Oligoribonuclease is a central feature of cyclic diguanylate signaling in Pseudomonas aeruginosa

Significance

Many bacteria possess enzymes that synthesize and degrade the intracellular second messenger cyclic diguanylate (c-di-GMP). Bacteria use this molecule to relay environmental signals into physiological responses that control motility, virulence, and biofilm formation. There are two pathways for enzymatic c-di-GMP degradation. One of these pathways involves the production of an intermediate molecule called 5ʹ-phosphoguanylyl-(3ʹ,5ʹ)-guanosine (pGpG). Although many enzymes responsible for c-di-GMP degradation have been characterized, microbiologists have long sought those responsible for pGpG degradation. Here we identify that oligoribonuclease (Orn) mediates pGpG degradation and show that Orn is important for c-di-GMP signaling in the human pathogen Pseudomonas aeruginosa. This discovery reveals that nanoribonucleases, which have been considered housekeeping proteins crucial for mRNA turnover, also have a key role in c-di-GMP signaling.

Abstract

The second messenger cyclic diguanylate (c-di-GMP) controls diverse cellular processes among bacteria. Diguanylate cyclases synthesize c-di-GMP, whereas it is degraded by c-di-GMP–specific phosphodiesterases (PDEs). Nearly 80% of these PDEs are predicted to depend on the catalytic function of glutamate-alanine-leucine (EAL) domains, which hydrolyze a single phosphodiester group in c-di-GMP to produce 5ʹ-phosphoguanylyl-(3ʹ,5ʹ)-guanosine (pGpG). However, to degrade pGpG and prevent its accumulation, bacterial cells require an additional nuclease, the identity of which remains unknown. Here we identify oligoribonuclease (Orn)—a 3ʹ→5ʹ exonuclease highly conserved among Actinobacteria, Beta-, Delta- and Gammaproteobacteria—as the primary enzyme responsible for pGpG degradation in Pseudomonas aeruginosa cells. We found that a P. aeruginosa Δorn mutant had high intracellular c-di-GMP levels, causing this strain to overexpress extracellular polymers and overproduce biofilm. Although recombinant Orn degraded small RNAs in vitro, this enzyme had a proclivity for degrading RNA oligomers comprised of two to five nucleotides (nanoRNAs), including pGpG. Corresponding with this activity, Δorn cells possessed highly elevated pGpG levels. We found that pGpG reduced the rate of c-di-GMP degradation in cell lysates and inhibited the activity of EAL-dependent PDEs (PA2133, PvrR, and purified recombinant RocR) from P. aeruginosa. This pGpG-dependent inhibition was alleviated by the addition of Orn. These data suggest that elevated levels of pGpG exert product inhibition on EAL-dependent PDEs, thereby increasing intracellular c-di-GMP in Δorn cells. Thus, we propose that Orn provides homeostatic control of intracellular pGpG under native physiological conditions and that this activity is fundamental to c-di-GMP signal transduction.

Enzymes that are predicted to “make and break” second messenger cyclic diguanylate (c-di-GMP) have been identified in nearly every known bacterial phylum (1). These enzymes are found in many species in vast numbers and are associated with diverse cellular processes. In many bacteria, low levels of intracellular c-di-GMP up-regulate motility and virulence factor expression, whereas high levels promote extracellular polysaccharide (EPS) production, biofilm development, and cell cycle progression (1, 2).

c-di-GMP is enzymatically synthesized by proteins harboring glycine-glycine-aspartate-glutamate-phenylalanine (GGDEF) domains and degraded by glutamate-alanine-leucine (EAL) and histidine-aspartate-glycine-tyrosine-proline (HD-GYP) domain-containing proteins (3). EAL domains catalyze the asymmetric hydrolysis of c-di-GMP to yield the linear dinucleotide 5ʹ-phosphoguanylyl-(3ʹ,5ʹ)-guanosine (pGpG) (4). By contrast, HD-GYP domains degrade c-di-GMP to GMP (5–7). Degradation of c-di-GMP by an EAL-dependent pathway thus requires a second phosphodiesterase (PDE) to eliminate pGpG and recycle this dinucleotide into the cellular guanosine pool. Although the activity of this second enzyme was predicted to be an important part of c-di-GMP degradation more than 25 y ago (8), the identities of the proteins involved in degrading pGpG have remained enigmatic.

RNAs that are two to five nucleotides in length have been termed nanoRNAs (9). These short oligonucleotides, including pGpG, are degraded by nanoribonucleases (nanoRNAses) (9). One of the most likely candidates for pGpG degradation in vivo is oligoribonuclease (orn) (1). Escherichia coli Orn and its orthologs are required to complete the degradation of mRNA to mononucleotides. Although some bacteria possess multiple redundant nanoRNAses (10), Orn is essential for the viability of other species, including E. coli (9, 11). Thus, the role of Orn in bacterial physiology has been investigated primarily by using systems for the conditional expression or depletion of this enzyme.

Pseudomonas aeruginosa is a formidable opportunistic pathogen that has served as a model organism for studying biofilm formation. This developmental process is regulated by c-di-GMP (12). Recently, a method was devised to trigger the posttranslational degradation of Orn in P. aeruginosa. Work with this system demonstrated that nanoRNAs accumulate in vivo and that these oligonucleotides may prime RNA transcription, shifting transcriptional start sites and causing changes in promoter-specific gene expression (13). Nevertheless, because Orn depletion did not seem to cause expression changes in genes known to belong to a c-di-GMP regulon, an additional role for Orn in c-di-GMP signaling was not evident.

Here we report that P. aeruginosa PAO1 with a complete and precisely engineered deletion of orn is viable and, surprisingly, exhibits elevated levels of c-di-GMP. This finding provided a previously unidentified opportunity to investigate the function of Orn in bacterial c-di-GMP signaling. We present biochemical and genetic evidence that Orn degrades pGpG in vitro as well as in vivo. We also show that the addition of pGpG to EAL-dependent PDEs produced in cell lysates or purified in vitro inhibits c-di-GMP degradation and that this inhibition can be alleviated through addition of purified Orn. Collectively, these data suggest that Orn has a key role in P. aeruginosa c-di-GMP signal transduction. Finally, we demonstrate that the E. coli ortholog of Orn alleviates pGpG accumulation and restores wild-type (WT) phenotypes to a P. aeruginosa Δorn mutant, suggesting that orthologs of Orn may be a central feature of c-di-GMP signaling in other species of bacteria.

Results

Loss of orn Enhances Biofilm Formation by P. aeruginosa PAO1.

Transposon mutagenesis indicates that P. aeruginosa can withstand mutational inactivation of orn (PA4951) (14). Thus, to investigate the function of oligoribonuclease, we constructed an in-frame deletion of orn in P. aeruginosa PAO1. A prominent phenotype of Δorn mutant cells grown in shaken cultures was their tendency to clump and attach to the sides of the culture tubes (Fig. S1). This observation suggested that inactivation of orn may increase surface attachment and biofilm growth. To test this, we carried out microtiter dish biofilm formation assays. The Δorn mutant accumulated significantly increased biofilm biomass in microtiter plate wells relative to the ancestral PAO1 strain (Fig. 1A). We noted that the orn-G11::ISlacZ/hah transposon mutant, which was obtained from the P. aeruginosa PAO1 transposon mutant library (14), produced similar results (Fig. 1A). To investigate these findings further, we grew PAO1 and the Δorn mutant expressing GFP in continuous flow chambers. We saw that the Δorn mutant attached to the interior of flow cells more rapidly and produced mature biofilms with much greater thickness than the parental PAO1 strain (Fig. 1B). Introducing a WT copy of orn on a plasmid complemented all of these phenotypes (Fig. 1 and Fig. S1). Repeating the complementation analysis with an allele of orn bearing a site-directed mutation at a conserved catalytic residue (D11A, E13Q, or D162A) failed to reverse the phenotypes of the Δorn strain (Fig. 1A and Fig. S1). We conclude from these collective data that the ribonuclease activity of Orn suppresses P. aeruginosa aggregation, surface attachment, and biofilm formation.

Fig. S1.

Inactivation of orn causes P. aeruginosa to aggregate in liquid culture and this is c-di-GMP dependent. Transformation with a plasmid expressing the c-di-GMP–specific PDE PA2133 causes cells to disperse. Transformation with plasmid bearing an inactivating mutation (D11A, E13Q, D162A) in a predicted catalytic residue of Orn fails to complement the phenotype. Cultures were grown at 37 °C overnight in BM2 and then photographed. VC, vector control (pUCP18); wt, wild type.

