William R Schwan, Nicole L Flohr, Abigail R Multerer, Jordan C Starkey

Abstract

Key words:Escherichia coli; Flagella; GadE; Motility; Acid response; fliC

INTRODUCTION

In the United States, approximately 10.5 million women suffer from a urinary tract infection each year. Around 80% of urinary tract infection are caused by uropathogenicEscherichia coli(UPEC), resulting in over 100000 hospitalizations and an approximate cost of $ 3.5 billion per year[1-3]. UPEC sometimes ascend all of the way to the kidneys, causing life-threatening pyelonephritis in some of the women[2,3]. The ability ofEscherichia coli(E. coli) to move up the human urinary tract is due to the presence of flagella expressed by the bacteria[4-7].

Bacterial flagella allow the directional movement ofE. colibased upon a chemotactic response[8,9]. Several genes are involved in the expression of flagella,althoughfliCencodes the flagellin subunits that comprise the bulk of a flagellum structure[10]. Several studies have shown the importance of flagella in UPEC pathogenesis[4-7,11]. For instance, several studies have examined the prevalence of thefliCgene in UPEC strains. One study showed the prevalence of thefliCgene in UPEC strains varied from 84% (community-acquired) to 95% (nosocomial-acquired)[12],whereas another study reported that only 16% of the UPEC strains had thefliCgene[13]. Part of the disparity in the frequency offliCgene prevalence could be due to the respective primers used in each study. Certainly, UPEC flagella are critical for ascension out of the bladder into the kidneys of an animal host. Within a mouse or human urinary tract, UPEC are continuously bathed in urine. Typically, human and murine urine will have a slightly acidic pH and variations in osmolality[14-16], although the osmolality within murine urine is usually higher than human urine[15]. Hence, pH is one critical environmental factor found in the urinary tract.

WithinE. coli, homeostasis in an acidic environment is mediated by at least five acid response (AR) systems[17-21]. System two (AR2) is induced in stationary phase and requires a glutamate decarboxylase and a glutamate: γ-aminobutyric acid antiporter.AR2 is the predominant and best characterized of the five AR system pathways[22-25].The AR2 requires the antiporter GadC and two inducible glutamate decarboxylases:GadA and GadB. The antiporter is responsible for transporting glutamate into the cell while transporting the product of glutamate decarboxylation, glutamate: γaminobutyric acid, out of the cell[22,24-30]. GadE, belonging to the LuxR family of regulatory proteins[31], has been identified as the central transcriptional activator ofgadA/BC, and provides the primary means ofgadA/BCactivation[32,33]. Microarray studies done under acidic conditions originally identified theyhiEgene (renamedgadE), which was found to encode for this transcriptional regulator protein[31]. GadE binds to a 20-bp sequence (GAD box: 5’-TTAGGATTTTGTTATTTAAA-3’) located -63 bp from the transcriptional start site of bothgadAand thegadBCoperon and is necessary for expression of these genes under all conditions[28,34,35].

In this study, we have studied the role GadE may play inE. coliflagella expression.Through the use of agadEmutant, afliC-lacZreporter system, and a motility assay; we demonstrate that GadE regulates transcription offliCinE. coli, which in turn affects bacterial motility.

