Regulation of the Swarming Inhibitor disA in Proteus mirabilis

Abstract

The disA gene encodes a putative amino acid decarboxylase that inhibits swarming in Proteus mirabilis. 5′ rapid amplification of cDNA ends (RACE) and deletion analysis were used to identify the disA promoter. The use of a disA-lacZ fusion indicated that FlhD₄C₂, the class I flagellar master regulator, did not have a role in disA regulation. The putative product of DisA, phenethylamine, was able to inhibit disA expression, indicating that a negative regulatory feedback loop was present. Transposon mutagenesis was used to identify regulators of disA and revealed that umoB (igaA) was a negative regulator of disA. Our data demonstrate that the regulation of disA by UmoB is mediated through the Rcs phosphorelay.

INTRODUCTION

Proteus mirabilis is a Gram-negative bacillus and a causative agent of urinary tract infections in patients with abnormal urethras or requiring long-term catheterization (1, 2, 3). P. mirabilis is also known for its ability to swarm, a form of flagellum-mediated surface motility (4, 5). In liquid medium, P. mirabilis exists as a vegetative, peritrichously flagellated swimming cell. However, 3 to 4 h after being placed on a solid surface, the vegetative swimming cells differentiate into elongated, multinucleate, aseptate, hyperflagellated swarmer cells, as reviewed in reference 6. The swarmer cells aggregate to form multicellular rafts and move concentrically away from the central inoculum for approximately 1 to 2 h before dedifferentiating back to vegetative swimming cells (7). This cycle of differentiation and consolidation gives P. mirabilis its characteristic bull's-eye appearance on agar plates (4).

Flagellar biogenesis is tightly controlled in P. mirabilis through a hierarchically tiered regulatory cascade consisting of class I, II, and III gene clusters (reviewed in reference 8). Class I consists solely of the flagellar master regulator flhDC. The FlhD₄C₂ heterohexamer is the master swarming regulator and activates transcription of class II genes (9, 10, 11). Class II is comprised of genes needed to form the hook-basal body structure of the flagella as well as fliA, encoding the swarming sigma factor σ²⁸, and flgM, the corresponding anti-sigma factor. σ²⁸ is responsible for transcribing class III genes, including genes involved in chemotaxis, and the structural genes of the flagellar filament and motor. The energy expenditure to fully flagellate a swarmer cell and the cyclic aspect of swarming require that swarming be a tightly regulated process. Several signals inducing differentiation, such as the inhibition of flagellar rotation, accumulation of putrescine, and O-antigen contact with a solid surface, have been identified; however, the signals responsible for consolidation are poorly understood (12, 13, 14, 15).

A novel regulator of swarming, disA, that bears homology to aromatic amino acid decarboxylases was discovered by Stevenson et al. (16). Disruption of the disA gene resulted in a hyperswarming phenotype, whereas overexpression completely abolished swarming (16). Currently, the mechanism by which DisA inhibits swarming is unknown, but it inhibits FlhD₄C₂ at the posttranslational level, possibly by interfering with multimer formation (16). In addition, the mechanism of DisA-mediated inhibition is conserved in other Gram-negative enterics, in which DisA overexpression inhibited motility (17). The actual biochemical function of DisA is currently unknown, and metabolomic analysis of both the disA mutant and overexpressing strains did not reveal significant changes in the cellular levels of decarboxylated amino acids. In addition, the use of purified DisA and all possible amino acids did not reveal any products. However, the strong homology of DisA to tyrosine/phenylalanine decarboxylases, together with the fact that phenethylamine, but not tyramine, mimics the effect of DisA overexpression, suggests that DisA is a phenylalanine decarboxylase.

This study further defines the disA locus by identifying the transcriptional start site and begins the process of elucidating the regulation of disA. 5′ rapid amplification of cDNA ends (RACE) analysis and transcriptional lacZ fusions demonstrate that disA transcription begins at a thymine residue 70 bp upstream of the DisA start codon. Use of a disA-lacZ fusion demonstrated that FlhD₄C₂ does not have a significant role in disA expression. Transposon mutagenesis was used to identify UmoB as a negative regulator of disA. The umoB gene product is a negative regulator of the Rcs phosphorelay and has been previously implicated by our lab and others in swarming regulation (15, 18, 19, 20, 21, 22, 23, 24, 25). Our data indicate that the effect of the umoB mutation on disA expression is dependent upon the Rcs phosphorelay system. Taken together, our data indicate that a complex network is responsible for regulation of disA, allowing the cell more-precise control over the energy-intensive process of swarming.

