Sphingosine induction of the Pseudomonas aeruginosa hemolytic phospholipase C/sphingomyelinase (PlcH)

ABSTRACT

Hemolytic phospholipase C, PlcH, is an important virulence factor for Pseudomonas aeruginosa. PlcH preferentially hydrolyzes sphingomyelin and phosphatidylcholine, and this hydrolysis activity drives tissue damage and inflammation and interferes with the oxidative burst of immune cells. Among other contributors, transcription of plcH was previously shown to be induced by phosphate starvation via PhoB and the choline metabolite, glycine betaine, via GbdR. Here, we show that sphingosine can induce plcH transcription and result in secreted PlcH enzyme activity. This induction is dependent on the sphingosine-sensing transcriptional regulator SphR. The SphR induction of plcH occurs from the promoter for the gene upstream of plcH that encodes the neutral ceramidase, CerN, and transcriptional readthrough of the cerN transcription terminator. Evidence for these conclusions came from mutation of the SphR binding site in the cerN promoter, mutation of the cerN terminator, enhancement of cerN termination by adding the rrnB terminator, and reverse transcriptase PCR (RT-PCR) showing that the intergenic region between cerN and plcH is made as RNA during sphingosine, but not choline, induction. We also observed that, like glycine betaine induction, sphingosine induction of plcH is under catabolite repression control, which likely explains why such induction was not seen in other studies using sphingosine in rich media. The addition of sphingosine as a novel inducer for PlcH points to the regulation of plcH transcription as a site for the integration of multiple host-derived signals.

IMPORTANCE

PlcH is a secreted phospholipase C/sphingomyelinase that is important for the virulence of Pseudomonas aeruginosa. Here, we show that sphingosine, which presents itself or as a product of P. aeruginosa sphingomyelinase and ceramidase activity, leads to the induction of plcH transcription. This transcriptional induction occurs from the promoter of the upstream ceramidase gene generating a conditional operon. The transcript on which plcH resides, therefore, is different depending on which host molecule or condition leads to induction, and this may have implications for PlcH post-transcriptional regulation. This work also adds to our understanding of P. aeruginosa with host-derived sphingolipids.

INTRODUCTION

The secreted hemolytic phospholipase C/sphingomyelinase, PlcH, was one of the earliest identified Pseudomonas aeruginosa virulence factors (1). PlcH preferentially hydrolyzes phosphatidylcholine and sphingomyelin to yield phosphorylcholine and either diacylglycerol or ceramide, respectively (2). Unlike virulence systems commonly lost during chronic infection (3–5), PlcH production is retained in all clinical and environmental strains that have been reported, though levels vary between strains (2, 6–8). Experimentally, plcH mutants have decreased virulence in many infection models, and strains producing higher levels of PlcH cause more lung damage (8–12), due to a combination of cytotoxicity, pro-inflammatory properties, and surfactant damage (9, 13–17). As befits a secreted enzyme that likely functions as a public good, plcH transposon insertants do not typically have substantial competition defects in transposon sequencing studies (18–21), as mutants are presumably rescued by community production of PlcH. In humans, PlcH production is negatively correlated with patient outcomes and treatment efficacy in cystic fibrosis (CF) (6, 7).

Like many important virulence pathways, there are a number of inputs into PlcH regulation. GbdR activates transcription in response to glycine betaine (GB) or dimethylglycine and binds directly to the plcH upstream region ~90 bases upstream of the GB-dependent transcription start site (22–27). GB and dimethylglycine can lead to plcH induction directly, whereas choline can only induce plcH when metabolized to GB via the BetBA enzymes (22, 27). Transcription of plcH is also induced under conditions of phosphate starvation, transduced through the PhoBR system (25). Additionally, Anr can function as a repressor of plcH transcription under varying oxygen conditions (28) in response to metabolites from the GB catabolic pathway (29). Induction of plcH by GB is under catabolite repression control (26), but it is not known if that is direct or indirect. There is also a Vfr binding site predicted ~300 bases upstream of the plcH translational start site; however, no function arising from this site has been shown, and Vfr was previously ruled out as the mediator of plcH catabolite repression (26). The H-NS-like proteins MvaU and MvaT are known to directly or indirectly regulate expression from many loci in Pa, including plcH (30). PhrS has also been shown to negatively regulate plcH expression, likely indirectly (31).

In this study, we add to the already multifaceted transcriptional regulation of plcH by showing that the host-derived sphingolipid, sphingosine, is a novel inducer of plcH transcription and resultant extracellular PlcH. Rather than acting directly on the plcH promoter like GbdR and PhoB, sphingosine induction is the result of the activation of the cerN promoter via SphR and transcriptional readthrough from cerN to plcHR. This regulation was not previously observed when sphingosine induction of plcH was examined (32) because, as we show here, sphingosine induction of plcH is under catabolite repression control. Together with previously described regulatory pathways, these findings highlight plcH as a node for the integration of important host-derived metabolites.