MATERIALS AND METHODS

Bacterial strains, plasmids, and media

All of the bacterial strains and plasmids used in this study are listed in Table 1.E. colistrain NU149 is a clinical isolate obtained from a patient with cystitis[36]. TheE. colistrain DH5α was used to construct thefliC-lacZreporter system.E. colistrains MC4100(supplied by Linda Kenney) and EK227 (supplied by John Foster) were subsequently tested under various pH and osmotic conditions with thefliC-lacZreporter system.The ΔgadEstrain EF1007 and ΔgadE/pPCRScript AmpgadEstrain EF1083 were also supplied by John Foster. Multicopy plasmid pUJ9[37]and single copy plasmid pPP2-6[38]were used for cloning. The pUJ9 plasmid contains a promoterlesslacZgene and an ampicillin antibiotic resistance gene. Plasmid pPP2-6 is a single copy plasmid with a multiple cloning site that possesses a chloramphenicol resistance gene[38]. The pPCRScript AmpgadEplasmid had thegadEgene cloned into the multicopy plasmid pPCRScript Amp[33]. Luria Agar (LA) supplemented with 12.5 μg/mL chloramphenicol was used to grow the recombinantE. colicells containing the reporter system. Luria Bertani (LB) broth containing 1% glycerol at pH’s ranging from 5.5 to 8.0 was used to test pH ranges, and LB broth (pH 5.5 and pH 7.0, 1% glycerol, 0.1 mol/L Na3PO4buffering) coupled with osmotic variation of 0 to 400 mmol/L NaCl was used to gauge pH plus osmotic changes[38]. Under these growth conditions, the recombinantE. colistrains were assayed for β-galactosidase activity.

Construction of the fliC-lacZ fusion

Oligonucleotide primers FliC1 (5’-GAGAGAATTCGATGAAATACTTGCCATGC-3’)and FliC2 (5’-AGAAGGATCCAGACGCTGGATAGAACTC-3’) specific for a 397-bp segment of theE. colistrain NU149fliCpromoter were amplified with theBamHI andEcoRI restriction endonuclease sites flanking the DNA promoter sequence.Polymerase chain reaction (PCR) amplification using these primers was set up as follows: An initial denaturation of five minutes, then 35 cycles 1 min at 94 °C, 1 min at 55 °C, and 1 min at 72 °C, finishing with a 7 min elongation at 72 °C after the 35th cycle. Chromosomal DNA fromE. colistrain NU149 extracted with a PurElute Bacterial Genomic kit (Edge Biosystems, Mountain View, CA, United States) served as the template in the PCR. The 397-bp product was visualized on a 0.8% agarose gel containing ethidium bromide with a 1 kb ladder (New England Biolabs, Ipswich, MA,United States) served as the molecular weight standard.

The PCR amplified 397-bpfliCpromoter DNA fragment was passed through a Microcon 30 filter (MilliporeSigma, Burlington, VT, United States) to concentrate the DNA. Subsequently, the DNA was digested with the restriction endonucleasesEcoRI andBamHI (New England Biolabs). The digested DNA fragment was ligated toEcoRI/BamHI digested pUJ9 plasmid DNA and transformed into competent DH5α cells. The resulting transformants were selected on LA containing 100 μg/mL ampicillin and X-Gal (Promega, Madison, WI, United States). Blue colonies were screened for β-galactosidase[39]and the plasmid DNA was extracted with a QIAPrep kit (Qiagen, Valencia, CA, United States) to verify the appropriate size. One recombinant plasmid, pNK1-1, was carried further in the process. This plasmid DNA was digested with the restriction endonucleaseNotI (New England Biolabs) and ligated toNotI cut pPP2-6 DNA. Following ligation, the DNA was transformed into DH5α and clones were selected on LA containing 12.5 μg/mL chloramphenicol and X-Gal. One clone, pNK2-29, was selected forin vitroanalysis.

Galactosidase assays

Galactosidase assays were performed on DH5α/pNK2-29 and MC4100/pPP2-6 cells grown in LB media at various pH and in the presence and absence of NaCl at pH 5.5 and 7.0[39]. Bacteria were grown mid-logarithmically and β-galactosidase activity on the sodium dodecyl sulfate and CHCl3permeabilized cells. The mean values +standard deviation was calculated from at least three separate experiments for each bacterial strain.

