Characterization of the flexible genome complement of the commensal Escherichia coli strain A0 34/86 (O83 : K24 : H31)

Jana Hejnova1,2, Ulrich Dobrindt3, Radka Nemcova4, Christophe Rusniok1, Alojz Bomba4, Lionel Frangeul5, Jörg Hacker3, Philippe Glaser1, Peter Sebo2 and Carmen Buchrieser1

1 Unité de Génomique des Microorganismes Pathogènes and CNRS URA 2171, Institut Pasteur, 28 Rue du Dr. Roux, 75724 Paris, France
2 Laboratory of Molecular Biology of Bacterial Pathogens, Institute of Microbiology of the Czech Academy of Sciences, Videnska 1083, 142 20 Prague 4, Czech Republic
3 Institut für Molekulare Infektionsbiologie, Universität Würzburg, Röntgenring 11, 97070 Würzburg, Germany
4 Research Institute of Veterinary Medicine, Hlinkova 1/A, 040 01 Kosice, Slovakia
5 Plate-Forme 4 – Intégration et Analyse Génomique, Institut Pasteur, 28 Rue du Dr. Roux, 75724 Paris, France

Correspondence
Carmen Buchrieser
cbuch{at}pasteur.fr


   ABSTRACT
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS AND DISCUSSION
REFERENCES
 
Colonization by the commensal Escherichia coli strain A0 34/86 (O83 : K24 : H31) has proved to be safe and efficient in the prophylaxis and treatment of nosocomial infections and diarrhoea of preterm and newborn infants in Czech paediatric clinics over the past three decades. In searching for traits contributing to this beneficial effect related to the gut colonization capacity of the strain, the authors have analysed its genome by DNA–DNA hybridization to E. coli K-12 (MG1655) genomic DNA arrays and to ‘Pathoarrays’, as well as by multiplex PCR, bacterial artificial chromosome (BAC) library cloning and shotgun sequencing. Four hundred and ten E. coli K-12 ORFs were absent from A0 34/86, while 72 out of 456 genes associated with pathogenicity islands of E. coli and Shigella were also detected in E. coli A0 34/86. Furthermore, extraintestinal pathogenic E. coli-related genes involved in iron uptake and adhesion were detected by multiplex PCR, and genes encoding the HlyA and cytotoxic necrotizing factor toxins, together with 21 genes of the uropathogenic E. coli 536 pathogenicity island II, were identified by analysis of 2304 shotgun and 1344 BAC clone sequences of A0 34/86 DNA. Multiple sequence comparisons identified 31 kb of DNA specific for E. coli A0 34/86; some of the genes carried by this DNA may prove to be implicated in the colonization capacity of the strain, enabling it to outcompete pathogens. Among 100 examined BAC clones roughly covering the A0 34/86 genome, one reproducibly conferred on the laboratory strain DH10B an enhanced capacity to persist in the intestine of newborn piglets. Sequencing revealed that this BAC clone carried gene clusters encoding gluconate and mannonate metabolism, adhesion (fim), invasion (ibe) and restriction/modification functions. Hence, the genome of this clinically safe and highly efficient colonizer strain appears to harbour many ‘virulence-associated’ genes. These results highlight the thin line between bacterial ‘virulence’ and ‘fitness' or ‘colonization’ factors, and question the definition of enterobacterial virulence factors.


Abbreviations: BAC, bacterial artificial chromosome; CNF, cytotoxic necrotizing factor; EHEC, enterohaemorrhagic Escherichia coli; ExPEC, extraintestinal pathogenic Escherichia coli; HPI, high-pathogenicity island; IPEC, intestinal pathogenic Escherichia coli; PAI, pathogenicity island; UPEC, uropathogenic Escherichia coli

The GenBank/EMBL/DDBJ accession number for the sequence of the completely sequenced BAC insert (C4/1) is AJ829704.

A list of non-detectable ORFs of E. coli strain MG1655 in strain Colinfant in the order in which the ORFs appear in the MG1655 chromosome, together with a list of detectable virulence-associated or PAI-localized genes of pathogenic E. coli may be found in Supplementary Table S1 with the online version of this paper at http://mic.sgmjournals.org. A list of genes of the flexible E. coli genome complement identified in E. coli strain A0 34/86 by the combination of DNA array hybridization, Multiplex PCR and partial genome sequencing may be found in Supplementary Table S2.


   INTRODUCTION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS AND DISCUSSION
REFERENCES
 
Escherichia coli is a versatile bacterial species with many ecotypes. While commensal E. coli strains are a typical component of human colonic flora, the pathogenic types of E. coli can cause both enteric and diarrhoeal disease, as well as extraintestinal infections of the urinary tract, neonatal sepsis and meningitis (Kaper & Hacker, 1999). The diversity of E. coli pathotypes is due to the presence of specific subsets of virulence-associated genes, which are considered to be largely absent from normal-flora E. coli strains (Kaper & Hacker, 1999). These virulence genes are usually carried by a variety of pathogenicity islands (PAIs), bacteriophages, plasmids and/or transposons (Nataro & Kaper, 1998; Ochman & Jones, 2000). Moreover, the presence of specific subsets of genomic islands encoding fitness factors, found frequently also in pathogenic variants, appears to account for the diversity of commensal strains (Hacker & Carniel, 2001). The evolutionary forces to incorporate the diverse genetic units into the genome are presumably caused by providing an advantage in replication and survival in a particular ecological niche. Thus, enterobacterial genomes in particular consist of a core genome, conserved among the members of a species, and a flexible gene pool conferring strain-, pathotype- or ecotype- specific characteristics which allow adaptation to special conditions, such as colonization of specific niches or pathogenicity (Lan & Reeves, 2000). Given the flexibility in gene content and the possibility of the transfer of genes among different E. coli eco- or pathotypes in the intestinal tract, it is important to understand the genetic basis of their differences and the evolution of virulence, commensalism, and colonization capacities.

Recently, detailed characterization of the genetic organization of pathogenic and commensal E. coli strains was initiated by successive publication of E. coli genome sequences from two non-pathogenic K-12 strains, MG1655 and W3110, two enterohaemorrhagic E. coli (EHEC) isolates, O157 Sakai and EDL933, and one uropathogenic E. coli (UPEC) isolate, CFT073 (Blattner et al., 1997; Hayashi et al., 2001; Perna et al., 2001; Welch et al., 2002). Early comparisons of these genomes have already revealed that horizontal gene transfer in E. coli is far more extensive than anticipated. Compared to E. coli MG1655, the genome of the EHEC strain EDL933 was found to contain 1387 additional genes encoded in strain-specific clusters of diverse sizes (Perna et al., 2001). Further comparisons with the genome of the UPEC strain CFT073 revealed that the three completely sequenced strains share only 2996 (39·2 %) of the identified 7638 E. coli genes (Welch, 2001), the rest belonging to the flexible gene pool moving horizontally along a highly conserved framework of the core E. coli genome (Ochman & Jones, 2000; Welch, 2001). Given this variation in chromosome size, different E. coli strains can contain well over a megabase of unique DNA, accounting for specific traits distinguishing the members of the species.

