Proteus mirabilis UreR coordinates cellular functions required for urease activity

ABSTRACT

A hallmark of Proteus mirabilis infection of the urinary tract is the formation of stones. The ability to induce urinary stone formation requires urease, a nickel metalloenzyme that hydrolyzes urea. This reaction produces ammonia as a byproduct, which can serve as a nitrogen source and weak base that raises the local pH. The resulting alkalinity induces the precipitation of ions to form stones. Transcriptional regulator UreR activates expression of urease genes in a urea-dependent manner. Thus, urease genes are highly expressed in the urinary tract where urea is abundant. Production of mature urease also requires the import of nickel into the cytoplasm and its incorporation into the urease apoenzyme. Urease accessory proteins primarily acquire nickel from one of two nickel transporters and facilitate incorporation of nickel to form mature urease. In this study, we performed a comprehensive RNA-seq to define the P. mirabilis urea-induced transcriptome as well as the UreR regulon. We identified UreR as the first defined regulator of nickel transport in P. mirabilis. We also offer evidence for the direct regulation of the Ynt nickel transporter by UreR. Using bioinformatics, we identified UreR-regulated urease loci in 15 Morganellaceae family species across three genera. Additionally, we located two mobilized UreR-regulated urease loci that also encode the ynt transporter, implying that UreR regulation of nickel transport is a conserved regulatory relationship. Our study demonstrates that UreR specifically regulates genes required to produce mature urease, an essential virulence factor for P. mirabilis uropathogenesis.

IMPORTANCE

Catheter-associated urinary tract infections (CAUTIs) account for over 40% of acute nosocomial infections in the USA and generate $340 million in healthcare costs annually. A major causative agent of CAUTIs is Proteus mirabilis, an understudied Gram-negative pathogen noted for its ability to form urinary stones via the activity of urease. Urease mutants cannot induce stones and are attenuated in a murine UTI model, indicating this enzyme is essential to P. mirabilis pathogenesis. Transcriptional regulation of urease genes by UreR is well established; here, we expand the UreR regulon to include regulation of nickel import, a function required to produce mature urease. Furthermore, we reflect on the role of urea catalysis in P. mirabilis metabolism and provide evidence for its importance.

INTRODUCTION

Catheter-associated urinary tract infections (CAUTIs) are the most prevalent healthcare-associated infections, representing up to 40% of nosocomial infections globally (1). Over 500,000 CAUTIs are reported annually in the U.S., at an estimated cost of up to $1.7 billion in healthcare costs each year (2, 3). Urinary catheter placement is common in general hospitals, where 15%–25% of patients is catheterized at some point during their stay (4, 5). Upon insertion into the bladder of a patient, a catheter is immediately available for colonization by bacteria and other microbes. Members of the gut microbiota that transiently colonize the periurethral area can access the bladder by traversing the abiotic catheter surface (4, 6). This route evades natural host defenses such as micturition and can minimize contact with host cells (7). Once in the bladder, bacteria can establish cystitis and occasionally ascend to the kidneys to cause acute pyelonephritis. In severe cases, pyelonephritis can seed bacteria into the bloodstream, resulting in bacteremia in up to 25% of cases and a risk of sepsis (8).

Although CAUTIs are frequently polymicrobial, Morganellaceae family member Proteus mirabilis remains a key pathogen of interest (7, 9). A Gram-negative facultative anaerobe, P. mirabilis, is commonly isolated from CAUTIs, often as a member of a polymicrobial infection, and can further complicate infections by inducing urinary stone formation (7, 10). This organism is notable for its ability to swarm across the surface of catheters and its potent use of urease as a virulence factor (11, 12). Urease, a nickel-requiring metalloenzyme, has long been implicated in P. mirabilis uropathogenesis (13, 14). After nickel incorporation into the active site of the UreC subunit, mature urease hydrolyzes urea and produces ammonia, a nitrogen source and weak base. The resulting alkalinity induces the precipitation of polyvalent ions that crystalize into struvite and apatite urinary stones (15, 16). Urease-null P. mirabilis mutants cannot induce stones and exhibit significant fitness defects in a murine UTI model, indicating urease is required for pathogenesis (13). Urease activity directly enables P. mirabilis to generate a unique niche in the urinary tract; bacteria are both encapsulated within the stone during its formation and colonize its surface (15, 17). Stones can complicate CAUTI treatment by sheltering P. mirabilis from the immune system and antibiotic treatment (16, 18). In addition to inducing urinary stone formation, urease activity is the leading cause of crystalline biofilm catheter blockages (11, 19, 20). Importantly, urinary stones and crystalline catheter blockages can serve as reservoirs to continually seed infection (21).

Urea-dependent genetic regulation of the urease operon (ureDABCEFG) is well established. Urease expression is activated by UreR, an AraC/XylS family transcriptional regulator (22–24). The ure operon is adjacent to its regulator and organized in a head-to-head orientation. Expression of both transcripts is driven by a shared bi-directional promoter (Fig. 1A). When urea is present, UreR directly binds this metabolite and forms a dimer. Activated UreR dimers then bind two sites within the shared promoter, activating transcription in both directions (25, 26). Urea is the primary nitrogenous waste molecule of mammals and is highly concentrated in human urine (~400 mM) (27, 28). Thus, urease genes are highly expressed in the urinary tract.

Fig 1.

UreR rapidly responds to urea by inducing expression of urease genes. (A) Schematic of the urease operon that illustrates the function encoded by each gene. ureR encodes a transcriptional activator (red) that responds to urea and drives expression of ureDABCEFG. ureABC encode the structural genes of urease (yellow), a metalloenzyme that hydrolyzes urea to form ammonia and carbon dioxide. ureDEFG encode accessory proteins (orange) that facilitate nickel incorporation into the urease active site. Letters within the shapes depicting the accessory proteins indicate the protein’s name (i.e., “D” signifies UreD). Inactive UreR monomer and urease apoenzyme are depicted in gray. (B and C) ureD expression was measured using quantitative reverse-transcription PCR (qRT-PCR). (B) ureD is rapidly induced in the presence of urea. HI4320 was cultured to mid-logarithmic phase and supplemented with urea. Samples were removed 15, 30, or 60 minutes after urea addition to quantify ureD expression via qRT-PCR. (C) ureC::bla exhibits urea-induced ureD expression. HI4320, ureC::bla, and ureR::aphA were cultured to mid-logarithmic phase and supplemented with urea. Samples were removed 15 minutes after urea addition to quantify ureD expression via qRT-PCR. Significance was determined by one-way Analysis of Variance (ANOVA) with Tukey’s multiple comparison correction, **P < 0.01. In both panels, data were normalized to time-matched controls incubated without urea and expression of housekeeping gene rpoA. Each bar represents the mean of three to six biological replicates, and error bars represent standard error of the mean.

After the expression of urease genes, the three structural subunits of the apoenzyme (UreA, UreB, and UreC) assemble to form a homo-trimer of heterotrimers [(UreABC)₃] (29). During maturation, two nickel ions are incorporated into each UreC subunit to activate the holoenzyme. This function is performed by urease accessory proteins UreD, UreE, UreF, and UreG in an energy-dependent manner (Fig. 1A) (30, 31). Nickel must be transported into the cell before its delivery to the intracellular urease. P. mirabilis encodes two ABC nickel transporters, Ynt and Nik, that perform this function. A recent study demonstrated that of the two, Ynt has a higher affinity for nickel and is preferentially deployed during experimental UTI (32). In contrast to other transition metal co-factors such as iron, manganese, magnesium, and zinc, few enzymes in nature require nickel (33, 34). Nickel is also capable of poisoning enzymes that require other metal co-factors, necessitating tight regulation of its import. This supervision is provided by NikR, Nur, and NmtR in other species (35–37). However, a BLASTP search against all P. mirabilis genomes in GenBank failed to identify homologs for these nickel regulators, implying that this species uses an alternative regulatory mechanism (38). In this study, we demonstrate that UreR induces expression of both nickel transporters and directly regulates expression of nickel transporter Ynt, thereby coordinating the transcriptional program that produces active urease.

