An IncI1 Plasmid Contributes to the Adherence of the Atypical Enteroaggregative Escherichia coli Strain C1096 to Cultured Cells and Abiotic
http://www.100md.com
感染与免疫杂志 2006年第4期
Center for Vaccine Development Departments of Pediatrics Medicine Microbiology and Immunology, University of Maryland School of Medicine, Baltimore, Maryland
Universidade Federal de Sao Paulo, Sao Paulo, Brazil
Groupe de Genetique des Biofilms, Institut Pasteur, URA CNRS 2172, 25 rue du Docteur Roux, 75724 Paris Cedex 15, France
Bacteriology Section, Instituto de Salud Publica, Santiago, Chile
ABSTRACT
Enteroaggregative Escherichia coli (EAEC) is defined by a characteristic "stacked-brick" aggregative adherence (AA) pattern to cultured cells. In well-studied EAEC prototype strains (called typical EAEC strains), the AA phenotype requires aggregative adherence fimbriae (AAFs). However, previous studies suggest that known AAF alleles are not found in all EAEC strains. To define mechanisms contributing to adherence in an atypical strain, we studied EAEC strain C1096. An E. coli K12 derivative carrying two plasmids, designated pSERB1 and pSERB2, from C1096 adhered to cell lines and exhibited an AA pattern. Nucleotide sequence analysis of pSERB1 indicated that it is related to plasmids of the IncI1 incompatibility group. These plasmids encode genes involved in pilus-mediated conjugal transfer, as well as pilL-V, which encodes a second pilus of the type IV family. Insertional inactivation of the gene predicted to encode the major type IV pilin subunit (pilS) reduced conjugal transfer of the plasmid by 4 orders of magnitude. Adherence of the mutant strain to polystyrene and to HT29 cells was reduced by approximately 21% and 75%, respectively. In a continuous-flow microfermentor, the pilS inactivation reduced mature biofilm formation on a glass slide by approximately 50%. In addition, the simultaneous presence of both pSERB1 and pSERB2 plasmids promoted pilS-independent biofilm formation. We conclude that the IncI1 plasmid of EAEC C1096 encodes a type IV pilus that contributes to plasmid conjugation, epithelial cell adherence, and adherence to abiotic surfaces. We also observe that AA can be mediated by factors distinct from AAF adhesins.
INTRODUCTION
Enteroaggregative Escherichia coli (EAEC) is an emerging diarrheal pathotype defined by its characteristic "stacked-brick" aggregative adherence (AA) pattern to semiconfluent HEp-2 cells (30). This pattern results from adherence to cultured cells, other bacteria, and surrounding abiotic surfaces. For the prototype EAEC strain 042, this phenotype is mediated by the aggregative adherence fimbria (AAF) designated AAF/II. Three other AAF alleles are defined, and current data suggest that others exist (2, 9, 31). AAF fimbriae are necessary ex vivo for adherence of EAEC 042 to intestinal epithelia, abiotic surfaces, and cultured cells (8, 31, 42). Following colonization, EAEC may release cytotoxins, resulting in cellular damage (14, 40), and elicit an intestinal inflammatory response, which is stimulated at least in part by the bacterial flagellum (44).
Distribution studies previously identified numerous plasmid and chromosomal genes that are well conserved in EAEC strains isolated from distinct geographic locations. These include genes for dispersin (aap), which is involved in dispersing bacterial aggregates (41); aatA, which encodes a TolC-related transport system for dispersin (33); an AggR-regulated chromosomal gene (E. G. Dudley and J. P. Nataro, unpublished data); and the AraC-family transcriptional regulator aggR (32). These genes, including aggR itself (R. Sohoni and J. P. Nataro, unpublished data), are transcriptionally activated by AggR, and we hypothesize that the aggR regulon constitutes a package of linked virulence genes defining typical EAEC strains (30). Some clinical studies report that aggR-positive EAEC strains are isolated more frequently from individuals presenting with diarrhea and elevated fecal cytokine levels (19, 34, 35), supporting the hypothesis that the presence of the aggR regulon distinguishes pathogenic from nonpathogenic strains.
In contrast, atypical EAEC strains (30) lack all genes of the aggR regulon. It is suspected that the majority of these strains are nonpathogenic (6, 39). Atypical EAEC strains also lack genes for the currently defined AAFs, as do at least one-third of all EAEC strains (2). Therefore, either certain EAEC strains encode distantly related AAFs or there exist AAF-independent mechanisms for bacteria to adhere in an AA phenotype. Support for this latter hypothesis was recently reported for EAEC 236, which encodes an outer-membrane protein that confers AA upon E. coli K-12 (28).
Type IV pili are a group of well-studied extracellular structures defined by related phylogeny and a conserved pathway of pilus biogenesis (45). These pili are implicated in numerous phenotypes in gram-negative pathogens, including adherence (12, 47), twitching motility (26), DNA transformation (13), and conjugation (20). The role of a type IV pilus in promoting conjugation is best described for the Salmonella plasmid designated R64 (17, 49), which is a member of the IncI1 plasmid incompatibility group. The IncI1 family encodes a large number of genes mediating the synthesis of two different pili, defined as the thick pilus and the thin pilus (3). The former is essential for conjugal transfer of R64, while the thin pilus, which is categorized as type IV, is needed only when conjugation occurs in liquid medium (3). The thin pilus locus encompasses 14 genes, designated pilI through pilV, and 12 of these genes are essential for pilus biogenesis (49). A number of studies define the role of the thin pili in mediating surface interactions between bacteria (16-18); however, the IncI1-encoded type IV pilus has not been implicated in phenotypes apart from conjugation.
Here, we report the first studies on a proven pathogenic atypical EAEC strain, targeting the multiple-antibiotic-resistant strain implicated in a 1995 Serbian neonatal ward outbreak (5). This strain, designated C1096, is probe negative for aggR and all currently known AggR regulon members (9). We found the following: (i) EAEC C1096 carries a large plasmid (pSERB1) that encodes a functional type IV pilus and is related to plasmids of the IncI1 family; (ii) type IV pilus synthesis is required for optimal conjugation of the large plasmid; (iii) the presence of pSERB1 mediates adherence and biofilm formation of C1096 via expression of type IV pili; (iv) pSERB1, in cooperation with the small cryptic plasmid pSERB2, also mediates type IV pilus-independent biofilm formation; and (v) the genes encoding the type IV pilus are found in approximately 10% of both typical and atypical EAEC strains. This paper therefore reports two previously undescribed phenotypes conferred by an IncI1 plasmid: adherence to epithelial cells in a type IV pilus-dependent manner and adherence to abiotic surfaces via a type IV pilus-independent mechanism.
MATERIALS AND METHODS
Strains, plasmids, and antibiotics. Strains and plasmids used in this study are listed in Table 1. Bacteria used in colony hybridization experiments were described previously (9) or are clinical isolates from our laboratory stocks. Antibiotics, when used, were incorporated into solid and liquid media at the following concentrations: trimethoprim (TMP), 10 μg ml–1; ampicillin, 60 μg ml–1; chloramphenicol, 25 μg ml–1; kanamycin (KAN), 20 μg ml–1; streptomycin (STR), 50 μg ml–1; and nalidixic acid, 10 μg ml–1.
Cell culture techniques. All cell cultures were maintained in humidified 5% CO2 at 37°C. HEp-2 cells were grown in low-glucose Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% (vol/vol) fetal calf serum, penicillin (100 U ml–1), and STR (100 μg ml–1) (all from Gibco-BRL). HT29/C1 human colonic cells were grown in high-glucose DMEM supplemented with 10% fetal calf serum, apotransferrin (10 μg ml–1) (Sigma), STR (50 μg ml–1), and penicillin (50 U ml–1).
Molecular biology protocols. DNA manipulations were performed using standard laboratory protocols (38). Total DNA was isolated from bacterial strains by the cetyltrimethylammonium bromide-NaCl method (1). PCRs were performed using either the Elongase Amplification System (Invitrogen) or Taq DNA polymerase (New England Biolabs).
Colony hybridizations were performed following the method of Sambrook and Russell (38), using high-stringency hybridization conditions (60°C; 5x SSC [1x SSC is 0.15 M NaCl plus 0.015 M sodium citrate], 0.1% N-lauroylsarcosine, 0.5% blocking reagent, 0.02% sodium dodecyl sulfate in prehybridization and hybridization buffers) or low-stringency conditions (42°C; 10% formamide in prehybridization and hybridization buffers). Gene probes were synthesized using Ready-To-Go DNA Labeling Beads (Amersham Pharmacia Biotech, Inc.). Template DNA for pilL, pilN, pilQ, pilS, pilT, and pilV was synthesized by PCR, using the primer pairs 5'-CTTCTGCCGTCATTACTCA-3' and 5'-CATCGATGCT GTAGTTGACA-3', 5'-AGTTCTGACACGACACTTCC-3' and 5'-CTGGTTGATGGTTTTCAGAT-3', 5'-ACGTCATAAAGGGAACACAG-3' and 5'-GACCTGACCACTGTTGATTT-3', 5'-ATGAGCGTCATAACCTGTTC-3' and 5'-CTGTTGGTTTCCAGTTTGAT-3', 5'-TGTTATGCTCACTTCCTTCC-3' and 5'-CTAACGTCCCTGCATGTAA-3', and 5'-CTTTGGGCAGAGTTATACCA-3' and 5'-TTAATTGAGCGTTACACACG-3', respectively. Amplification conditions were as follows: 95°C for 3 min and 30 cycles of 94°C for 45 s, 50°C for 30 s, and 72°C for 1 min, followed by a final extension of 72°C for 7 min.
Southern hybridizations were used to confirm successful construction of the pilS mutant of C1096 (see below), using the pilS gene probe described above. Total DNA was digested with AflIII (New England Biolabs), and fragments were separated on a 0.7% agarose gel. DNA transfer to a Hybond-N+ membrane (Amersham), prehybridization, hybridization, and radioactive detection were performed using standard methods (38). The prehybridization and hybridization temperature was 60°C. To construct derivatives of E. coli DH5 carrying plasmids from strain C1096, an EZ::TN Insertion Kit (Epicenter Technologies) was used. A dihydrofolate reductase-encoding (dfrA) transposon was randomly inserted into CsCl-purified plasmids from strain C1096 in vitro, as recommended by the manufacturer. After the transposition reaction, plasmid pools were transformed into DH5 by electroporation and selected on LB plates containing TMP (LB+TMP).
Tagging of pSERB1 with a KAN resistance marker was achieved using a one-step recombination method (29). A PCR product encoding a KAN resistance marker and 50 nucleotides complementary to a putative noncoding region of pSERB1 on the 5' and 3' ends was generated using the primers 5'-AGACGATTGTTTTCCTCTCGTAGTTTTGTCCACTCTAAACAAAGTGTTATCATATGAATATCCTCCTTAG-3' and 5'-GATCACAGGAAATGGAATAAACACTGTAACTGTCAATGGAAAAGTGAAACGTGTAGGCTGGAGCTGCTTC-3'. The aph-encoding plasmid pKD4 (10) was used as template DNA, and the reaction conditions were 95°C for 2 min; 10 cycles of 94°C for 45 s, 55°C for 30 s, and 72°C for 90 sec; and 15 cycles of 94°C for 45 s, 68°C for 30 s, and 72°C for 90 s. The PCR product was transformed into E. coli (pSERB1, pSERB2, and pKM208), and integrants were selected for on LB+KAN plates. Loss of pKM208 was confirmed by replica plating colonies onto LB+KAN and LB+KAN+ampicillin plates. Integration of the PCR product in the expected location was confirmed using PCR and the primers 5'-GCGTCAATAAGAGACGTTGT-3' and 5'-ATAAACATGAACGGGCATAC-3'. Reaction conditions were 95°C for 10 min and 30 cycles of 94°C for 45 s, 55°C for 30 s, and 72°C for 2 min 15 s, followed by 72°C for 7 min. Template DNA consisted of 1 μl of an overnight grown bacteria culture. A 0.4-kb or 2.0-kb PCR product was generated when the template was pSERB1 or pSERB1::aph, respectively. This method was also followed to construct pSERB3.
Quantification of the pSERB1-encoded genes traN and traY and chloramphenicol acetyltransferase (cat), which is not encoded on pSERB1, was accomplished using real-time PCR. Total DNA from strains E. coli C1096 and C1096pilS::dfrA was isolated using the Instagene matrix (Bio-Rad). Real-time PCR was performed in a Chromo4 Real-Time PCR Detector (Bio-Rad), using a Quantitect SYBR Green PCR kit (QIAGEN). The following primer pairs were used: cat, 5'-CCACTCATCGCAGTACTGTT-3' and 5'-TCACTGGATATACCACCGTT-3'; traN, 5'-GGAATGTGACGGTCTATCTC-3' and 5'-GACTGTCGATTTCCTGACTC-3'; and traY, 5'-ATGGGGATAGTCTGGTTTCT-3' and 5'-TCCACTCACGACACTGAGTA-3'.
DNA sequencing and bioinformatic analysis. Plasmids from E. coli (pSERB1 and pSERB2) were isolated by CsCl density gradient centrifugation, sonicated using a Fisher Scientific 550 Sonic Dismembrator, and end repaired with T4 DNA polymerase, Klenow DNA polymerase, and T4 polynucleotide kinase (38). DNA was separated on a 0.6% agarose gel, and fragments between 3 to 4 kb were isolated with a QIAquick Gel Extraction Kit (QIAGEN). Fragments were ligated into SmaI-digested pUC18 and transformed into DH5. Sequencing reactions were performed using BigDye Terminator Sequencing Reagents (Applied Biosystems) and were analyzed using an Applied Biosystems 3700 DNA analyzer.
DNA sequences were assembled and edited using a Lasergene software package (DNASTAR, Inc.). Assembled contigs were annotated using the Artemis package (37), and open reading frame identification was assisted by the GeneMark prediction program (http://opal.biology.gatech.edu/GeneMark) using the training set generated for enterohemorrhagic E. coli EDL933. Sequence alignments and amino acid motifs were identified using BLAST (http://www.ncbi.nlm.nih.gov/BLAST) and the BLOSUM62 matrix. Nucleotide sequence overlaps between pSERB1 contigs and R64 were generated using BLASTALL from the NCBI website (http://www.ncbi.nlm.nih.gov) and were visualized using the Artemis Comparison Tool from the Sanger Centre (http://www.sanger.ac.uk).