MATERIALS AND METHODS

Bacterial growth conditions.

The bacterial strains and plasmids utilized are listed in Table 1. P. mirabilis and Escherichia coli were grown in Luria-Bertani (LB) broth (10 g tryptone, 5 g yeast extract, 5 g sodium chloride per liter) at 37°C with shaking at 250 rpm. For plate growth, E. coli and nonswarming P. mirabilis strains were grown on 1.5% agar; swarming strains of P. mirabilis were plated on 3% agar to inhibit motility. Concentrations of antibiotics for selection for E. coli were as follows: 25 μg/ml for streptomycin and chloramphenicol, 20 μg/ml for kanamycin, and 100 μg/ml for ampicillin. Concentrations of antibiotics for selection for P. mirabilis were as follows: 35 μg/ml streptomycin, 100 μg/ml chloramphenicol, 300 μg/ml ampicillin, 20 μg/ml kanamycin, and 15 μg/ml tetracycline. A total of 12 μg/ml 5-bromo-4-chloro-indolyl-β-d-galactopyranoside (X-Gal) was used to observe blue and white colonies unless otherwise stated.

Table 1.

Strains and plasmids

Strain or plasmid	Genotype or characteristicsa	Source or reference
Strains
E. coli
XL1	endA1 gyrA96(Nalr) thi-1 recA1 relA1 lac glnV44 F′ hsdR17(rΚ− mΚ+)	Laboratory stock
CC118	araD139 Δ(ara-leu)7697 ΔlacZ74 phoAΔ20 galE galK thi rpsE rpoB argE(Amp) recA1 λpir	31
SM10 λpir	thi thr leu tonA supE recA RP4-2Tc::Mu Kanr λpir	32
EC100D	TransforMax EC100D pir+ electrocompetent E. coli: F− mcrA Δ(mrr-hsdRMS-mcrBC) ϕ80dlacZΔM15 ΔlacX74 recA1 endA1 araD139 Δ(ara-leu)7697 galU galK λ− rpsL (Strr) nupG pir+ (DHFR)	Epicentre
DH5α	F− ϕ80dlacZΔM15 Δ(lacZYA-argF)U169 endA1 recA1 hsdR17(rK− mK−) deoR thi-1 supE44 λ− gyrA96 relA1	Laboratory stock
P. mirabilis
PM7002	Wild type; Tetr	ATCC
PM2199	PM7002 disA::mini-Tn5lacZ; Kanr	Laboratory stock
BB1	PM7002 umoB::mini-Tn5-Kanr/pQF50 + PdisA −1206 to +39-lacZ	This study
BB2	PM7002 umoB::mini-Tn5-Kanr rcsB::Strr/pQF50 + PdisA −1206 to +39-lacZ	This study
BB3	PM7002 umoB::mini-Tn5-Kanr rcsC::Strr/pQF50 + PdisA −1206 to +39-lacZ	This study
BB4	PM7002 flhC mutant/pQF50 + PdisA −1206 to +39-lacZ	This study
BB5	PM7002 rcsB::Strr/pQF50 + PdisA −1206 to +39-lacZ	This study
BB6	PM7002 rcsC::Strr/pQF50 + PdisA −1206 to +39-lacZ	This study
Plasmids
pQF50	Low copy number; Ampr	28
pACYC184	Low copy number; Chlr	33
pFDCH1	pACYC184 + flhDC	34
pKNG101	R6K-derived suicide vector; Strr	26
pBB1	pQF50 + PdisA −1206 to +39-lacZ	This study
pBB2	pQF50 + PdisA −69 to −1-lacZ	This study
pBB3	pQF50 + PdisA −1206 to +39(Δ-10)-lacZ	This study
pUmoB	pACYC184 + umoB	Laboratory stock
pBB4	pKNG101 + flhC mutant (frameshift mutation)	This study
pUBK	pKNG101 + umoB::Kanr	15
pRcsB	pKNG101 + rcsB::Strr	15
pRcsC	pKNG101 + rcsC (internal fragment)	15
pDisA	pKNG101 + disA (internal fragment)	Laboratory stock
pMDA	pACYC184 + disA	Laboratory stock

^aNal^r, nalidixic acid resistant; Kan, kanamycin; Str, streptomycin; Tet, tetracycline; Amp, ampicillin; Chl, chloramphenicol.