MATERIALS AND METHODS

Bacterial strains and culture conditions

The bacterial strains and constructs used in this study are described in Table 1. P. mirabilis type strain HI4320 was isolated from a nursing home patient with a long-term indwelling catheter and is the prototypical strain used to study this organism (39, 40). Bacteria were routinely cultured at 37°C with aeration (200 RPM). Routine overnight culture was performed in Luria Broth (LB; per liter: 10 g tryptone, 5 g yeast extract, and 0.5 g NaCl) or on LB medium solidified with 1.5% agar. Some experiments were performed in Minimal A medium [per Liter: 10.5 g K₂HPO₄, 4.5 g KH₂PO₄, 0.47 g sodium citrate, and 1.0 g (NH₄)₂SO₄; autoclave to sterilize and add 1 mL of 1 M MgSO₄, 10 mL of 20% glycerol, and 1 mL of 1% nicotinic acid] (41). As needed, medium was supplemented with 120 mM HEPES for buffering or with antibiotics (kanamycin, 25 µg/mL; ampicillin, 100 µg/mL; chloramphenicol, 20 µg/mL; and tetracycline, 15 µg/mL).

TABLE 1.

Strains and plasmids used in this study

Strain	Species	Description	Selection	Reference
HI4320	P. mirabilis	Wild type		(39)
ureC::bla	P. mirabilis	HI4320 background	amp	(13)
ureR::bla	P. mirabilis	HI4320 background	kan	(24)
ureR::kanR	P. mirabilis	ureR targetron mutant with kanamycin marker	kan	This study
ureRΔkanR	P. mirabilis	ureR targetron mutant, unmarked		This study
TOP10	Escherichia coli	Strain for routine cloning and protein expression		Invitrogen

Strain construction

Construction of a ureR mutant in strain HI4320 was performed the using the TargeTron system (Sigma) as previously described (45, 46). Briefly, a mutated intron was amplified with primers designed to add end sequences homologous with the insert site within the target vector. Next, the intron was cloned into vector pACD4K-C-loxP (NEB HiFi Assembly Mix) and the sequence was confirmed via whole plasmid sequencing (Nanopore by Eurofins). This vector was then transformed into HI4320 containing the T7 helper plasmid pAR1219 (42). After T7 induction, the resulting colonies were selected on kanamycin agar and sequenced to confirm the intron insertion in the 5′ region of ureR. The cre recombinase plasmid pQL123 was used to generate a kanamycin-sensitive insertional mutant (43).

Isolation of RNA

Overnight cultures of bacterial strains were sub-cultured in duplicate in 3 mL HEPES-buffered LB and incubated at 37°C with agitation. After 2 hours, one culture of each strain was supplemented with urea to a final concentration of 50 mM while the other culture remained unperturbed. LB, pH 7, was buffered with 120 mM HEPES. We experimentally determined that this concentration was sufficient to maintain an endpoint pH of 7 when cultures were spiked with 50 mM urea. For time point selection: 15, 30, and 60 minutes post-urea addition, 1 mL of each culture was preserved in 2 mL of RNAprotect Bacteria Reagent (Qiagen) for RNA isolation. For RNA sequencing and qRT-PCR: 15 minutes post-urea addition, 1 mL of each culture was preserved in 2 mL of RNAprotect. After incubating for ~15 minutes at room temperature, samples were centrifuged, supernatants were removed, and pellets were frozen at −20°C for up to 2 weeks to facilitate cell lysis. RNA was isolated using an RNeasy Kit (Qiagen). Samples were incubated with lysozyme (0.5 mg/mL) for 15 minutes with agitation to facilitate cell lysis. Subsequent RNA isolation steps were performed according to the manufacturer’s instructions. Genomic DNA was depleted using the Turbo-DNAfree Kit (Invitrogen). RNA samples were confirmed as DNA free via PCR prior to further use (Table S1).

Library preparation and sequencing

Library preparation and sequencing were performed by the University of Michigan Advanced Genomics Core. RNA quality was assessed using the TapeStation (Agilent) and Qubit RNA broad-range assay (Thermo Fisher). The median RNA Integrity Number for samples was 8.3, with a range of 7.8–8.7 (Table S2). Samples were prepared using the NEBNext rRNA Depletion Kit (bacteria) (NEB E7850X), xGen Broad-range RNA Library Prep (IDT 10010145), and xGen Normalase UDI Primer Plate 1 (IDT 10009795). An input of 740 ng of total RNA was ribosomal depleted using the rRNA Depletion module. The rRNA-depleted RNA was then fragmented per IDT’s protocol and then reverse transcribed into cDNA using random primers. The 3′ ends of the cDNA were then adenylated, and adapters were ligated. The products were purified and enriched by PCR to create the final cDNA library. Final libraries were checked for quality and quantity by Qubit hsDNA (Thermo Fisher) and LabChip (PerkinElmer). The samples were pooled and sequenced on the Illumina NovaSeq S4 (paired-end 150 bp) according to the manufacturer’s recommended protocols.

RNA sequencing data analysis

Geneious Prime 2022.1 software was used for data analysis (www.geneious.com). Reads were paired as fastq files and were imported into Geneious Prime using the average insert size for each sample (Table S2). The BBduk plug-in (47) was used to remove Illumina adapter sequences and perform quality trimming (maq = 20, maxns = 1, trimq = 14, and minlength = 20). Trimmed reads were mapped to the HI4320 genome (accession: NC_010554.1) using the Geneious read mapper. The 36-kb plasmid carried by HI4320 was excluded from our analysis; this plasmid is not widely carried by P. mirabilis isolates and primarily encodes machinery for plasmid replication and conjugation (7). Geneious calculated expression levels of genes in each sample, and the DeSeq2 plug-in (48) was used to calculate differential expression (fit type = parametric).

qRT-PCR

After RNA isolation, cDNA was synthesized using the iScript cDNA Synthesis Kit (Bio-Rad) according to the manufacturer’s protocol. Each qRT-PCR reaction contained 2 µL of diluted cDNAs (1:2 dilution), 150 nM of each primer, and 12.5 µL of 2× Power SYBR Green Master Mix (Thermo Fisher). Gene expression relative to housekeeping gene rpoA was calculated using the 2^−ΔΔCt method. After each run, melt curve analysis was used to confirm a lack of off-target amplification or primer dimer. Prior to use, the efficiency of each qRT-PCR primer pair was assessed via standard curve. Primers used in this study are listed in Table S1.

Bacterial in vitro growth

Overnight cultures of bacterial strains were diluted 1:100 into either LB or Minimal A medium. For growth analysis, 300 µL of each prepared culture was added to a 100-well plate in triplicate. Uninoculated blanks for each media condition were also included. Plates were incubated in a Bioscreen C growth analyzer for 24 hours at 37°C with intermediate shaking, and the optical density at 600 nm (OD₆₀₀) of each well was measured every 15 minutes. Data for three biological replicates of each condition were collected on separate days. For evaluating nitrogen source availability, control cultures were incubated in complete Minimal A medium. Bacteria were also cultured in Minimal A lacking ammonium sulfate and supplemented with 2% L-serine, 2% adenosine, 2% N-carbamoyl-L-cysteine, or 2% L-carbamoyl-L-glutamic acid. To measure growth in the presence of tetracycline, a proxy for tetJ expression, LB was supplemented with 15 µg/mL tetracycline HCl. For comparison, control cultures were incubated in unsupplemented LB to demonstrate growth kinetics in the absence of tetracycline. After subtracting the OD₆₀₀ of the appropriate blank, data were visualized in GraphPad Prism 10.