Construction of pSERB1 derivatives carrying pilS::dfrA and pilS::cat mutations. A plasmid containing dfrA inserted into pilS was constructed by a modification of the overlap PCR protocol of Merlin et al. (27). Total DNA (50 ng) from EAEC C1096 was used as a template in a 20-μl reaction mixture. A PCR fragment encoding the 5' end of pilS and upstream DNA was generated using the primers 5'-AAAAAAGTCGACCGGATTTTGGTAAAAAGAAA-3' (Nat1913) and 5'-CGCTCTTGCGGCCGCTTGGAACGGGTCATGCCTTTATCATCGTT-3' (Nat1916). A PCR fragment encoding the 3' end of pilS and downstream DNA was generated using primers 5'-AAAAAAGTCGACAGGCCTATGGTTTCTGTTTT-3' (Nat1914) and 5'-CCGTTCCAAGCGGCCGCAAGAGCGGGCAATGTGACGAATAAAGT-3' (Nat1915). The 3' end of the former product and the 5' end of the latter product overlap by 24 bp, and an NotI site is located within this overlap (underlined in the sequences of Nat1915 and Nat1916). Reaction conditions were 94°C for 30 s, followed by 30 cycles of 94°C for 30 s, 55°C for 30 s, and 68°C for 1 min. A second round of PCR was performed using approximately 20 ng of each PCR product from the prior reactions as template DNA and the primers Nat1913 and Nat1914. Reaction conditions were the same as before, except that an extension time of 2 min was used. The resulting PCR product was digested with SalI, ligated into similarly digested pCVD442, and transformed into DH5-pir. This plasmid was designated pPIL1096. Next, a PCR fragment encoding dihydrofolate reductase (dfrA) and 5' and 3' NotI sites was generated using primers 5'-AAAAAAGCGGCCGCGGATCACCGAAATCTTCATA-3' and 5'-AAAAAAGCGGCCGCGCATGCACGATTTGTAATAA-3' and the EZ::TN Transposon (Epicenter Technologies) as template DNA. The cycling conditions were the same as those used with primers Nat1914 and Nat1915.
The PCR product was digested with NotI and ligated to NotI-digested pPIL1096, forming pPIL1096dfrA. This plasmid was transformed into E. coli SM10-pir and was conjugated into EAEC C1096 (11). Sucrose-, chloramphenicol-, and TMP-resistant transconjugants were screened by PCR using primers Nat1913 and Nat1914; template DNA from wild-type C1096 generates a 2.7-kb product, whereas mutants containing the dfrA insert generate a 3.1-kb product. Putative mutants were further screened by Southern hybridization. The pilS::dfrA derivative of pSERB1 was tagged with aph as described above and was designated pSERB3.
A chloramphenicol acetyltransferase (cat) insertion into pilS was generated by amplifying cat using the primers 5'-AAAAAAGCGGCCGCGTGTAGGCTGGAGCTGCTTA-3' and 5'-ATATATGCGGCCGCCATATGAATATCCTCCTTAG-3' and pKD3 (10) as template DNA. Cycling conditions were 25 cycles of 94°C for 45 s, 55°C for 30 s, and 72°C for 70 s. The product was digested with NotI and ligated into similarly digested pPIL1096. The resulting plasmid was used as template DNA in a PCR using primers Nat1913 and Nat1914 under the conditions described previously. The resulting PCR product, encoding pilS::cat and flanking DNA, was transformed into DH5(pSERB1::aph, pSERB2, pKM208) and integrated by the one-step recombination method (29). Proper integration was verified by PCR using primers 5'-TACGTTTTACGAGTCCCTTG-3' and 5'-TGTCTGTTACACCGCAGATA-3'. Reaction conditions were 94°C for 45 s, 55°C for 30 s, and 72°C for 4 min. A wild-type product generated a PCR product of 2.7 kb, while a mutant generated a 3.4-kb product. The resulting plasmid was designated pSERB4.
Scanning electron microscopy. Bacteria were grown overnight in LB with appropriate antibiotics and were inoculated to 2% in high-glucose DMEM and grown in 24-well polystyrene cell culture dishes containing a 1-cm2 glass coverslip. After approximately 4 h of growth, the medium was removed, and nonadherent bacteria were removed by five washes with phosphate-buffered saline (PBS). Cells were fixed to the coverslips for 15 min with PBS containing 2% glutaraldehyde and 3 mM CaCl. Coverslips were sent to the Johns Hopkins School of Medicine Microscopy Facility for analysis.
Bacterial adherence assays. Adherence of bacteria to glass and polystyrene surfaces was performed using a previously described crystal violet-based method (42).
For qualitative adherence assays, cultured cells were grown on glass coverslips in 24-well tissue culture plates. T84 cells were grown to 100% confluence (8 to 10 days) and were subsequently incubated for an additional 2 to 3 days. HEp-2 and HT29 cells were grown to approximately 90% confluence. Overnight bacterial broth cultures were diluted 1:50 in culture medium supplemented to 1% methyl--D-mannopyranoside (to inhibit binding by type I fimbriae) and incubated with cells for 2.5 to 6 h at 37°C. After this period, cells were thoroughly washed with PBS to remove nonadherent bacteria, fixed in 70% methanol for 5 min, stained with 10% Giemsa for 15 min, and observed by light microscopy.
For quantitative adherence assays using plate counts, HT29 cells were grown to approximately 90% confluence. Cells were washed twice with PBS, and 1 ml of high-glucose DMEM supplemented to 1% methyl--D-mannopyranoside was added to each well. Overnight cultures of EAEC strains C1096 and C1096pilS::dfrA were diluted in PBS to an optical density at 600 nm (OD600) of 0.08 (approximately 2 x 106 CFU ml–1), and 10 μl of each was added to wells. After a 2.5-h incubation at 37°C, half of the coverslips were moved to a new 24-well plate and washed with five 1-ml aliquots of PBS. After the last PBS aliquot was aspirated, 1 ml of 0.1% Triton X-100 was added, mixed thoroughly, and allowed to stand for approximately 5 min. Serial dilutions were made on LB and LB+TMP plates to determine the total number of adherent bacteria and pilS mutant cells, respectively. To the other half of the culture plate (the half for which the coverslips were not removed), 10 μl of 10% Triton X-100 was added. Serial dilutions were made on LB and LB+TMP plates to determine the total number of bacteria and total number of pilS mutant cells, respectively, per well. The ratio of adherent wild-type to mutant bacteria was calculated by dividing the number of adherent wild-type C1096 bacteria by the number of adherent C1096pilS::dfrA bacteria and normalizing to the total CFU of each strain in the wells. In control experiments, no difference in CFU ml–1 was found when serial dilutions of C1096pilS::dfrA were plated on LB or LB+TMP, indicating that differences in adherence calculated for this strain were not due to suppression of growth of C1096pilS::dfrA by TMP. Additionally, no difference in the growth rate was observed between C1096 and C1096pilS::dfrA in DMEM.
For quantitative adherence assays using direct cell counting, HT29 cells and bacteria were grown, and infections were performed as described above. After the PBS washes, coverslips were fixed with 70% methanol for 5 min and stained with 10% Giemsa. Slides were visualized by light microscopy. A total of 100 HT29 cells were randomly chosen from each of five coverslips from the wild-type and mutant strain infections, and the number of bacteria adhering to each HT29 cell were counted. Data are reported as the number of wild-type bacteria adhering to 100 HT29 cells divided by the number of mutant bacteria adhering to 100 HT29 cells.
Biofilm formation assay in microfermentors. Sixty-milliliter microfermentors, each containing a removable Pyrex glass slide, were configured as continuous-flow culture bioreactors with a flow rate of 40 ml h–1. This flow rate minimizes planktonic growth of the bacteria, and most of the observed growth occurs on the available surfaces inside the microfermentors and on the glass slide (15). (For a more detailed explanation, see http://www.pasteur.fr/recherche/unites/Ggb/biofilmfermenter.htm.) Overnight cultures grown in 0.4% glucose M63B1 minimal medium supplemented with appropriate antibiotics were used to inoculate microfermentors containing the same medium to 1 OD600 equivalent. The bacteria were cultivated for 24 to 72 h. Biofilm biomass on the internal Pyrex glass slide was quantified by measuring the OD600 after resuspension of the slide biofilm in 10 ml of M63B1 medium. Total biofilm was measured at OD600 after resuspension of the total cell mass present within the microfermentor in 60 ml of M63B1 medium. All experiments were performed at least in triplicate.
Plasmid conjugation. Plasmid transfer for strain construction was performed by individually inoculating donor and recipient strains from an overnight culture to 1% in LB and growing them at 37°C for 1 h. Aliquots (250 μl) of donor and recipient strains were combined, centrifuged, and resuspended in 50 μl of LB. Cells were spotted on LB agar and incubated at 37°C for 6 h, and transconjugants were plated on selective medium. When DH5 or HB101 was the recipient strain, selection was performed on LB+KAN+nalidixic acid or LB+KAN+STR plates, respectively. Plasmid transfer was confirmed by agarose gel electrophoresis. The conjugation frequency of pSERB1::aph and pSERB3 from DH5 to HB101 on a solid surface was assessed similarly, except that the resuspended cells were placed on a 0.22-μM-pore-size filter overlaying an LB plate. Serial dilutions were made onto LB+KAN+STR plates to quantify the number of transconjugants and on LB+STR plates to quantify the number of recipients. Conjugation frequency is reported as the number of transconjugants per recipient.
Quantitative broth conjugations were performed by centrifuging 500-μl aliquots of overnight cultures, washing them twice with 1 ml of LB broth, and inoculating each strain into 2 ml of LB broth to an OD600 of approximately 0.03. Tubes inoculated with only one strain served as negative controls. Conjugations were carried out in 14-ml plastic tubes. Tubes were placed on their sides and rotated at 50 rpm for 6 h at 37°C. Plate counts were performed as described above for solid surface conjugation.
Nucleotide sequence accession numbers. The partial nucleotide sequence obtained for pSERB1 was deposited in the GenBank database under accession number AY686591. The full nucleotide sequence for pSERB2 was assigned accession number DQ269444.
RESULTS
EAEC C1096 adheres to epithelial cells and glass in an AA pattern. Previously, Cobeljic et al. implicated EAEC C1096 as the causative agent of a nonbloody diarrhea outbreak in Ni, Serbia. The authors reported that EAEC C1096 adhered mainly to the HEp-2 cells, with few bacteria adhering to the glass coverslips. Whereas prototype EAEC strain 042 requires AAF/II fimbriae for AA to abiotic surfaces and to human intestinal tissue (8, 42), C1096 does not encode genes for the expression of AAF/II or the distantly related fimbriae AAF/I (9). Additionally, total DNA from C1096 does not hybridize to a DNA probe (2) that identifies the usher subunit of AAF/I, AAF/II, AAF/III, and probably other as yet unidentified AAF alleles (data not shown). This suggests that AAF-independent determinants may be involved in C1096 adherence.
To characterize adherence factors of EAEC C1096, we began by confirming a previous report that this strain adhered to HEp-2 cells in an AA pattern, which is the defining characteristic of EAEC (Fig. 1A), and studying whether C1096 adhered to glass similarly to the prototype strain 042. In our hands, EAEC 042 and C1096 both displayed AA on HEp-2 cells (Fig. 1A and B, respectively). Strain C1096 also adhered to glass (Fig. 1D) in an AA pattern, although the pattern observed differed from the distinct honeycomb structure seen with EAEC 042 (Fig. 1C). Adherence to an abiotic surface was also measured quantitatively, with EAEC 042 and a biofilm-deficient derivative designated JS28 (42) serving as positive and negative controls, respectively. EAEC C1096 adhesion was weaker than that observed with strain 042 (Fig. 2), although it was still significantly higher than that seen with strain JS28. Thus, EAEC C1096 adheres to HT29 cells and abiotic surfaces by the AA phenotype, confirming the classification of C1096 as EAEC.
Plasmids from C1096 confer adherence upon DH5. EAEC C1096 carries at least six plasmids (Fig. 3). To determine if any plasmids encode adherence factors, we used an in vitro transposition reaction to randomly insert a dfrA-encoding transposon into the plasmids and transformed the tagged plasmids into E. coli DH5. All transformants screened carried at least one of the three lower-molecular-mass plasmids (1.6, 2.5, and 6.4 kb). One isolate was identified that carried one of the large plasmid species from C1096, in addition to a second plasmid that migrated in agarose gel electrophoresis at approximately 2.5 kb (Fig. 3, lane 3). The larger and smaller plasmid species from this strain were designated pSERB1 and pSERB2, respectively. Upon electroporation of plasmid pools from DH5(pSERB1, pSERB2) back into DH5, all TMP-resistant transformants carried pSERB2 only. This suggests that during strain construction, the 0.9-kb dfrA-encoding transposon inserted into the 1.6-kb plasmid from C1096, resulting in a 2.5-kb species, and this plasmid cotransformed with pSERB1 during electroporation of DH5.
DH5 transformants carrying different plasmid combinations from C1096 were screened for the ability to adhere to HEp-2 cells. DH5(pSERB1, pSERB2) adhered to cultured cells in the presence of 1% methyl--D-mannopyranoside (Fig. 4B). The bacteria were typically found in clusters, adhering to the cultured cells as well as to each other. DH5(pSERB1, pSERB2) adhered to a smaller number of HEp-2 cells than did C1096 and did not display the extensive adherence seen with strain C1096 (compare Fig. 4B with 1B). A much lower level of adherence was seen with wild-type DH5 (Fig. 4A) and with derivatives carrying any of the three low-molecular-mass plasmids from C1096 (data not shown). These observations suggest that at least one factor that contributes to the adherence of EAEC C1096 to HEp-2 cells is encoded on pSERB1 and/or pSERB2.
Adherence of DH5(pSERB1, pSERB2) to abiotic surfaces requires both plasmids. In order to test the adherence of DH5(pSERB1, pSERB2) on abiotic surfaces, we tested this strain for initial adhesion in microtiter plate assays and for mature biofilm formation in microfermentors. Although adherence of DH5(pSERB1, pSERB2) was not observed on polystyrene tissue culture plates (data not shown), DH5(pSERB1, pSERB2) formed a strong mature biofilm in a continuous-culture microfermentor, where the formation of a thick biofilm can be monitored on removable Pyrex glass slides between 48 to 72 h (Fig. 5A). To determine which plasmid was responsible for this phenotype, pSERB1 was tagged in a predicted noncoding region with the KAN resistance marker aph, creating pSERB1::aph, and pSERB2 was electroporated into both DH5 and DH5(pSERB1::aph). Adherence of DH5(pSERB1::aph) and DH5(pSERB2) to the spatula was significantly decreased compared to DH5(pSERB1::aph, pSERB2) (Fig. 5A and B), although adherence of each was greater than DH5. These results demonstrate that both plasmids are required for fully mature biofilm formation under the conditions tested and that pSERB1 and pSERB2 promote biofilm formation in a synergistic manner.
pSERB1 is related to IncI1 plasmids, and pSERB2 is a cryptic plasmid. To identify adherence and other possible virulence factors of EAEC C1096, the plasmids from DH5(pSERB1, pSERB2) were subjected to nucleotide sequence analysis. A total of 742 shotgun sequences were assembled into five contigs of 67.2, 7.2, 7.1, 2.5, and 1.1 kb, with an estimated sequence coverage of 5.9-fold.
A BLAST analysis indicated that the 2.5-kb contig encoded dfrA and, therefore, is derived from pSERB2. This contig also encoded two plasmid replication protein homologs. Further analysis of the pSERB2 sequence did not provide any additional insights into the requirement of both plasmids for the adherence phenotype observed with DH5(pSERB1::aph, pSERB2).