In this study, we have focused on the characterization of E. coli strain A0 34/86 of serotype O83 : K24 : H31, which was initially isolated from porcine faeces. This strain has been safely and effectively used in Czech paediatric clinics over the past three decades for prophylactic and therapeutic colonization of the intestine of several thousand premature and newborn infants at risk of nosocomial infection and diarrhoea (Lodinova et al., 1967, 1980; Lodinova-Zadnikova et al., 1998). Colonization with E. coli A0 34/86 significantly reduces the occurrence of gastrointestinal infection and death associated with nosocomial infection in preterm newborns, most likely by displacing intestinal pathogens in infected carriers and by assisting the re-establishment of a normal gut flora (Lodinova et al., 1980; Lodinova-Zadnikova et al., 1995, 1998). Colonization by E. coli A0 34/86 further results in important stimulation of local antibody formation in the gut and saliva of colonized infants (Lodinova-Zadnikova et al., 1991). Moreover, subjects colonized by this strain early after birth appear to be significantly less prone to repeated infections and development of allergies later in life (Lodinova-Zadnikova et al., 2003). E. coli strain A0 34/86 is a type 1 fimbriated strain which does not produce thermolabile or thermostable enterotoxins (Lodinova et al., 1980) but produces {alpha}-haemolysin (HlyA), which is considered a virulence factor of UPEC and EPEC.

Given its clinical safety and efficacy record, the A0 34/86 strain has been approved for routine use in the Czech and Slovak Republics as a live oral vaccine preparation for infants. Aiming at identifying factors accounting for its beneficial effects, in particular those conferring its high colonization capacity, and in view of the potential use of this strain as a live vector for oral delivery of mucosal vaccines, we report here the first analysis of its genome, as obtained by hybridization to DNA arrays, bacterial artificial chromosome (BAC) and shotgun library construction and sequencing, and in vivo screening for genes potentially accounting for intestinal colonization capacity.


   METHODS
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS AND DISCUSSION
REFERENCES
 
Bacterial strains and growth conditions.
E. coli DH10B (F mcrA {Delta}(mrr–hsdRMSmcrBC) f80dlacZ{Delta}M15 DlacX74 deoR recA1 endA1 araD139 {Delta}(ara, leu)7649 galU galK rspL nupG) was used for construction of the BAC library and E. coli XL2-Blue (recA1 endA1 gyrA96 thi-1 hsdR17 supE44 relA1 lac [F' proAB lacIqZ{Delta}M15 Tn10 (Tetr) Amy Camr]) was used for construction of the shotgun library, respectively. The E. coli A0 34/86 strain of serotype O83 : K24 : H31 is a natural porcine isolate initially deposited in the type-culture collection of the Institute of Hygiene and Epidemiology in Prague under the label A0 34/86. It was grown for the purpose of this work from the commercial and lyophilized live vaccine preparation ‘Colinfant Newborn’ supplied by Dyntec, Terezín, Czech Republic. The spontaneous nalidixic acid-resistant (NalR) mutant of A0 34/86 (ZKRL+) was isolated and kindly provided by M. Nováková. Unless indicated otherwise, bacteria were grown at 37 °C for 24 h on Luria–Bertani (LB) plates supplemented with the appropriate antibiotics.

Genome comparison by DNA–DNA hybridization.
Total genomic DNA of E. coli A0 34/86 was [33P]dATP-labelled and used to probe Panorama E. coli Gene arrays (Sigma-Genosys) and the ‘Pathoarray’, as described previously (Dobrindt et al., 2003). Hybridization experiments were repeated four times, using independently labelled DNA probes; the DNA from E. coli strain MG1655 served as positive control and that of the Staphylococcus aureus strain Wood 46 as negative control, respectively. Normalization of spot intensities and array data analysis were performed as described earlier (Dobrindt et al., 2003), and the ORFs were recorded as lacking/not detectable if the signal : noise ratio was below 1·0 in at least three of the four hybridization experiments.

Multiplex PCR.
The multiplex PCR assay for the presence of virulence-associated genes of extraintestinal pathogenic E. coli (ExPEC) was performed as previously described (Johnson & Stell, 2000).

PFGE.
Genomic DNA for PFGE analysis was prepared in agarose plugs, as previously described (Buchrieser et al., 1994), cleaved by I-CeuI (New England Biolabs) and separated on a CHEF-Dr III system (Bio-Rad) at 12 °C in 0·5x TBE buffer, at 6·5 kV cm–2 and with pulse times increasing from 5 to 50 s over a period of 22 h.

BAC and shotgun library construction.
The BAC library was prepared by cloning HindIII partially digested and size-separated genomic DNA of E. coli A0 34/86 into the pBeloBAC11 vector (CmR), as previously described (Buchrieser et al., 1999). The size distribution of inserts ranged between 40 and 100 kb, with a mean size of 70 kb, as judged from PFGE analysis of a representative sample of randomly picked BAC plasmids digested by NotI. Shotgun libraries of chromosomal and/or BAC C4/1 DNA sheared to fragments of 1–3 kb by nebulization were constructed in the pcDNA2.1 vector (Invitrogen), as previously described (Buchrieser et al., 1999).

DNA sequencing and sequence analysis.
Sequencing reactions were performed on purified plasmid DNA using the Taq BigDye Terminator cycle sequencing kit (Applied Biosystems), the ABI PRISM 377 automatic sequencer and the 3700 capillary DNA sequencer (Applied Biosystems). Sequencing reads were assembled using the PHRED (Ewing & Green, 1998) and PHRAP software (P. Green, unpublished data). Editing was performed with CONSED (Gordon et al., 1998) and annotation was performed using the ARTEMIS (Rutherford et al., 2000) and CAAT-box (Frangeul et al., 2004) software tools. BLAST (Altschul et al., 1997) and FASTA (Pearson, 1990; Pearson & Lipman, 1988) sequence similarity searches were performed using public databases (NCBI). The end sequences of BAC library clones were compared to the genome of E. coli K-12 using the Colibri server at the Institut Pasteur (http://genolist.pasteur.fr/Colibri/). The sequence of the completely sequenced BAC insert (C4/1) is available under the GenBank/EMBL accession number AJ829704. The contigs assembled from the shotgun and BAC sequences are available from our website at http://www.pasteur.fr/gmp/sitegmp/Escherichia_coli_A034-86.

Gut persistence assay in newborn piglets.
Piglets (three to four per group) were orally inoculated immediately after birth and prior to colostrum ingestion by depositing 2 ml of bacterial suspension, in a viscous slurry with dried whey, at the root of the piglet tongue using a syringe without a needle. The piglets were then placed with the sow and reared conventionally for the duration of the experiment. Samples of piglet colon content were withdrawn from individual animals at indicated times, diluted in a 10-fold series in PBS and plated on appropriate selective agar media for determination of bacterial counts (c.f.u.) per gram of faeces.