5′ rapid amplification of cDNA ends

To identify transcriptional start sites of the ynt and nik operons, we used a 5′ rapid amplification of cDNA Ends Kit (5′ RACE Version 2.0) according to the manufacturer’s instructions (Invitrogen). RNA samples isolated from P. mirabilis HI4320 cultured for 2 hours (approximate OD₆₀₀ range 0.4–0.6) and exposed to urea (see Isolation of RNA) were used for gene-specific cDNA synthesis. Gene-specific primers (Table S1) were designed with guidance from the kit’s manual. After the addition of a poly-C tail to the gene-specific cDNAs, these products were further amplified using a second gene-specific primer closer to the transcriptional start site and primers included in the kit. 5′ RACE cDNA products were cloned into the pCR2.1-TOPO vector using the TOPO TA Cloning Kit (Invitrogen) and transformed into electrocompetent E. coli TOP10 cells. Transformants were selected on ampicillin (100 µg/mL) and screened for inserts using X-gal. Insert sizes were screened via PCR to exclude the products of incomplete extension during cDNA synthesis (Table S1). Plasmids were isolated by mini-prep (Epoch) and Sanger sequenced (Eurofins) to locate the nik and ynt transcriptional start sites. The σ⁷⁰ binding sites (−10 and −35) were predicted using BPROM software from Softberry (49).

Expression and purification of UreR-myc-His

Expression and purification of UreR-myc-His were performed as previously described (44). Briefly, an overnight culture of TOP10/pCP016 was sub-cultured into 200 mL LB supplemented with ampicillin (100 µg/mL) and incubated at 37°C with agitation. After 4 hours, corresponding with mid-logarithmic phase (OD₆₀₀ = ~0.7), expression of UreR-myc-His was induced with 0.2% arabinose. Bacteria were cultured for 3 hours before pelleting by centrifugation (4,000 × g) and then stored at −20°C. The thawed pellet was resuspended in 10 mL lysis buffer (50 mM NaH₂PO₄, 300 mM NaCl, and 10 mM imidazole) and lysed via French press. Lysate was centrifuged at 7,500 × g for 10 minutes to remove cellular debris and at 150,000 × g for 1 hour to remove membranes. One milliliter of Ni-NTA slurry (Qiagen) was incubated with the lysate overnight at 4°C with agitation. After applying the slurry mixture to a column, the resin was rinsed with wash buffer (50 mM NaH₂PO₄, 300 mM NaCl, and 20 mM imidazole) and protein was collected in elution buffer (50 mM NaH₂PO₄, 300 mM NaCl, and 250 mM imidazole). Elution fractions were separated using a Mini-PROTEAN TGX Precast 10% polyacrylamide gel (Bio-Rad) to confirm purification of UreR-myc-His (~33 kDa).

Electrophoretic mobility shift assay

Probes to measure UreR binding at the ynt operon promoter were designed to include the predicted σ⁷⁰ binding site and adjacent upstream sequences (total length: 71 nt). Complementary oligonucleotides of each probe were synthesized with a 6-FAM dye covalently linked to the 5′ end (IDT). Labeled, complementary oligonucleotides comprising part of the 16S promoter and unlabeled ynt oligonucleotides were used in control reactions. All probe sequences are presented in Table S1. Complementary oligonucleotides were mixed in equimolar concentrations in duplex buffer (50 mM NaH₂PO₄ and 300 mM NaCl, pH 8) prior to annealing (heated to 95°C and cooled to 16°C in a stepwise manner via thermocycler: 5 minutes at 95°C; 1 minute at 90°C, 80°C, etc.; and 10 minutes at 16°C). Probes were incubated for 15 minutes with or without purified UreR-myc-His in a binding reaction containing 150 mM KCl, 200 mM urea, 250 ng poly[d(A-T)], 0.1 mM dithiothreitol, 0.1 mM EDTA, and 10 mM Tris base, pH 7.4. Protein-DNA duplexes were separated on a pre-electrophoresed 5% Criterion TBE Polyacrylamide Gel (Bio-Rad) in 0.5× TBE buffer at 4°C. Fluorescence imaging was performed on a Typhoon (Amersham).

Generating a Morganellaceae phylogeny

Morganellaceae family species names and reference strains were identified using NCBI taxonomy (https://www.ncbi.nlm.nih.gov/taxonomy). These 83 reference genomes were submitted to the Bacterial and Viral Bioinformatics Resource Center (BV-BRC) Codon Tree pipeline to generate a phylogenetic tree (settings: max allowed deletions = 0; max allowed duplications = 0; and single copy genes = 100) (50). Seventy-eight single-copy genes were identified by the pipeline and used to build a tree that was imported to iTOL for visualization and labeling (51).

Identification of UreR homologs

Urease loci were identified in representative genomes from each Morganellaceae family species in BV-BRC (50). Using the genome browser function, sequences upstream of each urease locus were assessed for the presence of an annotated AraC/XylS family transcriptional regulator.

To aid in the identification of any UreR homologs outside of the Morganellaceae family, we input the N-terminal domain of the UreR amino acid sequence (residues 1–164) into the Enzyme Function Initiative (EFI) enzyme similarity tool (52). The C-terminal helix-turn-helix domain was excluded to prevent the inclusion of candidate homologs solely based on homology within the AraC/XylS family DNA-binding domain. We employed standard settings with a taxonomy filter for species within the Bacterial domain. This analysis identified 86 UniProt IDs that represented 78 unique sequences of which the overwhelming majority were species within class Gammaproteobacteria (85/86 UniProt IDs). These sequences were transferred to the EFI Genome Neighborhood Tool (53) and filtered to identify sequences within 10 genes of a urease operon (n = 41). Using UniProt, sequences were aligned and a percent identity matrix was generated (54). The percent identity matrix was visualized with GraphPad Prism 10.0.

RESULTS

UreR rapidly responds to urea by inducing expression of urease genes

The production of urease, regulated by UreR, is a major virulence determinant in P. mirabilis uropathogenesis. Previous studies demonstrated that ureR is required for urease production and disruption of ureR results in a significant fitness disadvantage in a murine model of UTI (22–25). Additionally, UreR directly binds urea, a nitrogenous waste metabolite much more concentrated in human urine (~400 mM) than in other parts of the body (27, 28), for example, the normal range for urea in the blood is 1.8 to 7.1 mM (55). UreR rapidly responds to this cue upon entering the urinary tract and strongly induces urease gene expression (56). Due to the speed and magnitude of this response, we hypothesized that UreR regulates other genes required for rapidly adapting to the urinary tract. To address this hypothesis, we performed RNA sequencing to identify additional regulatory targets of UreR.

While it would be instructive to define the UreR regulon in human urine, the consequences of urease activity make this approach infeasible. When P. mirabilis is cultured in urine in a closed system (i.e., a test tube), urease-derived ammonia accumulates quickly. Ammonia production results in a dramatic elevation in pH that negatively impacts the growth rate of P. mirabilis and can precipitate urinary ions into crystals; both phenomena encumber the isolation of high-quality RNA. Instead of using urine, we performed experiments in LB medium containing or lacking urea, the specific inducer of UreR (57). Previous work demonstrated that 50 mM urea is sufficient to induce maximal urease expression and activity (25, 57). To prevent a urease-driven pH change in our cultures, we minimized the time cultured with urea and buffered the medium with 120 mM HEPES. Buffered LB enabled the separation of transcriptional changes driven by urea from those driven by a pH change. For these reasons, we performed experiments in buffered LB medium, pH 7, spiked with 50 mM urea during mid-logarithmic growth.

