ABSTRACT
Enteroaggregative Escherichia coli (EAEC) is an important cause of acute and persistent diarrhea. The defining stacked brick adherence pattern of Peruvian EAEC isolate 042 has previously been attributed to aggregative adherence fimbriae II (AAF/II), which confer aggregative adherence on laboratory E. coli strains. EAEC strains also show exceptional autoaggregation and biofilm formation, other phenotypes that have hitherto been ascribed to AAF/II. We report that EAEC 042 carries the heat-resistant agglutinin (hra1) gene, also known as hek, which encodes an outer membrane protein. Like AAF/II, the cloned EAEC 042 hra1 gene product is sufficient to confer autoaggregation, biofilm formation, and aggregative adherence on nonadherent and nonpathogenic laboratory E. coli strains. However, an 042 hra1 deletion mutant is not deficient in these phenotypes compared to the wild type. EAEC strain 042 produces a classic honeycomb or stacked brick pattern of adherence to epithelial cells. Unlike wild-type 042, the hra1 mutant typically does not form a tidy stacked brick pattern on HEp-2 cells in culture, which is definitive for EAEC. Moreover, the hra1 mutant is significantly impaired in the Caenorhabditis elegans slow kill colonization model. Our data suggest that the exceptional colonization of strain 042 is due to multiple factors and that Hra1 is an accessory EAEC colonization factor.
Enteroaggregative Escherichia coli (EAEC) was originally identified as the etiologic agent of persistent diarrhea in developing countries but is gaining increasing prominence for its role in a wider spectrum of diarrheal syndromes. EAEC strains have been implicated in acute as well as persistent diarrhea among adults and children (reviewed in references 25 and 40). A recent meta-analysis found that EAEC is significantly associated with disease in every group at high risk for diarrhea, including young children, human immunodeficiency virus-positive individuals, and visitors to developing countries (24). In addition to its association with disease in epidemiological studies in developing countries, EAEC has also been identified as a principal cause of diarrheal disease in Germany, the United Kingdom, and the United States (11, 26, 51).
Aggregative adherence is the defining characteristic of EAEC (38). EAEC strains adhere to the intestinal epithelium, and to epithelial cells in culture, in a characteristic two-dimensional “stacked brick” fashion. The pattern features bacteria adhering to the eukaryotic surface, other bacteria, and the solid substratum. Four types of fimbriae have so far been documented as conferring aggregative adherence (4, 14, 17, 37). Two noncontiguous plasmid loci containing the complete complement of genes encoding aggregative adherence fimbriae I (AAF/I) or AAF/II are sufficient to confer aggregative adherence on nonadherent E. coli (14, 49). The plasmid bearing type IV pili found in Serbian EAEC outbreak strain C1096 are also sufficient to confer a weak aggregative adherence phenotype on E. coli K-12 (17). AAF additionally play an essential role in production of a superfluous EAEC-associated biofilm, which could account for the association of these strains with persistent diarrhea in epidemiological studies (46).
Some categories of diarrheagenic pathogens have a conserved set of adhesins which allow them to overcome flushing across the intestinal epithelium. Typical enteropathogenic E. coli isolates, for example, all possess bundle-forming pili and the outer membrane adhesin intimin, whereas atypical enteropathogenic E. coli isolates possess intimin but not bundle-forming pili (reviewed in reference 10). EAEC strains, by contrast, are considerably heterogeneous. While many EAEC strains carry genes encoding one of the known aggregative adherence fimbriae, some EAEC do not harbor any known AAF even though they do demonstrate aggregative adherence (4, 7, 13, 14). This, and the presence of multiple adhesins in most mucosal colonizers (53), points to the likelihood of other EAEC adhesins. Imuta et al. recently implicated a TolC secreted factor in adherence (27), and Montiero-Neto et al. (33) described a 58-kDa nonstructural adhesin in O111:H12 EAEC. However, the former factor is only a contributor to aggregative adherence and the latter adhesin is not found in other EAEC. Overall, nonstructural EAEC adhesins have received little attention.
The outer membrane protein Tia was originally characterized as an invasin and later shown to confer adhesive properties on enterotoxigenic E. coli (ETEC) (20, 21). Fleckenstein et al. (21) observed that a tia gene probe hybridized to DNA from non-ETEC strains, one of which was EAEC strain 042. As the Southern blot data published by Fleckenstein et al. showed bands of different intensities, as well as size, between ETEC strain H10407, which carries tia, and EAEC strain 042, we hypothesized that the probe was recognizing a similar, rather than identical, gene (21).
