|Home | About | Journals | Submit | Contact Us | Français|
Autotransporters (AT) are widespread in Gram-negative bacteria, and many of them are involved in virulence. An open reading frame (APECO1_O1CoBM96) encoding a novel AT was located in the pathogenicity island of avian pathogenic Escherichia coli (APEC) O1's virulence plasmid, pAPEC-O1-ColBM. This 3.5-kb APEC autotransporter gene (aatA) is predicted to encode a 123.7-kDa protein with a 25-amino-acid signal peptide, an 857-amino-acid passenger domain, and a 284-amino-acid β domain. The three-dimensional structure of AatA was also predicted by the threading method using the I-TASSER online server and then was refined using four-body contact potentials. Molecular analysis of AatA revealed that it is translocated to the cell surface, where it elicits antibody production in infected chickens. Gene prevalence analysis indicated that aatA is strongly associated with E. coli from avian sources but not with E. coli isolated from human hosts. Also, AatA was shown to enhance adhesion of APEC to chicken embryo fibroblast cells and to contribute to APEC virulence.
The autotransporter (AT) proteins are a large and diverse family of extracellular virulence proteins of Gram-negative bacteria. All ATs share the same general structure and are comprised of three domains: an amino-terminal signal peptide; an α or passenger domain, which confers the function of the secreted protein; and a C-terminal β domain that mediates secretion through the outer membrane. The cardinal feature of conventional ATs is a long C-terminal translocator domain consisting of about 300 amino acids, in contrast to the very short C-terminal translocator domain (about 70 amino acids) of trimeric ATs that form highly stable trimers in the outer membrane (8). While all trimeric AT proteins identified so far display adhesive activity mediating bacterial interactions with either host cells or extracellular matrix (ECM) proteins, the conventional ATs that have been characterized to date have diverse functions, including adhesion, cytotoxicity, and lipase or protease activity (3, 6, 7, 46, 49, 54).
Temperature-sensitive hemagglutinin (Tsh) was the first AT described in avian pathogenic Escherichia coli (APEC), a pathogen which causes extraintestinal infections in turkeys, layers, and broilers (44). This conventional AT, which is encoded by a virulence plasmid, occurs as a 106-kDa extracellular protein and a 33-kDa outer membrane protein. Its passenger domain contains a 7-amino-acid serine protease motif that includes the active-site serine (S259), which has also been found in the secreted domain of IgA1 protease. Although Tsh did not show any IgA protease activity in vitro (51), it was involved in virulence through mediation of APEC's adherence to the air sacs of chickens (11). The gene encoding a second serine protease AT, termed the vacuolating autotransporter or Vat, was identified in a pathogenicity island (PAI) adjacent to the thrW tRNA gene in APEC (42). Vat has vacuolating cytotoxic activity similar to that of VacA of Helicobacter pylori and contributes to APEC virulence (48). Both tsh and vat are present in E. coli from avian sources and are also found in E. coli isolated from human hosts. In the present study, we identified and characterized a novel AT that is strongly associated with avian E. coli. This AT is encoded by the APEC autotransporter gene (aatA), which has been localized to the PAI found in the virulence plasmid (pAPEC-O1-ColBM; accession number NC_009837) of APEC O1, the first APEC strain to be completely sequenced (25, 26).
The strains and plasmids used in this study are listed in Table Table1.1. Well-characterized collections of strains of APEC, E. coli from the feces of apparently healthy chickens (27), human uropathogenic E. coli (UPEC) (47), and human neonatal meningitis-associated E. coli (NMEC) were used for gene prevalence studies (23). Strains were grouped phylogenetically using multiplex PCR. APEC O1, an O1:K1:H7 strain whose genome shares strong similarities with human extraintestinal pathogenic E. coli (ExPEC) genomes, was used to construct mutants and as a positive control in virulence and other functional assays. E. coli DH5α was employed as a negative control. Cells were routinely grown at 37°C in Luria-Bertani broth (LB) supplemented with kanamycin (Km) (50 mg ml−1), chloramphenicol (Cm) (25 mg ml−1), or ampicillin (Amp) (100 mg ml−1), unless otherwise specified. Chicken embryo fibroblast (CEF) cells (ATCC CRL-12203) were maintained in ATCC-formulated Dulbecco's modified Eagle's medium (DMEM) with 10% fetal bovine serum (FBS).
