The lpf Gene Cluster for Long Polar Fimbriae Is Not Involved in Adherence of Enteropathogenic Escherichia coli or Virulence of Citrobacter r
http://www.100md.com
感染与免疫杂志 2006年第1期
Center for Vaccine Development and Department of Microbiology and Immunology, University of Maryland School of Medicine, Baltimore, Maryland, 21201
Department of Microbiology and Immunology, University of Texas Medical Branch, Galveston, Texas 77555
Centre for Molecular Microbiology and Infection, Division of Cell and Molecular Biology, Imperial College London, London SW7 2AZ
Centre for Paediatric Gastroenterology, Royal Free Hospital, London NW3 2QG, United Kingdom
ABSTRACT
Using the enteropathogenic Escherichia coli (EPEC) genome sequence, we found that EPEC E2348/69 has an lpfABCDE gene cluster homologous (about 60% identical at the protein level) to the Salmonella long polar fimbria (LPF) operon. To determine whether this operon is essential for adherence, the lpfABCDE23 genes were deleted from EPEC strain E2348/69 by allelic exchange. Analysis of the resulting EPEClpfABCDE23 strain showed no change in adherence to HeLa cells or to human intestinal biopsy cells in the in vitro organ culture (IVOC) system compared to the wild type. Sera from volunteers experimentally infected with E2348/69 showed no antibody response to the major subunit protein, LpfA. These results suggested that the lpfE23 gene cluster is not necessary for EPEC adherence and attaching/effacing (A/E) lesion formation on human biopsy samples and is not expressed during human infection. We also identified an lpf gene cluster in Citrobacter rodentium strain ICC168 (lpfcr). A lpfAcr mutant of ICC168 retained wild-type adherence and A/E lesion-forming activity on HeLa cells. C3H/HeJ mice were infected with a wild-type C. rodentium strain and its lpfAcr isogenic mutant. Both strains were recovered at high levels in stools, and there were no significant differences between the groups both in terms of the number of CFU/organ (colon and cecum) and in terms of the amount of hyperplasia, as measured by weight. Similar results were observed in a second mouse strain, C57BL/6. These data suggest that in addition to playing no apparent role in EPEC pathogenesis, lpfcr is not required for C. rodentium virulence in either the C3H/HeJ or C57BL/6 mouse model.
INTRODUCTION
Enterohemorrhagic Escherichia coli (EHEC), enteropathogenic E. coli (EPEC), rabbit-specific enteropathogenic E. coli (REPEC), and Citrobacter rodentium are attaching/effacing (A/E) bacterial pathogens that attach to host intestinal epithelium and efface brush border microvilli, forming pathognomonic A/E lesions (6, 12). This characteristic adherence pattern is encoded by the locus of enterocyte effacement (LEE) pathogenicity island (23), which contains genes encoding the outer membrane protein intimin (18), its receptor Tir/EspE (7, 20), and a type III secretion system (17). EPEC, the first A/E pathogen to be described, also has a type IV pilus, bundle-forming pilus (BFP), as a potential adherence factor (13). While EPEC is a human pathogen, C. rodentium is a natural pathogen of mice and now is used as an animal model for studying virulence-associated factors of EPEC or EHEC (8, 31).
The long polar fimbriae (LPF), another potential adherence factor that was originally described in Salmonella, have been identified in other A/E bacteria, such as EHEC and REPEC. The LPF of Salmonella enterica serovar Typhimurium were shown to direct the attachment of bacteria to murine Peyer's patches in vivo (4, 5). Two nonidentical lpf loci are present in EHEC O157:H7 (35, 36), and one or both of them were found to be involved in microcolony formation, increased adherence to HEp-2 cells when cloned into a nonfimbriated K-12 strain, and the colonization by and persistence of E. coli O157:H7 in swine and sheep (19). The lpf141 locus of REPEC O15:H- strain 83/39 appeared to play a role in the early stages of REPEC infection of rabbits and was essential for the development of severe diarrhea (26). Recently, sequencing of the EPEC O127:H6 strain E2348/69 and C. rodentium ICC168 genomes revealed lpf loci in each of those pathogens (http://www.sanger.ac.uk/Projects/Escherichia_Shigella/ and http://www.sanger.ac.uk/Projects/C_rodentium, respectively), but it is unclear whether these loci are also functional in contributing to disease caused by these A/E pathogens. We therefore investigated the potential role of the EPEC and C. rodentium LPF in adherence and/or virulence.
MATERIALS AND METHODS
Strains and plasmids. E. coli E2348/69 (O127:H6) is a prototypic EPEC strain proven to cause diarrhea in volunteers (22). JPN15 is a derivative of E2348/69 cured of the EPEC adherence factor plasmid. E. coli HB101(pCVD462) (also known as pLEE) contains the cloned LEE pathogenicity island and confers A/E on HEp-2 and Caco-2 cells (24). pMAR7 is a Tn801 insertion derivative of the EPEC adherence factor plasmid from EPEC strain E2348/69 (O127:H6) (1). C. rodentium ICC168 is a wild-type strain shown to cause colonic epithelial cell hyperplasia in mice (8, 31). A reference collection of representative diarrheagenic E. coli clones known as the diarrheagenic E. coli (DEC) collection was screened for the presence of lpf genes. This collection is available from the STEC Center at Michigan State University (http://www.shigatox.net). Two of these strains, 572-56 and 3787-62, representative of the DEC1 and DEC2 clones, respectively, were further analyzed by mutating the lpf genes and testing for resulting changes in adherence as described below.
Cloning of the EPEC lpf gene cluster. The lpf gene cluster of EPEC strain E2348/69 was amplified by PCR with Easy-A high-fidelity PCR cloning enzyme (Stratagene) using specific primers lpf-Fto (5'-CTCGGTTTCGGTATCGGCAGC-3') and lpf-Rto (5'-CCAGCTGCGGCGGCCAAACA-3'). The 6,484-bp amplicon was cloned into pGEM-T easy (Promega) to yield pELP102, whose insert was sequenced. The ApaI-SacI fragment containing the amplified insert DNA of pELP102 was subcloned into the low-copy-number vector pWKS130 (38) to yield pELP130. A 1.1-kb EcoRI fragment containing the lpfA gene of pELP102 was subcloned into pSU18 (3) to yield plpfA2348F and plpfA2348R, in which the lpfA gene is oriented in the same and opposite direction, respectively, as the lacUV5 promoter.
Bacterial adhesion to cultured epithelial cells. Adherence of bacterial strains to cultured Hep-2, HeLa, T84, and Caco-2 epithelial cells was evaluated as previously described (35). Briefly, bacteria were grown statically in Luria-Bertani (LB) broth overnight at 37°C and inoculated at a multiplicity of infection of approximately 50:1 onto semiconfluent cultured epithelial cell monolayers grown on 24-well microtiter plates in Dulbecco's modified Eagle's medium (DMEM) with 10% fetal calf serum. Prior to infection, fresh DMEM without supplements such as mannose was added to cells. Bacteria and cells were incubated for 3 h and washed five times with 1 ml of sterile phosphate-buffered saline (PBS). The monolayers were fixed and stained with Giemsa's solution for microscopic evaluation. To quantify E. coli adherence, the infected monolayers were washed five times with PBS, and the adherent bacteria were recovered with 200 μl of 0.1% Triton X-100 in PBS buffer and plated on Luria agar plates containing the appropriate antibiotic. Data are expressed as the percentages of the bacterial inocula recovered from triplicate wells and represent at least three separate experiments. This protocol was used for all adherence assays unless otherwise described. The fluorescence actin staining (FAS) test was performed as previously described (21).
Organ culture adhesion assay. In vitro organ culture (IVOC) was performed as described previously (14), with ethical approval and informed consent, using small intestinal samples. Large-bowel mucosa was not used. The assay was terminated at 2 and 4 h on three occasions to determine "early" events. "Later" events were observed at 6 h on two occasions, and all other experiments were maintained for 8 h. Tissue culture medium was changed every 2 h, although some medium plus bacterial inoculum would remain within the foam insert supporting the tissue sample and in association with the tissue by surface tension. Each bacterial strain was examined in IVOC on at least three occasions with tissue from different children. Strain E2348/69 and an uninoculated specimen were included with each experimental culture to act as positive and negative controls, respectively.
Tissue processing. After the specified culture period, specimens were thoroughly washed three times to remove any nonadherent bacteria and then prepared for scanning electron microscopy (SEM) as described previously (14). Following fixation and dehydration, specimens were placed in 100% ethanol and critical point dried in liquid CO2 with an Emitech K850 critical point drying apparatus. Samples were then mounted on aluminum stubs and sputter coated with gold-palladium using a Polaron sputter coating apparatus. Specimens were observed with a JEOL JSM-5300 SEM at an accelerating voltage of 30 kV.
Construction of lpf mutants. In-frame lpf deletion mutants were constructed in three EPEC strains and one C. rodentium strain. The wild-type strains and the resulting mutant designations, in parentheses, were E2348/69 (lpfABCDE23), 572-56 (lpfABCD572), 3787-62 (lpfABCD3787), and C. rodentium (lpfABCDcr). Bacteria were cultured at 30°C for all constructions. A set of oligonucleotides, primer lpfD-F (5'-GGTCCGATCGGTACCCAACAACATC-3') and primer lpfA-R (5'-CAACCTGAGCGGTACCATAACC-3'), were used as described above to amplify DNA fragments of 5.0 kb (corresponding to the in-frame lpfABCDE23 deletion mutation and additional flanking regions from pELP102) by PCR. The corresponding oligonucleotides contained restriction sites for the endonuclease KpnI (shown in boldface in the primer sequences). This enzyme was used for self-ligation of the amplification product. The resulting 2.0-kb ApaI-SacI fragment (containing the in-frame lpfABCDE23 deletion) was subcloned into the plasmid pWM91 (25) to yield a recombinant suicide plasmid that was introduced by conjugation into EPEC strain E2348/69 and integrated into the chromosome. Further cultivation in the presence of 5% sucrose led to a second round of recombination and the excision of the plasmid with the wild-type lpfABCDE23 gene. Bacteria were confirmed to harbor a mutated lpfABCDE23 allele by PCR using the two primers lpf-F (5'-TTTGGGGCGCGATTGTACTGG-3') and lpf-R (5'-TTCACCAATAATGAAAACGAC-3'), which are specific to sequences outside of those for primers lpf-Fto and lpf-Rto, respectively. lpfABCD572 and lpfABCD3787 deletion mutants were constructed in the same manner as the lpfABCDE23 deletion mutant of EPEC E2348/69, except that 3787-62 (lpfABCD3787) was confirmed to harbor a mutated lpfABCD3787 allele by PCR using primers lpf-F2 (5'-TTACTGTTTCCCGTCACTGG-3') and lpf-R2 (5'-GAAGGTCATGTAATAGCTGGC-3'), which are also specific to sequences outside of those for primers lpf-Fto and lpf-Rto, respectively.
The lpfcr gene cluster of C. rodentium strain ICC168 was amplified by PCR with Easy-A high-fidelity PCR cloning enzyme (Stratagene) using primers CRlpfA-F (5'-AGATGATTTTCTCTGTTTTC-3') and CRlpfA-R (5'-GCGAGGCGCTGTCGCCCAGC-3'). The 3,298-bp amplicon was cloned into pGEM-T easy (Promega) to yield pCRlpfA. An in-frame deletion in lpfAcr was made using an inverse PCR approach with pCRlpfA as the template and a set of oligonucleotides, primer CRlpfA-F (5'-GCTTATGTCGGGTACCGCGTTCCG-3') and primer CRlpfA-R (5'-GTATCGAGGAGGTACCCGACAAAATTC-3'). The primers contain KpnI restriction sites (in boldface). The 5.9-kb product was cut with KpnI and religated, yielding the lpfAcr construct. A 2.9-kb SphI-PstI fragment containing lpfAcr was subcloned into the suicide plasmid pCACTUS (33). The lpfAcr suicide plasmid was introduced by electroporation into strain ICC168 and integrated into the chromosome by growth at 42°C. Plating in the presence of 5% sucrose led to the isolation of recombinants carrying the lpfAcr allele. Bacteria were confirmed to harbor the lpfAcr allele by PCR using the two outside primers CRlpfAout-F (5'-CCGAGGCGATGCCTAACGCC-3') and CRlpfAout-R (5'-GCTGTCAAAGATATCTGCGG-3'), which flank the region used for homologous recombination.