Persistence of the E. coli A0 34/86 (ZKRL+) strain in the piglet gut was assessed upon single-dose inoculation by 109 c.f.u. On specified days the counts of {alpha}-haemolytic and nalidixic acid-resistant (HlyA+ and NalR) bacteria in colon content samples were determined by plating on LB agar supplemented with 5 % sheep blood and nalidixic acid (100 µg ml–1). It was verified that prior to piglet colonization the flora of the sow was free of HlyA+ and NalR bacteria and that the counts of HlyA+ and NalR on LB plates correlated well with total counts of NalR E. coli from simultaneous platings on MacConkey agar containing nalidixic acid (100 µg ml–1). The identity of the HlyA+ and NalR bacteria was further verified by PCR analysis of randomly picked colonies using the primers specific for the hlyA and hlyB intergenic region of E. coli (5'-CTAATGCGGGCAGAGAAATAAAGT-3' and 5'-ATAAAGACGGCAGGGTAACACAC-3'). Screening for BAC clones which conferred enhanced persistence in the piglet gut to the K-12 laboratory E. coli strain DH10B was performed by inoculation of piglets with three different pools of BAC clones, designated A, B and A+B. Pool A contained an equivalent mix of 80 BAC clones, which were overlapping with respect to the E. coli K-12 genome sequence; pool B contained 21 equivalently represented BAC clones with end sequences absent from the E. coli K-12 genome sequence and absent from pool A; pool A+B was an equivalent mixture of pools A and B. The pools of selected BAC clones were prepared by mixing equal volumes of overnight 2 ml cultures of selected BAC library clones grown individually in LB medium with chloramphenicol (12·5 µg ml–1). The pooled cultures were pelleted and stored frozen at –76 °C until use for animal inoculation. Cells of E. coli DH10B strain carrying the empty pBeloBAC11 vector were used as a negative control inoculum. To verify that representation of the individual BAC clones in the initial pools was unbiased and that the inocula were not dominated by one or more BAC clones, serial dilutions of pools A and B were plated on LB media with chloramphenicol (12·5 µg ml–1), and BAC DNA was extracted from 96 randomly picked colonies per pool. After digestion with NotI, the diversity of BAC clones in the pool was first assessed by PFGE analysis, and the identity of the individual BAC clones was verified by determining the terminal sequences of the inserts. Newborn piglets (three to four per group) were inoculated with 1011 c.f.u. of the indicated pools of BAC clones, as described above, and the representation of the individual BACs within the colon content of individual piglets was assessed at days 4 and 6 after inoculation. Reisolated BAC plasmids were analysed using PFGE and partial fragment sequencing, or PCR with primer pairs specific for selected BAC clones.


   RESULTS AND DISCUSSION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS AND DISCUSSION
REFERENCES
 
The E. coli species exhibits a remarkable genome plasticity and variability in genome content. In contrast to the characterization of pathogenic E. coli variants, detailed studies of the genome content of probiotic and/or commensal E. coli strains are rare, and the analysis of the molecular basis of their beneficial traits is essentially missing. In this study, we attempted to provide a first characterization of the genome of E. coli strain A0 34/86 (O83 : K24 : H31), which has been used in several Czech clinics over the past three decades for the prophylactic and therapeutic colonization of preterm newborn infants at risk of nosocomial infection. To identify the flexible gene pool complement that may encode specific traits improving adaptability and may account for the colonization properties of this strain, we characterized E. coli strain A0 34/86 using different comparative genomics approaches.

Analysis of the genome content of E. coli A0 34/86 by hybridization to DNA arrays
To obtain a first characterization of the genome of E. coli A0 34/86, we compared it to that of E. coli K-12 strain MG1655 by hybridization to Panorama E. coli gene arrays and to the E. coli Pathoarray’. The latter contains probes specific for the typical virulence-associated genes of ExPEC, intestinal pathogenic E. coli (IPEC) and Shigella species. It also contains genes known to contribute to the fitness and adaptability of enteric bacteria and which may also be involved in the virulence of ExPEC and/or IPEC, or which are present on mobile genetic elements (Dobrindt et al., 2003). This analysis revealed that 410 ORFs of E. coli MG1655 (10 %) were not detectable in the A0 34/86 strain (Table 1, Fig. 1; see also Supplementary Table S1, available with the online version of this paper at http://mic.sgmjournals.org, which lists the 410 E. coli K-12 genes absent in A0 34/86 and their functions, where known). The majority (245) of the missing ORFs encode hypothetical, unclassified or unknown gene products. The E. coli A0 34/86 genome also exhibits diversity in ORFs representing mobile genetic elements, i.e. the prophages CP4-6, DLP12, e14, Rac, Qin, CP4-44, CPS-53, CPZ-55, CP4-57 and KpLEC, which are present in E. coli MG1655 but absent or variable in E. coli A0 34/86.


View this table:
[in this window]
[in a new window]
 
Table 1. Categories of E. coli K-12 (MG1655)-specific ORFs that were not detected in E. coli A0 34/86 strain by hybridization to DNA arrays

 


View larger version (11K):
[in this window]
[in a new window]
 
Fig. 1. Comparison of the genomes of E. coli MG1655 and A0 34/86 based on DNA array analysis. The individual chromosomes are displayed linearly as horizontal bars of equal length and missing/not detectable ORFs are marked by vertical lines between the two chromosomes. Positions of the ORFs refer to the location on the E. coli MG1655 chromosome, and the origin and terminus of chromosome replication, the respective positions of tRNA genes frequently used as chromosomal insertion sites of horizontally acquired DNA elements and the locations of the 10 prophages of strain MG1655 are indicated.

 
Hybridization to the Pathoarray revealed that 54 (25·4 %) of the 212 probes specific for ORFs located on the five PAIs (PAI I–V536) of UPEC 536 and 17 (17 %) of the 100 probes specific for ExPEC ‘virulence-associated’ genes present on the Pathoarray gave a positive signal with E. coli A0 34/86 genomic DNA (Table 2 and Supplementary Table S2, available with the online version of this paper at http://mic.sgmjournals.org; the latter gives a more detailed description of the genes shared by A034/86 and pathogenic E. coli, including BLAST probability values, gene accession numbers and the methods by which they were detected). These genes included determinants encoding different fimbrial adhesins (type 1, P and S fimbriae and putative F17-like fimbriae), the toxins {alpha}-haemolysin (hlyA) and cytotoxic necrotizing factor 1 (cnf1), the ibeA gene required for invasion of eukaryotic cells, different iron-uptake systems (yersiniabactin, salmochelin), as well as several ORFs encoding transposases or bacteriophage integrases. Furthermore, 3 out of 95 probes specific for virulence-associated or PAI-encoded genes of IPEC, such as the serine protease-encoding genes and a gene encoding the diffuse adherence fibrilla determinant (dafa) of the EHEC strains O157 Sakai and EDL933, gave significant hybridization signals with genomic DNA of the A0 34/86 strain. These results were further corroborated by multiplex PCR-based screening of typical ExPEC ‘virulence-associated’ genes (Table 2, Supplementary Table S2) (Johnson & Stell, 2000).


View this table:
[in this window]
[in a new window]
 
Table 2. Genes belonging to the flexible E. coli gene pool identified in E. coli strain A0 34/86 by DNA array hybridization, multiplex PCR and partial genome sequencing

Methods used for the detection of genes: B, sequencing of BAC library; S, sequencing of shotgun library; M, multiplex PCR assay; P, Pathoarray hybridization. Abbreviations: UPEC, uropathogenic E. coli; MENEC, meningitis E. coli; ExPEC, extraintestinal E. coli; EPEC, enteropathogenic E. coli. PAI association was according to Pathoarray hybridization probes.

 
DNA–DNA array hybridization and multiplex PCR assays allowed the detection of already-known sequenced genes. Given that 410 ORFs (~400 kb) of E. coli K-12 scored as missing from A0 34/86, and that the size of its genome was estimated by PFGE analysis of I-CeuI-digested chromosomal DNA at 4·8 Mb (data not shown), strain A0 34/86 could be expected to contain about 700 ‘non-K-12’ genes (~700 kb). We thus constructed a BAC library containing genomic DNA of E. coli A0 34/86 and used sample shotgun sequencing and sequencing of BAC insert termini to further characterize the gene content.