DNA manipulations and transformations were performed using standard methods (2). All restriction and DNA-modifying enzymes were purchased from New England Biolabs, Invitrogen, or Amersham Pharmacia and were used according to the suppliers' recommendations. Recombinant plasmids, PCR products, and restriction fragments were purified using plasmid miniprep, PCR cleanup, and gel extraction kits (Qiagen, Valencia, CA) as recommended by the supplier. Transformation of E. coli strains was routinely done using electroporation. DNA and amino acid sequence analyses were performed using DNASTAR Lasergene 8 software to predict conserved domains and using the search engine at http://blast.ncbi.nlm.nih.gov/Blast.cgi. The I-TASSER online server was used to predict the three-dimensional (3D) structure of the amino acid (55-57). The sequence from position 825 to position 1265 was predicted to be the translocator domain, and the sequence from position 26 to position 824 was predicted to be the passenger domain. Five final models of the passenger domain were generated using I-TASSER and evaluated further to select the most appropriate structure based on the energies calculated by four-body contact potentials (17) and the gapless threading method.
PCR was performed to detect the presence of aatA sequences in different E. coli strains. Primers 5′-TGGTAGTGTTTGGGGAGGAG-3′ and 5′-GCATTTCCTGCAGACAGGTT-3′ were used to amplify the AatA passenger domain. Reactions were carried out using Taq DNA polymerase (New England Biolabs) under the following conditions: 95°C for 1 min, followed by 30 cycles of 94°C for 30 s, 54°C for 30 s, and 72°C for 1 min and then extension at 72°C for 1 min. Specific amplification was confirmed using APEC O1 as a positive control and E. coli DH5α as a negative control.
The PBAD expression system was used for cloning and arabinose-inducible expression of aatA. The coding sequence of aatA (the translation start site in the forward primer is indicated below by italics) was amplified by PCR using genomic DNA of APEC O1 as the template. An Advantage 2 PCR kit was used in these experiments according to the manufacturer's directions (Clontech, Mountain View, CA). The primers used were forward primer aatAE-F (5′-GCCAGAGCTCAGGAGGAATTCATGAATAAGAATATACGAATTT-3′), which introduced a SacI site (underlined) and a ribosome binding site (bold type), and reverse primer aatAE-R (5′-ATCGTCTAGACCCAGCTAACCATGCCTTAT-3′), which introduced an XbaI site (underlined). The complete aatA gene was cloned into the expression vector pBAD18-cm using the SacI and XbaI sites created (19) to obtain pBAD aatA (Table (Table1).1). APEC O1 MaatA mutant strains harboring the empty plasmid pBAD18-cm and pBAD aatA were designated APEC O1 p1 and APEC O1 p2, respectively.
aatA was deleted using the method of Datsenko and Wanner (9). The chloramphenicol (Cm) resistance cassette in pKD3, flanked by 5′ and 3′ sequences of aatA, was amplified from genomic DNA of strain APEC O1 using primers aatAM-F (5′-GTTGATAAAAATGCATCACTAAAGAAAAAACAGTATGAATGTGTAGGCTGGAGCTGCTTCGA-3′) and aatAM-R (5′-TAAACAATATATTGCGAAGAATGTTCATAATGTAAAGAGTCATATGAATATCCTCCTTAG-3′) and was introduced into APEC O1 by homologous recombination using λ Red recombinase (the underlined portions of the primer sequences are identical to the flanking regions of the aatA gene). Successful ΔaatA::Cm mutation was confirmed by PCR, using primers flanking the aatA region. The chloramphenicol resistance cassette was cured by transforming plasmid pCP20 and selecting for a chloramphenicol-sensitive mutant strain. The ΔaatA derivative of APEC O1 was designated APEC O1 MaatA. The ΔaatA mutant strain APEC O1 MaatA was complemented by single-copy integration of plasmid pGP aatA. The aatA operon, including its putative promoter, was amplified by PCR using primers 5′-ATCGTCTAGACTCGCCACGGGAATATCTAC-3′ and 5′-CTAGGTCGACCCCAGCTAACCATGCCTTAT-3′, and the sequence was confirmed. pGP aatA was constructed by cloning the XbaI-SalI fragment (the underlined portions of the primer sequences are cut sites) containing the aatA operon into the same sites of suicide vector pGP704 (37). A strain that was resistant to ampicillin and was found to contain a full-length copy of the aatA gene, as confirmed by PCR, was designated APEC O1 CaatA. APEC O1 CaatA was conjugated from strain S17/pGP aatA to strain APEC O1 MaatA.