The remaining four contigs were all aligned by BLAST to plasmids of the IncI1 incompatibility group, including R64 (GenBank accession no. NC_005014) of Salmonella enterica serovar Typhimurium. EcoRI digests suggested that pSERB1 is approximately 82 kb in size and that the 7.2-kb, 1.1-kb, 7.1-kb, and 67.2-kb contigs (total, 82.6 kb) were aligned with nucleotides 48284 to 52639, 54932 to 56052, 56043 to 62909, and 63385 to 102094, respectively, of R64 (data not shown). These data strongly suggest that nearly all of the sequence of pSERB1 was obtained.
pSERB1 encodes a conjugal transfer locus and a type IV pilus system. R64 encodes a large number of genes responsible for the formation of two pilus types referred to as thin pili and thick pili (3). Thick pilus biosynthesis requires genes encoded on two distinct regions of R64: traBC, which encodes regulatory proteins, and a 31-kb region encoding 26 open reading frames, 22 of which are required for conjugal transfer in broth and on solid surfaces (23). Further analysis of the 67.2-kb contig revealed that homologs of these 22 proteins and TraBC were present on pSERB1, suggesting that pSERB1 may be a conjugative plasmid.
R64 thin pili are members of the type IV pilus family and are required for plasmid conjugation in broth only (22); biosynthesis requires 12 genes designated pilK-V (49). Genes similar to pilL-V are encoded on pSERB1 in the same order as on R64 (Fig. 6). Although no pilK was identified, characterization of type IV pili from Yersinia pseudotuberculosis and S. enterica serovar Typhi (7, 50) suggests that either some systems do not require pilK or certain type IV pilus-expressing organisms synthesize an unrelated protein with PilK activity. BLAST searches indicated that PilL-V are most closely related (48 to 91% amino acid identity) to a type IV pilus locus encoded on pO113, a plasmid isolated from a Shiga toxin-producing E. coli (STEC) strain (24).
Certain type IV pilus biogenesis systems encode two prepilin subunits, designated pilS (major subunit) and pilV (minor subunit). These pili are further divided into two classes based upon posttranslational processing events (25): type IVA, in which the prepilus has a short (usually 5 to 6 amino acids [aa]) leader sequence and the N-terminal amino acid residue of the mature protein is methylated phenylalanine; and type IVB, in which the prepilus has a longer (13 to 30 aa) signal sequence and the N-terminal amino acid residue is Met, Leu, Trp, or Ser. Based upon predicted signal sequence cleavage sites, PilS from EAEC C1096 is predicted to be type IVB (predicted leader sequence of 22 aa; mature PilS is predicted to have an N-terminal Met), and PilV is type IVA (predicted leader sequence of 8 aa; mature PilV is predicted to have an N-terminal Phe). The deduced amino acid sequence of PilS is aligned by BLAST only with the PilS homologs from pO113 (Fig. 6) and plasmid R721 (accession no. NP_065349; 43% identical to pilS from R64 over 173 aa).
EAEC C1096 expresses surface pili absent in a pilS::dfrA mutant. To test whether the pSERB1-encoded pil locus is expressed, dfrA was inserted into the plasmid-encoded copy of pilS by homologous recombination, creating EAEC C1096pilS::dfrA. Such a mutation in plasmid R64 eliminates the assembly of functional type IV pili (49), and by analogy the same phenotype was expected with pSERB1. Under magnification of x15,000, numerous pili of approximately 1 μm in length were observed extending from the surface of most bacterial cells (Fig. 7). The pili appear to tether bacteria together over a distance, and direct interaction of the pili with the glass coverslip was also noted. In sharp contrast, similar surface pili were not seen on the surface of EAEC C1096pilS::dfrA cells, although a minority of these cells still expressed much shorter pili (Fig. 7B).
pSERB1 is a conjugative plasmid that requires the type IV pilus for optimal conjugal transfer in broth cultures. As noted, our sequence analysis suggested that pSERB1 is an IncI1-type plasmid. This finding implies that pSERB1 is conjugative, as are other IncI1 plasmids, and that efficient conjugation in broth, but not on solid surfaces, is promoted by the type IV pilus (22). To test this, the pSERB1-derivative from EAEC C1096pilS::dfrA was tagged with aph, creating pSERB3, and this construct was transferred to DH5. When matings were performed in LB broth at 37°C, the transfer frequency of pSERB1::aph and pSERB3 to E. coli HB101 after 6 h was 2.6 x 10–2 and 6.7 x 10–6 transconjugants per recipient, respectively. When matings were performed on a 0.22-μM-pore-size filter overlaying LB agar, transfer frequencies were 1.3 x 10–3 and 3.2 x 10–3, which was not a statistically significant difference (P = 0.17).
Because of the cooperative role of pSERB1 and pSERB2 in biofilm formation, we also tested whether the transfer frequency of pSERB1::aph was altered by the presence of pSERB2. However, no detectable difference was noted in the transfer frequency of pSERB1::aph to E. coli HB101 when either DH5(pSERB1::aph, pSERB2) or DH5(pSERB1::aph) was used as the donor strain. This suggests that the biofilm enhancement phenotype is not the result of increased conjugal transfer of pSERB1.
The type IV pilus contributes to adherence of EAEC C1096 to HT29 cells and to abiotic surfaces. Next, we tested whether the type IV pilus mediates adherence of EAEC C1096 to cells and abiotic surfaces. Adherence to HT29 cells was quantified both as single-strain infections and as a coinfection. First, HT29 cells were coinfected with equivalent numbers of C1096 and the pilS::dfrA mutant strains. We performed coinfections because we found that C1096 detaches epithelial cells from the glass coverslips during infection, and therefore differences seen in experiments performed as single-strain infections could be due to differences in the numbers of detached HT29 cells. Data from four independent experiments revealed 2.9-fold more CFU for the wild-type compared to the mutant strain, or a 65% decrease in adherence (P = 0.02 by a Student's pairwise t test). Direct counting of bacteria adhering to HT29 cells by microscopic examination of single-strain infections supported this result, as 5.4-fold fewer pilS::dfrA bacteria (82% reduction) were counted adhering to HT29 cells than with the wild-type bacteria. Therefore, the two different methods of quantifying bacterial adherence to cultured cells were in agreement and suggested that EAEC C1096 adherence to epithelial cells is reduced by approximately 75% when the type IV pilus is inactivated.
Qualitative examination of adherence to glass of the mutant and wild-type strains indicated that the pattern of biofilm formation is not strongly affected by the pilS mutation. Nevertheless, quantitative assays revealed a slight (21%) but statistically significant reduction of biomass adhering in wells containing the mutant strain compared to the wild type (Fig. 2) (P = 0.009 by a Student's pairwise t test). In continuous-flow culture, the inactivation of pilS did not affect the total biofilm biomass in the microfermentor but reduced biofilm development on the vertical internal glass slide by approximately 50% (Fig. 8) (P = 0.014), indicating that a more fragile biofilm was formed by the C1096pilS mutant. Since the wild-type C1096 and its pilS mutant have similar growth rates in liquid culture, this suggests that the type IV pilus encoded on pSERB1 affects the initial stages of adhesion and/or promotes tight bacteria-bacteria interactions within a biofilm.
Many conjugative plasmids, including those of the IncI1 group, are known to mediate biofilm formation in bacteria via a conjugation-dependent mechanism (15). The contribution of the type IV pilus subunit was therefore further investigated in the microfermentor-biofilm model. As shown in Fig. 5, biofilm formation by strains DH5(pSERB3) and DH5(pSERB4, pSERB2) was not significantly different than that formed by strains DH5 (pSERB1::aph) and DH5(pSERB1::aph, pSERB2), respectively. Therefore, adherence in the microfermentor assay does not depend upon efficient plasmid conjugation.
Lastly, we wanted to confirm that phenotypes observed with E. coli C1096pilS::dfrA were not due to an altered copy number of pSERB1 or to altered expression of the pSERB1-encoded thick pilus. Real-time PCR using total DNA from C1096 and C1096pilS::dfrA as a template indicated that the quantity of traN and traY (two pSERB1-encoded genes) relative to that of cat (a gene not encoded on pSERB1) was the same for both strains. This indicates that the dfrA insertion does not affect plasmid copy number. Additionally, as reported above, the transfer frequency to E. coli HB101, when conjugation was performed on a 0.22-μM-pore-size filter, was comparable for DH5(pSERB1::aph) and DH5a(pSERB3). This suggests inactivation of pilS did not affect the synthesis of the thick pilus, which is essential for plasmid conjugation.
Taken together, these results indicate that the type IV pili from EAEC C1096 play a role in adherence to cultured epithelial cells and moderately contribute to both initial or late biofilm formation on abiotic surfaces. These data also suggest the presence of other adherence mechanisms in this strain.
Genes from the EAEC C1096 type IV pilus locus are found in both typical and atypical EAEC strains. Colony hybridizations were performed to determine whether other EAEC strains carry genes for type IV pilus biogenesis that are related to those found in EAEC C1096. A library consisting of 40 typical and 20 atypical EAEC strains was screened. Under both high-stringency (60°C hybridization) and low-stringency (42°C hybridization in 10% formamide) conditions, four of the typical EAEC strains were hybridization positive using the pilL, pilN, pilQ, pilS, pilT, and pilV probes (data not shown). These strains were previously designated H38-1, DS67-R2, 278-1, and 495-1 (9) and were isolated from Peru (H38-1), the Philippines (DS67-R2), and Thailand (278-1 and 495-1). Two atypical EAEC strains, designated strains "A" and 25-1, were also probe positive under high-stringency conditions with all six gene probes. Additionally, an atypical strain designated 6-1 was probe positive under high-stringency conditions only with pilN, pilL, and pilV, and a fourth strain designated DS61-R2 was only probe positive with pilV under low-stringency conditions. As expected, the prototype EAEC strain 042 was probe negative with all probes.
DISCUSSION
EAEC strains are defined as strains that adhere to HEp-2 cells in an aggregative pattern (30). While the majority of well-characterized strains adhere to abiotic surfaces (42), cultured cell lines (8, 31), and human intestinal tissue (8) via AAFs, it is clear that alternative mechanisms are also responsible for aggregative adherence patterns (2, 9, 28). While studying the adherence factors of EAEC strain C1096, we discovered that one of the large plasmids carried by this strain is related to members of the IncI1 incompatibility group and encodes genes for two previously described pili designated thick and thin (type IV) pili. We suggest here a model where the IncI1 plasmid-encoded type IV pili carried by EAEC C1096 contribute to, but are not essential for, adherence of this strain to cells and surfaces (Fig. 9). Additionally, this pilus mediates an increase in adherence of the bacteria to abiotic surfaces in a manner that is further enhanced by an unknown interaction with a small cryptic plasmid designated pSERB2 (Fig. 9A).
It is well established that IncI1-encoded type IV pili mediate conjugal transfer (Fig. 9A) of these plasmids (21). We are unaware, however, of any prior reports implicating the thin pilus from conjugative plasmids as a factor mediating adherence to intestinal epithelial cells (Fig. 9B); in fact Srimanote et al. (43) previously reported that the STEC plasmid pO113, which belongs to the IncI1 family, does not play a role in adherence to intestinal tissue. Some conjugative plasmids, including IncI1 members, also catalyze adherence to abiotic surfaces as the result of host-donor contact during DNA transfer (15, 36). Therefore, it is possible that the decreased adherence to HT29 cells observed with the C1096dfrA::pilS mutant is the result of decreased mating aggregate formation. However, given that C1096dfrA::pilS demonstrated a 65 to 82% decrease in adherence to HT29 cells, while only a 21% decrease in adherence to polystyrene, we favor a model in which the type IV pilus mediates adherence through direct contact with the eukaryotic cell membrane or other surface structure. The decreased adherence of the pilS mutant on polystyrene suggests that the pSERB1-encoded pilL-V may also augment biofilm levels or may accelerate biofilm formation, as suggested for the Vibrio cholerae type IV pilus (48). In summary, our data indicate that type IV pili encoded on some IncI1 plasmids may be multifunctional, serving roles in adherence to both host cells and abiotic surfaces and in plasmid conjugation.
We are interested in determining whether this observation can be extended to other IncI1 plasmids found in gram-negative bacterial pathogens. We would also like to further characterize the role of individual genes from the pilL-V locus and those encoding the thick pilus. However, given the antibiotic resistance profile of C1096 (including ampicillin, chloramphenicol, KAN, tetracycline, gentamicin, neomycin, and STR) (reference 5 and data not shown), the large number of plasmids present in this strain, and the current lack of knowledge concerning genetic manipulations in this strain, further studies are needed before investigators can use sophisticated genetic techniques that combine gene knockouts with complementation in trans using expression vectors. We are pursuing the development of such vectors rather than studying a more genetically tractable host, as C1096 is currently the only atypical EAEC strain that is a proven human pathogen.
Strains carrying IncI1 plasmids were previously shown to form biofilms in continuous-flow microfermentors (15). This phenotype required the presence of recipient bacteria, presumably because the expression of the plasmid-encoded conjugation apparatus and surface pili is repressed unless bacteria lacking the IncI1 plasmid are present in the population. We found that pure cultures of C1096 form a type IV pilus-mediated biofilm in the microfermentor, suggesting that this pSERB1-encoded pilus is expressed regardless of whether recipient bacteria are present. The continuous-flow culture experiments also showed that while both C1096 and C1096pilS::dfrA biofilms initially form on the internal vertical surfaces, the pilS mutant falls to the bottom of the microfermentor to a greater extent than the wild type during later stages of biofilm development. This suggests that the pilS mutant may form weaker bacteria-bacteria interactions, resulting in a more fragile biofilm under these conditions. Interestingly, biofilm formation by DH5(pSERB1::aph, pSERB2) in the microfermentor was more than fourfold greater than that of DH5(pSERB1::aph) (Fig. 9A), and this adherence did not depend upon the pSERB1-encoded type IV pilus or efficient conjugal transfer of pSERB1. Given the scarcity of genetic information found in the pSERB2 sequence, it is difficult to speculate as to the nature of this effect. While pSERB2 is mobilizable in the presence of pSERB1 (E. G. Dudley and J. P. Nataro, unpublished results), the frequency of pSERB2 conjugation is approximately 3 log units less than that of pSERB1. Therefore, it seems unlikely that the conjugative nature of pSERB2 alone contributes to increasing the number of mating aggregates. Plasmids related to pSERB2 from putative uropathogenic E. coli have been described by others (4), and thus studies concerning whether these plasmids modulate adherence in other species would be of interest.