GST-LpfAE23 expression. The lpfAE23 gene was amplified by PCR using the primers GST-lpfA (5'-CGCGGATCCGCTGATTCTGGTGATGGTAC-3') and 3572 (5'-ACCATACCGGACTTTTCAAT-3') and cloned into pGEX-2T (Promega) to yield pGST-lpfAE23. The corresponding oligonucleotide GST-lpfAE23 contained a restriction site for the endonuclease BamHI (shown in boldface in the primer sequence). To induce the GST-lpfAE23 fusion protein, 0.1 mM isopropyl--D-thiogalactopyranoside (IPTG) was added to a logarithmic-phase culture of DH5/pGST-lpfA and shaken for 2 h at 30°C. The GST-LpfAE23 fusion protein was purified by glutathione Sepharose 4B (Amersham Biosciences) and digested with thrombin protease per the manufacturer's instructions (Amersham Biosciences). The resultant proteins were analyzed as described previously (34). Briefly, GST-LpfAE23 was eluted, without thrombin treatment, by boiling with an equal volume of 2x sodium dodecyl sulfate (SDS) loading buffer. To obtain free LpfAE23, after incubation in PBS with thrombin protease, the supernatant and pellet beads were combined with an equal volume of 2x SDS loading buffer (the ratio of the supernatant to the pellet beads used was 1:1 [vol/vol]). Five microliters each of the prepared supernatant and pellet samples was resolved by SDS-12% polyacrylamide gel electrophoresis and stained with Coomassie brilliant blue. Antiserum against LpfAE23 was prepared by immunizing rabbits with the resulting purified LpfAE23 (Lampire Biological Laboratories Inc.).
Electron microscopy. Bacteria were grown at 37°C under various conditions including different media (DMEM, LB, MacConkey, and colonization factor antigen; broth versus agar), different incubation times (2, 3, 5 h, or overnight), and with or without HEp-2 cells to examine the presence of fimbriae using electron microscopy. The bacteria were recovered and allowed to adhere to a carbon-formvar-coated 300-mesh copper grid (Electron Sciences). The fimbriae were visualized by negative staining with 0.1% phosphotungstic acid, pH 7.4, and the grids were examined in a Philips CM 120 electron microscope (FEI Co.). In order to confirm that the sample was properly prepared, typical type 1 fimbriae of DH5 grown on LB agar were observed as a positive control.
RT-PCR. One milliliter of EPEC E2348/69 overnight culture in LB was inoculated into 10 ml of DMEM (Catalog No. 11995-065; Invitrogen) and statically incubated for 2 h in 37°C, 5% CO2. Total RNA was extracted by RNeasy Starter kit (QIAGEN). First-strand cDNA was synthesized by using a SuperScript First-Strand synthesis system for reverse transcription-PCR (RT-PCR) (Invitrogen) and the gene-specific primers BamHItirMR (5'-GCTGAGGCCGGATCCAATACCCG-3') for tir and 3572 for lpfAE23. The cDNA for tir and lpfAE23 was amplified by using a combination of primers SwaItirF (5'-CGTGTCAAATATTTTAAATAAAAG-3') and BamHItirMR as well as lpfARBS-F (5'-GGAGCTCGAGATGAAAAAGG-3') and lpfA-R, respectively. The resultant PCR products were loaded into 0.7% agarose with a 1-kb Plus DNA ladder (Invitrogen) and analyzed.
One milliliter of an overnight culture of C. rodentium in LB was inoculated into 10 ml of DMEM (Catalog No. 11995-065; Invitrogen) and statically incubated for 2 h at 37°C, 5% CO2. Total RNA was extracted using the RNeasy Starter kit (QIAGEN). First-strand cDNA was synthesized by using a SuperScript First-Strand synthesis system for RT-PCR (Invitrogen) and a gene-specific primer, CRlpfA-R. The cDNA for lpfABCcr was amplified using primers CRlpfA-F and CRlpfA-R. PCR products were loaded into 0.7% agarose with a 1-kb Plus DNA ladder (Invitrogen) and analyzed.
Human immune responses. A serological response against LpfA during the course of human infection was evaluated using convalescent human serum obtained from volunteers experimentally infected with E2348/69 in previous studies (9, 22, 32). The GST-LpfAE23 fusion protein was prepared and transferred to polyvinylidene difluoride membrane as described above, and immunoblots were performed as previously described (22) using horseradish peroxidase-conjugated goat antibody to human immunoglobulin G.
Mouse experiments. Male, specific-pathogen-free C3H/HeJ mice (6 to 8 weeks old) were purchased from Harlan Olac (Bicester, United Kingdom), and female, specific-pathogen-free C57BL/6 (6 to 8 weeks old) were purchased from B and K (Hull, United Kingdom). Mice were housed in individual ventilated cages with free access to food and water. Unanesthetized mice were orally gavaged with 200 μl of bacterial suspension using a gavage needle. The viable count of the inoculum was determined by retrospective plating on LB agar containing nalidixic acid to select for nalidixic acid-resistant C. rodentium.
In the infection experiments, bacteria were grown to stationary phase in LB broth plus nalidixic acid. For the C3H/HeJ mice, the bacteria were pelleted by centrifugation, washed, and then diluted 1:10 with PBS and gavaged into mice (approximately 1 x 107 CFU per mouse). The C57BL/6 mice required a higher inoculum, so 10 ml of stationary-phase bacteria was pelleted by centrifugation and resuspended in 1 ml of PBS and then gavaged into mice (approximately 1 x 109 CFU per mouse). Each strain was tested in groups of at least four mice per bacterial strain. Stool samples were recovered aseptically at various time points after inoculation, and the numbers of viable bacteria per gram of stool were determined by plating onto LB agar containing appropriate antibiotics. Mice were killed at either the peak of infection, 11 days postchallenge (C3H/HeJ mice), or once the infection had been cleared, 28 days postchallenge (C57BL/6 mice). In all cases, the distal 8 cm of colon was aseptically removed and weighed after removal of fecal pellets. Colons were then homogenized mechanically using a Seward 80 stomacher (London, United Kingdom), and the numbers of viable bacteria per colon were determined by plating onto LB agar containing appropriate antibiotics.
RESULTS
Identification and cloning of EPEC lpfE23 genes. The complete genome sequence of E2348/69 available at http://www.sanger.ac.uk/Projects/Escherichia_Shigella/ revealed a locus whose predicted protein products showed homology to the LPF proteins of S. enterica serovar Typhimurium. The organization of the EPEC lpfE23 locus and the similarity of the predicted EPEC LPF proteins to LPF proteins of other pathogens are shown in Fig. 1 and Table 1. The percent G+C composition of the EPEC lpf locus is 43.1, substantially lower than that of the E. coli K-12 genome (50.8%), suggesting that this locus was introduced by horizontal gene transfer. The lpf locus was inserted between the yhiW and yhiJ genes corresponding to minute 79.9 of the E. coli K-12 chromosome (http://www.uni-giessen.de/gx1052/ECDC/ecdc/ecmap75.htm). The same insertion site is seen with the lpf operons of S. enterica serovar Typhimurium, EHEC O157:H7 lpf1, and REPEC (26, 35). The EPEC lpfE23 gene cluster was amplified by PCR, and the amplicon was cloned to yield pELP102 (see Materials and Methods). The resulting cloned amplicon was sequenced, and the insert DNA sequence was identical to that determined in the genomic sequencing.
Adherence phenotype of an EPEC lpf mutant to IVOC. In order to address whether LPFE23 is associated with adherence of EPEC to human small intestinal tissue, an EPEClpfABCDE23 mutant was constructed and tested in IVOC. As shown in Fig. 2, an EPEClpfABCDE23 mutant strain adhered to IVOC in a manner similar to that seen with the wild type, indicating that EPEC LPFE23 is not essential for adherence to human intestinal tissue in the IVOC system.
Effect of lpfE23 on bacterial adherence to cultured epithelial cells. EPEC LPFE23 was also characterized for adherence to cultured epithelial cell lines. When EPEClpfABCDE23 was investigated for the ability to adhere to Caco-2 cells, it showed an adherence phenotype similar to that of parental wild-type EPEC (data not shown). Similar results were also seen in HEp-2 and T-84 cells under various experimental conditions, including different growth phase (log or stationary), static or shaking culture conditions, and growth in LB and DMEM (data not shown). To address the possibility that the function of LPFE23 is masked by other EPEC adherence factors, we investigated whether the cloned lpfE23 confers adherence to E. coli K-12. When the lpfE23 gene cluster was cloned in a high-copy vector, HB101(pELP102) did not show increased adherence to T-84 cells (4.7% ± 2.5%, CFU of output/input) compared to the parental HB101 strain containing the control vector only, which demonstrates minimal adherence (7.5% ± 3.8%). As previously observed, HB101(pMAR7) expressing BFP showed increased adherence to cultured epithelial cell lines (20.5% ± 6.0%). EPEClpfABCDE23 and E. coli K-12 containing the cloned lpfE23 genes on high-copy-number (pELP102) or low-copy-number (pELP130) cloning vectors were further tested using combinations of the following experimental conditions: different epithelial cells (HEp-2, Caco-2, T-84, and HeLa), different hosts for cloned lpfE23 [DH5 and HB101(pCVD462)], different multiplicities of infection (1 to 200), different bacterial culture conditions (growth phase with log or stationary, shaking versus static, DMEM versus LB), and different infection times (2 to 5 h). Regardless of the conditions tested, no increased adherence activity was detected due to the presence of the lpfE23 locus (data not shown).
Because no apparent role was observed for LPFE23 in the adherence of EPEC strain E2348/69 to four epithelial cell lines or human intestinal tissue, we mutated the lpfE23 genes in two other EPEC strains to determine whether E2348/69 was unusual in this regard. We first examined the presence of lpfE23 genes in the DEC collection of diarrheagenic E. coli clones and found lpfE23 sequences in representatives of the EPEC1, EHEC1, and EHEC2 groups (Table 2). The amplified fragments in these clones were the same size, 6.5 kb, as that seen with E2348/69. Strains of the EPEC2 group gave a mixed pattern, with two representative strains of DEC11 clones yielding a 6.5-kb fragment and two strains of the DEC12 group yielding a 3.0-kb fragment. We generated lpf mutants of two additional strains from the EPEC1 group, lpfABCD527 and lpfABCD3787, and tested them for adherence to HeLa cells. No differences were seen in the adherence of these mutants from that of the parent strains (data not shown).
Expression of lpfE23 genes in vitro. To investigate why the adherence phenotype associated with other LPF was not detected with EPEC LPFE23, the expression of the lpfE23 genes was analyzed by electron microscopy. LPF from three enteric pathogens, S. enterica serovar Typhimurium, EHEC, and REPEC, have previously been visualized using negative staining and electron microscopy (4, 26, 36). In each of these cases, visualizing LPF was much easier with the cloned lpf genes expressed in a nonfimbriated K-12 strain than in the wild-type strain. We used these same techniques to examine wild-type E2348/69 for LPF without success. When pELP102 containing the EPEC E2348/69 lpfE23 genes was transformed into nonfimbriated E. coli strain ORN172, we also failed to observe fimbrial structures (data not shown). We examined five different EPEC strains containing lpfE23 genes grown under a variety of different growth conditions but still did not observe fimbrial structures (data not shown).