APEC O1, APEC O1 MaatA, APEC O1 p1 (with arabinose), and APEC O1 p2 (with arabinose) cells were cultured until the optical density at 600 nm (OD600) was 1 and harvested. Bacterial fractionation was performed as previously described (16). Ten micrograms of protein was examined by electrophoresis on a 10% sodium dodecyl sulfate-polyacrylamide electrophoresis (SDS-PAGE) gel using standard methods (2). Immunodetection was performed following transfer to nitrocellulose membranes (Protran; Schleicher & Schuell) using a 1:1,000 dilution of polyclonal rabbit antiserum raised against the AatA protein. This antiserum was generated by Open Biosytems (Huntsville, AL) using 19 amino acids (DNMISGGYGIKQGGDAISG) of the AatA passenger domain conjugated with keyhole limpet hemocyanin (KLH).
Immunofluorescence microscopy analysis was performed as follows. Bacteria were grown at 37°C in LB in the presence of 0.2% arabinose until the OD600 was 1. Cells were pelleted by centrifugation and fixed with 3% paraformaldehyde for 10 min. Cells were saturated for 15 min with 0.5% bovine serum albumin (BSA) before incubation with a 1:1,000 dilution of the primary polyclonal rabbit antiserum raised against AatA. Cells were next incubated with a 1:500 dilution of the secondary polyclonal goat anti-rabbit serum coupled to Alexa 488. Cells were loaded onto 0.1% poly-l-lysine-treated immunofluorescence microscope slides. The slides were fixed with 3% paraformaldehyde, and then Prolong reagent containing 10 mg ml−1 of 4′,6-diamidino-2-phenylindole (DAPI) (Invitrogen) was added to fix coverslips to the slides. Finally, the slides were observed by epifluorescence microscopy using an Axiovert 200 inverted fluorescence microscope (Zeiss). Images were collected digitally using an AxioCam MR color camera and Axiovision AC imaging software (Zeiss). Images were prepared for presentation using Photoshop and Illustrator software (Adobe Systems).
The positive sera used in this study were obtained from 3-week-old chickens infected with live APEC O1 by intratracheal inoculation (1 × 107 CFU). These birds received a booster inoculation (1 × 107 CFU) of APEC O1 1 month later and were euthanized 2 weeks after the booster was administered. Negative-control sera were obtained from chickens sham inoculated with phosphate-buffered saline (PBS) (pH 7.2). For each group 10 samples were used to detect AatA antibody. Microtiter plates (Maxisorb; Nunc) were coated overnight at 4°C with 1 μg of the synthesized polypeptide DNMISGGYGIKQGGDAISG of AatA. Bovine serum albumin (BSA) (Sigma) was used as a negative control. Wells were washed twice with TBS (150 mM NaCl, 20 mM Tris; pH 7.5) and then blocked with TBS-2% skim milk for 1 h. After washing with TBS, chicken antisera were titrated horizontally across the plate starting with a 1:4 dilution, and the mixtures were incubated for 2 h. After three washes, 100 μl of secondary anti-rabbit horseradish peroxidase antibody in blocking buffer (diluted 1:1,000) was added and incubated for 3 h at room temperature. Finally, the chicken antibody against AatA was detected by adding 150 μl of 1-Strep ABTS [2,2′-azinobis(3-ethylbenzthiazoline-6-sulfonic acid), diammonium salt; Pierce], and the absorbance at 405 nm was determined with a plate reader. A sample was considered positive if the A405 was twice that of the negative control. The ELISA was repeated to determine the anti-AatA antibody titers of positive sera absorbed by whole-cell E. coli antigens made from APEC O1 MaatA.
The interaction of APEC O1, APEC O1 MaatA, APEC O1 p1, and APEC O1 p2 with cultured chicken embryo fibroblasts (CEF) was studied essentially as previously described (36). Briefly, CEF were cultured until they were confluent, and then the culture medium was removed and the cells were washed once with Eagle minimum essential medium (MEM) without fetal bovine serum (FBS). Bacteria were cultured until the OD600 was 0.3, and arabinose was added to a final concentration of 0.2%. Bacterial cells were incubated for 1 h to induce AatA expression. Then 106 CFU of bacteria, as confirmed by counting viable cells, was resuspended in MEM with 0.2% arabinose (to induce AatA expression) but without FBS, added to the monolayers of chicken fibroblasts, and incubated for 1 h at 39°C in the presence of 5% CO2 (a longer incubation time resulted in lysis of APEC O1 p2 because of AatA overexpression). Monolayers were washed three times with PBS to remove nonadherent bacteria, and the eukaryotic cells were lysed with 0.1% Triton X-100. The numbers of adherent and internalized bacteria were determined by plate counting of the lysed cell suspensions. The percentage of adherent bacteria was determined by dividing the number of adherent bacteria by the number of bacteria inoculated. All experiments were performed in triplicate.