Colony hybridization experiments indicated that approximately 10% of the EAEC strains screened encode genes related to pilL, pilN, pilQ, pilS, pilT, and pilV and that these genes are often inherited together. This distribution frequency is similar to that reported for the well-studied alleles of the AAF fimbriae, in which the AAF/I, AAF/II, and AAF/III are also detected only in a minority of EAEC isolates (2, 9). It is possible that type IV pili are more widely distributed in this EAEC library, assuming that some probe-negative strains encode distantly related genes. The prototype EAEC strain 042 was probe negative with all gene probes tested, and genes similar to pilL-V from pSERB1 are not found in the completed genome sequence of prototype EAEC strain 042 (http://www.sanger.ac.uk). As four of the probe-positive strains were also probe positive for aggR and other members of the aggR regulon (9), these data suggest that type IV pili related to those encoded on pSERB1 are not a defining characteristic of either typical or atypical EAEC but may serve as an accessory adhesin of EAEC. These genes are not as widely distributed as the related type IV pili of STEC, which were found in 16 of 19 strains that lack the locus of enterocyte effacement (43). Future work is directed at identifying additional adherence factors of atypical EAEC in order to determine whether this pathotype adheres to surfaces and cells via a common mechanism.
This study and previous work highlight a difficulty with the current method of defining the EAEC pathotype. The HEp-2 adherence assay is the current "gold standard" for identification; however, it has become clear in recent years that many strains exhibiting the EAEC "stacked-brick" pattern on cultured cells are distinct from the prototype strain 042 as determined by probe hybridization and/or multilocus enzyme electrophoresis (2, 9). While the HEp-2 adherence assay is a proven and reliable method for distinguishing EAEC from other E. coli pathotypes, genetic heterogeneity of isolates and variability of adherence patterns (2, 43, 46) highlight the need for molecular methods of identifying and characterizing EAEC strains. A growing body of evidence suggests that the EAEC 042 plasmid-encoded regulator AggR is a global regulator of virulence determinants (9, 35; E. G. Dudley and J. P. Nataro, unpublished data) and that genes of this regulon are detected more frequently in EAEC isolates from stools of patients with diarrhea than from healthy individuals (19, 35, 39). Therefore, the nomenclature of typical EAEC and atypical EAEC was recently proposed to differentiate EAEC strains carrying and lacking members of the aggR regulon, respectively (30), and genes of the aggR regulon were designated as markers defining typical EAEC. While the majority of atypical EAEC strains are believed to be nonpathogenic (39), outbreaks such as the one involving EAEC C1096 demonstrate that some atypical strains cause diarrhea and that AAFs are not the only factors that confer AA. The virulence factors of atypical EAEC are largely undefined, and this work represents the first study directed at defining the molecular pathogenesis of an AggR regulon-negative EAEC strain. Further studies to understand the pathogenic factors of EAEC C1096 and to characterize the similarities and differences between atypical EAEC strains are under way.
ACKNOWLEDGMENTS
This work was supported by grant AI033096 through the National Institute of Allergy and Infectious Diseases to J.P.N. J.-M. G. and P.L.-L. are supported by the Institut Pasteur and CNRS URA2172 and the Foundation BNP-PARIBAS grants.
We thank O. Colin Stine and Yuansha Chen (University of Maryland—Baltimore Department of Epidemiology and Preventative Medicine) for generating the pSERB1 shotgun sequence data and Brad Harris (Johns Hopkins School of Medicine) for taking the scanning electron microscopy pictures.
Present address: Institute of Child Health, University of Birmingham, B4 6LH Birmingham, United Kingdom.
REFERENCES
1. Ausubel, F. M., R. Brent, D. D. Moore, J. A. Smith, J. G. Siedman, and K. Struhl. 1999. Current protocols in molecular biology. John Wiley & Sons, New York, N.Y.
2. Bernier, C., P. Gounon, and C. Le Bouguenec. 2002. Identification of an aggregative adhesion fimbria (AAF) type III-encoding operon in enteroaggregative Escherichia coli as a sensitive probe for detecting the AAF-encoding operon family. Infect. Immun. 70:4302-4311.
3. Bradley, D. E. 1984. Characteristics and function of thick and thin conjugative pili determined by transfer-derepressed plasmids of incompatibility groups I1, I2, I5, B, K and Z. J. Gen. Microbiol. 130:1489-1502.
4. Burian, J., L. Guller, M. Macor, and W. W. Kay. 1997. Small cryptic plasmids of multiplasmid, clinical Escherichia coli. Plasmid 37:2-14.
5. Cobeljic, M., B. Miljkovic-Selimovic, D. Paunovic-Todosijevic, Z. Velickovic, Z. Lepsanovic, N. Zec, D. Savic, R. Ilic, S. Konstantinovic, B. Jovanovic, and V. Kostic. 1996. Enteroaggregative Escherichia coli associated with an outbreak of diarrhoea in a neonatal nursery ward. Epidemiol. Infect. 117:11-16.
6. Cohen, M. B., J. P. Nataro, D. I. Bernstein, J. Hawkins, N. Roberts, and M. A. Staat. 2005. Prevalence of diarrheagenic Escherichia coli in acute childhood enteritis: a prospective controlled study. J. Pediatr. 146:54-61.
7. Collyn, F., M.-A. Lety, S. Nair, V. Escuyer, A. Ben Younes, M. Simonet, and M. Marceau. 2002. Yersinia pseudotuberculosis harbors a type IV pilus gene cluster that contributes to pathogenicity. Infect. Immun. 70:6196-6205.
8. Czeczulin, J. R., S. Balepur, S. Hicks, A. Phillips, R. Hall, M. H. Kothary, F. Navarro-Garcia, and J. P. Nataro. 1997. Aggregative adherence fimbria II, a second fimbrial antigen mediating aggregative adherence in enteroaggregative Escherichia coli. Infect. Immun. 65:4135-4145.
9. Czeczulin, J. R., T. S. Whittam, I. R. Henderson, F. Navarro-Garcia, and J. P. Nataro. 1999. Phylogenetic analysis of enteroaggregative and diffusely adherent Escherichia coli. Infect. Immun. 67:2692-2699.
10. Datsenko, K. A., and B. L. Wanner. 2000. One-step inactivation of chromosomal genes in Escherichia coli K-12 using PCR products. Proc. Natl. Acad. Sci. USA 97:6640-6645.
11. Donnenberg, M., and J. Kaper. 1991. Construction of an eae deletion mutant of enteropathogenic Escherichia coli by using a positive-selection suicide vector. Infect. Immun. 59:4310-4317.
12. Donnenberg, M. S., J. A. Giron, J. P. Nataro, and J. B. Kaper. 1992. A plasmid-encoded type IV fimbrial gene of enteropathogenic Escherichia coli associated with localized adherence. Mol. Microbiol. 6:3427-3437.
13. Dubnau, D. 1999. DNA uptake in bacteria. Annu. Rev. Microbiol. 53:217-244.
14. Eslava, C., F. Navarro-Garcia, J. R. Czeczulin, I. R. Henderson, A. Cravioto, and J. P. Nataro. 1998. Pet, an autotransporter enterotoxin from enteroaggregative Escherichia coli. Infect. Immun. 66:3155-3163.
15. Ghigo, J. 2001. Natural conjugative plasmids induce bacterial biofilm development. Nature 412:442-445.
16. Ishiwa, A., and T. Komano. 2000. The lipopolysaccharide of recipient cells is a specific receptor for PilV proteins, selected by shufflon DNA rearrangement, in liquid matings with donors bearing the R64 plasmid. Mol. Gen. Genet. 263:159-164.
17. Ishiwa, A., and T. Komano. 2004. PilV adhesins of plasmid R64 thin pili specifically bind to the lipopolysaccharides of recipient cells. J. Mol. Biol. 343:615-625.
18. Ishiwa, A., and T. Komano. 2003. Thin pilus PilV adhesins of plasmid R64 recognize specific structures of the lipopolysaccharide molecules of recipient cells. J. Bacteriol. 185:5192-5199.
19. Jiang, Z. D., D. Greenberg, J. P. Nataro, R. Steffen, and H. L. DuPont. 2002. Rate of occurrence and pathogenic effect of enteroaggregative Escherichia coli virulence factors in international travelers. J. Clin. Microbiol. 40:4185-4190.
20. Kim, S., and T. Komano. 1997. The plasmid R64 thin pilus identified as a type IV pilus. J. Bacteriol. 179:3594-3603.
21. Komano, T. 1999. Shufflons: multiple inversion systems and integrons. Annu. Rev. Genet. 33:171-191.
22. Komano, T., N. Funayama, S. R. Kim, and T. Nisioka. 1990. Transfer region of IncI1 plasmid R64 and role of shufflon in R64 transfer. J. Bacteriol. 172:2230-2235.
23. Komano, T., T. Yoshida, K. Narahara, and N. Furuya. 2000. The transfer region of IncI1 plasmid R64: similarities between R64 tra and Legionella icm/dot genes. Mol. Microbiol. 35:1348-1359.
24. Leyton, D. L., J. Sloan, R. E. Hill, S. Doughty, and E. L. Hartland. 2003. Transfer region of pO113 from enterohemorrhagic Escherichia coli: similarity with R64 and identification of a novel plasmid-encoded autotransporter, EpeA. Infect. Immun. 71:6307-6319.
25. Manning, P. 1997. The tcp gene cluster of Vibrio cholerae. Gene 192:63-70.
26. Mattick, J. S. 2002. Type IV pili and twitching motility. Annu. Rev. Microbiol. 56:289-314.
27. Merlin, C., S. McAteer, and M. Masters. 2002. Tools for characterization of Escherichia coli genes of unknown function. J. Bacteriol. 184:4573-4581.
28. Monteiro-Neto, V., S. Y. Bando, C. A. Moreira-Filho, and J. A. Giron. 2003. Characterization of an outer membrane protein associated with haemagglutination and adhesive properties of enteroaggregative Escherichia coli O111:H12. Cell Microbiol. 5:533-547.
29. Murphy, K. C., and K. G. Campellone. 2003. Lambda Red-mediated recombinogenic engineering of enterohemorrhagic and enteropathogenic E. coli. BMC Mol. Biol. 4:11.
30. Nataro, J. P. 2003. Enteroaggregative Escherichia coli. In W. M. Scheld, J. M. Hughes, and B. E. Murray (ed.), Emerging infections 6. ASM Press, Washington, D.C.
31. Nataro, J. P., Y. Deng, D. R. Maneval, A. L. German, W. C. Martin, and M. M. Levine. 1992. Aggregative adherence fimbriae I of enteroaggregative Escherichia coli mediate adherence to HEp-2 cells and hemagglutination of human erythrocytes. Infect. Immun. 60:2297-2304.
32. Nataro, J. P., D. Yikang, D. Yingkang, and K. Walker. 1994. AggR, a transcriptional activator of aggregative adherence fimbria I expression in enteroaggregative Escherichia coli. J. Bacteriol. 176:4691-4699.
33. Nishi, J., J. Sheikh, K. Mizuguchi, B. Luisi, V. Burland, A. Boutin, D. J. Rose, F. R. Blattner, and J. P. Nataro. 2003. The export of coat protein from enteroaggregative Escherichia coli by a specific ATP-binding cassette transporter system. J. Biol. Chem. 278:45680-45689.
34. Okeke, I. N., A. Lamikanra, J. Czeczulin, F. Dubovsky, J. B. Kaper, and J. P. Nataro. 2000. Heterogeneous virulence of enteroaggregative Escherichia coli strains isolated from children in Southwest Nigeria. J. Infect. Dis. 181:252-260.
35. Pabst, W. L., M. Altwegg, C. Kind, S. Mirjanic, D. Hardegger, and D. Nadal. 2003. Prevalence of enteroaggregative Escherichia coli among children with and without diarrhea in Switzerland. J. Clin. Microbiol. 41:2289-2293.
36. Reisner, A., J. A. J. Haagensen, M. A. Schembri, E. L. Zechner, and S. Molin. 2003. Development and maturation of Escherichia coli K-12 biofilms. Mol. Microbiol. 48:933-946.
37. Rutherford, K., J. Parkhill, J. Crook, T. Horsnell, P. Rice, M.-A. Rajandream, and B. Barrell. 2000. Artemis: sequence visualization and annotation. Bioinformatics 16:944-945.
38. Sambrook, J., and D. W. Russell. 2001. Molecular cloning: a laboratory manual, 3rd ed. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y.
39. Sarantuya, J., J. Nishi, N. Wakimoto, S. Erdene, J. P. Nataro, J. Sheikh, M. Iwashita, K. Manago, K. Tokuda, M. Yoshinaga, K. Miyata, and Y. Kawano. 2004. Typical enteroaggregative Escherichia coli is the most prevalent pathotype among E. coli strains causing diarrhea in Mongolian children. J. Clin. Microbiol. 42:133-139.
40. Savarino, S. J., A. Fasano, J. Watson, B. M. Martin, M. M. Levine, S. Guandalini, and P. Guerry. 1993. Enteroaggregative Escherichia coli heat-stable enterotoxin 1 represents another subfamily of E. coli heat-stable toxin. Proc. Natl. Acad. Sci. USA 90:3093-3097.
41. Sheikh, J., J. R. Czeczulin, S. Harrington, S. Hicks, I. R. Henderson, C. LeBouguenec, P. Gounon, A. Phillips, and J. P. Nataro. 2002. A novel dispersin protein in enteroaggregative Escherichia coli. J. Clin. Investig. 110:1329-1337.
42. Sheikh, J., S. Hicks, M. Dall'Agnol, A. D. Phillips, and J. P. Nataro. 2001. Roles for Fis and YafK in biofilm formation by enteroaggregative Escherichia coli. Mol. Microbiol. 41:983-997.
43. Srimanote, P., A. W. Paton, and J. C. Paton. 2002. Characterization of a novel type IV pilus locus encoded on the large plasmid of locus of enterocyte effacement-negative Shiga-toxigenic Escherichia coli strains that are virulent for humans. Infect. Immun. 70:3094-3100.
44. Steiner, T. S., J. P. Nataro, C. E. Poteet-Smith, J. A. Smith, and R. L. Guerrant. 2000. Enteroaggregative Escherichia coli expresses a novel flagellin that causes IL-8 release from intestinal epithelial cells. J. Clin. Investig. 105:1769-1777.
45. Strom, M. S., and S. Lory. 1993. Structure-function and biogenesis of the type IV pili. Annu. Rev. Microbiol. 47:565-596.
46. Suzart, S., B. E. C. Guth, M. Z. Pedroso, U. M. Okafor, and T. A. T. Gomes. 2001. Diversity of surface structures and virulence genetic markers among enteroaggregative Escherichia coli (EAEC) strains with and without the EAEC DNA probe sequence. FEMS Microbiol. Lett. 201:163-168.
47. Taylor, R. K., V. L. Miller, D. B. Furlong, and J. J. Mekalanos. 1987. Use of phoA gene fusions to identify a pilus colonization factor coordinately regulated with cholera toxin. Proc. Natl. Acad. Sci. USA 84:2833-2837.
48. Watnick, P. I., and R. Kolter. 1999. Steps in the development of a Vibrio cholerae El Tor biofilm. Mol. Microbiol. 34:586-595.
49. Yoshida, T., S.-R. Kim, and T. Komano. 1999. Twelve pil genes are required for biogenesis of the R64 thin pilus. J. Bacteriol. 181:2038-2043.