Next, we performed a time course experiment to establish the kinetics of the UreR regulatory response to urea using a known UreR target. Overnight cultures of P. mirabilis type strain HI4320 in LB were subcultured into buffered LB in duplicate and cultured at 37°C with aeration. When the cultures reached mid-logarithmic phase, half of the cultures were spiked with urea to a final concentration of 50 mM. Samples were collected for RNA extraction 15, 30, and 60 minutes after urea addition. Samples containing or lacking urea at each time point were compared to measure urea-induced expression of ureD and ureR, the only known UreR targets at the start of this study. There was no significant increase in ureR expression at any time point assayed (Fig. S1). After 15 minutes of urea exposure, expression of ureD was 4.2 log₂-fold higher than in time-matched cultures lacking urea (Fig. 1B). This response waned at the 30 minute mark, where the urea-containing culture exhibited a 2.4 log₂-fold increase in ureD expression. At 60 minutes, expression of ureD was markedly reduced (0.8 log₂-fold) compared with expression at 15 minutes after addition of urea. These data indicated that, in the conditions tested, urea-induced expression of ureD peaks early after a urea spike and returns to baseline around 60 minutes after the addition of urea. To ensure we would see UreR-induced transcriptional changes in our RNA-seq experiment, we selected the 15-minute time point for sample collection.

One goal of this study was to define the UreR regulon using comparative transcriptomics. We sought to define differences in gene expression in wild-type HI4320 when compared with a mutant lacking UreR (ureR::aphA). Since UreR is induced by the presence of urea, we added urea to buffered medium to induce UreR regulatory activity. One complication of this approach is that UreR induction rapidly results in urease activity that produces ammonia, thereby increasing the local pH. Therefore, our RNA sequencing approach must successfully differentiate direct UreR targets from genes simply responding to the downstream effects of urease activity. Toward this end, we included a urease-null construct (ureC::bla) in our study. In previous work, our lab had demonstrated that the ureC::bla mutant is unable to make functional urease enzyme and exhibits a severe defect in a murine model of ascending UTI (12, 13). The ureC gene is the fourth gene in the ureDABCEFG operon (Fig. 1A). Thus, an insertion in ureC (Fig. S1) should have no effect on the expression of ureR or ureDAB. This enables the use of ureD induction as a proxy for UreR activity. Thus, we measured the urea-induced expression of ureD in each of our RNA-seq study strains as described above. As expected, urea induced the expression of ureD in wild-type HI4320 10.6-fold but had no effect on ureD expression in ureR::aphA (Fig. 1C). Without UreR, ureR::aphA is unable to respond to urea, further confirming that UreR is the sole urea-responsive regulator of urease genes. Expression of ureD was induced by urea in ureC::bla, indicating that UreR is active in this strain. In summary, wild-type HI4320 produces urease in response to UreR induction, whereas ureR::aphA is unable to produce urease due to disruption of ureR, and ureC::aphA cannot produce urease but retains UreR activity. Based on these results, we used these three constructs to genetically isolate UreR regulation to aid in the identification of its regulatory targets.

Identification of putative negatively regulated UreR targets

We first compared the transcriptomes of wild-type P. mirabilis HI4320 and ureR::aphA without induction to investigate whether we could detect any genes that were negatively regulated by UreR. Nine genes were significantly upregulated in ureR::aphA compared with the wild type, while a group of seven genes was significantly downregulated (Table 2). One notable gene upregulated in ureR::aphA was tetJ, an efflux pump that confers tetracycline resistance upon P. mirabilis (58). tetJ experienced an unexpected 6.6 log₂-fold increase in expression in ureR::aphA compared with the wild type. However, differential expression of tetJ was not observed in the ureC::bla transcriptome when cultured in LB (Table S3).

TABLE 2.

Differentially regulated genes in LB: HI4320 vs ureR::aphA

Locus tag	Old locus tag	Gene	Protein	Log2 ratio	P value
PMI_RS11870	PMI2399	tetJ	Tetracycline efflux MFS transporter Tet(J)	6.6	0
PMI_RS18285	PMI3675	dcuC	C4-dicarboxylate transporter	1.7	3.40E−13
PMI_RS06270	PMI1300	add	Adenosine deaminase	1.6	8.74E−11
PMI_RS13430	PMI2725		TetR/AcrR family transcriptional regulator	1.3	6.35E−07
PMI_RS01385	PMI0287		Zn-dependent hydrolase CDS	1.2	9.23E−07
PMI_RS15115	PMI3052	asnA	Aspartate—ammonia ligase	1.1	2.00E−06
PMI_RS03300	PMI0671	sdaB	L-Serine ammonia-lyase	1.1	5.73E−06
PMI_RS03965	PMI0807		Cold-shock protein	1.0	5.49E−05
PMI_RS17710	PMI3560		Sulfite exporter TauE/SafE family protein	1.0	4.28E−05
PMI_RS18315	PMI3681	ureR	AraC family transcriptional regulator	−1.1	2.76E−12
PMI_RS18335	PMI3685	ureC	Urease alpha subunit	−1.2	1.11E−15
PMI_RS15115	PMI3052	asnA	Aspartate—ammonia ligase	−1.3	4.15E−07
PMI_RS18330	PMI3684	ureB	Urease beta subunit	−1.4	1.54E−16
PMI_RS18320	PMI3682	ureD	Urease accessory protein UreD	−1.5	1.03E−16
PMI_RS18325	PMI3683	ureA	Urease gamma subunit	−1.5	1.10E−19
PMI_RS14875	PMI3008	gdhA	NADP-specific glutamate dehydrogenase	−1.7	8.04E−13

UreR does not regulate tetracycline efflux pump TetJ

Strong upregulation of tetJ in ureR::aphA when cultured in LB was confirmed via qRT-PCR (7.3 log₂-fold; Fig. 2A). We were perplexed by this result since there is no literature connecting UreR to the endogenous tetracycline resistance locus of P. mirabilis. Work in other species indicates that tetJ is constitutively repressed by its neighboring gene tetR, and this repression is alleviated by tetracycline (59). Importantly, the medium used to culture these strains for RNA isolation lacked tetracycline. One possible explanation is that tetJ expression is driven by PMI2725, a TetR family transcriptional regulator that was upregulated 1.3 log₂-fold in ureR::aphA. However, altered expression of PMI2725 in ureR::aphA was not detected by qRT-PCR (Fig. S2A). Next, we examined if increased tetJ could be driven by a strain-specific phenotype or the expression of aphA which encodes kanamycin resistance. To address this hypothesis, we constructed a second ureR mutant using targetron mutagenesis. In this method, a group II intron containing a kanamycin resistance gene is inserted into the 5′ end of the ureR gene to produce ureR::kan. We subsequently deleted the antibiotic resistance gene to form an unmarked mutant that retains a 1-kb insertion (ureR∆kan). We assayed wild-type HI4320 and all three ureR mutants for the ability to grow in the presence of tetracycline. All strains grew well in LB (Fig. 2B), but only ureR::aphA was able to immediately grow in LB containing 15 µg/mL tetracycline (Fig. 2C). In comparison, HI4320, ureR::kan, and ureR∆kan experienced an extended lag (~10 hours), which is consistent with growth patterns observed in tetracycline-resistant Escherichia coli (59). These data indicate that tetJ induction in the ureR::aphA mutant is not due to UreR but instead may be an artifact of strain construction or random mutation during laboratory handling. Furthermore, increased expression of tetJ was not driven solely by the presence of a kanamycin resistance gene, as the newly constructed ureR::kan behaved similarly as ureR∆kan. Moving forward, we used ureR∆kan to confirm that phenotypes observed in ureR::aphA were driven by UreR regulatory activity.

Fig 2.