Characterization of E. coli A0 34/86 flexible genome complement by partial sequencing and mapping of BAC clones on the core genome of E. coli K-12
An exhaustive BAC library of the A0 34/86 genome was prepared and characterized by sequencing of BAC insert termini. A total of 3936 BAC clones of E. coli A0 34/86 chromosomal DNA in the pBeloBAC11 vector was obtained. The BAC insert sizes ranged between 40 and 100 kb, with a mean insert size of ~70 kb, as assessed from PFGE analysis of a representative sample of NotI-digested BAC DNA (data not shown). The BAC library obtained thus represented approximately 30–50-fold coverage of the A0 34/86 genome.

A subset of 341 randomly picked BAC clones corresponding to approximately fivefold coverage of the genome, with 7·1 % of the clones (25 BACs) being represented twice, was subjected to sequencing of insert termini. These DNA sequences were analysed for similarity at the protein and nucleotide levels using the GenBank database. Out of the 682 BAC insert termini sequences, 537 could be mapped to the E. coli K-12 (MG1655) genome sequence. Furthermore, 51 BAC end sequences exhibited homology to genes encoding hypothetical proteins of unknown function present in the genomes of UPEC CFT073 and/or EHEC O157 : H7 strains. As summarized in Table 2, 48 BAC end sequences exhibited a significant homology to known ‘virulence-associated’ genes, including genes encoding the K1 capsule and LPS synthesis systems and fimbrial adhesins (fimA–H, F17), toxins (hlyA, cnf) and microcin (mch), confirming the detection of these genes by the Pathoarray hybridization experiments (Table 2, Supplementary Table S2). In contrast, 27 (5 %) did not match any sequence in the public databases. These BAC clones hence appear to carry inserts containing genomic DNA specific to E. coli A0 34/86.

A minimal overlapping set of BAC clones already represents a powerful tool for comparative and functional genomics. The 341 BAC clones were thus positioned on the map of the E. coli K-12 chromosome (Blattner et al., 1997) and an ordered physical map was established (Fig. 2). Of these clones, 318 covered altogether ~94 % of the E. coli K-12 genome, while 23 of the BACs (7 %) could not be positioned on the map because there was no homology of their insert termini sequences to the K-12 genomic sequence. Moreover, the 70 BAC clones having a single insert terminal sequence matching the K-12 chromosome could be exploited to roughly locate at least some of the A0 34/86-specific (‘non-K-12’) DNA on the framework of the E. coli K-12 core genome. As further shown in Fig. 2, inserts of this A0 34/86-specific genetic material appear to be scattered around the core E. coli genome in at least 10 major regions.



View larger version (23K):
[in this window]
[in a new window]
 
Fig. 2. Location of the E. coli A0 34/86-derived BAC clones carrying ‘non-K-12’ DNA on the framework of the E. coli K-12 chromosome. The overlapping set of BACs (outer circle) was positioned on the chromosomal map of E. coli K-12 (inner circle). The numbered regions with estimated insert sizes (where applicable) correspond to: 1, sequences specific to E. coli A0 34/86, including the haemolysin operon; 2, sequences found in E. coli PAI II536 and E. coli CFT073; 3, sequences found in E. coli O157 : H7; 4, E. coli A0 34/86-specific sequences; 5, sequences found in E. coli CFT073 and comprising the gene for the Pil V-like protein; 6, sequences found in E. coli CFT073; 7, sequences found in E. coli CFT073; 8, sequences found in E. coli CFT073, including the gene for heptosyltransferase (waaT); 9, sequences found in E. coli PAI II536 comprising the putative ORF2 superfamily I DNA helicase; 10, BAC C4/1 clone insert comprising the GimA island of E. coli K1 and the O157 : H7 and CFT073 sequences. The positions of tRNA genes frequently used as chromosomal insertion sites of horizontally acquired DNA elements are marked inside the circle of the E. coli K-12 chromosome.

 
In parallel, a shotgun library with small inserts was constructed and partially sequenced, in order to identify additional A0 34/86-specific genes. Sequences of 1152 shotgun clones with inserts of 1–3 kb were determined and assembled with the 641 BAC clone sequences using the PHRED and PHRAP software. By this approach, a roughly 1·2 Mb sequence of the E. coli A0 34/86 genome was acquired and assembled into 618 contigs (635 kb) and 867 singlets (555 kb), representing about one-fourth of the genome.

To identify the E. coli A0 34/86-specific sequences with respect to the already known E. coli genome sequences, we masked the E. coli K-12, E. coli O157 : H7 and E. coli CFT073 sequences, using Cross-match (http:/bozeman.mbt.washington.edu/). When comparing the A0 34/86 sequences with those of the E. coli K-12 genome, a total of 227 kb (19 %) of DNA sequence (134 contigs and 186 single sequences) was identified as present in E. coli A0 34/86 but absent from E. coli K-12 (Table 3). Sixty-seven of these 227 kb sequences were present in E. coli O157 : H7, determining 160 kb (103 contigs and 142 single sequences) as sequences specific to A0 34/86 (Table 3). Further comparison of A0 34/86 sequences with the merged K-12 and E. coli CFT073 (UPEC) genomes, resulted in 81 kb (50 contigs and 64 single sequences) of A0 34/86-specific sequence (Table 3). Finally, about 64 kb of DNA sequence (38 contigs and 56 single sequences) scored as A0 34/86-specific when compared to all three completely sequenced genomes of E. coli (K-12, O157, CFT073), as given in Table 3. Furthermore, 32 kb out of the 64 kb of these A0 34/86-specific sequences were found by comparison to public databases to have been previously identified in other E. coli strains (Table 2, Supplementary Table S2). Altogether, 32 kb of the A0 34/86 DNA sequences distributed in 18 contigs (14 kb) and 29 single sequences (18 kb) did not match any sequences in public databases and thus are new E. coli sequences, some of which might code particular functions of this strain.


View this table:
[in this window]
[in a new window]
 
Table 3. Specific E. coli A0 34/86 sequences with regard to the genome sequences of the E. coli strains K-12 (MG1655), O157 : H7 (Sakai and EDL933) and UPEC (CFT073) and to sequences present in the GenBank database

The size of the contigs and singlets varied from 100 bp to 3·5 kb.

 
Presence of adhesion, iron-acquisition and bacteriocin systems may contribute to the colonizing capacity of E. coli A0 34/86
In addition to the genes detected by the Pathoarray, low-coverage sequencing further revealed the presence of ‘virulence-associated’ genes in the A0 34/86 genome. These could code for toxins (EspC), bacteriocins (colicin V and microcin H47 synthesis, and transport systems), protectins involved in synthesis of the capsule (K1 and K5), polysialic acid and LPS (waaO), adhesins (P and S fimbriae, F17-like fimbrial adhesin, F1C fimbriae type IV pili), iron-uptake systems (salmochelin, Hbp haemoglobin protease), as well as genes encoding hypothetical proteins detected earlier in the EHEC and UPEC genomes. Among the E. coli A0 34/86 genes identified here, many code for factors that potentially contribute to the colonizing capacity of the strain, such as the genes for adherence and iron-uptake systems.