50. Zhang, X.-L., I. S. M. Tsui, C. M. C. Yip, A. W. Y. Fung, D. K.-H. Wong, X. Dai, Y. Yang, J. Hackett, and C. Morris. 2000. Salmonella enterica serovar Typhi uses type IVB pili to enter human intestinal epithelial cells. Infect. Immun. 68:3067-3073.(Edward G. Dudley, Cecilia)
Universidade Federal de Sao Paulo, Sao Paulo, Brazil
Groupe de Genetique des Biofilms, Institut Pasteur, URA CNRS 2172, 25 rue du Docteur Roux, 75724 Paris Cedex 15, France
Bacteriology Section, Instituto de Salud Publica, Santiago, Chile
ABSTRACT
Enteroaggregative Escherichia coli (EAEC) is defined by a characteristic "stacked-brick" aggregative adherence (AA) pattern to cultured cells. In well-studied EAEC prototype strains (called typical EAEC strains), the AA phenotype requires aggregative adherence fimbriae (AAFs). However, previous studies suggest that known AAF alleles are not found in all EAEC strains. To define mechanisms contributing to adherence in an atypical strain, we studied EAEC strain C1096. An E. coli K12 derivative carrying two plasmids, designated pSERB1 and pSERB2, from C1096 adhered to cell lines and exhibited an AA pattern. Nucleotide sequence analysis of pSERB1 indicated that it is related to plasmids of the IncI1 incompatibility group. These plasmids encode genes involved in pilus-mediated conjugal transfer, as well as pilL-V, which encodes a second pilus of the type IV family. Insertional inactivation of the gene predicted to encode the major type IV pilin subunit (pilS) reduced conjugal transfer of the plasmid by 4 orders of magnitude. Adherence of the mutant strain to polystyrene and to HT29 cells was reduced by approximately 21% and 75%, respectively. In a continuous-flow microfermentor, the pilS inactivation reduced mature biofilm formation on a glass slide by approximately 50%. In addition, the simultaneous presence of both pSERB1 and pSERB2 plasmids promoted pilS-independent biofilm formation. We conclude that the IncI1 plasmid of EAEC C1096 encodes a type IV pilus that contributes to plasmid conjugation, epithelial cell adherence, and adherence to abiotic surfaces. We also observe that AA can be mediated by factors distinct from AAF adhesins.
INTRODUCTION
Enteroaggregative Escherichia coli (EAEC) is an emerging diarrheal pathotype defined by its characteristic "stacked-brick" aggregative adherence (AA) pattern to semiconfluent HEp-2 cells (30). This pattern results from adherence to cultured cells, other bacteria, and surrounding abiotic surfaces. For the prototype EAEC strain 042, this phenotype is mediated by the aggregative adherence fimbria (AAF) designated AAF/II. Three other AAF alleles are defined, and current data suggest that others exist (2, 9, 31). AAF fimbriae are necessary ex vivo for adherence of EAEC 042 to intestinal epithelia, abiotic surfaces, and cultured cells (8, 31, 42). Following colonization, EAEC may release cytotoxins, resulting in cellular damage (14, 40), and elicit an intestinal inflammatory response, which is stimulated at least in part by the bacterial flagellum (44).
Distribution studies previously identified numerous plasmid and chromosomal genes that are well conserved in EAEC strains isolated from distinct geographic locations. These include genes for dispersin (aap), which is involved in dispersing bacterial aggregates (41); aatA, which encodes a TolC-related transport system for dispersin (33); an AggR-regulated chromosomal gene (E. G. Dudley and J. P. Nataro, unpublished data); and the AraC-family transcriptional regulator aggR (32). These genes, including aggR itself (R. Sohoni and J. P. Nataro, unpublished data), are transcriptionally activated by AggR, and we hypothesize that the aggR regulon constitutes a package of linked virulence genes defining typical EAEC strains (30). Some clinical studies report that aggR-positive EAEC strains are isolated more frequently from individuals presenting with diarrhea and elevated fecal cytokine levels (19, 34, 35), supporting the hypothesis that the presence of the aggR regulon distinguishes pathogenic from nonpathogenic strains.
In contrast, atypical EAEC strains (30) lack all genes of the aggR regulon. It is suspected that the majority of these strains are nonpathogenic (6, 39). Atypical EAEC strains also lack genes for the currently defined AAFs, as do at least one-third of all EAEC strains (2). Therefore, either certain EAEC strains encode distantly related AAFs or there exist AAF-independent mechanisms for bacteria to adhere in an AA phenotype. Support for this latter hypothesis was recently reported for EAEC 236, which encodes an outer-membrane protein that confers AA upon E. coli K-12 (28).
Type IV pili are a group of well-studied extracellular structures defined by related phylogeny and a conserved pathway of pilus biogenesis (45). These pili are implicated in numerous phenotypes in gram-negative pathogens, including adherence (12, 47), twitching motility (26), DNA transformation (13), and conjugation (20). The role of a type IV pilus in promoting conjugation is best described for the Salmonella plasmid designated R64 (17, 49), which is a member of the IncI1 plasmid incompatibility group. The IncI1 family encodes a large number of genes mediating the synthesis of two different pili, defined as the thick pilus and the thin pilus (3). The former is essential for conjugal transfer of R64, while the thin pilus, which is categorized as type IV, is needed only when conjugation occurs in liquid medium (3). The thin pilus locus encompasses 14 genes, designated pilI through pilV, and 12 of these genes are essential for pilus biogenesis (49). A number of studies define the role of the thin pili in mediating surface interactions between bacteria (16-18); however, the IncI1-encoded type IV pilus has not been implicated in phenotypes apart from conjugation.
Here, we report the first studies on a proven pathogenic atypical EAEC strain, targeting the multiple-antibiotic-resistant strain implicated in a 1995 Serbian neonatal ward outbreak (5). This strain, designated C1096, is probe negative for aggR and all currently known AggR regulon members (9). We found the following: (i) EAEC C1096 carries a large plasmid (pSERB1) that encodes a functional type IV pilus and is related to plasmids of the IncI1 family; (ii) type IV pilus synthesis is required for optimal conjugation of the large plasmid; (iii) the presence of pSERB1 mediates adherence and biofilm formation of C1096 via expression of type IV pili; (iv) pSERB1, in cooperation with the small cryptic plasmid pSERB2, also mediates type IV pilus-independent biofilm formation; and (v) the genes encoding the type IV pilus are found in approximately 10% of both typical and atypical EAEC strains. This paper therefore reports two previously undescribed phenotypes conferred by an IncI1 plasmid: adherence to epithelial cells in a type IV pilus-dependent manner and adherence to abiotic surfaces via a type IV pilus-independent mechanism.
MATERIALS AND METHODS
Strains, plasmids, and antibiotics. Strains and plasmids used in this study are listed in Table 1. Bacteria used in colony hybridization experiments were described previously (9) or are clinical isolates from our laboratory stocks. Antibiotics, when used, were incorporated into solid and liquid media at the following concentrations: trimethoprim (TMP), 10 μg ml–1; ampicillin, 60 μg ml–1; chloramphenicol, 25 μg ml–1; kanamycin (KAN), 20 μg ml–1; streptomycin (STR), 50 μg ml–1; and nalidixic acid, 10 μg ml–1.
Cell culture techniques. All cell cultures were maintained in humidified 5% CO2 at 37°C. HEp-2 cells were grown in low-glucose Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% (vol/vol) fetal calf serum, penicillin (100 U ml–1), and STR (100 μg ml–1) (all from Gibco-BRL). HT29/C1 human colonic cells were grown in high-glucose DMEM supplemented with 10% fetal calf serum, apotransferrin (10 μg ml–1) (Sigma), STR (50 μg ml–1), and penicillin (50 U ml–1).
Molecular biology protocols. DNA manipulations were performed using standard laboratory protocols (38). Total DNA was isolated from bacterial strains by the cetyltrimethylammonium bromide-NaCl method (1). PCRs were performed using either the Elongase Amplification System (Invitrogen) or Taq DNA polymerase (New England Biolabs).
Colony hybridizations were performed following the method of Sambrook and Russell (38), using high-stringency hybridization conditions (60°C; 5x SSC [1x SSC is 0.15 M NaCl plus 0.015 M sodium citrate], 0.1% N-lauroylsarcosine, 0.5% blocking reagent, 0.02% sodium dodecyl sulfate in prehybridization and hybridization buffers) or low-stringency conditions (42°C; 10% formamide in prehybridization and hybridization buffers). Gene probes were synthesized using Ready-To-Go DNA Labeling Beads (Amersham Pharmacia Biotech, Inc.). Template DNA for pilL, pilN, pilQ, pilS, pilT, and pilV was synthesized by PCR, using the primer pairs 5'-CTTCTGCCGTCATTACTCA-3' and 5'-CATCGATGCT GTAGTTGACA-3', 5'-AGTTCTGACACGACACTTCC-3' and 5'-CTGGTTGATGGTTTTCAGAT-3', 5'-ACGTCATAAAGGGAACACAG-3' and 5'-GACCTGACCACTGTTGATTT-3', 5'-ATGAGCGTCATAACCTGTTC-3' and 5'-CTGTTGGTTTCCAGTTTGAT-3', 5'-TGTTATGCTCACTTCCTTCC-3' and 5'-CTAACGTCCCTGCATGTAA-3', and 5'-CTTTGGGCAGAGTTATACCA-3' and 5'-TTAATTGAGCGTTACACACG-3', respectively. Amplification conditions were as follows: 95°C for 3 min and 30 cycles of 94°C for 45 s, 50°C for 30 s, and 72°C for 1 min, followed by a final extension of 72°C for 7 min.
Southern hybridizations were used to confirm successful construction of the pilS mutant of C1096 (see below), using the pilS gene probe described above. Total DNA was digested with AflIII (New England Biolabs), and fragments were separated on a 0.7% agarose gel. DNA transfer to a Hybond-N+ membrane (Amersham), prehybridization, hybridization, and radioactive detection were performed using standard methods (38). The prehybridization and hybridization temperature was 60°C. To construct derivatives of E. coli DH5 carrying plasmids from strain C1096, an EZ::TN
Tagging of pSERB1 with a KAN resistance marker was achieved using a one-step recombination method (29). A PCR product encoding a KAN resistance marker and 50 nucleotides complementary to a putative noncoding region of pSERB1 on the 5' and 3' ends was generated using the primers 5'-AGACGATTGTTTTCCTCTCGTAGTTTTGTCCACTCTAAACAAAGTGTTATCATATGAATATCCTCCTTAG-3' and 5'-GATCACAGGAAATGGAATAAACACTGTAACTGTCAATGGAAAAGTGAAACGTGTAGGCTGGAGCTGCTTC-3'. The aph-encoding plasmid pKD4 (10) was used as template DNA, and the reaction conditions were 95°C for 2 min; 10 cycles of 94°C for 45 s, 55°C for 30 s, and 72°C for 90 sec; and 15 cycles of 94°C for 45 s, 68°C for 30 s, and 72°C for 90 s. The PCR product was transformed into E. coli (pSERB1, pSERB2, and pKM208), and integrants were selected for on LB+KAN plates. Loss of pKM208 was confirmed by replica plating colonies onto LB+KAN and LB+KAN+ampicillin plates. Integration of the PCR product in the expected location was confirmed using PCR and the primers 5'-GCGTCAATAAGAGACGTTGT-3' and 5'-ATAAACATGAACGGGCATAC-3'. Reaction conditions were 95°C for 10 min and 30 cycles of 94°C for 45 s, 55°C for 30 s, and 72°C for 2 min 15 s, followed by 72°C for 7 min. Template DNA consisted of 1 μl of an overnight grown bacteria culture. A 0.4-kb or 2.0-kb PCR product was generated when the template was pSERB1 or pSERB1::aph, respectively. This method was also followed to construct pSERB3.
Quantification of the pSERB1-encoded genes traN and traY and chloramphenicol acetyltransferase (cat), which is not encoded on pSERB1, was accomplished using real-time PCR. Total DNA from strains E. coli C1096 and C1096pilS::dfrA was isolated using the Instagene matrix (Bio-Rad). Real-time PCR was performed in a Chromo4 Real-Time PCR Detector (Bio-Rad), using a Quantitect SYBR Green PCR kit (QIAGEN). The following primer pairs were used: cat, 5'-CCACTCATCGCAGTACTGTT-3' and 5'-TCACTGGATATACCACCGTT-3'; traN, 5'-GGAATGTGACGGTCTATCTC-3' and 5'-GACTGTCGATTTCCTGACTC-3'; and traY, 5'-ATGGGGATAGTCTGGTTTCT-3' and 5'-TCCACTCACGACACTGAGTA-3'.
DNA sequencing and bioinformatic analysis. Plasmids from E. coli (pSERB1 and pSERB2) were isolated by CsCl density gradient centrifugation, sonicated using a Fisher Scientific 550 Sonic Dismembrator, and end repaired with T4 DNA polymerase, Klenow DNA polymerase, and T4 polynucleotide kinase (38). DNA was separated on a 0.6% agarose gel, and fragments between 3 to 4 kb were isolated with a QIAquick Gel Extraction Kit (QIAGEN). Fragments were ligated into SmaI-digested pUC18 and transformed into DH5. Sequencing reactions were performed using BigDye Terminator Sequencing Reagents (Applied Biosystems) and were analyzed using an Applied Biosystems 3700 DNA analyzer.
DNA sequences were assembled and edited using a Lasergene software package (DNASTAR, Inc.). Assembled contigs were annotated using the Artemis package (37), and open reading frame identification was assisted by the GeneMark prediction program (http://opal.biology.gatech.edu/GeneMark) using the training set generated for enterohemorrhagic E. coli EDL933. Sequence alignments and amino acid motifs were identified using BLAST (http://www.ncbi.nlm.nih.gov/BLAST) and the BLOSUM62 matrix. Nucleotide sequence overlaps between pSERB1 contigs and R64 were generated using BLASTALL from the NCBI website (http://www.ncbi.nlm.nih.gov) and were visualized using the Artemis Comparison Tool from the Sanger Centre (http://www.sanger.ac.uk).
