federal Ministry of Education and Research (BMBF). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Research on avian pathogenic E. coli (APEC) has increased greatly over the years where the pathogen has been known to cause disease among chickens and other fowls which usually results in large economic losses for the poultry industry . APEC are mostly associated with infection of extraintestinal tissues in chickens, turkeys, ducks and other avian species. The most important disease syndrome associated with APEC begins as a respiratory tract infection and may be referred to as aerosacculitis or the air sac disease . This inevitably results in severe systemic infection leading to death of the animal infected. These pathogens have recently gained even more importance, now being classified under the category of extraintestinal pathogenic E. coli (ExPEC), a group which includes both human pathogens like the uropathogenic E. coli (UPEC) and newborn meningitic E. coli (NMEC) and animal pathogens, which in turn suggest the possibility of APEC having zoonotic potential –.
Microbial pathogenicity is a complex phenomenon encompassing diverse mechanisms. There are, however, several common strategies that pathogenic organisms use to sustain themselves and overcome host barriers, one of them being the firm adhesion of the micro-organism to host cells . Colonization is crucial to pathogenesis of bacteria, being the earliest stage during onset of the disease and the ability to adhere to host surfaces is by far the most vital step in the successful colonization by microbial pathogens , . The presence of adhesins are said to be essential to the first steps of bacterial pathogenicity . A distinct family of adhesins called the adhesive pili or fimbriae, encoded by adhesin-gene clusters and assembled by the chaperone-usher pathway are said to be ubiquitous in Gram-negative organisms . Some of the well known fimbriae present among ExPEC strains are Type I fimbriae (fim), P fimbriae or the pilus associated with pyelonephritis (pap), curli fibre (csg), S fimbriae or the sialic acid-specific fimbriae (sfa), F1C fimbriae (foc), afimbrial adhesin/Dr antigen-specific adhesin (afa/dra), heat-resistant agglutinin (hra) and others . Each of these adhesins recognizes a specific receptor although as a group they share common genomic organization, assembly and even quaternary structural traits .
The role of some of these adhesins during APEC infection is well recognized. It has been reported that bacterial colonization of the respiratory tract is mediated by fimbrial adhesins, and studies have shown that type 1 and P fimbriae are expressed in vivo by bacteria colonizing the lung, air sacs and internal organs of chickens during infection . Bacterial adhesion to pharyngeal epithelial cells of chickens was also found to be inhibited by the presence of D-mannose, demonstrating the role of type 1 fimbriae during the early stages of the disease . No expression of P fimbriae was seen in the chicken trachea, suggesting that they may be important only in later stages of infection . In another study, it was seen that 99% of E. coli isolated from diseased birds possessed the csgA gene responsible for curli biosynthesis . Furthermore curli fibres were found to be essential for the internalization of bacteria causing avian septicaemia as seen in vitro .
Although there is increased knowledge about fimbrial adhesins and other colonization factors including the role they might play during infection, further investigation is still required to completely elucidate the function of each adhesin and its specific role in pathogenesis. In addition, the availability of complete genome sequences for some ExPEC pathogens like UPEC strain UTI89 (O18:K1:H7) and APEC strain APEC_O1 (O1:K1:H7) has drawn our attention to the large increase in the number of genes with unknown function, possibly involved in adhesion and colonization, and whose contribution to virulence on the whole is presently merely hypothetical –. Determination of the function of these novel genes will lead to a deeper understanding of host-pathogen interactions during the initial stages of APEC infection.
Previously our laboratory applied Signature-tagged mutagenesis (STM) to APEC in a chicken systemic infection model which led to the successful identification of genes crucial to systemic infection in chickens, however no adhesins were identified –.
In the present study, we applied STM to APEC in a modified lung infection model , in order to identify novel genes functionally involved in the adhesion and colonization of the chicken lung during the infection process. The STM screen led, among others, to the identification of a mutant EA7F9 with a transposon insertion in a putative adhesin gene (yqi). In addition, a mutant with an insertion in a gene encoding the type 1 fimbriae regulatory protein was also identified. These two genes were the most obvious targets of the STM screen being adhesins, the primary structures in direct contact with host tissue during infection. Type 1 fimbriae are already well characterized for their role in ExPEC pathogenesis; however, yqi has until now never been described for its potential role during infection. Therefore, the new adhesin encoded by the gene yqi became the target of this study, and was further characterized for its role in the initiation of APEC infection. We describe for the first time, the identification and characterization of this novel adhesin in APEC, temporarily renamed ExPEC adhesin I (EA/I), and the essential role it plays in the initial colonization of the chicken lung during infection.
Attenuation of IMT5155 with Mutation in ExPEC Adhesin I (EA/I) Gene yqi
Among the many genes identified during the STM screen, was the ExPEC adhesin I (EA/I) gene, also annotated as yqi . In vitro competition assays of transposon-mutant EA7F9, with a mutation in gene yqi, versus wild type strain IMT5155 showed that there was no significant growth defect in vitro as determined by growth curves (Figure 1). An in vitro competition index of 1.2 was observed for EA7F9. On the other hand, the mutant was found to be attenuated in the chicken with an in vivo competition index of 0.6. This result confirmed the attenuation of mutant EA7F9 during STM screening.
Sequencing of the EA/I Gene Cluster Region in IMT5155
The putative yqi adhesin gene cluster was completely sequenced in APEC strain IMT5155. Genomic organization of the IMT5155 yqi adhesin gene cluster was as follows: the putative outer membrane usher preceded the periplasmic chaperone, followed by the adhesin gene. A conserved hypothetical protein gene precedes the usher gene which may code for the adhesin subunit protein although this has not yet been confirmed (Figure 2). This 4,975 bp region showed a hundred percent sequence identity with the UPEC and APEC yqi adhesin gene cluster in sequenced strains UTI89 (Acc. No: CP000243) and APEC_O1 (Acc. No: CP000468) . Comparing the sequence with 1096 bacterial genomes currently available in the public database, out of which 28 were Escherichia coli genomes, this adhesin gene cluster is only found among ExPEC strains including APEC, UPEC and NMEC and is not harboured by intestinal pathogenic E. coli like enteropathogenic E. coli (EPEC) or enterohaemorrhagic E. coli (EHEC) such as E. coli O157:H7 strains EDL933 and Sakai, or non pathogenic E. coli K-12 strains MG1655 and W3110.
EA/I Plays a Role in the Adhesion of APEC to Chicken Fibroblasts and Kidney Epithelial Cells In Vitro
In order to determine the effect of EA/I in APEC, adhesion assays were initially performed in vitro using chicken fibroblast cells. Strain EA7F9 with a disruption in the yqi gene via a transposon (STM generated mutant), and strain IMT5155Δyqi devoid of the ExPEC adhesin I gene were tested against the wild type pathogen IMT5155 and a negative control strain MG1655. Bacterial determination at two different time points revealed a reduction in the ability of both EA7F9 and IMT5155Δyqi to adhere to fibroblast cells up to about forty percent of the total adhesion by IMT5155 (Figure 3A). MG1655 as expected also showed decreased adhesion ability when compared with IMT5155. An average CFU was calculated from three independent wells, and results were reproducible between adhesion experiments. Statistical analysis of the CFU values showed a significant difference between wild type IMT5155 and mutants tested, with a p<0.005 and p<0.05 at 1.5 h and 3 h respectively (Figure 3A).