Creation of a ΔgadE mutation in uropathogenic E. coli strain NU149

To create a deletion mutation of thegadEgene, the red recombinase system described by Datsenko and Wanner[40]was used. Briefly, the primer pair GadE1(5’GATGACATATTCGAAACGATAACGGCTAAGGAGCAAGTTTGTGTAGGCTGG AGCTGCTTCG-3’) and GadE2 (5’TCGTCATGCCAGCCATGAATTTCAGTTGCTTATGTCCTGACATATGAATATCCTCCTTAG-3’) was used to create a PCR product, using pKD4 plasmid DNA as a template. The PCR conditions that were used were an initial denaturation at 95 °C for 5 min followed by 35 cycles of 95 °C, 1 min; 57°C, 1 min; and 72 °C, 2 min. The resulting PCR product was concentrated and separated on a 0.8% agarose gel, cut out, and the DNA extracted from the agarose gel.With this purified PCR product, an electroporation was performed on strain NU149/pKD46 cells as described previously[40], selecting for transformants on LA with 40 μg/mL kanamycin. One transformant, NU149gadE, was chosen for further analysis. To remove the kanamycin resistance gene, plasmid pCP20 was introduced into NU149gadEby electroporation. The resulting strain was processed as noted previously[6]. To confirm thegadEdeletion, a PCR-based assay was used with the GadE5 (5’-ACAGGGCTTTTGGCAGTTGAA-3’) and GadE6 (5’-AAATATTAGCGTCGACGTGA-3’) primers. The PCR conditions that were used were an initial denaturation at 95 °C for 5 min followed by 30 cycles of 95 °C, 1 min; 57 °C, 1 min; and 72 °C, 2 min. This ΔgadEmutation was complemented by electroporating the pPCRScript AmpgadEplasmid into NU149gadEand selecting for transformants on LA with 100 μg/mL ampicillin. The wild-type NU149 strain was used a positive control andStaphylococcus aureusgenomic DNA was used as a negative control.

Table 1 Bacterial strains and plasmids used in the study

Construction of ΔlacZ mutation in uropathogenic E. coli strain NU149

To construct the ΔlacZmutation in UPEC strain NU149, the procedure described above was used. The LacZ1 (5’-CCTTACGCGAAATACGGGCAGACATGGCCTGCCCGGTTAT

TACATATGAATATCCTCCTTAG-3’) and LacZ2 (5’-TGGAATTGTGAGCGGATAACAA

TTTCACACAGGAAACAGCTTGTGTAGGCTGGAGCTGCTTCG-3’) primer pair were used to create the PCR product using the amplification conditions noted above.To confirm the ΔlacZmutation, the LacZ3 (5’-ATGAAACGCCGAGTTAACGC-3’) and LacZ4 (5’-AGCTGGCGTAATAGCGAAGA-3’) primers were used in the PCR amplification conditions described above. Plasmid pNK2-29 was electroporated into strain NU149 and colonies were selected on MacConkey containing 12.5 mg/mL chloramphenicol.

Soft agar assay for motility

A soft agar motility test was performed as previously described[41]for the wild-typevs gadEmutant and complemented mutant analysis. Each strain was inoculated into the center of the agar plate and the amount of bacterial spread measured after 24 h postinoculation. The motility assays were repeated two more times on separate days.

Statistical analyses

A two-tailed Student’st-test was used to calculate statistical variation with aP＜ 0.05 considered significant.

RESULTS

Examination of the fliC::lacZ fusion at different pHs

To assess whether pH affected the transcription of ourfliC-lacZfusion plasmid, the pH of buffered LB medium was adjusted to 5.5 to 8.0 by using 0.1 M Na3PO4buffering and glycerol to maintain the pH[38]. The resulting media were inoculated with MC4100/pNK2-29 and theβ-galactosidase activities of mid-logarithmic-phase cells were determined. The optimal pH forfliCexpression was found to be at pH 7.0 (1111 Miller units; Table 2). As the pH shifted to the acidic range,fliCtranscription declined until there was a significant 3.9-fold difference observed comparingfliCtranscription at pH 7.0 compared to pH 5.5 (288 Miller units,P＜ 0.01). When the pH of the buffered LB was raised into the alkaline range, there was a slight decline infliCtranscription that was 1.5-fold lower at pH 8.0 (738 Miller units,P＜ 0.05)vsgrowth in pH 7.0 medium. These results indicate that pH alone affectsfliCtranscription.