Agglutination of the bacteria by latex particles containing the Pap receptor and by specific agglutination with an anti-PrfA serum showed that the pap pilus operon is expressed by E. coli A0 34/86, at least in vitro (Table 4). Moreover, agglutination with anti-FocA serum revealed that the F1C fimbriae are also expressed by the A0 34/86 strain (Table 4). It has been shown that P fimbriae, which bind to galactose-{alpha}-1,4-galactose molecules on uroepithelial cells (Baga et al., 1985), are implicated in increased persistence of E. coli in the colon (Wold et al., 1992). Thus they may be contributing to the intestine colonization capacity of the A0 34/86 strain.


View this table:
[in this window]
[in a new window]
 
Table 4. Qualitative assessment of the expression of various ‘virulence factors' detected in E. coli A0 34/86

 
Possession of multiple specialized iron-acquisition systems is a prerequisite for the successful multiplication of virulent and commensal bacteria in the host, and such systems thus likely contribute also to the gut colonization capacity of E. coli A0 34/86 strain. Multiple iron-uptake systems, such as enterobactin, aerobactin and salmochelin, and the fungal siderophores ferrichrome and coprogen, have indeed been found in other E. coli strains (Otto et al., 1998). Besides these systems, yet another system encoding yersiniabactin (Ybt) siderophore synthesis and uptake appears to be present in the A0 34/86 strain, as detected here by the sequences of several irp and ybt genes. The yersiniabactin system was first described as part of the so-called high-pathogenicity island (HPI) of Yersinia species (Carniel et al., 1996), and it was subsequently found also in a variety of E. coli strains, which have acquired this system by horizontal transfer mediated by a P4-like bacteriophage (Bach et al., 2000; Buchrieser et al., 1998; Dobrindt et al., 2002). We also identified the tsh or hbp gene that encodes Hbp, a serine protease cleaving haemoglobin (Otto et al., 1998). Hbp is a member of the IgA protease family of secreted autotransporters, which play diverse roles in bacterial ‘virulence’ (Dozois et al., 2000), and its activity against haemoglobin suggests that Hbp might be part of a specific iron-uptake system. Many bacteria, and pathogenic strains in particular, can indeed use haem compounds as a source of iron (Otto et al., 1998).

In a similar way to iron-acquisition systems, the capacity to produce bacteriocins may also contribute to the colonization capacity and competitiveness of E. coli strains. The mchEFXI genes encoding the microcin H47 secretion and immunity determinant (Rodriguez et al., 1999) are present in the A0 34/86 strain. Moreover, production of H47 microcin/colicin activity by strain A0 34/86 was detected in vitro when using an E. coli DH5{alpha} indicator strain (Table 4). Furthermore, the cvaA and cvaB genes required for secretion of the small proteinaceous colicin V were present in strain A0 34/86 (Gilson et al., 1990). Production of colicin and microcin may also contribute to the persistence and competitiveness of E. coli A0 34/86 in the gastrointestinal environment.

The E. coli strain A0 34/86 is further known to produce the RTX (repeat in toxin) {alpha}-haemolysin HlyA, and the activity of the cytotoxic necrotizing factor 1 (CNF1) in A0 34/86 cellular lysates was revealed by using a HeLa cell-based assay (Table 4). The presence and expression of these different adhesin and toxin genes, usually present in UPEC, makes A0 34/86 resemble a uropathogenic rather than a commensal isolate. Furthermore, the results of a triplex-PCR typing method (Clermont et al., 2000) support this finding. Commensal strains belong mainly to the phylogenetic lineage A of E. coli, whereas the A0 34/86 strain belongs, according to the triplex-PCR method, to the B2 group (data not shown), which comprises mainly extraintestinal pathogenic strains and, in particular, uropathogenic isolates.

In vivo selection for E. coli A0 34/86 gut persistence factors
The availability of an ordered BAC library opened the way to in vivo screening for genes which could be involved in the gut colonization capacity of the E. coli A0 34/86 strain and which could potentially confer an enhanced persistence in the gut on the BAC carrier strain DH10B (rough, derivative of K-12). Screening experiments were hence performed in the natural porcine host of the A0 34/86 strain using the model of bacterial survival in the gut of newborn piglets. To approximate the prophylactic colonization of newborn infants under the competitive pressure of maternal flora and of the nosocomial strains in the clinics, the animals were reared conventionally with the sow, following inoculation by the bacterial suspensions.

First, the colonization capacity of the E. coli A0 34/86 strain in newborn piglets was characterized, using a NalR variant of A0 34/86 (ZKRL+) that could be unambiguously traced in samples of piglet colon content by plating on blood agar and MacConkey plates containing nalidixic acid (100 µg ml–1). As documented by the typical result shown in Fig. 3(a), upon a single oral inoculation of newborn animals by 109 c.f.u. of E. coli A0 34/86 (NalR), the strain rapidly colonized the piglet gut, reaching ~1010 c.f.u. per gram colon content within 24 h, and persisted at >108 c.f.u. g–1 for a week after a single inoculation. Moreover, the A0 34/86 strain was still detectable in the piglet colon content (~104 c.f.u. g–1) 4 weeks after a single inoculation, despite the strong competitive pressure of the intestinal flora of the sow, which was uncontrollably ingested by the piglets with sow excrement over the duration of the experiment. It is also noteworthy that control piglets inoculated with buffer became rapidly contaminated and colonized by the A0 34/86 (NalR) strain to quite high levels, when reared together with the inoculated animals (Fig. 3a). These results confirmed the high colonization capacity of E. coli A0 34/86 in newborn piglets, and showed that this animal model can be used for in vivo selection of BAC clones contributing to the high capacity of the A0 34/86 strain to persist in the gut.



View larger version (76K):
[in this window]
[in a new window]
 
Fig. 3. In vivo selection of clones expressing potential colonization factors of A0 34/86 in the model of intestinal colonization of newborn piglets. (a) Colonization of newborn piglet colon by the E. coli A0 34/86(NalR) strain in animals (three to four per group) inoculated by 109 c.f.u. of the strain shortly after birth and prior to colostrum ingestion ({blacklozenge}). Control piglets that received only buffer ({blacktriangleup}) were reared with the same sow and also rapidly became colonized (cross-contaminated) by the A0 34/86(NalR) strain. {blacksquare}, total E. coli. Mean values per group of animals are given and the results are representative of three independent experiments. (b) Groups of three to four newborn piglets were inoculated prior to colostrum ingestion by 1011 c.f.u. of the BAC clone pools A, B and A+B, and the colon content of individual animals was sampled daily over an interval of a week to follow colonization and to reisolate the persisting BAC clones carried by the E. coli DH10B host. Mean values from two independent colonization experiments are given. {blacklozenge}, pool A; {bullet}, pool B; {blacktriangleup}, pools A+B. (c) One example of the PFGE analysis of BAC plasmids extracted from single colonies of BAC clones recovered from the colon content of piglets inoculated with BAC pools A and B. The clones were randomly picked from selection plates and the extracted BAC DNA was digested with NotI to estimate the diversity and size of their respective inserts. M, molecular mass marker LR PFG; L, molecular mass marker ‘Lambda Ladder’ PFG. The migration of the NotI fragment of BAC C4/1 is shown by the arrow to the right of the figure.