Differences between IMT5155 and IMT5155Δyqi were statistically significant with a p<0.005 at 1.5 h and p<0.05 at 3 h (A). Bacterial adhesion to polarized Madin Darby canine kidney (MDCK-1) cells 3 h after infection with an MOI = 100. The difference between IMT5155Δyqi and IMT5155Δyqi (pDSK602:yqi) was significant with a p<0.04 (B).
To further confirm the role of yqi in APEC in vitro, adhesion assays were carried out using polarized Madin-Darby Canine-Kidney (MDCK-1) epithelial cells. Strains IMT5155Δyqi and IMT5155Δyqi (pDSK602:yqi), a strain carrying a modified expression plasmid with the yqi gene, that is, the complemented mutant, were tested against IMT5155, the positive control, and IMT11327, a negative control strain of avian origin lacking the gene yqi. A reduction in colonization by IMT5155Δyqi was seen, up to twenty percent of the total adhesion by IMT5155, 3 h after infection (Figure 3B). The complemented strain IMT5155Δyqi (pDSK602:yqi) regained its ability to adhere to MDCK-1 cells by introduction of the yqi gene when compared with its deletion mutant IMT5155Δyqi with a significance of p<0.04. All results were reproducible in consecutive adhesion experiments.
Effect of EA/I in Colonization of the Chicken Lung In Vivo
Colonization of the chicken lung in vivo was studied by infecting 5-week old chickens intra-tracheally, and isolating bacteria from the lungs 24 h after infection. For this purpose, two different infection set ups were made use of, including the lung colonization model of infection and the systemic infection model. IMT5155 and IMT11327 served as positive and negative controls respectively. The results of the in vivo experiments confirmed the observations made in vitro in cell culture models described above. When chickens were infected with a dose of 106 CFU of IMT5155Δyqi in a model used to study colonization abilities of the chicken lung by various strains, a distinct reduction in re-isolated bacterial numbers was observed as compared to IMT5155 as depicted in figure 4A. Differences between strains were statistically significant with a p<0.05. The average lung score in chickens infected with IMT5155Δyqi was 1.63 in comparison to 1.71 and 1.21 in chickens infected with IMT5155 and IMT11327 respectively. All chickens infected did not exhibit any clinical symptoms.
Differences between IMT5155 and IMT5155Δyqi were statistically significant with a p<0.05 (n = 6). Strain IMT11327 is the negative control (A). Bacterial colonization of the chicken lungs 24 h after intra-tracheal infection with 109 CFU of bacteria. Differences between IMT5155 and IMT5155Δyqi were statistically significant with a p<0.02 (n = 6). Strain IMT11327 is the negative control (B).
Moreover, when chickens were infected with a higher dose of bacteria, that is, 109 CFU, in a model designed to induce systemic infection, there still was a significant difference in the colonization of the lung by strains IMT5155 and IMT5155Δyqi as seen in figure 4B with a p<0.02 . The average lung score in chickens infected with IMT5155Δyqi was found to be 1.52 in comparison to 2.4 and 0.6 in chickens infected with IMT5155 and IMT11327 respectively. In both infection models, the ability of IMT11327 to colonize the chicken lung was much less than both IMT5155 and IMT5155Δyqi as seen in figure 4.
Effect of EA/I during Systemic Infection in Chickens
Systemic infection can be induced in chickens by infecting 5-week old birds intra-tracheally with a pathogenic strain with an infection dose of 109 CFU. As mentioned before, infection of chickens with this infection dose resulted in significant reduction in bacterial colonization of the chicken lung. When bacteria were re-isolated from chicken internal organs including spleen, kidneys, heart, liver and brain, there was a reduction in bacterial numbers when infected with IMT5155Δyqi as compared to IMT5155 as seen in figure 5, however only marginal. The only exception was the brain, where almost no bacteria were isolated when infected with IMT5155Δyqi as compared to IMT5155. All organ scores were considerably reduced in chickens infected with IMT5155Δyqi as compared to chickens infected with IMT5155 as seen in table 1.
Absence of columns indicates that no bacteria were isolated from the organ.
Electron Microscopy Reveals Fimbrial-Like Appendages Associated with ExPEC Adhesin I (yqi) Gene Cluster
The ExPEC adhesin I (yqi) 4,975 bp gene cluster coding for the putative subunit, chaperone, usher and adhesin was successfully cloned and over-expressed in an afimbriate strain E. coli AAEC189. Negative staining of strain AAEC189 (pKESK:yqi_4975_XB) revealed the expression of short fimbrial like appendages forming on the outer membrane of the bacterial cell (Figure 6A–C). These appendages were about 0.04 µm long and 0.005 µm thick and were not detected in the afimbriate strain AAEC189, that is, the negative control (Figure 6D). The wild type strain IMT5155, harbouring other adhesins like Type 1 fimbriae and curli, in addition to ExPEC adhesin I, was used as a positive control for the staining method, and long fimbriae with a length of about 0.5 µm (Figure 6E) were observed which were morphologically different from the fimbrial structures observed in strain AAEC189 (pKESK:yqi_4975_XB).
Electron micrographs show negatively stained afimbriate strain E. coli AAEC189 (pKESK:yqi_4975_XB) over-expressed with the yqi adhesin gene cluster at a magnification of 45,000x, 65,000x and 100,000x (A–C), negative control afimbriate strain E. coli AAEC189 (D) and wild type fimbriated E. coli strain IMT5155 (E). The arrows indicate the location of the fimbriae.
Prevalence of EA/I among ExPEC Reveals Its Potential as a Factor of Virulence
A collection of ExPEC strains, intestinal pathogenic E. coli, and avian and human non pathogenic strains available at the Institute of Microbiology and Epizootics, Freie Universität Berlin, was screened to determine the presence of the yqi gene among these strains. Out of a total of 588 ExPEC isolates tested for the presence of yqi, including 406 APEC, 138 UPEC, 25 NMEC and 19 Septicaemia associated E. coli (SePEC), 368 isolates were found to be positive for yqi, which amounts to a total of 62.5% of pathogenic isolates harbouring the yqi gene (Table 2). Among APEC isolates alone, 221 isolates were found to be positive making it a total of 54.4% positive for yqi in this group. Among UPEC isolates 91 isolates were found to be positive in a total percentage of 65.9% positive for yqi and among NMEC isolates 15 were found to be positive amounting to 60.0% positive for yqi. Finally, among SePEC isolates, 10 were positive in a percentage of 52.6%. From a total of 159 non pathogenic isolates tested, including those from avian and human origin, 31 were found to be positive for yqi, which accounts for only 19.4% of the non pathogenic strains tested.
Results were compared with MLST/EcoR data deposited in the publicly available database (www.mlst.net), of which 292 of the pathogenic isolates tested were known for their sequence type (ST) and 607 were known for their EcoR groups according to Herzer et al , which was carried out as part of a separate study at the Institute of Microbiology and Epizootics. Of the isolates positive for yqi, 68 isolates belonged to the EcoR group A (18.9%), 4 to B1 (1.1%), 255 to B2 (70.8%) and 33 to D (9.1%) making B2 the predominant group for isolates harbouring yqi. Among isolates negative for yqi, 122 isolates belonged to EcoR group A (49.3%), 29 to B1 (11.7%), 42 to B2 (17.0%) and 54 to D (21.8%) where A is the predominant group for isolates lacking yqi.