We have determined that EAEC strain 042 harbors a gene encoding the heat-resistant agglutinin 1 (hra1), a hemagglutinin originally reported from an O9:H10:K99 porcine ETEC strain. Hra1 has also been reported from uropathogenic E. coli strains and neonatal meningitis E. coli strain RS218, in which context it is otherwise known as Hek (19, 48). (The hek nomenclature was introduced after hra1, to delineate the form of the gene found in invasive human pathogens from that of a porcine isolate [19].) A role for the outer membrane protein Hra1/Hek in adherence by neonatal meningitis E. coli has recently been defined (19).
Although hra1/hek has been reported from multiple pathogens, its role in colonization and virulence has only been conclusively studied in the neonatal meningitis E. coli strain RS218 (19). In this paper, we demonstrate that the EAEC hra1 gene is sufficient to confer colonization-associated phenotypes, including aggregative adherence and biofilm formation, on laboratory E. coli strains. Intriguingly, we find that although it confers these phenotypes on K-12 and is expressed in 042, hra1 is not required for in vitro colonization-associated phenotypes demonstrated by 042. The hra1 gene is, however, essential for the formation of a true stacked brick pattern in EAEC and for optimal in vivo colonization in a Caenorhabditis elegans model.
MATERIALS AND METHODS
Bacterial strains.The strains used in this study are listed in Table 1. EAEC strain 042 is a Peruvian EAEC isolate that elicited diarrhea in three of five adult volunteers during a human challenge study (36). Most molecular pathogenesis studies have focused on this strain, and its genome has recently been completed (http://www.sanger.ac.uk/Projects/Escherichia_Shigella/ ). E. coli K-12 strain TOP10 (Invitrogen) was used as host for cloned genes. Pseudomonas aeruginosa PA14 and E. coli OP50 and genome-sequenced K-12 MG1655 were employed as controls in some experiments. Plasmids with an oriR6K origin were maintained in EC100D pir-116 (Epicentre), and E. coli SM10λpir was used as a conjugative donor. Bacteria were cultured in Luria-Bertani (LB) broth or LB agar and bacterial strains were maintained in LB-glycerol (1:1) at −80°C. Antibiotics, ampicillin (100 μg/ml), chloramphenicol (30 μg/ml), tetracycline (25 μg/ml), or neomycin (50 μg/ml) were added for selection when required.
Strains and plasmids used in this work
General molecular biology procedures.Standard molecular biology procedures were employed (43). DNA amplification was performed using 1 unit of recombinant Taq polymerase enzyme, 2 mM magnesium chloride, PCR buffer (Invitrogen), and 1 μM oligonucleotide primer in each reaction mixture. All amplifications began with a 2-minute hot start at 94°C followed by 30 cycles of denaturing at 94°C for 30 s, annealing for 30 s at 5°C below primer annealing temperature, and extending at 72°C for 1 min for every kilobase of DNA. PCRs were templated with boiled bacterial colonies, plasmid, or genomic DNA. Oligonucleotide primer sequences are listed in Table 2. Unless otherwise stated, ligations were performed using Quick T4 ligase enzyme (New England Biolabs), and clones and plasmids were transformed into chemically competent E. coli K-12 DH5α or TOP10 cells. Transformation into EAEC strains was accomplished by electroporation using 2-mm cuvettes with a Bio-Rad Micropulser, according to the manufacturer's instructions.
Oligonucleotide primers used in this study
Cloning the hra1 gene from EAEC strain 042.The hra1 gene and about 0.6 kb of flanking sequence on either side were amplified from EAEC strain 042 (using primers hra1up and hra1dn) and cloned into pGEMT (Promega) to produce pTHra1. The clone was sequenced and 160 bp of sequence was deleted upstream of the hra1 promoter with SspI and SphI to excise the fragment and remove an upstream NcoI restriction site. The fragment was subcloned into SspI-SphI sites of pBR322 to produce pBJ1. The expression and outer membrane localization of Hra1 in TOP10 strains carrying pTHra1 and pBJ1 was verified by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) of fractionated bacteria (9).