Fig. 1.

Loss of orn increases P. aeruginosa surface attachment and biofilm development. (A) Biofilm formation in microtiter plates. Values represent the means and SDs of three biological replicates. *P ≤ 0.05 and ***P ≤ 0.001 vs. WT with Student t test. (B) Biofilm development in flow cells. Each square on the grid is 20 × 20 μm. VC, vector control.

A Δorn Mutant Strain Has Elevated Intracellular c-di-GMP and Overexpresses Extracellular Polymers.

Because Orn is predicted to participate in a two-step pathway for the degradation of c-di-GMP, we postulated that the increased biofilm growth of the Δorn mutant strain might result from elevated intracellular c-di-GMP concentrations. To test this hypothesis, we extracted nucleotide pools from WT and Δorn cells and quantified c-di-GMP using LC tandem MS (LC-MS/MS). We observed that cultures of Δorn cells produced approximately twofold more c-di-GMP than the ancestral PAO1 strain (Fig. 2A). WT levels of c-di-GMP in the Δorn strain were restored by genetic complementation (Fig. 2A). To further corroborate the involvement of c-di-GMP in regulating the biofilm associated traits of the Δorn mutant, a plasmid encoding an EAL-dependent PDE (PA2133) was transformed into this strain. This plasmid eliminated c-di-GMP–related phenotypes of the Δorn mutant strain (Fig. 1A and Fig. S1) and reduced intracellular c-di-GMP (Fig. 2A). These data indicate that inactivation of orn increases intracellular c-di-GMP and that this is associated with increased biofilm formation.

Fig. 2.

Loss of orn increases intracellular c-di-GMP and pGpG. Intracellular levels of guanosine nucleotides were measured by LC-MS/MS. (A) c-di-GMP. (B) pGpG. (C) GMP. Values represent the means and SDs of three biological replicates. *P ≤ 0.05, **P ≤ 0.005, and ***P ≤ 0.001 vs. WT + vector control (VC) with Student t test. Because Δorn cells aggregate in culture, cell numbers have been calculated by determining the total protein in cultures and dividing this by the protein mass of single P. aeruginosa cell (see SI Results for a detailed calculation).

In P. aeruginosa, c-di-GMP regulates the expression of genes for synthesis of the biofilm matrix, including extracellular polysaccharides (pel and psl) and an adhesin (cdrA) (15). We observed that transcription of these genes was up-regulated 1.7- to 2.6-fold by loss of orn (Fig. S2 A–C). Deletions in pelF and pslD abolished biofilm formation (Fig. S2D) and aggregation (Fig. S2E) by the Δorn strain (SI Results). These data indicate that loss of orn transcriptionally up-regulates a hallmark c-di-GMP regulon and that this contributes to increased production of extracellular polymers in the Δorn strain.

Fig. S2.

The P. aeruginosa Δorn mutant overexpresses genes for extracellular polymers, and this causes aggregation and overproduction of biofilm. (A–C) RT-PCR was used to quantify relative transcript levels for (A) pelA, (B) pslA, and (C) cdrA. Values represent the means and SDs of three biological replicates each performed in three technical replicates, as normalized to the ampR gene in RNA from WT + VC samples. Relative expression levels were calculated using the comparative 2^-ΔΔCT method (27). (D) Biofilm formation by the Δorn strain depends on the extracellular polysaccharides PEL and PSL. Values represent the means and SDs of three biological replicates. *P ≤ 0.05 and **P ≤ 0.005 with Student t test. (E) Aggregation by the Δorn mutant is PEL and PSL dependent. Cultures were grown at 37 °C overnight in BM2 and then photographed.

Inactivation of orn Increases Intracellular pGpG.

Orn is expected to function in cellular nucleotide recycling. The putative biochemical function of Orn and the disruption of c-di-GMP signaling in the Δorn strain led us to hypothesize that the loss of orn might radically alter intracellular pools of pGpG and GMP. To test this, we extracted nucleotide pools from WT and Δorn strains and measured pGpG and GMP using LC-MS/MS. Loss of orn resulted in a fivefold decrease in cellular GMP (Fig. 2C). Although pGpG was not detected in WT cells, there was a massive increase in this dinucleotide in the Δorn strain (Fig. 2B). In fact, these data suggest that Δorn cells had intracellular pGpG concentrations that far exceeded those of GMP. We estimated that WT P. aeruginosa had an average intracellular concentration of 134 μM GMP, undetectable levels of pGpG, and 1.7 µM c-di-GMP, whereas the Δorn strain had ∼26 μM GMP, 88 μM pGpG, and 3.3 µM c-di-GMP (see SI Results for calculation). These data suggest that Orn has a key role in degrading pGpG in vivo.

Orn Efficiently Degrades pGpG in Vitro.

The biochemical activity of Orn was first characterized for E. coli nearly 40 y ago (16, 17). These studies indicated that Orn is a manganese (Mn²⁺)-dependent 3ʹ→5ʹ exonuclease that produces 5ʹ-phosphorylated ribonucleotide monomers from polyribonucleotides (16). The rate at which E. coli Orn degraded RNA oligomers was inversely proportional to the chain length of the substrate, and moreover, the presence of guanine at the 5ʹ or 3ʹ end seemed to be strongly inhibitory to enzyme activity (17). Previous work also suggested that Orn has relatively low activity against a GpG substrate (17).

P. aeruginosa Orn has high identity (67%) with the E. coli protein. Because our data indicate that Orn has a crucial role in degrading pGpG, we sought to reassess the biochemical functionality of Orn. To do this, we first compared pGpG degradation by cell lysates from WT and Δorn strains. We observed that Δorn cell lysates had diminished pGpG degradation activity and that this exonuclease activity could be restored by genetic complementation (Fig. 3A). By contrast, providing an allele of orn with a mutation at the catalytic site (D11A) failed to restore this exonuclease function to lysates (Fig. 3A).

Fig. 3.

Orn preferentially degrades purine dinucleotides in vitro. (A) Degradation of exogenously added pGpG (20 μM) after 60 min in cell lysates as measured by LC-MS/MS. A representative independent experiment is shown, and each datum point represents the mean and SE for three technical replicates. (B) Degradation of pGpG and production of GMP by 100 ng Orn in vitro as assessed by LC-MS/MS. Each datum point is the mean and SD of three replicates and is expressed as the percentage yield of GMP, or the amount of pGpG or c-di-GMP remaining. (C–F) Degradation and semiquantitation of Cy5-pentamers. Each semiquantitation is one representative of three technical replicates.

Next we produced and purified recombinant P. aeruginosa Orn. The exonuclease activity of this enzyme was assessed by tracking the time-dependent degradation of ³³P- or Cy5-labeled RNA oligomers. In the presence of Mn²⁺, Orn degraded a ³³P-labeled RNA 24-mer (Fig. S3A). Similar to the work with E. coli, P. aeruginosa Orn showed a notable predilection for degrading nanoRNAs (i.e., two to five ribonucleotides) (Fig. S3A). Next, the preference of P. aeruginosa Orn for specific nanoRNAs was evaluated. Because dinucleotides and trinucleotides can be difficult to synthesize, and because guanosine homopolymers can form stable quadruplexes that may interfere with enzyme assays, the activity of Orn against RNA homo- and heteropentamers was investigated. To begin, we examined the kinetics of 5ʹ-Cy5-AAAAA-3ʹ, 5ʹ-Cy5-UUAAA-3ʹ, 5ʹ-Cy5-GGAAA-3ʹ, and 5ʹ-Cy5-CCAAA-3ʹ degradation. Although Orn degraded all of these pentamers, this enzyme displayed a preference for purine substrates and bias against pyrimidines (Fig. 3 C–F). 5ʹ-Cy5-CCAAA-3ʹ was degraded the least efficiently of all these substrates (Fig. 3F). Corresponding with this observation, a 5ʹ-Cy5-CCCCC-3ʹ homopentamer was degraded much less efficiently than the RNA heteropentamer containing purines at the 3ʹ end (Fig. S3B). The stepwise degradation of pentamers into intermediates also informed the relative proclivity of Orn for degrading RNA homodimers. Orn degraded 5ʹ-Cy5-AA-3ʹ most rapidly, followed by 5ʹ-Cy5-GG-3ʹ and 5ʹ-Cy5-UU-3ʹ. The degradation of 5ʹ-Cy5-CC-3ʹ was inefficient and the production of 5ʹ-Cy5-C-3ʹ monomers was barely detectable in these assays (Fig. 3 C–F and Fig. S3B).

Fig. S3.