The failure to visualize EPEC LPFE23 is consistent with our inability to detect a difference in adherence to epithelial cell lines or to IVOC. We therefore examined expression of LPF by protein and mRNA analysis. To generate antisera against LPFE23, the lpfAE23 gene was cloned into the expression vector pGEX-2T, which creates a GST-LpfAE23 fusion under the transcriptional control of the Ptac promoter contained in this vector. The addition of IPTG induced the expression of a protein with the expected size of GST-LpfAE23, and digestion of this protein with thrombin protease resulted in a protein with the expected size of LpfAE23 (data not shown). This protein was purified and used to raise rabbit antisera against LpfAE23 as described in Materials and Methods. However, when the lpfAE23 gene was cloned into a standard plasmid vector under its native promoter, no expression of LpfAE23 protein was detected by immunoblotting with anti-LpfAE23 antibody (data not shown). Two different K-12 strains (ORN172 and DH5) containing the cloned lpfE23 locus were grown under a variety of different expression conditions, including one in which LpfA of O157:H7 was detected with specific antibody (35), but no LpfAE23 could be detected by immunoblotting. The pSU18 vector contains a lacUV5 promoter, but expression of clones in the same or opposite orientation of this promoter with or without IPTG still yielded no LpfAE23 expression. When the EPEC E2348/69 wild-type strain was incubated at 37°C in DMEM, which are standard conditions for expression of EPEC adherence factors, only a faint band of lpfAE23 mRNA was detected by RT-PCR at a substantially lower level than was observed with tir mRNA in the same experiment (Fig. 3). These results indicate that the lpfAE23 gene, encoding the major fimbrial subunit, is poorly expressed, if at all, under the experimental conditions used in our adherence and electron microscopy studies.
Expression of lpfE23 genes in vivo. To investigate whether there was a serological response against LPFE23 during the course of human infection, convalescent-phase sera of 10 volunteers previously infected with E. coli strain E2348/69 (9, 22, 32) were tested against GST-LpfAE23 fusion protein in immunoblots. None of the volunteers showed an immune response to LpfAE23 in the convalescent-phase sera obtained 28 days after infection other than a faint, nonspecific band that was unchanged in intensity between preinfection and postinfection sera (data not shown).
Identification of C. rodentium lpfcr genes. Another lpf operon of an A/E organism was identified in the genome sequence of C. rodentium strain ICC168 (http://www.sanger.ac.uk/Projects/C_rodentium/). The predicted LpfAcr, LpfBcr, and LpfCcr proteins have amino acid similarities ranging from 32 to 46% to Lpf proteins in EPEC, REPEC, E. coli O157, and Salmonella (Fig. 1B, Table 1). No protein with striking homology to LpfD was seen in C. rodentium, but a predicted protein of size comparable to LpfD was encoded in the same location downstream of LpfCcr (Fig. 1C). LpfDcr does share 28% identity with a fimbrial adhesin precursor of Photorhabdus luminescens subsp. laumondii TTO1 (11). Additionally, a fimbrial domain was detected in the C terminus of LpfDcr using the "Conserved Domain Search" program (http://www.ncbi.nlm.nih.gov). LpfEcr also did not show high homology with other LpfE proteins (with a maximum of 22% identity seen with LpfE of E. coli O157 and Salmonella). However, this is consistent with LpfD and LpfE being poorly conserved among the lpf loci of O157:H7, EPEC, and REPEC (26, 35, 36). The G+C composition of the lpfcr locus is 54%, which is slightly lower than that of the 9 kb flanking it to one side and the 70 kb to the other side (both 57%). The sequences flanking lpfcr did not show high homology to E. coli K-12, indicating that the lpfcr gene cluster of C. rodentium strain ICC168 does not map to the same chromosomal location as the lpf operon of S. enterica serovar Typhimurium, EPEC, EHEC O157:H7, and REPEC (26, 35).
Virulence of the C. rodentium lpfAcr mutant in the C3H/HeJ mouse model. To address whether LPF is associated with the virulence of C. rodentium in mice, a ICC168lpfAcr mutant was constructed. In initial in vitro studies, lpfAcr had no effect on the ability of ICC168 to adhere to HeLa cells or the ability to produce a positive fluorescent actin staining (FAS) test indicative of A/E lesion formation (data not shown). Thus, LpfAcr is not required for in vitro adherence and A/E lesion formation on HeLa cells. To examine potential effects of lpfAcr on virulence, groups of four mice were inoculated with the wild-type C. rodentium strain and the lpfAcr mutant. The lpfAcr mutant behaved like the wild type. Both strains were recovered at high levels in stools collected at days 1, 4, 6, and 8 (Fig. 4A). At day 10, two mice infected with the wild-type strain and two mice infected with the lpfAcr mutant died. The remaining two mice in each group were sacrificed on day 11, and the colon and cecum were removed. The numbers of viable cells in the colon were similar in both groups, as was the extent of hyperplasia as measured by weight (data not shown).
Virulence of the C. rodentium lpfcr mutant in the C57BL/6 mouse model. The experimental infection was repeated in C57BL/6 mice, which are less susceptible to infection than C3H/HeJ mice when infected with C. rodentium. The milder response of BL/6 allows the length of infection to be extended (37), thereby providing an opportunity to detect a more subtle defect in virulence that may go undetected in the C3H/HeJ mice. Figure 4B shows the numbers of CFU/gram of stool shed by mice infected with the wild type and the lpfAcr mutant over 28 days. Colons were taken at day 28, and at this time point all the citrobacteria had been cleared from both the colon and the cecum. The average weights of colons ± standard errors are the following: wild-type-infected mice, 0.224 ± 0.014 g; lpfAcr-infected mice, 0.242 ± 0.009 g; and uninfected mice, 0.147 ± 0.008 g. The C57BL/6 mice do not exhibit as much hyperplasia as the C3H/HeJ mice in response to infection; however, there was significant hyperplasia in the distal 1 to 2 cm of the colon in mice infected with both the wild type and the lpfAcr mutant. Collectively, these data suggest that lpfcr is not required for virulence in our mouse models.
Expression of lpfcr genes in vitro. The expression of lpfcr was examined to confirm that these genes were expressed. When the C. rodentium wild-type strain was incubated at 37°C in DMEM, no expression of LpfAcr protein was detected in whole-cell extracts by immunoblotting with anti-EPEC LpfA antibody. Since the proteins share low similarity, it is possible that the anti-EPEC LpfA sera did not detect LpfAcr. Nevertheless, lpfABCcr mRNA was detected by RT-PCR using primers specific to lpfAcr and lpfC (Fig. 5A) but not when the 5' primer was outside the lpfAcr coding region (Fig. 5B). Thus, lpfcr appears to be at least transcribed under these conditions. Furthermore, this result suggests that lpfABCcr comprises an operon.
DISCUSSION
Our in vitro studies did not indicate a role for LPFE23 in adherence or A/E lesion formation by EPEC. Similarly, mutation of C. rodentium lpfcr did not alter the in vitro adherence of this organism to cultured epithelial cells. In addition, we examined expression of LpfAE23 using immunoblots of E2348/69 and K-12 containing cloned lpfE23 genes grown under a variety of cultural conditions and found no evidence of expression using antisera raised against the cloned EPEC lpfAE23 gene product. There are at least two possible reasons why LPFE23 is apparently not expressed in EPEC. The first possibility is that some of the LpfABCDEE23 proteins are inactive and therefore cannot assemble into mature fimbriae. EHEC O157:H7 contains two lpf loci, lpf1 and lpf2, whose predicted protein products share sequence identities ranging from 24 to 37% (36). However, neither of these operons corresponds exactly with the intact Salmonella lpf operon. The EHEC lpf1 operon has two smaller lpfC open reading frames (ORFs) compared to one large lpfC ORF in Salmonella, and the lpf2 operon has two smaller lpfB ORFs compared to one large ORF in Salmonella and lacks the lpfE gene. Despite the apparent lack of intact lpf operons, the cloned EHEC lpf1 and lpf2 loci still produced fimbriae in K-12, although the morphologies more closely resembled the E. coli type 1 fimbriae (lpf1) and a thin fibrilla-like structure (lpf2) rather than the long polar fimbriae of Salmonella (35, 36). Perhaps chaperone/usher proteins encoded in K-12 were utilized in assembly of these fimbriae. In contrast, the EPEC lpfE23 operon appears to be intact and is predicted to encode five proteins very similar in size to the Salmonella Lpf proteins (Table 1). The numbers of amino acid residues in the five predicted EPEC LpfE23 subunits are 178, 217, 859, 351, and 176 for LpfAE23 through LpfEE23 compared to 178, 232, 842, 355, and 175 residues for the Salmonella Lpf proteins. A phylogenetic analysis using CLUSTAL (15) of Lpf proteins described to date indicates that the EPEC proteins are more closely related to the Salmonella proteins than are most E. coli Lpf proteins (data not shown). No structural information is available on any of the Lpf proteins, so it is impossible to predict whether key residues essential for assembly and function are altered in the EPEC proteins. A second possibility for our failure to detect a phenotype for the EPEC lpfE23 locus is poor transcription or translation of the operon. Using standard prediction tools (2, 29, 30), more than one potential 70 promoter plus a reasonable Shine-Dalgarno sequence appear to be located upstream of the lpfAE23 gene. However, RT-PCR analysis of transcripts extracted from wild-type EPEC grown under in vitro conditions suitable for demonstrating adherence to epithelial cell lines showed only a very weak lpfAE23 transcript (Fig. 3). Expression of Lpf was not improved by growing the organism under a variety of conditions. Meanwhile, in vitro characterization of lpf genes in other A/E pathogenic E. coli strains has presented a mixed picture. As noted above, both EHEC O157:H7 lpf operons produced fimbriae in K-12. Mutation of lpf1 in the native EHEC O157:H7 resulted in slight, nonsignificant reductions in adherence to HeLa cells along with decreased microcolony formation (35). Mutation of the lpf2 locus in O157:H7 caused no detectable effect in adherence to HeLa cells but did slightly diminish adherence to Caco-2 cells at early time points (36). The REPEC strain 83/39 also contains two lpf loci, but mutation of one or both loci caused no alteration in adherence to CHO-K1 cells (26).
In any in vitro system used to study the contribution of potential adherence factors, the inability to detect a phenotype with an isogenic mutant or with the cloned adhesin genes in K-12 can always be attributed to cultural conditions or tissue culture cell lines that do not accurately reproduce the relevant in vivo situation. Since C. rodentium is a natural mouse pathogen, characterization of lpf in this model should provide further insight into the function of this putative adhesin. However, when tested in two different mouse strains, C3H/HeJ and C57BL/6, the lpfAcr mutant was indistinguishable from the wild-type parent with regard to fecal shedding, colonic hyperplasia, or numbers of organisms associated with intestinal tissue. As with in vitro adherence studies, the contribution of lpf genes in other A/E pathogenic E. coli strains in animal models is variable. REPEC strain 83/39 is a natural pathogen of rabbits, and Newton et al. (26) tested mutants in one or both lpf operons in this host species and found no significant differences in body weight or mortality rate between the wild type and mutants. However, animals receiving the wild type had more severe diarrhea and slightly better intestinal colonization in the early stages of disease. In addition, intimate bacterial adherence to the gut mucosa was indistinguishable in morphology, location, and number between the wild type and the lpf mutants of REPEC 83/39 (26). For EHEC O157:H7, the most important animal reservoir is cattle, although animals do not get sick from this organism. Jordan et al. (19) studied lpfA1 and lpfA2 mutants of O157:H7 in sheep, conventional pigs, and gnotobiotic piglets. In gnotobiotic piglets, a model of early colonization, the lpfA1 mutant caused fewer attaching/effacing lesions in the spiral colon than the wild-type parent. Using conventional pigs and sheep to study long-term colonization over 60 days, lpf mutants were recovered at lower levels than those for the parent strain (19). Thus, LPF appears to contribute to the long-term persistence, rather than virulence, of O157:H7 in species such as sheep (19), which have been implicated as an important reservoir of O157:H7 in the farm environment (27). Thus, although we failed to detect a role for LPF in EPEC or C. rodentium pathogenesis, experience with other A/E pathogenic E. coli strains indicates that LPF contributes to intestinal colonization and disease in some hosts but to a substantially lesser degree than does the intimin adhesin, which is essential for formation of the A/E lesion. Therefore, it appears that LPF is not an essential virulence factor but may be an accessory virulence determinant in some A/E pathogens.
lpf genes also have been identified in non-A/E pathogenic E. coli, although the genes have undergone considerable divergence (16). Mutation of the lpf genes in EHEC O113:H21, which is a more significant cause of hemolytic-uremic syndrome in Australia than O157:H7, resulted in a significant reduction in adherence to CHO-K1 cells (10). An lpf mutant of a meningitis-associated extraintestinal pathogenic E. coli strain was significantly reduced in adherence to a kidney cell line (HEK 293) (16). LPF may therefore play a more important role in disease caused by E. coli strains that lack intimin and the rest of the LEE than in disease caused by E. coli strains that possess intimin and the LEE, such as EPEC and C. rodentium.