Construction of plasmids and strains.

All plasmids were introduced by electroporation into PM7002 as follows. PM7002 was grown in 30 ml of LB to an optical density at 600 nm (OD₆₀₀) of 0.4 to 0.6, and cells were harvested by centrifugation at 3,000 × g for 5 min at 4°C. Pellets were washed and resuspended in cold 10% glycerol with a final resuspension volume of 60 μl per electroporation. Cells were electroporated in cold cuvettes (Bio-Rad Gene Pulser cuvette; 0.2-cm diameter) using a Bio-Rad MicroPulser electroporator set to the E2 value for bacterial electroporation. Cells were recovered from the cuvette using 200 μl of prewarmed LB broth and incubated with shaking at 37°C for a minimum of 3 h for all plasmids harboring ampicillin resistance and a minimum of 1 h for all other antibiotics. Cells were plated on LB with appropriate antibiotic and agar concentrations and incubated at 37°C to allow growth of transformants.

For PCR amplification of selected genes, Phusion Hot Start II high-fidelity DNA polymerase (Thermo Scientific) was used. Validation of vector constructs and identification of transposon insertion sites were done through Beckman Coulter QuickLane sequencing according to the recommendations of the manufacturer. All primers are listed in Table 2.

Table 2.

Primers

Name	Sequence (5′–3′)	Use
GSP1	GGATGACGTGCAATCGCCATCGGCAG	5′ RACE
GSP2	GCTGATAATGTTTTTAAC	5′ RACE
GSP3	CCAGCAGCTAATGAATAA	5′ RACE
disPro Full Fwd	CCCAAGCGTCGACCGTAAAATAAACTCAATTCTGATTAAAATTGATAACAAAAATTTATATATGGATCCTTGGC	PdisA−69-−1::lacZ
disPro Full Rev	GCCAAGGATCCATATATAAATTTTTGTTATCAATTTTAATCAGAATTGAGTTTATTTTACGGTCGACGCTTGGG	PdisA−69-−1::lacZ
PdisA Fwd	ATCAAGGATCCATGAAGATATCGCTTTACCG	PdisA−1206-+39::lacZ
PdisA Rev	ATCAAGCGTCGACCTGCAGCACTCAGACAGG	PdisA−1206-+39::lacZ
dis-10Fwd	GATTCTTACTCATCATGTAGCGCCGGCCTTAAAAAGAGACTAATTATT	PdisA−1206-+39Δ-10::lacZ
dis-10Rev	AATAATTAGTCTCTTTTTAAGGCCGGCGCTACATGATGAGTAAGAATC	PdisA−1206-+39Δ-10::lacZ

To construct pBB1, primers PdisA Fwd and PdisA Rev were used to amplify the 1,245-bp region upstream of disA. The product was ligated into the SmaI site of pQF50, verified, and used to transform PM7002. Plasmid pBB2 was constructed by annealing 1 μg disPro Full Fwd and disPro Full Rev at 72°C for 10 min. The product was digested with SalI and BamHI and ligated into pQF50 digested with the same enzymes. The plasmid was validated by sequencing before transformation into PM7002. Plasmid pBB3 was constructed by performing site-directed mutagenesis on pBB1 using the primers dis-10Fwd and dis-10Rev. The QuikChangeII site-directed mutagenesis kit (Agilent Technologies) was used to mutate the eight base pairs at the putative −10 region upstream (TATATCAT to CGCCGGCC) as described in the user manual with the following modifications: 54 ng of pBB1 isolated from XL1 was used as a double-stranded DNA (dsDNA) template. Eighteen rounds of amplification were performed at 68°C with 1 min of extension per 500 bp (16.5 min). A total of 1 μl of 10 μM dis-10Fwd and dis-10Rev primers was used in the reactions. EC100D was transformed with 1 μl of the mutagenesis product and plated on 1.5% LB agar plates containing 100 μg/ml ampicillin. Plasmid pBB4 was constructed by digesting pBC plus flhDC with EcoNI, which digests once in the middle of the flhC gene. The product was blunt ended with T4 polymerase (Roche) as described by Promega. The blunt-ended vector was ligated to itself and digested a second time with EcoNI to enrich the sample for plasmids that had been successfully religated in the first reaction. This product was transformed into XL1 and selected by plating on chloramphenicol. Sequencing validated the presence of the single base pair insertion in flhC that resulted from religation. pBC with the flhC mutant was digested with PvuII and SalI, and the excised product was ligated to the suicide vector pKNG101 digested by SmaI with SalI. The proper flhC mutation in pBB4 was validated by sequencing before transformation into PM7002.