Orn is a calcium-insensitive exoribonuclease that selectively degrades purine rich nanoRNAs. (A) Degradation of the 24-mer 5ʹ-³³P-CACACACACACACACACACACACA-3ʹ. The red arrow indicates the preferential elimination of nanoRNAs. M, Millenium RNA marker; H, acid hydrolysis; -, control with no enzyme. (B) Orn is a poor degrader of cytosine rich nanoRNA substrates. Degradation of 5′-Cy5-CCCCC-3′ is relatively slow compared to other substrates (compare with Fig. 3 C–F). (C and D) Degradation of pGpG by Orn is not sensitive to calcium (Ca²⁺). A concentration of 1 mM CaCl₂ was added to reaction mixtures, and this did not affect nanoRNA degradation kinetics for (C) 5′-Cy5-GGAAA-3′ or (D) 5′-Cy5-UUAAA-3′ (compare with Fig. 3 D and E). Each semiquantitation is one representative of three technical replicates.

Previously, calcium sensitivity has been used to distinguish between EAL-dependent PDEs and the putative enzymes for pGpG degradation. Specifically, EAL-dependent PDEs were found to be inhibited by 1 mM Ca²⁺, whereas the putative PDEs for pGpG degradation were insensitive to this cation (8). We noted that the ribonuclease activity of Orn was Ca²⁺-insensitive (Fig. S3 C and D).

Last, we directly assessed the ability of recombinant Orn to degrade pGpG. Under the tested conditions, LC-MS/MS indicated that 100 ng of clean, purified Orn degraded pGpG to GMP with an initial reaction rate of at least ∼10⁻⁷ mol/min (Fig. 3B). Recombinant Orn exhibited no activity against c-di-GMP (Fig. 3B). Overall, we conclude from these data that Orn has a preference for purine rich nanoRNA substrates and that this enzyme can readily degrade pGpG in vitro.

pGpG Slows c-di-GMP Degradation in Cell Lysates and Inhibits the Activity of Purified RocR.

The buildup of an enzymatic end product frequently leads to negative feedback that inhibits the enzyme that produced it. We thus hypothesized that pGpG-dependent product inhibition of c-di-GMP–specific PDEs may be a mechanism accounting for elevated c-di-GMP levels in the Δorn strain. However, testing this hypothesis is a daunting problem because there are many EAL domain-containing proteins in P. aeruginosa. To address this challenge we began with cell lysates, which contain all of the PDE proteins expressed by the cell. Because WT cell lysates had robust nanoRNAse activity that degraded exogenous pGpG (Fig. 3A), we worked with lysates from Δorn cells. Moreover, we examined Δorn strains transformed with a plasmid for overexpressing either the phenotype variant regulator (PvrR) or PA2133 EAL domain-dependent PDEs or the appropriate vector control.

c-di-GMP was added to aliquots of cell lysates and its degradation was quantified by LC-MS/MS in the presence or absence of exogenous pGpG. Cells transformed with PvrR and PA2133 produced lysates with increased c-di-GMP degradation activity relative to the vector control. However, in all cases, c-di-GMP degradation was significantly inhibited by the addition of 20 µM pGpG (Fig. 4A).

Fig. 4.

pGpG inhibits c-di-GMP degradation in cell lysates and inhibits activity of EAL-dependent PDEs in vitro. C-di-GMP was measured by LC-MS/MS unless otherwise noted in this legend. (A) Degradation of c-di-GMP (70 nM) after 60 min in cell lysates with and without added pGpG (20 µM). (B) Degradation kinetics of c-di-GMP (70 nM) in lysates of Δorn cells overexpressing PA2133 with and without the addition of pGpG (20 µM). Inhibition was alleviated by adding 100 ng of purified Orn. (C) C-di-GMP and pGpG concentration-dependent inhibition of c-di-GMP degradation in lysates of Δorn cells overexpressing PA2133 (measured at 60 min). Where indicated, 100 ng of purified Orn was added. (D) Degradation of MANT-c-di-GMP (400 nM) by purified RocR (1 µM) in vitro, with and without the addition of pGpG (400 µM) or purified Orn (1 µM). In all panels, a representative independent experiment is shown, and each datum point represents the mean and SD for three technical replicates.

To further investigate pGpG-dependent inhibition of c-di-GMP degradation, we focused on lysates from Δorn cells overexpressing PA2133, which had the greatest c-di-GMP degradation activity (Fig. 4A). An analysis of degradation kinetics suggested that the addition of pGpG extended the half-life of c-di-GMP from minutes to hours in these lysates (Fig. 4B). Adding purified recombinant Orn to cell lysates rapidly eliminated pGpG from the reaction mixture (Fig. S4A) and completely alleviated inhibition (Fig. 4B). Increasing concentrations of pGpG over a range of 2–20 μM increased inhibition (Fig. 4C), and moreover, increasing the amount of c-di-GMP 10-fold reduced the level of inhibition (Fig. 4C). These observations suggest that competitive inhibition may at least partially explain pGpG-dependent product inhibition of PA2133 and c-di-GMP degradation in cell lysates.

Fig. S4.

Orn eliminates pGpG from cell lysates and pGpG inhibits the activity of purified RocR in vitro. (A) pGpG levels in the same cell lysates illustrated in Fig. 4B (relative to a 20 µM starting concentration in pGpG treated groups). (B) Inhibition of RocR c-di-GMP degradation activity by the addition of pGpG (80 µM, 60-s reaction) as measured by LC-MS/MS. Each datum point or bar represents the mean and SE of three technical replicates from a representative independent experiment.

Last, we purified recombinant RocR, one of the best-characterized EAL domain-dependent PDEs from P. aeruginosa (18), and sought evidence for pGpG-dependent product inhibition of this PDE in vitro. Here we used a fluorescent N-methylisatoic anhydride (MANT) conjugated c-di-GMP analog to track RocR PDE activity (19). Similar to work with cell lysates, the addition of pGpG inhibited c-di-GMP degradation, and this inhibition was alleviated by the addition of purified Orn (Fig. 4D). These results were corroborated by endpoint LC-MS/MS measurements of RocR-dependent c-di-GMP degradation in a separate set of reactions (Fig. S4B).

In summary, due to the high concentrations of pGpG used in these assays, we propose that pGpG-dependent inhibition of EAL domain-dependent PDEs may be physiologically relevant specifically in the context of the Δorn strain. However, a corollary from these collective data is that Orn exonuclease activity is essential for the elimination of pGpG in WT cells and that this homeostatic function is crucial for the normal catalytic functioning of EAL domain-dependent PDEs.

Orn Is Highly Conserved Among Bacteria but Is Not Likely the Only NanoRNAse Important for c-di-GMP Signal Transduction.

Orn is essential for the viability of E. coli, another paradigm species for studying c-di-GMP signal transduction (2). Therefore, we used interspecies genetic complementation to test whether E. coli and P. aeruginonsa Orn proteins are functional orthologs (SI Results). We observed that P. aeruginosa Orn could rescue the growth defect of an E. coli conditional orn mutant (Fig. S5A) and that E. coli Orn was able to reverse the phenotypes of the P. aeruginosa Δorn mutant (Fig. S5 B–E). These data suggest that E. coli Orn has the capacity to function in the two-step pathway for c-di-GMP degradation.

Fig. S5.

Interspecies genetic complementation suggests that E. coli and P. aeruginosa orn genes are functional orthologs. (A) P. aeruginosa orn rescues the conditional growth defect of an E. coli orn mutant. (B) E. coli orn eliminates the hyper-biofilm formation phenotype of the P. aeruginosa Δorn strain. E. coli orn also restores WT levels of intracellular guanosine metabolites. (C) c-di-GMP. (D) pGpG. (E) GMP. VC, vector control (pUCP18). Values represent the means and SDs of three biological replicates. *P ≤ 0.05, **P ≤ 0.005, and ***P ≤ 0.001 vs. WT + VC with Student t test. Cell numbers have been calculated by determining the total protein in cultures and dividing this by the protein mass of single P. aeruginosa cell (see SI Results for a detailed calculation).

It is thus tempting to speculate that orthologs of Orn may function in c-di-GMP signal transduction in other organisms, and therefore, a tBLASTn survey of Orn was performed for 3,109 complete genomes in the National Center for Biotechnology Information (NCBI) database (as of October 31, 2013). This search identified proteins from 568 additional bacterial species with >50% identity to P. aeruginosa PAO1 Orn (Dataset S1). A Bayesian analysis of Orn sequences placed these orthologs into four phyla (Fig. S6A). This primarily includes Proteobacteria (especially Beta-, Delta- and Gammaproteobacteria) and Actinobacteria, but also single species representing Gemmatimonadetes and Fibrobacteres.