Comparison of wild-type and ureR mutant gene expression in the absence of urea induction. (A) HI4320 and ureR::aphA were grown to mid-logarithmic phase in LB medium. Expression of tetJ, sdaC, add, and PMI0287 in ureR::aphA was measured via qRT-PCR. Data were normalized to expression in HI4320 and housekeeping gene rpoA. Parametric data (tetJ, sdaC) were analyzed via one-sample t-test while nonparametric data (add, PMI0287) were analyzed via Wilcoxon signed-rank test, both with a hypothetical mean of 1. HI4320, ureR::aphA, ureR::kan, and ureR∆kan were cultured in LB (B) or LB supplemented with 15 µg/mL tetracycline (C). HI4320, ureC::bla, and ureR::aphA were cultured in Minimal A medium (D) or minimal medium where the nitrogen source was replaced with 0.2% L-serine (E) or 0.2% adenosine (F). All growth experiments were performed at 37°C with aeration for 24 hours. Graphed data are the means of three biological replicates, and error bars represent standard error of the means.

Disruption of ureR subtly alters expression of ammonia metabolism genes

When examining the list of differentially regulated genes in ureR::aphA as compared with the wild type, a group of genes stood out as being involved in nitrogen metabolism (Table 2). As expected, urease genes (ureDABC) and ureR were significantly downregulated in the mutant compared with the wild type. Repression of urease genes eliminates ammonia production driven by urease activity. Accordingly, ureR::aphA also downregulated two ammonia assimilation genes that incorporate intracellular ammonia, gdhA (−1.7 log₂-fold) and asnA (−1.3 log₂-fold). Glutamate dehydrogenase (GdhA) is a key ammonia assimilation enzyme that catalyzes the conversion of 2-oxoglutarate and ammonium to form L-glutamate (60). gdhA is strongly upregulated by P. mirabilis during experimental UTI, and is hypothesized to assimilate ammonia produced by urease during infection (61). Downregulation of gdhA may indicate subtle differences in nitrogen assimilation in ureR::aphA compared with the wild type. Also downregulated was aspartate-ammonia ligase (asnA), which catalyzes the conversion of ammonia and L-aspartate to L-asparagine. Neither of these gene expression changes were detected by qRT-PCR (Fig. S2A), perhaps due to the small magnitude and different normalization methods in RNA-seq vs qRT-PCR.

Conversely, transcripts for enzymes that generate ammonia/ammonium were upregulated in the ureR::aphA mutant compared with the wild type. An operon encoding both a serine transporter (sdaC) and a serine deaminase (sdaB) was upregulated 1.1 log₂-fold. After serine is imported by SdaC, ammonium is liberated from serine by SdaB to form pyruvate (62). This reaction provides a nitrogen source (ammonium) and produces a byproduct (pyruvate) that can enter the TCA cycle or branched chain amino acid biosynthesis pathways (56, 60). Additionally, SdaC was recently shown to mediate self-recognition in P. mirabilis during swarming (63). Also upregulated are adenosine deaminase (add, 1.6 log₂-fold) and a putative N-carbamoyl-L-amino acid hydrolase (PMI0287, 1.2 log₂-fold). These enzymes release ammonia as byproducts of their reactions. Once again, we were unable to observe significant expression changes by qRT-PCR (Fig. 2A). However, we hypothesized that the substrates of these enzymes would be more readily available nitrogen sources for ureR::aphA than for the wild type due to the increased expression of gdhA, add, and PMI0287 in ureR::aphA. To test this possibility, we cultured bacteria in Minimal A medium or Minimal A where the nitrogen source (ammonium sulfate) was replaced by L-serine, adenosine, or two commercially available N-carbamoyl-L-amino acids (glutamic acid and cysteine). In all conditions, glycerol was provided as a nonfermentable carbon source. No differences in growth rates of these strains were observed in Minimal A (Fig. 2D). Similarly, no differences in growth were observed when the nitrogen source was L-serine (Fig. 2E) or adenosine (Fig. 2F). L-Serine is an important nutrient for P. mirabilis during UTI (64), but this is the first report that P. mirabilis can use adenosine as a sole nitrogen source. Wild-type HI4320 was unable to grow in Minimal A where either N-carbamoyl-L-glutamic acid or N-carbamoyl-L-cysteine was the sole nitrogen source (Fig. S2B). However, these metabolites cannot be entirely ruled out as substrates of PMI0287; perhaps, P. mirabilis simply does not have transporters capable of importing these substrates to the cytoplasm.

Previous work in our lab indicates that urease exhibits low levels of constitutive activity in LB without supplementing with urea (12). The elimination of this baseline activity through disruption of ureR may prompt ureR::aphA to downregulate certain ammonia incorporation genes (gdhA and asnA) and upregulate alternate ammonia production genes (sdaBC, add, and PMI0287) as a compensatory mechanism. These data indicate a potential role for urease in cellular metabolism outside of the context of pathogenesis.

The P. mirabilis urea-induced transcriptome includes urease genes and nickel importers

Next, we characterized the urea-induced transcriptome of wild-type HI4320 to identify genes as potential urea-responsive UreR targets. In the wild type, a total of 11 genes experienced at least a twofold upregulation (or 1 log₂-fold change) and were statistically significant (adjusted P values < 0.05) (Table 3). There were no downregulated genes that met these criteria, reinforcing that UreR is a positive regulator when bound to urea. Seven of these upregulated genes are in the urease operon (ureDABCEFG) and were induced 1.2 to 1.9 log₂-fold. This group served as an internal control; based on extensive literature, we expected the urease operon to be upregulated in the presence of urea (22, 23, 25, 56, 57, 65, 66). The remaining four genes are within two operons that encode nickel transport systems: yntA (1.7 log₂-fold), yntB (1.3 log₂-fold), PMI2948 (1.1 log₂-fold), and nikA (1.2 log₂-fold). Notably, nickel is a cofactor required for urease enzyme maturation (31, 67–72). Every gene in the ynt and nik operons experienced a statistically significant upregulation, but only 4/11 genes were induced at least twofold. Transcripts per million values (a metric for overall transcript abundance) indicated that transcripts of genes closer to the beginning of each operon were more abundant than later operon genes for both ynt and nik (Fig. S3). This is in line with the understanding that genes at the beginning of an operon are generally more highly expressed (due in some cases to mRNA stability of long transcripts) and, therefore, more strongly activated in the presence of an inducer. Thus, the ynt and nik transporter operons exhibited urea-responsive expression.

TABLE 3.

Urea-induced transcriptome in HI4320

Locus tag	Old locus tag	Gene	Protein	Log2 ratio	P value
PMI_RS18335	PMI3685	ureC	Urease alpha subunit	1.9	2.69E−61
PMI_RS18340	PMI3686	ureE	Urease accessory protein UreE	1.8	5.03E−52
PMI_RS18330	PMI3684	ureB	Urease beta subunit	1.8	8.12E−50
PMI_RS18320	PMI3682	ureD	Urease accessory protein UreD	1.7	1.39E−48
PMI_RS12440	PMI2516	yntA	ABC transporter substrate-binding protein	1.7	4.85E−53
PMI_RS18345	PMI3687	ureF	Urease accessory protein UreF	1.6	2.02E−43
PMI_RS18325	PMI3683	ureA	Urease gamma subunit	1.6	8.39E−40
PMI_RS12435	PMI2515	yntB	ABC transporter permease	1.3	3.23E−32
PMI_RS18350	PMI3688	ureG	Urease accessory protein UreG	1.2	1.05E−25
PMI_RS14575	PMI2948	PMI2948	Class I SAM-dependent methyltransferase	1.2	1.59E−33
PMI_RS14570	PMI2947	nikA	ABC transporter substrate-binding protein	1.1	1.86E−23

If induction of nickel transporters was driven by UreR and not a byproduct of urease activity, we would expect a similar expression pattern in the urease-null ureC::bla mutant. We addressed this hypothesis by determining the urea-induced transcriptome of ureC::bla. Seven genes were induced by the presence of urea in ureC::bla, and no genes were downregulated (Table 4). The top two upregulated genes were ureA (2.5 log₂-fold) and ureD (2.1 log₂-fold), which are both upstream of the ureC insertion that renders this strain urease-null (Fig. S4). Also upregulated were nickel transport genes yntA (2.0 log₂-fold), yntB (1.5 log₂-fold), yntE (1.0 log₂-fold), nikA (1.0 log₂-fold), and nikB (1.1 log₂-fold). The operons that are urea responsive in wild-type HI4320 remained urea responsive in ureC::bla, indicating that this phenotype is a direct response to UreR regulatory activity and not urease enzymatic activity.

