PafR, a Novel Transcription Regulator, Is Important for Pathogenesis ...

PafR, a Novel Transcription Regulator, Is Important for Pathogenesis in Uropathogenic Escherichia coli Mordechai Baum,a Mobarak Watad,a Sara N. Smith,b Christopher J. Alteri,b Noa Gordon,a Ilan Rosenshine,a Harry L. Mobley,b Orna Amster-Chodera Department of Microbiology and Molecular Genetics, IMRIC, The Hebrew University Faculty of Medicine, Jerusalem, Israela; Department of Microbiology and Immunology, University of Michigan Medical School, Ann Arbor, Michigan, USAb

The metV genomic island in the chromosome of uropathogenic Escherichia coli (UPEC) encodes a putative transcription factor and a sugar permease of the phosphotransferase system (PTS), which are predicted to compose a Bgl-like sensory system. The presence of these two genes, hereby termed pafR and pafP, respectively, has been previously shown to correlate with isolates causing clinical syndromes. We show here that deletion of both genes impairs the ability of the resulting mutant to infect the CBA/J mouse model of ascending urinary tract infection compared to that of the parent strain, CFT073. Expressing the two genes in trans in the two-gene knockout mutant complemented full virulence. Deletion of either gene individually generated the same phenotype as the double knockout, indicating that both pafR and pafP are important to pathogenesis. We screened numerous environmental conditions but failed to detect expression from the promoter that precedes the paf genes in vitro, suggesting that they are in vivo induced (ivi). Although PafR is shown here to be capable of functioning as a transcriptional antiterminator, its targets in the UPEC genome are not known. Using microarray analysis, we have shown that expression of PafR from a heterologous promoter in CFT073 affects expression of genes related to bacterial virulence, biofilm formation, and metabolism. Expression of PafR also inhibits biofilm formation and motility. Taken together, our results suggest that the paf genes are implicated in pathogenesis and that PafR controls virulence genes, in particular biofilm formation genes.

U

rinary tract infection (UTI) is the most common form of extraintestinal bacterial infection acquired by humans, and uropathogenic Escherichia coli (UPEC) is the most common cause for UTI (1) (2). UPEC colonizes the bladder but may subsequently ascend the ureters and establish a secondary infection in the kidney. Left untreated, UPEC can elicit serious complications, including septicemia and death. Up to 30% of patients with UTI experience recurrent episodes (3). Only a few serogroups of E. coli cause a high proportion of infections. The UPEC strain CFT073, originally isolated from the blood and urine of a patient with acute pyelonephritis (4), is considered a prototype UPEC strain (5, 6). The virulence of CFT073 has been reproduced in the CBA mouse model of ascending UTI (4), where it can cause acute pyelonephritis (7). Sequencing of the CFT073 genome revealed that many of the open reading frames encode hypothetical or putative proteins (8), suggesting that many virulence determinants that play a role in UTI pathogenesis are still unknown. Using comparative genomic hybridization analysis of a panel of uropathogenic and fecal/commensal E. coli isolates, 131 genes were found to be present only in UPEC strains (9). The large number of hypothetical and unannotated genes among these UPEC-specific genes reinforces the notion that many urovirulence factors remain uncharacterized. Significantly, the UPEC-specific genes are not necessarily typical urovirulent genes, such as genes encoding fimbriae, hemolysin, cytotoxic necrotizing factor, autotransporters, or iron acquisition systems. The identification of a relatively high number of genes encoding predicted transcriptional regulators among the UPEC-specific genes suggests that focusing on this family of genes may reveal novel information about the mechanisms underlying urovirulence (9). The genome of CFT073 contains 13 large genomic islands (9). The role of the majority of these genomic islands in pathogenesis has been studied by their systematic deletion, one at a time. Dele-

October 2014 Volume 82 Number 10

tion of three islands—metV, aspV, or asnT—reduced the ability of the resulting mutant constructs to colonize the urinary tract compared to that of the wild-type strain, as indicated by transurethral cochallenge experiments in CBA/J mice. The only genomic island whose deletion resulted in inability to colonize the urinary tract of mice during an independent challenge was the metV island. Further deletion constructs demonstrated that the region that contains the genes c3405 to c3410 (c3405-c3410) is responsible for this attenuation (10). The c3405-c3409 genes appear to constitute an operon, which we hereby term paf, found exclusively in extraintestinal pathogenic E. coli (ExPEC) strains and primarily in UPEC strains (10). When a sequence that contains the first two genes of this operon, c3409 and c3408 (here designated pafR and pafP, respectively), was used as a probe for DNA hybridization, it was identified in strains associated with clinical syndromes of cystitis (81%) and acute pyelonephritis (79%) but rarely in fecal strains isolated from healthy people (19%) (11). Sequence analysis of the pafR gene revealed that the predicted PafR protein shares 55% amino acid sequence similarity and 30% identity with BglG, which represents a family of factors that prevent premature termination of transcription at intrinsic termina-

Received 16 January 2014 Returned for modification 23 February 2014 Accepted 23 July 2014 Published ahead of print 28 July 2014 Editor: S. M. Payne Address correspondence to Orna Amster-Choder, [email protected]. Supplemental material for this article may be found at http://dx.doi.org/10.1128 /IAI.00086-14. Copyright © 2014, American Society for Microbiology. All Rights Reserved. doi:10.1128/IAI.00086-14

Infection and Immunity

p. 4241– 4252

iai.asm.org

4241

Baum et al.

tors preceding sugar utilization genes and operons in Gram-negative as well as Gram-positive bacteria (12–14). The activity of BglG-like factors is regulated by phosphotransferase system (PTS) sugar permeases, often encoded from the same operon (15). Intriguingly, the pafP gene, which is adjacent to pafR, encodes a putative PTS sugar permease, which shares 53% sequence similarity and 42% identity to MalX, a maltose/glucose-specific transporter from E. coli. Based on these annotations, the PafR and PafP proteins (previously designated R1 and R2) are predicted to form a Bgl-like sensory system that responds to the presence of certain PTS sugars in the environment by preventing premature termination of transcription at intrinsic terminators (15). Some BglG-like factors regulate transcription of their own genes or operons, whereas some antiterminate transcription at other genetic loci. Significantly, the paf operon does not contain internal transcription terminators, suggesting that PafR regulates expression of other genes or operons in the CFT073 genome. Here we have shown that the paf genes are important for pathogenesis, since their knockout results in a dramatic reduction in colonization in the murine model, whereas their expression in trans restores full virulence. Our results further suggest that the paf genes are induced in vivo. Using microarray analysis to identify the targets of the PafR transcriptional antiterminator in the UPEC genome, we found that PafR affects expression of genes traditionally related to virulence, biofilm formation, and metabolism. Moreover, expression of PafR inhibits biofilm formation and bacterial motility. Hence, PafR is a novel transcription factor that controls virulence genes, especially biofilm formation genes. MATERIALS AND METHODS Bacterial strains and growth conditions. CFT073 is a highly virulent UPEC strain isolated from the urine and blood of a woman with UTI (4). The ⌬pafRP and ⌬pafP strains are derivatives of CFT073, in which the pafRP or pafP gene has been replaced with a kanamycin cassette, using the method developed by Datsenko and Wanner (16). In short, CFT073 bacteria carrying the pKD46 helper plasmid were cultured at 30°C until they reached an optical density at 600 nm (OD600) of 0.4 to 0.55. Bacteria were transformed with a PCR fragment that contained the kanamycin resistance cassette, which was amplified with oligonucleotides that contained, in addition to the Kanr sequences, 50 nucleotides (nt) that are homologous to the two ends of the genes to be replaced. The ⌬pafR mutant is a CFT073 derivative in which the Tn5 transposon, which carries the kanamycin resistance cassette, has been inserted 6 nt after the beginning of the pafR open reading frame. MA152, used for the antitermination assay, is an E. coli K-12 strain that carries a chromosomal bglG-lacZ fusion (17). The E. coli K-12 strain used in this study was MG1655. Bacteria were cultured in Luria-Bertani (LB) medium or M9 minimal medium, prepared as described before (18). Artificial urine medium (AUM) was prepared as described previously (19). Soft agar plates, used for the motility assay, contained 1% tryptone, 0.5% NaCl, and 0.25% agar with 0.5 mM isopropyl-␤-D-thiogalactopyranoside (IPTG) and were prepared the day before use and left at room temperature. Ampicillin (100 ␮g/ml), kanamycin (30 ␮g/ml), and chloramphenicol (25 ␮g/ml) were included in the medium for growing strains containing plasmids that confer resistance to one of these antibiotics. To assay the ability of the CFT073 and K-12 strains that express PafR to grow on maltose, single colonies were streaked on M9 agar plates containing 0.22 mM maltose as a sole carbon source and 0.5 mM IPTG and incubated at 37°C overnight. Competition of bacterial growth in vitro. Overnight cultures of CFT073 and the ⌬pafRP strain were diluted 1:200 and grown to an OD600 of 0.2. Equal amounts (1 ⫻ 108) of the two cultures were added to 20 ml of fresh medium, and the mixed culture was grown at 37°C. Samples were