To determine the virulence of the bacteria of interest, 1-day-old chicks were inoculated intratracheally with 0.1 ml of a bacterial suspension containing ~5 × 108 CFU ml−1 of APEC O1, APEC O1 MaatA, or APEC O1 CaatA in accordance with an Institutional Animal Care and Use Committee-approved protocol. Birds used as negative controls received 0.1 ml of PBS by the same route. The chicks were monitored for 7 days. Deaths were recorded, and the survivors were euthanized and examined for macroscopic lesions. Lesion scores for the air sacs and combined lesion scores for the pericardium and liver were determined as described by Lamarche et al. (29), except that lesions in the caudal thoracic air sacs were scored from 0 to 3 using the following criteria: 0, normal and clear; 1, mild cloudiness and thickness; 2, moderate cloudiness and thickness accompanied by serous exudates or fibrin spots; and 3, extensive cloudiness and thickness accompanied by muco-or fibrinopurulent exudates. Bacteria were reisolated from livers, hearts, air sacs, and brains of the dead birds, and multiplex PCR (27) was used to confirm that they were the inoculated strain.
The lethality of APEC O1, APEC O1 MaatA, and APEC O1 CaatA for chicken embryos was assessed by inoculating overnight washed bacterial cultures (~500 CFU) into the allantoic cavities of 15-day-old embryonated specific-pathogen-free eggs. Twenty embryos were used for each organism tested. E. coli DH5α was used as the negative-control strain. PBS-inoculated and uninoculated embryos were also used as controls. Embryo deaths were recorded after 24 h, and the assay was performed twice.
Statistical analyses were performed with the Graphpad Software package (GraphPad Software, La Jolla, CA). A one-way analysis of variance (ANOVA) was used in the analysis of adherence data, an unpaired t test was performed for the ELISA results, a Student t test was used to analyze the lesion scores, and a Fisher exact test was used to analyze the results of the embryo lethality assay. For the in vitro assays, mean values were obtained using a minimum of three independent values. Statistical significance was established using P values of <0.05.
Analysis of the sequence of the APEC virulence plasmid pAPEC-O1-ColBM revealed a 116-kb PAI (25). The conserved region of this PAI harbors all known important virulence genes of APEC plasmids, including etsABC, which are genes encoding a putative ABC transporter system; sitABCD, which are genes encoding another ABC transporter system involved in iron and manganese transport; iucABCD and iutA, which are genes encoding the aerobactin siderophore system; iss, the increased-serum-survival gene; iroBCDEN, which are genes of the salmochelin siderophore system; eitABCD, which are genes encoding a putative iron transporter system; and the temperature-sensitive hemagglutinin gene tsh. The APEC autotransporter (aatA) gene (APECO1_O1CoBM96) is between tsh and eitABCD and is annotated as a gene encoding a putative adhesin with similarity to the HMM PF03212 protein family. Both upstream and downstream sequences flanking aatA include insertion sequences (ISs) and transposases. These flanking mobile elements and the lower G+C content of aatA (42%) than of the APEC O1 genome (50.5%) and the ColBM plasmid (49.6%) suggest that aatA was acquired by horizontal gene transfer. This 3,498-bp gene encodes a hypothetical 1,166-amino-acid protein with a theoretical molecular mass of 123.7 kDa. Using the SignalP 3.0 signal peptide prediction software (http://www.cbs.dtu.dk/services/SignalP/), the signal sequence was identified as a sequence that is 25 amino acids long and has a potential cleavage site between residues A25 and Q26 (Fig. (Fig.1A).1A). Amino acid sequence analysis of APECO1_O1CoBM96 using the protein motif search function of DNASTAR Lasegene 8 showed that the C-terminal amino acids 883 to 1166 are predicted to form the β domain of an AT. The C terminus shows similarity to various outer membrane proteins and β domains of ATs (Fig. (Fig.1A).1A). Analysis of the probable membrane topology of the protein revealed that hydrophobic amino acids account for more than 50% of this domain and that there are several different stretches of hydrophobic residues with intermittent hydrophilic residues, which together form the antiparallel β sheets and the connecting loops of the outer membrane β barrel of the translocator. The putative passenger (α) domain was predicted to be from amino acid 26 to amino acid 882, and it shared no significant amino acid sequence identity with the α domain of any other characterized AT. A low level of amino acid sequence similarity in the α domain was restricted to two regions in an AidA-I adhesin-like protein from E. coli O157:H7 strain Sakai (20) (Fig. (Fig.1A).1A). The 3D structure of AatA was predicted using the I-TASSER online server (55-57). The