Out of the 31 non pathogenic isolates found to be positive for yqi, 24 (77.4%) were found to belong to EcoR group B2. Additionally 17 of the non pathogenic strains positive for yqi were isolated from the intestinal tract of healthy humans. These strains were positive for K1 capsule antigen and also belonged to EcoR group B2.
When comparing the distribution of the yqi gene among ExPEC strains with available MLST data, it was observed that strains positive for yqi are mostly allotted to sequence types 12, 68, 73, 95, 104, 140, 141, 349, 355, 358, 372, 390, 420, 913. All strains belonging to sequence types 141, 372 and sequence type complex 95 were positive for yqi.
Among 153 intestinal pathogenic E. coli tested in this study, none of the isolates were found to be positive for yqi as seen in table 2.
Due to the exclusive association of the putative fimbrial adhesin gene yqi with extraintestinal pathogenic E. coli (ExPEC), that is, its high prevalence among ExPEC isolates compared with non pathogenic E. coli, and complete absence in intestinal pathogenic E. coli, the adhesin was temporarily designated ExPEC adhesin I (EA/I).
EA/I Is Evolving under Positive Selection
The ExPEC adhesin I gene (yqi) was sequenced in many strains and compared with existing multilocus sequence typing (MLST) data. MLST is a nucleotide sequence based approach for the unambiguous characterization of isolates of bacteria using sequences of internal fragments of seven house-keeping genes . The population structure of microbial species with intermediate levels of recombination can thus be revealed by allele-based analyses . Wirth et al. described that sequence polymorphisms could define unique sequences for each of the seven house-keeping gene loci, which are referred to as alleles and each unique combination of alleles is assigned a sequence type (ST) number . Related STs are assigned to so-called ST complexes, using the principles of the eBurst algorithm : each ST complex includes at least three STs that differ from their nearest neighbour by no more than two of the seven loci while ST complexes differ from each other by three or more loci, and STs that did not match the criteria for inclusion within an ST complex are simply referred to by their ST designation .
The evolutionary history of the yqi gene was inferred using the Maximum Parsimony method. Sequencing of the ExPEC adhesin I (yqi) gene in strains belonging to the various sequence types (ST) and sequence type complexes (STC) including STC10, STC12, STC73, STC95, ST141 and ST372, and calculation of the distances of the adhesin gene locus between these strains revealed distribution of the strains tested into two groups (Figure 7). The value of replicate trees in which the associated taxa clustered together in the bootstrap test was 99% for the two groups identified. One group included strains belonging to the ST73 complex, while all other strains were assigned to a second group. Sequence homology was observed among strains belonging to a particular sequence type or sequence type complex as is the case with ST141, ST372, STC12 and STC73 within the sequence type or sequence type complex (Figure 7). One exception to the rule was strain IMT15008, ST73, STC73 which showed variations in its yqi gene sequence compared to other strains in the ST73 complex, and could be better assigned to the group harbouring other strains tested. Interestingly, only strains belonging to sequence type complex 95 showed mutations in the yqi gene sequence. Therefore, we determined the ratio of the non-synonymous mutation rate (Dn) to the synonymous mutation rate (Ds) within this complex using DnaSP 4.50.3 software. Dn/Ds < or = 1 indicates purifying or neutral selection, favouring amino acid substitutions . Our data show a non-synonymous mutation rate of 0.00095496 and a synonymous mutation rate of 0.00024 resulting in a Dn/Ds ratio of 3.979, indicating strong positive selection for structural evolution of the adhesin gene yqi within the sequence type 95 complex.
A necessary step in the successful colonization and progression of disease by microbial pathogens is the ability to adhere to host surfaces . The present study made use of a lung colonization chicken infection model in order to identify APEC genes specific to early stages of infection including adhesion to and colonization of the host by its pathogen via signature-tagged mutagenesis. As anticipated, we identified fimbrial proteins including the type 1 fimbrial regulatory protein, and other novel genes explicitly involved in the early stages of APEC pathogenesis as determined by competition assays with mutants. Our most interesting finding was a novel yqi adhesin gene which was seen to play a very significant role in the colonization of the avian lung by APEC. The newly identified putative fimbrial adhesin encoded by gene yqi, will be temporarily called ExPEC adhesin I (EA/I) until the specific host receptor for this adhesin has been identified in order to allow better classification and nomenclature of the novel E. coli fimbrial adhesin.
Adhesins are known to facilitate host colonization by mediating the first and crucial interaction with host tissue . In certain cases pathogens have been reported to have a substantial amount of different adhesins expressed at one time or another . In E. coli, two of the best characterized adhesins are the type 1 fimbriae and the P fimbriae which are regulated by the fim and pap operons . Genome sequencing of prototypic cystitis UPEC strain UTI89 has now revealed ten different chaperone-usher adhesin systems, among which the auf, yad, yfc, yqi, yeh and fml operons still remain to be characterized .
Interestingly, we identified the UPEC yqi adhesin gene in an STM screen specially designed to detect factors that would play a crucial role in the initiation of APEC infection. As stated before, the APEC yqi gene showed complete sequence identity with the UPEC yqi and is part of a 4,975 bp adhesin gene cluster yet uncharacterized. When comparing this adhesin gene sequence with publicly available E. coli genomes, the adhesin was found to be present among ExPEC strains including APEC and UPEC ,  and is not harboured by non-pathogenic E. coli nor by any of the intestinal pathogenic E. coli like enteropathogenic E. coli (EPEC) or enterohaemorrhagic E. coli (EHEC), making it unique to extraintestinal infections. This result therefore validates the temporary name ExPEC adhesin I (EA/I) allotted to the new adhesin.
Fimbrial adhesins share common genetic organization, in that the adhesin regulatory genes precede the major subunit gene, which is followed by the periplasmic chaperone, outer membrane usher, and finally the adhesin genes . This gene cluster organization is seen in type 1 fimbriae (fim), S fimbriae (sfa), F1C fimbriae (foc) and the Dr-antigen recognizing fimbriae (dra) . Organization of the yqi adhesin gene cluster differs by having the positions of the usher and chaperone inverted, which is also true for the pap adhesin gene cluster of the P fimbriae , although the reason for this, be it functional or biological, is still unclear.
To date, there are no studies implying the role of ExPEC adhesin I (yqi) in the colonization of host tissues during infection. Previously, it has been shown that the fimH adhesin of the type 1 fimbriae, and the papG adhesin of the P fimbriae are the receptor specific binding proteins that are located at the tip of the pilus structure . It has also been reported that deletion of the papG gene had no effect on pilus formation; however, the pili isolated from a papG deletion strain were not adhesive . We now have evidence which shows that deletion of the EA/I gene, yqi, in APEC strain IMT5155 results in a significant decrease in the colonization of fibroblasts and epithelial cells in vitro and of the chicken lung in vivo. It is possible that the Yqi adhesin is the receptor-specific protein responsible for initial attachment to host cells; however, this receptor is yet to be identified.