4242

iai.asm.org

removed every 24 h for 2 days, diluted, and plated on plates with and without antibiotics. The numbers of colonies that represent each strain were calculated by counting the number of colonies that grew in the presence of kanamycin, which represent the ⌬pafRP strain, and subtracting this number from the number of colonies that grew in the absence of the antibiotic, which represent the CFT073 strain. Then, the actual number of bacteria in each culture was calculated from the dilution plated. Plasmids. pGEN-MCS (20) was used as a vector in the mouse complementation assay. pGEN-pafRP was constructed by amplifying the pafR and pafP genes, together with the 235 bp that precedes pafR, by PCR and ligating the resulting fragment to pGEN-MCS that had been cleaved with EcoRI and SphI. pRO-gfp, used for monitoring expression from the putative paf promoter, was constructed by amplifying a 347-bp-long DNA fragment (282 bp upstream of the pafR gene and the first 65 bp of pafR), which includes the paf putative promoter (Ppaf), by PCR and ligating it to pIRI, a plasmid carrying a promoterless gfp gene (21), which had been cleaved with BamHI and XbaI. pIRI served as a negative control for gfp expression, whereas pSA11 (22) was used as a positive control. pQE-SG, used for pafR expression, was constructed by amplifying pafR and cloning it into pQE-31 (Qiagen), the latter being used as vector-only control. pMN25 (17), which expresses bglG constitutively, was used as a positive control in the antitermination assay. pKD4 (16), which served as a template for amplifying the kanamycin resistance cassette, and pKD46, which expresses lambda Red, were used for the construction of the ⌬pafRP and ⌬pafP strains (16). Murine model of ascending urinary tract infection. Six- to 8-weekold female CBA/J mice (20 to 22 g; Jackson Laboratories) were anesthetized with ketamine-xylazine and inoculated transurethrally over a 30-s period with a 50-␮l bacterial suspension per mouse using a sterile polyethylene catheter (inside diameter [i.d.], 0.28 mm; outside diameter [o.d.], 0.61 mm) connected to an infusion pump (Harvard Apparatus). To measure relative fitness, overnight LB cultures for CFT073 and mutant constructs were collected by centrifugation and resuspended in sterile phosphate-buffered saline (PBS), mixed 1:1, and adjusted to deliver 2 ⫻ 108 CFU per mouse. Dilutions of each inoculum were spiral plated onto LB with and without kanamycin using an Autoplate 4000 instrument (Spiral Biotech) to determine the input CFU/ml. After 48 h postinfection (hpi), mice were sacrificed by overdose with isoflurane, and the bladders, kidneys, and spleens were aseptically removed, weighed, and homogenized in sterile culture tubes containing 3 ml of PBS using an Omni mechanical homogenizer (Omni International). Appropriate dilutions of the homogenized tissue were then spiral plated onto duplicate LB plates with and without kanamycin to determine the output CFU/g of tissue. Plate counts obtained on kanamycin were subtracted from those on plates lacking antibiotic to determine the number of wild-type bacteria. Competitive indices were calculated by dividing the ratio of wild type to mutant at 48 hpi by the ratio of wild-type to mutant input CFU/ml. Groups of ⱖ5 mice per cochallenge were used to determine defects in fitness. For assessment of virulence, mice were inoculated singly with CFT073 or the mutant. Statistically significant differences in colonization (P ⬍ 0.05) were determined using a two-tailed Wilcoxon matched-pairs test. All animal protocols were approved by the University Committee on Use and Care of Animals at the University of Michigan Medical School. Phenotypic microarray analysis. An overnight culture of CFT073 transformed with pRO-gfp was diluted in fresh M9 minimal medium without a carbon source and normalized to an OD600 between 0.06 and 0.09. Aliquots (120 ␮l) of this culture were transferred to each well in the phenotypic microarray plate (Biolog) and cultured in the fluorometer (Victor3V; PerkinElmer) for 1 h at 37°C with shaking. Subsequently, 100 ␮l was transferred from each well of the phenotypic microarray plate to a new 96-well microplate and covered with 50 ␮l of mineral oil. The new plate was placed in the fluorometer, and the gfp levels were monitored every 30 min for 72 h. ␤-Galactosidase assay. Strains of interest were cultured overnight in M9 minimal medium while shaking at 37°C. Bacteria were diluted 1:100 into a


PafR Is a Novel UPEC Regulator

fresh medium and incubated until an OD600 of ⬇0.4 to 0.5 was reached. Measurement of ␤-galactosidase activity was carried out as described previously (18). Statistical analysis was performed by using the Student t test. Microarray analysis. Bacteria were cultured in LB with 1 mM IPTG until an OD600 of 0.6 was reached, harvested, and resuspended in 4 ml of ice-cold CI buffer (25 mM HEPES, 10 mM magnesium acetate [MgOAc], 50 mM potassium acetate [KOAc], 250 mM NH4Cl, 0.1 mM EDTA, and 0.1 mM dithiothreitol [DTT]). RNA was purified from 250 ␮l of the resuspended cells (in biological duplicate samples) using TRI Reagent (Sigma). The quality of the purified RNA was assessed by gel electrophoresis. Conversion of RNA to cDNA and DNA labeling and hybridization to the Affymetrix GeneChip E. coli Genome 2.0 Array were performed according to the Affymetrix Expression Analysis Technical Manual (available at http://media.affymetrix.com/support/downloads /manuals/expression_analysis_technical_manual.pdf). The cDNA samples were prepared from biological duplicates. The Affymetrix chips contain inherent controls, since each probe pair consists of a perfectmatch probe, which has a sequence exactly complementary to that of the particular gene, thus measuring the expression of the gene, and a mismatch probe, which differs from the perfect-match probe by a single base substitution at the center base position, thus determining the background and nonspecific hybridization; these values are later subtracted by the GeneChip Expression Console software program from the perfect-match probe’s intensity to determine the absolute or specific intensity value for each probe set (http://medicine.yale.edu/keck /ycga/microarrays/affymetrix/index.aspx). Microarray data analysis. Analysis of the Affymetrix microarray data was carried out at the Hebrew University (HU) Center for Genomic Technologies using the Spotfire software program. Extraction of LPS. Lipopolysaccharide (LPS) extraction was performed as described previously (23). In short, bacteria were cultured in LB under static conditions for 48 h at 37°C in the presence of 0.5 mM IPTG and normalized to an OD600 of 1, and LPS was extracted using an LPS extraction kit (Intron Biotechnology, Boca Raton, FL) according to the manufacturer’s protocol. The LPS extract was separated on a 12.5% SDSpolyacrylamide gel under reducing conditions and stained by using a silver stain kit for LPS (161-0443; Bio-Rad) according to the manufacturer’s instructions, with the exception that periodic acid was added to the fixation solution to a final concentration of 0.7% (24). The LPS profile was photographed by the Bio-Rad Gel Doc XR imaging system and processed by the ImageJ software program. Real-time PCR. Bacteria were cultured in the presence of 0.5 mM IPTG until reaching an OD600 of 0.6, and samples of 0.5 ml were removed. RNA was extracted by using the TRI Reagent kit (Ambion). RNA concentration and purity were determined using a NanoDrop machine (NanoDrop Technologies), and 1 ␮g of RNA from each sample was treated with RQ1 RNase-free DNase (Promega) for 1 h. To check for DNA contamination, 1 ␮l from each RNA preparation was amplified by PCR with DreamTaq green DNA polymerase (Thermo Scientific), using the appropriate primers. cDNA was synthesized by using a RevertAid First Strand cDNA synthesis kit (Thermo Scientific). Quantification of cDNA was performed by real-time PCR using Thermo ScientificABsolute Blue qPCR SYBR green ROX mix (Thermo Scientific) with a Rotor-Gene 3000 A system (Corbett) according to the manufacturer’s instructions. The level of gapA was used to normalize the expression data for each gene. The expression data were analyzed using the 2-⌬⌬CT quantification methods. Primers sequence were as follow: gapA_F, GTTGGTGTTGACGTTGTC; gapA_R, GCGCCTTTAACGAACATC; pafP_F, GCACTTGGTGGTAAA GAG; pafP_R, CTTTAGTGCCGATAATGACTTG; c1492_F, CTTGCAC TTTCTCACAGTG; c1492_R, CAACTCATTTATTACAGGCTC; wrbA_ F, GCAGAAGTTGTCGTTAAGC; wrbA_R, GAAGGTACGCATTTG ACC; frdB_F, CGAAGTATCACCAGTTCTCC; frdB_R, CCATACGCTC CTTCTTACC. Biofilm formation assay. The ability of bacteria to form biofilm was assayed as described previously (25) with a few modifications. Briefly,