ACKNOWLEDGMENTS
We thank Mary-Jane Lombardo, Laura Leverton, and Kristen J. Kanack for critical reading of the manuscript and Diana Gomez and Asaomi Kuwae for technical assistance and suggestions.
This study was supported by NIH grant numbers AI21657 and DK58957, the Naito Foundation, and The Wellcome Trust. A.G.T. was supported by institutional funds from the UTMB John Sealy Memorial Endowment Fund for Biomedical Research.
REFERENCES
1. Baldini, M. M., J. B. Kaper, M. M. Levine, D. C. Candy, and H. W. Moon. 1983. Plasmid-mediated adhesion in enteropathogenic Escherichia coli. J. Pediatr. Gastroenterol. Nutr. 2:534-538.
2. Barrick, D., K. Villanueba, J. Childs, R. Kalil, T. D. Schneider, C. E. Lawrence, L. Gold, and G. D. Stormo. 1994. Quantitative analysis of ribosome binding sites in E. coli. Nucleic Acids Res. 22:1287-1295.
3. Bartolome, B., Y. Jubete, E. Martinez, and F. de la Cruz. 1991. Construction and properties of a family of pACYC184-derived cloning vectors compatible with pBR322 and its derivatives. Gene 102:75-78.
4. Baumler, A. J., and F. Heffron. 1995. Identification and sequence analysis of lpfABCDE, a putative fimbrial operon of Salmonella typhimurium. J. Bacteriol. 177:2087-2097.
5. Baumler, A. J., R. M. Tsolis, and F. Heffron. 1996. The lpf fimbrial operon mediates adhesion of Salmonella typhimurium to murine Peyer's patches. Proc. Natl. Acad. Sci. USA 93:279-283.
6. Clarke, S. C., R. D. Haigh, P. P. Freestone, and P. H. Williams. 2003. Virulence of enteropathogenic Escherichia coli, a global pathogen. Clin. Microbiol. Rev. 16:365-378.
7. Deibel, C., S. Kramer, T. Chakraborty, and F. Ebel. 1998. EspE, a novel secreted protein of attaching and effacing bacteria, is directly translocated into infected host cells, where it appears as a tyrosine-phosphorylated 90 kDa protein. Mol. Microbiol. 28:463-474.
8. Deng, W., Y. Li, B. A. Vallance, and B. B. Finlay. 2001. Locus of enterocyte effacement from Citrobacter rodentium: sequence analysis and evidence for horizontal transfer among attaching and effacing pathogens. Infect. Immun. 69:6323-6335.
9. Donnenberg, M. S., C. O. Tacket, S. P. James, G. Losonsky, J. P. Nataro, S. S. Wasserman, J. B. Kaper, and M. M. Levine. 1993. Role of the eaeA gene in experimental enteropathogenic Escherichia coli infection. J. Clin. Investig. 92:1412-1417.
10. Doughty, S., J. Sloan, V. Bennett-Wood, M. Robertson, R. M. Robins-Browne, and E. L. Hartland. 2002. Identification of a novel fimbrial gene cluster related to long polar fimbriae in locus of enterocyte effacement-negative strains of enterohemorrhagic Escherichia coli. Infect. Immun. 70:6761-6769.
11. Duchaud, E., C. Rusniok, L. Frangeul, C. Buchrieser, A. Givaudan, S. Taourit, S. Bocs, C. Boursaux-Eude, M. Chandler, J. F. Charles, E. Dassa, R. Derose, S. Derzelle, G. Freyssinet, S. Gaudriault, C. Medigue, A. Lanois, K. Powell, P. Siguier, R. Vincent, V. Wingate, M. Zouine, P. Glaser, N. Boemare, A. Danchin, and F. Kunst. 2003. The genome sequence of the entomopathogenic bacterium Photorhabdus luminescens. Nat. Biotechnol. 21:1307-1313.
12. Frankel, G., A. D. Phillips, I. Rosenshine, G. Dougan, J. B. Kaper, and S. Knutton. 1998. Enteropathogenic and enterohaemorrhagic Escherichia coli: more subversive elements. Mol. Microbiol. 30:911-921.
13. Giron, J. A., A. S. Ho, and G. K. Schoolnik. 1991. An inducible bundle-forming pilus of enteropathogenic Escherichia coli. Science 254:710-713.
14. Hicks, S., G. Frankel, J. B. Kaper, G. Dougan, and A. D. Phillips. 1998. Role of intimin and bundle-forming pili in enteropathogenic Escherichia coli adhesion to pediatric intestinal tissue in vitro. Infect. Immun. 66:1570-1578.
15. Higgins, D. G., J. D. Thompson, and T. J. Gibson. 1996. Using CLUSTAL for multiple sequence alignments. Methods Enzymol. 266:383-402.
16. Ideses, D., D. Biran, U. Gophna, O. Levy-Nissenbaum, and E. Z. Ron. 2005. The lpf operon of invasive Escherichia coli. Int. J. Med. Microbiol. 295:227-236.
17. Jarvis, K. G., J. A. Giron, A. E. Jerse, T. K. McDaniel, M. S. Donnenberg, and J. B. Kaper. 1995. Enteropathogenic Escherichia coli contains a putative type III secretion system necessary for the export of proteins involved in attaching and effacing lesion formation. Proc. Natl. Acad. Sci. USA 92:7996-8000.
18. Jerse, A. E., J. Yu, B. D. Tall, and J. B. Kaper. 1990. A genetic locus of enteropathogenic Escherichia coli necessary for the production of attaching and effacing lesions on tissue culture cells. Proc. Natl. Acad. Sci. USA 87:7839-7843.
19. Jordan, D. M., N. Cornick, A. G. Torres, E. A. Dean-Nystrom, J. B. Kaper, and H. W. Moon. 2004. Long polar fimbriae contribute to colonization by Escherichia coli O157:H7 in vivo. Infect. Immun. 72:6168-6171.
20. Kenny, B., R. DeVinney, M. Stein, D. J. Reinscheid, E. A. Frey, and B. B. Finlay. 1997. Enteropathogenic E. coli (EPEC) transfers its receptor for intimate adherence into mammalian cells. Cell 91:511-520.
21. Knutton, S., T. Baldwin, P. H. Williams, and A. S. McNeish. 1989. Actin accumulation at sites of bacterial adhesion to tissue culture cells: basis of a new diagnostic test for enteropathogenic and enterohemorrhagic Escherichia coli. Infect. Immun. 57:1290-1298.
22. Levine, M. M., J. P. Nataro, H. Karch, M. M. Baldini, J. B. Kaper, R. E. Black, M. L. Clements, and A. D. O'Brien. 1985. The diarrheal response of humans to some classic serotypes of enteropathogenic Escherichia coli is dependent on a plasmid encoding an enteroadhesiveness factor. J. Infect. Dis. 152:550-559.
23. McDaniel, T. K., K. G. Jarvis, M. S. Donnenberg, and J. B. Kaper. 1995. A genetic locus of enterocyte effacement conserved among diverse enterobacterial pathogens. Proc. Natl. Acad. Sci. USA 92:1664-1668.
24. McDaniel, T. K., and J. B. Kaper. 1997. A cloned pathogenicity island from enteropathogenic Escherichia coli confers the attaching and effacing phenotype on E. coli K-12. Mol. Microbiol. 23:399-407.
25. Metcalf, W. W., W. Jiang, L. L. Daniels, S. K. Kim, A. Haldimann, and B. L. Wanner. 1996. Conditionally replicative and conjugative plasmids carrying lacZ alpha for cloning, mutagenesis, and allele replacement in bacteria. Plasmid 35:1-13.
26. Newton, H. J., J. Sloan, V. Bennett-Wood, L. M. Adams, R. M. Robins-Browne, and E. L. Hartland. 2004. Contribution of long polar fimbriae to the virulence of rabbit-specific enteropathogenic Escherichia coli. Infect. Immun. 72:1230-1239.
27. Ogden, I. D., M. MacRae, and N. J. Strachan. 2005. Concentration and prevalence of Escherichia coli O157 in sheep faeces at pasture in Scotland. J. Appl. Microbiol. 98:646-651.
28. Paulozzi, L. J., K. E. Johnson, L. M. Kamahele, C. R. Clausen, L. W. Riley, and S. D. Helgerson. 1986. Diarrhea associated with adherent enteropathogenic Escherichia coli in an infant and toddler center, Seattle, Washington. Pediatrics 77:296-300.
29. Reese, M. G. 2001. Application of a time-delay neural network to promoter annotation in the Drosophila melanogaster genome. Comput. Chem. 26:51-56.
30. Reese, M. G. 2000. Computational prediction of gene structure and regulation in the genome of Drosophila melanogaster. Ph.D thesis. University of California—Berkeley/University of Hohenheim, Stuttgart, Germany.
31. Schauer, D. B., and S. Falkow. 1993. Attaching and effacing locus of a Citrobacter freundii biotype that causes transmissible murine colonic hyperplasia. Infect. Immun. 61:2486-2492.
32. Tacket, C. O., M. B. Sztein, G. Losonsky, A. Abe, B. B. Finlay, B. P. McNamara, G. T. Fantry, S. P. James, J. P. Nataro, M. M. Levine, and M. S. Donnenberg. 2000. Role of EspB in experimental human enteropathogenic Escherichia coli infection. Infect. Immun. 68:3689-3695.
33. Tamano, K., S. Aizawa, E. Katayama, T. Nonaka, S. Imajoh-Ohmi, A. Kuwae, S. Nagai, and C. Sasakawa. 2000. Supramolecular structure of the Shigella type III secretion machinery: the needle part is changeable in length and essential for delivery of effectors. EMBO J. 19:3876-3887.
34. Tatsuno, I., H. Kimura, A. Okutani, K. Kanamaru, H. Abe, S. Nagai, K. Makino, H. Shinagawa, M. Yoshida, K. Sato, J. Nakamoto, T. Tobe, and C. Sasakawa. 2000. Isolation and characterization of mini-Tn5Km2 insertion mutants of enterohemorrhagic Escherichia coli O157:H7 deficient in adherence to Caco-2 cells. Infect. Immun. 68:5943-5952.
35. Torres, A. G., J. A. Giron, N. T. Perna, V. Burland, F. R. Blattner, F. Avelino-Flores, and J. B. Kaper. 2002. Identification and characterization of lpfABCC'DE, a fimbrial operon of enterohemorrhagic Escherichia coli O157:H7. Infect. Immun. 70:5416-5427.
36. Torres, A. G., K. J. Kanack, C. B. Tutt, V. Popov, and J. B. Kaper. 2004. Characterization of the second long polar (LP) fimbriae of Escherichia coli O157:H7 and distribution of LP fimbriae in other pathogenic E. coli strains. FEMS Microbiol. Lett. 238:333-344.
37. Vallance, B. A., W. Deng, K. Jacobson, and B. B. Finlay. 2003. Host susceptibility to the attaching and effacing bacterial pathogen Citrobacter rodentium. Infect. Immun. 71:3443-3453.