APEC strain IMT5155 has been previously tested for the presence of important adhesins like fim, pap, crl, sfa, tsh, afa, dra, foc and others, and has been found to harbour only the type 1 fimbria (fim), curli (csg) and temperature-sensitive haemagglutinin (tsh) genes . The absence of P, F1C, S and Dr fimbriae in IMT5155, which are known for their role in UPEC pathogenesis , made it apparent that there might in fact be unidentified adhesins that participate crucially during initiation of infection. We observed that in both infection models used in this study, there was always a substantial reduction in the colonization of the chicken lung when infecting chickens with IMT5155Δyqi. Therefore, we believe that ExPEC adhesin I makes a significant contribution to total bacterial adhesion and colonization of the chicken lung by APEC strain IMT5155 during infection. We further observed that there was only a slight reduction in bacterial numbers isolated from internal organs of chickens when infected with IMT5155Δyqi in the systemic infection model. This is not unique, because IMT5155 harbours a number of known virulence genes like iron acquisition genes (chuA, fyuA, ireA, sitD), serum resistance genes and protectins (iss, ompA, traT, neuC) and invasion related genes (gimB, ibeA) which play a role in the initiation and prolongation of disease and which may on the whole contribute to systemic infection in chickens . Besides with an infection dose of 109 CFU as used in the systemic infection model, the chicken lung is overwhelmed with bacteria, and therefore, there is very little room left for specific adhesion to take place in contrast to the lung infection model. This could also be a reason for the larger numbers of bacteria in internal organs during systemic infection. However, we still observe a significant bacterial reduction in the lung even with a high infection dose, which suggest the existence of specific receptors for ExPEC adhesin I on the lung epithelium, which account for a significant amount of fimbrial receptors on the lung. Due to the alternative fimbriae still present in the knock-out mutant of ExPEC adhesin I, attachment to other receptors on the surface of the lung would still occur strongly, particularly in the systemic infection model with its high infection dose, thus in turn initiating a systemic infection in this model.
An interesting result was the prevalence of EA/I among a relatively large collection of strains, particularly its high incidence among ExPEC strains like APEC, UPEC, NMEC and septicaemia associated E. coli (SePEC) strains, infrequent occurrence among non-pathogenic strains and complete absence among intestinal pathogenic E. coli strains. It has been previously reported that distinctive strains of E. coli responsible for most cases of urinary tract infection, sepsis and newborn meningitis are derived predominantly from E. coli phylogenetic group B2 . Additionally, diverse studies show that virulent clonal groups are derived mainly from phylogenetic group B2 and to a lesser extent from group D , . Our observations show that 66.1% of the isolates positive for yqi belong to the phylogenetic group B2 in contrast to only 11.5% of isolates negative for yqi belonging to this group.
Furthermore, we found that all of the isolates belonging to sequence types 95, 140, 141 and 372 were positive for yqi. ST95 and ST140 belong to the sequence type 95 complex. It is known that well defined pathogens are associated with specific STs or ST complexes, for example, ST95 complex contains related pathogenic bacteria of serogroups O1, O2, and O18 that express the K1 polysaccharide, that is, the K1 isolates , . Of immense interest, therefore, was the evolutionary analysis of the adhesin gene sequences from strains belonging to different sequence types, which showed that within a particular sequence type complex (STC), the adhesin sequence was homologous which is the case for STC12 and STC73 as well as other sequence types like ST372, ST141 and ST358 with a single exception, strain IMT15008. Since this strain had no unique characteristic that differentiated it from other strains within the ST73 complex, a possible explanation could be that the yqi gene in strain IMT15008 is the result of a recombination event.
Interestingly, within the ST95 complex, the adhesin gene sequence showed the presence of single nucleotide polymorphisms (SNPs) which were confirmed as a positive selection on the gene within this complex. Mutations producing functional modifications are called pathogenicity-adaptive or pathoadaptive, and are often SNPs producing amino acid replacements in proteins essential for a pathogen's success and it has been shown that two major adhesins of extraintestinal pathogenic E.coli (ExPEC) – type 1 and P fimbrial adhesins – acquire structural SNPs at a rapid rate, and this adaptation constitutes a major factor in the pathoadaptive microevolution and genetic diversification of ExPEC clonal groups . It is possible, that ExPEC adhesin I also undergoes structural mutations or pathoadaptive mutations, particularly, among strains belonging to the ST95 complex, which further confirms the importance of this adhesin within this highly pathogenic complex. Not surprisingly therefore, is the fact that the wild type strain used in this study, IMT5155, also belongs to the ST95 complex and harbours the most number of SNPs in the gene coding for ExPEC adhesin I as can be seen from the branching pattern on the dendrogram.
It therefore makes sense to assume that ExPEC adhesin I could perhaps play a very specific role in ExPEC pathogenesis, especially during the early colonization steps of infection, although this still needs to be proven for UPEC and NMEC strains. However, an interesting addition to this story is the decrease in adhesion to kidney epithelial cells as shown by our in vitro adhesion experiments. This would indicate the importance of yqi in UPEC, which particularly colonize the bladder epithelium during urinary tract infection (UTI). Furthermore, it provides evidence for the zoonotic potential of APEC, a topic of great interest in the field of ExPEC.
The presence of a single virulence factor hardly ever makes a strain virulent, while a combination of virulence factors usually determines its ability to cause disease . Therefore, much like countless other genes that are classified into the class of virulence factors when exceedingly prevalent among pathogens, the yqi gene could also eventually be an addition to this category based on its regular presence in highly pathogenic ExPEC strains in contrast to non pathogenic strains. One must keep in mind, that the meager group of Afaecal strains and human non pathogenic strains that were found to be positive for yqi in this study are no classical non-pathogenic strains as seen in a previous study . Ewers et al. reported that a number of Afaecal E. coli strains have characteristics typical of human and animal ExPEC, and that some nonoutbreak strains are capable of causing systemic disease in immunocompetent 5-week old chickens, suggesting the avian intestine reservoir hypothesis . Furthermore it was reported that these strains pose a zoonotic risk because they could be transferred directly from birds to humans or serve as a genetic pool for ExPEC strains. With regard to the non pathogenic human isolates found to be positive for yqi, it is interesting to note that though these strains were isolated from the intestinal tract of healthy humans they express the K1 polysaccharide otherwise associated with highly pathogenic strains.
Therefore, taken together, the results of the ExPEC adhesin I (yqi) prevalence study in E. coli are in fact intriguing, since the search for novel virulence factors still goes on. Nevertheless, they do not imply that EA/I would be solely responsible for virulence of ExPEC strains, rather only contribute to pathogenesis on the whole.
An interesting result in this study was the successful expression of ExPEC adhesin I in vitro. We hypothesized that cloning of the putative adhesin gene cluster coding for the putative subunit protein, putative usher and chaperone proteins and putative adhesin, which together are believed to be responsible for expression of fimbrial structures, when successfully cloned in an afimbriate E. coli strain would enable the expression of EA/1 fimbriae visible under the electron microscope. Using negative staining and transmission electron microscopy, our hypothesis was indeed confirmed, in that, short fimbrial like structures were detected in the afimbriate strain overexpressed with the adhesin gene cluster, which were not observed in the afimbriate strain, or negative control. The wild type strain IMT5155 used as a positive control revealed long fimbrial structures representing all fimbrial adhesins harboured by the strain including ExPEC adhesin I which explains the difference in length of fimbrial structures observed. The short fimbrial structures observed in the afimbriate strain overexpressed with ExPEC adhesin I will have to be confirmed using specific antibody and immunogold staining in the future; however, this study provides preliminary evidence for the fimbrial structures of this newly identified fimbrial adhesin. ExPEC adhesin I is therefore confirmed to be a novel fimbrial adhesin, worthy of further studies, which will indeed enable a better understanding of the interactions between adhering ExPEC and their host cells.