Purification of Hra1 and Western blotting. E. coli K-12 TOP10 containing the high-copy-number hra1 clone pTHra1 was cultured in LB. The cells were washed in phosphate-buffered saline (PBS) and lysed by sonication. Outer membranes were purified as described by Chart et al. (9). Outer membrane proteins were fractionated on a preparative SDS-PAGE column in a Minicell (Bio-Rad) according to the manufacturer's instructions. Fractions containing a 25-kDa protein that was absent in the outer membrane preparations from TOP10 cells carrying the pGEMT vector were combined and concentrated using Centricon columns. They were then used to raise polyclonal antibodies in rabbits by Covance (Princeton, NJ). For Western blotting, proteins separated on 12 or 14% SDS-PAGE gels were transferred to polyvinylidene difluoride membranes (Immobilon; Millipore) using a Trans-Blot semidry transfer cell (Bio-Rad) according to the manufacturer's instructions. Membranes were blocked overnight in 5% skim milk in PBS and reacted with a 1 in 5,000 dilution of anti-Hra1 antiserum. Following reaction of bound antibody with goat anti-rabbit horseradish peroxidase antibody (Pierce) and SuperSignal West Femto maximum sensitivity chemilumniscence substrate (Pierce), bands were visualized by chemilumnescence.
Construction of a nonpolar isogenic hra1 mutant in strain 042.The hra1 clone, pBJ1, was digested with NcoI, which excised a 395-bp central, predicted surface-exposed encoding region of the gene. The deleted region was replaced with a promoterless aphA-3 cassette amplified from pUC18K (31). In this context, the aphA-3 gene is preceded by translational stop codons in all three reading frames and followed by a consensus ribosomal binding site. The resulting nonpolar deletion construct was recombined into the 042 chromosome using suicide vector pCVD442. Kanamycin- and sucrose-resistant but ampicillin-sensitive mutant candidates were selected, and deletion of the hra1 gene was confirmed by PCR, which generated appropriate-sized fragments with five primer pairs located inside and outside hra1 and the aphA-3 cassette (see Fig. S1 in the supplemental material). The presence of host strain markers tetA, cat, and pAA bearing CVD432 in the mutant was also confirmed (data not shown; see Fig. S1 in the supplemental material). Plasmid pBJ1, which contains the hra1 gene in a pBR322 vector, was used to complement the mutant.
Autoaggregation assay.To measure autoaggregation, we performed an assay that quantifies bacterial settling rates over time (22). Overnight cultures of each strain were adjusted to the same optical density at 600 nm (OD600). Five milliliters of each adjusted culture was placed into two separate tubes. One tube remained static and the other was vortexed before each OD measurement. The tubes were incubated without shaking at 37°C. At designated time points, 0.5 ml was removed from within 2 cm of the surface of the culture and the OD600 was measured. This assay was performed using both LB medium and high-glucose Dulbecco's minimum essential medium (DMEM; Invitrogen).
Biofilm formation.Biofilm formation was tested in LB medium and high-glucose DMEM, the latter of which has been shown to be optimal for biofilm formation by EAEC strain 042 (46). Biofilms were observed and quantified by fixing and staining with crystal violet (42). Briefly, to measure biofilm formation, 10 μl of overnight culture was added to 1 ml of test medium in a 24-well plate. Plates were incubated without shaking, or with rocking, at 37°C. At designated time points, culture medium was aspirated, each well was washed three times with PBS and contents were fixed for 10 min with 75% ethanol. The wells were allowed to dry completely. Fixed biofilms were visualized by staining with 0.5% crystal violet for 10 min, washing with water, and viewing with a stereoscopic zoom microscope. For biofilm quantification, crystal violet was eluted with 1 ml of a 3:1 solution of ethanol-acetone. The OD570 of the eluted crystal violet was measured. Data were analyzed by an unpaired Student's t test.
Air-water interface biofilms.We used the Henk's method (23) to collect and observe bacteria at the air-water interface without disrupting their arrangement. A 0.5-ml aliquot of an overnight culture was added to sterile cylindrical 125-mm by 65-mm Pyrex dishes containing 500 ml of high-glucose DMEM. Cultures were grown undisturbed at room temperature for 3 days. Collection slides were coated with were collodion, a hydrophobic adhesive polymer, as described by Henk (23). The slides were dipped horizontally and smoothly into the culture, immersed briefly in sterile water, and then set to dry overnight. The slides were stained with gentian violet stain for 10 min, rinsed with water, air dried, and observed by light microscopy.
