Development of a suite of Proteus mirabilis-derived urea-inducible promoters

ABSTRACT

Catheter-associated urinary tract infections (CAUTIs) are a significant burden on healthcare systems, accounting for up to 40% of hospital-acquired infections globally. A prevalent CAUTI pathogen, Proteus mirabilis, is an understudied Gram-negative bacterium. One sequela of P. mirabilis CAUTI is the production of urinary stones, which complicates treatment and clearing of the infection. Stone formation is induced by the activity of urease, a nickel-metalloenzyme that is regulated by UreR in a urea-dependent manner. As urea is abundant in the urinary tract, urease genes are highly expressed during experimental UTI. We sought to leverage the urease promoter to create an expression system that would enable urea-inducible expression of genes during in vitro experiments as well as during experimental UTI. During preliminary studies, we observed unexpectedly high levels of basal expression of the urease promoter. This was somewhat dependent on the presence of regulator UreR. To further develop this expression system, we generated a series of reporter constructs to assess the impact of specific promoter elements on promoter activity in the presence and absence of urea. Elements of interest included known regulatory binding sites, alternative translational start sites, and single-nucleotide polymorphisms identified through comparative genomics. This work describes a suite of urea-inducible promoters, constructed during this study, that exhibit a variety of expression dynamics, providing a customizable platform for gene expression.

IMPORTANCE

Urea is an inexpensive molecule that can easily be supplied during in vitro experiments. A urea-inducible promoter would also be activated by environments where urea naturally occurs, such as in the urinary tract. Thus, the development of a urea-inducible system for selective gene expression is of great interest to the field of uropathogenesis as it would enable selective gene induction during experimental urinary tract infection. This expression system would also have important applications for recombinant protein production in biotech and manufacturing.

INTRODUCTION

In recent years, major advances in the field of bacterial pathogenesis have followed the development of genetic technologies. For example, new mutagenesis methods enable the manipulation of genomes from bacteria that were previously considered intractable (1–5). Additionally, random mutagenesis via transposon insertion potentiated the identification of genes involved in bacterial fitness and virulence (6, 7). However, not every gene can be characterized through deletion or insertional mutagenesis; indeed, essential genes are required for viability and are difficult to study using deletion analysis. Characterizing the function of essential genes requires additional genetic techniques such as transient repression or inducible overexpression of a transcript. Several gene expression platforms exist to support these studies (8–10). However, these expression platforms can be difficult to adapt for use in in vivo models of pathogenesis. Published studies leveraging tetracycline-induced promoters have added this antibiotic to drinking water or food supplied to mice (10). However, this approach may produce secondary effects through altering the mouse microbiome. Furthermore, precise control of tetracycline concentrations within the mouse is not achievable. Thus, there remains a need for additional gene expression systems that respond to inducers that are either easily supplied to animal models or are already present at relevant organ sites.

One approach for developing new inducible promoters to study essential genes is by adapting promoters that exist in nature. Indeed, this is the origin of commonly used expression platforms including the IPTG-inducible lac promoter, arabinose-inducible ara promoter, and tetracycline-responsive promoters (8–10). In this study, we present the development of a urea-inducible expression system based on a promoter in the genome of Proteus mirabilis, a uropathogen.

P. mirabilis, a causative agent of complicated urinary tract infections (UTIs), can induce the formation of urinary stones via the activity of the urease enzyme. Genetic regulation of the urease production operon is well-characterized. The regulator of urease production, UreR, is encoded directly upstream of the urease operon (ureDABCEFG) in a “head to head” orientation (11, 12). These two transcripts are driven by a urea-responsive bidirectional promoter located between ureR and ureD (13). When urea is present, UreR directly binds this metabolite, dimerizes, and binds at two sites within the bidirectional promoter (14). Urea is highly abundant in the urinary tract of mice (~1.2 M), ensuring that urease genes are highly expressed in a commonly used murine model of UTI (15, 16). Furthermore, urea is an inexpensive, readily available chemical that can easily be added to media in a laboratory setting. Thus, the urea-inducible promoter, P_ureD, is an excellent candidate for development as a urea-inducible expression system. For scientists interested in the urinary tract, this expression platform could be used to facilitate in vivo gene expression in this organ system.

During the process of evaluating P_ureD as a platform for driving urea-responsive gene expression in P. mirabilis, a higher than expected promoter activity was observed in the absence of urea. Perhaps this was due to a change in the host species; most of the previous work characterizing P_ureD expression kinetics was examined from a plasmid in an Escherichia coli strain that lacks the regulatory machinery of this promoter’s native P. mirabilis environment (12, 17–19).

In this study, we sought to characterize P_ureD in its native species with the goal of developing a more tightly regulated urea-inducible expression system. In addition to gaining new insights into the regulation of P_ureD, we present a suite of P_ureD-derived promoters that offer reduced basal expression and a range of urea-induced activity levels.

RESULTS

P_ureD exhibits high levels of activity in the absence of urea

The ureR-ureD intergenic region encodes a bidirectional promoter that drives urea-inducible expression of both ureR and ureDABCEFG (Fig. 1A). Urea-dependent induction of this promoter in both directions is facilitated by the transcriptional regulator UreR, which directly binds urea, dimerizes, and activates expression (13, 18–22). P_ureD contains two UreR binding sites that we will refer to as UreR1 and UreR2 throughout this study (14). UreR1 overlaps with the −35 site that facilitates RNA polymerase binding at P_ureR, while UreR2 overlaps with the −35 site of P_ureD (14). This bidirectional promoter also features an H-NS binding site between UreR1 and UreR2. H-NS, a well-established negative regulator generally binds at AT-rich sites across the chromosome (23). This particular H-NS binding site negatively regulates the activity of P_ureR and P_ureD (17, 24). While examining this promoter, we also noticed a putative open reading frame at −73 to −150 in relation to ureD. This putative ORF predicted the synthesis of a 25 amino acid protein with the sequence MQFIFIHTLPNIHFIIFSVILNDII. This putative protein was cloned and FLAG-tagged, but was not detected via immunoblotting. Thus, we were unable to conclude whether this putative protein was translated.

Fig 1.

UreR contributes to basal activity of PureD. (A) Schematic depicting elements of the urease promoter of P. mirabilis. (B) Schematic of reporter constructs pMF2 and pMF6. (C) The promoter activity of each construct was measured via Miller assay after growth to mid-logarithmic phase in buffered medium or buffered medium supplemented with 100 mM urea. Significance was determined via two-way ANOVA with Tukey’s multiple comparisons correction (*P < 0.05; **P < 0.01, ****P < 0.0001). (D) To measure the urea-responsiveness of each construct, the induction ratio (Miller units_induced/Miller units_uninduced) was calculated. pMF2 and pMF6 were assessed in a wild-type background (grey bars) or ureR mutant (white bars). Statistically significant changes in urea-responsiveness were assessed via ordinary Kruskal-Wallis with Dunn’s correction for multiple comparisons (**P < 0.01).

In this study, we sought to develop P_ureD as a urea-inducible expression system that could be leveraged to selectively induce genes in P. mirabilis during experimental UTI. However, previous work to characterize this promoter was performed using plasmid constructs in E. coli, a species that does not typically encode urease and provides a regulatory environment distinct from P. mirabilis. In addition to encoding unique sets of transcriptional regulators, these two organisms exhibit a marked difference in GC-content. E. coli genomes typically contain 50% GC-content, while P. mirabilis type-strain HI4320 contains 39% GC-content (25, 26). This difference is particularly notable for H-NS-regulated P_ureD given that H-NS binds to AT-rich regions. We hypothesized that the kinetics of expression may be different in P. mirabilis.