TABLE 4.

Urea-induced transcriptome in ureC::aphA

Locus tag	Old locus tag	Gene	Protein	Log2 ratio	P value
PMI_RS18325	PMI3683	ureA	Urease alpha subunit	2.5	8.25E−38
PMI_RS18320	PMI3682	ureD	Urease accessory protein UreD	2.4	1.62E−26
PMI_RS12440	PMI2516	yntA	ABC transporter substrate-binding protein	2.0	3.63E−29
PMI_RS12435	PMI2515	yntB	ABC transporter permease	1.5	1.23E−16
PMI_RS14565	PMI2946	nikB	ABC transporter permease subunit	1.1	1.69E−07
PMI_RS12420	PMI2515	yntE	ATP-binding cassette domain-containing protein	1.0	3.88E−08
PMI_RS14570	PMI2947	nikA	Nickel ABC transporter substrate-binding protein	1.0	1.65E−06

Before continuing this line of inquiry, we validated the urea responsiveness of the ynt and nik operons by qRT-PCR (Fig. 3). The ynt operon was upregulated by the addition of urea in the wild type (2.7 log₂-fold) and ureC::bla (4.8 log₂-fold), but not in ureR::aphA or ureR∆kan. Urea-induced expression of ynt was significantly different (P < 0.05) when comparing either strain encoding active UreR (wild type or ureC::bla) with either strain lacking UreR (ureR::aphA or ureR∆kan). Similarly, urea-responsive expression of the nik operon was observed in the wild type (2.8 log₂-fold) and ureC::bla (4.0 log₂-fold), but not in ureR::aphA or ureR∆kan. qRT-PCR data measuring nik expression experienced greater variability between samples compared with the ynt expression data. Thus, the only comparison that exhibited a statistically significant difference in nik induction was ureC::bla vs ureR::aphA. Like in our RNA-seq study, ynt experienced greater urea-responsive induction than nik when measured by qRT-PCR; perhaps, this contributed to the sample-to-sample variation in nik expression and fewer significant comparisons. Taken together, our transcriptomic data provide evidence that UreR regulates nickel transport in addition to its well-characterized regulation of urease.

Fig 3.

UreR is required for urea-induced expression of nickel transport genes. HI4320, ureC::bla, and ureR::aphA were cultured to mid-logarithmic phase and supplemented with urea. After 15 minutes, samples were removed to quantify expression of ynt (A) or nik (B) via qRT-PCR. Data were normalized to time-matched controls incubated without urea and expression of housekeeping gene rpoA. Each bar represents the mean of 4–6 biological replicates, and error bars represent standard error of the means. Significance of ynt and log-transformed nik data was determined by one-way ANOVA with Tukey’s multiple comparisons correction. P: *<0.05 and **<0.01.

Finally, we calculated the urea-induced transcriptome for ureR::aphA. No genes experienced significant expression changes. Thus, UreR was the only urea-responsive regulator active in the conditions tested, confirming that UreR is required for P. mirabilis to coordinate a genetic response to the presence of urea. Taken together, these three datasets indicate that UreR is a positive regulator of both urease and nickel transport in P. mirabilis.

Identification of transcriptional start sites of nickel transport operons

A hallmark of the AraC/XylS family of transcriptional regulators, to which UreR belongs, is that binding sites are directly upstream and slightly overlapping with the −35 sites in target promoters. The two known UreR binding sites within the ureR and ureD bidirectional promoter fit within this paradigm. To inform the selection of DNA sequences to use to assess whether UreR directly regulates nickel transport genes in P. mirabilis, we performed 5′ RACE to identify the transcriptional start sites of the ynt and nik operons. As part of this protocol, a poly-C tail is added to the products of gene-specific cDNA synthesis to facilitate further amplification and cloning. We mapped the ynt transcriptional start site to either 24 bp (“T”) or 25 bp (“G”) upstream of the operon’s first start codon (Fig. 4A). Since the poly-C tail is added to the reverse complement of the mRNA encoding these operons, we were unable to discriminate between these two possible start sites. We encountered a similar problem when mapping the nik transcriptional start site; transcription of this operon begins either 187 or 188 bp before the first start codon (Fig. S5). Given that the 5′ untranslated region for nik is considerably longer than that of ynt, we also confirmed that there were no in-frame classical or alternate start codons (“ATG,” GTG,” and “TTG”) upstream of the annotated start codon. Finally, we identified the likely −10 and −35 sites of each promoter (depicted in Fig. 4A; Fig. S5).

Fig 4.

UreR directly binds the ynt promoter. (A) A schematic of the ynt promoter. The transcriptional start site was determined via 5′ RACE using cDNAs from HI4320 cultured with urea. The predicted −10 and −35 sites are indicated. The sequence highlighted in blue was used as a probe in the electrophoretic mobility shift assays (Pynt). (B) Gel shift reaction demonstrating that purified UreR-myc-His binds the ynt promoter. A FAM-labeled fragment encoding part of the ynt promoter (Pynt-FAM; 71 bp) was incubated with increasing amounts of UreR-myc-His (lanes 2–5), producing a shifted band. The addition of 10-fold excess unlabeled Pynt reduced the intensity of the shifted band (lane 6). When UreR-myc-His was incubated with a FAM-labeled fragment encoding part of the 16S promoter (P16S-FAM; 71 bp), no shifted band appeared (lanes 7–8).

UreR directly regulates Ynt nickel transporter

To determine if UreR directly binds the ynt operon promoter, we conducted electrophoretic mobility shift assays using purified recombinant UreR-myc-His. The 5′ RACE results informed the design of a 71-bp fragment of the ynt promoter that includes the predicted −35 site. We annealed complementary, 5′ 6-FAM-labeled single-stranded DNA oligonucleotides to form the probe Pynt-FAM. Purified UreR-myc-His was incubated with 6 nM Pynt-FAM in binding buffer. Unbound probe was separated from probe bound by UreR-myc-His via electrophoresis using a 5% polyacrylamide gel. Incubation of the probe with increasing amounts of UreR-myc-His produced a shifted band, indicating the formation of UreR-myc-His:Pynt-FAM complexes (Fig. 4B, lanes 2–5). Excess unlabeled probe (Pynt) was able to outcompete this binding reaction, as indicated by the increased intensity of the band representing unbound probe (Fig. 4B, lane 3 vs lane 6). UreR-myc-His was unable to bind a labeled 16S promoter fragment (P16S-FAM), indicating that the binding interaction between UreR-myc-His and Pynt-FAM is specific (Fig. 4B, lanes 7–8). These data indicate that UreR directly binds the ynt promoter, thereby expanding the known direct targets of UreR to include nickel transport.