HEp-2 adherence assay.The HEp-2 adherence assay originally described by Cravioto et al. (12) was used with modifications for delineating aggregative adherence (55). HEp-2 cell monolayers were cultured overnight in 8-well chamber slides (for qualitative tests) or 24-well plates (for quantitative assays) to 50% confluence in high-glucose DMEM with fetal bovine serum, streptomycin, and penicillin (Invitrogen). Bacteria were cultured in LB broth without shaking at 37°C overnight. On the day of the adherence assay, the HEp-2 cells were washed three times with PBS. Growth medium was replaced with high-glucose DMEM containing 1% mannose (without fetal bovine serum and antibiotics). At 3 hours, culture medium was aspirated and each well was washed three times with PBS. The cells were fixed for 20 min with 70% methanol and then stained for 20 min with a 1:40 dilution of Giemsa in PBS. Adherence patterns were observed using oil immersion light microscopy at a 1,000× magnification. As adhesins sometimes produce an exaggerated effect with TOP10 and DH5α, which do not adhere at all, E. coli strain MG1655 was also used as host for adherence assays employing plasmid-borne hra1. A quantitative adherence assay protocol was adapted from that described by Torres et al. (52) with some modifications. Wells on the 24-well plate were infected with 20 μl of each strain. At each designated time point, the wells containing infected HEp-2 cells were washed three times with PBS. A 200-μl volume of 0.1% Triton X-100 in PBS was added to each well to break up the adherent monolayer. The 24-well plate was incubated at room temperature for 15 min. Dilutions of the lysate were plated out on LB (with ampicillin for selection when appropriate) and viable counts were incubated after overnight incubation at 37°C. Colonies were counted and the CFU/ml of medium was computed for each well.
C. elegans slow kill assay.The slow kill assay used was adapted from that reported by Aballay et al. (2) and Tan et al. (50) as optimized for EAEC strains (J. Hwang, L. Mattei, L. VanArendonk, P. Meneely, and I. Okeke, submitted for publication). Briefly, 10 hermaphroditic N2 worms at the L4 stage were seeded onto plates of nematode growth agar 24 h after they had been surface inoculated with test or control bacterial cultures. Worms were transferred to fresh bacterial lawns every 48 h to distinguish generations. Survival was documented every 24 h and worms were considered dead when they no longer responded to touch. The strains Pseudomonas aeruginosa PA14 and E. coli OP50 were used as positive and negative controls, respectively, and each independent assay was repeated in quadruple. Data from the assays were analyzed using Kaplan-Meier statistics, and significant differences were inferred from chi-square analyses performed using the PEPI version 4.0 SURVIVAL program.
RESULTS
EAEC 042 carries an hra1 gene.Based on Southern blotting data from Fleckenstein et al. (21), we hypothesized that the 042 genome contained a tia homolog. BLAST revealed that the predicted product of the tia gene is 67% identical to Hra1/Hek from an O9:H10:K99 animal pathogen E. coli, uropathogenic E. coli strain 536, and neonatal meningitis E. coli strain RS218. The genes showed strong similarity at the 5′ and 3′ ends as well as within predicted integral membrane β-sheet regions. They are however variable at predicted surface-exposed regions that could confer function. We initially designed primers agglF and agglR (Table 2) to anneal to conserved regions at the ends of known tia and hra1/hek genes and amplify the intervening region. The 042 amplicon was cloned and sequenced, and the sequence we obtained was identical to that of open reading frame Ec042-3176 of the now-available sequence of the gene from the 042 genome project. Hek/Hra1 alleles from EAEC 042, neonatal meningitis E. coli strain RS218, and uropathogenic E. coli strains 536 and UTI189, as well as an O9:H10:K99 porcine isolate are nearly identical and differ mostly in N-terminal signal sequences that would be cleaved from the functional protein (see Fig. S2 in the supplemental material). The predicted product of the hra1/hek gene from EAEC strain 042 was found to show 98% identity with Hra1 from the O99 animal pathogen and 90% identity to Hra1/Hek alleles from the three invasive human isolates.