Fig. S6.

Phylogenetic distribution of Orn proteins among bacterial taxa. (A) Unrooted Bayesian inference dendrogram of 167 bacterial Orn proteins. Genus and species shown are colored by class (see graphical legend). A total of 80% of all basal nodes had 0.60–0.79 posterior probability (PP) confidence values and 90% of distal nodes had higher (0.8–1.0) PP values. The Bacillus subtilis NrnA sequence (shown in black, NP_390803) was selected as the outgroup based on related function and low sequence identity (>10%) with other Orn homologs. (B) Co-occurrence of Orn, EAL, and HD-GYP genes among 1,025 genome-sequenced bacterial species.

A cofrequency analysis indicated that the number of genome-sequenced bacterial species possessing both Orn and EAL domain proteins (23%) was slightly greater than the number of those possessing Orn but no EAL domain proteins (13%) (Fig. S6B). At first glance this seems to suggest coselection between Orn and EAL domain proteins. However, Orn may have a crucial role in completing the degradation of RNA (11), and thus it is no surprise that some bacterial species that have Orn do not possess EAL domain proteins. What is more intriguing, however, is that many bacterial species that possess EAL-dependent PDEs do not have an apparent ortholog of Orn (Fig. S6B). Indeed, 15 additional well-sampled bacterial phyla that are thought to possess EAL domain proteins (1) do not have Orn. These phyla include members of Alphaproteobacteria such as Gluconacetobacter xylinus, which is the species in which c-di-GMP was discovered (8). Moreover, a large number of all genome-sequenced bacterial species (10%) are predicted to possess EAL domain proteins yet have neither Orn nor HD-GYP domain proteins (Fig. S6B), the latter of which have been postulated to possess functional redundancy to Orn (Discussion). This finding raises the intriguing possibility that there may be additional, phylogenetically distinct types of nanoRNAses that participate in c-di-GMP signal transduction in the place of oligoribonuclease.

SI Results

Loss of orn Differentially Regulates a Hallmark c-di-GMP Regulon.

In P. aeruginosa, c-di-GMP regulates the expression of the pel, psl, and cdr operons (12, 15), which encode genes for extracellular polysaccharides (PEL and PSL) and a biofilm matrix protein (CdrA). Because Δorn cells display elevated intracellular c-di-GMP, we hypothesized that loss of orn should activate the transcription of pel, psl, and cdr genes. To test this, we began by engineering a Δorn mutant that had additional deletions in pelF and pslD. A ΔornΔpelFΔpslD triple mutant neither clumped in shaken culture tubes nor formed biofilms in microtiter dishes (Fig. S2). Levels of pel, psl, and cdr transcripts were also quantified using RT-PCR. These measurements revealed 2.1-, 1.7-, and 2.6-fold increased transcription for pelA, pslA, and cdrA in the Δorn background (Fig. S2). This transcriptional up-regulation could be reversed by genetic complementation or by overexpressing PA2133 from a plasmid (Fig. S2). We conclude from these data that loss of orn transcriptionally up-regulates a hallmark c-di-GMP regulon and that this results in overexpression of extracellular polymers, causing Δorn cells to aggregate and overproduce biofilm.

Estimation of Intracellular Nucleotide Concentrations.

Several of the bacterial strains used in this study (specifically those with a Δorn mutation) aggregated in shaking culture. This aggregation interfered with accurate OD₆₀₀ and viable cell counts, which are standard measures used to determine the number of bacterial cells in a culture. We thus used protein concentration as a proxy for cell counts. Therefore, to determine intracellular concentrations of nucleotides, protein concentration was measured in one aliquot and nucleotides extracted from another aliquot of the same culture. We then estimated cell number by dividing the total amount of protein by the average protein mass of a P. aeruginosa cell (m_cell = 0.27 × 10⁻¹² g; SI Materials and Methods). Thus, the following equation was used to estimate intracellular concentrations of guanosine metabolites ([G_cell]):

$$\left[{\text{G}}_{\text{cell}}\right]=\frac{\left[\frac{{\text{G}}_{\text{total}}\left(\text{mol}\right)}{{\text{m}}_{\text{total}}\left(\text{g}\right)}\hspace{0.17em}\times {\text{m}}_{\text{cell}}\left(\text{g}\right)\right]}{{\text{V}}_{\text{cell}}\left(\text{L}\right)}.$$

Here, G_total was the total moles of guanosine metabolite in the culture aliquot, m_total was the total grams of protein in another aliquot of the same culture, and V_cell was the volume of a single P. aeruginosa cell (2.2 × 10⁻¹⁵ L, approximated by light microscopy; SI Materials and Methods). The numerator of this equation is the data displayed in Fig. 2 and Fig. S5 (converted to a human friendly scale).

Using this approach, we estimate that WT P. aeruginosa cells contain 134 ± 14 μM GMP, undetectable levels of pGpG, and 1.7 ± 0.3 μM c-di-GMP under the tested growth conditions. By contrast, Δorn cells contained 26 ± 3 μM GMP, 88 ± 5 μM pGpG, and 3.3 ± 0.5 μM c-di-GMP.

E. coli orn Complements the P. aeruginosa Δorn Strain and Degrades pGpG in Vivo.

Our data provide evidence that Orn has an important role in P. aeruginosa c-di-GMP signal transduction. Although orn is evolutionarily conserved among many bacteria, testing this role in other species is challenging because Orn may be essential for viability. This challenge is especially pertinent to E. coli, another paradigm species for studying c-di-GMP signal transduction (2). Given this quandary and the previous report that E. coli Orn may not degrade diguanosine substrates efficiently in vitro (17), we turned to interspecies genetic complementation to look for a potential function of the E. coli ortholog of orn in c-di-GMP signaling. Here, we postulated that E. coli orn may complement the defects of a P. aeruginosa orn mutant, and vice versa. To test this hypothesis, we began with an E. coli strain in which the native promoter of orn was replaced with ptetO (9). This strain had a considerable growth defect in the absence of the inducer anhydrotetracycline (Atc). P. aeruginosa orn expressed ectopically from a plasmid restored the growth of this strain in the absence of Atc (Fig. S5A). Analogously, E. coli orn expressed from a plasmid abolished the hyper-biofilm formation phenotype of the P. aeruginosa Δorn strain and eliminated the accumulation of intracellular pGpG (Fig. S5 B–E). Altogether, these data suggest that orthologs of Orn from other bacterial species, including E. coli, have the capacity to function in the two-step pathway for c-di-GMP degradation.

Discussion

c-di-GMP was originally identified as an allosteric activator of cellulose synthase in G. xylinus (8). This work recognized that c-di-GMP is degraded in two steps. Deactivation of c-di-GMP in G. xylinus protein fractions occurred when a Ca²⁺-sensitive phosphodiesterase, termed PDE-A, cleaved c-di-GMP to yield the inactive open dimer, pGpG. Subsequently, a second Ca²⁺-insensitive phosphodiesterase, termed PDE-B, split this product into two molecules of GMP. It was soon discovered that EAL domains mediate PDE-A activity (20); however, the enzymes responsible for PDE-B activity have remained elusive.

Identifying the enzymes responsible for PDE-B activity has been complicated because bacteria possess many types of proteins that could have a role in pGpG degradation. Biochemical investigations of several of these proteins have not provided evidence supporting their role in pGpG degradation. For example, EAL domain proteins may cleave pGpG; however, this activity occurs in vitro at a rate that is orders of magnitude slower than that of c-di-GMP degradation (21). This mechanism may explain the reduction of pGpG in P. aeruginosa Δorn cells transformed with a plasmid that highly overexpresses PA2133 (Fig. 2). However, overexpression of PA2133 was still insufficient to restore native levels of pGpG to Δorn cells (Fig. 2B). HD-GYP domain proteins might also degrade pGpG; however, whereas HD-GYP domains may degrade pGpG in vitro (7), the relevance of this activity has not been substantiated in vivo. Finally, nanoRNAses may degrade pGpG. It has been previously hypothesized that Orn mediates PDE-B activity (1); however, Orn was first studied in E. coli, and investigations of this enzyme have been hampered because it is essential for E. coli viability (11).

Unlike E. coli, P. aeruginosa can withstand a loss-of-function mutation in orn. This discovery has enabled us to revisit the hypothesis that Orn mediates PDE-B activity. Here, we have provided multiple lines of evidence that Orn has a key role in not only pGpG degradation, but also c-di-GMP signaling in vivo. Although there may be other PDEs that degrade pGpG, it is clear from this work that Orn has a role in guanosine homeostasis that is fundamental to c-di-GMP signal transduction in P. aeruginosa (Fig. 5).