bacteria from an overnight culture were diluted 1:100 into fresh LB medium containing 0.5 mM IPTG and dispensed into sterile 96-well polystyrene plates. Plates were incubated at 37°C for 48 h without agitation. After 48 h, the medium was gently removed, and the wells were washed three times by immersing the plate gently in a container with tap water. After decanting the water, the wells were stained with 200 ␮l of 1% crystal violet for 15 min at room temperature. The wells were washed gently as described previously, and the amount of crystal violet in each well was quantified by solubilizing it in 80% ethanol and 20% acetone and recording the absorbance at 600 nm, using a plate reader (Victor3V; PerkinElmer). Motility assay. Bacterial motility was evaluated as described elsewhere (26). In short, overnight cultures were diluted 1:100 in fresh medium and cultured to an OD600 of 0.9 to 1.1. Cultures were stabbed into the middle of soft agar plates with a sterile inoculating needle. Plates were incubated at 30°C for approximately 12 h, and the diameter of the colony, which is correlative to bacterial motility (27), was measured. Fluorescence microscopy. CFT073 bacteria containing pRO-gfp (carrying Ppaf::gfp), pIRI (negative control), or pSA11 (positive control) were cultured to an OD600 of 0.6 in minimal medium containing a 15 ␮M concentration of one of the following carbon sources: maltose, glucose, succinate, salicin, or fructose. Expression of GFP from pSA11 was induced with 10 ␮M IPTG. Aliquots of 0.5 ml were removed from each culture, gently centrifuged, and resuspended in PBS (80 mM disodium phosphate [Na2HPO4], 20 mM NaH2PO4, and 100 mM NaCl). Cells were visualized and photographed using an Axioplan2 microscope (Zeiss) equipped with a CoolSnap HQ camera (Photometrics, Roper Scientific). Image processing was performed using ImageJ software (28). Immunoblot analysis. Cells were cultured in LB or in M9 minimal medium containing maltose as a sole carbon source until reaching an OD600 of 0.4. Bacteria from 2 ml culture were harvested by centrifugation and resuspended in 0.1 ml of 20% sucrose in 0.03 M Tris-HCl. Lysozyme (10 ␮l of 1 mg/ml) in EDTA, pH 8, was added, and the samples were incubated for 30 min on ice. Bacteria were frozen in liquid nitrogen and thawed at room temperature four times, followed by sonication. Samples were centrifuged, and the supernatant was collected. Equal amounts of protein were separated on 12.5% SDS-polyacrylamide gels. Purified maltose binding protein (MBP) (NEB) was loaded as a positive control. Gels were subjected to Western blot analysis essentially as described previously (29). Incubation with the primary antibody (anti-MBP, monoclonal, diluted 1:3,000) was carried out for 1 h at room temperature (RT), and incubation with the secondary antibodies (goat anti-mouse, diluted 1:20,000) was for 30 min at RT. The membrane was also reacted with anti-RpoA antibodies to verify that equal amounts of cell extract were analyzed in each case. Antibody-bound proteins were detected by an enhanced chemiluminescence (ECL) light-based detection procedure (Amersham Pharmacia) and visualized by using the LAS 3000 luminescent image analyzer (Fuji).

RESULTS

Deletion of pafR and/or pafP impairs the ability of CFT073 to colonize the urinary tract. Based on DNA hybridization analysis of UPEC strains isolated from clinical isolates, the presence of the pafR and pafP genes (previously referred to as c3409 and c3408, respectively) in CFT073 were found to be significantly associated with isolates causing clinical syndromes (10, 11). To directly examine the involvement of these genes in UTI, we tested the effect of deletion of these genes on the pathogenicity of CFT073 in a mouse model of ascending UTI. First, we performed an independent challenge experiment by transurethrally inoculating CBA/J mice with equal numbers of bacteria of either CFT073 (wild type) or an isogenic mutant strain, in which we deleted the pafR and pafP genes (⌬pafRP). Twenty-four hours postinoculation, urine samples were collected and the bladder, kidney, and spleen were removed, homogenized and plated. The numbers of wild-type and ⌬pafRP bacterial cells (CFU) per ml of urine and per gram of

iai.asm.org 4243

Baum et al.

FIG 1 pafR and pafP contribute to pathogenesis of uropathogenic E. coli. Comparative colonization levels, presented as competitive index, in the urine, bladder,

and kidneys of CBA/J mice 48 h after transurethral inoculation with a mixture (cochallenge) of wild-type CFT073 with the ⌬pafRP strain (n ⫽ 19) (A), the ⌬pafRP strain carrying a plasmid expressing pafRP (n ⫽ 10) (B), the ⌬pafR strain (n ⫽ 10) (C), or the ⌬pafP strain (n ⫽ 9) (D). Bars indicate the median competitive index. P values for cochallenge infections were determined using the Wilcoxon matched-pair test.

homogenized bladder, kidney, or spleen tissues were compared. The results, presented in Fig. S1A in the supplemental material, revealed a significant decrease in CFU in the kidneys of mice infected with the ⌬pafRP mutant compared to results for the wildtype strain (P ⫽ 0.054). These results suggest that deletion of pafRP reduces the ability of CFT073 to fully colonize in the host. To further examine the possibility that pafR and pafP contribute to bacterial virulence, we performed a cochallenge experiment, during which CBA/J mice were transurethrally inoculated with a mixture of the wild-type strain and the ⌬pafRP construct at a 1:1 ratio. Bacterial counts in the urine, bladder, and kidney were measured as described above. The results, presented in Fig. 1A, demonstrate that the mutant strain was significantly outcompeted by the wild type in the urine (P ⬍ 0.0001), bladder (P ⬍ 0.0001), and kidneys (P ⬍ 0.0009). Importantly, the median levels of colonization of the ⌬pafRP strain were at the limit of detection in all cases (see Fig. S1B in the supplemental material), indicating that the pafR and pafP genes have an important role in UPEC colonization throughout the urinary tract. Of note, we ruled out the possibility that the observed difference are due to reduced growth of the ⌬pafRP strain compared to that of the wild type, since the ⌬pafRP strain was not outcompeted by CTF073 when the strains were cultured together in a flask (see Fig. S2A).

4244

iai.asm.org

To corroborate that the observed impairment in colonization was due to the deletion of pafR and pafP, we tested whether this effect could be complemented by expressing the two genes in trans in the ⌬pafRP construct. To this end, we repeated the cochallenge experiment with a mixture of the wild-type strain and the ⌬pafRP construct, the latter complemented with a plasmid carrying pafR and pafP. The results (Fig. 1B) show that expression of the pafRP genes in trans restored colonization of the ⌬pafRP construct in the bladder and kidney to a level similar to that of the wild-type (i.e., numbers of CFU/g not different between the wild type and the complemented mutant). These results confirm that the pafR and pafP genes contribute to colonization of the urinary tract and their deletion impairs UPEC virulence. Finally, to examine the importance of pafR and pafP individually for UPEC infection, we deleted them one at a time from the CFT073 chromosome and tested the ability of the resulting mutant constructs to compete with CFT073 in a cochallenge experiment. The results of these experiments are shown in Fig. 1C and D. The ⌬pafR mutant was significantly outcompeted by CFT073 in the urine, kidney, and bladder (P ⫽ 0.0098, P ⫽ 0.0527, and P ⫽ 0.0020, respectively). The ⌬pafP mutant was also outcompeted by CFT073 in the urine and kidney (P ⫽ 0.0391 and P ⫽ 0.0039, respectively), although it was not significantly outcompeted in the