38. Wang, R. F., and S. R. Kushner. 1991. Construction of versatile low-copy-number vectors for cloning, sequencing and gene expression in Escherichia coli. Gene 100:195-199.(Ichiro Tatsuno, Rosanna M)
Department of Microbiology and Immunology, University of Texas Medical Branch, Galveston, Texas 77555
Centre for Molecular Microbiology and Infection, Division of Cell and Molecular Biology, Imperial College London, London SW7 2AZ
Centre for Paediatric Gastroenterology, Royal Free Hospital, London NW3 2QG, United Kingdom
ABSTRACT
Using the enteropathogenic Escherichia coli (EPEC) genome sequence, we found that EPEC E2348/69 has an lpfABCDE gene cluster homologous (about 60% identical at the protein level) to the Salmonella long polar fimbria (LPF) operon. To determine whether this operon is essential for adherence, the lpfABCDE23 genes were deleted from EPEC strain E2348/69 by allelic exchange. Analysis of the resulting EPEClpfABCDE23 strain showed no change in adherence to HeLa cells or to human intestinal biopsy cells in the in vitro organ culture (IVOC) system compared to the wild type. Sera from volunteers experimentally infected with E2348/69 showed no antibody response to the major subunit protein, LpfA. These results suggested that the lpfE23 gene cluster is not necessary for EPEC adherence and attaching/effacing (A/E) lesion formation on human biopsy samples and is not expressed during human infection. We also identified an lpf gene cluster in Citrobacter rodentium strain ICC168 (lpfcr). A lpfAcr mutant of ICC168 retained wild-type adherence and A/E lesion-forming activity on HeLa cells. C3H/HeJ mice were infected with a wild-type C. rodentium strain and its lpfAcr isogenic mutant. Both strains were recovered at high levels in stools, and there were no significant differences between the groups both in terms of the number of CFU/organ (colon and cecum) and in terms of the amount of hyperplasia, as measured by weight. Similar results were observed in a second mouse strain, C57BL/6. These data suggest that in addition to playing no apparent role in EPEC pathogenesis, lpfcr is not required for C. rodentium virulence in either the C3H/HeJ or C57BL/6 mouse model.
INTRODUCTION
Enterohemorrhagic Escherichia coli (EHEC), enteropathogenic E. coli (EPEC), rabbit-specific enteropathogenic E. coli (REPEC), and Citrobacter rodentium are attaching/effacing (A/E) bacterial pathogens that attach to host intestinal epithelium and efface brush border microvilli, forming pathognomonic A/E lesions (6, 12). This characteristic adherence pattern is encoded by the locus of enterocyte effacement (LEE) pathogenicity island (23), which contains genes encoding the outer membrane protein intimin (18), its receptor Tir/EspE (7, 20), and a type III secretion system (17). EPEC, the first A/E pathogen to be described, also has a type IV pilus, bundle-forming pilus (BFP), as a potential adherence factor (13). While EPEC is a human pathogen, C. rodentium is a natural pathogen of mice and now is used as an animal model for studying virulence-associated factors of EPEC or EHEC (8, 31).
The long polar fimbriae (LPF), another potential adherence factor that was originally described in Salmonella, have been identified in other A/E bacteria, such as EHEC and REPEC. The LPF of Salmonella enterica serovar Typhimurium were shown to direct the attachment of bacteria to murine Peyer's patches in vivo (4, 5). Two nonidentical lpf loci are present in EHEC O157:H7 (35, 36), and one or both of them were found to be involved in microcolony formation, increased adherence to HEp-2 cells when cloned into a nonfimbriated K-12 strain, and the colonization by and persistence of E. coli O157:H7 in swine and sheep (19). The lpf141 locus of REPEC O15:H- strain 83/39 appeared to play a role in the early stages of REPEC infection of rabbits and was essential for the development of severe diarrhea (26). Recently, sequencing of the EPEC O127:H6 strain E2348/69 and C. rodentium ICC168 genomes revealed lpf loci in each of those pathogens (http://www.sanger.ac.uk/Projects/Escherichia_Shigella/ and http://www.sanger.ac.uk/Projects/C_rodentium, respectively), but it is unclear whether these loci are also functional in contributing to disease caused by these A/E pathogens. We therefore investigated the potential role of the EPEC and C. rodentium LPF in adherence and/or virulence.
MATERIALS AND METHODS
Strains and plasmids. E. coli E2348/69 (O127:H6) is a prototypic EPEC strain proven to cause diarrhea in volunteers (22). JPN15 is a derivative of E2348/69 cured of the EPEC adherence factor plasmid. E. coli HB101(pCVD462) (also known as pLEE) contains the cloned LEE pathogenicity island and confers A/E on HEp-2 and Caco-2 cells (24). pMAR7 is a Tn801 insertion derivative of the EPEC adherence factor plasmid from EPEC strain E2348/69 (O127:H6) (1). C. rodentium ICC168 is a wild-type strain shown to cause colonic epithelial cell hyperplasia in mice (8, 31). A reference collection of representative diarrheagenic E. coli clones known as the diarrheagenic E. coli (DEC) collection was screened for the presence of lpf genes. This collection is available from the STEC Center at Michigan State University (http://www.shigatox.net). Two of these strains, 572-56 and 3787-62, representative of the DEC1 and DEC2 clones, respectively, were further analyzed by mutating the lpf genes and testing for resulting changes in adherence as described below.
Cloning of the EPEC lpf gene cluster. The lpf gene cluster of EPEC strain E2348/69 was amplified by PCR with Easy-A high-fidelity PCR cloning enzyme (Stratagene) using specific primers lpf-Fto (5'-CTCGGTTTCGGTATCGGCAGC-3') and lpf-Rto (5'-CCAGCTGCGGCGGCCAAACA-3'). The 6,484-bp amplicon was cloned into pGEM-T easy (Promega) to yield pELP102, whose insert was sequenced. The ApaI-SacI fragment containing the amplified insert DNA of pELP102 was subcloned into the low-copy-number vector pWKS130 (38) to yield pELP130. A 1.1-kb EcoRI fragment containing the lpfA gene of pELP102 was subcloned into pSU18 (3) to yield plpfA2348F and plpfA2348R, in which the lpfA gene is oriented in the same and opposite direction, respectively, as the lacUV5 promoter.
Bacterial adhesion to cultured epithelial cells. Adherence of bacterial strains to cultured Hep-2, HeLa, T84, and Caco-2 epithelial cells was evaluated as previously described (35). Briefly, bacteria were grown statically in Luria-Bertani (LB) broth overnight at 37°C and inoculated at a multiplicity of infection of approximately 50:1 onto semiconfluent cultured epithelial cell monolayers grown on 24-well microtiter plates in Dulbecco's modified Eagle's medium (DMEM) with 10% fetal calf serum. Prior to infection, fresh DMEM without supplements such as mannose was added to cells. Bacteria and cells were incubated for 3 h and washed five times with 1 ml of sterile phosphate-buffered saline (PBS). The monolayers were fixed and stained with Giemsa's solution for microscopic evaluation. To quantify E. coli adherence, the infected monolayers were washed five times with PBS, and the adherent bacteria were recovered with 200 μl of 0.1% Triton X-100 in PBS buffer and plated on Luria agar plates containing the appropriate antibiotic. Data are expressed as the percentages of the bacterial inocula recovered from triplicate wells and represent at least three separate experiments. This protocol was used for all adherence assays unless otherwise described. The fluorescence actin staining (FAS) test was performed as previously described (21).
Organ culture adhesion assay. In vitro organ culture (IVOC) was performed as described previously (14), with ethical approval and informed consent, using small intestinal samples. Large-bowel mucosa was not used. The assay was terminated at 2 and 4 h on three occasions to determine "early" events. "Later" events were observed at 6 h on two occasions, and all other experiments were maintained for 8 h. Tissue culture medium was changed every 2 h, although some medium plus bacterial inoculum would remain within the foam insert supporting the tissue sample and in association with the tissue by surface tension. Each bacterial strain was examined in IVOC on at least three occasions with tissue from different children. Strain E2348/69 and an uninoculated specimen were included with each experimental culture to act as positive and negative controls, respectively.
Tissue processing. After the specified culture period, specimens were thoroughly washed three times to remove any nonadherent bacteria and then prepared for scanning electron microscopy (SEM) as described previously (14). Following fixation and dehydration, specimens were placed in 100% ethanol and critical point dried in liquid CO2 with an Emitech K850 critical point drying apparatus. Samples were then mounted on aluminum stubs and sputter coated with gold-palladium using a Polaron sputter coating apparatus. Specimens were observed with a JEOL JSM-5300 SEM at an accelerating voltage of 30 kV.
Construction of lpf mutants. In-frame lpf deletion mutants were constructed in three EPEC strains and one C. rodentium strain. The wild-type strains and the resulting mutant designations, in parentheses, were E2348/69 (lpfABCDE23), 572-56 (lpfABCD572), 3787-62 (lpfABCD3787), and C. rodentium (lpfABCDcr). Bacteria were cultured at 30°C for all constructions. A set of oligonucleotides, primer lpfD-F (5'-GGTCCGATCGGTACCCAACAACATC-3') and primer lpfA-R (5'-CAACCTGAGCGGTACCATAACC-3'), were used as described above to amplify DNA fragments of 5.0 kb (corresponding to the in-frame lpfABCDE23 deletion mutation and additional flanking regions from pELP102) by PCR. The corresponding oligonucleotides contained restriction sites for the endonuclease KpnI (shown in boldface in the primer sequences). This enzyme was used for self-ligation of the amplification product. The resulting 2.0-kb ApaI-SacI fragment (containing the in-frame lpfABCDE23 deletion) was subcloned into the plasmid pWM91 (25) to yield a recombinant suicide plasmid that was introduced by conjugation into EPEC strain E2348/69 and integrated into the chromosome. Further cultivation in the presence of 5% sucrose led to a second round of recombination and the excision of the plasmid with the wild-type lpfABCDE23 gene. Bacteria were confirmed to harbor a mutated lpfABCDE23 allele by PCR using the two primers lpf-F (5'-TTTGGGGCGCGATTGTACTGG-3') and lpf-R (5'-TTCACCAATAATGAAAACGAC-3'), which are specific to sequences outside of those for primers lpf-Fto and lpf-Rto, respectively. lpfABCD572 and lpfABCD3787 deletion mutants were constructed in the same manner as the lpfABCDE23 deletion mutant of EPEC E2348/69, except that 3787-62 (lpfABCD3787) was confirmed to harbor a mutated lpfABCD3787 allele by PCR using primers lpf-F2 (5'-TTACTGTTTCCCGTCACTGG-3') and lpf-R2 (5'-GAAGGTCATGTAATAGCTGGC-3'), which are also specific to sequences outside of those for primers lpf-Fto and lpf-Rto, respectively.
The lpfcr gene cluster of C. rodentium strain ICC168 was amplified by PCR with Easy-A high-fidelity PCR cloning enzyme (Stratagene) using primers CRlpfA-F (5'-AGATGATTTTCTCTGTTTTC-3') and CRlpfA-R (5'-GCGAGGCGCTGTCGCCCAGC-3'). The 3,298-bp amplicon was cloned into pGEM-T easy (Promega) to yield pCRlpfA. An in-frame deletion in lpfAcr was made using an inverse PCR approach with pCRlpfA as the template and a set of oligonucleotides, primer CRlpfA-F (5'-GCTTATGTCGGGTACCGCGTTCCG-3') and primer CRlpfA-R (5'-GTATCGAGGAGGTACCCGACAAAATTC-3'). The primers contain KpnI restriction sites (in boldface). The 5.9-kb product was cut with KpnI and religated, yielding the lpfAcr construct. A 2.9-kb SphI-PstI fragment containing lpfAcr was subcloned into the suicide plasmid pCACTUS (33). The lpfAcr suicide plasmid was introduced by electroporation into strain ICC168 and integrated into the chromosome by growth at 42°C. Plating in the presence of 5% sucrose led to the isolation of recombinants carrying the lpfAcr allele. Bacteria were confirmed to harbor the lpfAcr allele by PCR using the two outside primers CRlpfAout-F (5'-CCGAGGCGATGCCTAACGCC-3') and CRlpfAout-R (5'-GCTGTCAAAGATATCTGCGG-3'), which flank the region used for homologous recombination.