RESULTS

Sphingosine promotes PlcH production via transcriptional induction

In unrelated studies to understand sphingosine-dependent transcriptional responses in multiple bacterial species, we conducted sphingosine induction of P. aeruginosa PAO1 in minimal media and observed that transcript counts from plcH and plcR were much more abundant in the presence of sphingosine. The plcHR operon encodes hemolytic phospholipase C/sphingomyelinase, PlcH, and its chaperone, PlcR. The cerN gene, which encodes the neutral ceramidase (33, 34), is immediately upstream of plcHR and in the same orientation (Fig. 1A). As sphingosine was not a previously known inducer of PlcH, we chose to further examine this induction route.

Fig 1.

Spingosine induction of PlcH. (A) Genomic region containing plcH, plcR, and cerN with gene numbers from both PAO1 and PA14 with components important for this study diagramed and labeled. (B) Sphingosine induces extracellular PlcH activity in both PAO1 and PA14 as measured by nitrophenylphosphorylcholine (NPPC) hydrolysis. “a” denotes P < 0.0001 comparing ∆sphR to wild-type (WT) in the presence of sphingosine using two-way ANOVA with Tukey’s post-test. All ∆plcH conditions are significantly lower than WT and ∆sphR in both strains. Error bars represent standard deviation. (C) β-Galactosidase activity from a P_plcH-lacZYA plasmid-encoded reporter [pMW22, location noted with a blue bar in panel (A)] showing that choline leads to induction in both WT and ∆sphR, while sphingosine does not alter expression and loss of sphR has no impact (two-way ANOVA with Sidak’s post-test). (D) Diagram of the chromosomal locus for strains examined in Fig. 1E and F showing the integration of the nanoluciferase open reading frame in between plcH and plcR, generating a synthetic plcH-nLuc-plcR operon. (E and F) Sphingosine induction of PlcH activity and transcription requires sphR, and loss of sphR on the chromosome can be complemented by sphR expressed from a plasmid (pSphR) for both PlcH enzyme activity (E) and nLuc activity (F). Statistical significance noted as ^**P < 0.01 and ^***P < 0.001 using a two-way ANOVA with Tukey’s post-test, and the noted comparisons are to the WT pEV in the respective condition. The sphR complement is not significantly different than WT in any condition. For panels (C, E, and F), all data points are shown and are colored by experiment with white circles for all replicates from experiment #1, gray from experiment #2, and black from experiment #3. Only the means for each experiment are used in the statistical analyses for these panels (n = 3 per condition). Gene organization diagram generated with BioRender. Abbreviations: bs, binding site; sph, sphingosine; tss, transcription start site; term, transcriptional terminator; P, pyruvate (control); C, choline; S, sphingosine; pEV, empty vector pMQ80.

We first tested if sphingosine induction of plcHR leads to increased secreted PlcH as measured by an enzymatic assay based on NPPC hydrolysis. Figure 1B shows that choline and sphingosine can both independently drive increased PlcH enzyme activity in the supernatant and that sphingosine induction of PlcH activity was not observed in an sphR deletion. We had previously used the region immediately upstream of the plcH gene as a negative control for electrophoretic mobility shift assays (EMSAs) testing SphR binding (35), and in support of that, the P_plcH-lacZYA transcriptional reporter shows no response to sphingosine but responds to choline induction (Fig. 1C), as we have previously shown (23). We then integrated a codon-optimized nanoluciferase gene (nLuc) into the plcHR locus to function as a chromosomal reporter: plcH-nLuc-plcR (Fig. 1D). The integration of nLuc at this position did not alter PlcH secreted activity to either sphingosine or choline (Fig. 1E; Data Set S2), and measurement of the chromosomally encoded nLuc showed induction by sphingosine (Fig. 1F). Though the pattern of nLuc expression mirrors PlcH activity, nLuc appears to be very stable and accumulates in cells during overnight growth, which decreases the dynamic range of this reporter when starting from the high cell inoculum we use to simultaneously measure supernatant NPPC hydrolysis activity within this short timeframe.

Sphingosine induction of plcH transcription is due to transcriptional readthrough from the P_cerN promoter

Upstream of the plcHR operon is cerN, encoding a neutral ceramidase, which is a known target for sphingosine-dependent transcriptional induction (32) controlled by SphR (Fig. 1A) (35). To determine whether the overall organization of this region, including the cerN promoter, was needed for plcH induction by sphingosine, we tested a plasmid-borne P_cerN-fLuc-P_plcH-nLuc reporter (Fig. 2A, bottom). Wild-type cells carrying this reporter showed that induction nanoluciferase activity in the presence of sphingosine was not induced in response to sphingosine in an sphR deletion mutant. Choline induction of nanoluciferase activity was not impacted by sphR deletion (data not shown).

Fig 2.