bladder. Of note, differential colonization of the different parts of the urinary tract by mutants of specific genes is often observed in the murine model of ascending UTI, for example, a mutant in the hma-encoded heme binding protein exhibited a similar behavior (30). This is explained by substantial differences between the kidneys and bladder, with the first environment being highly vascularized and well oxygenated, whereas the second is less vascularized and carries a microaerobic environment during an infection. Taken together, the results described thus far indicate that the pafR gene, as well as the pafP gene, contributes to the fitness of CFT073 in the urinary tract. Defining the conditions that induce expression of the paf genes. To study the conditions that lead to expression of the paf genes outside the host, we constructed a transcriptional fusion of the paf promoter and green fluorescent protein (gfp). To this end, we cloned a fragment that contains 282 bp upstream of the pafR gene and the first 65 bp of pafR, upstream of gfp gene. This region is predicted to contain the promoter of the paf operon, here termed Ppaf, with the putative ⫺10 and ⫺35 elements starting at ⫺153 bp and ⫺173 bp upstream of the pafR open reading frame, respectively. The ability of the pafRP genes, preceded by the 235 bp before pafR, to complement the ⌬pafRP strain in a cochallenge experiment in mice (see above) indicates that this region indeed contains a promoter. We introduced the recombinant plasmid carrying the Ppaf::gfp fusion into CFT073 and monitored expression of the gfp gene in the transformed cells by measuring fluorescence intensity during growth. GFP could hardly be detected when the cells were cultured in LB. To find environmental conditions under which expression from Ppaf can be induced in vitro, we monitored the level of GFP under the following growth conditions, some known to prevail in the urinary tract or to induce virulence genes in general: (i) growth in urine, (ii) different temperatures, (iii) different pH, (iv) different iron concentrations, (v) different Ca⫹2 concentrations, (vi) anaerobic and aerobic conditions, (vii) different growth phases, and (viii) different carbon sources. Based on these experiments, the following conclusions could be drawn: first, under all conditions, except for growth in the presence of certain carbon sources, expression of Ppaf:: gfp was not detected. Second, mild expression of Ppaf::gfp was detected in the presence of maltose, in line with the predicted role of PafP as a maltose permease, and in the presence of fructose, N-acetyl-glucosamine, salicin, and succinate but not in the presence of glucose (Fig. 2A), suggesting that the promoter is subjected to catabolite repression. To further explore the effect of carbon sources on expression from Ppaf, we monitored expression of Ppaf::gfp in the presence of 190 carbon sources, using phenotypic microarray (PM) technology (31, 32). The PM analysis allowed us to conduct a simultaneous, high-throughput comparison of the GFP levels in cells grown with different carbon sources in the same microplate. The results in Fig. 2B show that expression of Ppaf::gfp was detected in the presence of certain carbon sources, again including maltose; Fig. 2B shows representative results, and the others are listed in Table S1 in the supplemental material. However, expression of Ppaf::gfp was not detected in the presence of most carbon sources, represented by adonitol and dulcitol in Fig. 2B (see Table S1) (below 300,000 fluorescence units). Of note, even under conditions that allowed GFP detection, its expression level from Ppaf was very low, as indicated by the fluorescence microscopy images presented in Fig. 2A. Finally, we tested the effect of maltose on pafR expression by comparing the levels of the chromosome-encoded pafR tran-


scripts in CFT073 cells, cultured in artificial urine medium (AUM) and in AUM supplemented with maltose, by real-time PCR. The results, presented in Fig. S3, show that the difference between the pafR transcript levels produced under the two conditions is not significant. Together, our results suggest that the putative promoter of the paf operon is mildly responsive to certain sugars, including maltose, and raise the possibility that the paf genes are ivi (in vivo-induced) genes, i.e., they are expressed mainly in the host (33). PafR can function as a transcriptional antiterminator. The pafR and pafP genes encode putative transcriptional antiterminator and maltose/glucose-specific phosphotransferase system (PTS) permease, respectively (see the introduction). To examine whether PafR (previously designated R1) can function as an antiterminator, we examined its ability to antiterminate transcription of the bgl operon in E. coli K-12 in vivo. For this purpose, we expressed PafR from a plasmid in the E. coli K-12 strain MA152, which is deleted for the bgl operon and carries a chromosomal fusion of the bgl promoter and terminator to the lacZ gene. Our positive control was expression of the bgl-lacZ fusion in the presence of the bgl operon native antiterminator, BglG, which has also been expressed from a plasmid. As expected, lacZ expression was induced in the presence of BglG, as indicated by comparing the level of ␤-galactosidase in its presence and absence (BglG and vector only, respectively, in Fig. 3A). Importantly, expression of the PafR protein from CFT073 in the heterologous MA152 strain also resulted in expression of ␤-galactosidase (compare PafR and vector-only results in Fig. 3A), albeit at a level that is lower than the level obtained in the presence of the endogenous antiterminator BglG (compare results for PafR and BglG in Fig. 3A). These results indicate that the PafR protein from CFT073 can function as an antiterminator of transcription. The lack of identifiable terminator sequences within the paf operon raises the possibility that PafR regulates transcription of other genes or operons in CFT073. Transcriptome analysis of PafR expression in CFT073. The predicted function of PafR as a transcription factor and its ability to antiterminate transcription of the bgl operon in E. coli K-12 prompted us to try identifying genes whose expression is affected by PafR. To detect the transcriptomic changes that PafR may induce, we used microarray analysis. To this end, we isolated total RNA from three strains, wild-type CFT073, the ⌬pafRP strain, and CFT073, that expresses PafR from a plasmid (here referred to as WT, KO, and PafR-expressing, respectively) and used them to prepare cDNA libraries (in biological duplicates) that were hybridized to the Affymetrix GeneChip E. coli Genome 2.0 microarray, which represents the entire CFT073 genome (each probe set represents 11 antisense and 11 mismatch probes for each CFT073 gene; see Materials and Methods). Of note, the WT, KO, and PafR-expressing strains demonstrated the same growth rates (see Fig. S2A and B in the supplemental material), ruling out the possibility that the differences between their microarray profiles are due to differential growth. To validate the microarray results, we chose two genes that were upregulated and two that were downregulated and quantified their expression level in the ⌬pafRP, WT, WT with vector only, and PafR-expressing strains by real-time PCR. The results in Fig. S4 show that the levels of pafP and c1492 transcripts, which were upregulated in the PafR-expressing strain, and of frdB and wrbA transcripts, which were downregulated in the PafR-expressing strain, demonstrated the same tendency in

iai.asm.org 4245

FIG 2 Monitoring expression of GFP from Ppaf, the putative paf promoter, in CFT073 cells. (A) Expression of Ppaf::gfp in the presence of the indicated carbon sources was imaged by fluorescence microscopy. (B) Expression of Ppaf::gfp in the presence of 190 different carbon sources was monitored in phenotypic microarray (PM) plates. Fluorescence units represent the recorded fluorescence normalized to cell density. Shown are results with representative carbon sources. The results with all carbon sources are presented in Table S1 in the supplemental material.

4246

iai.asm.org



FIG 3 Regulatory activities of PafR. (A) PafR is capable of preventing premature termination of transcription. Plasmids carrying the E. coli K-12 bglG gene, the CFT073 pafR gene, or the cloning vector only were introduced into MA152, an E. coli K-12 strain which is deleted for the bgl operon and carries a chromosomal fusion of the bgl promoter and terminator to the lacZ gene. The ability of the plasmid-encoded proteins to enable lacZ expression was tested by measuring ␤-galactosidase levels. (B) PafR is a major regulator. Shown are groups of genes that were differentially expressed in PafR-expressing CFT073 bacteria and in ⌬pafRP bacteria, as indicated by oligonucleotide-based microarray analysis. Black bars, upregulated genes; gray bars, downregulated genes. The lists of all upregulated and downregulated genes in PafR-expressing CFT073 compared to those in ⌬pafRP cells are presented in Data Sets S1 and S2, respectively, in the supplemental material.

the quantitative PCR experiment, confirming the reliability of the microarray results. Notably, among the top 50 genes that were either upregulated or downregulated in the PafR expression strain (see Tables S2 and S3, respectively, in the supplemental material), 36 and 8 were unannotated hypothetical proteins, respectively. Below, we discuss genes belonging to various gene groups that displayed at least a 1.66-fold differential expression (Fig. 3B; see also Data Sets S1 and S2). The first result that emerged from the microarray analysis is that the expression profiles obtained with the WT and KO strains were very similar, exhibiting no expression of the paf genes (the same as genes not present in CFT073) and comparable differences from the PafR-expressing strain. These results confirm our supposition that the paf genes are ivi genes, that is, they are not expressed under routine growth conditions outside the host. Hence, in Data Sets S1 and S2 in the supplemental material, we present the differential expression results obtained with the PafR-expressing strain versus those with the KO strain only. In total, expression of 1,026 genes, 70% of which have suggested annotation, was affected by PafR expression. Expression of


647 genes was upregulated, whereas expression of 379 genes was downregulated, in the PafR-expressing strain compared to results for the KO strain, respectively. Most of the PafR-induced up- or downregulated genes can be classified into several functional groups (Fig. 3B). Significantly, many of these genes are known to be associated with virulence. First, various genes whose protein products are associated with fimbria production and cell adhesion were mostly upregulated in response to PafR expression, including different fimbrial subunits, fimbrial chaperones, and P pilusencoding genes (see Data Set 1, under “Adhesion”). Second, hemolysin genes, encoding type 1 secretion machinery that secretes the hemolysin A (HlyA) toxin (34), were also upregulated in response to PafR expression (see Data Set 1, under “Toxins”). Third, we observed significant changes in the expression of several genes associated with biofilm formation (Fig. 4A); two of these genes, yahA and ymgB (ariR), both among the top 50 upregulated genes, with yahA being at the top of this list, are expected to negatively regulate biofilms. Then again, we also observed changes in the expression of a large number of genes associated with metabolism, with a significant number of these genes being among the top 50

iai.asm.org 4247

Baum et al.