GST-LpfAE23 expression. The lpfAE23 gene was amplified by PCR using the primers GST-lpfA (5'-CGCGGATCCGCTGATTCTGGTGATGGTAC-3') and 3572 (5'-ACCATACCGGACTTTTCAAT-3') and cloned into pGEX-2T (Promega) to yield pGST-lpfAE23. The corresponding oligonucleotide GST-lpfAE23 contained a restriction site for the endonuclease BamHI (shown in boldface in the primer sequence). To induce the GST-lpfAE23 fusion protein, 0.1 mM isopropyl--D-thiogalactopyranoside (IPTG) was added to a logarithmic-phase culture of DH5/pGST-lpfA and shaken for 2 h at 30°C. The GST-LpfAE23 fusion protein was purified by glutathione Sepharose 4B (Amersham Biosciences) and digested with thrombin protease per the manufacturer's instructions (Amersham Biosciences). The resultant proteins were analyzed as described previously (34). Briefly, GST-LpfAE23 was eluted, without thrombin treatment, by boiling with an equal volume of 2x sodium dodecyl sulfate (SDS) loading buffer. To obtain free LpfAE23, after incubation in PBS with thrombin protease, the supernatant and pellet beads were combined with an equal volume of 2x SDS loading buffer (the ratio of the supernatant to the pellet beads used was 1:1 [vol/vol]). Five microliters each of the prepared supernatant and pellet samples was resolved by SDS-12% polyacrylamide gel electrophoresis and stained with Coomassie brilliant blue. Antiserum against LpfAE23 was prepared by immunizing rabbits with the resulting purified LpfAE23 (Lampire Biological Laboratories Inc.).
Electron microscopy. Bacteria were grown at 37°C under various conditions including different media (DMEM, LB, MacConkey, and colonization factor antigen; broth versus agar), different incubation times (2, 3, 5 h, or overnight), and with or without HEp-2 cells to examine the presence of fimbriae using electron microscopy. The bacteria were recovered and allowed to adhere to a carbon-formvar-coated 300-mesh copper grid (Electron Sciences). The fimbriae were visualized by negative staining with 0.1% phosphotungstic acid, pH 7.4, and the grids were examined in a Philips CM 120 electron microscope (FEI Co.). In order to confirm that the sample was properly prepared, typical type 1 fimbriae of DH5 grown on LB agar were observed as a positive control.
RT-PCR. One milliliter of EPEC E2348/69 overnight culture in LB was inoculated into 10 ml of DMEM (Catalog No. 11995-065; Invitrogen) and statically incubated for 2 h in 37°C, 5% CO2. Total RNA was extracted by RNeasy Starter kit (QIAGEN). First-strand cDNA was synthesized by using a SuperScript First-Strand synthesis system for reverse transcription-PCR (RT-PCR) (Invitrogen) and the gene-specific primers BamHItirMR (5'-GCTGAGGCCGGATCCAATACCCG-3') for tir and 3572 for lpfAE23. The cDNA for tir and lpfAE23 was amplified by using a combination of primers SwaItirF (5'-CGTGTCAAATATTTTAAATAAAAG-3') and BamHItirMR as well as lpfARBS-F (5'-GGAGCTCGAGATGAAAAAGG-3') and lpfA-R, respectively. The resultant PCR products were loaded into 0.7% agarose with a 1-kb Plus DNA ladder (Invitrogen) and analyzed.
One milliliter of an overnight culture of C. rodentium in LB was inoculated into 10 ml of DMEM (Catalog No. 11995-065; Invitrogen) and statically incubated for 2 h at 37°C, 5% CO2. Total RNA was extracted using the RNeasy Starter kit (QIAGEN). First-strand cDNA was synthesized by using a SuperScript First-Strand synthesis system for RT-PCR (Invitrogen) and a gene-specific primer, CRlpfA-R. The cDNA for lpfABCcr was amplified using primers CRlpfA-F and CRlpfA-R. PCR products were loaded into 0.7% agarose with a 1-kb Plus DNA ladder (Invitrogen) and analyzed.
Human immune responses. A serological response against LpfA during the course of human infection was evaluated using convalescent human serum obtained from volunteers experimentally infected with E2348/69 in previous studies (9, 22, 32). The GST-LpfAE23 fusion protein was prepared and transferred to polyvinylidene difluoride membrane as described above, and immunoblots were performed as previously described (22) using horseradish peroxidase-conjugated goat antibody to human immunoglobulin G.
Mouse experiments. Male, specific-pathogen-free C3H/HeJ mice (6 to 8 weeks old) were purchased from Harlan Olac (Bicester, United Kingdom), and female, specific-pathogen-free C57BL/6 (6 to 8 weeks old) were purchased from B and K (Hull, United Kingdom). Mice were housed in individual ventilated cages with free access to food and water. Unanesthetized mice were orally gavaged with 200 μl of bacterial suspension using a gavage needle. The viable count of the inoculum was determined by retrospective plating on LB agar containing nalidixic acid to select for nalidixic acid-resistant C. rodentium.
In the infection experiments, bacteria were grown to stationary phase in LB broth plus nalidixic acid. For the C3H/HeJ mice, the bacteria were pelleted by centrifugation, washed, and then diluted 1:10 with PBS and gavaged into mice (approximately 1 x 107 CFU per mouse). The C57BL/6 mice required a higher inoculum, so 10 ml of stationary-phase bacteria was pelleted by centrifugation and resuspended in 1 ml of PBS and then gavaged into mice (approximately 1 x 109 CFU per mouse). Each strain was tested in groups of at least four mice per bacterial strain. Stool samples were recovered aseptically at various time points after inoculation, and the numbers of viable bacteria per gram of stool were determined by plating onto LB agar containing appropriate antibiotics. Mice were killed at either the peak of infection, 11 days postchallenge (C3H/HeJ mice), or once the infection had been cleared, 28 days postchallenge (C57BL/6 mice). In all cases, the distal 8 cm of colon was aseptically removed and weighed after removal of fecal pellets. Colons were then homogenized mechanically using a Seward 80 stomacher (London, United Kingdom), and the numbers of viable bacteria per colon were determined by plating onto LB agar containing appropriate antibiotics.
RESULTS
Identification and cloning of EPEC lpfE23 genes. The complete genome sequence of E2348/69 available at http://www.sanger.ac.uk/Projects/Escherichia_Shigella/ revealed a locus whose predicted protein products showed homology to the LPF proteins of S. enterica serovar Typhimurium. The organization of the EPEC lpfE23 locus and the similarity of the predicted EPEC LPF proteins to LPF proteins of other pathogens are shown in Fig. 1 and Table 1. The percent G+C composition of the EPEC lpf locus is 43.1, substantially lower than that of the E. coli K-12 genome (50.8%), suggesting that this locus was introduced by horizontal gene transfer. The lpf locus was inserted between the yhiW and yhiJ genes corresponding to minute 79.9 of the E. coli K-12 chromosome (http://www.uni-giessen.de/gx1052/ECDC/ecdc/ecmap75.htm). The same insertion site is seen with the lpf operons of S. enterica serovar Typhimurium, EHEC O157:H7 lpf1, and REPEC (26, 35). The EPEC lpfE23 gene cluster was amplified by PCR, and the amplicon was cloned to yield pELP102 (see Materials and Methods). The resulting cloned amplicon was sequenced, and the insert DNA sequence was identical to that determined in the genomic sequencing.
Adherence phenotype of an EPEC lpf mutant to IVOC. In order to address whether LPFE23 is associated with adherence of EPEC to human small intestinal tissue, an EPEClpfABCDE23 mutant was constructed and tested in IVOC. As shown in Fig. 2, an EPEClpfABCDE23 mutant strain adhered to IVOC in a manner similar to that seen with the wild type, indicating that EPEC LPFE23 is not essential for adherence to human intestinal tissue in the IVOC system.
Effect of lpfE23 on bacterial adherence to cultured epithelial cells. EPEC LPFE23 was also characterized for adherence to cultured epithelial cell lines. When EPEClpfABCDE23 was investigated for the ability to adhere to Caco-2 cells, it showed an adherence phenotype similar to that of parental wild-type EPEC (data not shown). Similar results were also seen in HEp-2 and T-84 cells under various experimental conditions, including different growth phase (log or stationary), static or shaking culture conditions, and growth in LB and DMEM (data not shown). To address the possibility that the function of LPFE23 is masked by other EPEC adherence factors, we investigated whether the cloned lpfE23 confers adherence to E. coli K-12. When the lpfE23 gene cluster was cloned in a high-copy vector, HB101(pELP102) did not show increased adherence to T-84 cells (4.7% ± 2.5%, CFU of output/input) compared to the parental HB101 strain containing the control vector only, which demonstrates minimal adherence (7.5% ± 3.8%). As previously observed, HB101(pMAR7) expressing BFP showed increased adherence to cultured epithelial cell lines (20.5% ± 6.0%). EPEClpfABCDE23 and E. coli K-12 containing the cloned lpfE23 genes on high-copy-number (pELP102) or low-copy-number (pELP130) cloning vectors were further tested using combinations of the following experimental conditions: different epithelial cells (HEp-2, Caco-2, T-84, and HeLa), different hosts for cloned lpfE23 [DH5 and HB101(pCVD462)], different multiplicities of infection (1 to 200), different bacterial culture conditions (growth phase with log or stationary, shaking versus static, DMEM versus LB), and different infection times (2 to 5 h). Regardless of the conditions tested, no increased adherence activity was detected due to the presence of the lpfE23 locus (data not shown).
Because no apparent role was observed for LPFE23 in the adherence of EPEC strain E2348/69 to four epithelial cell lines or human intestinal tissue, we mutated the lpfE23 genes in two other EPEC strains to determine whether E2348/69 was unusual in this regard. We first examined the presence of lpfE23 genes in the DEC collection of diarrheagenic E. coli clones and found lpfE23 sequences in representatives of the EPEC1, EHEC1, and EHEC2 groups (Table 2). The amplified fragments in these clones were the same size, 6.5 kb, as that seen with E2348/69. Strains of the EPEC2 group gave a mixed pattern, with two representative strains of DEC11 clones yielding a 6.5-kb fragment and two strains of the DEC12 group yielding a 3.0-kb fragment. We generated lpf mutants of two additional strains from the EPEC1 group, lpfABCD527 and lpfABCD3787, and tested them for adherence to HeLa cells. No differences were seen in the adherence of these mutants from that of the parent strains (data not shown).
Expression of lpfE23 genes in vitro. To investigate why the adherence phenotype associated with other LPF was not detected with EPEC LPFE23, the expression of the lpfE23 genes was analyzed by electron microscopy. LPF from three enteric pathogens, S. enterica serovar Typhimurium, EHEC, and REPEC, have previously been visualized using negative staining and electron microscopy (4, 26, 36). In each of these cases, visualizing LPF was much easier with the cloned lpf genes expressed in a nonfimbriated K-12 strain than in the wild-type strain. We used these same techniques to examine wild-type E2348/69 for LPF without success. When pELP102 containing the EPEC E2348/69 lpfE23 genes was transformed into nonfimbriated E. coli strain ORN172, we also failed to observe fimbrial structures (data not shown). We examined five different EPEC strains containing lpfE23 genes grown under a variety of different growth conditions but still did not observe fimbrial structures (data not shown).