We have identified an adhesin that plays a significant role in the initial stages of APEC infection. Previous studies with other adhesins, like the fimH adhesin, have shown that antibodies elicited against the adhesin can impede colonization, block infection and prevent disease, that is, prophylactic vaccination with adhesins can inhibit bacterial infections . In other studies, it was found that vaccination with Gal-Gal pili or the P fimbrial vaccines prevented pyelonephritis by piliated E. coli in a murine model and in monkeys , . Therefore, blocking initial stages of infection may be one of the most effective strategies to prevent bacterial infections. Furthermore, it would be interesting to identify the specific receptor to which the EA/I binds. Blocking of specific receptors could also be a means of preventing early infection stages. It is our aim to test ExPEC adhesin I for its vaccine potential in chickens making use of our well established infection models, in order to come one step closer to the final goal of any research in the field of infectious pathogens, namely the prevention of infection and disease.
Materials and Methods
All animal experiments were approved by the “Landesamt fuer Gesundheit und Soziales” (LAGeSo) (G 0220/06) and chickens were killed according to animal welfare norms (Reg. 0220/06).
Bacterial Strains, Plasmids and Growth Conditions
All E. coli strains and plasmids used are listed in table 3. The wild type APEC isolate IMT5155 (O2:K1:H5) was obtained from internal organs of a laying hen clinically diagnosed with systemic APEC infection during an outbreak in the northern part of Germany in the year 2000 , . The strain possesses a number of virulence-associated genes typical of ExPEC strains, among others, curli fibre gene (csgA), type 1 fimbriae (fimC), temperature-sensitive haemagglutinin (tsh), heme receptor gene (chuA), outer membrane protein (ompA), invasion of brain endothelium (ibeA), genetic island associated with newborn meningitis (gimB), colicin V plasmid (cvaC) and is classified into the B2 phylogenetic group, and multi locus sequence type 140 of the ST95 complex . An additional E. coli strain IMT11327 (Ont:H16, ST295, ST23 complex), isolated from the intestine of a clinically healthy chicken, was used as a non-virulent control in both in vivo and in vitro assays, while E. coli - K12 strain (MG1655) was used as a negative control in vitro. IMT5155 was used for all genetic manipulation studies. All isolates used in the prevalence studies are part of a strain collection in our laboratory obtained from various sources , , , .
Bacterial strains were routinely cultured at 37°C in Luria-Bertani (LB) broth and on LB agar plates with appropriate antibiotics when required in the following concentrations: Kanamycin (Kan), 50 µg/ml; Nalidixin (Nal), 30 µg/ml; Ampicillin (Amp), 50 µg/ml; Chloramphenicol (Cm), 30 µg/ml; Spectinomycin (Spec) 50 µg/ml. All strains were stored at −70°C in LB broth with 10% (v/v) glycerol until further use.
All oligonucleotide primers are listed in table 4. Unless otherwise specified, E. coli isolates, from which DNA was to be used as the template for PCR amplification, were grown overnight in Luria-Bertani broth at 37°C and DNA was released from whole organisms by boiling for 10 minutes. After centrifugation, 2 µl of the supernatant was taken as template DNA and added to a 25 µl reaction mixture containing 0.1 µl of each primer pair in a 10 pmol concentration, 0.3 µl 10 mM of the four deoxynucleoside triphosphates (Sigma-Aldrich Chemie GmbH, Munich Germany), 2.5 µl of 10x PCR buffer, 1 µl of 50 mM magnesium chloride and 1 unit of Taq-Polymerase (Rapidozym GmbH, Berlin Germany). The samples were subjected to 25 cycles of amplification in a thermal cycler (GeneAmp PCR system, Applied Biosystems, Darmstadt, Germany). The amplification products were analyzed by gel electrophoresis on a 1.5% agarose gel (Biodeal, Markkleeberg, Germany), stained with ethidium bromide and photographed on exposure to UV.
Establishment of a Lung Colonization Model of Infection
A modified lung colonization model of infection was established based on the existing systemic infection model, for the purpose of screening STM mutant pools for adhesion and colonization factors as described previously according to animal welfare norms . The lung infection model differed from the systemic infection model in that, the infection dose was lowered to 106 CFU such that no systemic infection was induced.
Briefly, five-week old white leghorn specific pathogen free (SPF) chickens (Lohmann Tierzucht GmbH, Cuxhaven, Germany) were used for infection purposes. Groups of 10 chickens were infected intra-tracheally with a 0.5 ml suspension containing 106, 105, 104 and 103 CFU of the virulent strain IMT5155 respectively. IMT11327 was used as a negative control during infection. Chickens were euthanized 24 h post infection, and a clinical score was determined which monitored the infection ranging from score 0 (no symptoms) to score 4 (severe symptoms). An organ lesion score was likewise determined with a minimum and maximum score of 1 and 5 in the lungs respectively depicting the severity of lesions in the lungs, and a score of 0 (no hyperplasia) or 1 (hyperplasia) in the spleen. Bacteria were isolated from the lungs and spleen as follows: Organ samples were weighed, suspended in phosphate-buffered saline (1 ml/g) and homogenized with an Ultra-Turrax apparatus. Serial dilutions were plated out onto LB agar plates which were then incubated at 37°C for 24 h. Colonies were then counted to determine the CFU per gram in each organ. An infection dose of 106 was made the infection dose of choice in subsequent infections, since the number of re-isolated bacteria with this infection dose was optimal for STM analyses (data not shown).
Transposon pUTmini-Tn5 km2 Mutagenesis and Identification of a Novel Adhesin
Insertion mutagenesis of APEC strain IMT5155 was performed randomly by using transposon pUTmini-Tn5km2 as previously described . A transposon mutant library of strain IMT5155 was generated in this study and screened as described previously by Li et al , in the chicken lung colonization model of infection in search for colonization and adhesion factors crucial to APEC infection. A mutant EA7F9, was selected in this screen and the disrupted gene yqi, encoding a putative adhesin, was identified using arbitrary PCR also described previously . This adhesin, temporarily renamed ExPEC adhesin I (EA/I) was further characterized for its role in APEC pathogenesis.
Functional Characterization of ExPEC Adhesin I (EA/I)
In vitro and in vivo competition assays.
In vitro and in vivo competition assays were performed by mixing cultures of mutant and wild type strains with an OD600 = 1 in a ratio of 1∶1. For in vitro assays, the bacterial mixture was incubated in LB broth for 4 h at 37°C and then plated onto LB plates with and without Kanamycin and Nalidixin. For in vivo assays, four chickens were infected with 106 CFU/ml of this bacterial mixture. At 24 h post infection chickens were euthanized, lungs homogenized and serial dilutions plated onto LB plates with and without Kanamycin and Nalidixin for selection of the mutant and wild type strain respectively. A competitive index (CI) was calculated by dividing the output ratio (CFU mutant: CFU wild type) by the input ratio (CFU mutant: CFU wild type) at 0 h and 4 h or 24 h for in vitro and in vivo assays respectively.