Construction of pSERB1 derivatives carrying pilS::dfrA and pilS::cat mutations. A plasmid containing dfrA inserted into pilS was constructed by a modification of the overlap PCR protocol of Merlin et al. (27). Total DNA (50 ng) from EAEC C1096 was used as a template in a 20-μl reaction mixture. A PCR fragment encoding the 5' end of pilS and upstream DNA was generated using the primers 5'-AAAAAAGTCGACCGGATTTTGGTAAAAAGAAA-3' (Nat1913) and 5'-CGCTCTTGCGGCCGCTTGGAACGGGTCATGCCTTTATCATCGTT-3' (Nat1916). A PCR fragment encoding the 3' end of pilS and downstream DNA was generated using primers 5'-AAAAAAGTCGACAGGCCTATGGTTTCTGTTTT-3' (Nat1914) and 5'-CCGTTCCAAGCGGCCGCAAGAGCGGGCAATGTGACGAATAAAGT-3' (Nat1915). The 3' end of the former product and the 5' end of the latter product overlap by 24 bp, and an NotI site is located within this overlap (underlined in the sequences of Nat1915 and Nat1916). Reaction conditions were 94°C for 30 s, followed by 30 cycles of 94°C for 30 s, 55°C for 30 s, and 68°C for 1 min. A second round of PCR was performed using approximately 20 ng of each PCR product from the prior reactions as template DNA and the primers Nat1913 and Nat1914. Reaction conditions were the same as before, except that an extension time of 2 min was used. The resulting PCR product was digested with SalI, ligated into similarly digested pCVD442, and transformed into DH5-pir. This plasmid was designated pPIL1096. Next, a PCR fragment encoding dihydrofolate reductase (dfrA) and 5' and 3' NotI sites was generated using primers 5'-AAAAAAGCGGCCGCGGATCACCGAAATCTTCATA-3' and 5'-AAAAAAGCGGCCGCGCATGCACGATTTGTAATAA-3' and the EZ::TN
The PCR product was digested with NotI and ligated to NotI-digested pPIL1096, forming pPIL1096dfrA. This plasmid was transformed into E. coli SM10-pir and was conjugated into EAEC C1096 (11). Sucrose-, chloramphenicol-, and TMP-resistant transconjugants were screened by PCR using primers Nat1913 and Nat1914; template DNA from wild-type C1096 generates a 2.7-kb product, whereas mutants containing the dfrA insert generate a 3.1-kb product. Putative mutants were further screened by Southern hybridization. The pilS::dfrA derivative of pSERB1 was tagged with aph as described above and was designated pSERB3.
A chloramphenicol acetyltransferase (cat) insertion into pilS was generated by amplifying cat using the primers 5'-AAAAAAGCGGCCGCGTGTAGGCTGGAGCTGCTTA-3' and 5'-ATATATGCGGCCGCCATATGAATATCCTCCTTAG-3' and pKD3 (10) as template DNA. Cycling conditions were 25 cycles of 94°C for 45 s, 55°C for 30 s, and 72°C for 70 s. The product was digested with NotI and ligated into similarly digested pPIL1096. The resulting plasmid was used as template DNA in a PCR using primers Nat1913 and Nat1914 under the conditions described previously. The resulting PCR product, encoding pilS::cat and flanking DNA, was transformed into DH5(pSERB1::aph, pSERB2, pKM208) and integrated by the one-step recombination method (29). Proper integration was verified by PCR using primers 5'-TACGTTTTACGAGTCCCTTG-3' and 5'-TGTCTGTTACACCGCAGATA-3'. Reaction conditions were 94°C for 45 s, 55°C for 30 s, and 72°C for 4 min. A wild-type product generated a PCR product of 2.7 kb, while a mutant generated a 3.4-kb product. The resulting plasmid was designated pSERB4.
Scanning electron microscopy. Bacteria were grown overnight in LB with appropriate antibiotics and were inoculated to 2% in high-glucose DMEM and grown in 24-well polystyrene cell culture dishes containing a 1-cm2 glass coverslip. After approximately 4 h of growth, the medium was removed, and nonadherent bacteria were removed by five washes with phosphate-buffered saline (PBS). Cells were fixed to the coverslips for 15 min with PBS containing 2% glutaraldehyde and 3 mM CaCl. Coverslips were sent to the Johns Hopkins School of Medicine Microscopy Facility for analysis.
Bacterial adherence assays. Adherence of bacteria to glass and polystyrene surfaces was performed using a previously described crystal violet-based method (42).
For qualitative adherence assays, cultured cells were grown on glass coverslips in 24-well tissue culture plates. T84 cells were grown to 100% confluence (8 to 10 days) and were subsequently incubated for an additional 2 to 3 days. HEp-2 and HT29 cells were grown to approximately 90% confluence. Overnight bacterial broth cultures were diluted 1:50 in culture medium supplemented to 1% methyl--D-mannopyranoside (to inhibit binding by type I fimbriae) and incubated with cells for 2.5 to 6 h at 37°C. After this period, cells were thoroughly washed with PBS to remove nonadherent bacteria, fixed in 70% methanol for 5 min, stained with 10% Giemsa for 15 min, and observed by light microscopy.
For quantitative adherence assays using plate counts, HT29 cells were grown to approximately 90% confluence. Cells were washed twice with PBS, and 1 ml of high-glucose DMEM supplemented to 1% methyl--D-mannopyranoside was added to each well. Overnight cultures of EAEC strains C1096 and C1096pilS::dfrA were diluted in PBS to an optical density at 600 nm (OD600) of 0.08 (approximately 2 x 106 CFU ml–1), and 10 μl of each was added to wells. After a 2.5-h incubation at 37°C, half of the coverslips were moved to a new 24-well plate and washed with five 1-ml aliquots of PBS. After the last PBS aliquot was aspirated, 1 ml of 0.1% Triton X-100 was added, mixed thoroughly, and allowed to stand for approximately 5 min. Serial dilutions were made on LB and LB+TMP plates to determine the total number of adherent bacteria and pilS mutant cells, respectively. To the other half of the culture plate (the half for which the coverslips were not removed), 10 μl of 10% Triton X-100 was added. Serial dilutions were made on LB and LB+TMP plates to determine the total number of bacteria and total number of pilS mutant cells, respectively, per well. The ratio of adherent wild-type to mutant bacteria was calculated by dividing the number of adherent wild-type C1096 bacteria by the number of adherent C1096pilS::dfrA bacteria and normalizing to the total CFU of each strain in the wells. In control experiments, no difference in CFU ml–1 was found when serial dilutions of C1096pilS::dfrA were plated on LB or LB+TMP, indicating that differences in adherence calculated for this strain were not due to suppression of growth of C1096pilS::dfrA by TMP. Additionally, no difference in the growth rate was observed between C1096 and C1096pilS::dfrA in DMEM.
For quantitative adherence assays using direct cell counting, HT29 cells and bacteria were grown, and infections were performed as described above. After the PBS washes, coverslips were fixed with 70% methanol for 5 min and stained with 10% Giemsa. Slides were visualized by light microscopy. A total of 100 HT29 cells were randomly chosen from each of five coverslips from the wild-type and mutant strain infections, and the number of bacteria adhering to each HT29 cell were counted. Data are reported as the number of wild-type bacteria adhering to 100 HT29 cells divided by the number of mutant bacteria adhering to 100 HT29 cells.
Biofilm formation assay in microfermentors. Sixty-milliliter microfermentors, each containing a removable Pyrex glass slide, were configured as continuous-flow culture bioreactors with a flow rate of 40 ml h–1. This flow rate minimizes planktonic growth of the bacteria, and most of the observed growth occurs on the available surfaces inside the microfermentors and on the glass slide (15). (For a more detailed explanation, see http://www.pasteur.fr/recherche/unites/Ggb/biofilmfermenter.htm.) Overnight cultures grown in 0.4% glucose M63B1 minimal medium supplemented with appropriate antibiotics were used to inoculate microfermentors containing the same medium to 1 OD600 equivalent. The bacteria were cultivated for 24 to 72 h. Biofilm biomass on the internal Pyrex glass slide was quantified by measuring the OD600 after resuspension of the slide biofilm in 10 ml of M63B1 medium. Total biofilm was measured at OD600 after resuspension of the total cell mass present within the microfermentor in 60 ml of M63B1 medium. All experiments were performed at least in triplicate.
Plasmid conjugation. Plasmid transfer for strain construction was performed by individually inoculating donor and recipient strains from an overnight culture to 1% in LB and growing them at 37°C for 1 h. Aliquots (250 μl) of donor and recipient strains were combined, centrifuged, and resuspended in 50 μl of LB. Cells were spotted on LB agar and incubated at 37°C for 6 h, and transconjugants were plated on selective medium. When DH5 or HB101 was the recipient strain, selection was performed on LB+KAN+nalidixic acid or LB+KAN+STR plates, respectively. Plasmid transfer was confirmed by agarose gel electrophoresis. The conjugation frequency of pSERB1::aph and pSERB3 from DH5 to HB101 on a solid surface was assessed similarly, except that the resuspended cells were placed on a 0.22-μM-pore-size filter overlaying an LB plate. Serial dilutions were made onto LB+KAN+STR plates to quantify the number of transconjugants and on LB+STR plates to quantify the number of recipients. Conjugation frequency is reported as the number of transconjugants per recipient.
Quantitative broth conjugations were performed by centrifuging 500-μl aliquots of overnight cultures, washing them twice with 1 ml of LB broth, and inoculating each strain into 2 ml of LB broth to an OD600 of approximately 0.03. Tubes inoculated with only one strain served as negative controls. Conjugations were carried out in 14-ml plastic tubes. Tubes were placed on their sides and rotated at 50 rpm for 6 h at 37°C. Plate counts were performed as described above for solid surface conjugation.
Nucleotide sequence accession numbers. The partial nucleotide sequence obtained for pSERB1 was deposited in the GenBank database under accession number AY686591. The full nucleotide sequence for pSERB2 was assigned accession number DQ269444.
RESULTS
EAEC C1096 adheres to epithelial cells and glass in an AA pattern. Previously, Cobeljic et al. implicated EAEC C1096 as the causative agent of a nonbloody diarrhea outbreak in Ni, Serbia. The authors reported that EAEC C1096 adhered mainly to the HEp-2 cells, with few bacteria adhering to the glass coverslips. Whereas prototype EAEC strain 042 requires AAF/II fimbriae for AA to abiotic surfaces and to human intestinal tissue (8, 42), C1096 does not encode genes for the expression of AAF/II or the distantly related fimbriae AAF/I (9). Additionally, total DNA from C1096 does not hybridize to a DNA probe (2) that identifies the usher subunit of AAF/I, AAF/II, AAF/III, and probably other as yet unidentified AAF alleles (data not shown). This suggests that AAF-independent determinants may be involved in C1096 adherence.
To characterize adherence factors of EAEC C1096, we began by confirming a previous report that this strain adhered to HEp-2 cells in an AA pattern, which is the defining characteristic of EAEC (Fig. 1A), and studying whether C1096 adhered to glass similarly to the prototype strain 042. In our hands, EAEC 042 and C1096 both displayed AA on HEp-2 cells (Fig. 1A and B, respectively). Strain C1096 also adhered to glass (Fig. 1D) in an AA pattern, although the pattern observed differed from the distinct honeycomb structure seen with EAEC 042 (Fig. 1C). Adherence to an abiotic surface was also measured quantitatively, with EAEC 042 and a biofilm-deficient derivative designated JS28 (42) serving as positive and negative controls, respectively. EAEC C1096 adhesion was weaker than that observed with strain 042 (Fig. 2), although it was still significantly higher than that seen with strain JS28. Thus, EAEC C1096 adheres to HT29 cells and abiotic surfaces by the AA phenotype, confirming the classification of C1096 as EAEC.
Plasmids from C1096 confer adherence upon DH5. EAEC C1096 carries at least six plasmids (Fig. 3). To determine if any plasmids encode adherence factors, we used an in vitro transposition reaction to randomly insert a dfrA-encoding transposon into the plasmids and transformed the tagged plasmids into E. coli DH5. All transformants screened carried at least one of the three lower-molecular-mass plasmids (1.6, 2.5, and 6.4 kb). One isolate was identified that carried one of the large plasmid species from C1096, in addition to a second plasmid that migrated in agarose gel electrophoresis at approximately 2.5 kb (Fig. 3, lane 3). The larger and smaller plasmid species from this strain were designated pSERB1 and pSERB2, respectively. Upon electroporation of plasmid pools from DH5(pSERB1, pSERB2) back into DH5, all TMP-resistant transformants carried pSERB2 only. This suggests that during strain construction, the 0.9-kb dfrA-encoding transposon inserted into the 1.6-kb plasmid from C1096, resulting in a 2.5-kb species, and this plasmid cotransformed with pSERB1 during electroporation of DH5.
DH5 transformants carrying different plasmid combinations from C1096 were screened for the ability to adhere to HEp-2 cells. DH5(pSERB1, pSERB2) adhered to cultured cells in the presence of 1% methyl--D-mannopyranoside (Fig. 4B). The bacteria were typically found in clusters, adhering to the cultured cells as well as to each other. DH5(pSERB1, pSERB2) adhered to a smaller number of HEp-2 cells than did C1096 and did not display the extensive adherence seen with strain C1096 (compare Fig. 4B with 1B). A much lower level of adherence was seen with wild-type DH5 (Fig. 4A) and with derivatives carrying any of the three low-molecular-mass plasmids from C1096 (data not shown). These observations suggest that at least one factor that contributes to the adherence of EAEC C1096 to HEp-2 cells is encoded on pSERB1 and/or pSERB2.
Adherence of DH5(pSERB1, pSERB2) to abiotic surfaces requires both plasmids. In order to test the adherence of DH5(pSERB1, pSERB2) on abiotic surfaces, we tested this strain for initial adhesion in microtiter plate assays and for mature biofilm formation in microfermentors. Although adherence of DH5(pSERB1, pSERB2) was not observed on polystyrene tissue culture plates (data not shown), DH5(pSERB1, pSERB2) formed a strong mature biofilm in a continuous-culture microfermentor, where the formation of a thick biofilm can be monitored on removable Pyrex glass slides between 48 to 72 h (Fig. 5A). To determine which plasmid was responsible for this phenotype, pSERB1 was tagged in a predicted noncoding region with the KAN resistance marker aph, creating pSERB1::aph, and pSERB2 was electroporated into both DH5 and DH5(pSERB1::aph). Adherence of DH5(pSERB1::aph) and DH5(pSERB2) to the spatula was significantly decreased compared to DH5(pSERB1::aph, pSERB2) (Fig. 5A and B), although adherence of each was greater than DH5. These results demonstrate that both plasmids are required for fully mature biofilm formation under the conditions tested and that pSERB1 and pSERB2 promote biofilm formation in a synergistic manner.
pSERB1 is related to IncI1 plasmids, and pSERB2 is a cryptic plasmid. To identify adherence and other possible virulence factors of EAEC C1096, the plasmids from DH5(pSERB1, pSERB2) were subjected to nucleotide sequence analysis. A total of 742 shotgun sequences were assembled into five contigs of 67.2, 7.2, 7.1, 2.5, and 1.1 kb, with an estimated sequence coverage of 5.9-fold.
A BLAST analysis indicated that the 2.5-kb contig encoded dfrA and, therefore, is derived from pSERB2. This contig also encoded two plasmid replication protein homologs. Further analysis of the pSERB2 sequence did not provide any additional insights into the requirement of both plasmids for the adherence phenotype observed with DH5(pSERB1::aph, pSERB2).