 
Thus, different pools of BAC clones were used to inoculate the animals for screening experiments. Pool A comprised an equivalent mixture of 80 overlapping BAC clones in E. coli DH10B, covering essentially the entire ‘K-12-like’ portion of the E. coli A0 34/86 genome (‘core genome’), while pool B consisted of an equivalent mixture of 21 BAC clones, covering mainly the ‘non-K-12’ DNA specific to the A0 34/86 genome. Prior to animal inoculation, it was verified by PCR and PFGE analysis that the composition of the BAC clone pools was not biased by overrepresentation of any of the selected BAC clones (data not shown). Groups of three to four newborn piglets per BAC pool were then inoculated by 1011 c.f.u. of the BAC clone pools propagated in E. coli DH10B, and animal colonization was followed over 6 days, as shown in Fig. 3(b). Samples of piglet colon content were plated on MacConkey agar containing 20 µg chloramphenicol ml–1, to select for E. coli clones carrying the BAC plasmids (CmR). BACs from 20 individual colonies isolated per animal on days 4 and 6 after inoculation were extracted. Diversity of the recovered BAC plasmids was assessed by PFGE analysis of NotI-digested DNA and by sequencing of their insert termini. As illustrated by one of the PFGE analysis results shown in Fig. 3(c), no specific enrichment (positive selection) could be observed for any of the individual BAC clones present in pool A upon passage through the piglet intestine, suggesting that pool A did not contain any clone that expressed factors significantly enhancing persistence of the DH10B carrier strain. By contrast, in animals inoculated with pool B, or even with the mixture of pools A and B, one BAC clone, labelled C4/1, was found to be repeatedly enriched with respect to other BAC clones on days 4 and 6 after inoculation (Fig. 3c). In the first in vivo experiment, BAC C4/1 was found 28 times among the 96 sequence-characterized BACs (29 %) isolated on days 4 or 6 from animals inoculated with pool B, while in the second experiment it was found 10 times (35 %) among the 28 sequence-characterized BACs recovered on day 4. Moreover, BAC C4/1 was also represented 13 times in the pool of 19 BACs recovered on day 6 from animals inoculated with a mixture of pools A and B. By contrast, all other BAC clones present in the inocula were reisolated only once or twice per sample of colon content of the colonized animals on days 4 or 6. As the individual BAC clones were equivalently represented in the inoculation pools, and no in vitro enrichment of any of the BAC clones could be observed upon three consecutive passages of the inoculated BAC pools in vitro without antibiotic selection (data not shown), it can be concluded that the BAC C4/1 clone was positively and specifically selected in the piglet gut, most likely because it carried E. coli A0 34/86 genes that conferred an enhanced capacity to persist in the piglet gut to the BAC carrier strain DH10B.

Complete sequence determination and analysis of BAC clone C4/1
To characterize the potential colonization factor genes of A0 34/86 carried by BAC C4/1, its entire sequence (55·5 kb) was determined. As illustrated in Fig. 4, the insert of BAC C4/1 carries several gene clusters that might be contributing to the intestinal persistence of E. coli. These comprise the fim gene cluster, which is present in both pathogenic and commensal E. coli strains and encodes production of type I fimbriae, which largely account for the adhesive properties of the family Enterobacteriaceae. Interestingly, the sequence of the tip adhesin, FimH, of A0 34/86 appeared to be identical to that of the highly adhesive pathogenic CFT073 strain, suggesting that the A0 34/86 fimbriae may be contributing to the enhanced retention on intestinal mucosa of the DH10B clone carrying the BAC C4/1.



View larger version (16K):
[in this window]
[in a new window]
 
Fig. 4. Genetic organization of the insert of BAC C4/1 (55·5 kb). The ORFs deduced from sequencing and annotation of the BAC insert are indicated below the corresponding ORF symbols. The red region indicates the operon encoding the fimbriae; the blue region represents the GimA genetic island first identified in an E. coli K1 neonatal meningitis (ECNM) strain (Huang et al., 2001); sequences from the region indicated in green are found in part also in K-12 isolates and constitute, together with the orange region, which harbours the restriction and modification system genes, a segment of approximately 20 kb that is found partly in EHEC O157 : H7 strains and is almost completely preserved in the UPEC CFT073 isolate.

 
Downstream of the fim operon, a gluconate and mannonate metabolism gene cluster (gntP, uxuA, uxuB and uxuR) was identified, and the central region of the BAC C4/1 insert contained a 20 kb genetic island, GimA, absent from E. coli K-12. The GimA island was previously identified in an E. coli K1 strain causing neonatal meningitis (ECNM) and carries 14 genes presumably involved in energy metabolism and carbon source regulation (Huang et al., 2001). It also contains the gene for the 50 kDa IbeA protein, which appears to play a key role in mediating the invasion of K1 E. coli into human brain microvascular endothelial cells and which may also mediate bacterial passage of the blood–brain barrier (Huang et al., 2001). The entire GimA ‘meningitis' island is present in the E. coli A0 34/86 genome adjacent to the fim operon, as in some ECNM strains (Huang et al., 1995). Interestingly, the GimA island was found on BAC C4/1 that was positively selected for colonization persistence in newborn piglets. However, the preliminary experiments indicate that deletion of the ibeA gene from the chromosome of A0 34/86 does not affect the capacity of the strain to colonize the porcine intestine (data not shown). A recent study by Bonacorsi et al. (2003), involving 132 neonatal meningitis E. coli isolates, indeed showed that only E. coli belonging to the B2 group contained the ibeA gene and the gimA operon, while a number of other isolates belonging to the A lineage of E. coli and lacking determinants such as portions of PAI III536 (sfa/foc and iro; 34 %), the gimA island (ibeA and ptnC; 38 %) and PAI II536 (hly, cnf1 and hra; 10 %), can also cause meningitis similar to that caused by the E. coli strains from the B2 lineage. Thus, additional, as yet unknown, factors appear to be involved in the pathophysiology of meningitis caused by E. coli, and the GimA and PAI island genes mentioned above appear not to be necessary for virulence. This would be in line with the observation that genes for such factors can also be found in the avirulent E. coli strain A0 34/86, which also belongs to the B2 lineage (Fig. 3).

The rest of the BAC C4/1 insert consisted of a 25 kb segment almost identical to a chromosomal region of the uropathogenic isolate CFT073, containing the hypothetical yji genes, the yjj osmotic adaptation genes and several chemotaxis and transporter protein-coding genes. These also appear to be partly present in the K-12 strain MG1655 and are largely present in isolates of the EHEC E. coli O157 : H7. More interestingly, a predicted type I restriction and modification system was encoded on this segment of the C4/1 BAC, with the hsdSMR (ecoE specificity, data not shown) genes similar to that of the EHEC (O157 : H7) and UPEC (CFT073) strains. The ecoE genes were expressed by BAC C4/1 in vitro, and conferred on the DH10B carrier strain the capacity to restrict differently methylated {lambda} phages (Table 4). As the DH10B carrier strain has all E. coli K-12 restriction modification systems (hsdRMS, mcr and mrr) disabled for the purpose of stable propagation of large heterologous DNA segments, the ecoE restriction modification system expressed by C4/1 BAC could contribute to the enhanced persistence of the C4/1 clone in the animal intestine by bringing about an enhanced protection of DH10B against bacteriophage attack. However, to confirm a role for the C4/1-encoded genes in intestinal colonization, defined mutants of A034/86 and corresponding trans-complemented strains will be constructed and then characterized in the porcine model.