The EAEC hra1 gene confers colonization-associated phenotypes on E. coli K-12.Enteroaggregative E. coli is able to autoaggregate, a property that has heretofore been attributed to aggregative adherence fimbriae, which confer this phenotype on laboratory E. coli strains (14, 37). We cloned the EAEC 042 hra1 gene under the control of its own promoter into the SspI and SphI sites, within the tet gene, of pBR322. The resulting plasmid (pBJ1) was able to confer surface expression of the protein and autoaggregation on E. coli K-12 strain TOP10 (Fig. 1). Deletion of a 395-bp predicted surface-exposed region of the molecule abolished Hra1 expression and autoaggregation activity (Fig. 1 and 2). As shown in Fig. 3, the hra1 clone was also able to confer the enhanced biofilm formation on E. coli K-12, and the central predicted surface-exposed region of the gene was required for this property. EAEC strain 042 can form a stacked brick biofilm at the air-water interface when grown at room temperature in a 3-day static high-glucose DMEM culture. The hra1 gene was also able to confer this phenotype on E. coli K-12 (Fig. 3).
(A) Coomasie blue-stained SDS-PAGE gel of outer membrane preparations of E. coli TOP10 transformed with hra1 in pBR322-pBJ1 (lane 2), the medium-copy-number pBR322 vector (lane 3), hra1 in pGEMT-pTHra (lane 4), and the high-copy-number vector pGEMT (lane 5). Lane 1, prestained marker (Bio-Rad). (B) Anti-Hra1 Western blot of electrophoresed cell lysates of E. coli TOP10 carrying pBR322 (lane 1), hra1 clone pBJ1 (lanes 2 and 3), hra1 with the central 395 bp deleted (pBJ2; lane 4), pBJ2 with an inserted aphA-3 casette (pBJ4; lane 5), and pTHra1 (lane 6).
Autoaggregation in LB broth. Overnight cultures were incubated statically for 10 h. The optical density of culture medium withdrawn from 1 to 2 cm from the surface was measured at 600 nm (open squares). For each test strain, a parallel tube (filled triangles) was vortexed before measurements were taken. Autoaggregation ability is proportionate to the distance between the vortexed and static culture plots. (A) EAEC strain 042; (B) E. coli K-12 TOP-10 carrying the pBR322 vector; (C) E. coli K-12 TOP-10 carrying the hra1 gene cloned into pBR322 (pBJ1); (D) E. coli K-12 TOP-10 carrying a pBJ1 derivative with the predicted surface-exposed region of Hra1 deleted (pBJ2).
Biofilm formation. (A) Eight-hour biofilms produced in DMEM on polystyrene surfaces. (B) Biofilm formation quantified at 8, 24, and 48 h by crystal violet staining and elution. Biofilms were quantified as the OD570 of eluted crystal violet, plotted on a logarithmic scale. Strains tested were E. coli K-12 TOP-10 containing pBJ1, the hra1 clone (gray bars), pBJ2, pBJ1 with the surface-exposed region deleted (dotted bars), or pBR322 vector (unshaded bars) as well as wild-type EAEC strain 042 (hatched bars). (C) Surface biofilms formed at the air-water interface in DMEM. Test strains were EAEC strain 042, E. coli K-12 TOP-10 (pBJ1), E. coli K-12 TOP-10 (pBJ1), and negative control E. coli K-12 TOP-10 carrying the pBR322 vector. Values for pBJ2 and pBR322 in TOP-10 were not significantly different from each other (P > 0.05). pBJ1 in TOP-10 produced a significantly greater biofilm than pBJ2 and pBR322 in TOP-10 at 8, 24, and 48 h (P < 0.05). Values obtained for 042 were significantly greater than those for pBJ1 in TOP-10, as well as pBR322 in TOP-10, at all time points (P < 0.05).
E. coli K-12 strains bearing hra1 demonstrate aggregative adherence.Aggregative adherence, the signature phenotype of enteroaggregative E. coli, has been previously attributed to AAF. We were able to demonstrate that hra1 is sufficient to confer a stacked brick pattern of adherence on E. coli K-12. The extent of adherence conferred on E. coli K-12 by pBJ1 was lower than seen in wild-type EAEC strains; however, the adherence pattern was aggregative. Deletion of the 395-bp surface-exposed region abolished the phenotype (Fig. 4).