The remaining four contigs were all aligned by BLAST to plasmids of the IncI1 incompatibility group, including R64 (GenBank accession no. NC_005014) of Salmonella enterica serovar Typhimurium. EcoRI digests suggested that pSERB1 is approximately 82 kb in size and that the 7.2-kb, 1.1-kb, 7.1-kb, and 67.2-kb contigs (total, 82.6 kb) were aligned with nucleotides 48284 to 52639, 54932 to 56052, 56043 to 62909, and 63385 to 102094, respectively, of R64 (data not shown). These data strongly suggest that nearly all of the sequence of pSERB1 was obtained.
pSERB1 encodes a conjugal transfer locus and a type IV pilus system. R64 encodes a large number of genes responsible for the formation of two pilus types referred to as thin pili and thick pili (3). Thick pilus biosynthesis requires genes encoded on two distinct regions of R64: traBC, which encodes regulatory proteins, and a 31-kb region encoding 26 open reading frames, 22 of which are required for conjugal transfer in broth and on solid surfaces (23). Further analysis of the 67.2-kb contig revealed that homologs of these 22 proteins and TraBC were present on pSERB1, suggesting that pSERB1 may be a conjugative plasmid.
R64 thin pili are members of the type IV pilus family and are required for plasmid conjugation in broth only (22); biosynthesis requires 12 genes designated pilK-V (49). Genes similar to pilL-V are encoded on pSERB1 in the same order as on R64 (Fig. 6). Although no pilK was identified, characterization of type IV pili from Yersinia pseudotuberculosis and S. enterica serovar Typhi (7, 50) suggests that either some systems do not require pilK or certain type IV pilus-expressing organisms synthesize an unrelated protein with PilK activity. BLAST searches indicated that PilL-V are most closely related (48 to 91% amino acid identity) to a type IV pilus locus encoded on pO113, a plasmid isolated from a Shiga toxin-producing E. coli (STEC) strain (24).
Certain type IV pilus biogenesis systems encode two prepilin subunits, designated pilS (major subunit) and pilV (minor subunit). These pili are further divided into two classes based upon posttranslational processing events (25): type IVA, in which the prepilus has a short (usually 5 to 6 amino acids [aa]) leader sequence and the N-terminal amino acid residue of the mature protein is methylated phenylalanine; and type IVB, in which the prepilus has a longer (13 to 30 aa) signal sequence and the N-terminal amino acid residue is Met, Leu, Trp, or Ser. Based upon predicted signal sequence cleavage sites, PilS from EAEC C1096 is predicted to be type IVB (predicted leader sequence of 22 aa; mature PilS is predicted to have an N-terminal Met), and PilV is type IVA (predicted leader sequence of 8 aa; mature PilV is predicted to have an N-terminal Phe). The deduced amino acid sequence of PilS is aligned by BLAST only with the PilS homologs from pO113 (Fig. 6) and plasmid R721 (accession no. NP_065349; 43% identical to pilS from R64 over 173 aa).
EAEC C1096 expresses surface pili absent in a pilS::dfrA mutant. To test whether the pSERB1-encoded pil locus is expressed, dfrA was inserted into the plasmid-encoded copy of pilS by homologous recombination, creating EAEC C1096pilS::dfrA. Such a mutation in plasmid R64 eliminates the assembly of functional type IV pili (49), and by analogy the same phenotype was expected with pSERB1. Under magnification of x15,000, numerous pili of approximately 1 μm in length were observed extending from the surface of most bacterial cells (Fig. 7). The pili appear to tether bacteria together over a distance, and direct interaction of the pili with the glass coverslip was also noted. In sharp contrast, similar surface pili were not seen on the surface of EAEC C1096pilS::dfrA cells, although a minority of these cells still expressed much shorter pili (Fig. 7B).
pSERB1 is a conjugative plasmid that requires the type IV pilus for optimal conjugal transfer in broth cultures. As noted, our sequence analysis suggested that pSERB1 is an IncI1-type plasmid. This finding implies that pSERB1 is conjugative, as are other IncI1 plasmids, and that efficient conjugation in broth, but not on solid surfaces, is promoted by the type IV pilus (22). To test this, the pSERB1-derivative from EAEC C1096pilS::dfrA was tagged with aph, creating pSERB3, and this construct was transferred to DH5. When matings were performed in LB broth at 37°C, the transfer frequency of pSERB1::aph and pSERB3 to E. coli HB101 after 6 h was 2.6 x 10–2 and 6.7 x 10–6 transconjugants per recipient, respectively. When matings were performed on a 0.22-μM-pore-size filter overlaying LB agar, transfer frequencies were 1.3 x 10–3 and 3.2 x 10–3, which was not a statistically significant difference (P = 0.17).
Because of the cooperative role of pSERB1 and pSERB2 in biofilm formation, we also tested whether the transfer frequency of pSERB1::aph was altered by the presence of pSERB2. However, no detectable difference was noted in the transfer frequency of pSERB1::aph to E. coli HB101 when either DH5(pSERB1::aph, pSERB2) or DH5(pSERB1::aph) was used as the donor strain. This suggests that the biofilm enhancement phenotype is not the result of increased conjugal transfer of pSERB1.
The type IV pilus contributes to adherence of EAEC C1096 to HT29 cells and to abiotic surfaces. Next, we tested whether the type IV pilus mediates adherence of EAEC C1096 to cells and abiotic surfaces. Adherence to HT29 cells was quantified both as single-strain infections and as a coinfection. First, HT29 cells were coinfected with equivalent numbers of C1096 and the pilS::dfrA mutant strains. We performed coinfections because we found that C1096 detaches epithelial cells from the glass coverslips during infection, and therefore differences seen in experiments performed as single-strain infections could be due to differences in the numbers of detached HT29 cells. Data from four independent experiments revealed 2.9-fold more CFU for the wild-type compared to the mutant strain, or a 65% decrease in adherence (P = 0.02 by a Student's pairwise t test). Direct counting of bacteria adhering to HT29 cells by microscopic examination of single-strain infections supported this result, as 5.4-fold fewer pilS::dfrA bacteria (82% reduction) were counted adhering to HT29 cells than with the wild-type bacteria. Therefore, the two different methods of quantifying bacterial adherence to cultured cells were in agreement and suggested that EAEC C1096 adherence to epithelial cells is reduced by approximately 75% when the type IV pilus is inactivated.
Qualitative examination of adherence to glass of the mutant and wild-type strains indicated that the pattern of biofilm formation is not strongly affected by the pilS mutation. Nevertheless, quantitative assays revealed a slight (21%) but statistically significant reduction of biomass adhering in wells containing the mutant strain compared to the wild type (Fig. 2) (P = 0.009 by a Student's pairwise t test). In continuous-flow culture, the inactivation of pilS did not affect the total biofilm biomass in the microfermentor but reduced biofilm development on the vertical internal glass slide by approximately 50% (Fig. 8) (P = 0.014), indicating that a more fragile biofilm was formed by the C1096pilS mutant. Since the wild-type C1096 and its pilS mutant have similar growth rates in liquid culture, this suggests that the type IV pilus encoded on pSERB1 affects the initial stages of adhesion and/or promotes tight bacteria-bacteria interactions within a biofilm.
Many conjugative plasmids, including those of the IncI1 group, are known to mediate biofilm formation in bacteria via a conjugation-dependent mechanism (15). The contribution of the type IV pilus subunit was therefore further investigated in the microfermentor-biofilm model. As shown in Fig. 5, biofilm formation by strains DH5(pSERB3) and DH5(pSERB4, pSERB2) was not significantly different than that formed by strains DH5 (pSERB1::aph) and DH5(pSERB1::aph, pSERB2), respectively. Therefore, adherence in the microfermentor assay does not depend upon efficient plasmid conjugation.
Lastly, we wanted to confirm that phenotypes observed with E. coli C1096pilS::dfrA were not due to an altered copy number of pSERB1 or to altered expression of the pSERB1-encoded thick pilus. Real-time PCR using total DNA from C1096 and C1096pilS::dfrA as a template indicated that the quantity of traN and traY (two pSERB1-encoded genes) relative to that of cat (a gene not encoded on pSERB1) was the same for both strains. This indicates that the dfrA insertion does not affect plasmid copy number. Additionally, as reported above, the transfer frequency to E. coli HB101, when conjugation was performed on a 0.22-μM-pore-size filter, was comparable for DH5(pSERB1::aph) and DH5a(pSERB3). This suggests inactivation of pilS did not affect the synthesis of the thick pilus, which is essential for plasmid conjugation.
Taken together, these results indicate that the type IV pili from EAEC C1096 play a role in adherence to cultured epithelial cells and moderately contribute to both initial or late biofilm formation on abiotic surfaces. These data also suggest the presence of other adherence mechanisms in this strain.
Genes from the EAEC C1096 type IV pilus locus are found in both typical and atypical EAEC strains. Colony hybridizations were performed to determine whether other EAEC strains carry genes for type IV pilus biogenesis that are related to those found in EAEC C1096. A library consisting of 40 typical and 20 atypical EAEC strains was screened. Under both high-stringency (60°C hybridization) and low-stringency (42°C hybridization in 10% formamide) conditions, four of the typical EAEC strains were hybridization positive using the pilL, pilN, pilQ, pilS, pilT, and pilV probes (data not shown). These strains were previously designated H38-1, DS67-R2, 278-1, and 495-1 (9) and were isolated from Peru (H38-1), the Philippines (DS67-R2), and Thailand (278-1 and 495-1). Two atypical EAEC strains, designated strains "A" and 25-1, were also probe positive under high-stringency conditions with all six gene probes. Additionally, an atypical strain designated 6-1 was probe positive under high-stringency conditions only with pilN, pilL, and pilV, and a fourth strain designated DS61-R2 was only probe positive with pilV under low-stringency conditions. As expected, the prototype EAEC strain 042 was probe negative with all probes.
DISCUSSION
EAEC strains are defined as strains that adhere to HEp-2 cells in an aggregative pattern (30). While the majority of well-characterized strains adhere to abiotic surfaces (42), cultured cell lines (8, 31), and human intestinal tissue (8) via AAFs, it is clear that alternative mechanisms are also responsible for aggregative adherence patterns (2, 9, 28). While studying the adherence factors of EAEC strain C1096, we discovered that one of the large plasmids carried by this strain is related to members of the IncI1 incompatibility group and encodes genes for two previously described pili designated thick and thin (type IV) pili. We suggest here a model where the IncI1 plasmid-encoded type IV pili carried by EAEC C1096 contribute to, but are not essential for, adherence of this strain to cells and surfaces (Fig. 9). Additionally, this pilus mediates an increase in adherence of the bacteria to abiotic surfaces in a manner that is further enhanced by an unknown interaction with a small cryptic plasmid designated pSERB2 (Fig. 9A).
It is well established that IncI1-encoded type IV pili mediate conjugal transfer (Fig. 9A) of these plasmids (21). We are unaware, however, of any prior reports implicating the thin pilus from conjugative plasmids as a factor mediating adherence to intestinal epithelial cells (Fig. 9B); in fact Srimanote et al. (43) previously reported that the STEC plasmid pO113, which belongs to the IncI1 family, does not play a role in adherence to intestinal tissue. Some conjugative plasmids, including IncI1 members, also catalyze adherence to abiotic surfaces as the result of host-donor contact during DNA transfer (15, 36). Therefore, it is possible that the decreased adherence to HT29 cells observed with the C1096dfrA::pilS mutant is the result of decreased mating aggregate formation. However, given that C1096dfrA::pilS demonstrated a 65 to 82% decrease in adherence to HT29 cells, while only a 21% decrease in adherence to polystyrene, we favor a model in which the type IV pilus mediates adherence through direct contact with the eukaryotic cell membrane or other surface structure. The decreased adherence of the pilS mutant on polystyrene suggests that the pSERB1-encoded pilL-V may also augment biofilm levels or may accelerate biofilm formation, as suggested for the Vibrio cholerae type IV pilus (48). In summary, our data indicate that type IV pili encoded on some IncI1 plasmids may be multifunctional, serving roles in adherence to both host cells and abiotic surfaces and in plasmid conjugation.
We are interested in determining whether this observation can be extended to other IncI1 plasmids found in gram-negative bacterial pathogens. We would also like to further characterize the role of individual genes from the pilL-V locus and those encoding the thick pilus. However, given the antibiotic resistance profile of C1096 (including ampicillin, chloramphenicol, KAN, tetracycline, gentamicin, neomycin, and STR) (reference 5 and data not shown), the large number of plasmids present in this strain, and the current lack of knowledge concerning genetic manipulations in this strain, further studies are needed before investigators can use sophisticated genetic techniques that combine gene knockouts with complementation in trans using expression vectors. We are pursuing the development of such vectors rather than studying a more genetically tractable host, as C1096 is currently the only atypical EAEC strain that is a proven human pathogen.
Strains carrying IncI1 plasmids were previously shown to form biofilms in continuous-flow microfermentors (15). This phenotype required the presence of recipient bacteria, presumably because the expression of the plasmid-encoded conjugation apparatus and surface pili is repressed unless bacteria lacking the IncI1 plasmid are present in the population. We found that pure cultures of C1096 form a type IV pilus-mediated biofilm in the microfermentor, suggesting that this pSERB1-encoded pilus is expressed regardless of whether recipient bacteria are present. The continuous-flow culture experiments also showed that while both C1096 and C1096pilS::dfrA biofilms initially form on the internal vertical surfaces, the pilS mutant falls to the bottom of the microfermentor to a greater extent than the wild type during later stages of biofilm development. This suggests that the pilS mutant may form weaker bacteria-bacteria interactions, resulting in a more fragile biofilm under these conditions. Interestingly, biofilm formation by DH5(pSERB1::aph, pSERB2) in the microfermentor was more than fourfold greater than that of DH5(pSERB1::aph) (Fig. 9A), and this adherence did not depend upon the pSERB1-encoded type IV pilus or efficient conjugal transfer of pSERB1. Given the scarcity of genetic information found in the pSERB2 sequence, it is difficult to speculate as to the nature of this effect. While pSERB2 is mobilizable in the presence of pSERB1 (E. G. Dudley and J. P. Nataro, unpublished results), the frequency of pSERB2 conjugation is approximately 3 log units less than that of pSERB1. Therefore, it seems unlikely that the conjugative nature of pSERB2 alone contributes to increasing the number of mating aggregates. Plasmids related to pSERB2 from putative uropathogenic E. coli have been described by others (4), and thus studies concerning whether these plasmids modulate adherence in other species would be of interest.
Colony hybridization experiments indicated that approximately 10% of the EAEC strains screened encode genes related to pilL, pilN, pilQ, pilS, pilT, and pilV and that these genes are often inherited together. This distribution frequency is similar to that reported for the well-studied alleles of the AAF fimbriae, in which the AAF/I, AAF/II, and AAF/III are also detected only in a minority of EAEC isolates (2, 9). It is possible that type IV pili are more widely distributed in this EAEC library, assuming that some probe-negative strains encode distantly related genes. The prototype EAEC strain 042 was probe negative with all gene probes tested, and genes similar to pilL-V from pSERB1 are not found in the completed genome sequence of prototype EAEC strain 042 (http://www.sanger.ac.uk). As four of the probe-positive strains were also probe positive for aggR and other members of the aggR regulon (9), these data suggest that type IV pili related to those encoded on pSERB1 are not a defining characteristic of either typical or atypical EAEC but may serve as an accessory adhesin of EAEC. These genes are not as widely distributed as the related type IV pili of STEC, which were found in 16 of 19 strains that lack the locus of enterocyte effacement (43). Future work is directed at identifying additional adherence factors of atypical EAEC in order to determine whether this pathotype adheres to surfaces and cells via a common mechanism.