The failure to visualize EPEC LPFE23 is consistent with our inability to detect a difference in adherence to epithelial cell lines or to IVOC. We therefore examined expression of LPF by protein and mRNA analysis. To generate antisera against LPFE23, the lpfAE23 gene was cloned into the expression vector pGEX-2T, which creates a GST-LpfAE23 fusion under the transcriptional control of the Ptac promoter contained in this vector. The addition of IPTG induced the expression of a protein with the expected size of GST-LpfAE23, and digestion of this protein with thrombin protease resulted in a protein with the expected size of LpfAE23 (data not shown). This protein was purified and used to raise rabbit antisera against LpfAE23 as described in Materials and Methods. However, when the lpfAE23 gene was cloned into a standard plasmid vector under its native promoter, no expression of LpfAE23 protein was detected by immunoblotting with anti-LpfAE23 antibody (data not shown). Two different K-12 strains (ORN172 and DH5) containing the cloned lpfE23 locus were grown under a variety of different expression conditions, including one in which LpfA of O157:H7 was detected with specific antibody (35), but no LpfAE23 could be detected by immunoblotting. The pSU18 vector contains a lacUV5 promoter, but expression of clones in the same or opposite orientation of this promoter with or without IPTG still yielded no LpfAE23 expression. When the EPEC E2348/69 wild-type strain was incubated at 37°C in DMEM, which are standard conditions for expression of EPEC adherence factors, only a faint band of lpfAE23 mRNA was detected by RT-PCR at a substantially lower level than was observed with tir mRNA in the same experiment (Fig. 3). These results indicate that the lpfAE23 gene, encoding the major fimbrial subunit, is poorly expressed, if at all, under the experimental conditions used in our adherence and electron microscopy studies.
Expression of lpfE23 genes in vivo. To investigate whether there was a serological response against LPFE23 during the course of human infection, convalescent-phase sera of 10 volunteers previously infected with E. coli strain E2348/69 (9, 22, 32) were tested against GST-LpfAE23 fusion protein in immunoblots. None of the volunteers showed an immune response to LpfAE23 in the convalescent-phase sera obtained 28 days after infection other than a faint, nonspecific band that was unchanged in intensity between preinfection and postinfection sera (data not shown).
Identification of C. rodentium lpfcr genes. Another lpf operon of an A/E organism was identified in the genome sequence of C. rodentium strain ICC168 (http://www.sanger.ac.uk/Projects/C_rodentium/). The predicted LpfAcr, LpfBcr, and LpfCcr proteins have amino acid similarities ranging from 32 to 46% to Lpf proteins in EPEC, REPEC, E. coli O157, and Salmonella (Fig. 1B, Table 1). No protein with striking homology to LpfD was seen in C. rodentium, but a predicted protein of size comparable to LpfD was encoded in the same location downstream of LpfCcr (Fig. 1C). LpfDcr does share 28% identity with a fimbrial adhesin precursor of Photorhabdus luminescens subsp. laumondii TTO1 (11). Additionally, a fimbrial domain was detected in the C terminus of LpfDcr using the "Conserved Domain Search" program (http://www.ncbi.nlm.nih.gov). LpfEcr also did not show high homology with other LpfE proteins (with a maximum of 22% identity seen with LpfE of E. coli O157 and Salmonella). However, this is consistent with LpfD and LpfE being poorly conserved among the lpf loci of O157:H7, EPEC, and REPEC (26, 35, 36). The G+C composition of the lpfcr locus is 54%, which is slightly lower than that of the 9 kb flanking it to one side and the 70 kb to the other side (both 57%). The sequences flanking lpfcr did not show high homology to E. coli K-12, indicating that the lpfcr gene cluster of C. rodentium strain ICC168 does not map to the same chromosomal location as the lpf operon of S. enterica serovar Typhimurium, EPEC, EHEC O157:H7, and REPEC (26, 35).
Virulence of the C. rodentium lpfAcr mutant in the C3H/HeJ mouse model. To address whether LPF is associated with the virulence of C. rodentium in mice, a ICC168lpfAcr mutant was constructed. In initial in vitro studies, lpfAcr had no effect on the ability of ICC168 to adhere to HeLa cells or the ability to produce a positive fluorescent actin staining (FAS) test indicative of A/E lesion formation (data not shown). Thus, LpfAcr is not required for in vitro adherence and A/E lesion formation on HeLa cells. To examine potential effects of lpfAcr on virulence, groups of four mice were inoculated with the wild-type C. rodentium strain and the lpfAcr mutant. The lpfAcr mutant behaved like the wild type. Both strains were recovered at high levels in stools collected at days 1, 4, 6, and 8 (Fig. 4A). At day 10, two mice infected with the wild-type strain and two mice infected with the lpfAcr mutant died. The remaining two mice in each group were sacrificed on day 11, and the colon and cecum were removed. The numbers of viable cells in the colon were similar in both groups, as was the extent of hyperplasia as measured by weight (data not shown).
Virulence of the C. rodentium lpfcr mutant in the C57BL/6 mouse model. The experimental infection was repeated in C57BL/6 mice, which are less susceptible to infection than C3H/HeJ mice when infected with C. rodentium. The milder response of BL/6 allows the length of infection to be extended (37), thereby providing an opportunity to detect a more subtle defect in virulence that may go undetected in the C3H/HeJ mice. Figure 4B shows the numbers of CFU/gram of stool shed by mice infected with the wild type and the lpfAcr mutant over 28 days. Colons were taken at day 28, and at this time point all the citrobacteria had been cleared from both the colon and the cecum. The average weights of colons ± standard errors are the following: wild-type-infected mice, 0.224 ± 0.014 g; lpfAcr-infected mice, 0.242 ± 0.009 g; and uninfected mice, 0.147 ± 0.008 g. The C57BL/6 mice do not exhibit as much hyperplasia as the C3H/HeJ mice in response to infection; however, there was significant hyperplasia in the distal 1 to 2 cm of the colon in mice infected with both the wild type and the lpfAcr mutant. Collectively, these data suggest that lpfcr is not required for virulence in our mouse models.
Expression of lpfcr genes in vitro. The expression of lpfcr was examined to confirm that these genes were expressed. When the C. rodentium wild-type strain was incubated at 37°C in DMEM, no expression of LpfAcr protein was detected in whole-cell extracts by immunoblotting with anti-EPEC LpfA antibody. Since the proteins share low similarity, it is possible that the anti-EPEC LpfA sera did not detect LpfAcr. Nevertheless, lpfABCcr mRNA was detected by RT-PCR using primers specific to lpfAcr and lpfC (Fig. 5A) but not when the 5' primer was outside the lpfAcr coding region (Fig. 5B). Thus, lpfcr appears to be at least transcribed under these conditions. Furthermore, this result suggests that lpfABCcr comprises an operon.
DISCUSSION
Our in vitro studies did not indicate a role for LPFE23 in adherence or A/E lesion formation by EPEC. Similarly, mutation of C. rodentium lpfcr did not alter the in vitro adherence of this organism to cultured epithelial cells. In addition, we examined expression of LpfAE23 using immunoblots of E2348/69 and K-12 containing cloned lpfE23 genes grown under a variety of cultural conditions and found no evidence of expression using antisera raised against the cloned EPEC lpfAE23 gene product. There are at least two possible reasons why LPFE23 is apparently not expressed in EPEC. The first possibility is that some of the LpfABCDEE23 proteins are inactive and therefore cannot assemble into mature fimbriae. EHEC O157:H7 contains two lpf loci, lpf1 and lpf2, whose predicted protein products share sequence identities ranging from 24 to 37% (36). However, neither of these operons corresponds exactly with the intact Salmonella lpf operon. The EHEC lpf1 operon has two smaller lpfC open reading frames (ORFs) compared to one large lpfC ORF in Salmonella, and the lpf2 operon has two smaller lpfB ORFs compared to one large ORF in Salmonella and lacks the lpfE gene. Despite the apparent lack of intact lpf operons, the cloned EHEC lpf1 and lpf2 loci still produced fimbriae in K-12, although the morphologies more closely resembled the E. coli type 1 fimbriae (lpf1) and a thin fibrilla-like structure (lpf2) rather than the long polar fimbriae of Salmonella (35, 36). Perhaps chaperone/usher proteins encoded in K-12 were utilized in assembly of these fimbriae. In contrast, the EPEC lpfE23 operon appears to be intact and is predicted to encode five proteins very similar in size to the Salmonella Lpf proteins (Table 1). The numbers of amino acid residues in the five predicted EPEC LpfE23 subunits are 178, 217, 859, 351, and 176 for LpfAE23 through LpfEE23 compared to 178, 232, 842, 355, and 175 residues for the Salmonella Lpf proteins. A phylogenetic analysis using CLUSTAL (15) of Lpf proteins described to date indicates that the EPEC proteins are more closely related to the Salmonella proteins than are most E. coli Lpf proteins (data not shown). No structural information is available on any of the Lpf proteins, so it is impossible to predict whether key residues essential for assembly and function are altered in the EPEC proteins. A second possibility for our failure to detect a phenotype for the EPEC lpfE23 locus is poor transcription or translation of the operon. Using standard prediction tools (2, 29, 30), more than one potential 70 promoter plus a reasonable Shine-Dalgarno sequence appear to be located upstream of the lpfAE23 gene. However, RT-PCR analysis of transcripts extracted from wild-type EPEC grown under in vitro conditions suitable for demonstrating adherence to epithelial cell lines showed only a very weak lpfAE23 transcript (Fig. 3). Expression of Lpf was not improved by growing the organism under a variety of conditions. Meanwhile, in vitro characterization of lpf genes in other A/E pathogenic E. coli strains has presented a mixed picture. As noted above, both EHEC O157:H7 lpf operons produced fimbriae in K-12. Mutation of lpf1 in the native EHEC O157:H7 resulted in slight, nonsignificant reductions in adherence to HeLa cells along with decreased microcolony formation (35). Mutation of the lpf2 locus in O157:H7 caused no detectable effect in adherence to HeLa cells but did slightly diminish adherence to Caco-2 cells at early time points (36). The REPEC strain 83/39 also contains two lpf loci, but mutation of one or both loci caused no alteration in adherence to CHO-K1 cells (26).
In any in vitro system used to study the contribution of potential adherence factors, the inability to detect a phenotype with an isogenic mutant or with the cloned adhesin genes in K-12 can always be attributed to cultural conditions or tissue culture cell lines that do not accurately reproduce the relevant in vivo situation. Since C. rodentium is a natural mouse pathogen, characterization of lpf in this model should provide further insight into the function of this putative adhesin. However, when tested in two different mouse strains, C3H/HeJ and C57BL/6, the lpfAcr mutant was indistinguishable from the wild-type parent with regard to fecal shedding, colonic hyperplasia, or numbers of organisms associated with intestinal tissue. As with in vitro adherence studies, the contribution of lpf genes in other A/E pathogenic E. coli strains in animal models is variable. REPEC strain 83/39 is a natural pathogen of rabbits, and Newton et al. (26) tested mutants in one or both lpf operons in this host species and found no significant differences in body weight or mortality rate between the wild type and mutants. However, animals receiving the wild type had more severe diarrhea and slightly better intestinal colonization in the early stages of disease. In addition, intimate bacterial adherence to the gut mucosa was indistinguishable in morphology, location, and number between the wild type and the lpf mutants of REPEC 83/39 (26). For EHEC O157:H7, the most important animal reservoir is cattle, although animals do not get sick from this organism. Jordan et al. (19) studied lpfA1 and lpfA2 mutants of O157:H7 in sheep, conventional pigs, and gnotobiotic piglets. In gnotobiotic piglets, a model of early colonization, the lpfA1 mutant caused fewer attaching/effacing lesions in the spiral colon than the wild-type parent. Using conventional pigs and sheep to study long-term colonization over 60 days, lpf mutants were recovered at lower levels than those for the parent strain (19). Thus, LPF appears to contribute to the long-term persistence, rather than virulence, of O157:H7 in species such as sheep (19), which have been implicated as an important reservoir of O157:H7 in the farm environment (27). Thus, although we failed to detect a role for LPF in EPEC or C. rodentium pathogenesis, experience with other A/E pathogenic E. coli strains indicates that LPF contributes to intestinal colonization and disease in some hosts but to a substantially lesser degree than does the intimin adhesin, which is essential for formation of the A/E lesion. Therefore, it appears that LPF is not an essential virulence factor but may be an accessory virulence determinant in some A/E pathogens.
lpf genes also have been identified in non-A/E pathogenic E. coli, although the genes have undergone considerable divergence (16). Mutation of the lpf genes in EHEC O113:H21, which is a more significant cause of hemolytic-uremic syndrome in Australia than O157:H7, resulted in a significant reduction in adherence to CHO-K1 cells (10). An lpf mutant of a meningitis-associated extraintestinal pathogenic E. coli strain was significantly reduced in adherence to a kidney cell line (HEK 293) (16). LPF may therefore play a more important role in disease caused by E. coli strains that lack intimin and the rest of the LEE than in disease caused by E. coli strains that possess intimin and the LEE, such as EPEC and C. rodentium.