Sequencing of a 4,975 bp EA/1 gene cluster region in IMT5155.
In order to sequence the EA/I gene cluster (yqi) in IMT5155, the 4,975 bp region was amplified using primers IMT-P3045 and IMT-P3046 and the PCR product was cloned into pCR2.1 TOPO vector according to the standard TOPO cloning manual (Invitrogen GmbH, Karlsruhe, Germany). The cloned product was transformed into E. coli TOP10 by electroporation and plated out on LB with Kanamycin to select for positive clones. Colonies were tested for the presence of the yqi gene cluster (4,975 bp) using standard primers IMT-P1560 and IMT-P1561. The plasmid containing the insert was isolated from the host strain and sequenced commercially by LGC's AGOWA Genomics, Berlin, Germany.
Generation of an isogenic yqi mutant.
To generate a knock-out mutant of the fimbrial adhesin in IMT5155, the yqi gene was replaced with a Chloramphenicol resistance cassette using the lambda red recombinase system . The Chloramphenicol acetyl transferase (CAT) gene was amplified from plasmid pKD3 using PCR with primers IMT-P2510 and IMT-P2511 (Table 4), part of which have sequence homology to the yqi flanking region. The PCR product was purified on a 1% agarose gel and 2 µl of the sample was transformed by electroporation into IMT5155 containing the lambda red recombinase expression plasmid pKD46. After electroporation, samples were incubated at 28°C for 1 h in SOC broth and plated on LB agar with Chloramphenicol to select for positive clones (CAT). After overnight incubation at 37°C, transformants were selected and tested for loss of the yqi gene using PCR with primer pairs IMT-P718/IMT-P2558 and IMT-P719/IMT-P2559 (Table 4).
Complementation of yqi.
For complementation and over-expression studies the IMT5155 yqi gene (1050 bp) PCR product was cloned into pCR2.1 TOPO vector according to the standard TOPO cloning manual (Invitrogen GmbH, Karlsruhe, Germany) using primers IMT-P2910 and IMT-P2911 (Table 4) with restriction enzyme recognition sites EcoRI and HindIII, and transformed into E. coli TOP10 via electroporation. Positive colonies were selected on LB agar with Kanamycin and tested for the presence of yqi using standard primers IMT-P1560 and IMT-P1561 (Table 4). The TOPO vector with the IMT5155 yqi insert was commercially sequenced by LGC's AGOWA Genomics, Berlin, Germany to determine that the sequence was in frame. The pCR2.1 TOPO:yqi vector and expression vector pDSK602  were then digested with restriction enzymes EcoRI and HindIII for 1 h at 37°C, and ligated using T4 DNA ligase overnight at 4°C. Three microliters of the ligation mix were then electroporated into E. coli TOP10 and plated out on LB agar containing spectinomycin. Colonies were tested for the presence of yqi using PCR with primers IMT-P2910 and IMT-P2911 (Table 4). The modified plasmid pDSK602 with the yqi insert was isolated from E. coli TOP10 and transformed into IMT5155Δyqi to complement the gene deleted. In adhesion and colonization assays, the complemented strain was induced with IPTG in a final concentration of 100 mM for expression of the protein.
Adhesion assays in vitro with Chicken Fibroblast (CEC) and polarized Madin Darby Canine Kidney (MDCK-1) epithelial cell lines.
Chicken fibroblast (CEC-32) cells  were used between passages 6 and 9 and were seeded into 12-well microtitre plates at a density of ∼2×105 cells per well and incubated at 37°C in an atmosphere containing 5% CO2 without antibiotics prior to adhesion assays. Minimal essential cell culture medium (PanTM Biotech GmbH, Aidenbach, Germany) with 5% Foetal calf serum (FCS) (PanTM Biotech GmbH, Aidenbach, Germany) was used to grow cells. Monolayers were used after 6 days incubation.
Madin Darby Canine Kidney (MDCK-1)  cells were used between passages 1 and 5. Cells were grown in Dulbecco's modified eagle's medium (DMEM) (PanTM Biotech GmbH, Aidenbach, Germany) with 10% FCS and were incubated at 37°C in 5% CO2. Transwell filter units (Costar) (Sigma-Aldrich Chemie GmbH, Munich Germany) contained a 1.12 cm2 porous filter membrane (0.4-µm pores) that had been treated for tissue culture. Filter units were incubated in 12-well microtitre plates (Costar) and were placed in DMEM containing 10% FCS for 1 h, at 37°C before seeding. 40 µl of a trypsinized MDCK cell suspension were added to each Transwell unit. Monolayers were used after 4 days incubation at which time there were around ∼3×105 MDCK cells per filter.
For adhesion assays, bacteria were added to the appropriate wells in triplicate in medium without FCS at an MOI of 100, that is, ∼2–3×107 bacteria per well. Microtitre plates were centrifuged at 250 x g for 10 minutes, and then incubated for 1.5 h and 3 h for CEC cells, and 3 h for MDCK cells, after which the supernatant was discarded; cells were washed thrice with Phosphate buffered saline (PBS) and then plated out on LB agar to determine the number of adherent bacteria in each well.
Animal infection studies.
Animal experiments were also carried out to determine the colonization ability of strain IMT5155Δyqi in comparison to IMT5155. IMT11327 was used as the negative control. Briefly, in an assay involving the lung colonization chicken model, groups of 6 five-week old chickens each were infected intratracheally with a bacterial suspension containing 106 bacteria per ml. At 24 h post infection, chickens were euthanized and dissected. A clinical and organ score was recorded and the lungs and spleen were homogenized.
In the chicken systemic infection model, groups of 6 five-week old chickens were infected intratracheally with a bacterial suspension containing 109 bacteria per ml. At 24 h post infection chickens were euthanized, dissected and a clinical and organ score was recorded. The lungs, heart, liver, kidneys, spleen and brain were homogenized.
All homogenates were appropriately diluted and plated out on LB agar and LB agar with antibiotics where required, to determine the number of bacteria colonizing the chicken lung and number of bacteria in internal organs during systemic infection.
Over-expression of the 4,975 bp EA/I adhesin gene cluster in fimbrial negative E. coli strain AAEC189.
The 4,975 bp EA/1 adhesin gene cluster was amplified and the PCR product cloned into pCR2.1 vector according to the standard TOPO cloning manual (Invitrogen GmbH, Karlsruhe, Germany) using primers IMT-P3259 and IMT-P3260 with restriction enzyme cutting sites XbaI and BamHI, and transformed into electro-competent E. coli TOP10 cells via electroporation. Positive clones were selected on LB agar with Kanamycin and tested for the presence of the 4,975 bp gene cluster using standard primers IMT-P1560 and IMT-P1561.
The plasmid pCR2.1-TOPO:yqi_4975_XB and expression vector pKESK-22 were digested with restriction enzymes XbaI and BamHI, and ligated using T4 DNA ligase overnight at 4°C and finally electroporated into electro-competent E.coli TOP10 cells and plated out on LB agar containing spectinomycin. Colonies were tested for the presence of the 4,975 bp adhesin gene cluster using PCR with primers IMT-P3138 and IMT-P3139.
Plasmid pKESK:yqi_4975_XB was isolated from E. coli TOP10 and transformed into electro-competent E. coli AAEC189 (afimbriate E. coli strain) cells, to over-express the adhesin gene cluster.