Strain BB1 was obtained by transposon mutagenesis of PM7002/pBB1 as described in “Transposon mutagenesis.” BB4 was constructed by conjugating PM7002 with SM10 harboring pBB4. Exconjugants were plated on tetracycline and streptomycin to select for P. mirabilis containing a Campbell-type insertion of pBB4. A colony was grown in LB broth in the absence of antibiotics and plated on 10% sucrose to identify colonies that had successfully excised pKNG101. Recombinants containing the flhC mutant allele were validated by Southern blotting and transformed with pBB1. Strain BB1 was cured of pBB1 and then mated with SM10 containing pRcsB or pRcsC to obtain the umoB rcsB and umoB rcsC double mutants. Mutants were transformed with pBB1 to yield BB2 and BB3.

5′ RACE.

5′ RACE was performed on 5 μl PM7002 RNA according to the 5′ RACE system for rapid amplification of cDNA ends, version 2.0 (Invitrogen), methods section with a few modifications. cDNA was synthesized according to the alternative protocol for first-strand cDNA synthesis. Nested PCR was performed using the AUAP primer provided and GSP3 in 40 cycles with 30 s of annealing at 59°C and 45 s of extension using Phusion Hot Start II high-fidelity DNA polymerase. DNA was ligated to the EcoRV site of pBC and transformed into DH5α for blue/white screening. White colonies were cultured, and insertion of 5′ RACE products was verified by restriction digestion. Clones harboring inserts were sequenced using the universal T7 primer to identify potential transcriptional start sites. pBB2 and pBB3 were constructed based on the sequences of the cloned 5′ RACE products to determine the transcriptional start site.

Transposon mutagenesis.

Strain BB1 was transformed by electroporation with EZ-Tn5 <Kan-2> Tnp Transposome (Epicentre). Transformants were selected on 3% LB containing ampicillin, kanamycin, and X-Gal (60 μg/ml). Colonies with increased or decreased blue color were cultured as described above and assayed for β-galactosidase activity (see below). Southern blotting was performed to identify the segment of the chromosome where the transposon was inserted. This region was subcloned into pBC and sequenced using primers provided with the transposome to identify the specific site of insertion.

Construction of an umoB disruption by a Campbell-type insertion.

An umoB disruption was constructed in PM2199 disA::mini-Tn5lacZ by cloning a PCR-derived fragment internal to the umoB coding region obtained using the primers 5′-CGTCATCTAGAGCGGTAGAGATCCATATTCC and 5′-CGTCAGGATCCGGCCCTTGCTTGATAACATG into the suicide plasmid pKNG101 (26). The construct was then mobilized into PM2199 by a filter mating with E. coli SM10 containing the plasmid. Exconjugants were selected on LB plates containing 35 μg/ml streptomycin and 15 μg/ml tetracycline. The correct disruption of umoB was verified by Southern blotting.

β-Galactosidase assays.

Overnight cultures were grown, and optical densities (ODs) were normalized to the lowest density culture. A total of 200 μl of normalized culture was spread on a 1.5% LB plate and incubated at 37°C for 2 or 4 h. These time points were chosen to assess expression before (T2) and at the peak of (T4) swarming. Cells were collected from plates either by resuspension in 1 ml of fresh LB (two-hour harvest) or by washing one-quarter of the plate with 500 μl of fresh LB (four-hour harvest). The OD₆₀₀ was recorded, and a portion of the culture was pelleted and frozen at −20°C overnight. Pellets were lysed by chloroform/SDS treatment and assayed as previously described (27). Plates with 25 mM phenethylamine (PEA) were made using an appropriate volume of phenethylamine (Sigma-Aldrich catalog no. 241008 [50 ml] or Acros Organics catalog no. 156491000), and pH was adjusted to 7.