Fig. 5.

Guanosine biochemistry fundamental to c-di-GMP signal transduction in P. aeruginosa. Solid lines indicate reactions and enzymes in which physiological relevance has been substantiated in vivo. The dashed lines indicate a reaction that may be mediated by other unknown cellular enzymes or through the activity of the named enzyme, which has been inferred by sequence or structural similarity. Gmk and Ndk denote guanylate kinase and nucleoside diphosphate kinase, respectively.

Investigations carried out in the 1970s examined the activity of E. coli Orn against a variety of dinucleotide substrates and estimated that Orn had relatively low activity against GpG (17). Contrary to that work, our investigation suggests that P. aeruginosa Orn has a relatively high capability for degrading purine dimers such as pApA and pGpG (Fig. 3 C and D). The basis for this discrepancy in relative reaction rates is not clear. However, our direct measurements of pGpG in cells lacking Orn (Fig. 2 and Fig. S5), as well as the ability of E. coli Orn to complement the P. aeruginosa Δorn mutation (Fig. S5), suggest that Orn-dependent degradation of pGpG has a biologically relevant function, especially in the context of c-di-GMP signaling.

The mutational inactivation of orn is pleiotropic (13). Among other phenotypes, this results in the accumulation of nanoRNAs—including pGpG—that may differentially prime transcription (13). We reasoned that overexpressing the c-di-GMP–specific PDE PA2133 could distinguish the subset of genes and phenotypes differentially regulated by c-di-GMP from other changes due to loss of orn. Overexpression of PA2133 eliminated intracellular c-di-GMP in the Δorn strain (Fig. 2). As expected, this also abolished biofilm hyperproduction (Fig. 1), aggregation (Fig. S1), and overexpression of pelA, pslA, and cdrA (Fig. S2).

A key observation is that pGpG inhibits the degradation of c-di-GMP in lysates of Δorn cells overexpressing EAL-dependent PDEs, as well as purified RocR in vitro (Fig. 4). A question that arises from this work is whether or not this product inhibition may lead to elevated c-di-GMP in vivo. Data in this report do not provide compelling evidence that this mechanism would be operable under normal circumstances. For instance, pGpG could not be detected in nucleotide extracts from WT P. aeruginosa by LC-MS/MS (Fig. 2). Also, although it is possible that Orn activity may be differentially regulated under certain conditions, there is little evidence for differential Orn expression in the Gene Expression Ominbus database (22). Last, we investigated orn gene expression by RT-PCR in planktonic cultures and biofilms. There was no more than a 1.7-fold change in orn transcription measured between any two conditions tested (Fig. S7).

Fig. S7.

Transcription of orn in biofilms and planktonic cells. Planktonic cells were grown in BM2. Biofilm cells were grown in a silicone-elastomer tube with an initial attachment period of 2 h followed by continuous flow in BM2 medium for 24 h. orn transcripts were normalized to PA1684, a predicted ortholog of 1,2-dihydroxy-3-keto-5-methylthiopentene dioxygenase (mtnD). Results shown are the mean of three technical replicates from each of three independent experiments. Error bars represent the SDs. **P ≤ 0.005 vs. logarithmic growth or stationary phase with Student t test.

However, by contrast we propose that pGpG-dependent inhibition of PDEs is a likely scenario for the Δorn strain. In contrast to WT cells, intracellular pGpG in Δorn cells was greater than >80 μM, exceeding levels of GMP (Fig. 2 and SI Results). LC-MS/MS measurements indicated that the molar ratio of pGpG to c-di-GMP was nearly 30:1 in Δorn cells (Fig. 2). Similar stoichiometric ratios of pGpG to c-di-GMP (between 25:1 and 1,000:1) inhibited EAL domain-dependent PDEs in cell lysates (even when those proteins were overexpressed; Fig. 4B) and inhibited the activity of purified RocR in vitro (Fig. 4D and Fig. S4B). Thus, pGpG-dependent inhibition of EAL domain PDEs may be relevant specifically in the context of the Δorn strain. We acknowledge that not all PDE containing enzymes in the cell may be subject to feedback inhibition by pGpG. However, if it happens for even a few, this may tip the balance of DGC and PDE activity to favor the buildup of c-di-GMP in the Δorn mutant.

The corollary of these results is that Orn is responsible for low levels of pGpG in WT P. aeruginosa. These data suggest that homeostatic control of pGpG by Orn enables the responsiveness of c-di-GMP signaling networks by preventing feedback inhibition of PDEs under normal physiological conditions. Thus, we conclude that Orn is the primary PDE in P. aeruginosa that catalyses a crucial final step in c-di-GMP degradation. This finding reveals the possibility that Orn and other nanoRNAses, which have long been considered housekeeping enzymes essential for mRNA degradation (11), may participate in other RNA-based signal transduction processes in many species.

Materials and Methods

Bacterial Strains, Plasmids, and Growth Conditions.

Bacterial strains, plasmids, and primers are described in Tables S1 and S2. Microbiological media, growth conditions, antibiotic selection, and molecular methods for plasmid and strain construction are described in SI Materials and Methods.

Table S1.

Strains and plasmids used in this study

Strain or plasmid	Description	Source
E. coli strains
BL21(DE3)	ompT gal dcm fhuA2 hsdS λ(DE3 [sBamHIo ΔEcoRI-B int::(lacI::PlacUV5-T7 gene1) ind1nin5])	New England Biolabs
NiCo21(DE3)	can::CBD fhuA2 lon ompT gal (λ DE3) dcm arnA::CBD slyD::CBD glmS6Ala ∆hsdS λ DE3 = λ sBamHIo ∆EcoRI-B int::(lacI::PlacUV5::T7 gene1) i21 ∆nin5	New England Biolabs
DH5α	fhuA2 Δ(argF-lacZ)U169 phoA glnV44 Φ80Δ (lacZ)M15 gyrA96 recA1 relA1 endA1 thi-1 hsdR17	Zymogen
MG1655	ilvG, rfb-50, rph-1	Coli Genetic Stock Center
S17.1 (λpir)	thi, pro, hsdR, recA::RP4-2-Tc::Mu aphA::Tn7, λ-pir, Smr, Tpr	(41)
SM10	thi, thr, leu, tonA, lacy, supE, recA::RP4-2-Tc::Mu aphA::Tn7, Kmr	(41)
UM341	CF10230 ptetO::orn, Tcr, Kmr	(9)
P. aeruginosa strains
PAO1	wild type	(42)
DC2	PAO1 orn::aacC1, Gmr	This study
DC16	PAO1 ΔpelF	This study
DC17	PAO1 Δorn	This study
DC26	PAO1 ΔpelF ΔpslD	This study
DC42	PAO1 Δorn ΔpelF ΔpslD	This study
DC52	PAO1 Δorn ΔpelF	This study
PW9334	PAO1 orn-G11::ISlacZ/hah	(14)
Plasmids
pBAD18	Low copy number cloning vector, arabinose inducible, Apr	(43)
pDC12	pBAD18::orn (P. aeruginosa ortholog), Apr	This study
pUM407	pBAD18 with a C-term His×6 tag, Apr	(9)
pDC15	pUM407::orn-His×6 (P. aeruginosa ortholog), Apr	This study
pUM408	pUM407::orn-His×6 (E. coli ortholog), Apr	(10)
pET26b::RocR	Expression vector for RocR-His×6	(18)
pHKT2	pBBR1Tp::gfp, Tpr	(44)
pEX18Gm	Allelic exchange vector, oriT, sacB, Gmr	(37)
pEX18Gm::ΔpelF	pEX18Gm with an in-frame deletion allele for pelF, Gmr	(39)
pHL129	pEX18Gm with an in-frame deletion allele for pslD, Gmr	(38)
pJJH14	pEX18Gm with an in-frame deletion allele for orn, Gmr	This study
pUCP18	E. coli - P. aeruginosa shuttle vector, Apr/	(36)
pDC2	pUCP18::orn (P. aeruginosa ortholog), Apr	This study
pDC57	pDC2 encoding OrnD11A, Apr	This study
pDC58	pDC2 encoding OrnE13Q, Apr	This study
pDC59	pDC2 encoding OrnD162A, Apr	This study
pDC7	pUCP18::PA2133, Apr	This study
pDC14	pUCP18::orn (E. coli ortholog), Apr	This study
pMRP9-1	pUCP18::gfp, Apr	(45)

Ap, ampicillin; Gm, gentamicin; K_m, kanamycin; Sm, spectinomycin, Tc, tetracycline; Tp, trimethoprim.