Adherence of bacteria to HEp-2 cells in culture after a standard 3-h assay. (A) Uninfected HEp-2 cells; (B) EAEC strain 042, demonstrating classic aggregative adherence; (C) pBR322 vector in E. coli K-12 TOP-10; (D) hra1 clone pBJ1 in TOP-10; (E) hra1 deletion construct, pBJ2, in TOP-10; (F) E. coli K-12 strain MG1655; (G) MG1655 carrying pBJ1; (H) MG1655 carrying pBJ2.
An EAEC 042 hra1 deletion mutant does not show quantitative deficiencies in autoaggregation, biofilm formation, and HEp-2 adherence.We determined that the hra1 gene could confer autoaggregation, biofilm formation, and aggregative adherence on nonpathogenic bacteria; however, these phenotypes are also conferred by AAF/II (14). We therefore constructed a mutation in the hra1 gene in an EAEC 042 background to evaluate the contribution of the outer membrane protein to these colonization-associated phenotypes. We hypothesized that the hra1 mutant would show a quantitative reduction in its ability to demonstrate these phenotypes. Although Hra1 can be detected by Western blotting in the wild-type strain 042 and not the mutant, we saw no quantitative difference in autoaggregation when the wild-type 042 strain was compared to its hra1 isogenic mutant (Fig. 5). We did see enhanced autoagglutination in the complemented strain, but this is likely due to expression of hra1 from a multicopy vector. Similarly, biofilm formation between the wild-type and mutant strains occurred to a similar degree although the mutant formed a slightly smaller biofilm at early time points and this observation could be complemented in trans (Fig. 5; see also Fig. S3 in the supplemental material). Visualization of the polystyrene surface revealed that the wild type and complement formed surface-adherent clumps on the polystyrene surface as early as 3 h but these appeared at later time points in the mutant. We also performed biofilm experiments in LB and minimal media, at low (5.5) and neutral (7.0) pH, and with shear. The hra1 mutant was not significantly defective in biofilm formation under any of these conditions (data not shown). There was also no quantitative difference in HEp-2 adherence at 1 or 3 hours between the wild-type 042 strain (6.2 × 108 ± 1.5 × 108 and 7.0 × 109 ± 5.7 × 108, respectively) and its hra1 mutant (4.8 × 108 ± 1.5 × 108 and 1.6 × 109 ± 6.9 × 108, respectively) (P > 0.05).
(A) Western blot of whole-cell lysates of 042 (lane 1), hra1 mutant SB1 (lane 2), SB1 complemented with the hra1 clone (pBJ1), pBJ1 in TOP10 (lane 3), and pBJ2, the hra1 deletion construct, in TOP10 (lane 5). (B to D) Autoaggregation in DMEM of 042 (B), its isogenic hra1 mutant (C), and SB1 and the transcomplement SB1(pBJ1) (D) in DMEM. For panel E, biofilm formation by 042, SB1, and SB1(pBJ1) in DMEM was quantified by crystal violet staining and elution. Differences between the wild type and mutant were not significant at any time point (P > 0.05).
The EAEC 042 hra1 mutant departs from the strict stacked brick pattern of adherence.Although the number of bacterial CFU adhering to HEp-2 cells in culture was not significantly different in the hra1 mutant compared to the wild-type EAEC 042 strain, we observed important differences in the pattern of adherent cells. Significantly, the hra1 mutant bacteria were often disarrayed around HEp-2 cells, while wild-type 042 bacteria arrayed in the stacked brick formation that is characteristic for EAEC (Fig. 6). The mutant phenotype could be complemented by supplying the hra1 gene in trans.
(A) The most commonly exhibited HEp-2 adherence patterns of 042, its isogenic hra1 mutant, SB1, and the transcomplement SB1(pBJ1) at 3 hours. (B) The number of HEp-2 cells around which bacteria demonstrated the strict stacked brick pattern, a disheveled arrangement, or a mixed pattern that could not be objectively categorized as one or the other. For each strain, adherence patterns around 25 HEp-2 cells were counted.
The EAEC 042 hra1 mutant is deficient in C. elegans slow killing.We observed no quantitative difference in adherence but a striking difference in adherence pattern in the hra1 mutant. There were no differences in autoaggregation and only subtle variations in biofilm formation at early time points. We therefore wished to determine whether these apparently small but marked differences had in vivo consequences. Although there is no nonhuman vertebrate model of EAEC infection, we have developed a C. elegans model in which EAEC strains colonize the nematode gut and demonstrate enhanced slow killing. In this model system, strain 042 colonizes and kills worms less efficaciously than some other EAEC strains but shows a detectable difference in this property compared to nonpathogenic E. coli (Hwang et al., submitted). In the C. elegans slow kill assay, worms fed on OP50, the negative control strain, have a median survival time of at least 8.3 days, while worms fed on EAEC strain 042 have a median survival time of only 5 days. The hra1 mutant (SB1) showed an intermediate level of colonization with a significantly increased median survival time of 7 days, compared to 5 days for the wild-type strain (P = 0.013) (Fig. 7).