RESULTS

Identification of the disA promoter.

The 5′ end of the disA transcript was identified by 5′ RACE performed on total RNA harvested from wild-type PM7002. Sequencing of the 5′ RACE PCR products returned two potential transcriptional start sites, one located 8 bp upstream of the disA open reading frame (ORF) (Fig. 1A, open arrow, designated −8) and a second site located 70 bp upstream of the disA ORF (Fig. 1A, closed arrow, designated −70). To determine if active promoters were present upstream of these potential start sites, various fragments were cloned into pQF50 (28) to create transcriptional lacZ fusions. These constructs were transformed into PM7002 and assayed for β-galactosidase activity (Fig. 2).

Fig 1.

Construction of transcriptional lacZ fusions. (A) The region upstream of disA is shown with the noncoding region comprised of 511 bp, designated −511 to +1 with the disA ORF beginning at +1. The bent arrows indicate potential transcriptional start sites identified by 5′ RACE. The black arrow (−70) is located 70 bp upstream of the disA ORF. The open arrow (−8) is located eight base pairs upstream of the disA ORF. (B) Plasmid pBB1 contains a region extending from 39 bp into the disA gene to 1,206 bp upstream of the disA gene (including 695 bp from the PMI1208 gene) that is fused to a promoterless lacZ gene in pQF50. (C) Plasmid pBB2 contains the region from −69 to −8 that is fused to lacZ in pQF50. (D) Plasmid pBB3 is identical to pBB1 but contains an 8-bp substitution at the putative −10 region, changing the sequence from TATATCAT to CGCCGGCC (see Materials and Methods) (Table 1). (E) Sequence upstream of the disA open reading frame. The start codon is underlined, and the putative transcriptional start sites located 8 bp and 70 bp upstream of the ATG start codon are shown in bold font. The proposed −10 promoter element is shaded. Three Rcs binding sites with homology to the E. coli consensus sequence (TAAGAATAATCCTA) are underlined, and mismatched bases are noted with lighter font.

Fig 2.

Transcriptional activity of promoter fragments. 5′ RACE returned two potential transcriptional start sites, −70 and −8 (Fig. 1). Various promoter fragments were cloned upstream of a promoterless lacZ gene in pQF50 as shown in Fig. 1. LB agar plates were inoculated with overnight cultures of each strain that were adjusted to identical optical densities, and cells were harvested off plates at 2 and 4 h after plating. The data shown are representative of two independent experiments, with samples assayed in triplicate. An asterisk indicates a P value of <0.05.

Plasmid pBB1 contains a fragment extending from −1,206 to +39 relative to the ATG start of the disA gene (Table 1; Fig. 1B). This fragment contains both potential transcriptional start sites as well as part of the disA and PMI1208 open reading frames (Fig. 1B). The expression of β-galactosidase from pBB1 was 816-fold and 50-fold higher than in cells containing the pQF50 vector with no promoter inserted at 2 and 4 h after plating, respectively (Fig. 2). PM7002 containing plasmid pBB2 with the disA region from −69 to −1 exhibited no β-galactosidase activity, indicating the absence of a promoter upstream of the −8 transcriptional start site (Fig. 1C and 2). This indicated that the functional disA promoter region was upstream of the transcriptional start site originating 70 bp upstream of the disA ATG start codon. To verify this, site-directed mutagenesis was used to change base pairs at the −10 sequence from TATATCAT to CGCCGGCC. The resulting plasmid pBB3 contains these altered base pairs in the context of the full-length disA region present in pBB1. Plasmid pBB3 exhibited a 4.3-fold reduction and a 10.7-fold reduction in β-galactosidase activity 2 and 4 h after being plated on agar surfaces, respectively, compared to pBB1, indicating that altering nucleotides in the −10 region severely decreased the overall promoter activity.

FlhD₄C₂ does not regulate disA.