Table S2.

Primers used in this study

Primer	Sequence
Plasmid sequencing
M13-Universal-F	GTA AAA CGA CGG CCA G
M13-Universal-R	CAG GAA ACA GCT ATG AC
Construction of E. coli–P. aeruginosa shuttle vectors
DC103_ornF01	GAC GGA ATT CAT GCA GAA CCC GCA GAA CCT
DC104_ornR01	GAC GGG ATC CTC AGA GCT TGA TGA AGT GGT CG
DC121_PA2133F01	GAC GGA ATT CGT GAA CGG TTC CCC ACA GGC
DC122_PA2133R01	GAC GGG ATC CTC ACC CCT GGC GGC TCG CCA
DC140_orneF01	GAC GGA ATT CAT GAG TGC CAA TGA AAA CAA CCT
DC141_orneR01	GAC GGT CGA CTT ACA GCT TGA TAA AAT GCT CGC
Site directed mutagenesis of orn (in pDC2)
DC356_a32c_R	CCG GTC ATC TCC AGg gcG ATC CAG ATA AGG T
DC357_a32c_F	ACC TTA TCT GGA TCg ccC TGG AGA TGA CCG G
DC358_g37c_R	AGG CCG GTC ATc tgC AGG TCG ATC CAG
DC359_g37c_F	CTG GAT CGA CCT Gca gAT GAC CGG CCT
DC360_a485c_t486c_R	ATC GAC TCG CGG ATg gcG TCC AGC GCC AGG
DC361_a485c_t486c_F	CCT GGC GCT GGA Cgc cAT CCG CGA GTC GAT
Construction of the Orn-His×6 expression vector
DC129_ornF02	GGG GGA ATT CAC CAT GCA GAA CCC GCA GAA CCT
DC130_ornR02	GGG GGT CGA CGA GCT TGA TGA AGT GGT CGC
DC131_ornR03	GGG GTC TAG ATC AGA GCT TGA TGA AGT GGT CG
DC132_seqF01	AGA TTA GCG GAT CCT ACC TG
DC133_seqR01	ATC TTC TCT CAT CCG CCA AAA CAG
Construction of allelic exchange vectors and verification of mutants
JJH98_ornupF1	ATC CGG AAG CTT GAG CGA TCA CCT GGC CG
JJH99_ornupR1	GGA TCA GAG CTT GAT GAA GTG GAT AAG GTT CTG CGG GTT CTG
JJH100_orndownF1	GAC CAC TTC ATC AAG CTC TGA TC
JJH101_orndownR1	ATC CGG CTG CAG GGA CAA CGC CGA ACT CGC
KMC179_pslDF1	CCG AGG TCT ACC ATT CCC ACG
KMC180_pslDR1	GAA CTT GGT GCG CTT CCA CAG
KMC181_pelFF1	CTG GTA CTG GGA ACT GGC CTA CC
KMC182_pelFR1	CAC GCT GAC GAT CGA CAG CAC
RNA purification and gene expression by RT-PCR
ornF_RT	CTG GAT CGA CCT GGA GAT GAC
ornR_RT	TCA CTG TCG GTG ACG ATG GT
JJH600_rplUF01	CGC AGT GAT TGT TAC CGG TG
JJH601_rplUR01	AGG CCT GAA TGC CGG TGA TC
JJH1459_ampR-F1-RT	GCG CCA TCC CTT CAT CG
JJH1460_ampR-R1-RT	GAT GTC GAC GCG GTT GTT G
JJH1461_pslA-F1-RT	AAG ATC AAG AAA CGC GTG GAA T
JJH1462_pslA-R1-RT	TGT AGA GGT CGA ACC ACA CCG
JJH1463_cdrA-F1-RT	TCA ACC CCA ACG AGA TCA AGA
JJH1464_cdrA-R1-RT	CGA AGC CCT TCC AGT TGA TG
JJH1467_pelA-F1-RT	CCT TCA GCC ATC CGT TCT TCT
JJH1467_pelA-F1-RT	TCG CGT ACG AAG TCG ACC TT
PA1684-HKG-RT-2F	TGA GCA GCC TTA CCG TCT ATC A
PA1684-HKG-RT-88R	CCA GGG TCG AAG CGA TGT

Bold indicates a restriction site; italics denote a region of reverse complementary for SOE-PCR; underlined sequences are complementary to the target amplicon; and lowercase denotes a nucleotide change for site-directed mutagenesis.

Static and Flow Cell Biofilm Experiments.

Crystal violet quantification of biofilms grown in basal medium 2 (BM2) in static microtiter plates was carried out using established protocols (23). Flow cell biofilms were grown in 1% tryptic soy broth (TSB) and then imaged using a Leica TCS SPE confocal microscope as described in SI Materials and Methods. Z-stacks of 2D confocal images were rendered in three dimensions using Imaris (Bitplane). Last, biofilms were cultivated using a silicone tube model (24) in BM2 at 37 °C for gene expression measurements (SI Materials and Methods).

Nucleotide Extractions and Quantification by LC-MS/MS.

Nucleotide pools were extracted from cultures grown for ∼5 h at 37 °C in BM2 using ice-cold acetonitrile/methanol/water [40%/40%/20% (vol/vol/vol)] according to established procedures (25). Protein concentrations were determined from the cell pellets of 1-mL aliquots from these cultures using a bicinchoninic acid (BCA) Protein Assay kit (Pierce). GMP, pGpG, and c-di-GMP concentrations were determined by LC-MS/MS using a cXMP or [¹³C¹⁵N]c-di-GMP internal standard as previously described (25).

RNA Isolation, cDNA Synthesis, and RT-PCR.

Approximately 1.0 × 10⁹ cells grown in BM2 were harvested at an OD₆₀₀ = 0.5, mixed with RNAprotect Bacteria Reagent (Qiagen), and stored at −80 °C. Isolation of total RNA and assessment of its quantity and quality were performed as previously described (26) (SI Materials and Methods). First-strand synthesis was carried out using the qScript cDNA Supermix kit (Quanta Biosciences). Quantitative PCR (27) measurements were made with a CFX96 Touch Real-Time PCR Detection System (BioRad) using SsoAdvanced Universal SYBR Green Supermix (BioRad).

Protein Production.

Recombinant Orn and RocR were produced by using 0.2% l-arabinose and 0.1 mM isopropyl-β-d-1-thiogalactopyranoside to induce the expression of Orn-His×6 (from pDC15) and RocR-His×6 (from pET26b::rocR) (18) in E. coli BL21(DE3) and Nico21(DE3), respectively. Cells were harvested and Orn-His×6 or RocR-His×6 were purified using established protocols for HPLC with a HisTrap HP column (GE Healthcare) (SI Materials and Methods).

RNA Degradation Assays.

Exonuclease assays were carried out using ³³P- and Cy5-labled RNA oligomers (9) electrophoresed and imaged in 40% polyacrylamide gels as described in SI Materials and Methods.

PDE Activity Assays.

Established procedures for PDE assays using whole cell lysates (28) and MANT-c-di-GMP (19) are detailed in SI Materials and Methods.

Interspecies Genetic Complementation.

E. coli UM341 (9), in which ptetO drives orn expression, was transformed with plasmids expressing either E. coli or P. aeruginosa orn, or the appropriate vector control. Genetic complementation was evaluated by assessing bacterial growth with or without of anhydrotetracyline (Atc).

Phylogenetic Analyses.

Bioinformatics tools (29–33) used to identify orn orthologs and domain cofrequencies and to construct gene phylogenies are described in SI Materials and Methods.

SI Materials and Methods

Growth Media and Antibiotic Selection.