Effects of pH and osmotic conditions together on fliC::lacZ transcription

In an environment such as the the human or murine urinary tract, fluctuations in both pH and osmolarity can occur[14-16]. To determine if the combination of acidic pH and high osmolarity affectfliCtranscription, MC4100/pNK2-29 was grown in buffered pH with variation in both the pH (5.5 and 7.0) and the osmolarity (0 to 400 mmol/L NaCl). When MC4100/pNK2-29 was grown in pH 7.0/low-osmolarity (0 mmol/L NaCl) LB,fliCtranscription was the highest (1,132 Miller units, Table 3). An increase in the osmolarity to 400 mmol/L NaCl in the pH 7.0 LB causedfliCtranscription to significantly fall by 2.5-fold (454 Miller units,P＜ 0.01) compared to growth in the pH 7.0 low-osmolarity LB.E. coliwith the pNK2-29 plasmid grown in pH 5.5/lowosmolarity conditions displayedfliCtranscription of 308 Miller units (Table 3);however,fliCtranscription dropped almost 5-fold to 62 Miller Units (P＜ 0.01) as the osmolarity increased to 400 mmol/L NaCl. A comparison offliCtranscription inE.coligrown in pH 7.0/low-osmolarity LB to theE. colipopulation grown in pH 5.5/high-osmolarity LB showed a highly significant 18.2-fold change (P＜ 0.001).Thus, a growth environment possessing both an acidic pH and high osmolarity substantially repressedfliCtranscription in theE. coliK-12 strain.

To determine if the samefliCtranscriptional changes occurred in a UPEC strain, a ΔlacZmutation was created in UPEC strain NU149. The pNK2-29 plasmid containing thefliC-lacZfusion was moved intoE. colistrain NU149 LacZ1 and the same environmental conditions tested for theE. coliK-12 strain were used. Growth of NU149 LacZ1/pNK2-29 in pH 7.0 with no added NaCl displayed the highestfliCtranscription (1353 Miller Units, Table 3), whereasfliCtranscription significantly fell 3.06-fold when the strain was grown in pH 5.5 LB (442 Miller Units,P＜ 0.01). An increase in the osmolarity to 400 mM NaCl in pH 7.0 LB causedfliCtranscription to fall 2.77-fold (489 Miller Units,P＜ 0.01). Moreover, the growth of NU149 LacZ1/pNK2-29 in pH 5.5 LB with 400 mM added NaCl showed the lowest level of fliC transcription (147 Miller Units) that was 9.2-fold lower than when grown in pH 7.0 no added salt medium (P＜ 0.01). Overall, thefliCtranscription results in the UPEC strain mirrored theE. coliK-12 strain’s results.

Transcription of fliC was affected by the gadE mutation in E. coli grown in acidic pH media

As shown above, acidic pH growth conditions led to lowerfliCtranscription compared to transcription in neutral pH growth conditions. Previous work has shown that the glutamate decarboxylase system is critical for acid resistance inE. coliand GadE is an important regulator of this AR system[31-33]. We then asked whether GadE might also regulatefliCtranscription under acidic growth conditions. We examined anE. coliK-12 wild-type strain, agadEmutant strain as well as a complementedgadEmutant strain all of which contained thefliC-lacZpNK2-29 plasmid. The strains were grown in buffered LB set at pH 5.5 or 7.0 with (400 mmol/L) or without (0 mmol/L)added NaCl and monitored for galactosidase activity. Derepression offliCtranscription occurred in thegadEmutant grown in acidic pH LB (Table 4). After growth in pH 5.5/low-osmolarity medium, thegadEmutant strain (1742 Miller units)exhibited a 3.2-fold increase infliCtranscription, compared to the wild-type strain(540 Miller units,P＜ 0.001), which indicated that GadE repressedfliCunder acidic conditions. Complementation with an intactgadEgene reduced the activity below the wild-type levels to 295 Miller units, below even wild-type levels, confirming the repressive effect of GadE onfliCexpression. The repressive effect of GadE on fliC expression was reduced in pH 7.0/low-osmolarity medium with thegadEmutant strain showing only slightly higherfliCtranscription (2196 Miller units)vsthegadE+wild-type strain (1520 Miller units,P＜ 0.01). However, when the growth conditions were changed to a high osmolarity environment (400 mmol/L NaCl), thegadEmutation had no significant effect onfliCtranscription (540 Miller units). A change to a pH 5.5/ high-osmolarity environment caused a further repression offliCtranscription (165 Miller units,P＜ 0.05) that was significant.