Sphingosine induction of PlcH/plcH dependent on the cerN promoter. (A) Dual-luciferase assay results from strains carrying the plasmid diagramed below the graph. The plasmid results in a P_cerN-fLuc-P_plcH-nLuc dual reporter that mimics the chromosomal locus without the cerN or plcH coding sequences. Sphingosine induction from the plcH promoter is dependent on sphR with significance tested using two-way ANOVA with Sidak’s post-test with * denoting P < 0.05. (B and C) Mutation of five bases in the distal SphR binding site in the chromosomal cerN promoter region (yellow highlighted residues) leads to loss of extracellular PlcH activity (B) and resultant plcH-nLuc activity (C). Significance tested using two-way ANOVA and Dunnett’s post-test with WT pyruvate as the comparator, with *** denoting P < 0.001. For all panels, all collected data points are shown and are colored by experiment with white circles for all replicates from experiment #1, gray from experiment #2, and black from experiment #3. Only the means for each experiment are used in the statistical analyses for these panels (n = 3 per condition). Gene organization diagram generated with BioRender. Abbreviations: mut, mutant.

Mutation of the SphR binding site in the cerN promoter at the native locus leads to the elimination of sphingosine-dependent induction of extracellular PlcH activity and plcH transcriptional induction (Fig. 2B and C). Sphingosine induction of plcH requires SphR regulation from the cerN promoter (Fig. 2B and C) and does not require any part of the cerN coding sequence (Fig. 2A) that strongly suggests the existence of a conditional operon that includes cerN, plcH, and plcR leading to a cerNplcHplcR polycistron.

To determine the potential for transcriptional readthrough from the cerN promoter, we conducted reverse transcriptase PCR to examine the existence of RNA containing cerN, plcH, and the cerN-plcH intergenic region starting from the 3′ end of cerN open reading frame and continuing into the 5′ end of the plcH open reading frame (Fig. 3A). The RT-PCR supports the existence of this transcript and suggests that the cerN transcriptional terminator may not be a complete block to transcriptional extension.

Fig 3.

Evidence for conditional operon containing cerN and plcH. Reverse transcriptase PCR demonstrates a contiguous RNA from the 3′ end of the cerN coding sequence to the 5′ end of the plcH coding sequence. The templates were total RNA isolated from choline induction (cho) or sphingosine induction (sph) with a positive control of genomic DNA (G) and a negative control without RNA. Reverse transcriptase addition into the cDNA synthesis mix is noted in the RT label with + designating inclusion of reverse transcriptase. In agreement with all presented data and previously published work, the cerN coding sequence is detected only in RNA from sphingosine induction, the plcH coding sequence is detected in RNA from both choline and sphingosine induction, and the region spanning the cerN-plcH intergenic region (cerN-plcH ig) is detected only in RNA from sphingosine induction. Approximate binding sites for primers and span of the products noted in the diagram below the gel images; diagram is not to scale.

To test whether the predicted cerN terminator functioned as such, we tested its ability to prevent transcriptional readthrough using a synthetic operon on a plasmid containing P_BAD-gfpmut3-ig_cerNplcH-nLuc compared to a version that lacked the distal half of the predicted terminator stem loop. Mutation of the cerN terminator allowed completely permissive transcriptional readthrough, while the wild-type cerN terminator greatly inhibited readthrough (Fig. 4A). The presence of sphingosine did not impact the degree of termination (Fig. 4A) nor did heterologous testing in Escherichia coli (data not shown), and thus, the terminator appears to be unregulated in this context. Additionally, to rule out the potential that an anti-terminator in the cerN coding sequence would typically prevent termination and would not be detectable with our nLuc reporter, we constructed the same cerN terminator stem loop mutation on the chromosome. Deletion of the distal half of the cerN terminator stem loop led to a large increase in PlcH enzyme induction in response to sphingosine, demonstrating that the cerN terminator functions as such on the chromosome (Fig. 4B). To examine the impact of transcriptional terminator strength on sphingosine-dependent plcH induction, we inserted the rrnB transcriptional terminator immediately upstream of the native cerN transcriptional terminator on the chromosome. Adding this additional transcription terminator substantially reduced sphingosine induction of PlcH enzyme activity (Fig. 4C) and abrogated nanoluciferase induction from the plcH-nLuc-plcR locus (Fig. 4D). Together, these data support the function of the cerN terminator in the cerN-plcH intergenic region but suggest it likely functions as a static dose control system.

Fig 4.