The class I activator FlhD₄C₂ has a central role in activating gene expression during swarming, and our lab previously demonstrated that disA expression increases during swarming (16). To address the role of FlhD₄C₂ in regulating disA expression, a null allele in flhC was constructed as described above (Materials and Methods), resulting in strain BB4. As expected, BB4 was unable to swarm (data not shown). The loss of flhC did not have a statistically significant effect on the expression of a disA-lacZ fusion (pBB1) when cells were assayed either 2 or 4 h after plating on agar surfaces (data not shown). Furthermore, when flhDC was overexpressed from a medium-copy-number plasmid, disA expression was not altered in a statistically significant manner (data not shown). These data indicate that FlhD₄C₂, the master regulator of swarming, does not have a role in regulating disA expression.

Role of phenethylamine and autoregulation in disA expression.

The predicted product of the DisA decarboxylase is phenethylamine (PEA), and previous work demonstrated that exogenous PEA inhibited swarming and flagellar gene expression in a manner similar to disA overexpression (16). We assessed the effect of PEA on disA expression in PM2199, containing a single-copy transcriptional disA-lacZ fusion generated by the insertion of mini-Tn5lacZ1 into the chromosomal copy of disA. This strain was used because the disA gene is inactivated, thereby reducing the intracellular levels of the putative product phenethylamine and allowing for a more sensitive assessment of the effects of exogenous phenethylamine. The presence of various concentrations of phenethylamine decreased disA expression in a dose-dependent manner at T4, with 3.1-fold repression seen at 25 mM, 2.7-fold repression at 16 mM, 1.7-fold repression at 8 mM, and 1.3-fold repression at 4 mM (Fig. 3A). The presence of phenethylamine had little effect in cells at T2 (Fig. 3). To determine if disA expression was subject to autoregulation, the disA gene was overexpressed in trans (Fig. 3B). The overexpression of disA in PM2199 decreased disA-lacZ expression 1.3-fold at 4 h after plating on agar but had no effect at 2 h.

Fig 3.

Role of autoregulation in disA expression. (A) The effect of phenethylamine on disA expression was assayed in PM2199 containing a single-copy disA-lacZ fusion generated by an insertion of mini-Tn5lacZ1 into the disA coding region. In this strain, the disA gene is inactivated by the transposon insertion. (B) disA-lacZ expression is measured in PM2199 cells containing either the vector pACYC184 or pACYC184 plus disA. Cells were harvested off plates as described for Fig. 2. Data shown are representative of two independent experiments, with samples assayed in triplicate. An asterisk indicates a P value of <0.05.

UmoB is a negative regulator of disA.

To identify potential regulators of disA, transposon mutagenesis was used to create random mutations in PM7002/pBB1 and colonies were screened on X-Gal plates for those with increased expression from the disA-lacZ fusion. This yielded an insertion in the umoB (igaA) gene encoding an integral membrane protein that has been shown in members of the Enterobacteriaceae to act as an inhibitor of the Rcs phosphorelay (reviewed in reference 29). The transposon insertion in umoB resulted in a 2.5- to 3-fold increase in disA expression (Fig. 4). The levels of disA expression were reduced to those of the wild type in the presence of a plasmid containing the cloned umoB gene (data not shown). An umoB mutation (umoB::Str) was also independently constructed in PM2199, in which the disA-lacZ fusion is in a single copy, and a similar 2.5-fold increase in expression was observed (Fig. 4).

Fig 4.

disA expression in umoB mutants. (A) The expression of disA-lacZ was monitored from plasmid pBB1 in wild-type PM7002 and an umoB::Kan mutant. (B) The effect of the umoB mutation on a single-copy disA-lacZ fusion was examined in PM2199 or the corresponding umoB::Str mutant. The expression of disA was measured by β-galactosidase expression in cells harvested 2 and 4 h after plating on agar plates as described for Fig. 2. Data shown are representative of two independent experiments, with samples assayed in triplicate. An asterisk indicates a P value of <0.05.

The umoB mutation alters disA expression via the Rcs phosphorelay.

The effect of UmoB on disA expression, as well as its established role as an inhibitor of the Rcs phosphorelay, led us to investigate if the Rcs phosphorelay was involved in regulating disA. Single mutations in rcsB and rcsC, the Rcs response regulator and sensor kinase, and the rcsB umoB and rcsC umoB double mutants were constructed and designated BB5, BB6, BB2, and BB3, respectively (Table 1). Activity of disA was measured in samples harvested 2 h and 4 h after plating by β-galactosidase assays, as described earlier. The data demonstrate that an rcsB or rcsC single mutant had no effect on disA expression; however, the umoB rcsB and umoB rcsC double mutants mitigated the effect of an umoB single mutant, returning disA expression to wild-type levels (Fig. 5).