Conclusions
Genomic characterization of E. coli strain A0 34/86 revealed the presence of a broad range of genes classified so far as enterobacterial ‘virulence factors’. However, these factors may instead contribute to the fitness and colonization capacity of this commensal strain. The detected genomic islands may provide a selective advantage under specific environmental conditions, and might contribute to the high colonization capacity of this strain in humans and piglets. Our results are in line with the recent observations of Dobrindt and colleagues that a rather high proportion of ORFs encoding putative ‘virulence-associated’ or PAI-localized genes are also present in non-pathogenic commensal and probiotic E. coli isolates (Grozdanov et al., 2004). However, in contrast to E. coli A0 34/86, the probiotic E. coli strain Nissle lacks virulence factors such as {alpha}-haemolysin, P-fimbrial adhesins and the semirough lipopolysaccharide phenotype but, like E. coli A0 34/86, it does express microcins, different iron-uptake systems, adhesins and proteases, which thus may support its survival and its successful colonization of the human gut and not be virulence factors as defined for other E. coli strains (Grozdanov et al., 2004). These results clearly suggest that for the species E. coli the terms ‘virulence’, ‘fitness' and ‘colonization factor’ are overlapping, and that the relationship between virulence and commensalism in E. coli is far from being fully understood. It is tempting to speculate that expression of these ‘virulence-associated’ factors may account for the capacity of the A0 34/86 strain to outcompete pathogens and to immunize the host by stimulating the previously observed strong and polyclonal secretory IgA response (Lodinova & Jouja, 1977; Lodinova-Zadnikova et al., 1991), which in turn might explain its clinical efficacy when used as a live oral vaccine against nosocomial infections and diarrhoeal disease of newborn infants.


   ACKNOWLEDGEMENTS
 
The authors are indebted to Lubomir Grozdanov for performance of the microcin activity assay (Würzburg), to Marie Weiserova for the {lambda} phage restriction assays on DH10B carrying BAC C4/1, to J. P. Nougayréde and E. Oswald for conducting the CNF activity assay and M. Weiserova for EcoE, as well as to the students of the Genome Analysis Course of the Institut Pasteur who contributed to the sequencing of BAC C4/1. The gift of polyclonal anti-FocA serum by A. S. Khan (Würzburg University) is gratefully acknowledged. This work was supported by a joint French–Czech ‘thèse-en-cotutelle’ fellowship to J. H. and by funds from Institut Pasteur, Deutsche Forschungsgemeinschaft (SFB479, TP A1) and grant No. S5020311 of the Grant Agency of the Czech Academy of Sciences.


   REFERENCES
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS AND DISCUSSION
REFERENCES
 
Altschul, S. F., Madden, T. L., Schaffer, A. A., Zhang, J., Zhang, Z., Miller, W. & Lipman, D. J. (1997). Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res 25, 3389–3402.[Abstract/Free Full Text]

Bach, S., de Almeida, A. & Carniel, E. (2000). The Yersinia high-pathogenicity island is present in different members of the family Enterobacteriaceae. FEMS Microbiol Lett 183, 289–294.[CrossRef][Medline]

Baga, M., Goransson, M., Normark, S. & Uhlin, B. E. (1985). Transcriptional activation of a pap pilus virulence operon from uropathogenic Escherichia coli. EMBO J 4, 3887–3893.[Abstract]

Blattner, F. R., Plunkett, G. R., Bloch, C. A. & 14 other authors (1997). The complete genome sequence of Escherichia coli K-12. Science 277, 1453–1474.[Abstract/Free Full Text]

Bonacorsi, S., Clermont, O., Houdouin, V. & 7 other authors (2003). Molecular analysis and experimental virulence of French and North American Escherichia coli neonatal meningitis isolates: identification of a new virulent clone. J Infect Dis 187, 1895–1906.[CrossRef][Medline]

Buchrieser, C., Weagant, S. D. & Kaspar, C. W. (1994). Molecular characterization of Yersinia enterocolitica by pulsed-field gel electrophoresis and hybridization of DNA fragments to ail and pYV probes. Appl Environ Microbiol 60, 4371–4379.[Abstract]

Buchrieser, C., Brosch, R., Bach, S., Guiyoule, A. & Carniel, E. (1998). The high-pathogenicity island of Yersinia pseudotuberculosis can be inserted into any of the three chromosomal asn tRNA genes. Mol Microbiol 30, 965–978.[CrossRef][Medline]

Buchrieser, C., Rusniok, C., Frangeul, L., Couve, E., Billault, A., Kunst, F., Carniel, E. & Glaser, P. (1999). The 102-kilobase pgm locus of Yersinia pestis: sequence analysis and comparison of selected regions among different Yersinia pestis and Yersinia pseudotuberculosis strains. Infect Immun 67, 4851–4861.[Abstract/Free Full Text]

Carniel, E., Guilvout, I. & Prentice, M. (1996). Characterization of a large chromosomal "high-pathogenicity island" in biotype 1B Yersinia enterocolitica. J Bacteriol 178, 6743–6751.[Abstract/Free Full Text]

Clermont, O., Bonacorsi, S. & Bingen, E. (2000). Rapid and simple determination of the Escherichia coli phylogenetic group. Appl Environ Microbiol 66, 4555–4558.[Abstract/Free Full Text]

De Rycke, J., Nougayrede, J. P., Oswald, E. & Mazars, P. (1997). Interaction of Escherichia coli producing cytotoxic necrotizing factor with HeLa epithelial cells. Adv Exp Med Biol 412, 363–366.[Medline]

Dobrindt, U., Blum-Oehler, G., Nagy, G., Schneider, G., Johann, A., Gottschalk, G. & Hacker, J. (2002). Genetic structure and distribution of four pathogenicity islands (PAI I(536) to PAI IV(536)) of uropathogenic Escherichia coli strain 536. Infect Immun 70, 6365–6372.[Abstract/Free Full Text]

Dobrindt, U., Agerer, F., Michaelis, K. & 7 other authors (2003). Analysis of genome plasticity in pathogenic and commensal Escherichia coli isolates by use of DNA arrays. J Bacteriol 185, 1831–1840.[Abstract/Free Full Text]

Dozois, C. M., Dho-Moulin, M., Bree, A., Fairbrother, J. M., Desautels, C. & Curtiss, R., 3rd (2000). Relationship between the Tsh autotransporter and pathogenicity of avian Escherichia coli and localization and analysis of the Tsh genetic region. Infect Immun 68, 4145–4154.[Abstract/Free Full Text]

Ewing, B. & Green, P. (1998). Base-calling of automated sequencer traces using phred. II. Error probabilities. Genome Res 8, 186–194.[Abstract/Free Full Text]

Frangeul, L., Glaser, P., Rusniok, C., Buchrieser, C., Duchaud, E., Dehoux, P. & Kunst, F. (2004). CAAT-Box, contigs-Assembly and Annotation Tool-Box for genome sequencing projects. Bioinformatics 20, 790–797.[Abstract/Free Full Text]

Gilson, L., Mahanty, H. K. & Kolter, R. (1990). Genetic analysis of an MDR-like export system: the secretion of colicin V. EMBO J 9, 3875–3894.[Abstract]

Gordon, D., Abajian, C. & Green, P. (1998). Consed: a graphical tool for sequence finishing. Genome Res 8, 195–202.[Abstract/Free Full Text]

Grozdanov, L., Raasch, C., Schulze, J., Sonnenborn, U., Gottschalk, G., Hacker, J. & Dobrindt, U. (2004). Analysis of the genome structure of the non-pathogenic probiotic Escherichia coli Strain NIssle 1917. J Bacteriol 186, 5432–5441.[Abstract/Free Full Text]