P. aeruginosa was grown at 37 °C in lysogeny broth (LB), tryptic soy broth (TSB), Vogel-Bonner minimal medium (VBMM), or basal medium 2 (BM2), and E. coli was grown in LB. Semisolid media were prepared by adding 1.0% noble agar to VBMM, and 1.5% agar to BM2 or LB. TSB was purchased from Difco and prepared according to the manufacturer’s directions. LB contained, per liter of ultrapure water, 10 g tryptone, 5.0 g yeast extract, and 5.0 g NaCl. VBMM was prepared as a 10× concentrate, which contained per liter of ultrapure water, 2.0 g MgSO₄⋅7 H₂O, 20 g citric acid, 100 g K₂HPO₄, and 35 g NaNH₄HPO₄⋅4 H₂O and was adjusted to pH 7.0 and sterilized by filtration. The 10× VBMM solution was diluted as needed in sterile ultrapure water. BM2 was prepared as a 10× concentrate, which contained per liter of ultrapure water, 69.7 g K₂PO₄, 29.9 g KH₂PO₄, and 9.25 g (NH₄)₂SO₄ and was adjusted to pH 7.0 and sterilized by filtration. Subsequently, a 10× glucose solution, containing 40.0 g/L glucose, and a 50× MgSO₄ solution, containing 50 mM MgSO₄⋅7 H₂O, were prepared separately in ultrapure water and sterilized by filtration. One liter of ready-to-use 1× BM2 was prepared by adding 100 mL of 10× BM2, 100 mL of 10× glucose, and 20 mL of 50× MgSO₄ to 780 mL of sterile ultrapure water.

Where appropriate, antibiotics were added to maintain or select for plasmids as follows: for E. coli, ampicillin (Ap) at 100 μg/mL, gentamicin (Gm) at 10 μg/mL, and kanamicin (K_m) at 50 μg/mL; for P. aeruginosa, carbenicillin (Cb) at 300 μg/mL, trimethoprim (Tp) at 300 μg/mL, and Gm at 100 μg/mL The inducers Atc and l-arabinose were added at 250 ng/mL and 2.0 mg/mL, respectively, as required.

Molecular Methods.

All basic molecular and microbiological techniques were executed according to standard protocols (34). Genomic DNA (gDNA) was purified using the DNeasy Blood and Tissue Kit (Qiagen), plasmids were purified with the QIAprep Spin Miniprep Kit (Qiagen), and DNA was extracted from agarose gels with the QIAEX II Gel Extraction Kit (Qiagen). Comparable kits for nucleotide purification were also purchased from BioBasics. All restriction enzymes were purchased from New England Biolabs, and all primers were obtained from Integrated DNA Technologies. c-di-GMP, pGpG, and MANT-c-di-GMP were purchased from Biolog. Transformations of P. aeruginosa were carried out using established protocols for electroporation (35).

Plasmid Construction.

For the purposes of complementation analyses and intracellular c-di-GMP depletion, P. aeruginosa orn, PA2133, and E. coli orn were cloned, respectively, into the E. coli–P. aeruginosa shuttle vector pUCP18 (36). To begin, the ORF of P. aeruginosa orn was amplified by PCR with primers DC103 and DC104, PA2133 with primers DC121 and DC122, and E. coli orn with primers DC140 and DC141 (Table S2). These primers were tailed with EcoRI and BamHI sites for P. aeruginosa orn and PA2133, and with EcoRI and SalI for E. coli orn. PCR products for P. aeruginosa orn, PA2133, and E. coli orn were gel extracted, then restricted and ligated in pUCP18 by standard methods, creating pDC2, pDC7, and pDC14, respectively (Table S1).

For the purpose of genetic complementation of the E. coli orn conditional mutant (UM341), the P. aeruginosa orn orthologue was cloned into the E. coli shuttle vector pBAD18 (Table S1). To begin, the orn ORF was amplified by PCR with primers DC129 and DC131 (Table S2). These primers were tailed with EcoRI and XbaI sites. The PCR product for orn was gel extracted, then restricted and ligated into pBAD18 by standard methods, creating the plasmid pDC12 (Table S1). Site-directed mutations to pDC2, which targeted the predicted active site residues of Orn, were constructed using the QuikChange Lightening Site-directed Mutagenesis Kit (Agilent Technologies) according to the manufacturer’s directions. Here, primer pairs DC356 and DC357, DC358 and DC359, and DC360 and DC361 were used to introduce D11A, E13Q, and D162 mutations into Orn, respectively, resulting in plasmids pDC57, pDC58, and pDC59, respectively. Mutations were confirmed by Sanger sequencing.

For the purpose of constructing a clean, unmarked deletion of orn in the PAO1 genome, a deletion allele was created in vitro and inserted into the pEX18Gm (37) allelic exchange vector. The deletion allele was built by removing an in-frame segment of coding sequence, constituting >95% of the gene, from the target ORF. Assembly of this deletion allele was accomplished by joining two PCR products, which were amplified using primers that targeted the adjacent regions upstream (JJH98 and JJH99) and downstream (JJH100 and JJH101) of orn, via splicing by overlapping extension (SOE) PCR. The upstream forward (JJH98) and downstream reverse (JJH101) primers used to generate this deletion allele were tailed with HindIII and PstI sites. Subsequently, the deletion allele was restricted and ligated into the pEX18Gm allelic exchange vector, generating pJJH14. Allelic exchange vectors with ΔpslD (pHL129) and ΔpelF (pEX18Gm::ΔpelF) alleles were obtained from previous studies (38, 39).

For the purpose of protein production, an arabinose inducible expression vector for Orn C-terminally tagged with His×6 was created by using PCR to amplify P. aeruginosa orn (with primers DC129 and DC130; Table S2), and subsequently, EcoRI/SalI was used to restrict and ligate this fragment into pUM407 (9), creating the plasmid pDC15. This plasmid was sequence verified using primers DC132 and DC133 (Table S2).

Strain Construction.

Gene replacement was carried out using established tools (39) and methods (37). Allelic exchange vectors (pJJH14, pHL129, and pEX18Gm::ΔpelF; Table S1) were delivered from the donor strain E. coli S17.1 (λ pir) into P. aeruginosa by conjugation by standard procedures (19). Merodiploids were selected on VBMM agar containing 100 μg/mL Gm, and counter-selection was carried out on no salt LB agar that contained 15% sucrose. Clean deletions of orn, pslD, and pelF were identified by PCR using primers specific for the genomic regions flanking the target gene. To verify the mutation, these PCR products were also sent for Sanger sequencing. All unmarked KOs were verified for sensitivity to Gm and resistance to sucrose. Multiple gene deletions were generated by repeating this gene replacement process using the strains that were generated with single mutations.

Estimation of Single-Cell Volume by Light Microscopy.

The dimensions of P. aeruginosa PAO1 cells were measured directly by light microscopy. To do this, cells were grown in BM2 to an OD₆₀₀ = 0.5 and then diluted 10-, 100- or 1,000-fold in fresh medium. A 5-μL aliquot from each of these diluted cultures was transferred onto 1-cm² agarose pads (made from 1% agarose in BM2 and set in BioRad Mini-PROTEAN Tetra Cell Casting Module) placed on glass slides. Glass coverslips were placed on top of the agarose pads. Cells were viewed at 1,000× magnification via white light microscopy using an Olympus IX70 Wide-Field Microscope. Images were captured using Volocity (Perkin-Elmer), and the point tool was used to measure the width and length of 100 individual cells. This approach revealed an average cell length of 2.67 ± 0.30 µm and width of 1.02 ± 0.15 µm. Subsequently, the volume of the average P. aeruginosa cell was mathematically approximated to that of a cylinder (V_cell = πr²h), which was 2.17 × 10⁻¹⁵ L (or ∼2.2 fL).

Estimation of Single-Cell Protein Mass.

The average protein mass of a single P. aeruginosa cell (m_cell) was estimated by counting the number of viable cells and measuring the amount of protein in WT exponential phase cultures (harvested from BM2 at an OD₆₀₀ = 0.5). Under these conditions, viable cell counting estimated the population size at 7.8 × 10⁸ CFU/mL. In tandem with this assay, cells from 1-mL aliquots of culture were harvested by centrifugation (∼13,000 × g for 2 min). Cell pellets were solubilized in 100 μL Laemmli buffer (BioRad) at 95 °C for 20 min. Protein concentrations were determined using the Pierce 660 nm Protein Assay with Ionic Detergent Compatibility Reagent (Fisher Scientific) and prediluted BSA standards. These cultures yielded 2.1 × 10⁻⁴ g protein/mL, or ∼0.27 × 10⁻¹² g protein per cell. This result is close to the expected value because an E. coli cell, which has half the volume of a P. aeruginosa cell (∼1 fL), has about half the protein: the dry mass of an E. coli cell is estimated at 0.28 × 10⁻¹² g and ∼52.4% is protein (0.14 × 10⁻¹² g) (40).

Flow Cell Biofilms.

Biofilm flow chambers were inoculated with an overnight culture of P. aeruginosa, which was diluted to an OD₆₀₀ of 0.15 in 1% TSB, according to established protocols (24). Flow was initiated after 1 h using a peristaltic pump and a flow rate of ∼10 mL/h. An upright Leica SPE TCS Laser Scanning Confocal Microscope (LSCM) was used to image WT and Δorn strains expressing GFP from pMRP9-1, as well as the Δorn strain complemented with pDC2, which expressed GFP from pHKT2 (Table S1). GFP was excited at 488 nm, and fluorescence emission was collected in the range of 504–530 nm.