Fig 5.

Regulation of disA via UmoB is through the Rcs phosphorelay. Expression of disA was measured by β-galactosidase expression from plasmid pBB1 in wild-type PM7002, an umoB::Kan mutant (BB1), an rcsB::Str mutant (BB5), an rcsC::Str mutant (BB6), and rcsB umoB and rcsC umoB double mutants (BB2 and BB3, respectively). Data shown are representative of at least two independent experiments, with samples assayed in triplicate. An asterisk indicates a P value of <0.05.

DISCUSSION

In this study, regulation of the disA locus was characterized by using both single-copy and plasmid-based transcriptional lacZ fusions, which allowed us to (i) identify the promoter region, (ii) determine that FlhD₄C₂ does not have a role in disA regulation, (iii) address the role of autoregulation via phenethylamine and disA overexpression, and (iv) identify a regulatory mutation in umoB that alters disA expression via the Rcs phosphorelay. Interestingly, the levels of β-galactosidase from the single-copy disA-lacZ fusion in PM2199 were higher than those from the disA-lacZ fusion in multicopy (pBB1) (compare Fig. 2 and 3). There are several possible explanations for this. First, the lacZ gene is translated from different ribosome binding sites (RBS) in each construct; the RBS in mini-Tn5lacZ is from the trp operon of E. coli, and in pBB1 (pQF50 vector), it is derived from the lpp gene. Second, there may be cis-acting regulatory sequences that are missing from the promoter region cloned into pBB1 that are required for full expression.

Mutations in rcsB and rcsC in conjunction with an umoB mutation demonstrated that the increased disA expression observed in an umoB mutant background was completely abrogated when the Rcs phosphorelay was nonfunctional. The observation that disA expression was not decreased in either an rcsB or rcsC single mutant was unexpected but may be due to the presence of the disA-lacZ fusion in pQF50, where the effect of the Rcs phosphorelay, and specifically the RcsB response regulator, on disA expression may be masked by the multicopy nature of the disA-lacZ fusion. The increased RcsB activity in the umoB mutant may be enough to still see regulatory changes in the multicopy disA-lacZ fusion. In support of this, recent data from our lab have identified disA as an RcsB-activated gene by RNA-Seq analysis (our unpublished data). The fact that RcsB activated FlhD₄C₂ yet FlhD₄C₂ did not have a role in disA regulation indicates that RcsB may directly bind disA to mediate regulation. In support of this, there are three possible RcsB binding sites upstream of the disA −10 region (Fig. 1E).

Previous research and sequence homology led to the prediction that phenethylamine, the decarboxylated form of phenylalanine, or a similar molecule was the product of DisA (16). We have provided evidence that a negative feedback loop is present for disA regulation. First, disA expression was decreased in PM2199 when disA was overexpressed (Fig. 3). Second, the addition of exogenous phenethylamine inhibited disA expression (Fig. 3). However, the magnitude of the repression differed under each condition, 3-fold with phenethylamine at 25 mM versus 1.4-fold with disA overexpressed. This may be due to differences in the intracellular levels of phenethylamine, with higher levels present during growth with 25 mM and potentially lower levels when disA is overexpressed. Consistent with this, a strong dose-dependent effect on disA repression was observed with phenethylamine concentrations ranging from 4 to 25 mM (Fig. 3).

One purpose of this negative feedback loop may be to downregulate disA expression after the peak levels of expression have been reached at 3 to 4 h into the swarming cycle. The subsequent decrease in DisA activity would then prepare cells for the next cycle of swarming by relieving the inhibition of FlhD₄C₂. It is also possible that extracellular phenethylamine encountered in the environment may have a role in regulating disA expression. It has been proposed that phenethylamine concentrations in the low millimolar range may be encountered in the intestinal tract and that this may influence swarming in the intestine (30). A similar mechanism for the control of disA expression may exist in the urinary tract. However, this is highly speculative because, to our knowledge, the levels of phenethylamine in the urinary tract are unknown. The ability to fine-tune disA expression by both a negative feedback loop and the Rcs phosphorelay would provide more-precise control of disA expression required for the transition between the swarming and consolidation phases.