translocator domain has 12 antiparallel strands that form a β barrel with a hydrophilic core inside. As shown in Fig. Fig.1B,1B, the length and diameter of the hydrophilic core are around 30 Å and 25 Å, respectively, as determined by measuring the distances of the C-alpha atoms between Glu1095 and Met933 and between Ser1016 and Ala1119, respectively. If side chain atoms were considered, the diameters should be smaller. For the passenger domain, the I-TASSER online server generated five potential models. We used the four-body contact potential approach, which was developed in 2007 to identify the likely native structure of AatA from thousands of computer-generated models (17). The energy scores for five models were determined to be 196.6, 197.9, 185.4, 103.3, and 253.4. Model 4 was chosen as the best representation since it had the lowest energy score, which made it the most stable conformation. The 3D structure of the passenger domain is shown above the translocator domain in Fig. Fig.1B.1B. The parallel strands form an 80-Å rod-like shape in the middle of the passenger domain, and the loops form the C-terminal head. The predicted structure of the passenger domain has some similarity to the structure of the heme binding protein (Hbp), which is the passenger domain of an AT hemoglobin protease from pathogenic E. coli (40).
Our structural predictions for AatA suggested that AatA might be a novel conventional AT. It is known that the passenger domain of ATs is either secreted into the bacterial medium or displayed at the bacterial cell surface. To localize the passenger domain, an aatA isogenic mutant (APEC O1 MaatA) of APEC O1 was generated using the method of Datsenko and Wanner (9). The secreted proteins of mutant strain APEC O1 MaatA and the wild type were compared by performing SDS-PAGE. No differences between the protein patterns of the two strains were detected, nor were any differences found in the membrane proteins (data not shown). We reasoned that this failure to detect AatA might be due to a low level of expression of AatA. To circumvent this possible problem, we constructed an inducible expression plasmid containing aatA in which aatA is under the control of an arabinose-inducible (pBAD) promoter. In this pBAD aatA plasmid, one ribosome binding site was added upstream of the translation start site, while the structural gene (from the translation start ATG) of the aatA mutant was exactly the same as the wild-type structural gene (see Materials and Methods). Thus, the AatA protein expressed from this plasmid would be the same in terms of localization and function. Protein analysis of the inducible expression mutant revealed a band at the expected molecular mass (120 kDa) in the outer membrane preparations of the APEC O1 p2 strain which was absent from the preparations of the negative-control strain APEC O1 p1 when it was induced by arabinose (Fig. (Fig.2A).2A). To verify the induced protein band of APEC O1 p2, we raised polyclonal rabbit antiserum against a polypeptide fragment synthesized using the theoretical amino acid sequence of AatA. This band reacted specifically with antibodies directed against the polypeptide (DNMISGGYGIKQGGDAISG) in the passenger domain of AatA (Fig. (Fig.2B).2B). To demonstrate the surface localization of AatA, we performed immunofluorescence microscopy. AatA antiserum reacted with intact APEC O1 p2 cells expressing AatA. The surface of individual bacteria was labeled in a mottled ring-like staining pattern, confirming that the N-terminal region of AatA was effectively translocated to the cell surface (Fig. (Fig.3).3). In contrast, no evidence of AatA production was observed on the surfaces of APEC O1 and APEC O1 MaatA cells (Fig. (Fig.33).
The widespread occurrence of vat in ExPEC strains prompted us to investigate the prevalence of aatA in a well-characterized collection of E. coli isolates of avian and human origin. For this purpose, a pair of primers was designed to amplify the α-domain region of aatA. An aatA fragment of the correct size was amplified from 182 of 452 APEC strains (40.3%) and 52 of 106 avian fecal commensal E. coli (AFEC) strains (49.1%), but only 4 of 200 UPEC strains (2.0%) and 8 of 91 NMEC strains (8.8%) were positive. Thus, aatA is significantly more likely to be present in E. coli strains from avian sources (P < 0.001). Further analysis revealed that 69.6% (93/135) of phylogenetic group D E. coli strains in our APEC strain collection were aatA+, while only 24.1% of APEC strains belonging to phylogenetic group A, 31.5% of APEC strains belonging to phylogenetic group B1, and 32.5% of APEC strains belonging to phylogenetic group B2 were positive for aatA. Furthermore, 51.6% of the PCR-positive APEC strains belonged to phylogenetic group D, while 22.0%, 12.6%, and 13.8% of the PCR-positive APEC strains belonged to groups A, B1, and B2, respectively.