Hacker, J. & Carniel, E. (2001). Ecological fitness, genomic islands and bacterial pathogenicity. A Darwinian view of the evolution of microbes. EMBO Rep 2, 376–381.[Abstract/Free Full Text]

Hayashi, T., Makino, K., Ohnishi, M. & 19 other authors (2001). Complete genome sequence of enterohemorrhagic Escherichia coli O157 : H7 and genomic comparison with a laboratory strain K-12. DNA Res 8, 11–22.[Medline]

Huang, S. H., Wass, C., Fu, Q., Prasadarao, N. V., Stins, M. & Kim, K. S. (1995). Escherichia coli invasion of brain microvascular endothelial cells in vitro and in vivo: molecular cloning and characterization of invasion gene ibe10. Infect Immun 63, 4470–4475.[Abstract]

Huang, S. H., Chen, Y. H., Kong, G., Chen, S. H., Besemer, J., Borodovsky, M. & Jong, A. (2001). A novel genetic island of meningitic Escherichia coli K1 containing the ibeA invasion gene (GimA): functional annotation and carbon-source-regulated invasion of human brain microvascular endothelial cells. Funct Integr Genomics 1, 312–322.[CrossRef][Medline]

Johnson, J. R. & Stell, A. L. (2000). Extended virulence genotypes of Escherichia coli strains from patients with urosepsis in relation to phylogeny and host compromise. J Infect Dis 181, 261–272.[CrossRef][Medline]

Kaper, J. B. & Hacker, J. (1999). Pathogenicity Islands and Other Mobile Genetic Elements of Diarrheagenic Escherichia coli. Chapter 3 in Pathogenicity Islands and Other Mobile Virulence Elements, pp. 33–58. Washington, DC: American Society for Microbiology.

Lan, R. & Reeves, P. R. (2000). Intraspecies variation in bacterial genomes: the need for a species genome concept. Trends Microbiol 9, 396–401.[CrossRef]

Lodinova, R. & Jouja, V. (1977). Antibody production by the mammary gland in mothers after artificial oral colonisation of their infants with a non-pathogenic strain E. coli O83. Acta Paediatr Scand 66, 705–708.[Medline]

Lodinova, R., Jouja, V. & Lanc, A. (1967). Influence of the intestinal flora on the development of immune reactions in infants. J Bacteriol 93, 797–800.[Medline]

Lodinova, R., Jouja, V., Vinsova, N., Vocel, J. & Melkova, J. (1980). New attempts and possibilities in prevention and treatment of intestinal coli-infections in infants. Czech Med 3, 47–58.[Medline]

Lodinova-Zadnikova, R., Slavikova, M., Tlaskalova-Hogenova, H., Adlerberth, I., Hanson, L. A., Wold, A., Carlsson, B., Svanborg, C. & Mellander, L. (1991). The antibody response in breast-fed and non-breast-fed infants after artificial colonization of the intestine with Escherichia coli O83. Pediatr Res 29, 396–399.[Abstract]

Lodinova-Zadnikova, R., Tlaskalova, H., Korych, B. & Bartakova, Z. (1995). The antibody response in infants after oral administration of inactivated and living E. coli vaccines and their protective effect against nosocomial infections. Adv Exp Med Biol 371B, 1431–1438.[Medline]

Lodinova-Zadnikova, R., Sonnenborn, U. & Tlaskalova, H. (1998). Probiotics and E. coli infections in man. Vet Q 20, S78–81.

Lodinova-Zadnikova, R., Cukrowska, B. & Tlaskalova-Hogenova, H. (2003). Oral administration of probiotic Escherichia coli after birth reduces frequency of allergies and repeated infections later in life (after 10 and 20 years). Int Arch Allergy Immunol 131, 209–211.[Medline]

Nataro, J. P. & Kaper, J. B. (1998). Diarrheagenic Escherichia coli. Clin Microbiol Rev 11, 142–201.[Abstract/Free Full Text]

Ochman, H. & Jones, I. B. (2000). Evolutionary dynamics of full genome content in Escherichia coli. EMBO J 19, 6637–6643.[Abstract/Free Full Text]

Otto, B. R., van Dooren, S. J., Nuijens, J. H., Luirink, J. & Oudega, B. (1998). Characterization of a hemoglobin protease secreted by the pathogenic Escherichia coli strain EB1. J Exp Med 188, 1091–1103.[Abstract/Free Full Text]

Pearson, W. R. (1990). Rapid and sensitive sequence comparison with FASTP and FASTA. Methods Enzymol 183, 63–98.[Medline]

Pearson, W. R. & Lipman, D. J. (1988). Improved tools for biological sequence comparison. Proc Natl Acad Sci U S A 85, 2444–2448.[Abstract]

Perna, N. T., Plunkett, G., Burland, V. & 25 other authors (2001). Genome sequence of enterohaemorrhagic Escherichia coli O157 : H7. Nature 409, 529–533.[CrossRef][Medline]

Rodriguez, E., Gaggero, C. & Lavina, M. (1999). The structural gene for microcin H47 encodes a peptide precursor with antibiotic activity. Antimicrob Agents Chemother 43, 2176–2182.[Abstract/Free Full Text]

Roe, A. J., Currie, C., Smith, D. G. E. & Gally, D. L. (2001). Analysis of type 1 fimbriae expression in verotoxigenic Escherichia coli: a comparison between serotypes O157 and O26. Microbiology 147, 145–152.[CrossRef][Medline]

Rutherford, K., Parkhill, J., Crook, J., Horsnell, T., Rice, P., Rajandream, M. A. & Barrell, B. (2000). Artemis: sequence visualization and annotation. Bioinformatics 16, 944–955.[Abstract]

Welch, R. A. (2001). RTX toxin structure and function: a story of numerous anomalies and few analogies in toxin biology. Curr Top Microbiol Immunol 257, 85–111.[Medline]

Welch, R. A., Burland, V., Plunkett, G., 3rd & 16 other authors (2002). Extensive mosaic structure revealed by the complete genome sequence of uropathogenic Escherichia coli. Proc Natl Acad Sci U S A 99, 17020–17024.[Abstract/Free Full Text]

Weiserova, M., Janscak, P., Benada, O., Hubacek, J., Zinkevich, V. E., Glover, S. W. & Firman, K. (1993). Cloning, production and characterisation of wild type and mutant forms of the R.EcoK endonucleases. Nucleic Acids Res 21, 373–379.[Abstract]

Wold, A. E., Caugant, D. A., Lidin-Janson, G., de Man, P. & Svanborg, C. (1992). Resident colonic Escherichia coli strains frequently display uropathogenic characteristics. J Infect Dis 165, 46–52.[Medline]

Received 8 July 2004; revised 29 September 2004; accepted 12 October 2004.



This Article
Abstract
Full Text (PDF)
HTML Page - index.htslp
Alert me when this article is cited
Alert me if a correction is posted
Citation Map
Services
Email this article to a friend
Similar articles in this journal
Similar articles in PubMed
Alert me to new issues of the journal
Download to citation manager
Google Scholar
Articles by Hejnova, J.
Articles by Buchrieser, C.
Articles citing this Article
PubMed
PubMed Citation
Articles by Hejnova, J.
Articles by Buchrieser, C.
Agricola
Articles by Hejnova, J.
Articles by Buchrieser, C.


HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
INT J SYST EVOL MICROBIOL MICROBIOLOGY J GEN VIROL
J MED MICROBIOL ALL SGM JOURNALS
Copyright © 2005 Society for General Microbiology.