Role of the predicted cerN terminator in transcriptional termination. (A and B) The predicted cerN terminator does promote transcriptional termination. (A) Expression of nLuc from a synthetic operon containing gfpmut3 and nLuc with the wild-type cerN terminator or a mutant with the distal leg of the terminator stem loop deleted (see diagram below data). WT PA14 containing these plasmids were grown in a morpholinepropanesulfonic acid (MOPS) pyruvate (P), with 0.1% L-arabinose (A), or with both L-arabinose and 100 µM sphingosine (AS). (B) Generation of the same cerN terminator mutation as in (A) at the chromosomal cerN locus (small red x on diagram below graph) results in much greater PlcH induction in response to sphingosine than in WT. Note the scale difference in NPPC hydrolysis between panels (B and C). (C and D) Addition of the rrnB terminator upstream of the predicted cerN terminator leads to loss of significant extracellular PlcH activity (C) and resultant plcH-nLuc activity (D). For panels (A, C, and D), significance tested using two-way ANOVA and Dunnett’s post-test with WT pyruvate as the comparator, with ^*P < 0.05, ^**P < 0.01, and ^****P < 0.0001. For panel (B), significance tested by two-way ANOVA and Tukey’s post-test comparing all means, with ^***P < 0.001. For all panels, all of the collected data points are shown and are colored by experiment with white circles for all replicates from experiment #1, gray from experiment #2, and black from experiment #3. Only the means for each experiment are used in the statistical analyses for these panels (n = 3 per condition). Gene organization diagram generated with BioRender. Abbreviations: ig, intergenic; term, terminator.

Sphingosine induction of PlcH is under catabolite repression control

PlcH was shown to be unresponsive to sphingosine in a previous study conducted in a peptone and yeast extract-containing medium (32), while our study uses a minimal medium, which made us suspect catabolite repression control. PlcH induction by choline is under catabolite repression control while PlcH induction by phosphate starvation is not (26). To test if sphingosine:SphR-dependent induction of PlcH production is under catabolite repression control, we measured NPPC production during growth in minimal media with pyruvate, glucose, or succinate as primary carbon sources. Both choline induction and sphingosine induction of PlcH activity are repressed during growth on glucose and succinate compared to pyruvate (Fig. 5). The responsible mechanism for this regulation is currently unknown, but we speculate on possible mechanisms in Discussion.

Fig 5.

Extracellular PlcH activity is under catabolite repression control. Both choline and sphingosine induction of PlcH activity is repressed when cells are grown with either glucose or succinate as sole carbon sources. Pyruvate is permissive to PlcH induction by both compounds and is known to not trigger catabolite repression control. Analysis by two-way ANOVA with Dunnett’s post-test using pyruvate induction in pyruvate media as the comparator (far left bar) shows that only choline and sphingosine induction in pyruvate media is significantly different (P < 0.0001). The trend toward induction for sphingosine in glucose and succinate media is small (<50% increase) and not statistically significant. All of the collected data points are shown and are colored by experiment, with white circles for all replicates from experiment #1, light gray from experiment #2, dark gray from experiment #3, and black from experiment #4. Only the means for each experiment are used in the statistical analyses for these panels (n = 4 per condition).

DISCUSSION

PlcH is an important virulence factor for P. aeruginosa capable of initiating inflammation, damaging pulmonary surfactant, and direct cell lysis (2, 6, 8, 11, 12, 16, 36–38). Here, we identified an additional host-derived molecule driving PlcH induction—sphingosine. Sphingosine induction of plcH transcription is mediated by the sphingosine-binding regulator SphR in the context of transcriptional readthrough from the upstream cerN promoter (see model in Fig. 6). This was surprising given the previous report that sphingosine was unable to induce PlcH activity (32). However, this is likely due to catabolite repression of PlcH expression in rich media as discussed further below. Sphingosine induction of PlcH is a second positive feedback loop controlling its expression, along with the GB-GbdR controlled positive feedback loop. Given the potential that these two systems, Anr regulation, and PhoB regulation may be functional at the same time in the host, an important ongoing question is how these signals interplay, integrate, or impact each other for the ultimate production of PlcH during infection. Response to these signals could potentially be synergistic, interfering, temporally separated within a single cell, or physically separated among sub-populations of cells. The regulatory path(s) critical for PlcH induction in the host has not been determined, but based on the current tally of regulators, one might expect some functional redundancy.

Fig 6.

Current knowledge of transcription at the cerN-plcHR locus under different inducing conditions. The predicted mRNA transcripts are shown as the wavy lines beneath the diagram of the chromosomal region (5′ on the left for all transcripts). Based on data presented herein, the bulk of transcription from the cerN promoter stops at the cerN terminator, but some proportion continues to include the plcHR reading frames, leading to the larger transcript at the bottom of the figure.

The terminator stem loop downstream of the cerN open reading frame is predicted to be a strong Rho-independent terminator with a ∆G of −21.2 kcal/mol [as calculated from reference (39)]. We were, therefore, initially surprised that sphingosine induction of plcHR occurred via transcriptional activation at the cerN promoter. When we tested the cerN terminator downstream of a strong heterologous promoter (P_BAD), it showed robust but incomplete repression of readthrough, and the removal of half of the stem loop resulted in fully permissive transcription (Fig. 4A). The same effect was seen with the mutation of the stem loop at the native locus, where cerN terminator mutation led to much more PlcH induction in response to sphingosine compared to WT (Fig. 4B), which also suggests that there is no native anti-terminator in the 3′ end of the cerN coding sequence (and thus absent for Fig. 4A). Sphingosine does not alter terminator function in E. coli or P. aeruginosa, and we propose that the terminator functions as a static dose regulator controlling the ratio of cerN to plcHR. The importance of maintaining relative dosing between ceramidase and sphingomyelinase/phospholipase C production is not known.