For AatA to contribute to APEC virulence, it must be expressed in the host. We could not detect AatA in APEC O1 when it was cultured in vitro. In order to determine if it was expressed in vivo, we developed an ELISA to measure anti-AatA generated in APEC-infected chickens. The ELISA results showed that AatA did elicit generation of an antibody in vivo with a titer as high as 27.4 on average, while the titers of the sera from control chickens were significantly lower (22 on average) (Fig. (Fig.4).4). To rule out the possibility that the increase in the ELISA titer was due to greater production of an overall IgG response following infection with APEC O1 compared to the group that received only PBS, we preabsorbed anti-AatA antiserum with whole-cell E. coli antigens made from the mutant strain (APEC O1 MaatA) and performed the ELISA again. No significant difference between preabsorbed and nonabsorbed sera was detected (Fig. (Fig.4).4). Thus, the increase in the ELISA titer was not caused by the background effect. These results provide indirect evidence that AatA is expressed in vivo, where it is available to contribute to the pathogenesis of colibacillosis.
AatA showed some amino acid sequence similarity to certain regions of AidA-I, an adhesin that contributes to autoaggregation, biofilm formation, and adherence and that is associated with virulence in many Gram-negative pathogens. Therefore, we investigated whether AatA could mediate autoaggregation, biofilm formation, and cell adherence. No differences in autoaggregation and biofilm formation between the APEC O1 ΔaatA mutant and the wild type were detected, and induction of aatA expression in APEC O1 p2 (following arabinose induction) did not lead to an increased capacity for autoaggregation or biofilm formation. To test whether AatA plays a role in adherence, CEF cell monolayers were infected with APEC O1, APEC O1 MaatA, the expression-inducible strain APEC O1 P2, and the empty plasmid control strain APEC O1 P1. Enumeration of bacteria adhering to cells did not reveal significant differences between APEC O1 MaatA and the wild-type APEC O1 strain in adherence to this cell line. However, APEC O1 p2, following induction of AatA, showed 5-fold-greater efficiency in adherence to CEF cells than the wild type, the APEC O1 MaatA mutant, and the empty plasmid control strain APEC O1 P1 (P < 0.005) (Fig. (Fig.5).5). We also included noninduction (without arabinose) and glucose repression (with 0.2% glucose) controls for APEC O1 p2. No significant difference in adherence capacity among the wild type, the mutant strain, APEC O1 p1, and these noninduction and glucose repression controls was found (data not shown). These results suggest that the production of AatA by APEC O1 significantly enhances the capacity of this strain to adhere to CEF cells.
Since aatA expression enhanced adherence to CEF cells, it was regarded as a putative virulence gene. To test the contributions of this gene to APEC virulence, the abilities of APEC O1, APEC O1 MaatA, and the complemented strain APEC O1 CaatA to cause disease in chickens were compared. Groups of 1-day-old birds were inoculated intratracheally with APEC O1, the isogenic mutant, or the complemented strain. All three strains were able to invade and infect deeper tissues, to generate gross lesions, and to cause a systemic infection and death. However, compared with the aatA mutant, the wild-type strain, APEC O1, caused earlier death, a higher level of mortality, and more severe lesions (Table (Table2).2). Within 48 h after infection, 6 of 12 chicks in the wild-type group had died, and four had severe lesions in the heart, liver, and air sacs (slight lesions were observed in two chicks that died within 24 h after infection). Bacteria could be reisolated from these organs and also from the brains of all dead chickens. After 7 days of infection, all survivors were euthanized. Among the survivors, two chickens in the wild-type group were found to have severe lesions in the heart, liver, and air sacs. On average, chickens in the wild-type group had lesion scores of 1.9 for the air sacs and 2.6 for the liver and heart. In contrast, only 4 of 12 chickens in the aatA mutant group died during the 7-day period, and the chickens infected with the mutant had significantly (P < 0.05) lower lesion scores (1.33 for the air sacs and 1.83 for the liver and heart) than the chickens infected with the wild type. The aatA-complemented derivative APEC O1 CaatA exhibited wild-type levels of virulence (Table (Table2).2). Despite some delay in the average time to death, the complemented strain caused mortality and gross lesions of air sacculitis and pericarditis-perihepatitis that were similar to those caused by the wild-type parent.