UreR-mediated regulation of urease and nickel transport may be conserved in other bacterial species

Genes encoding urease are found throughout the tree of life in bacteria, fungi, plants, algae, and some invertebrates and are regulated in a variety of ways (14, 73). For example, urease can be constitutively expressed (Morganella morganii) or induced by environmental signals such as acid stress (Helicobacter pylori), low nitrogen levels (Klebsiella aerogenes), or a combination of acid stress and high glucose (Streptococcus salivarius) (73–76). Historically, urea-responsive regulation of urease by UreR has been limited to P. mirabilis and a plasmid carried by Providencia stuartii (73, 77, 78). More recently, UreR-regulation of urease was also observed on a mobile element in Vibrio parahaemolyticus (79). Due to the phylogenetic distance between Proteus and Vibrio genera, we were interested in characterizing the distribution of UreR-containing urease loci. To start, we focused our efforts on the Morganellaceae family that is home to both Proteus and Providencia genera. After identifying representative genomes of 83 Morganellaceae family members (Table S4), we used BV-BRC to identify urease loci with an AraC/XylS family transcriptional regulator. This approach allowed us to identify UreR homologs that lacked an annotated gene name. We identified 15 species that encode UreR homologs across three genera (Fig. 5A, red). All species within the Proteus (11/11) and Cosenzaea (1/1) genera encode UreR, as well as three Providencia species (3/12). No UreR homologs were identified within Arsenophonus, Moellerella, Morganella, Photorhabdus, and Xenorhabdus genera (Fig. S6). These UreR homologs are highly conserved at the amino acid sequence level with >68% identity across all homologs (Fig. 5B). Two distinct clusters exhibited >80% identity. One cluster contained all Proteus and Cosenzaea UreR homologs while the other was composed of Providencia UreR homologs. In addition to sequence divergence, the Providencia UreR locus also featured a distinct operon organization (Fig. 5C). A complete ynt operon was encoded immediately downstream of UreR, condensing the two P. mirabilis UreR targets into a single genetic locus (Fig. 5C). For comparison, P. mirabilis ureR and yntA are 1.2 Mb apart (40). We next determined the distribution of ynt and nik transport systems among UreR-encoding species. This analysis revealed that 15/15 identified species also encode a ynt nickel transporter. Conversely, the nik transporter was present in Proteus and Cosenzaea but absent from Providencia species (12/15).

Fig 5.

UreR homologs were identified in 15 Morganellaceae family species. (A) Phylogeny of 83 Morganellaceae family species built by the BV-BRC Codon Tree tool using 78 single-copy genes. Species highlighted in red encode a predicted UreR-regulated urease locus, and the Photorhabdus and Xenorhabdus clades have been collapsed for legibility. The full tree is depicted in Fig. S6. The branch length scale represents the mean number of substitutions per site. (B) Percent identity analysis of Morganellaceae family UreR homologs. (C) Schematics of three distinct UreR-regulated urease loci identified in this study.

Although effective, examining representative genomes of Morganellaceae family members did not identify the plasmid-encoded UreR homolog characterized in P. stuartii isolate BE2467 (26, 77, 80). Upon further investigation, the genome of the representative strain used in our analysis (CAVP490) lacks this plasmid. A deeper dive into P. stuartii genomes on BV-BRC identified 32 urease operons across 30 P. stuartii isolates. Of these, only seven operons encoded the UreR-regulated organization of P. stuartii BE2467 (Fig. 5C). Interestingly, this plasmid-borne urease locus has the same gene organization as the Providencia-type chromosomal urease locus.

Recently, UreR was identified in some V. parahaemolyticus strains (79, 81), and this ureR locus presents a third operon organization; the ynt operon is encoded immediately upstream of UreR between ureR and ureDABCEFG (Fig. 5C). Notably, a key virulence factor for this species is encoded immediately downstream of ureR (81). The TDH-related toxin (thr) is one of two hemolysins heavily involved in V. parahaemolyticus pathogenesis (82, 83). Nearby transposases and a mobile element protein hint that this locus was likely acquired via horizontal gene transfer (81).

Given the evidence that the P. mirabilis urease locus has mobilized in at least two instances (via plasmid and transposase), we sought to locate additional UreR homologs outside of the Morganellaceae family. We used the Enzyme Function Initiative’s enzyme similarity tool (EFI-EST) (52, 53) to identify candidate UreR homologs. This web-based bioinformatics tool leverages UniProt databases to identify proteins similar to a user-provided sequence. The output of this tool is easily transferred to EFI’s genome neighborhood tool, which provides the genomic context for each hit (52, 53). Although this tool was designed for enzymes, we modified the workflow to identify transcriptional regulators closely related to UreR. To curb the identification of proteins solely based on a shared AraC/XylS family DNA-binding domain, we submitted the 164 amino acids that comprise the N-terminal urea-binding dimerization domain to EFI-EST. This approach yielded 87 putative UreR homologs, 43 of which were adjacent to a urease operon (within 10 genes). Our analysis expanded the distribution of urease-adjacent UreR homologs from four genera to six: Proteus, Providencia, Cosenzaea, Vibrio, Shigella, and Shewanella (Table S5). Analysis revealed that these UreR homologs shared >40% identity with HI4320 UreR (Table S5). Taken together, these data indicate that UreR is found in a subset of Gammaproteobacteria-class species. UreR has been found on a plasmid and near mobile genetic element genes, indicating the locus has mobilized at least once. The fusion of the ure locus to the ynt operon indicates that UreR regulation of nickel transport is likely conserved across UreR loci.

DISCUSSION

Urinary stone formation is a major source of P. mirabilis pathology and is a direct result of urease activity. After urease genes are transcribed and translated, the structural subunits (UreA, UreB, and UreC) assemble into a homotrimer of heterotrimers, forming the nickel-free urease apoenzyme (UreABC)₃ (73). Urease accessory protein UreE acquires nickel from the Ynt ABC transporter via a naturally occurring 6xHis sequence. These nickel cations are then transferred to a protein complex comprising UreD, UreF, and UreG. Finally, the UreDFG complex incorporates nickel into the urease apoenzyme in an energy-dependent manner, forming mature urease (29, 30, 72). After enzyme maturation, cytosolic urease hydrolyzes urea to form ammonia and carbon dioxide. Ammonia can serve as a nitrogen source or diffuse out of the cell, increasing the local pH. This increased alkalinity precipitates polyvalent cations and anions that are usually soluble in urine, forming struvite and carbon apatite (84). Over time, this leads to the formation of urinary stones that both encapsulate P. mirabilis and form a colonizable surface (15). The urinary stone niche enables immune evasion and serves as a reservoir of bacteria that can continually seed infection (18, 21). Generating these stones requires the coordination of three cellular activities: production of urease structural subunits, acquisition of nickel from the extracellular environment, and enzyme maturation via nickel incorporation. Here, we demonstrated that UreR serves as the coordinator of these functions through activating expression of the ure, ynt, and nik operons. Thus, UreR is the architect of urinary stone construction.

Most AraC/XylS family transcriptional regulators directly sense a metabolite and act as positive regulators, and UreR is no exception. P. mirabilis encounters high concentrations of urea while in the urinary tract (400 mM on average in humans), which is thought to access the bacterial cytoplasm via diffusion (27, 28). What follows is a UreR-driven regulatory cascade that results in urease activity (Fig. 6). Upon binding urea, UreR dimerizes and binds at target promoters (25, 26, 44, 85). Prior to this study, there were two defined UreR binding sites within a bidirectional promoter activating ureDABCEFG expression as well as autoregulation of ureR. Our work illustrates that UreR also activates expression of nickel transporters ynt and nik. Regulation of ynt is facilitated by direct binding of its promoter. UreR is the first regulator of nickel transport to be identified in P. mirabilis, a species that lacks homologs of established nickel regulators NikR, Nur, and NmtR (35–37). Additionally, P. mirabilis UreR exemplifies a relatively simple regulatory network in which regulation of urease and nickel import is coordinated by a single regulator and metabolite. No other inducers of P. mirabilis urease have been identified (57). Comparatively, H. pylori urease is regulated by two-component systems ArsRS and FlgRS as well as the nickel transport regulator NikR (76, 86, 87). There is also some evidence of posttranslational regulation of H. pylori urease (76). Thus, UreR represents a streamlined, urea-inducible mechanism for regulating urease and nickel transport genes.