E. coli strains AAEC189, AAEC189 (pKESK:yqi_4975_XB) and IMT5155 were grown in 5 ml Brain Heart infusion broth after inoculating with 150 µl of the respective overnight cultures. Strains AAEC189 and IMT5155 were used as negative and positive controls respectively. Strain AAEC189 (pKESK:yqi_4975_XB) was additionally given Kanamycin in its growth medium, and induced with 0.1 M IPTG after 30 minutes of growth. All three strains were grown to an OD600 of 2.5, allowing an induction time of 2.5 h for strain AAEC189 (pKESK:yqi_4975_XB). One millilitre of bacterial culture was centrifuged at 8000 x g and the bacterial pellet was washed thrice with 1x PBS and then resuspended in 500 µl 1x PBS. Twenty microlitres of bacterial suspension were applied to Formvar-coated 200- mesh copper grids and stained with 1% Uranyl acetate for 2 mins. Negatively stained preparations were examined on a Zeiss EM900 microscope.
Prevalence of the EA/1 gene yqi among ExPEC, intestinal pathogenic E. coli (IPEC) and commensals.
To determine the prevalence of yqi among E. coli strains, 406 avian pathogenic E. coli (APEC) strains, 138 uropathogenic E. coli (UPEC), 25 Newborn meningitic E. coli (NMEC), 19 Septicaemia associated E. coli (SePEC), 153 intestinal pathogenic E. coli (IPEC) and 159 non pathogenic strains, including faecal strains from clinically healthy chickens (Afaecal) and non pathogenic strains isolated from the intestinal tract of healthy humans, were tested for the presence of the yqi gene using standard PCR reactions with primers IMT-P2512 and IMT-P2513 by way of amplification of a 400 bp region of the adhesin gene as described under PCR analyses. IMT5155 was used as a positive control, and IMT11327 as a negative control for all PCR reactions. Results were observed as a single clear band on an agarose gel and recorded as positive and negative for all strains respectively.
Sequencing of the EA/1 gene yqi and evolutionary analysis.
A total of 77 strains representing multi locus sequence types (MLST) ST12, ST73, ST95, ST104, ST135, ST140, ST141, ST358, ST363, ST368, ST372, ST390, ST416, ST417 and ST421 were selected for sequencing of the yqi gene (1050 bp), from a previous study at the Institute of Microbiology and Epizootics involving MLST analysis, using primer pairs IMT-P3706/IMT-P3707 and IMT-P2512/IMT-P2559. The PCR products were sequenced commercially by LGC's AGOWA Genomics, Berlin, such that a double stranded sequence was obtained for each strain. Sequences were analysed using Kodon software available from Applied Maths. A phylogenetic tree showing distances between strains of different STs was calculated by a maximum parsimony algorithm using Kodon Software from Applied Maths. Bootstrap values were computed with 100 replicates. The adhesin gene sequence of the strains APEC_O1 (APEC) (Acc. No: CP000468), CFT073 (UPEC) (Acc. No: AE014075), UTI89 (UPEC) (Acc. No: CP000243) was obtained from the available nucleotide database, and compared with sequenced strains in this study. Rates of non-synonymous (Dn) and synonymous (Ds) mutation were calculated using DnaSP 4.50.3 software  for the sequenced adhesin gene among strains belonging to the ST95 complex in order to determine the Dn/Ds ratio for each locus as described previously .
All statistical analysis for in vivo animal experiments and in vitro cell culture experiments were carried out using the software SPSS (Statistical package for the social sciences), version 15.0 by carrying out the non-parametric Mann-Whitney U-Test and the students t-test at the 95% significance level (p<0.05).
We would like to thank Prof. Dr. Bernd Kaspers (LMU, Munich, Germany), Alexander Karlas (MPI-IB, Berlin, Germany) and Dr. Ulrich Dobrindt (University of Würzburg) for kindly providing chicken fibroblast (CEC-32) cell line, Madin-Darby canine-kidney (MDCK-1) cell line and afimbriate E. coli strain AAEC189 respectively. We would also like to thank Ludwig Gröbler for technical assistance. Special thanks to Dr. Wilfried Bleiβ, Department of Parasitology, Humboldt University, Berlin for the electron microscopy photographs.
Conceived and designed the experiments: EMA CE DG GL LHW. Performed the experiments: EMA. Analyzed the data: EMA CE TH. Contributed reagents/materials/analysis tools: RP. Wrote the paper: EMA. Initiated the study and project: RP LHW.
- 1. Janβen T, Schwarz C, Preikschat P, Voss M, Philipp HC, et al. (2001) Virulence-associated genes in avian pathogenic Escherichia coli (APEC) isolated from internal organs of poultry having died from colibacillosis. Int J Med Microbiol 291: 371–378.
- 2. Dho-Moulin M, Fairbrother JM (1999) Avian pathogenic Escherichia coli (APEC). Vet Res 30: 299–316.
- 3. Kariyawasam S, Scaccianoce JA, Nolan LK (2007) Common and specific genomic sequences of avian and human extraintestinal pathogenic Escherichia coli as determined by genomic subtractive hybridization. BMC Microbiol 7: 81.
- 4. Ewers C, Li G, Wilking H, Kiessling S, Alt K, et al. (2007) Avian pathogenic, uropathogenic, and newborn meningitis-causing Escherichia coli: How closely related are they? Int J Med Microbiol 297: 163–176.
- 5. Johnson JR, Gajewski A, Lesse AJ, Russo TA (2003) Extraintestinal pathogenic Escherichia coli as a cause of invasive nonurinary infections. J Clin Microbiol 41: 5798–5802.
- 6. Krogfelt KA (1991) Bacterial adhesion: genetics, biogenesis, and role in pathogenesis of fimbrial adhesins of Escherichia coli. Rev Infect Dis 13: 721–735.
- 7. Martindale J, Stroud D, Moxon ER, Tang CM (2000) Genetic analysis of Escherichia coli K1 gastrointestinal colonization. Mol Microbiol 37: 1293–1305.
- 8. Finlay BB, Falkow S (1997) Common themes in microbial pathogenicity revisited. Microbiol Mol Biol Rev 61: 136–169.
- 9. Fantinatti F, Silveira WD, Castro AF (1994) Characteristics associated with pathogenicity of avian septicaemic Escherichia coli strains. Vet Microbiol 41: 75–86.
- 10. Wright KJ, Hultgren SJ (2006) Sticky fibers and uropathogenesis: bacterial adhesins in the urinary tract. Future Microbiology 1: 75–87.
- 11. Pourbakhsh SA, Dho-Moulin M, Bree A, Desautels C, Martineau-Doize B, et al. (1997) Localization of the in vivo expression of P and F1 fimbriae in chickens experimentally inoculated with pathogenic Escherichia coli. Microb Pathog 22: 331–341.
- 12. Chanteloup NK, Dho-Moulin M, Esnault E, Bree A, Lafont JP (1991) Serological conservation and location of the adhesin of avian Escherichia coli type 1 fimbriae. Microb Pathog 10: 271–280.
- 13. Gophna U, Barlev M, Seijffers R, Oelschlager TA, Hacker J, et al. (2001) Curli fibers mediate internalization of Escherichia coli by eukaryotic cells. Infect Immun 69: 2659–2665.