Toward this end, a reporter construct was generated to measure the activity of this promoter in P. mirabilis. The backbone of this construct was derived from pGEN-MCS, a stable plasmid routinely used for in vivo complementation of P. mirabilis (27, 28). The reporter construct pMF2-lacZYA encoded the 415 bp directly upstream of the ureD coding sequence driving expression of the lactose utilization operon lacZYA (Fig. 1B). The lac operon encodes three genes that enable the import and cleavage of lactose and its analog, o-nitrophenyl-β-galactoside (ONPG): β-galactosidase (lacZ), a permease (lacY), and a transacetylase (lacA). Cleavage of ONPG produces a yellow by-product, enabling the quantification of reporter activity by a widely used colorimetric assay (29). Importantly, P. mirabilis type strain HI4320 does not encode a lac operon in its genome; thus, all observed β-galactosidase activity was directly attributable to the reporter construct (30).

Using this approach, the activity of pMF2-lacYZA when cultured in buffered LB or buffered LB supplemented with 100 mM urea was measured via Miller assay. The use of a buffering agent was necessary as urease-driven urea hydrolysis produces a pH increase due to production of ammonia, and data were normalized to the activity of pMF2-lacZYA cultured in buffered LB. Surprisingly, pMF2-lacZYA exhibited considerable activity when cultured in the absence of urea (Fig. 1C; median raw Miller units = 267). This level of basal activity was not observed when this promoter was studied in E. coli (uninduced range = 19 ± 7 Miller units) indicating that P. mirabilis contains additional factors that impact P_ureD expression that are not found in E. coli (13, 19, 31). This finding underscored the importance of elucidating expression kinetics of a promoter in the context of its native species. This construct was also assessed in a ureR mutant background (ureR::aphA). In the absence of ureR, basal expression of pMF2-lacZYA was significantly reduced, indicating that UreR contributed to the elevated P_ureD activity observed in P. mirabilis without urea induction (Fig. 1C). The induction ratios for this construct in each genetic background were calculated to measure the fold-change in activity in media lacking urea versus containing urea. This metric allowed us to compare the urea-responsiveness of each construct. pMF2-lacZYA in HI4320 yielded a 1.73-fold increase in activity in the presence of urea, while the same construct was no longer responsive to urea when in a ureR mutant (Fig. 1D). These results indicated that UreR contributes to the basal expression observed in the wild-type genetic background.

Even in the absence of UreR, RNA polymerase retained the ability to recognize P_ureD and drive expression. This basal activity was also apparent when wild-type HI4320, a ureR mutant, and a urease null strain were cultured in minimal medium where urea was the sole carbon source (Fig. S1). Disruption of the urease operon was sufficient to produce a marked growth defect that can be rescued by a secondary urea catalysis pathway (the mum operon) in a concentration-dependent manner (32). In comparison, the ureR mutant exhibited a very slight growth defect with the lowest urea concentration (0.1%). Our results indicated that although the ureR mutant produces very little urease enzyme (19), this basal expression was sufficient to support growth in medium where urea was the sole nitrogen source.

Given these results, we concluded that P_ureD required further optimization prior to its use as a urea-inducible expression system. We sought to produce a more tightly regulated urea-inducible expression system by identifying sequence changes that reduce the basal expression of P_ureD. Toward this end, a series of reporter constructs were generated to determine if the basal activity of P_ureD could be reduced while maintaining strong urea-induced activation. As with pMF2-lacZYA, all constructs were derived from pGEN-MCS.

First, we hypothesized that the inclusion of the RNA polymerase binding sites in P_ureR may impact basal expression. Although activation of the partial P_ureR present in pMF2-lacZYA was incapable of producing its native coding transcript, it could provide a second binding site for transcription initiation machinery, thus artificially increasing the local concentration of proteins required for initiating transcription in the ureD direction. To address this hypothesis, we generated reporter construct pMF6-lacZYA, which encoded the −308 to −1 region of P_ureD upstream of lacZYA (Fig. 1B). Using the same reporter assay, the promoter strength of pMF6-lacZYA was measured in wild type and a ureR mutant. pMF6-lacZYA exhibited significantly lower levels of basal expression than pMF2-lacZYA in both wild type (66 vs 100 relative Miller units, P = 0.0009) and the ureR mutant (69 vs 36 relative Miller units, P = 0.0012, Fig. 1C). Furthermore, pMF6-lacZYA retained the same level of activity as pMF2-lacZYA in the presence of urea (Fig. 1C). Although the induction ratio increased from 1.7-fold to 2.4-fold with the deletion of the ureR transcriptional start site, this was not a statistically significant difference (Fig. 1D). Thus, pMF6-lacZYA successfully reduced basal activity in comparison to pMF2-lacZYA while maintaining a robust response to urea.

The UreR2 binding site was necessary and sufficient to confer urea-inducibility

Several regulatory elements have been noted in P_ureD (Fig. 1A). To characterize additional sequence elements that modulate the activity of the pMF6-lacZYA reporter, a series of reporter constructs were generated for assessment via Miller assay. These elements include two UreR binding sites (UreR1 and UreR2) that are separated by an H-NS binding site. A reporter construct lacking all three known regulatory sites, pMF6∆Reg-35-lacZYA, was produced (Fig. 2A). A reporter that replaced the H-NS binding site with a spacer sequence was also generated. Preserving the spacing between the two UreR binding sites allowed us to specifically distinguish between altered reporter activity driven by the H-NS binding site or a spacing requirement between the two UreR binding sites (Fig. 2A; pMF6∆H-NS::spacer-lacZYA). As is typical for this regulator, the H-NS binding site in P_ureD is an AT-rich sequence (35 nt, 89% AT); designing the spacer via scrambling the existing sequence did not produce a marked change in the sequence. Instead, a scrambled spacer the same size as the H-NS binding site was designed to reflect the frequencies of each nucleotide within the ureR-ureD intergenic region (35 nt, 56% AT). Finally, pMF6-based reporter constructs with selective deletions of each individual UreR binding site were produced (Fig. 2A; pMF6∆UreR1-lacZYA and pMF6∆UreR2-35-lacZYA). As with pMF6∆Reg-35- lacZYA, pMF6∆UreR2-35-lacZYA featured a partial deletion that eliminated UreR binding at this locus but preserved the −35 site. As in Fig. 1, data were normalized to the activity of pMF2-lacZYA cultured in buffered LB; moving forward, bars depicting the normalizing condition were excluded from graphs to avoid redundancy.

Fig 2.

UreR binding sites play distinct roles in urease expression. (A) Schematic depicting reporter constructs pMF6-lacZYA, pMF6∆UreR1-lacZYA, pMF6∆UreR2-35-lacZYA, pMF6∆H-NS::spacer-lacZYA, and pMF6∆Reg-35-lacZYA. (B) The individual and cumulative effects of regulator binding sites on promoter activity was assessed via Miller assay after growth to mid-logarithmic phase (OD₆₀₀ = 0.5–0.9) in buffered medium or buffered medium supplemented with 100 mM urea. Data were normalized to the activity of pMF2-lacZYA cultured in buffered medium. Significance was determined via two-way ANOVA with Tukey’s multiple comparisons correction (*P < 0.05; **P < 0.01, ****P < 0.0001). (C) To measure the urea-responsiveness of each construct, the induction ratio was calculated. Statistical significance was assessed via ordinary Kruskal-Wallis with Dunn’s correction for multiple comparisons (*P < 0.05, **P < 0.01, ***P < 0.001).