FIG 4 Expression of PafR in UPEC affects biofilm formation and motility. (A) An array of biofilm-related genes is upregulated in PafR-expressing CFT073 cells compared to results for ⌬pafRP cells, as indicated by oligonucleotide-based microarray analysis. (B) Biofilm formation by wild-type CFT073 and PafR-expressing cells, assessed by the crystal violet coloration assay. (C) Average diameters of CFT073 and PafR-expressing colonies on soft agar plates. (D) Colonies of CFT073 and PafR-expressing cells grown for 12 h on soft agar plates.

upregulated and downregulated genes. Examples include a large group of genes involved in carbon metabolism and a large group of electron transport genes, including metal ion binding/iron-sulfur cluster binding, which were mainly downregulated upon PafR expression. These results might reflect preference for the tricarboxylic acid (TCA) cycle over electron transfer as an energy-gaining mechanism under conditions of increased carbohydrate metabolism. Notably, expression of the entire maltose (mal) regulon was downregulated in the PafR-expressing strain (Fig. 5A), further supporting the involvement of the paf genes in maltose sensing and metabolism. Moreover, expression of PafR in CFT073 but not in E. coli K-12 inhibited growth on maltose as a sole carbon source (Fig. 5B). To substantiate this involvement and to verify that the observed changes in RNA level reflect changes in the protein level, we monitored the level of the maltose binding protein (MBP), encoded by the malE gene, using Western blot analysis. The levels of MBP in the WT and KO cells were comparable when these were cultured in LB or in minimal medium supplemented with maltose (Fig. 5C and D, respectively). However, the MBP level dropped upon PafR expression in CFT073 (Fig. 5C) (only MBP from cells cultured in LB is shown in the Western blot, since cells expressing PafR did not grow in the presence of maltose as a carbon source, as shown in Fig. 5B). In contrast, the MBP level in E. coli K-12 was not affected by PafR expression, as shown in Fig. 5E, confirming that PafR-mediated downregulation of the maltose genes is a UPEC-specific phenotype. Upregulation of genes responsible for the synthesis of lipo-

4248

iai.asm.org

polysaccharide (LPS), e.g., rfaI, rfaJ, and rfaY, and of the rfaH gene, which encodes a transcription factor that positively regulates their expression, in the PafR-expressing strain (35) (see Data Set S1 in the supplemental material) motivated us to compare the LPS profiles of the different strains that were analyzed in the microarray experiment. To this end, LPS was isolated from equal numbers of CFT073, ⌬pafRP, and PafR-expressing cells and separated on a reducing SDS-polyacrylamide gel. The results in Fig. S5 show that the LPS profile of PafR-expressing cells exhibits visible differences from those of CFT073 and KO both in the bands’ pattern and intensity, which correlate with the microarray results for upregulation of the rfa genes. Taken together, our results suggest that PafR is a transcription factor that regulates transcription of an array of genes associated with virulence and metabolism in UPEC. PafR inhibits biofilm formation. The microarray results, presented in Fig. 4A, show significant changes in biofilm-related genes in response to PafR expression. This, together with the reported linkage between virulence, recurrent UTI infection, and biofilms (36), encouraged us to ask whether PafR directly affects the ability of UPEC to form biofilms. To answer this question, we used the crystal violet coloration assay to quantitatively assess the ability of cells to form biofilms. First, we monitored the kinetic of biofilm formation by CFT073. The results in Fig. S6 in the supplemental material show that the biofilm capability reaches its peak after 2 days; the biofilms were dispersed between days 3 and 4. We then compared the biofilm formation capability of the PafR-ex-



FIG 5 PafR downregulates the level of mal gene transcripts and proteins in UPEC. (A) The entire mal regulon is downregulated in PafR-expressing CFT073

compared to expression in ⌬pafRP bacteria, as indicated by oligonucleotide-based microarray analysis. (B) CFT073 but not E. coli K-12 bacteria, which express PafR from a plasmid, cannot grow on minimal medium plates that contain maltose as a sole carbon source. (C and D) Maltose-binding protein (MBP), detected by Western blotting, in ⌬pafRP, wild-type, and PafR-expressing CFT073 bacteria cultured in LB (C) or in minimal medium with maltose as a sole carbon source (D). (E) MBP is expressed in comparable levels in E. coli K-12 bacteria that express PafR and in those that do not express it, as indicated by Western blotting with anti-MBP antibodies. Purified MBP was used as a positive control.

pressing strain to that of CFT073 2 days after the beginning of growth. The results in Fig. 4B demonstrate that PafR expression strongly inhibits biofilm formation by CFT073 (compare WT to “PafR-expressing”). These results confirm the involvement of PafR in biofilm formation. PafR affects bacterial motility. The strong correlation between biofilm formation and bacterial motility, i.e., that motility is necessary for the primary attachment of bacteria to a surface and for the dispersal of the biofilm colony (37, 38), and the reported dependence on motility for ascending the urinary tract and on sessility for establishing efficient infection (20, 39) motivated us to examine whether PafR expression affects motility of CFT073. For this purpose, we stabbed soft agar plates with CFT073 expressing or not expressing PafR. After incubation at 30°C for 12 h, the diameter of the colony, which correlates with bacterial motility, was measured. The results in Fig. 4C and the example in Fig. 4D clearly show that expression of PafR in CFT073 dramatically decreases motility. Taken together, our results show that PafR affects CFT073 pathogenicity in general and the ability of this strain to be motile and to form biofilm in particular. DISCUSSION

This study reports the identification of a novel transcriptional regulator found exclusively in ExPEC and primarily in UPEC strains and demonstrates that it contributes to colonization of the urinary tract. We termed this regulator PafR (patho-film regulator), since our results imply that this predicted transcriptional


antiterminator, capable of functioning as such (Fig. 3A), affects expression of an array of virulence genes in general and inhibits biofilm formation in particular. We suggest that PafR functions together with PafP, a putative PTS maltose permease encoded by the neighboring gene. Our study clearly shows the importance of pafRP for UPEC virulence in the urinary tract, since their knockout from the UPEC chromosome (⌬pafRP) reduced the ability of UPEC to infect mice in an independent challenge and in a cochallenge experiment with wild-type UPEC. Virulence capacity could be restored by expressing the pafRP genes in trans. Our results are compatible with the results of a previous study that investigated the contribution of the different genomic islands in CFT073 to pathogenesis (10). The metV island, which carries the paf genes, was the only island whose deletion prevented colonization of the urinary tract of CBA/J mice during an independent challenge. Cochallenge experiments identified the region that contains the putative paf operon (c3405-c3409) in the metV island as significantly associated with pathogenesis. Our results further demonstrate the importance of each of the first two paf genes for pathogenesis and for bacterial fitness in the urinary tract, since the UPEC constructs deleted for the individual genes, the ⌬pafR and ⌬pafP strains, were outcompeted by CFT073 in a cochallenge experiment. Our extensive efforts to identify the environmental stimuli that induce expression of the paf genes, which included testing many conditions that prevail during virulence and specifically during UTI as well as conducting a high-throughput screening of a wide

iai.asm.org 4249

Baum et al.