Fig 6.

UreR regulates all genes required to produce active urease. A schematic of our proposed model of UreR regulation. When urea enters the cell, UreR binds this metabolite and dimerizes. Dimerized UreR directly regulates the ure and ynt operons (and potentially the nik operon), activating expression of these genes. The Ynt high-affinity nickel transporter imports nickel, which is transferred to accessory protein UreE. Nickel chaperones UreD, UreF, and UreG receive nickel from UreE and incorporate this cofactor into the urease apoenzyme. Then, mature urease hydrolyzes available urea, forming ammonia and carbon dioxide products.

P. mirabilis encodes two distinct nickel-specific ABC transport systems that deliver nickel to urease and two (NiFe) hydrogenases (32). Ynt more efficiently imports nickel into the cytoplasm and delivers nickel to urease during UTI (32). Furthermore, a ynt mutant can only produce functional urease if the medium is supplemented with 10 µM nickel, which is well above what is physiologically available in urine (32). Our work characterizing the distribution of UreR among bacterial species further supports a preference for Ynt. First, all Morganellaceae family species that encode a UreR homolog also encode Ynt. Conversely, only species in the Proteus and Cosenzaea genera encode Nik. Notably, Proteus and Cosenzaea genera are closely related; the sole species in the Cosenzaea genus (C. myxofaciens) was considered a member of Proteus until a recent reclassification (88). Additional evidence of a preference for Ynt is provided by two mobilized UreR-regulated urease loci. One locus is on a plasmid that has been found in three species within Enterobacterales (26). The other locus is adjacent to transposases in V. parahaemolyticus, implying the presence of a mobile genetic element (81). Both of these loci encode ynt operons immediately adjacent to and on the same strand as ureR, indicating they may be co-transcribed. This finding highlights a preference for the Ynt transporter and also implies that UreR regulation of ynt is conserved across species.

Despite its high concentration in urine, P. mirabilis does not seem to recognize urea as a cue specifically signaling urinary tract entry. Rather than orchestrating broad transcriptional changes to adapt to the urinary tract environment, UreR specifically activates genes that support urease activity. For a species so proficient in colonizing the catheterized urinary tract, we were surprised that UreR regulates such a narrow set of genes in response to urea. In addition to urease, P. mirabilis encodes a second, energy-dependent pathway for urea catalysis. In this pathway, urea is first converted to allophanate by biotinylated urea carboxylase (MumC) and subsequently hydrolyzed by allophanate hydrolase (MumH) (89). This urea catabolism mechanism requires manganese rather than nickel. Preference for one mechanism over the other may be driven by the availability of metal cofactors (nickel and manganese), biotin biosynthesis, and intracellular ATP levels. Given urease’s relatively low affinity for urea, the preferred urea catabolism pathway could be influenced by the concentration of urea (90). P. mirabilis tends to minimize functional redundancy within its genome, so encoding two urea catalysis mechanisms is unusual (40). Thus, we posit that urea breakdown is an important function within P. mirabilis metabolism.

There are a few possibilities as to why urea catalysis is crucial to P. mirabilis. First, perhaps urea is a nitrogen source broadly available across niches harboring P. mirabilis. We have little knowledge regarding the environmental reservoirs of P. mirabilis, but urea is excreted as a nitrogenous waste product from mammals (91). Further environmental sampling studies are required to understand which niches are inhabited by P. mirabilis and determine if urea is available in these niches. Alternatively, the importance of urea catalysis may be driven by endogenous urea production. We have previously noted that P. mirabilis is able to both synthesize urea and break it down (40). Urea production is driven by the polyamine biosynthesis pathway; converting agmatine to putrescine forms a urea byproduct. Notably, putrescine is synthesized for incorporation into the lipopolysaccharide (LPS) core of P. mirabilis (92). Thus, LPS biosynthesis represents a potential source of intracellular urea production. There is also ample evidence that putrescine is an important cue in swarming motility, in which elongated P. mirabilis cells form coordinated rafts to swarm across a surface. Genetic disruption of polyamine biosynthesis results in swarming defects that are rescued by supplementation with putrescine (93). In addition to synthesizing putrescine, wild-type swarming requires the ability to import putrescine via transporter PlaP (94). There is some evidence that swarming drives urease activity; however, urease genes are not highly expressed during swarming (61, 95). Assaying swarmer cells for urease activity was performed on whole cell suspensions in buffer containing urea (95). With this experimental setup, it is challenging to discern whether urea hydrolysis is driven by urease present in swarmer cells prior to the assay start or by urease induced by the urea-containing medium. Alternatively, the observed increase in urease activity during swarming could be driven by membrane remodeling in swarmer cells (96). Perhaps, swarmer cell membrane components impact the rate at which urea or nickel enters the cytoplasm, thereby altering urease activity. Further work is required to define the relationship between these two major P. mirabilis phenotypes.

Once urea has been produced, catalysis of this metabolite serves two functions. First, urea catalysis prevents the accumulation of intracellular urea, which can be toxic at high levels. Additionally, urea catalysis releases nitrogen that would otherwise be trapped in urea, forming ammonia which can re-enter cellular metabolism. In general, P. mirabilis seems to have more tools for nitrogen cycling than other enteric bacteria. Urea hydrolysis is clearly an important function in P. mirabilis metabolism, and perhaps, urease plays a greater role in nitrogen cycling than is currently appreciated. Further investigation of the role of urease in nitrogen metabolism may yield interesting insights into this organism’s physiology.

In summary, we have expanded the UreR regulon to include nickel import using RNA-seq. We demonstrated that UreR directly binds the promoter of nickel importer Ynt, thereby identifying the first regulator of nickel transport in P. mirabilis. Our investigation into the conservation of UreR identified two instances where the ure locus is linked to the ynt locus, implying that UreR regulation of ynt is a conserved regulatory relationship. Our study did not identify negatively regulated UreR targets. However, we also recognize that we used a single time point, urea concentration, and growth medium to define the UreR regulon. Additional regulatory targets could be identified if these variables are altered. Our findings indicate that UreR selectively activates the cellular activities required to produce mature urease, furthering our understanding of UreR as a regulator of a key P. mirabilis virulence mechanism.

DATA AVAILABILITY

Raw RNA-sequencing reads are available through NIH BioProject accession number PRJNA1071682.

SUPPLEMENTAL MATERIAL

The following material is available online at https://doi.org/10.1128/jb.00031-24.

Figure S1. jb.00031-24-s0001.TIF.

Urea-induced expression of ureR is not detectable by qRT-PCR.

Figure S2. jb.00031-24-s0002.tif.

Supporting data for nitrogen assimilation experiments.

Figure S3. jb.00031-24-s0003.tif.

The first two genes in each nickel transport operon experienced the greatest expression change in response to urea.

Figure S4. jb.00031-24-s0004.tif.

Insertional mutants used in RNA-seq study.

Figure S5. jb.00031-24-s0005.tif.

A schematic of the nik operon promoter.

Figure S6. jb.00031-24-s0006.TIF.

Morganellaceae family phylogeny.

Supplemental legends. jb.00031-24-s0007.docx.

Legends for supplemental figures and tables.

Supplemental tables. jb.00031-24-s0008.xlsx.

Tables S1 to S5.

ASM does not own the copyrights to Supplemental Material that may be linked to, or accessed through, an article. The authors have granted ASM a non-exclusive, world-wide license to publish the Supplemental Material files. Please contact the corresponding author directly for reuse.