Table 2 Effect of pH on fliC::lacZ gene transcription in Escherichia coli strain MC4100/pNK2-29 grown in buffered Luria Bertani media

A gadE mutation affects uropathogenic E. coli motility

The data above suggested that GadE may repressfliCtranscription whenE. coliis grown under acidic pH conditions. Since transcriptional differences do not always translate into protein level differences or functional differences, the effects of agadEmutation onE. colimotility was next tested. First, motility was tested using theE. coliK-12 strain EF227 (wild-type), EK1007 (gadEmutation), and EF1083 (gadEmutation complemented with the pPCRScript AmpgadEplasmid). All strains were grown in pH 5.5 buffered LB and spotted onto motility agar plates. Wild-typeE. colistrain EF227 displayed an 8.33 mm spread diameter, whereas strain EF1007 showed a significantly larger spread diameter of 45 mm (P＜ 0.001, Table 5). When thegadEmutation was complemented in strain 1083, the spread diameter dropped below the wild-type level (6.67 mm diameter).

AgadEmutation was also created in the uropathogenicE. coliclinical isolate NU149 using a λred recombinase system. The NU149, NU149gadE, and NU149gadE/pPCRScript AmpgadEstrains were grown in pH 5.5 buffered LB and spotted onto motility agar plates. Wild-typeE. colistrain NU149 had a 10.67 mm spread diameter, whereas strain NU149gadEhad a 57.34 mm spread diameter that was significantly wider (P＜ 0.05). Complementation of thegadEmutation brought the spread diameter back down to a wild-type level (7.00 mm). These results indicate that GadE also affects UPEC motility.

DISCUSSION

The production of flagella in UPEC is vital for their pathogenesis in a human host,enabling the bacteria to ascend the urinary tract[4-7,11]. A transcriptome study of a UPEC strain in the murine urinary tract over time demonstrated that several genes that are involved in flagella biosynthesis and chemotaxis, including thefliCstructural gene, had their transcription down-regulated in this environment[42]. Within the urinary tract, theE. coliencounter an environment that typically has a slightly acidic pH and osmotic changes that increase as the bacteria move into the kidneys of the host[14-16].E. coliis able to survive in acidic pH environments that include the human and murine urinary tracts because of AR systems that include the glutamate decarboxylase system[15-18]. GadE is an important protein that regulates this AR system[31-33]. Since GadE is important for regulating genes in one AR system, could the GadE regulator of the glutamate decarboxylase AR system also be involved in the down-regulation offliCin uropathogenicE. coligrowing in the murine urinary tract?

To answer the question above, we designed afliC-lacZreporter system on a single copy number plasmid to measurefliCtranscription withinE. coligrowing in various environments that might be encountered in the urinary tract. Our results showedfliCtranscription fell in bothE. colistrains grown in a pH 5.5 environment compared to a neutral pH environment, suggesting one or more proteins produced byE. coligrowing in an acidic pH environment repressesfliCtranscription. A previous study revealed a substantial drop in motility byE. coligrown in an acidic environmentvsa neutral pH environment[43]that correlates with our experimental observations in this study. Moreover,E. coligrowth in a high salt concentration medium also caused repression offliCtranscription. Liet al[44]observed thatE. coligrown in a high-osmolarity medium were less motile compared toE. coligrown in a low-osmolarity medium.

Table 3 Effect of osmolarity on fliC::lacZ gene transcription in Escherichia coli grown in buffered pH 5.5 and 7.0 Luria Bertani media with different osmolarities

A combination of pH changes and osmolarity changes was also examined using ourfliC-lacZsystem. In a low pH/high-osmolarity medium, the growingE. coliexhibited an additive level of repression offliCtranscription that is in line with the previous transcriptome study[42].