Since no statistically significant differences in mortality were noted in the chicken tests, we performed a chicken embryo lethality assay (ELA). The results of the ELA showed that the mortality rates for embryos with APEC O1 MaatA and wild-type strain APEC O1 were 20% (4/20) and 65% (13/20), respectively. No deaths occurred in the uninoculated group or in the groups inoculated with the negative-control strain DH5α or PBS. The mortality rate for the complemented strain was similar to that for the wild-type parent (60%, 12/20). The difference in mortality between the wild type and the mutant was significant (P < 0.05).
The vast majority of APEC strains harbor large virulence plasmids that encode both known and unknown phenotypes that contribute to virulence (27). In this study, we described a novel AT, AatA, encoded by the PAI of APEC O1's virulence plasmid, pAPEC-O1-ColBM, which mediates adherence to chicken fibroblasts and contributes to virulence. Gram-negative bacteria have evolved several specialized secretion pathways. The simplest and most widespread of these transport pathways is the AT or type V pathway. All ATs have the same general structure and are comprised of three domains: an amino-terminal leader peptide; an α or passenger domain, which confers the function of a secreted protein; and a carboxy-terminal domain (β domain) that mediates secretion through the outer membrane. AatA is predicted to have a β domain consisting of 284 amino acids, suggesting that it is a conventional AT, not a trimeric AT. The primary difference between it and a trimeric AT is that its C terminus, which is about 70 amino acids long, is sufficient for translocating the passenger domain across the outer membrane. In contrast, the translocator domains of conventional ATs consist of ~300 amino acids. For conventional ATs, like Pet (13), EspP (5), and Tsh (51), the passenger domain may be processed and released into the extracellular milieu, or it may remain in contact with the bacterial surface via a noncovalent interaction with the β domain after cleavage (e.g., Ag43  and AIDA-I ). In the case of AatA, neither autoproteolytic cleavage of the passenger domain nor cleavage by a membrane-bound protease was observed. However, we demonstrated that the intact 120-kDa protein is present in the outer membrane fraction but not in cell supernatant (data not shown). It is unclear whether cleavage requires special conditions, and how cleavage of the passenger domain from the translocation unit occurs is still not known (21).
Since its discovery in the late 1980s, the AT family has been expanded continuously (1, 31, 54). AatA is the third AT from APEC identified and the second AT which is encoded in the PAI of APEC O1's virulence plasmid. Flanking aatA are two mobile genetic elements, IS2 and IS629, suggesting that this gene was acquired horizontally. The gene encoding Tsh (temperature-sensitive hemagglutinin), the first serine protease AT of the Enterobacteriaceae (SPATE) described for APEC, is also in pAPEC-O1-ColBM's PAI. This gene occurs in 50% to 63.2%, of APEC strains but is rarely found in commensal E. coli strains (15, 45). tsh has also been found in UPEC strains, and its prevalence ranges from 4.0% to 4.5% (15, 45). Vat, another AT first described in APEC, is closely related to Tsh and shares 78% identity at the amino acid level, and the genes encoding these proteins have regions where there are high levels of nucleotide identity. Vat is also encoded by a gene located in a PAI. The vat gene is widespread in APEC (39.8%) and human ExPEC strains, including UPEC (54.5%) and NMEC (50%) strains (15). Unlike vat, aatA has been associated predominantly with avian E. coli strains, including APEC strains (40.3%) and AFEC strains (49.1%), but not with human ExPEC strains, and it occurs in only 2% of UPEC strains and 8.8% of NMEC strains. For this reason, we named the protein avian E. coli AT (AatA). The host specificity of this gene suggests that it may play a significant role in the pathogenesis of disease caused by APEC. Also, the similar prevalence of aatA in APEC and AFEC should not be taken as an indication that aatA is not a virulence gene. Ewers et al. (14) recently showed that AFEC strains exhibiting characteristics typical of APEC were capable of causing disease in immunocompetent chickens, meaning that the prevalence data cannot rule out the possibility that aatA is involved in APEC pathogenesis.
Several ATs encoded in genomic islands have previously been shown to be phylogenetically distributed (45). vat was significantly linked with phylogenetic group B2, as nearly all strains classified as group B2 strains, irrespective of their origins, were vat positive. In contrast, the conjugative virulence plasmid containing tsh has wider distribution among all strains regardless of phylogenetic group. Interestingly, aatA, which is also located on a plasmid, is significantly associated with isolates assigned to phylogenetic group D; 70% of APEC strains belonging to phylogenetic group D are aatA+, and more than one-half of all strains that were aatA+ belonged to phylogenetic group D. Similar to the group B2 strains, E. coli strains belonging to phylogenetic group D are experimentally and epidemiologically associated with extraintestinal infections (24, 43). A possible explanation for this phylogenetic position-specific distribution is that the plasmid containing aatA is limited largely to certain groups of strains due to incompatibility issues and selective pressure for retention of other plasmids (45).