This study and previous work highlight a difficulty with the current method of defining the EAEC pathotype. The HEp-2 adherence assay is the current "gold standard" for identification; however, it has become clear in recent years that many strains exhibiting the EAEC "stacked-brick" pattern on cultured cells are distinct from the prototype strain 042 as determined by probe hybridization and/or multilocus enzyme electrophoresis (2, 9). While the HEp-2 adherence assay is a proven and reliable method for distinguishing EAEC from other E. coli pathotypes, genetic heterogeneity of isolates and variability of adherence patterns (2, 43, 46) highlight the need for molecular methods of identifying and characterizing EAEC strains. A growing body of evidence suggests that the EAEC 042 plasmid-encoded regulator AggR is a global regulator of virulence determinants (9, 35; E. G. Dudley and J. P. Nataro, unpublished data) and that genes of this regulon are detected more frequently in EAEC isolates from stools of patients with diarrhea than from healthy individuals (19, 35, 39). Therefore, the nomenclature of typical EAEC and atypical EAEC was recently proposed to differentiate EAEC strains carrying and lacking members of the aggR regulon, respectively (30), and genes of the aggR regulon were designated as markers defining typical EAEC. While the majority of atypical EAEC strains are believed to be nonpathogenic (39), outbreaks such as the one involving EAEC C1096 demonstrate that some atypical strains cause diarrhea and that AAFs are not the only factors that confer AA. The virulence factors of atypical EAEC are largely undefined, and this work represents the first study directed at defining the molecular pathogenesis of an AggR regulon-negative EAEC strain. Further studies to understand the pathogenic factors of EAEC C1096 and to characterize the similarities and differences between atypical EAEC strains are under way.
ACKNOWLEDGMENTS
This work was supported by grant AI033096 through the National Institute of Allergy and Infectious Diseases to J.P.N. J.-M. G. and P.L.-L. are supported by the Institut Pasteur and CNRS URA2172 and the Foundation BNP-PARIBAS grants.
We thank O. Colin Stine and Yuansha Chen (University of Maryland—Baltimore Department of Epidemiology and Preventative Medicine) for generating the pSERB1 shotgun sequence data and Brad Harris (Johns Hopkins School of Medicine) for taking the scanning electron microscopy pictures.
Present address: Institute of Child Health, University of Birmingham, B4 6LH Birmingham, United Kingdom.
REFERENCES
1. Ausubel, F. M., R. Brent, D. D. Moore, J. A. Smith, J. G. Siedman, and K. Struhl. 1999. Current protocols in molecular biology. John Wiley & Sons, New York, N.Y.
2. Bernier, C., P. Gounon, and C. Le Bouguenec. 2002. Identification of an aggregative adhesion fimbria (AAF) type III-encoding operon in enteroaggregative Escherichia coli as a sensitive probe for detecting the AAF-encoding operon family. Infect. Immun. 70:4302-4311.
3. Bradley, D. E. 1984. Characteristics and function of thick and thin conjugative pili determined by transfer-derepressed plasmids of incompatibility groups I1, I2, I5, B, K and Z. J. Gen. Microbiol. 130:1489-1502.
4. Burian, J., L. Guller, M. Macor, and W. W. Kay. 1997. Small cryptic plasmids of multiplasmid, clinical Escherichia coli. Plasmid 37:2-14.
5. Cobeljic, M., B. Miljkovic-Selimovic, D. Paunovic-Todosijevic, Z. Velickovic, Z. Lepsanovic, N. Zec, D. Savic, R. Ilic, S. Konstantinovic, B. Jovanovic, and V. Kostic. 1996. Enteroaggregative Escherichia coli associated with an outbreak of diarrhoea in a neonatal nursery ward. Epidemiol. Infect. 117:11-16.
6. Cohen, M. B., J. P. Nataro, D. I. Bernstein, J. Hawkins, N. Roberts, and M. A. Staat. 2005. Prevalence of diarrheagenic Escherichia coli in acute childhood enteritis: a prospective controlled study. J. Pediatr. 146:54-61.
7. Collyn, F., M.-A. Lety, S. Nair, V. Escuyer, A. Ben Younes, M. Simonet, and M. Marceau. 2002. Yersinia pseudotuberculosis harbors a type IV pilus gene cluster that contributes to pathogenicity. Infect. Immun. 70:6196-6205.
8. Czeczulin, J. R., S. Balepur, S. Hicks, A. Phillips, R. Hall, M. H. Kothary, F. Navarro-Garcia, and J. P. Nataro. 1997. Aggregative adherence fimbria II, a second fimbrial antigen mediating aggregative adherence in enteroaggregative Escherichia coli. Infect. Immun. 65:4135-4145.
9. Czeczulin, J. R., T. S. Whittam, I. R. Henderson, F. Navarro-Garcia, and J. P. Nataro. 1999. Phylogenetic analysis of enteroaggregative and diffusely adherent Escherichia coli. Infect. Immun. 67:2692-2699.
10. Datsenko, K. A., and B. L. Wanner. 2000. One-step inactivation of chromosomal genes in Escherichia coli K-12 using PCR products. Proc. Natl. Acad. Sci. USA 97:6640-6645.
11. Donnenberg, M., and J. Kaper. 1991. Construction of an eae deletion mutant of enteropathogenic Escherichia coli by using a positive-selection suicide vector. Infect. Immun. 59:4310-4317.
12. Donnenberg, M. S., J. A. Giron, J. P. Nataro, and J. B. Kaper. 1992. A plasmid-encoded type IV fimbrial gene of enteropathogenic Escherichia coli associated with localized adherence. Mol. Microbiol. 6:3427-3437.
13. Dubnau, D. 1999. DNA uptake in bacteria. Annu. Rev. Microbiol. 53:217-244.
14. Eslava, C., F. Navarro-Garcia, J. R. Czeczulin, I. R. Henderson, A. Cravioto, and J. P. Nataro. 1998. Pet, an autotransporter enterotoxin from enteroaggregative Escherichia coli. Infect. Immun. 66:3155-3163.
15. Ghigo, J. 2001. Natural conjugative plasmids induce bacterial biofilm development. Nature 412:442-445.
16. Ishiwa, A., and T. Komano. 2000. The lipopolysaccharide of recipient cells is a specific receptor for PilV proteins, selected by shufflon DNA rearrangement, in liquid matings with donors bearing the R64 plasmid. Mol. Gen. Genet. 263:159-164.
17. Ishiwa, A., and T. Komano. 2004. PilV adhesins of plasmid R64 thin pili specifically bind to the lipopolysaccharides of recipient cells. J. Mol. Biol. 343:615-625.
18. Ishiwa, A., and T. Komano. 2003. Thin pilus PilV adhesins of plasmid R64 recognize specific structures of the lipopolysaccharide molecules of recipient cells. J. Bacteriol. 185:5192-5199.
19. Jiang, Z. D., D. Greenberg, J. P. Nataro, R. Steffen, and H. L. DuPont. 2002. Rate of occurrence and pathogenic effect of enteroaggregative Escherichia coli virulence factors in international travelers. J. Clin. Microbiol. 40:4185-4190.
20. Kim, S., and T. Komano. 1997. The plasmid R64 thin pilus identified as a type IV pilus. J. Bacteriol. 179:3594-3603.
21. Komano, T. 1999. Shufflons: multiple inversion systems and integrons. Annu. Rev. Genet. 33:171-191.
22. Komano, T., N. Funayama, S. R. Kim, and T. Nisioka. 1990. Transfer region of IncI1 plasmid R64 and role of shufflon in R64 transfer. J. Bacteriol. 172:2230-2235.
23. Komano, T., T. Yoshida, K. Narahara, and N. Furuya. 2000. The transfer region of IncI1 plasmid R64: similarities between R64 tra and Legionella icm/dot genes. Mol. Microbiol. 35:1348-1359.
24. Leyton, D. L., J. Sloan, R. E. Hill, S. Doughty, and E. L. Hartland. 2003. Transfer region of pO113 from enterohemorrhagic Escherichia coli: similarity with R64 and identification of a novel plasmid-encoded autotransporter, EpeA. Infect. Immun. 71:6307-6319.
25. Manning, P. 1997. The tcp gene cluster of Vibrio cholerae. Gene 192:63-70.
26. Mattick, J. S. 2002. Type IV pili and twitching motility. Annu. Rev. Microbiol. 56:289-314.
27. Merlin, C., S. McAteer, and M. Masters. 2002. Tools for characterization of Escherichia coli genes of unknown function. J. Bacteriol. 184:4573-4581.
28. Monteiro-Neto, V., S. Y. Bando, C. A. Moreira-Filho, and J. A. Giron. 2003. Characterization of an outer membrane protein associated with haemagglutination and adhesive properties of enteroaggregative Escherichia coli O111:H12. Cell Microbiol. 5:533-547.
29. Murphy, K. C., and K. G. Campellone. 2003. Lambda Red-mediated recombinogenic engineering of enterohemorrhagic and enteropathogenic E. coli. BMC Mol. Biol. 4:11.
30. Nataro, J. P. 2003. Enteroaggregative Escherichia coli. In W. M. Scheld, J. M. Hughes, and B. E. Murray (ed.), Emerging infections 6. ASM Press, Washington, D.C.
31. Nataro, J. P., Y. Deng, D. R. Maneval, A. L. German, W. C. Martin, and M. M. Levine. 1992. Aggregative adherence fimbriae I of enteroaggregative Escherichia coli mediate adherence to HEp-2 cells and hemagglutination of human erythrocytes. Infect. Immun. 60:2297-2304.
32. Nataro, J. P., D. Yikang, D. Yingkang, and K. Walker. 1994. AggR, a transcriptional activator of aggregative adherence fimbria I expression in enteroaggregative Escherichia coli. J. Bacteriol. 176:4691-4699.
33. Nishi, J., J. Sheikh, K. Mizuguchi, B. Luisi, V. Burland, A. Boutin, D. J. Rose, F. R. Blattner, and J. P. Nataro. 2003. The export of coat protein from enteroaggregative Escherichia coli by a specific ATP-binding cassette transporter system. J. Biol. Chem. 278:45680-45689.
34. Okeke, I. N., A. Lamikanra, J. Czeczulin, F. Dubovsky, J. B. Kaper, and J. P. Nataro. 2000. Heterogeneous virulence of enteroaggregative Escherichia coli strains isolated from children in Southwest Nigeria. J. Infect. Dis. 181:252-260.
35. Pabst, W. L., M. Altwegg, C. Kind, S. Mirjanic, D. Hardegger, and D. Nadal. 2003. Prevalence of enteroaggregative Escherichia coli among children with and without diarrhea in Switzerland. J. Clin. Microbiol. 41:2289-2293.
36. Reisner, A., J. A. J. Haagensen, M. A. Schembri, E. L. Zechner, and S. Molin. 2003. Development and maturation of Escherichia coli K-12 biofilms. Mol. Microbiol. 48:933-946.
37. Rutherford, K., J. Parkhill, J. Crook, T. Horsnell, P. Rice, M.-A. Rajandream, and B. Barrell. 2000. Artemis: sequence visualization and annotation. Bioinformatics 16:944-945.
38. Sambrook, J., and D. W. Russell. 2001. Molecular cloning: a laboratory manual, 3rd ed. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y.
39. Sarantuya, J., J. Nishi, N. Wakimoto, S. Erdene, J. P. Nataro, J. Sheikh, M. Iwashita, K. Manago, K. Tokuda, M. Yoshinaga, K. Miyata, and Y. Kawano. 2004. Typical enteroaggregative Escherichia coli is the most prevalent pathotype among E. coli strains causing diarrhea in Mongolian children. J. Clin. Microbiol. 42:133-139.
40. Savarino, S. J., A. Fasano, J. Watson, B. M. Martin, M. M. Levine, S. Guandalini, and P. Guerry. 1993. Enteroaggregative Escherichia coli heat-stable enterotoxin 1 represents another subfamily of E. coli heat-stable toxin. Proc. Natl. Acad. Sci. USA 90:3093-3097.
41. Sheikh, J., J. R. Czeczulin, S. Harrington, S. Hicks, I. R. Henderson, C. LeBouguenec, P. Gounon, A. Phillips, and J. P. Nataro. 2002. A novel dispersin protein in enteroaggregative Escherichia coli. J. Clin. Investig. 110:1329-1337.
42. Sheikh, J., S. Hicks, M. Dall'Agnol, A. D. Phillips, and J. P. Nataro. 2001. Roles for Fis and YafK in biofilm formation by enteroaggregative Escherichia coli. Mol. Microbiol. 41:983-997.
43. Srimanote, P., A. W. Paton, and J. C. Paton. 2002. Characterization of a novel type IV pilus locus encoded on the large plasmid of locus of enterocyte effacement-negative Shiga-toxigenic Escherichia coli strains that are virulent for humans. Infect. Immun. 70:3094-3100.
44. Steiner, T. S., J. P. Nataro, C. E. Poteet-Smith, J. A. Smith, and R. L. Guerrant. 2000. Enteroaggregative Escherichia coli expresses a novel flagellin that causes IL-8 release from intestinal epithelial cells. J. Clin. Investig. 105:1769-1777.
45. Strom, M. S., and S. Lory. 1993. Structure-function and biogenesis of the type IV pili. Annu. Rev. Microbiol. 47:565-596.
46. Suzart, S., B. E. C. Guth, M. Z. Pedroso, U. M. Okafor, and T. A. T. Gomes. 2001. Diversity of surface structures and virulence genetic markers among enteroaggregative Escherichia coli (EAEC) strains with and without the EAEC DNA probe sequence. FEMS Microbiol. Lett. 201:163-168.
47. Taylor, R. K., V. L. Miller, D. B. Furlong, and J. J. Mekalanos. 1987. Use of phoA gene fusions to identify a pilus colonization factor coordinately regulated with cholera toxin. Proc. Natl. Acad. Sci. USA 84:2833-2837.
48. Watnick, P. I., and R. Kolter. 1999. Steps in the development of a Vibrio cholerae El Tor biofilm. Mol. Microbiol. 34:586-595.
49. Yoshida, T., S.-R. Kim, and T. Komano. 1999. Twelve pil genes are required for biogenesis of the R64 thin pilus. J. Bacteriol. 181:2038-2043.
50. Zhang, X.-L., I. S. M. Tsui, C. M. C. Yip, A. W. Y. Fung, D. K.-H. Wong, X. Dai, Y. Yang, J. Hackett, and C. Morris. 2000. Salmonella enterica serovar Typhi uses type IVB pili to enter human intestinal epithelial cells. Infect. Immun. 68:3067-3073.(Edward G. Dudley, Cecilia)