C. elegans survival curves for worms fed OP50 (negative control), EAEC 042, and the hra1 mutant SB1. Slow kill rates of SB1 were significantly lower than with 042 and greater than with OP50 (P < 0.05).
DISCUSSION
Deletion of aafA, which encodes the structural subunit of AAF/II, was earlier reported to abolish adhesive phenotypes (14, 46); however, EAEC strain 042 possesses and expresses an accessory outer membrane adhesin. The adhesin is encoded by a chromosomally located hra1 gene, which can confer autoagglutination and superior biofilm formation properties on laboratory strains, similar to hra1 from the O9:H10:H99 animal pathogen and neonatal meningitis strain RS218 (18, 19, 29). Fleckenstein et al. (21) originally reported that the tia probe hybridized to 042 genomic DNA, and some have proposed that this strain harbors the enterotoxigenic E. coli tia invasin gene (49). We report that the tia probe hybridized to a tia homolog, hra1 (otherwise known as hek). The EAEC 042 hra1 gene confers the ability to adhere in the characteristic aggregative pattern that defines the EAEC category. As this phenotype can also be conferred by AAF/II (14), we sought to assess the specific contribution of hra1 to the exceptional colonization-associated phenotypes of 042. Consistent with the dominant role played by AAF/II in EAEC adherence (14, 46), deletion of the hra1 gene from strain 042 does not abrogate or appreciably reduce autoagglutination, biofilm formation, or HEp-2 adherence ability. The hra1 gene appears to be dispensable overall in these in vitro phenotypes, suggesting some redundancy in 042 adhesins.
Differences in hra1-mediated activity in K-12 and EAEC 042 strains are in part due to copy number of the plasmid-borne gene in the former, as supported by exaggerated effects in the 042 trans-complemented mutant. However, since Western blotting of outer membrane preparations shows that the wild-type strain does express the gene, and this gene makes a significant contribution to adherence patterns and slow kill in C. elegans, we speculate that posttranslational interactions may be more important in determining the contributions from hra1. For example, fimbriae have been reported to sterically inhibit adhesive properties of nonstructural surface protein antigen 43 (22). Similarly, capsules can block autoaggregation mediated by antigen 43 and AIDA-1 (44). It is possible that hra1 is occluded when the 042 strain is piliated and heavily capsulated. The paradigm is further complicated by type I fimbriae, which Moreira et al. (34) also showed contribute to the robustness of the 042 stacked brick pattern. Our study further demonstrates the importance of evaluating putative adhesins in their natural host background as well as in surrogate strains in order to gauge potential contributions to in vivo colonization.
Careful observation of adherence patterns demonstrates that although hra1 is not required for HEp-2 adherence in 042, it is nonetheless required to produce a strict stacked brick formation. The significance of patterning is unclear but the finding has implications for definition of the category. For example, it is possible that some strains that possess aggregative adherence fimbriae may in fact not be detected as EAEC in studies that use the HEp-2 adherence test alone. Similarly, strains that possess hra1 but not other EAEC genes might also be classed as EAEC. In essence, the definition of enteroaggregative E. coli may require refining.
The C. elegans slow kill assay has been employed to study pathogenicity in enteric pathogens, including pathogenic E. coli (2, 15, 30). Virulence in this model has been shown by other investigators to correlate with the presence or absence of key virulence and colonization factors as well as the genes that regulate them (1, 2, 15, 30). Although the in vitro effects of deleting hra1 appear to be at best minimal, we observed a statistical significant reduction of the slow kill rate of the hra1 mutant compared to wild-type 042. In spite of the fact that, compared to other EAEC, strain 042 is a relatively poor colonizer of C. elegans (Hwang et al., submitted), we were able to detect a significant colonization defect in the hra1 mutant. It is possible that rapid colonization is a prerequisite for the full slow kill effect in C. elegans, that the true importance of hra1 (relative to AAF/II and other adhesins) is not possible to detect in vitro, that an ordered pattern of adherence promotes colonization, or that some combination of these factors plays out in vivo. It is also possible that an as-yet-undetermined mechanism, such as the availability of a specific receptor in the worm, could account for the failure to detect a role for Hra1 in vitro.