Responding to changes in their circumstances, bacteria modulate their patterns of gene expression by upregulating genes that are specifically required for survival. Thus, it is likely that bacterial genes that are upregulated during infection or under conditions simulating infection compared to routine culture are candidate virulence genes. AatA could not be detected when the wild-type APEC O1 strain was cultured in LB. However, antibodies against AatA were detected in infected chickens, suggesting that AatA might be involved in the pathogenesis of colibacillosis. Similarly, a UPEC-specific trimeric AT, UpaG, was shown to have limited expression in vitro (54). Also, unpublished microarray data collected in our lab demonstrated that aatA was upregulated more than 10-fold when the organism was cultured in chicken serum, suggesting that aatA is an in vivo-induced gene (34, 50). Further sequence analysis of aatA revealed that the intergenic region between aatA and its upstream gene is very A-T rich, with several putative H-NS nucleation sites containing six or more matches with the 10-bp H-NS consensus sequence (data not shown) (30). These features coincide with features reminiscent of H-NS-repressed loci (10). The H-NS protein is an abundant global repressor that controls genes related to pathogenicity and stress responses. The gene encoding it is acquired by horizontal gene transfer and is found in PAIs (18, 22, 32, 35, 39, 52). H-NS silence may be relieved in vivo by a mechanism that is not yet clear (38). Thus, the possibility that a failure to detect AatA expression in vitro might be due to H-NS protein silencing warrants further study.
Unlike trimeric ATs, known conventional ATs function as cytotoxins, enterotoxins, immunoglobin proteases, mucinases, heme-binding proteins, and adhesins in E. coli and other Gram-negative bacteria (3, 6, 7, 46, 49, 54). AatA shares low amino acid sequence similarity in the α domain with two regions identified in an AidA-I adhesin-like protein from E. coli O157:H7 that contributes to a diffuse adhesion phenotype (4). Like AidA-I, AatA acts as an adhesin. Induction of expression of AatA was shown to enhance adhesion of APEC to CEF cells, although no difference in adherence between the wild type and the aatA deletion mutant was found. The lack of an adherence phenotype in the wild type may be simply due to silencing of expression of AatA in vitro, which is consistent with our SDS-PAGE and Western blot results. Tsh, a conventional AT, contributes to the early stages of infection, including colonization of air sacs (11), and EspP, an AT of E. coli O157:H7, has been shown to influence intestinal colonization of calves (12). Unlike AIDA-I, AatA did not function in autoaggregation or biofilm formation. The functional differences may due to the structural difference in the passenger domain of these ATs, since in silico analyses and empirical data from circular dichroism spectra suggest that the passenger domain of AIDA-I has a significant amount of β-strand structure (28).
Comparison of the pathogenicities of the aatA mutant strain, the wild-type strain, and the complemented strain revealed significant differences in organ lesions. The wild-type and complemented strains caused earlier and more deaths and more severe lesions in internal organs than the aatA mutant. Also, the results of the ELA showed that there were significant differences in mortality between the wild type and the mutant groups. These results, together with the results of the adherence assay, suggest that AatA may make a significant contribution to APEC virulence through bacterial adherence to host tissues. The fact that the aatA deletion mutant still caused death and lesions suggests that multiple factors contribute to adherence. Thus, deletion of only one adhesin does not appear to totally abolish the virulence of APEC O1. In our preliminary tests (data not shown) subcutaneous inoculation (1 × 106 CFU/chicken) of 1-day-old chickens was used to assess virulence. In this case, major differences in mortality and lesion scores between strains were not found (data not shown), indicating that this route of infection could not be used to detect differences in the capacities of strains to colonize the trachea and lung, probably because the bacteria were able to bypass the respiratory tract completely (53). Therefore, the difference between the results obtained with these two models also indicates that AatA may play a role in the early stages of infection, such as colonization of the trachea and lung.
This work was supported by the Iowa Livestock Health Advisory Council (ILHAC) and the USDA NRICGP Microbial Functional Genomics Program (grant 20083560418805).
We thank Paul M. Mangiamele for technical assistance and Kathy Mou, Ashraf Hussein, and Musafiri Karama for their indispensable help with the animal tests.
Editor: V. J. DiRita
Published ahead of print on 22 December 2009.