- 14. Chen SL, Hung CS, Xu J, Reigstad CS, Magrini V, et al. (2006) Identification of genes subject to positive selection in uropathogenic strains of Escherichia coli: a comparative genomics approach. Proc Natl Acad Sci U S A 103: 5977–5982.
- 15. Johnson TJ, Kariyawasam S, Wannemuehler Y, Mangiamele P, Johnson SJ, et al. (2007) The genome sequence of avian pathogenic Escherichia coli strain O1:K1:H7 shares strong similarities with human extraintestinal pathogenic E. coli genomes. J Bacteriol 189: 3228–3236.
- 16. Wiles TJ, Kulesus RR, Mulvey MA (2008) Origins and virulence mechanisms of uropathogenic Escherichia coli. Exp Mol Pathol 85: 11–19.
- 17. Li G, Ewers C, Laturnus C, Diehl I, Alt K, et al. (2008) Characteriziation of a yjjQ mutant of avian pathogenic Escherichia coli (APEC). Microbiology 154: 1082–1093.
- 18. Hensel M, Shea JE, Gleeson C, Jones MD, Dalton E, et al. (1995) Simultaneous identification of bacterial virulence genes by negative selection. Science 269: 400–403.
- 19. Chiang SL, Mekalanos JJ, Holden DW (1999) In vivo genetic analysis of bacterial virulence. Annu Rev Microbiol 53: 129–154.
- 20. Li G, Laturnus C, Ewers C, Wieler LH (2005) Identification of genes required for avian Escherichia coli septicemia by signature-tagged mutagenesis. Infect Immun 73: 2818–2827.
- 21. Antao EM, Glodde S, Li G, Sharifi R, Homeier T, et al. (2008) The chicken as a natural model for extraintestinal infections caused by avian pathogenic Escherichia coli (APEC). Microb Pathog 45: 361–369.
- 22. Herzer PJ, Inouye S, Inouye M, Whittam TS (1990) Phylogenetic distribution of branched RNA-linked multicopy single-stranded DNA among natural isolates of Escherichia coli. J Bacteriol 172: 6175–6181.
- 23. Maiden MC, Bygraves JA, Feil E, Morelli G, Russell JE, et al. (1998) Multilocus sequence typing: a portable approach to the identification of clones within populations of pathogenic microorganisms. Proc Natl Acad Sci U S A 95: 3140–3145.
- 24. Wirth T, Falush D, Lan R, Colles F, Mensa P, et al. (2006) Sex and virulence in Escherichia coli: an evolutionary perspective. Mol Microbiol 60: 1136–1151.
- 25. Feil EJ, Li BC, Aanensen DM, Hanage WP, Spratt BG (2004) eBURST: inferring patterns of evolutionary descent among clusters of related bacterial genotypes from multilocus sequence typing data. J Bacteriol 186: 1518–1530.
- 26. Weissman SJ, Chattopadhyay S, Aprikian P, Obata-Yasuoka M, Yarova-Yarovaya Y, et al. (2006) Clonal analysis reveals high rate of structural mutations in fimbrial adhesins of extraintestinal pathogenic Escherichia coli. Mol Microbiol 59: 975–988.
- 27. Hanson MS, Brinton CC Jr (1988) Identification and characterization of E. coli type-1 pilus tip adhesion protein. Nature 332: 265–268.
- 28. Krogfelt KA, Bergmans H, Klemm P (1990) Direct evidence that the FimH protein is the mannose-specific adhesin of Escherichia coli type 1 fimbriae. Infect Immun 58: 1995–1998.
- 29. Jones CH, Jacob-Dubuisson F, Dodson K, Kuehn M, Slonim L, et al. (1992) Adhesin presentation in bacteria requires molecular chaperones and ushers. Infect Immun 60: 4445–4451.
- 30. Johnson JR, Russo TA (2002) Uropathogenic Escherichia coli as agents of diverse non-urinary tract extraintestinal infections. J Infect Dis 186: 859–864.
- 31. Johnson JR, Russo TA (2005) Molecular epidemiology of extraintestinal pathogenic (uropathogenic) Escherichia coli. Int J Med Microbiol 295: 383–404.
- 32. Mokady D, Gophna U, Ron EZ (2005) Virulence factors of septicemic Escherichia coli strains. Int J Med Microbiol 295: 455–462.
- 33. Ewers C, Antao EM, Diehl I, Philipp HC, Wieler LH (2009) Intestine and environment of the chicken as reservoirs for extraintestinal pathogenic Escherichia coli strains with zoonotic potential. Appl Environ Microbiol 75: 184–192.
- 34. Wizemann TM, Adamou JE, Langermann S (1999) Adhesins as targets for vaccine development. Emerg Infect Dis 5: 395–403.
- 35. Pecha B, Low D, O'Hanley P (1989) Gal-Gal pili vaccines prevent pyelonephritis by piliated Escherichia coli in a murine model. Single-component Gal-Gal pili vaccines prevent pyelonephritis by homologous and heterologous piliated E. coli strains. J Clin Invest 83: 2102–2108.
- 36. Roberts JA, Hardaway K, Kaack B, Fussell EN, Baskin G (1984) Prevention of pyelonephritis by immunization with P-fimbriae. J Urol 131: 602–607.
- 37. Mordhorst IL, Claus H, Ewers C, Lappann M, Schoen C, et al. (2009) O-acetyltransferase gene neuO is segregated according to phylogenetic background and contributes to environmental desiccation resistance in Escherichia coli K1. Environ Microbiol.
- 38. Sankar TS, Neelakanta G, Sangal V, Plum G, Achtman M, et al. (2009) Fate of the H-NS-repressed bgl operon in evolution of Escherichia coli. PLoS Genet 5: e1000405.
- 39. Datsenko KA, Wanner BL (2000) One-step inactivation of chromosomal genes in Escherichia coli K-12 using PCR products. Proc Natl Acad Sci U S A 97: 6640–6645.
- 40. Murillo J, Shen H, Gerhold D, Sharma A, Cooksey DA, et al. (1994) Characterization of pPT23B, the plasmid involved in syringolide production by Pseudomonas syringae pv. tomato PT23. Plasmid 31: 275–287.
- 41. Kaaden OR, Lange S, Stiburek B (1982) Establishment and characterization of chicken embryo fibroblast clone LSCC-H32. In Vitro 18: 827–834.
- 42. Balcarova-Stander J, Pfeiffer SE, Fuller SD, Simons K (1984) Development of cell surface polarity in the epithelial Madin-Darby canine kidney (MDCK) cell line. Embo J 3: 2687–2694.
- 43. Rozas J, Sanchez-DelBarrio JC, Messeguer X, Rozas R (2003) DnaSP, DNA polymorphism analyses by the coalescent and other methods. Bioinformatics 19: 2496–2497.
- 44. Blattner FR, Plunkett G 3rd, Bloch CA, Perna NT, Burland V, et al. (1997) The complete genome sequence of Escherichia coli K-12. Science 277: 1453–1474.
- 45. Blomfield IC, McClain MS, Eisenstein BI (1991) Type 1 fimbriae mutants of Escherichia coli K12: characterization of recognized afimbriate strains and construction of new fim deletion mutants. Mol Microbiol 5: 1439–1445.