The induction of plcH by GB/GbdR has long been known to be under catabolite repression control (26), but the direct target of this regulation has not been identified. We show that plcH induction by sphingosine is also under catabolite repression control (Fig. 5), which likely explains why sphingosine induction of PlcH was not detected in media containing peptone and yeast extract (32). In contrast, induction of plcH by PhoB is not catabolite-repressed (26), and ceramidase expression is sphingosine-inducible even in the presence of rich media (32). Looking at the sequence in the mRNAs related to plcH induction, there are AAnAAnAA motifs that are known to mediate post-transcriptional catabolite repression (40, 41) in the plcH 5′ untranslated region (5′ UTR) within the PhoB-induced plcH transcript and thus not likely functional targets for catabolite repression control. The cerN 5′ UTR has an AAnAAnAA motif, as do gbdR and sphR. cerN induction by sphingosine is unaffected by rich media, which suggests cerN and sphR are not direct targets of catabolite repression. Looking at the cerN-plcH intergenic region, there are three AAnAAnAA motifs between the cerN stop codon and the GbdR-dependent plcH transcriptional start site. Catabolite repression, thus, could impact the plcHR side of the conditional operon without directly impacting cerN expression, though that prediction has yet to be formally tested. However, it is also important to note that catabolite repression control is often lost in P. aeruginosa strains from the CF lung, and loss of strong catabolite repression is one of the phenotypes of lasR mutants (3, 4).

The cerN promoter has been shown to be one of the intergenic loci under pathoadaptive selection in the CF lung (42). The mutations reported therein are in or just downstream of the SphR binding site, depending on the particular strain. This could suggest selection for modulation of sphingosine-dependent cerN, and thus plcHR, induction in the context of CF, though the functional consequences of these mutations for cerN or plcH regulation have not been tested. It is important to note that while CF strains do not lose plcH activity, PlcH activity does vary widely, suggesting mechanisms to control the dynamic range of PlcH activity in these strains.

PlcH is not homologous to the best-studied phospholipases C, including those from Bacillus and Clostridium, but is rather a distinct enzyme in its own PC-PLC/phosphatase class (43). PlcHs are not unique to P. aeruginosa, but it is the species in which most studies of virulence alteration and regulation have been conducted. Within the genus Pseudomonas, plcH is carried only by P. aeruginosa and two strains of Pseudomonas denitrificans, while strictly environmental and plant pathogenic members lack this hemolytic phospholipase [orthology via reference (44)]. However, PlcH orthologs are found in a number of bacterial pathogens including Mycobacterium tuberculosis, Francisella tularensis, and Burkholderia pseudomallei (43). In B. pseudomallei, there are two distinct PlcH orthologs encoded in the genome with generally overlapping enzymatic capacity but non-redundant functions in virulence models (45). Whether the regulation of PlcH orthologs in these species is similar to that of P. aeruginosa, including identities of the activating molecules, is mostly unknown, though choline-dependent induction has been shown for B. pseudomallei (45).

The regulation of PlcH is multifaceted, and there are likely additional unknown pathways to control the expression of this important virulence factor. One potential is alternate regulation at the cerN promoter, which, in addition to sphingosine, has been shown to be induced in response to PC (32, 46). However, cerN is not part of the PhoB regulon (47) or the GbdR regulon (22) or induced by palmitate (32, 46) or diacylglycerol (32), pointing to some other moiety of PC as an independent inducing agent for cerN and, potentially, plcH. Finally, beyond the level of transcription, there may be mechanisms for post-transcriptional, translational, or post-translational regulation that have not been previously appreciated. While post-transcriptional and translational control has not been reported, PlcH export is mediated by the twin-arginine translocon (TAT) secretion system (48–50), the chaperones PlcR1 and PlcR2 (51), and the Xcp/Gsp type 2 secretion system (49). There is likely additional regulatory control at one or more of these steps to regulate the amount of PlcH exported and alter PlcH localization or activity once exported.