range of carbon sources, led us to conclude that by and large, the promoter that precedes the pafR gene is not expressed under routine laboratory growth conditions. Even growth in artificial urine supplemented with an excess of maltose was shown to be insufficient to induce pafR expression from the UPEC chromosome, as shown by real-time PCR. The mild expression detected in the presence of some carbon sources, including maltose, can supposedly be attributed to pafP, which follows pafR, which encodes a predicted PTS maltose permease. PafR and PafP are likely to compose a Bgl-like sensory system (15) that might be involved in sensing and consumption of carbohydrates that are available in the urinary tract, thus facilitating colonization. This idea is supported by the many genes that are involved in carbohydrate metabolism, whose expression has been affected by PafR expression in the microarray analysis. Possible reasons for observing only mild induction in vitro are that the carbon sources that we tested are similar but not identical to the natural inducer(s) of this system or that the induction of this system is multifactorial, with maltose/maltotriose being just one component of the signal. Another possible explanation for our inability to identify conditions for significant expression of the paf genes in vitro is that they are ivi (in vivoinduced) genes, i.e., their expression is specifically induced in host tissues (33). We favor this explanation for the following reasons. (i) Lack of paf gene expression in vitro is supported by the high similarity between the expression profiles of wild-type and ⌬pafRP strains obtained in the microarray analysis, i.e., both strains exhibited no expression of the paf genes and comparable differences from the PafR-expressing strain. (ii) Expression of pafRP in the host is indicated by the reduced ability of bacteria deleted for either one of these genes or both to cause UTI in mice and the restoration of infection ability by expressing pafRP in trans. (iii) Upregulation of pafR and pafP expression in vivo (19.5and 12-fold, respectively) has been demonstrated by sequence analysis of RNA that was isolated directly from the urine of women with UPEC-mediated UTI, compared to UPEC cultured in urine (H. Mobley, unpublished results). Apparently, expression of these genes is required for host colonization and/or infection but not for growth in the flask. The linkage of PafR to maltose is further supported by our findings that PafR downregulates expression of the entire maltose regulon (Fig. 5A) and prevents growth of CFT073 in minimal medium containing maltose as a sole carbon source (Fig. 5B). The finding that PafR-mediated growth inhibition is specific for UPEC and does not occur in the nonpathogenic E. coli K-12 strain suggests that this is a pathogenesis-related phenotype. Previous studies in different bacterial species already linked maltose and the mal regulon to bacterial virulence. In Vibrio cholerae, the mal regulon is required for full virulence and maltose inhibits production and secretion of major virulence factors, including cholera toxin (40). In enteropathogenic E. coli (EPEC), the cytopathic effect caused by the virulence factor EspB depends on the presence of LamB, an outer membrane maltoporin (41). LamB in enteroinvasive E. coli (EIEC) and its homologue Omp48 in Aeromonas veronii are critical for bacterial adhesion to epithelial cells and seem to mediate adhesion to host cells (42, 43). In group A Streptococcus, mutations in malT and malE impaired the capacity to grow in human saliva, to metabolize ␣-glucans, and to infect the mouse oropharynx (44, 45), whereas in Actinobacillus pleuropneumoniae, MalT and LamB play a role in persistence and in pathogenesis of the respiratory tract (46). Taken together, the linkage between maltose and bac-

4250

iai.asm.org

terial virulence, although not well understood yet, suggests that the connection of PafR to mal operon expression and maltose on the one hand and that to pathogenesis on the other hand are interdependent. Besides genes of the mal regulon, our microarray analysis identified several groups of pathogenesis-related genes whose expression is affected by PafR. First, 26 genes related to adherence to host cells, required for successful infection by bacterial pathogens (47), were upregulated by PafR. To attach to specific niches within the urinary tract, UPEC expresses multiple adhesins, e.g., P pili, encoded by the pap gene cluster, and type 1 fimbriae, encoded by the fim gene cluster. Evidently, a number of genes from both clusters and many additional adhesins were upregulated in response to PafR expression (see Data Set S1 in the supplemental material). Another virulence-associated group of genes that was upregulated by PafR consists of the hemolysin genes, essential for the formation of type 1 secretion machinery that secretes the hemolysin toxin. Notably, the list of virulence-associated genes that were upregulated upon PafR expression includes many genes that were upregulated in CFT073 isolated from the urine of infected mice (48), including type 1 fimbriae, capsule, and LPS genes, reinforcing the idea that the paf genes play a role in UPEC colonization, growth, and survival in the urinary tract. Taken together, PafR emerges as a global regulator of virulence factors, reminiscent of RfaH, which enhances expression of various virulence factors, and its loss attenuates virulence of UPEC strain 536 in mice (49). An additional PafR-upregulated group that stands out is that of the biofilm formation genes (Fig. 4A), associated with host cell attachment (50). This group contains genes that are involved in controlling the c-di-GMP concentration (51), in generating the biofilm-inhibiting signal indole (52, 53), and in forming the extracellular amyloid curli (54), as well as toxin-antitoxin genes that influence biofilm formation and fitness (55). The effect of PafR on expression of biofilm-related genes gains further significance when considered together with our finding that PafR directly inhibits formation of biofilms. Upregulation of biofilm-inhibiting genes and various regulators of biofilms is in accord with the negative effect of PafR. The fact that expression the gene encoding the main curli subunit, csgA, has not been upregulated in our microarray analysis is also in line with the inhibiting effect of PafR on biofilm formation. Inhibition of both biofilm formation and motility by PafR (Fig. 4B to D) might seem contradictory at first, since biofilm cells are nonmotile, so one may expect biofilm inhibition to be due to an increase in bacterial motility. However, motility is a prerequisite for biofilm formation, since it is essential for the first steps in biofilm formation, as well as biofilm dispersal. To attach to a surface and move along it in order to initiate a biofilm, bacteria found in fluid need to overcome hydrodynamic and repulsive electrostatic forces of the surface itself. To increase the probability of interacting with the surface, bacteria use passive and active motility mechanisms (56). It has been shown that motility per se, rather than chemotaxis or flagellum-mediated attachment, is required for biofilm formation (57, 58). Moreover, the ability to form biofilms has been found to be directly linked to bacterial swimming ability (59). Notably, PafR-induced inhibition of motility is not due to repression of flhDC, the master regulator of flagellar synthesis (60), or of the other flagellar and chemotaxis genes, since our microarray results show that their expression profiles are comparable in the presence and absence of PafR in the cells (see



Fig. S7 in the supplemental material). Rather, various biofilmregulating genes exert their effect by regulating motility, e.g., genes controlling c-di-GMP levels. Hence, biofilm inhibition in PafRexpressing cells is at least partly due to an inhibition of motility that decreases the probability that the bacteria will stably interact with the surface. We suggest that PafR regulates the delicate balance between motility and biofilm formation, thus orchestrating the sequence of events that leads to successful infection. While bacterial virulence certainly depends on bacterial adhesins, toxins, and other virulence factors, bacterial metabolism during infection is of great importance, since it contributes to persistence. The changes in carbon utilization genes observed upon PafR expression suggest that the paf system might be involved in sensing and consumption of nutrients that are available in the urinary tract, thus facilitating colonization. Among the genes that were significantly influenced by PafR are those encoding the tricarboxylic acid (TCA) cycle, which converts succinate to fumarate. Our microarray results show that the complete sdh operon, which codes for the succinate dehydrogenase enzyme of the TCA cycle, was considerably upregulated in the PafR-expressing strain relative to results for the wild type (see Fig. S8 in the supplemental material). This finding agrees with findings of a previous study showing that a UPEC strain deleted for the sdhB gene shows reduced fitness in vivo (61). Moreover, genes coding for six out of the eight enzymes that compose the TCA cycle were upregulated in PafR-expressing cells (see Fig. S8). Taken together with the many downregulated electron transport genes, including metal ion binding/iron-sulfur cluster binding in the PafR-expressing strain (see Data Set S2), the contribution of PafR to bacterial colonization and fitness might involve favoring the TCA cycle over electron transfer for energy gaining, under conditions of increased carbohydrate metabolism. Finally, many of the genes that were upregulated in the PafR-expressing cells are transcriptional regulators. Obviously, different niches in the human body require different gene expression patterns (62) that apparently rely on different sensors and regulators that control environmental adaptation (63). Therefore, it is not surprising that many regulators, such as those that belong to two-component sensory systems, have been shown to control expression of virulence genes, e.g., QseBC, which regulates virulence of UPEC and EHEC by regulating central metabolic circuits, including carbon metabolism (64). Significantly, the short list of UPEC-specific genes contains a relatively high number of genes encoding predicted transcriptional regulators (9), reinforcing the important role of these factors in urovirulence-specialized mechanisms. ACKNOWLEDGMENTS We thank Anat Nussbaum-Shochat and Michael Gluhoded for technical help. We appreciate helpful discussions with past and present members of the O.A.-C. lab. This research was supported by FIRST grant no. 1450/04, awarded to O.A.-C. by the Israel Science Foundation (ISF), by grant no. 3-3974/4174, awarded to O.A.-C. by the Chief Scientist Office of the Ministry of Health, Israel, and by grant no. 2009232, awarded to O.A.-C. and H.L.M. by the United States-Israel Binational Science Foundation (BSF).