As expected, pMF6∆Reg-35-lacZYA exhibited significantly lower activity than pMF6-lacYZA in both LB (19 vs 67 relative Miller units, P = 0.0002) and LB supplemented with the inducer urea (18 vs 174 relative Miller units, P < 0.0001, Fig. 2B). Deletion of the entire regulatory region abrogated the urea-responsiveness of P_ureD (Fig. 2C). We also evaluated pMF6∆Reg-35-lacZYA in a mutant lacking ureR. Again, pMF6∆Reg-35-lacZYA was significantly less active than pMF6-lacYZA in both LB (18 vs 67 relative Miller units, P < 0.0001) and LB supplemented with the inducer urea (15 vs 174 relative Miller units, P < 0.0001, Fig. 2B). Thus, deletion of the regulatory binding sites dramatically reduced activity, regardless of the presence of urea or UreR. This construct provided a useful baseline for determining the relative contributions of the UreR1, UreR2, and H-NS binding sites to expression driven by P_ureD. Additionally, our results did not indicate the presence of additional UreR binding sites.

Next, we evaluated the contribution of the H-NS binding site to P_ureD. Published data indicate that H-NS is a negative regulator of P_ureR (17, 24). Since activity of P_ureR drives the expression of the urease regulator, H-NS indirectly regulates urease production. The role of H-NS in specifically regulating P_ureD has not been characterized. Previous work also determined that UreR can dislodge H-NS bound to P_ureR (17). Interestingly, the activity of pMF6∆H-NS::spacer-lacZYA in our reporter assay phenocopied pMF6∆Reg-35-lacZYA (Fig. 2B and C). This was rather unexpected; H-NS is a well-characterized negative regulator, and disrupting H-NS binding was expected to increase promoter activity.

Data examining the UreR binding sites indicated that each site played different roles in the transcriptional regulation of P_ureD. The activity of pMF6∆UreR1-lacZYA was not significantly different from the activity of pMF6-lacZYA, regardless of whether the medium contained urea or not (Fig. 2B). Both pMF6-lacZYA and pMF6∆UreR1-lacZYA exhibited a similar increase in activity in the presence of urea (2.7-fold and 2.4-fold, respectively; Fig. 2C). Thus, UreR1 was dispensable for conferring urea-inducibility to P_ureD. Conversely, pMF6∆UreR2-35-lacZYA exhibited significantly lower basal levels of P_ureD (Fig. 2B), indicating that this binding site contributed to basal expression. Furthermore, pMF6∆UreR2-35-lacZYA did not recognize urea as an inducing cue (Fig. 2B and C). The combined observations of both UreR deletion constructs indicated that the UreR2 binding site was both necessary and sufficient for conferring urea-inducibility to P_ureD. Based on these data, we hypothesized that the UreR1 binding site contributes to urea-induced activation P_ureR rather than P_ureD. These are the first data demonstrating that these two binding sites play distinct regulatory roles in this bidirectional promoter.

Alternate start codon usage impacts reporter activity

Upon a close examination of P_ureD, we also noticed three in-frame alternate start codons for ureD at 6, 15, and 33 nucleotides upstream of the predicted translational start codon of ureD, which has not been experimentally verified. To assess the impact that forced use of these alternate start codons would have on P_ureD activity, two groups of reporter constructs were generated (Fig. 3A). In the first group, the ATG of lacZ was moved to the location of the alternate start codon (pMF6∆6-lacZYA, pMF6∆15-lacZYA, and pMF6∆33-lacZYA). This allowed us to functionally assess each locus for the presence of DNA elements required for recognition by the P. mirabilis translational machinery (e.g., Shine-Dalgarno sequence). Constructs in the second group forced the use of these alternate start sites (pMF6∆6-GTG-lacZYA, pMF6∆15-TTG-lacZYA, and pMF6∆33-TTG-lacZYA). Pairing these two sets of constructs enabled us to distinguish between alternate codons that were nonfunctional due to their genomic context from those that were nonfunctional due to the alternate start codon itself (TTG or GTG).

Fig 3.

Alternate start codon usage impacts promoter activity. (A) Schematic of pMF6 and pMF6-derived reporter constructs driving expression from alternate translational start positions at –6, –15, and −33 using ATG (∆6, ∆15, and ∆33) or the native alternate start codon (∆6-GTG, ∆15-TTG, and ∆33-TTG). (B and C) The promoter strength of constructs depicted in (A) was assessed via Miller assay after growth to mid-logarithmic phase in buffered medium or buffered medium supplemented with 100 mM urea. Data were normalized to the activity of pMF2-lacZYA cultured in buffered medium. Statistical significance was determined using two-way ANOVA with Tukey’s multiple comparisons correction (*P < 0.05; **P < 0.01, ****P < 0.0001). (D and E) The induction ratio for each construct was quantified to compare the urea-inducibility of the reporter constructs. Significance was determined via (D) Kruskal-Wallis with Dunn’s correction for multiple comparisons (*P < 0.05, ***P < 0.001) or (E) ordinary one-way ANOVA with the Dunnett correction for multiple comparisons (****P < 0.0001). To determine whether the increased activity of pMF6∆6-GTG-lacZYA was driven by increased transcription or translation, we measured the expression of lacZ via qRT-PCR (F) and the production of LacZ via Western blotting (G). Samples were cultured in the same conditions as in (B and C). qRT-PCR data were normalized to lacZ transcript from pMF6-lacZYA cultured in buffered medium and to housekeeping gene rpoA. Expression was assessed as significantly upregulated or downregulated compared to pMF6-lacZYA using one-sample t test with a hypothetical mean of 1 (*P < 0.05; **P < 0.01). Comparative analysis was calculated using the unpaired t test (**P < 0.01). In (G), a cross-reactive protein bound by α-LacZ served as a loading control.

We measured the promoter activity of all constructs in LB and LB supplemented with urea. Evaluation of pMF6∆6-lacZYA, pMF6∆15-lacZYA, and pMF6∆33-lacZYA indicated that there was variability in the ability of the alternate start codon loci to be recognized for translation (Fig. 3B). pMF6∆6 exhibited significantly lower activity compared to pMF6-lacZYA in both LB (16 vs 67 relative Miller units, P = 0.0017) and in LB supplemented with urea (57 vs 174 relative Miller units, P < 0.0001). This did not correspond to a significant change in the urea-inducibility of the construct (Fig. 3D). Conversely, the −15 and −33 sites were easily recognized by the translation machinery. pMF6∆15-lacZYA exhibited reduced basal activity compared to pMF6 (31 vs 67 relative Miller units; P = 0.0440), while pMF6∆33-lacZYA was not significantly different (Fig. 3B). Both pMF6∆15-lacZYA and pMF6∆33-lacZYA maintained urea-induced activity and exhibited significantly increased urea-induction ratios (Fig. 3D). These data indicated that the basal activity of pMF6 can be further reduced by using an ATG start codon to initiate translation directly after −15 or −33.

Interestingly, the phenotypes of the group of constructs that forced translation using the alternate translational start codons were reversed. Converting the translational start codon from ATG to TTG significantly reduced the basal activity of pMF6∆15-TTG-lacZYA and pMF6∆33-TTG-lacZYA in comparison to pMF6-lacZYA (7 and 12 vs 67 relative Miller units, Fig. 3C). Decreased translation from the TTG codon also produced a significant reduction in promoter activity when induced with urea (Fig. 3C). The induction ratios for these constructs were also altered; both pMF6∆15-TTG-lacZYA and pMF6∆33-TTG-lacZYA exhibited significantly higher induction ratios compared to pMF6-lacZYA (5-fold and 7-fold vs 2.5; Fig. 3E). Although the −15 and −33 translational start sites featured functional ribosome binding sites (as measured with pMF6∆15-lacZYA and pMF6∆33-lacZYA), the TTG start codon was poorly recognized by the ribosome at these loci. Conversely, forced use of GTG as a start codon at the −6 site enhanced reporter activity in both induced and uninduced conditions (Fig. 3C). pMF6∆6-GTG-lacZYA exhibited significantly increased activity compared to pMF6-lacZYA in LB (166 vs 67 relative Miller units) and LB supplemented with 100 mM urea (637 vs 174 relative Miller units). Despite the marked increase in activity, the induction ratios were not significantly different (Fig. 3E). Taken together, the results in Fig. 3B and C indicate that while each alternate start site locus can support translation, the overall magnitude of translational activity is dependent on the start codon.