MATERIALS AND METHODS

Strains and growth conditions

Pseudomonas aeruginosa PA14 and isogenic mutant strains were maintained on lysogeny broth–Lennox formulation (LB) or Pseudomonas isolation agar (PIA) plates with 50 µg/mL gentamicin added when appropriate. The Escherichia coli strains used in this study were maintained on LB plates or liquid LB supplemented with 10 or 7 µg/mL gentamicin, respectively. During genetic manipulations, P. aeruginosa was selected, and E. coli was eliminated, using PIA plates supplemented with 50 µg/mL gentamicin. Prior to transcriptional induction studies, P. aeruginosa was grown overnight in a MOPS medium (52) modified as previously described (53) and supplemented with 20 mM pyruvate and 5 mM glucose. E. coli used for induction studies was grown overnight in a MOPS medium with 10% LB (vol/vol), 5 mM glucose, and 7 µg/mL gentamicin.

Chemicals and notes on sphingosine stability and solubility

All media and standard chemicals were purchased from ThermoFisher or Sigma. Sphingosine was purchased as powder from Avanti Polar Lipids and solubilized in 95% ethanol to make 50 mM stocks. The lot number, stock solution age, and sphingosine deposition into the induction assay all seem to impact the absolute amount of sphingosine-dependent induction, leading to varying levels of induction between experiments shown here. As a lipid with labile function groups, the sphingosine stock solution age appeared to be most important, with the ability to induce plcH and other sphingosine-responsive genes going down as the stock became older. Evaporation of the ethanol vehicle in multiwell dishes by air-drying vs a gentle stream of nitrogen gas also showed differential induction, with nitrogen gas drying yielding higher induction levels for similar age stocks.

Strains, general genetic manipulation, and cloning

All strains and plasmids are listed in Table 1, and genetic manipulations are described below. All PCRs were conducted using Q5 DNA polymerase (NEB). Primer sequences are listed in the supplemental data (Data Set 1, Tab 1). Synthetic double-stranded DNA fragments were synthesized by IDT (as gBlocks), and their sequences are listed in the supplemental data (Data Set 1, Tab 2). All plasmids created for this study were sequenced using Plasmidsaurus (Eugene, OR, USA), and select genomes were sequenced using SeqCoast (Portsmouth, NH, USA). PA14 and PAO1 strains with in-frame chromosomal deletions of plcHR (PA0843-0844) and sphR (PA5324) were generated previously (35). For plasmid carriage, P. aeruginosa and E. coli strains were transformed via electroporation and selected for growth on PIA with 50 µg/mL gentamicin and LB with 10 µg/mL gentamicin, respectively.

TABLE 1.

Strains and plasmids used in this study

Strains
Lab strain ID	Genotype	Plasmid	Source
	E. coli NEB5 alpha	None	NEB
	E. coli S17 λpir	None	(54)
MJ79	Pseudomonas aeruginosa PAO1 wild type	None	(55)
MJ532	∆sphR in PAO1	None	(35)
MJ143	∆plcHR in PAO1	None	(16)
MJ984	P. aeruginosa PA14 wild type (new stock of MJ101)	None	(10)
MJ590	∆sphR in PA14	None	This study
MJ144	∆plcHR in PA14	None	This study
JR221	P. aeruginosa PA14 wild type	pMW239	This study
JR223	∆sphR in PA14	pMW239	This study
JR186	plcH-nLuc-plcR in PA14	None	This study
JR188	Wild-type revertant from JR186 generation	None	This study
JR189	plcH-nLuc-plcR in MJ590	None	This study
JR191	Wild-type revertant from JR189 generation	None	This study
JR195	PcerN mutant in PA14	None	This study
JR196	Wild-type revertant from JR195 generation	None	This study
JR210	PcerN mutant in PA14	None	This study
JR213	Wild-type revertant from JR210 generation	None	This study
JR225	cerN-term+rrnB term, plcH-nLuc-plcR	None	This study
JR227	Wild-type revertant from JR225 generation	None	This study
JR243	PA14 wt (MJ984)	pMW240	This study
JR246	PA14 wt (MJ984)	pMW241	This study
MJ1074	cerN terminator mutation in PA14	None	This study
MJ1076	Wild-type revertant from MJ1074 generation	None	This study

Plasmid construction

The pMW22 plasmid was generated in a previous study (23), while the remaining reporter plasmids were constructed as described below.

The p-P_cerN-fLuc-P_plcH-nLuc vector was built using compatible end ligation between a BamH1 cut synthetic fragment PcerN-Fluc-PplcH-nLuc and BamH1 cut and phosphatase-treated pUCP22. Correct cloning and directionality of the insert were determined by EcoR1 digest prior to complete sequencing.

The WT and terminator mutant versions of p-P_BAD-gfpmut3-cerN-ig-plcH-nLuc were built using HiFi (NEB) assembly using either synthetic gene fragment cerN-wtTerm-nLuc or cerN-mutTerm-nLuc and HindIII cut pMQ80, which puts the cerN terminator and the cerN-plcH intergenic-driven nLuc downstream of the P_BAD-controlled gfpmut3 in pMQ80. Diagnostic digest with NdeI was used to assess correct assembly prior to complete sequencing.