Two environmental variables are at play in a low pH/high-osmolarity environment. To adapt to acidic pH conditions,E. colirely on AR systems and their corresponding regulators, such as GadE. On the other hand, the OmpR-EnvZ twocomponent system is the main osmotic stress regulatory system inE. coli[45]. OmpR has been shown to regulate flagella expression[46,47]and is likely partially responsible for repressingfliCtranscription in the high-osmolarity environment that we tested.Furthermore, OmpR-regulated genes are tied to the acid response inE. coliand

Salmonella enterica[48,49].

Since GadE is a central player in AR system regulation, we examinedfliCtranscription and motility ingadEmutant strainsvsthe wild-type strains. By deleting thegadEgene,E. coli fliCtranscription was derepressed, particularly inE. coligrowing in an acidic pH environment. Complementation of thegadEmutation with thegadEgene on a multicopy plasmid caused additional suppression offliCtranscription that was below wild-type levels. Furthermore, a ΔgadEmutation in K-12 and UPEC strains led to significantly greater motility compared to the wild-type strain. Together, these data suggest that GadE repressesfliCtranscription either by directly binding to thefliCpromoter to repress transcription or acting in an indirect manner by influencing expression of FlhD that in turn regulatesfliC[50,51]. However, GadE does not appear to affect osmotic control offliCtranscription.

What would be the advantage of a loss of flagella expression inE. coligrowing in the human kidney? Flagella protruding from the surface ofE. colicells represent a target of the host’s immune system. FlagellatedE. colicells are more likely to be phagocytized than no-flagellated cells[52].E. colithat have reached the kidneys would be in a low pH/high-osmolarity environment where the flagella are no longer needed and may in fact be a detriment to their survival. Through the regulatory effects of the GadE and OmpR proteins,fliCtranscription may be shut down, causing the bacterial cells to lose their flagella and be able to hide behind their anti-phagocytic capsules.

Table 4 Assessing a gadE and mutation and complementation on fliC::lacZ gene transcription in Escherichia coli grown in buffered pH 5.5 and 7.0 Luria Bertani media with different osmolarities

Table 5 Motility of Escherichia coli strains NU149 and EK227, their gadE mutants, and complemented gadE mutants grown in pH 5.5 Luria Bertani

ARTICLE HIGHLIGHTS

Research background

UropathogenicEscherichia coli(UPEC) is the number one cause of urinary tract infection in women. Motility driven by the action of flagella is critical for UPEC pathogenesis. HowEscherichia coli(E. coli) adapts to a low pH/high osmolarity environment is essential for the species survival. Acid tolerance systems, such as the System two system, are important for UPEC survival in a low pH environment.

Research motivation

Our key problem to be solved was whether GadE, a part of the acid response two system,regulates transcription of thefliCgene, and in turn, UPEC motility.

Research objectives

Determine whether GadE regulatedfliCtranscription and subsequent motility of theE. coli.

Research methods

We created afliC-lacZreporter system on a single-copy number plasmid and measured bgalactosidase levels in both a K-12 and UPEC clinical isolate. Furthermore, motility was assessed in bothE. colistrains by inoculating wild-type,gadEmutant, and complementedgadEmutant strains onto motility agar.

Research results

Transcription offliCwas significantly lower inE. coligrown in pH 5.5 Luria Bertani compared to pH 7.0 Luria Bertani. A mutation in thegadEgene led to higherfliCexpression in that strainvswild-type bacteria. Motility was significantly higher in thegadEmutant strain compared to the wild-type strain.

Research conclusions

We confirmed thatfliCtranscription was down-regulated inE. coli grown in a low pH/high osmolarity environment compared to a neutral pH/low osmolarity environment. GadE appears to either directly or indirectly regulatefliCtranscription inE. coli.

Research perspectives

Future work could be done to affirm the GadE regulation of flagella expression inE. coli.

ACKNOWLEDGEMENTS

The authors wish to thank John Foster for a critical reading of the manuscript. We wish to thank Linda Kenney (MC4100) and John Foster (EK227, EF1007, EF1083) for strains used in this study.