We were surprised that a single nucleotide change that converted the start codon from ATG (pMF6∆6) to GTG (pMF6∆GTG) produced such a dramatic difference in promoter activity. Next, we sought to determine the mechanism driving the enhanced promoter activity. Toward this end, qRT-PCR was performed to measure lacZ transcript levels and immunoblotting to measure LacZ protein production. Given that pMF6∆6-GTG-lacZYA alters the translational start codon of P_ureD, increased promoter activity was expected to be driven by increased protein production. In accordance with data from our reporter assay, pMF6∆6-lacZYA exhibited significantly lower lacZ expression compared to pMF6-lacZYA in media lacking urea (P = 0.0160, Fig. 3F). Additionally, the qPCR data clearly demonstrate that pMF6∆6-GTG-lacZYA produced more lacZ transcript than pMF6∆6-lacZYA in medium containing urea (P = 0.0003, Fig. 3F). This expression pattern corresponded to the LacZ protein levels observed in each of these strains. pMF6∆6-GTG-lacZYA basal promoter activity produced more LacZ compared to pMF6-lacZYA and pMF6∆6-lacZYA (Fig. 3G; lanes 3, 5, and 7). This increased LacZ production was also observed when urea was supplied to induce promoter activity (Fig. 3G; lanes 4, 6, and 8). Thus, increased mRNA transcript drove the increased promoter activity observed in pMF6∆6-GTG-lacZYA in both media conditions.

The ureD promoter exhibits little variation among P. mirabilis isolates

After characterizing existing sequence features of P_ureD that can alter this promoter’s activity, we were curious if sequence homology could be used to identify naturally occurring single-nucleotide polymorphisms (SNPs) that impact expression. To aid in this endeavor, sequences of the ureR-ureD intergenic region were procured from 470 strains with high-quality genomes publicly available through the Bacterial and Viral Bioinformatics Resource Center (BV-BRC), (Table S1). Approximately 66% of strains were isolated from humans (309/470), although 15 strains lacked an identifiable isolation source.

Next, a multiple sequence alignment of these sequences was generated using Clustal Omega (33). To visualize the alignment, a Sequence Logo graph of the ureR-ureD intergenic region was produced (Fig. S2). This 493 bp region was highly conserved among P. mirabilis strains, and a position frequency matrix was used to locate SNPs (Table S2). To increase the likelihood of selecting SNPs that were not the result of sequencing error, we narrowed our focus to the most abundant SNPs. Six out of 15 SNPs occurred in at least 2% of strains (10/470), which were selected for further evaluation. Using this method, six unique SNPs in the ureR-ureD intergenic region were selected for further evaluation (Table 1). Integrating SNP loci and strain information enabled the categorization of the abundant SNPs into three distinct groups: (i) co-occurring [T(−493)C, G(−308)A, T(−149)C, and A(−146)G] mutations (n = 16); (ii) a T(−91)A mutation (n = 80); and (iii) a G(−16)T mutation (n = 15). Since the aim of this approach was to identify single nucleotide loci that impact P_ureD activity, we focused on the two SNPs found in isolation (Fig. 4A through C). These two SNPs were primarily found in human isolates, increasing our interest in them (Group 2, 15/15; Group 3, 50/80). Conversely, a minority of isolates carrying the co-occurring SNPs identified in Group 1 were cultured from humans (6/15). Notably, the HI4320 ureR-ureD intergenic region does not encode either of the two SNPs selected for further study.

TABLE 1.

Strains used in this study

Strain	Species	Description	Selection	Reference
HI4320	P. mirabilis	Wild type		(26)
ureR::bla	P. mirabilis	Insertional mutant in ureR constructed in HI4320	kan	(21)
ureC::aphA	P. mirabilis	Insertional mutant in ureC constructed in HI4320	amp	(34)
TOP10	E. coli	Strain for routine cloning		Invitrogen

Fig 4.

P. mirabilis SNP confers decreased response to urea. (A) Schematic depicts reporter constructs pMF6-lacZYA, pMF6-G(−16)T-lacZYA, and pMF6-T(−91)A-lacZYA. (B and C) SeqLogo graphs depict the relative abundance of nucleotides at each SNP locus in 470 P. mirabilis genomes. (D) The effect of each SNP on promoter activity was assessed via Miller assay after growth to mid-logarithmic phase (OD₆₀₀ = 0.5–0.9) in buffered medium or buffered medium supplemented with 100 mM urea. Data were normalized to the activity of pMF2-lacZYA cultured in buffered medium. Statistical significance was determined using two-way ANOVA with Tukey’s multiple comparisons correction (***P < 0.001). (E) SNPs had no effect on the urea-responsiveness of each promoter as calculated by ordinary one-way ANOVA with Dunnett’s multiple comparisons test.

To evaluate the effects of each SNP on P_ureD activity14, site-directed mutagenesis was performed on pMF6-lacZYA at sites −16 (G > T) and −91 (T > A) to produce pMF6-G(−16)T-lacZYA and pMF6-T(−91)A-lacZYA. The effects of each point mutation on promoter strength were measured via reporter assay. In this assay, pMF6-G(−16)T-lacZYA exhibited similar levels of basal activity as pMF6-lacZYA (63 vs 67 relative Miller units, Fig. 4D). Similarly, the activity in medium containing urea was not significantly different from pMF6-lacZYA (156 vs 174 relative Miller units, Fig. 4D). As expected, the induction ratios were equivalent (Fig. 4E). However, the T > A mutation at −91 produced a different outcome; pMF6-T(−91)A-lacZYA exhibited slightly reduced levels of basal activity compared to pMF6-lacZYA (54 vs 67 relative Miller units, Fig. 4D), and significantly reduced activity in medium containing urea (116 vs 174 relative Miller units, Fig. 4D). The reduction in urea-driven activity was insufficient to produce a significantly reduced induction ratio (Fig. 4E).

We successfully leveraged sequence conservation to identify functionally relevant SNPs within the ureD promoter. Of the two SNPs tested, only the higher frequency SNP altered the activity of P_ureD. Additional research is required to determine if the T(−91)A SNP meaningfully alters the virulence of isolates that carry it.

Conservation of the ureD promoter among UreR-encoding species

Our previous work identified 14 additional Morganellaceae family species that encode a UreR-regulated urease locus (35). Despite sharing this locus that is associated with uropathogenesis, P. mirabilis is the most predominant uropathogen within this group. Given that P. mirabilis pathogenesis is tightly linked to its ability to produce urease, we next sought to identify SNPs specific to P. mirabilis. The sequence encoding the ureR-ureD intergenic region was obtained from the reference genomes for each species and performed a multiple sequence alignment. The sequences from each species were visualized using the NCBI Multiple Sequence Alignment Viewer, in which each horizontal bar represents the sequence between ureR and ureD in each species (Fig. 5A). It was immediately apparent that the urease bidirectional promoter varies between species. Most of the Morganellaceae species that encode a UreR-regulated urease locus belonged to the Proteus genus (11/15). Sequences from the Providencia and Cosenzaea genera encode an increased number of less abundant nucleotides, which was reflective of both the overrepresentation of species from the Proteus genus in this alignment and the evolutionary distance from Providencia and Cosenzaea to Proteus. In general, the Proteus and Cosenzaea urease promoter sequences were longer than those from Providencia.