Interestingly, although the central 150 residues of Hra1 from EAEC strain 042, which include predicted surface-exposed loops designated by Fagan et al. (18), are most similar to the corresponding region of Hra1/Hek variants from invasive human isolates, its N-terminal signal sequence is identical to that of Hra1 from O9:H10:K99 E. coli strain (see Fig. S2 in the supplemental material). These findings suggest that functional regions may have evolved convergently or that recombination among the agglutinins may have occurred in the recent past. The similarities and differences among the agglutinins are reminiscent of the mix-and-match nature of the Wolbachia surface proteins from insect endosymbionts (3), which are homologous to Hra1. The modular nature of Hra1/Hek homologs suggests that an agreed-upon nomenclature is needed. As we, for the first time, have defined a role for hra1 in a noninvasive human enteric pathogen and because the gene we studied is more closely related to hra1 than hek overall, we have chosen to use the older nomenclature in recognition of the earliest work done on this family of proteins. Our data and those of other investigators point to related but distinct roles for the agglutinin genes in colonization by their cognate E. coli strains. However, the data suggest a more important role for factors other than hra1 polymorphisms in producing a less marked role for hra1 in EAEC infection than that of hra1/hek showed in neonatal meningitis E. coli (19). RS218 is an invasive strain and in this background, hra1/hek appears to be a principal adhesin and invasin (19). EAEC colonizes the intestinal mucosa, and we have demonstrated that in spite of exaggerated phenotypes that are conferred on nonpathogenic E. coli in trans, in EAEC strain 042, the hra1 gene plays an accessory role in this process. In addition to host- and niche-specific factors, context and therefore the nature of pathogenicity would also be influenced by other adhesins and surface factors expressed by the relevant E. coli strain. This in turn has important repercussions for EAEC, a heterogeneous category of diarrheagenic E. coli, because different strains carry different combinations of adhesins.
Although an AAF/II mutant is severely deficient in adherence (14), a recent study reported that deletion of the gene encoding outer membrane transporter TolC resulted in a marked but incomplete defect in autoaggregation, biofilm formation, and aggregative adherence (27). In this study, we have identified yet another factor in aggregative adherence. The quantitative contribution from hra1 appears to be even less marked than that of tolC, but deletion of the gene is disruptive to the aggregative pattern and the gene itself is sufficient to confer the aggregative adherence phenotype. As AAF/II interacts with the antiaggregative protein dispersin (45, 54) and TolC with an as-yet-unidentified secreted factor (27), aggregative adherence, which was once looked at as a simple, specific phenotype conferred by fimbriae alone, now appears to be a complex, mutifactorial phenotype with subtle redundancies. Accessory colonization factors, such as hra1, provide an explanation the long-acknowledged observation that aggregative adherence fimbrial content and type do not correlate with absolute adherence pattern (5, 28, 41). EAEC strains carry different adherence genes and the in vivo contributions of separate factors and their interaction with one another may be an as-yet-unstudied contributor to the in vivo heterogeneous virulence seen with this pathotype.
ACKNOWLEDGMENTS
This study was supported by National Science Foundation RUI grant MCB no. 0516591. I.N.O. is a Branco Weiss Fellow of the Society-in-Science, ETHZ, Zürich, Switzerland. We thank the Howard Hughes Medical Institute for support of undergraduate multicultural science program researcher J.H. and interdisciplinary scholars S.B., B.B.J., and A.S.G. This study was assisted by access to in-process sequence data produced by the Escherichia coli and Shigella spp. Comparative Sequencing Group at the Sanger Institute, which can be accessed at http://www.sanger.ac.uk/Projects/Escherichia_Shigella/ .
We are grateful to F. Ausubel for Pseudomonas aeruginosa PA14. We thank Karl Johnson for helpful discussions and Philip Meneely for C. elegans stocks, as well as Katherine Unger, Lucy M. Joo, and Andrew Kim for technical assistance.
FOOTNOTES
- Received 29 December 2008.
- Accepted 20 May 2009.
- Copyright © 2009 American Society for Microbiology