REFERENCES 1. Wiles TJ, Kulesus RR, Mulvey MA. 2008. Origins and virulence mechanisms of uropathogenic Escherichia coli. Exp. Mol. Pathol. 85:11–19. http://dx.doi.org/10.1016/j.yexmp.2008.03.007. 2. Foxman B. 2010. The epidemiology of urinary tract infection. Nat. Rev. Urol. 7:653– 660. http://dx.doi.org/10.1038/nrurol.2010.190.


3. Mobley H, Warren J. 1996. Urinary tract infections: molecular pathogenesis and clinical management. ASM Press, Washington, DC. 4. Mobley HL, Green DM, Trifillis AL, Johnson DE, Chippendale GR, Lockatell CV, Jones BD, Warren JW. 1990. Pyelonephritogenic Escherichia coli and killing of cultured human renal proximal tubular epithelial cells: role of hemolysin in some strains. Infect. Immun. 58:1281–1289. 5. Johnson JR, Stell AL. 2000. Extended virulence genotypes of Escherichia coli strains from patients with urosepsis in relation to phylogeny and host compromise. J. Infect. Dis. 181:261–272. http://dx.doi.org/10.1086 /315217. 6. Johnson JR, Stell AL, Delavari P, Murray AC, Kuskowski M, Gaastra W. 2001. Phylogenetic and pathotypic similarities between Escherichia coli isolates from urinary tract infections in dogs and extraintestinal infections in humans. J. Infect. Dis. 183:897–906. http://dx.doi.org/10.1086/319263. 7. Kao JS, Stucker DM, Warren JW, Mobley HL. 1997. Pathogenicity island sequences of pyelonephritogenic Escherichia coli CFT073 are associated with virulent uropathogenic strains. Infect. Immun. 65:2812–2820. 8. Welch RA, Burland V, Plunkett G, III, Redford P, Roesch P, Rasko D, Buckles EL, Liou SR, Boutin A, Hackett J, Stroud D, Mayhew GF, Rose DJ, Zhou S, Schwartz DC, Perna NT, Mobley HL, Donnenberg MS, Blattner FR. 2002. Extensive mosaic structure revealed by the complete genome sequence of uropathogenic Escherichia coli. Proc. Natl. Acad. Sci. U. S. A. 99:17020 –17024. http://dx.doi.org/10.1073/pnas.252529799. 9. Lloyd AL, Rasko DA, Mobley HL. 2007. Defining genomic islands and uropathogen-specific genes in uropathogenic Escherichia coli. J. Bacteriol. 189:3532–3546. http://dx.doi.org/10.1128/JB.01744-06. 10. Lloyd AL, Henderson TA, Vigil PD, Mobley HL. 2009. Genomic islands of uropathogenic Escherichia coli contribute to virulence. J. Bacteriol. 191: 3469 –3481. http://dx.doi.org/10.1128/JB.01717-08. 11. Guyer DM, Kao JS, Mobley HL. 1998. Genomic analysis of a pathogenicity island in uropathogenic Escherichia coli CFT073: distribution of homologous sequences among isolates from patients with pyelonephritis, cystitis, and catheter-associated bacteriuria and from fecal samples. Infect. Immun. 66:4411– 4417. 12. Houman F, Diaz-Torres MR, Wright A. 1990. Transcriptional antitermination in the bgl operon of E. coli is modulated by a specific RNA binding protein. Cell 62:1153–1163. http://dx.doi.org/10.1016/0092-8674 (90)90392-R. 13. van Tilbeurgh H, Declerck N. 2001. Structural insights into the regulation of bacterial signalling proteins containing PRDs. Curr. Opin. Struct. Biol. 11:685– 693. http://dx.doi.org/10.1016/S0959-440X(01)00267-6. 14. Mahadevan S, Wright A. 1987. A bacterial gene involved in transcription antitermination: regulation at a rho-independent terminator in the bgl operon of E. coli. Cell 50:485– 494. http://dx.doi.org/10.1016/0092 -8674(87)90502-2. 15. Amster-Choder O. 2005. The bgl sensory system: a transmembrane signaling pathway controlling transcriptional antitermination. Curr. Opin. Microbiol. 8:127–134. http://dx.doi.org/10.1016/j.mib.2005.02.014. 16. Datsenko KA, Wanner BL. 2000. One-step inactivation of chromosomal genes in Escherichia coli K-12 using PCR products. Proc. Natl. Acad. Sci. U. S. A. 97:6640 – 6645. http://dx.doi.org/10.1073/pnas.120163297. 17. Mahadevan S, Reynolds AE, Wright A. 1987. Positive and negative regulation of the bgl operon in Escherichia coli. J. Bacteriol. 169:2570 – 2578. 18. Miller JH. 1972. Experiments in molecular genetics. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY. 19. Brooks T, Keevil CW. 1997. A simple artificial urine for the growth of urinary pathogens. Lett. Appl. Microbiol. 24:203–206. http://dx.doi.org /10.1046/j.1472-765X.1997.00378.x. 20. Lane MC, Alteri CJ, Smith SN, Mobley HL. 2007. Expression of flagella is coincident with uropathogenic Escherichia coli ascension to the upper urinary tract. Proc. Natl. Acad. Sci. U. S. A. 104:16669 –16674. http://dx .doi.org/10.1073/pnas.0607898104. 21. Friedberg D, Umanski T, Fang Y, Rosenshine I. 1999. Hierarchy in the expression of the locus of enterocyte effacement genes of enteropathogenic Escherichia coli. Mol. Microbiol. 34:941–952. http://dx.doi.org/10 .1046/j.1365-2958.1999.01655.x. 22. Schlosser-Silverman E, Elgrably-Weiss M, Rosenshine I, Kohen R, Altuvia S. 2000. Characterization of Escherichia coli DNA lesions generated within J774 macrophages. J. Bacteriol. 182:5225–5230. http://dx.doi .org/10.1128/JB.182.18.5225-5230.2000. 23. Palaniyandi S, Mitra A, Herren CD, Lockatell CV, Johnson DE, Zhu X, Mukhopadhyay S. 2012. BarA-UvrY two-component system regulates

iai.asm.org 4251

Baum et al.

24. 25.

26.

27.

28. 29.

30.

31. 32. 33. 34. 35. 36.

37. 38. 39.

40. 41.

42. 43.

44.