Fig 5.

Identification of Morganellaceae SNPs within P_ureD. The intergenic region between ureR and ureD from 15 Morganellaceae species that encode a UreR-regulated urease locus were aligned using Clustal Omega. The resulting alignment was viewed in NCBI’s MSA Viewer and colored to highlight variations from the consensus sequence in red. Gaps in coverage are depicted in white. Below the alignment schematic, the frequencies of nucleotides within the regulatory region of this bidirectional promoter are depicted using a SeqLogo graph. The HI4320 sequence is depicted underneath the SeqLogo, and loci where HI4320 differs from the consensus sequence are emphasized in underlined, bold text. The −10 and −35 sites facilitating RNA polymerase binding (white), UreR binding sites (yellow), and H-NS binding sites (blue) are also depicted. SNPs within the UreR2 binding site selected for further analysis are indicated with arrows.

To gain a more granular view of sequence conservation across these 15 species, a SeqLogo graph depicting the relative frequency of nucleotides at each position was generated (Fig. S3). As expected, the region that encodes the regulatory elements of the urease promoter exhibited stronger conservation than other regions (Fig. S3; Fig. 5B). Within this 132-nucleotide region, P. mirabilis encoded a less abundant nucleotide at 21 loci (Fig. 5B). Two of these SNPs were within the UreR2 binding site, which has been demonstrated as necessary and sufficient to confer urea-inducibility to P_ureD (Fig. 2). Additionally, our work indicated that UreR also drives the basal activity of P_ureD. Due to their location within the UreR2 binding site, we hypothesized that these two SNPs alter the activity of P_ureD. Thus, these loci were selected for functional characterization.

To characterize these SNPs, site-directed mutagenesis was performed on pMF6-lacZYA to produce plasmids where each locus was independently converted from the P. mirabilis wild-type nucleotide to the most abundant Morganellaceae nucleotide at each site. This produced two constructs to evaluate via reporter assay: pMF6-C256T-lacZYA and pMF6-G265A-lacZYA (Fig. 6A). The promoter activity of pMF6-C256T-lacZYA was not significantly different than pMF6-lacZYA in either LB or LB supplemented with 100 mM urea (Fig. 6B). However, pMF6-G265A-lacZYA did exhibit altered activity. The activity of pMF6-G265A-lacZYA was significantly higher than pMF6-lacZYA, but only when urea was supplied (174 vs 203 relative Miller units). These results suggest that an “A” at residue 265 increased the ability of UreR to bind the UreR2 site, thereby increasing urea-induced promoter activity. Although the increased activity was statistically significant, it did not produce a significant change in the induction ratio (Fig. 6C).

Fig 6.

Morganellaceae SNP enhances urea-responsive activity of P_ureD. (A) Schematic depicting reporter constructs pMF6-lacZYA, pMF6-C256T-lacZYA, and pMF6-G265A-lacZYA. (B) The effect of each SNP on promoter activity was assessed via Miller assay after growth to mid-logarithmic phase (OD₆₀₀ = 0.5–0.9) in buffered medium or buffered medium supplemented with 100 mM urea. Data were normalized to the activity of pMF2-lacZYA cultured in buffered medium. Significance was determined via two-way ANOVA with Tukey’s multiple comparisons correction (*P < 0.05). (C) SNPs had no effect on the urea-responsiveness of each promoter as calculated by ordinary one-way ANOVA with Dunnett’s multiple comparisons test.

DISCUSSION

Inducible gene expression systems have been leveraged for many applications, from recombinant protein production to studying essential genes (10, 36, 37). Most of these expression platforms have been adapted from bacterial promoters that respond to a specific environmental cue such as lactose (P_lac) or arabinose (P_BAD). In this study, we aimed to develop a promoter from P. mirabilis into a urea-responsive expression system. In contrast to existing inducible promoters, a P_ureD-based promoter responds to an environmental cue that can be easily leveraged in experimental models of pathogenesis. A urea-responsive expression platform is particularly useful for studying the urinary tract and could provide selective induction of genes in in vivo UTI models. Urea could also serve as a readily available, cost-effective inducer in the context of pharmaceutical manufacturing.

One research group has already explored the use of the ureD promoter for recombinant protein production in E. coli (38). Their study paired severely abbreviated version of P_ureD that lacked much of the 5′ untranslated region (−167 to −1) with the ribosome binding site of the pET-15b vector to drive protein expression (38). Their promoter construct also included the ureR coding region to enable urea-inducible expression in E. coli. The authors demonstrated the application of this promoter to produce a variety of protein products with a focus on protein quantity. They primarily controlled protein production through modulating the concentration of urea (38). Our study provides an expanded set of urea-inducible promoters that exhibit a variety of expression kinetics in response to 100 mM urea. This concentration of urea was sufficient to induce maximum urease activity (13, 39). The expression kinetics of our constructs could be further modified by inducing lower concentrations of urea. In in vitro studies of urease-producing bacteria, the concentration of urea selected for induction must be balanced with sufficient buffering to prevent a urease-driven pH increase. Modulating the magnitude of urea-induced activity through manipulating the promoter sequence can facilitate appropriate buffering and enables the use of this expression system in environments where the urea concentration cannot be precisely controlled, such as the in urinary tract.

During our studies, we also identified a putative open reading frame in the 5′ untranslated region of P_ureD. This putative ORF was cloned and tagged in an attempt to determine if it is translated into protein via immunoblot. Although the putative ORF could not be detected, perhaps this small, putative protein is only translated under specific environmental conditions. Alternatively, the short putative coding region may encode a regulatory RNA or simply be an untranslated region of the mRNA transcript. If the latter is true, P_ureD provides another example of P. mirabilis encoding, on average, longer 5′ untranslated regions than other species such as E. coli and Klebsiella pneumoniae (25–35 bp) (40–43). Additionally, P. mirabilis encodes regulatory elements that are hundreds of nucleotides from the protein-coding sequences they regulate, indicating that this species may also prefer extended promoters (41, 42, 44).

At the start of our work developing P_ureD into a urea-inducible expression system, the expression kinetics of P_ureD were thought to be firmly established based on several studies performed in E. coli (12, 17–19). However, it quickly became apparent that the activity of this promoter differs in E. coli as compared to its native species, P. mirabilis. In contrast to E. coli, where P_ureD exhibited very little activity in media lacking urea, P_ureD exhibited a substantial level of expression that is independent of urea-induction in P. mirabilis. A further search of the literature provided additional evidence for differential P_ureD activity in E. coli versus P. mirabilis. One study leveraged a UreD-GFP translational fusion to monitor the progression of a P. mirabilis experimental UTI (31). Immunoblotting revealed that, unlike E. coli, P. mirabilis produced UreD-GFP in vitro in media lacking urea (31). The quantity of UreD-GFP induced by 500 mM urea was markedly higher in E. coli than in P. mirabilis (31). Indeed, UreD-GFP production in P. mirabilis at 50 mM urea was not markedly different from production at 500 mM urea (31). Given the differences in expression kinetics between P. mirabilis and E. coli, the promoters we present in this study may require further optimization for use in other organisms. Importantly, ureR must be supplied via plasmid to enable urea-inducible expression in species that do not encode a chromosomal ureR.