ACKNOWLEDGMENTS
We thank Mary-Jane Lombardo, Laura Leverton, and Kristen J. Kanack for critical reading of the manuscript and Diana Gomez and Asaomi Kuwae for technical assistance and suggestions.
This study was supported by NIH grant numbers AI21657 and DK58957, the Naito Foundation, and The Wellcome Trust. A.G.T. was supported by institutional funds from the UTMB John Sealy Memorial Endowment Fund for Biomedical Research.
REFERENCES
1. Baldini, M. M., J. B. Kaper, M. M. Levine, D. C. Candy, and H. W. Moon. 1983. Plasmid-mediated adhesion in enteropathogenic Escherichia coli. J. Pediatr. Gastroenterol. Nutr. 2:534-538.
2. Barrick, D., K. Villanueba, J. Childs, R. Kalil, T. D. Schneider, C. E. Lawrence, L. Gold, and G. D. Stormo. 1994. Quantitative analysis of ribosome binding sites in E. coli. Nucleic Acids Res. 22:1287-1295.
3. Bartolome, B., Y. Jubete, E. Martinez, and F. de la Cruz. 1991. Construction and properties of a family of pACYC184-derived cloning vectors compatible with pBR322 and its derivatives. Gene 102:75-78.
4. Baumler, A. J., and F. Heffron. 1995. Identification and sequence analysis of lpfABCDE, a putative fimbrial operon of Salmonella typhimurium. J. Bacteriol. 177:2087-2097.
5. Baumler, A. J., R. M. Tsolis, and F. Heffron. 1996. The lpf fimbrial operon mediates adhesion of Salmonella typhimurium to murine Peyer's patches. Proc. Natl. Acad. Sci. USA 93:279-283.
6. Clarke, S. C., R. D. Haigh, P. P. Freestone, and P. H. Williams. 2003. Virulence of enteropathogenic Escherichia coli, a global pathogen. Clin. Microbiol. Rev. 16:365-378.
7. Deibel, C., S. Kramer, T. Chakraborty, and F. Ebel. 1998. EspE, a novel secreted protein of attaching and effacing bacteria, is directly translocated into infected host cells, where it appears as a tyrosine-phosphorylated 90 kDa protein. Mol. Microbiol. 28:463-474.
8. Deng, W., Y. Li, B. A. Vallance, and B. B. Finlay. 2001. Locus of enterocyte effacement from Citrobacter rodentium: sequence analysis and evidence for horizontal transfer among attaching and effacing pathogens. Infect. Immun. 69:6323-6335.
9. Donnenberg, M. S., C. O. Tacket, S. P. James, G. Losonsky, J. P. Nataro, S. S. Wasserman, J. B. Kaper, and M. M. Levine. 1993. Role of the eaeA gene in experimental enteropathogenic Escherichia coli infection. J. Clin. Investig. 92:1412-1417.
10. Doughty, S., J. Sloan, V. Bennett-Wood, M. Robertson, R. M. Robins-Browne, and E. L. Hartland. 2002. Identification of a novel fimbrial gene cluster related to long polar fimbriae in locus of enterocyte effacement-negative strains of enterohemorrhagic Escherichia coli. Infect. Immun. 70:6761-6769.
11. Duchaud, E., C. Rusniok, L. Frangeul, C. Buchrieser, A. Givaudan, S. Taourit, S. Bocs, C. Boursaux-Eude, M. Chandler, J. F. Charles, E. Dassa, R. Derose, S. Derzelle, G. Freyssinet, S. Gaudriault, C. Medigue, A. Lanois, K. Powell, P. Siguier, R. Vincent, V. Wingate, M. Zouine, P. Glaser, N. Boemare, A. Danchin, and F. Kunst. 2003. The genome sequence of the entomopathogenic bacterium Photorhabdus luminescens. Nat. Biotechnol. 21:1307-1313.
12. Frankel, G., A. D. Phillips, I. Rosenshine, G. Dougan, J. B. Kaper, and S. Knutton. 1998. Enteropathogenic and enterohaemorrhagic Escherichia coli: more subversive elements. Mol. Microbiol. 30:911-921.
13. Giron, J. A., A. S. Ho, and G. K. Schoolnik. 1991. An inducible bundle-forming pilus of enteropathogenic Escherichia coli. Science 254:710-713.
14. Hicks, S., G. Frankel, J. B. Kaper, G. Dougan, and A. D. Phillips. 1998. Role of intimin and bundle-forming pili in enteropathogenic Escherichia coli adhesion to pediatric intestinal tissue in vitro. Infect. Immun. 66:1570-1578.
15. Higgins, D. G., J. D. Thompson, and T. J. Gibson. 1996. Using CLUSTAL for multiple sequence alignments. Methods Enzymol. 266:383-402.
16. Ideses, D., D. Biran, U. Gophna, O. Levy-Nissenbaum, and E. Z. Ron. 2005. The lpf operon of invasive Escherichia coli. Int. J. Med. Microbiol. 295:227-236.
17. Jarvis, K. G., J. A. Giron, A. E. Jerse, T. K. McDaniel, M. S. Donnenberg, and J. B. Kaper. 1995. Enteropathogenic Escherichia coli contains a putative type III secretion system necessary for the export of proteins involved in attaching and effacing lesion formation. Proc. Natl. Acad. Sci. USA 92:7996-8000.
18. Jerse, A. E., J. Yu, B. D. Tall, and J. B. Kaper. 1990. A genetic locus of enteropathogenic Escherichia coli necessary for the production of attaching and effacing lesions on tissue culture cells. Proc. Natl. Acad. Sci. USA 87:7839-7843.
19. Jordan, D. M., N. Cornick, A. G. Torres, E. A. Dean-Nystrom, J. B. Kaper, and H. W. Moon. 2004. Long polar fimbriae contribute to colonization by Escherichia coli O157:H7 in vivo. Infect. Immun. 72:6168-6171.
20. Kenny, B., R. DeVinney, M. Stein, D. J. Reinscheid, E. A. Frey, and B. B. Finlay. 1997. Enteropathogenic E. coli (EPEC) transfers its receptor for intimate adherence into mammalian cells. Cell 91:511-520.
21. Knutton, S., T. Baldwin, P. H. Williams, and A. S. McNeish. 1989. Actin accumulation at sites of bacterial adhesion to tissue culture cells: basis of a new diagnostic test for enteropathogenic and enterohemorrhagic Escherichia coli. Infect. Immun. 57:1290-1298.
22. Levine, M. M., J. P. Nataro, H. Karch, M. M. Baldini, J. B. Kaper, R. E. Black, M. L. Clements, and A. D. O'Brien. 1985. The diarrheal response of humans to some classic serotypes of enteropathogenic Escherichia coli is dependent on a plasmid encoding an enteroadhesiveness factor. J. Infect. Dis. 152:550-559.
23. McDaniel, T. K., K. G. Jarvis, M. S. Donnenberg, and J. B. Kaper. 1995. A genetic locus of enterocyte effacement conserved among diverse enterobacterial pathogens. Proc. Natl. Acad. Sci. USA 92:1664-1668.
24. McDaniel, T. K., and J. B. Kaper. 1997. A cloned pathogenicity island from enteropathogenic Escherichia coli confers the attaching and effacing phenotype on E. coli K-12. Mol. Microbiol. 23:399-407.
25. Metcalf, W. W., W. Jiang, L. L. Daniels, S. K. Kim, A. Haldimann, and B. L. Wanner. 1996. Conditionally replicative and conjugative plasmids carrying lacZ alpha for cloning, mutagenesis, and allele replacement in bacteria. Plasmid 35:1-13.
26. Newton, H. J., J. Sloan, V. Bennett-Wood, L. M. Adams, R. M. Robins-Browne, and E. L. Hartland. 2004. Contribution of long polar fimbriae to the virulence of rabbit-specific enteropathogenic Escherichia coli. Infect. Immun. 72:1230-1239.
27. Ogden, I. D., M. MacRae, and N. J. Strachan. 2005. Concentration and prevalence of Escherichia coli O157 in sheep faeces at pasture in Scotland. J. Appl. Microbiol. 98:646-651.
28. Paulozzi, L. J., K. E. Johnson, L. M. Kamahele, C. R. Clausen, L. W. Riley, and S. D. Helgerson. 1986. Diarrhea associated with adherent enteropathogenic Escherichia coli in an infant and toddler center, Seattle, Washington. Pediatrics 77:296-300.
29. Reese, M. G. 2001. Application of a time-delay neural network to promoter annotation in the Drosophila melanogaster genome. Comput. Chem. 26:51-56.
30. Reese, M. G. 2000. Computational prediction of gene structure and regulation in the genome of Drosophila melanogaster. Ph.D thesis. University of California—Berkeley/University of Hohenheim, Stuttgart, Germany.
31. Schauer, D. B., and S. Falkow. 1993. Attaching and effacing locus of a Citrobacter freundii biotype that causes transmissible murine colonic hyperplasia. Infect. Immun. 61:2486-2492.
32. Tacket, C. O., M. B. Sztein, G. Losonsky, A. Abe, B. B. Finlay, B. P. McNamara, G. T. Fantry, S. P. James, J. P. Nataro, M. M. Levine, and M. S. Donnenberg. 2000. Role of EspB in experimental human enteropathogenic Escherichia coli infection. Infect. Immun. 68:3689-3695.
33. Tamano, K., S. Aizawa, E. Katayama, T. Nonaka, S. Imajoh-Ohmi, A. Kuwae, S. Nagai, and C. Sasakawa. 2000. Supramolecular structure of the Shigella type III secretion machinery: the needle part is changeable in length and essential for delivery of effectors. EMBO J. 19:3876-3887.
34. Tatsuno, I., H. Kimura, A. Okutani, K. Kanamaru, H. Abe, S. Nagai, K. Makino, H. Shinagawa, M. Yoshida, K. Sato, J. Nakamoto, T. Tobe, and C. Sasakawa. 2000. Isolation and characterization of mini-Tn5Km2 insertion mutants of enterohemorrhagic Escherichia coli O157:H7 deficient in adherence to Caco-2 cells. Infect. Immun. 68:5943-5952.
35. Torres, A. G., J. A. Giron, N. T. Perna, V. Burland, F. R. Blattner, F. Avelino-Flores, and J. B. Kaper. 2002. Identification and characterization of lpfABCC'DE, a fimbrial operon of enterohemorrhagic Escherichia coli O157:H7. Infect. Immun. 70:5416-5427.
36. Torres, A. G., K. J. Kanack, C. B. Tutt, V. Popov, and J. B. Kaper. 2004. Characterization of the second long polar (LP) fimbriae of Escherichia coli O157:H7 and distribution of LP fimbriae in other pathogenic E. coli strains. FEMS Microbiol. Lett. 238:333-344.
37. Vallance, B. A., W. Deng, K. Jacobson, and B. B. Finlay. 2003. Host susceptibility to the attaching and effacing bacterial pathogen Citrobacter rodentium. Infect. Immun. 71:3443-3453.
38. Wang, R. F., and S. R. Kushner. 1991. Construction of versatile low-copy-number vectors for cloning, sequencing and gene expression in Escherichia coli. Gene 100:195-199.(Ichiro Tatsuno, Rosanna M)