virulence of uropathogenic E. coli CFT073. PLoS One 7:e31348. http://dx .doi.org/10.1371/journal.pone.0031348. Hitchcock PJ, Brown TM. 1983. Morphological heterogeneity among Salmonella lipopolysaccharide chemotypes in silver-stained polyacrylamide gels. J. Bacteriol. 154:269 –277. O’Toole GA, Kolter R. 1998. Initiation of biofilm formation in Pseudomonas fluorescens WCS365 proceeds via multiple, convergent signalling pathways: a genetic analysis. Mol. Microbiol. 28:449 – 461. http://dx.doi .org/10.1046/j.1365-2958.1998.00797.x. Lane MC, Lockatell V, Monterosso G, Lamphier D, Weinert J, Hebel JR, Johnson DE, Mobley HL. 2005. Role of motility in the colonization of uropathogenic Escherichia coli in the urinary tract. Infect. Immun. 73: 7644 –7656. http://dx.doi.org/10.1128/IAI.73.11.7644-7656.2005. Cooper LA, Simmons LA, Mobley HL. 2012. Involvement of mismatch repair in the reciprocal control of motility and adherence of uropathogenic Escherichia coli. Infect. Immun. 80:1969 –1979. http://dx.doi.org/10 .1128/IAI.00043-12. Collins TJ. 2007. ImageJ for microscopy. Biotechniques 43:25–30. http: //dx.doi.org/10.2144/000112505. Nussbaum-Shochat A, Amster-Choder O. 1999. BglG, the transcriptional antiterminator of the bgl system, interacts with the beta’ subunit of the Escherichia coli RNA polymerase. Proc. Natl. Acad. Sci. U. S. A. 96: 4336 – 4341. http://dx.doi.org/10.1073/pnas.96.8.4336. Hagan EC, Mobley HL. 2009. Haem acquisition is facilitated by a novel receptor Hma and required by uropathogenic Escherichia coli for kidney infection. Mol. Microbiol. 71:79 –91. http://dx.doi.org/10.1111/j.1365 -2958.2008.06509.x. Quan JA, Schneider BL, Paulsen IT, Yamada M, Kredich NM, Saier MH, Jr. 2002. Regulation of carbon utilization by sulfur availability in Escherichia coli and Salmonella typhimurium. Microbiology 148:123–131. Bochner B. 1993. Advances in the identification of bacteria and yeast. Am. Clin. Lab. 12:6. Mahan MJ, Slauch JM, Mekalanos JJ. 1993. Selection of bacterial virulence genes that are specifically induced in host tissues. Science 259:686 – 688. http://dx.doi.org/10.1126/science.8430319. Thomas S, Holland IB, Schmitt L. 2014. The type 1 secretion pathway— the hemolysin system and beyond. Biochim. Biophys. Acta 1843:1629 – 1641. http://dx.doi.org/10.1016/j.bbamcr.2013.09.017. Roncero C, Casadaban MJ. 1992. Genetic analysis of the genes involved in synthesis of the lipopolysaccharide core in Escherichia coli K-12: three operons in the rfa locus. J. Bacteriol. 174:3250 –3260. Garofalo CK, Hooton TM, Martin SM, Stamm WE, Palermo JJ, Gordon JI, Hultgren SJ. 2007. Escherichia coli from urine of female patients with urinary tract infections is competent for intracellular bacterial community formation. Infect. Immun. 75:52– 60. http://dx.doi.org/10.1128 /IAI.01123-06. Karatan E, Watnick P. 2009. Signals, regulatory networks, and materials that build and break bacterial biofilms. Microbiol. Mol. Biol. Rev. 73:310 – 347. http://dx.doi.org/10.1128/MMBR.00041-08. Beloin C, Roux A, Ghigo JM. 2008. Escherichia coli biofilms. Curr. Top. Microbiol. Immunol. 322:249 –289. Spurbeck RR, Tarrien RJ, Mobley HL. 2012. Enzymatically active and inactive phosphodiesterases and diguanylate cyclases are involved in regulation of motility or sessility in Escherichia coli CFT073. mBio 3(5): e00307-12. http://dx.doi.org/10.1128/mbio.00307-12. Lang H, Jonson G, Holmgren J, Palva ET. 1994. The maltose regulon of Vibrio cholerae affects production and secretion of virulence factors. Infect. Immun. 62:4781– 4788. Kumar SS, Sankaran K, Haigh R, Williams PH, Balakrishnan A. 2001. Cytopathic effects of outer-membrane preparations of enteropathogenic Escherichia coli and co-expression of maltoporin with secretory virulence factor, EspB. J. Med. Microbiol. 50:602– 612. Valkonen KH, Veijola J, Dagberg B, Uhlin BE. 1991. Binding of basement-membrane laminin by Escherichia coli. Mol. Microbiol. 5:2133– 2141. http://dx.doi.org/10.1111/j.1365-2958.1991.tb02143.x. Vazquez-Juarez RC, Romero MJ, Ascencio F. 2004. Adhesive properties of a LamB-like outer-membrane protein and its contribution to Aeromonas veronii adhesion. J. Appl. Microbiol. 96:700 –708. http://dx.doi.org/10 .1111/j.1365-2672.2004.02177.x. Shelburne SA, Davenport MT, Keith DB, Musser JM. 2008. The role of complex carbohydrate catabolism in the pathogenesis of invasive

4252

iai.asm.org

45.

46.

47. 48.

49.

50.

51. 52. 53.

54. 55.

56. 57. 58.

59. 60.

61. 62. 63. 64.

streptococci. Trends Microbiol. 16:318 –325. http://dx.doi.org/10 .1016/j.tim.2008.04.002. Shelburne SA, III, Keith DB, Davenport MT, Horstmann N, Brennan RG, Musser JM. 2008. Molecular characterization of group A Streptococcus maltodextrin catabolism and its role in pharyngitis. Mol. Microbiol. 69:436 – 452. http://dx.doi.org/10.1111/j.1365-2958.2008.06290.x. Lone AG, Deslandes V, Nash JH, Jacques M, MacInnes JI. 2009. malT knockout mutation invokes a stringent type gene-expression profile in Actinobacillus pleuropneumoniae in bronchoalveolar fluid. BMC Microbiol. 9:195. http://dx.doi.org/10.1186/1471-2180-9-195. Pizarro-Cerda J, Cossart P. 2006. Bacterial adhesion and entry into host cells. Cell 124:715–727. http://dx.doi.org/10.1016/j.cell.2006.02.012. Snyder JA, Haugen BJ, Buckles EL, Lockatell CV, Johnson DE, Donnenberg MS, Welch RA, Mobley HL. 2004. Transcriptome of uropathogenic Escherichia coli during urinary tract infection. Infect. Immun. 72: 6373– 6381. http://dx.doi.org/10.1128/IAI.72.11.6373-6381.2004. Nagy G, Dobrindt U, Schneider G, Khan AS, Hacker J, Emody L. 2002. Loss of regulatory protein RfaH attenuates virulence of uropathogenic Escherichia coli. Infect. Immun. 70:4406 – 4413. http://dx.doi.org/10.1128 /IAI.70.8.4406-4413.2002. Jain A, Gupta Y, Agrawal R, Khare P, Jain SK. 2007. Biofilms–a microbial life perspective: a critical review. Crit. Rev. Ther. Drug Carrier Syst. 24:393– 443. http://dx.doi.org/10.1615/CritRevTherDrugCarrierSyst.v24 .i5.10. Povolotsky TL, Hengge R. 2012. ‘Life-style’ control networks in Escherichia coli: signaling by the second messenger c-di-GMP. J. Biotechnol. 160:10 –16. http://dx.doi.org/10.1016/j.jbiotec.2011.12.024. Wood TK. 2009. Insights on Escherichia coli biofilm formation and inhibition from whole-transcriptome profiling. Environ. Microbiol. 11:1–15. http://dx.doi.org/10.1111/j.1462-2920.2008.01768.x. Zhang XS, Garcia-Contreras R, Wood TK. 2007. YcfR (BhsA) influences Escherichia coli biofilm formation through stress response and surface hydrophobicity. J. Bacteriol. 189:3051–3062. http://dx.doi.org/10.1128/JB .01832-06. Barnhart MM, Chapman MR. 2006. Curli biogenesis and function. Annu. Rev. Microbiol. 60:131–147. http://dx.doi.org/10.1146/annurev .micro.60.080805.142106. Kim Y, Wang X, Ma Q, Zhang XS, Wood TK. 2009. Toxin-antitoxin systems in Escherichia coli influence biofilm formation through YjgK (TabA) and fimbriae. J. Bacteriol. 191:1258 –1267. http://dx.doi.org/10 .1128/JB.01465-08. Donlan RM. 2002. Biofilms: microbial life on surfaces. Emerg. Infect. Dis. 8:881– 890. http://dx.doi.org/10.3201/eid0809.020063. Pratt LA, Kolter R. 1998. Genetic analysis of Escherichia coli biofilm formation: roles of flagella, motility, chemotaxis and type I pili. Mol. Microbiol. 30:285–293. http://dx.doi.org/10.1046/j.1365-2958.1998.01061.x. Genevaux P, Muller S, Bauda P. 1996. A rapid screening procedure to identify mini-Tn10 insertion mutants of Escherichia coli K-12 with altered adhesion properties. FEMS Microbiol. Lett. 142:27–30. http://dx.doi.org /10.1111/j.1574-6968.1996.tb08402.x. Wood TK, Gonzalez Barrios AF, Herzberg M, Lee J. 2006. Motility influences biofilm architecture in Escherichia coli. Appl. Microbiol. Biotechnol. 72:361–367. http://dx.doi.org/10.1007/s00253-005-0263-8. Gauger EJ, Leatham MP, Mercado-Lubo R, Laux DC, Conway T, Cohen PS. 2007. Role of motility and the flhDC operon in Escherichia coli MG1655 colonization of the mouse intestine. Infect. Immun. 75:3315– 3324. http://dx.doi.org/10.1128/IAI.00052-07. Alteri CJ, Smith SN, Mobley HL. 2009. Fitness of Escherichia coli during urinary tract infection requires gluconeogenesis and the TCA cycle. PLoS Pathog. 5:e1000448. http://dx.doi.org/10.1371/journal.ppat.1000448. Eisenreich W, Dandekar T, Heesemann J, Goebel W. 2010. Carbon metabolism of intracellular bacterial pathogens and possible links to virulence. Nat. Rev. Microbiol. 8:401– 412. http://dx.doi.org/10.1038/nrmicro2351. Stock AM, Robinson VL, Goudreau PN. 2000. Two-component signal transduction. Annu. Rev. Biochem. 69:183–215. http://dx.doi.org/10 .1146/annurev.biochem.69.1.183. Hadjifrangiskou M, Kostakioti M, Chen SL, Henderson JP, Greene SE, Hultgren SJ. 2011. A central metabolic circuit controlled by QseC in pathogenic Escherichia coli. Mol. Microbiol. 80:1516 –1529. http://dx.doi .org/10.1111/j.1365-2958.2011.07660.x.