Further work characterizing the growth kinetics of wild-type P. mirabilis HI4320 with a ureR mutant (ureR::aphA) and a urease-null strain (ureC::bla) indicated that the basal P_ureD activity in ureR::aphA was sufficient to support robust growth in a minimal medium where urea was the sole nitrogen source (Fig. S3). Conversely, a strain that cannot synthesize urease (ureC::bla) exhibited a significant lag. This phenotype was rescued with an increased concentration of urea, likely a consequence of a second urea catabolism pathway encoded by the mum operon (PMI3556-PMI3559). The mum pathway is energy-dependent and requires two enzymes (32). The observed growth phenotypes also imply that the mum system is a less efficient urea hydrolysis mechanism compared to urease, as the ureR mutant could grow easily on 0.1% urea while the urease null strain exhibited a significant lag (~10 h). However, mum does not appear to be active during UTI as this operon was not highly expressed during experimental UTI, nor was it identified as a fitness factor via transposon-sequencing (7, 15).

This basal activity was only in part driven by UreR. Although the deletion of ureR produced a significant decrease in promoter activity (Fig. 1C), 70% of urea-independent activity remained. Furthermore, our growth kinetic studies indicated that a ureR mutant exhibits sufficient basal expression of urease genes to support growth in a medium where urea is the sole nitrogen source. Thus, additional factors contribute to the basal activity of P_ureD. With this knowledge, we set out to identify other promoter elements that modulate P_ureD activity with the goal of developing more tightly regulated urea-inducible expression system. Ideally, this system would exhibit minimal basal expression while retaining robust urea-induced expression. To achieve this goal, a series of reporter constructs was generated, and their activities in the presence and absence of urea were evaluated (Table 2).

TABLE 2.

Summary of urea-inducible promoter phenotypes

Promoter	Basal activity (% of pMF6a)	Urea-induced activity (% of pMF6b)	Induction ratioc
pMF6	100%	100%	2.5
pMF6∆6	24%	33%	3.6
pMF6∆15	46%	85%	4.9
pMF6∆33	57%	94%	4.3
pMF6∆6-GTG	246%	366%	3.5
pMF6∆15-GTG	8%	23%	5.3
pMF6∆33-GTG	12%	43%	6.5
pMF6-T(−91)A	80%	67%	2.2
pMF6-G(270)A	118%	117%	2.6

^a Calculated by dividing the average Miller units of modified promoter in LB by average Miller units of pMF6 in LB.

^b Calculated by dividing the average Miller units of modified promoter in LB + urea by average Miller units of pMF6 in LB + urea.

^c Calculated by dividing the average Miller units in LB by average Miller units in LB + urea for each promoter.

Several reporter constructs exhibited reduced basal activity of a P_ureD-derived promoter. Notably, P_ureD is one-half of a bidirectional promoter between ureR and ureD. The RNA polymerase binding sites of P_ureR and P_ureD are separated by 71 nucleotides and drive transcription outward from each end of the bi-directional promoter. Surprisingly, removal of the P_ureR RNA polymerase binding site significantly reduced the basal activity of our P_ureD-derived promoter despite lacking the appropriate orientation to drive expression from P_ureD (Fig. 1C). Perhaps, a second RNA polymerase binding site increases expression by helping transcription initiation machinery to localize to this promoter. Alternatively, perhaps this sequence impacts the local three-dimensional structure of DNA, thereby altering expression.

Next, we evaluated the known binding sites for transcriptional regulators by further mutating pMF6-lacZYA. Within P_ureD lies two UreR binding sites, which are separated by an H-NS binding site. H-NS recognizes sequences where high AT content induces a slight bend to the DNA; H-NS binding exaggerates this bend to fold DNA back on itself and prevent transcription (23). Based on the extensive literature characterizing H-NS as a negative regulator, we were surprised when deletion of the H-NS binding site reduced promoter activity. Perhaps the UreR binding sites extend into the H-NS binding site, which could facilitate the ability of these two regulators to displace each other when bound to the urease promoter. Alternatively, perhaps the inherent curvature in the AT-rich site contributes to P_ureD recognition by UreR.

Analysis of our reporter constructs determined that the ureD-proximal UreR binding site (UreR2) is both necessary and sufficient for urea-responsive regulation of P_ureD by UreR, while the ureR-proximal binding site (UreR1) is dispensable. With some transcriptional regulators, supplying additional binding sites can increase the magnitude of the expression change driven by that regulator. These data indicate that this may not be the case with UreR.

During an examination of P_ureD, three alternate translational start sites that would provide in-frame translation at the –6, –15, and −33 positions were identified (Fig. 3A). Reporter assays demonstrated that all three of these loci were capable of supporting translation initiation, although it remains unclear whether these alternate start sites are used by P. mirabilis in nature. Use of these codons would add 2, 5, or 11 amino acids to the N-terminus of UreD, potentially modulating its function or ability to interact with other proteins. AlphaFold 3 produced a “low” or “very low” confidence model of the first seven amino acids of UreD, which is not altered by including any amino acids that would result from alternate codon usage (45, 46). Further studies of the structure and function of UreD may provide additional insights into the role of these alternate start codons.

Characterizing the activity of the observed alternate transcriptional start sites yielded urea-responsive promoters with varying expression dynamics. pMF6∆15-lacZYA and pMF6∆33-lacZYA, two constructs that move the ATG codon of lacZ to the locus of each alternate start codon, exhibited reduced basal activity of the promoter while maintaining the same magnitude of urea-induced expression. Thus, these two promoters adapted from P_ureD provide reduced background expression without sacrificing the magnitude of urea-induction (Table 1). Interestingly, forced use of the naturally occurring start codon at the −15 and −33 sites (pMF6∆15-GTG-lacZYA and pMF6∆33-GTG-lacZ) dramatically reduced translation in both media conditions relative to pMF6-lacZYA. This phenotype was also exhibited by pMF6∆6-lacZYA, a construct that drives translation from the −6 alternate start codon locus using an ATG codon. These three promoter constructs would be appropriate for applications requiring a more subtle change in expression in an in vivo model, perhaps for the study of transcriptional regulators or recombinant proteins that are toxic when overproduced. Forcing the use of the alternate start site at −6 yielded a third phenotype; the promoter activity was significantly enhanced in both media conditions. This change also produced a 3.5-fold increase in urea-induced activity compared to pMF6-lacZYA. This increased activity resulted from an increased transcript, perhaps due to an increase in transcript stability. Thus, the pMF6∆6-GTG promoter may be suitable for producing large quantities of recombinant protein.

Our work to develop this promoter into a urea-inducible expression system revealed that single nucleotide changes can profoundly impact expression. The pair of promoters for each alternate translational start locus vary by a single nucleotide within the translational start codon and yet exhibit unique expression kinetics. Furthermore, we identified functional SNPs that altered gene expression using sequence conservation studies. Using this approach, SNPs that reduce [T(−91)A] or enhance [G(270)A] expression specifically in response to urea were identified (Table 1). Interestingly, only one of those SNPs lies within a UreR binding site. Further work is required to determine the mechanism driving expression changes in response to single nucleotide changes. Perhaps these changes impact transcript stability or impact transcript recognition by the ribosome. Additionally, expanding the number of genomes used for this analysis could identify more functional SNPs within the ureD promoter. It may be worth revisiting this analysis as more genomes of lesser-studied Morganellaceae species become available.

In summary, we present a suite of urea-responsive promoters derived from the urease promoter of P. mirabilis (Table 1). Several of these promoters exhibited low activity in the absence of urea induction but maintained maximal induction when urea was added. Furthermore, we identified alternate translational start sites and single nucleotide changes that impact the overall output from the promoter as well as the magnitude of the response to urea. We anticipate that generating a suite of urea-responsive promoters with varying degrees of activity and urea-inducibility will provide a customizable platform for inducible expression system suitable for many applications. We are particularly excited about its potential as a tool for investigating P. mirabilis and other uropathogens. The study of P. mirabilis has been hampered by technical challenges in genome manipulation, and our study provides a new approach that may yield further insights into this organism’s biology. Furthermore, a urea-inducible expression system can be adapted to provide selective gene expression in in vivo models of uropathogenesis.