General allelic exchange and chromosomal alterations

All allelic exchange constructs were generated with upstream and downstream flanking sequences to whichever inserted, deleted, or mutated element was being altered (described in detail below) using the pMQ30 as the non-replicative and counter-selectable vector (57). After cloning the construct into the pMQ30 backbone, it was transformed into S17 λpir as a conjugative donor. Donor E. coli were mixed with their appropriate recipient strain, and conjugation was allowed to occur in spots on LB plates at 30°C overnight. Conjugation spots were resuspended and plated on PIA with 50 µg/mL gentamicin to select for single-crossover integrants. After restreaking on PIA with 50 µg/mL gentamicin, single-crossover integrants were moved to LB without selection for 3 h prior to plating on LB with 5% sucrose and without salt at 30°C overnight. Sucrose-resistant colonies were screened for loss of gentamicin resistance prior to PCR screening to determine whether each double-crossover colony was a mutant or WT revertant.

The allelic exchange vector for mutation of the SphR binding site in the cerN promoter was built using HiFi assembly (NEB) from two PCR products amplified from PA14 genomic DNA and HindIII and KpnI cut pMQ30. The upstream homology region with the targeted mutation was amplified with primers 2813 and 2815, while the downstream homology region was amplified with primers 2816 and 2814. Correct assembly was assessed by SspI digest prior to complete sequencing. Plasmids matching the predicted sequence were transformed into S17 λpir, and allelic exchange was conducted as described above. The strains with P_cerN engineered at the native site were verified by PCR with primers 2817 and 2818, which yielded a 600-bp fragment that, in the mutant promoter only, is cleavable by SspI into 448- and 152-bp fragments. This resulted in strain JR195 (P_cerN mutant in PA14), JR210 (P_cerN mutant in the plcH-nLuc-plcR background), and their respective wild-type revertants (JR196 and JR213). We sequenced the genome of strain JR195 and its parent strain, and the only difference was the anticipated alteration of the cerN promoter.

The allelic exchange vector for insertion of nLuc as part of the plcHR operon was built using HiFi assembly (NEB) from two PCR products generated from PA14 genomic DNA, the gBlock-nLuc-untagged synthetic fragment, and HindIII and KpnI cut pMQ30. The plcH-side homology region was amplified with primers 2833 and 2834. The plcR-side homology region was amplified with primers 2835 and 2836 for PAO1 and 2837 and 2836 for PA14, to retain strain-specific single-nucleotide differences in the plcH-plcR intergenic region. Correct assembly was assessed by EcoR1 and NdeI double digest prior to complete sequencing. Plasmids matching the predicted sequence were transformed into S17 λpir, and allelic exchange was conducted as described above. The strains with plcH-nLuc-plcR engineered at their native site were verified by PCR with primers 2896 and 2897 and induction of nLuc expression in response to 2 mM choline (described below), resulting in strains JR186 (PA14 wild type, plcH-nLuc-plcR), JR189 (PA14 ∆sphR, plcH-nLuc-plcR), JR189 (PA14 ∆sphR, plcH-nLuc-plcR), JR210 (P_cerN mutation, plcH-nLuc-plcR), and JR213 (P_cerN wild-type revertant, plcH-nLuc-plcR). Wild-type revertants (lacking nLuc) were saved for use as non-luminescent controls for reporter assays.

The allelic exchange vector for the insertion of rrnB upstream of the predicted cerN transcriptional terminator on the chromosome was built using HiFi assembly (NEB) from two PCR products generated from PA14 genomic DNA, the rrnB T1 synthetic fragment, and HindIII and KpnI cut pMQ30. The cerN-side homology region was amplified with primers 2846 and 2847. The plcH-side homology region was amplified with primers 2848 and 2849. Correct assembly was assessed by BamHI and KpnI double digest prior to complete sequencing. Plasmids matching the predicted sequence were transformed into S17 λpir, and allelic exchange was conducted as described above. Donor E. coli were mixed with recipient PA14 strains JR186 (wild-type, plcH-nLuc-plcR) and JR188 (wild-type, plcHR revertant). The successful insertion of rrnB upstream of the cerN transcriptional terminator sequence was then verified by PCR using primers 2853 and 2854, as well as induction of nLuc expression in response to 2 mM choline (described below), generating strains JR225 (cerN-rrnB-plcH-nLuc-plcR) and JR228 (cerN-rrnB-plcHR). Revertant strains, which lack the rrnB insertion, were saved for use as controls in reporter assays.

The allelic exchange vector for mutation of the predicted cerN transcriptional terminator on the chromosome was built using HiFi assembly (NEB) from two PCR products generated from PA14 genomic DNA, the cerN-mutTerm-Chr gBlock synthetic fragment, and HindIII and KpnI cut pMQ30. The cerN-side homology region was amplified with primers 2916 and 2917. The plcH-side homology region was amplified with primers 2918 and 2919. Correct assembly was assessed by KpnI digest prior to complete sequencing. Plasmids matching the predicted sequence were transformed into S17 λpir, and allelic exchange was conducted as described above. Donor E. coli were mixed with recipient PA14 strain MJ984 (wild-type PA14). The successful mutation of the cerN transcriptional terminator sequence was then verified by PCR using primers 2864 and 2863. Revertant strains, which lack the cerN terminator mutation, were saved for use as controls in reporter assays.