MATERIALS AND METHODS

Strains and media

Bacterial strains and constructs used in this study are described in Tables 3 and 4. P. mirabilis prototypical type strain HI4320 was isolated from a patient with a long-term indwelling catheter (47). Bacteria were routinely cultured at 37°C with aeration (200 RPM) 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 at 37°C. Some experiments were performed in Minimal A medium; per liter: 10.5 g of K₂HPO₄, 4.5 g of KH₂PO₄, 0.47 g of sodium citrate, and 1.0 g of (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 (48). As needed, liquid medium was supplemented with 250 mM HEPES or 50 µg/mL ampicillin, and solid medium was supplemented with 100 µg/mL ampicillin.

TABLE 3.

Plasmids used in this study

Plasmid	Description	Antibiotic marker	Reference
pGEN-MCS		Ampicillin	(49)
pMF2	pGEN-MCS backbone encoding the −416 to −1 section of the ureD promoter	Ampicillin	This study
pMF2-lacZYA	pMF2 driving expression of lacZYA	Ampicillin	This study
pMF6-lacZYA	pGEN-MCS backbone encoding the −308 to −1 section of the ureD promoter driving lacZYA expression	Ampicillin	This study
pMF6∆6-lacZYA	pMF6 modified to encode −308 to −7 section of the ureD promoter driving lacZYA expression from an ATG codon	Ampicillin	This study
pMF6∆6-TTG-lacZYA	pMF6 modified to encode −308 to −7 section of the ureD promoter driving lacZYA expression from a TTG codon	Ampicillin	This study
pMF6∆15-lacZYA	pMF6 modified to encode −308 to −16 section of the ureD promoter driving lacZYA expression from an ATG codon	Ampicillin	This study
pMF6∆15-TTG-lacZYA	pMF6 modified to encode −308 to −16 section of the ureD promoter driving lacZYA expression from a TTG codon	Ampicillin	This study
pMF6∆33-lacZYA	pMF6 modified to encode −308 to −34 section of the ureD promoter driving lacZYA expression from an ATG codon	Ampicillin	This study
pMF6∆33-GTG-lacZYA	pMF6 modified to encode −308 to −34 section of the ureD promoter driving lacZYA expression from a GTG codon	Ampicillin	This study
pMF6∆UreR1-lacZYA	pMF6 modified by deleting the ureR-proximal UreR binding site	Ampicillin	This study
pMF6∆UreR2-35-lacZYA	pMF6 modified by deleting the ureD-proximal UreR binding site while preserving the −35 site	Ampicillin	This study
pMF6∆H-NS::spacer-lacZYA	pMF6 modified by replacing the H-NS binding site with a random spacer sequence	Ampicillin	This study
pMF6∆Reg-35-lacZYA	pMF6 modified by deleting UreR and H-NS binding sites while preserving the −35 site	Ampicillin	This study
pMF6-G(−91)T-lacZYA	pMF6 with a point mutation at the −91 site	Ampicillin	This study
pMF6-T(−16)A-lacZYA	pMF6 with a point mutation at the −16 site	Ampicillin	This study
pMF6-C261T-lacZYA	pMF6 with a point mutation at the −238 site, which corresponds to position 261 in the Morganellaceae promoter alignment	Ampicillin	This study
pMF6-G270A-lacZYA	pMF6 with a point mutation at the −229 site, which corresponds to position 270 in the Morganellaceae promoter alignment	Ampicillin	This study
pMF6-ureD-FLAG	pMF6 driving expression of a FLAG-tagged ureD	Ampicillin	This study
pMF6-ORF-FLAG	pMF6 encoding −308 to −73 of PureD followed by a FLAG tag	Ampicillin	This study

TABLE 4.

Select SNPs identified in the ureR-ureD intergenic region among P. mirabilis isolates

Positiona	Ab	T	C	G
−16	–c	3.2%	–	96.8%
−91	17%	83%	–	–
−146	96.6%	–	–	3.4%
−149	–	96.6%	3.4%	–
−308	3.4%	–	–	96.6%
−493	–	96.6%	3.4%	–

^a Position in reference to the ureD start codon.

^b Percentage of 470 P. mirabilis isolates encoding a particular nucleotide at the position listed in the first column.

^c None of the 470 P. mirabilis isolates encoded this nucleotide at the position listed in the first column.

Molecular cloning of reporter constructs

Molecular cloning was performed in E. coli TOP10. The 416 nucleotides directly upstream of ureD were amplified by PCR from P. mirabilis HI4320 genomic DNA and Gibson cloned into the multiple cloning site of linear pGEN-MCS using HiFi Assembly Mix (NEB) to form pMF2 (50). pGEN-MCS is a low-copy plasmid encoding a p15A origin, ampicillin resistance, and multiple plasmid maintenance systems (par hok sok mok parM parR) (49). The pMF2 backbone was amplified by PCR, and lacZYA was amplified from E. coli CFT073 genomic DNA with primers that added sequence homologous to the pMF2 backbone at the site of insertion. These fragments were cloned using HiFi Assembly mix (NEB) to form pMF2-lacZYA. Site-directed mutagenesis of pMF2-lacZYA was accomplished by a combination of Gibson assembly method and Single Primer Reactions In Parallel (SPRINP) (51). Primers and methods for each reporter construct are presented in Table S3. To construct pMF6-C261T and pMF6-G270A, Gibson assemblies were performed with three fragments: synthesized promoters with overhangs (Table S4), PCR product of lacZYA, and digested pGEN-MCS. The plasmids were sequence confirmed via nanopore sequencing (Plasmidsaurus, Eurofins) before transformation into HI4320. Plasmids were maintained with 100 µg/mL ampicillin on solid agar and 50 µg/mL ampicillin in liquid medium. pMF6-ureD-FLAG was produced via Gibson cloning with linearized pGEN-MCS and a PCR fragment encoding −308 to −1 of P_ureD and the ureD coding sequence amplified from HI4320 gDNA using primers that added a 3′ FLAG tag and stop codon. Similarly, pMF6-ORF-FLAG encoding the −308 to −73 region of P_ureD was amplified from HI4320 gDNA using primers that added a 3′ FLAG tag and stop codon.

Miller assays

Miller assays were performed as described (52). Briefly, overnight cultures of reporter constructs were diluted to an OD₆₀₀ = 0.05 in 2 mL LB buffered with 250 mM HEPES, pH 7. For each construct, a second subculture was prepared with in buffered LB containing 100 mM urea. All cultures were cultured to mid-logarithmic phase. Absorbance at 600 nm (OD₆₀₀) for each culture was recorded prior to permeabilization in permeabilization solution [100 mM Na₂HPO₄, 20 mM KCl, 2 mM MgSO₄, 0.4 mg/mL sodium deoxycholate, 0.8 mg/mL cetyltrimethylammonium bromide (CTAB), and 5.4 µL/mL β-mercaptoethanol]. Permeabilized cultures and substrate solution [1 mg/mL o-nitrophenyl-β-galactoside (ONPG), 60 mM Na₂HPO₄, 40 mM NaH₂PO₄, and 2.7 µL/mL β-mercaptoethanol] were brought to 37°C prior to the start of the reaction. Substrate solution was added to permeabilized cells, and the mixtures were incubated at room temperature until a visible accumulation of the yellow by-product of ONPG catalysis by β-galactosidase. At this time, 700 µL of freshly made 1 M Na₂CO₃ was added to stop the reaction. Tubes were centrifuged at 21,000 × g for 5–10 minto pellet cell debris prior to measuring the color change using OD₄₂₀. Miller units were calculated according to the equation below. Six biological replicates were performed for each construct, and Miller units were normalized to the pMF2 construct cultured in buffered LB.