Induction for reporter and enzymatic assays

Three reporter constructs were used in this study to observe SphR-sphingosine induction of plcH transcription: the chromosomal plcH-nLuc-plcR reporter, the plasmid-borne pMQ80::P_BAD-gfpmut3-ig_cerNplcH-nLuc and mutant version reporters, and the plasmid-borne P_cerN-fLuc-P_plcH-nLuc reporter. P. aeruginosa strains were grown overnight in MOPS minimal media with 20 mM sodium pyruvate and 5 mM glucose prior to induction. Cells were collected via centrifugation, washed in MOPS, and resuspended in MOPS with 20 mM pyruvate or in MOPS with 20 mM pyruvate plus 2 mM choline or 100 µM sphingosine (Avanti Polar Lipids). All inductions took place at 37°C and were shaken at 170 rpm for 4 h.

nLuc assay

We observed the induction of the chromosomal plcH-nLuc-plcR reporter in nanoluciferase reporter assays using the Nano-Glo Luciferase Assay kit (Promega). The assay was conducted as described by the manufacturer with some modifications reported here. Prior to reading luminescence, induced cells were collected and incubated with 3 mg/mL lysozyme for 30 minutes at 37°C (as the kit does not lyse bacterial cells). Lysed cells were transferred to white (flat clear-bottom) 96-well plates and mixed 1:1 with the Nano-Glo reagent. Luminescence and optical density (OD₆₀₀) were measured using the Synergy 2 H1 BioTek plate reader. The background luminescence of plcH-plcR revertant P. aeruginosa (no reporter) was subtracted from all values.

Experiments using the plasmid-borne P_BAD-gfpmut3-mutig_cerNplcH-nLuc reporter were conducted in a similar manner as described above with the addition of measuring GFP fluorescence (485 excitation and 528 nm emission).

Phospholipase C activity assay

Phospholipase C activity was measured by observing the hydrolysis of the synthetic substrate NPPC based upon the Kurioka and Matsuda method (58), modified as we have described previously (23), but using a final concentration of 10 mM NPPC. Briefly, P. aeruginosa was grown overnight and induced as described in the Induction for reporter and enzymatic section above. NPPC hydrolysis was measured by monitoring the absorbance at 410 nm over time. Phospholipase C activity was reported in micromoles of ρ-nitrophenol generated per minute of reaction per optical density (OD₆₀₀). ρ-Nitrophenol concentration was calculated using the extinction coefficient of 17,700 M⁻¹ cm⁻¹ (59).

RT-PCR for operon determination

RNA was isolated as described above from PA14 induced in MOPS pyruvate, MOPS pyruvate with 100 µM sphingosine, and MOPS pyruvate with 2 mM choline. cDNA was generated using the Induro RT reverse transcriptase system (NEB). PCRs were conducted using Q5 DNA polymerase (NEB) for cerN (primers 2850 and 2851), plcH (primers 2858 and 2859), and the cerN-plcH intergenic region (primers 2866 and 2867) and products separated by agarose gel electrophoresis. PA14 genomic DNA was the positive amplification control, while the negative amplification control has no added template. Control for DNA contamination of the RNA was done by leaving out the reverse transcriptase from the cDNA creation reaction.

Data acquisition, data analysis, data visualization, and statistics

All absorbance, luminescence, and fluorescence measures were done using either BioTek Synergy2 or BioTek H1 multimode plate readers and data acquired using Gen5 (BioTek). Data were exported to Microsoft Excel, and all normalization and fold change calculations were done in Excel. The individual data values used in each figure are available in Supplementary material (Data Set 2). For visualization and statistical analysis, data were moved to GraphPad Prism. Specific statistical tests and their comparator groups are noted in the figure legends. We have opted to show all collected data points using color coding to separate the different experiments however, and importantly, statistical testing was done only using the means from each experiment. Diagrams of chromosomal loci and plasmid constructs were generated in BioRender and used here under the academic licensing terms.

DATA AVAILABILITY

All data pertinent to this study are contained within and complete measurements for all technical replicates for all experiments are contained in whole in the supplemental material.

SUPPLEMENTAL MATERIAL

The following material is available online at https://doi.org/10.1128/jb.00382-23.

Data Set S1. jb.00382-23-s0001.xlsx.

Primers and synthetic nucleotides used in this study.

Data Set S2. jb.00382-23-s0002.xlsx.

All data used to generate figures and statistics in this study.

Supplemental legends. jb.00382-23-s0003.docx.

Legends for Data Sets S1 and S2.

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.