$${\displaystyle \mathrm{M}\mathrm{i}\mathrm{l}\mathrm{l}\mathrm{e}\mathrm{r}\mathrm{u}\mathrm{n}\mathrm{i}\mathrm{t}\mathrm{s}={1000}^{\ast }{\displaystyle \frac{(\mathrm{O}{\mathrm{D}}_{420})}{(\mathrm{O}{\mathrm{D}}_{600}{)}^{\ast }(0.02\mathrm{m}\mathrm{L}{)}^{\ast }(\mathrm{E}\mathrm{l}\mathrm{a}\mathrm{p}\mathrm{s}\mathrm{e}\mathrm{d}\mathrm{t}\mathrm{i}\mathrm{m}\mathrm{e}\mathrm{o}\mathrm{f}\mathrm{r}\mathrm{e}\mathrm{a}\mathrm{c}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}\mathrm{i}\mathrm{n}\mathrm{m}\mathrm{i}\mathrm{n}\mathrm{u}\mathrm{t}\mathrm{e}\mathrm{s})}}}$$

Identification of P. mirabilis SNPs

The sequence conservation of the urease promoter region across P. mirabilis isolates was assessed using P. mirabilis genome sequences publicly available on BV-BRC (n = 793) (53, 54). Duplicate entries as well as genomes with a BV-BRC “Poor Quality” rating, CheckM completion score <95%, CheckM contamination score >2%, or fine consistency score <95% were excluded. Four hundred seventy genomes met these quality standards (Table S1). The sequence of the ureR-ureD intergenic region was obtained for each isolate. During this process, we determined that ureR is misannotated in all BV-BRC P. mirabilis genomes. Thus, we trimmed sequences to exclude the experimentally determined UreR coding sequence (19). Sequences were aligned using Clustal Omega (33). The alignment revealed a single highly dissimilar sequence from strain CRK0056. Submitting the ureR-ureD intergenic region of CRK00056 to the NCBI BLAST tool revealed higher similarity to sequences in the Proteus vulgaris, Proteus terrae, Proteus columbae, and Proteus penneri genomes than to the sequence of P. mirabilis HI4320. Thus, we determined that CRK00056 likely belongs to another Proteus species and excluded its sequence from further analysis. Sequence conservation within the ureR-ureD promoter region was visualized via Sequence Logo graph created with WebLogo (55). To identify SNPs, we used multiple sequence alignment to generate a position frequency matrix with BuddySuite. This approach yielded 15 SNPs (Table S2). We narrowed our focus to the most abundant SNPs (present in at least 2% of strains; 6/15 SNPs; Table 4). Two of these SNPs were found in primarily human-associated P. mirabilis isolates and were selected for site-directed mutagenesis.

Identification of Morganellaceae-family SNPs

P. mirabilis is a member of the Morganellaceae-family of bacteria composed of Arsenophonus, Cosenzaea, Moellerella, Morganella, Photorhabdus, Proteus, Providencia, and Xenorhabdus genera. Within this family, 14 additional species encode a chromosomal UreR-regulated urease locus. The ureR-ureD intergenic region from representative genomes from each species was obtained from BV-BRC. Sequences were aligned with Clustal Omega and visualized using the NCBI Multiple Sequence Alignment Viewer (33, 56). To assess the relative frequencies of bases at each nucleotide, a Sequence Logo graph was created with WebLogo (Fig. S3) (55). Within the UreR2 binding site, the most abundant base at each nucleotide was compared to the base present in strain HI4320. At three loci within UreR2, HI4320 did not encode the same base as the majority of the other Morganellaceae species, one of which was within the −35 site (13). The other two loci (−16 and −91) were chosen for site-directed mutagenesis.

RNA isolation

Overnight cultures of bacterial strains were subcultured into both 3 mL of buffered LB and 3 mL of buffered LB supplemented with 100 mM urea. Cultures were incubated at 37°C with agitation and cultured to mid-logarithmic phase (OD₆₀₀ = 0.4–0.8). After reaching this growth phase, 1 mL of culture was preserved in 2 mL of RNAprotect Bacteria Reagent (Qiagen) for RNA isolation. After incubating for 15 min at room temperature, samples were centrifuged, supernatants removed, and pellets frozen at −20°C for at least one day. Since P. mirabilis is challenging to lyse, this freeze-thaw cycle was included to facilitate cell lysis. RNA was isolated using an RNeasy kit (Qiagen). Samples were also incubated with lysozyme (0.5 mg/mL) for 15 min 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).

Quantitative reverse-transcriptase PCR

After RNA isolation, cDNA was synthesized using the iScript cDNA synthesis kit (Bio-Rad) according to the manufacturer’s protocol. In parallel, cDNA synthesis reactions were prepared without reverse transcriptase for each sample to ensure no genomic DNA was present in the samples. Each qRT-PCR reaction contained 2 µL of cDNA, 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 (57). 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.

In vitro growth kinetics

Overnight cultures of bacterial strains were diluted 1:100 in 1 mL of Minimal A Medium in which the nitrogen source, ammonium sulfate, was replaced with 0.1%, 0.3%, or 0.5% urea by volume. To analyze growth kinetics, 300 µL of each prepared culture was added to a 100-well plate in triplicate. Blanks for each urea concentration were also included. Plates were incubated in a Bioscreen C growth analyzer for 24 h with intermediate shaking, and the OD₆₀₀ of each well was measured every 15 min. Data for three biological replicates of each condition were collected on separate days. After subtracting the OD₆₀₀ of the appropriate blank, data were visualized in GraphPad Prism 10.

Immunoblot

Overnight cultures of bacterial strains were subcultured into both 3 mL buffered LB and 3 mL buffered LB supplemented with 100 mM urea. Cultures were incubated at 37°C with agitation and cultured to mid-logarithmic phase (OD₆₀₀ = 0.4–0.8). At this time, 1 mL of each culture was centrifuged (8–10 min at 8,000 × g or 21,100 × g). The pellet was resuspended in a volume of 1× sample buffer [5% glycerol, 58 mM Tris (pH = 6.8), 2% SDS, 0.1 M DTT, and 0.00002% bromophenol blue] equal to 0.1*(OD₆₀₀). Samples were incubated at 95°C for 10 min prior to moving to −20°C storage. Samples were boiled again at 95°C for 10 min and separated using a precast 7.5% SDS-polyacrylamide gel (Bio-Rad). Proteins were transferred to a PVDF membrane (Millipore Sigma), and Ponceau staining was used to confirm the successful transfer. Proteins were detected with anti-LacZ-HRP (Fisher Scientific 501991083) or anti-FLAG (Invitrogen MA1-91878) in conjunction with a secondary HRP-conjugated antibody (Dako P0260). Chemiluminescent signal was produced using the ECL Plus Western blotting detection system (Thermofisher).

SUPPLEMENTAL MATERIAL

The following material is available online at https://doi.org/10.1128/aem.01273-24.

Figure S1. aem.01273-24-s0001.tiff.

Differential use of urea as a nitrogen source.

Figure S2. aem.01273-24-s0002.tiff.

Alignment of the ureR-ureD intergenic region from 470 P. mirabilis genomes.

Figure S3. aem.01273-24-s0003.tiff.

Alignment of the ureR-ureD intergenic region from 15 Morganellaceae species.

Supplemental legends. aem.01273-24-s0004.docx.

Legends for Fig. S1 to S3 and Tables S1 to S4.

Supplemental tables. aem.01273-24-s0005.xlsx.

Tables S1 to S4.

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.