Tube Biofilms.

Biomass from tube biofilms was harvested for the purpose of orn gene expression studies. A silicone-elastomer tube biofilm system was assembled according to established methods (24). P. aeruginosa biofilms were grown in BM2 at 37 °C, and a flow rate of 10 mL/h was maintained using a peristaltic pump. Biofilms were collected by longitudinally dissecting the silicone tube with a razor blade and then by removing biomass using a cell scraper. Biofilms were suspended in 1 mL BM2 medium containing RNAprotect Bacteria Reagent.

RNA Isolation for RT-PCR.

Cells were lysed using one of two methods. For pelA, pslA, and cdrA expression, cell pellets that were stored in RNAprotect Bacteria Reagent were thawed, suspended in QIAzol Lysis Reagent, and disrupted by bead beating using standard protocols (26). Alternatively, for analysis of orn expression, cell pellets were thawed, suspended in 90 µL of 1 mg/mL lysozyme solution, treated with 10 µL proteinase K solution (Qiagen), and suspended in 1 mL of TRI Reagent (Sigma-Aldrich).

Regardless of the method of cell lysis, RNA was isolated using chloroform extraction and purified using an RNeasy Mini Kit (Qiagen). RNA was treated with Turbo DNase (Ambion) and purified using an RNeasy MinElute Cleanup Kit (Qiagen). Quality of the purified RNA was assessed on 1% agarose gels prepared with NorthernMax-Gly-10× buffer (Ambion). Elimination of genomic DNA was verified by PCR using the Expand Long Template PCR System (Roche) with primers for rplU (Table S2). The estimated detection limit of this PCR-based quality control assay was ∼25 pg gDNA per reaction.

Recombinant Production and Purification of Orn.

E. coli BL21(DE3) transformed with pDC15 was grown at 30 °C to an OD₆₀₀ of 0.7–0.9. Expression of C-terminally tagged Orn-His×6 was induced by the addition of l-arabinose, after which cultures were incubated at 30 °C for an additional 2 h. Cells were harvested by centrifugation at 5,000 × g for 10 min, washed once with PBS (pH 8.0), and then collected again by centrifugation. Pellets were flash frozen in liquid nitrogen and stored at −80 °C for up to 1 mo. Frozen pellets were thawed and suspended in 12.5 mL binding buffer (PBS, pH 8.0, containing 10 mM imidazole and cOmplete EDTA-free protease inhibitor mixture). Cells were lysed by incubating on ice with 0.5 mg/mL lysozyme, followed by high-frequency sonication using ten 30-s pulses with 30-s intervals for heat dissipation. Cell debris was removed by centrifugation at 15,000 × g for 30 min at 4 °C. Recombinant Orn-His×6 was purified in one batch on a 1 mL HisTrap HP Ni-affinity Column (GE Healthcare) according to the manufacturer’s directions. The column was washed with 5 mL wash buffer (PBS, pH 8.0, containing 20 mM imidazole), and subsequently, protein bound to the column was eluted with 50 mL elution buffer (PBS, pH 8.0, containing 250 mM imidazole and cOmplete EDTA-free protease inhibitor mixture) in 1-mL fractions. The fractions were analyzed by SDS/PAGE. Polyacrylamide (PAA) gels were stained with Coomassie blue, or alternatively, transferred to a nitrocellulose membrane and probed with horseradish peroxidase (HRP)-conjugated anti-His×6 antibodies (Roche). Antibody-bound proteins were detected using an enhanced chemiluminescence (ECL) substrate and standard procedures for Western blotting. The fraction with highest yield and >95% purity was dialyzed against 50 mM Hepes (pH 7.5) and 100 mM NaCl. Purified Orn-His×6 was stable and was stored at 4 °C.

Recombinant Production and Purification of RocR.

Recombinant RocR-His×6 was produced from E. coli NiCo21 (DE3) transformed with pET26b::RocR using an established procedure (18). Initially, an inoculum was grown at 30 °C for 24 h, which was then diluted 1:100 in 400 mL of superbroth with 50 µg/mL K_m in 2 L baffled flasks. Five of these production cultures were grown to an OD₆₀₀ of 0.85 at 37 °C on a platform shaker set to 250 rpm. Next, the flasks were cooled at 28 °C for 15 min, and the cultures were induced with 100 µM isopropyl-β-d-1-thiogalactopyranoside (IPTG). The induced cultures were incubated 16 h at 28 °C at 250 rpm, and the cells were harvested by centrifugation at 3,000 × g for 30 min at 4 °C. The cells were washed twice, once in 400 mL cold Lysis-Bind-Wash (LBW) buffer [50 mM Tris⋅HCl, 250 mM NaCl, 10 mM imidazole, 5% (vol/vol) glycerol, 0.5 mM DTT, pH 7.5], and once with 80 mL of cold LBW buffer, each time being pelleted by centrifugation at 3,000 × g for 20 min at 4 °C. The resulting cell pellets were frozen at −80 °C and were stored for up to 1 wk. To lyse the cells, pellets were thawed on ice and suspended in 3 mL LBW buffer per gram of cells. One pellet of cOmplete Mini-EDTA Free Protease Inhibitor (Roche) was added per 10 mL of cells. Cells were lysed using a probe-tip sonicator (12 cycles of 45-s sonication followed by 1-min cool down on ice). The cell lysate was centrifuged for 1 h at 3,000 × g to remove cellular debris and then at 20,000 × g for 1 h to yield a soluble fraction, which was diluted to 7 mg/mL in LBW buffer, flash frozen in liquid nitrogen, and stored at −80 °C.

To purify the RocR-His×6 protein, the soluble fraction was first thawed in an ice water bath and filtered with a 0.45-µm syringe filter (PALL Acrodisc; Supor Membrane). Next, 10 mL soluble fraction was loaded onto a 5-mL HisTrap FF column (GE Healthcare) using LBW buffer. Subsequently, the protein was eluted with elute buffer (50 mM Tris⋅HCl, 250 mM NaCl, 250 mM imidazole, 5% glycerol, 0.5 mM DTT, pH 7.5) using a block elution profile of 18%, 74%, and 100% elute buffer. The protein fractions were pooled and concentrated with a 10K Amicon Ultra-15 centrifugal filter column. The protein was then desalted by diluting the protein with storage buffer (50 mM Tris⋅HCl, 250 mM NaCl, 25 mM KCl, 5 mM MgCl₂, 40% glycerol, 1 mM DTT, pH 8.0) and concentrated with a second 10K Amicon Ultra-15 centrifugal filter. After desalting, the purified RocR-His×6 protein (which was >10 mg/mL) was stored at −20 °C.

Synthesis of RNA Oligomers.

The RNA 24-mer (5′-CACACACACACACACACACACACA-3′) was labeled at the 5′-end using γ-³³P-ATP. Radioactive labeling was accomplished using the MirVana Probe and Marker Kit (Ambion) to set up a 93-µL reaction containing 450 pmol of the 24-mer, 20 pmol γ-³³P-ATP, 90 pmol ATP, and 4.5 µL T4 polynucleotide kinase. The reaction was incubated for 1 h at 37 °C. The reaction was quenched with 10 µL of 10 mM EDTA (pH 8.0) and then incubated at 95 °C for 2 min. The γ-³³P-labeled 24-mer was purified using a NucAway Spin Column (Ambion). The γ-³³P-ATP-labeling of RNA standards was carried out using the Decade Marker System (Ambion). Cy5-labeled RNA homo- and heteropentamers were purchased from Integrated DNA Technologies.

Exonuclease Assays.

The nuclease activity of Orn-His×6 was assessed in reaction buffer, which contained 50 mM 1,3-bis(tris(hydroxymethyl)methylamino)propane (BTP, pH 8.0), 5 mM MgCl₂, and 0.1% (wt/vol) BSA as previously described (9). Approximately 75 pmol of ³³P- or Cy5-labled RNA substrate and 100 ng of clean purified Orn-His×6 were added to 45 µL reaction buffer to give a final reaction volume of 50 µL. Reactions were heat inactivated for 3 min at 95 °C. Reactions were loaded on denaturing 22% (24-mer) or 40% (pentmers) PAA gels that contained 7 M urea and were electrophoresed in 2× Tris-borate-EDTA (TBE) buffer.

Quantitative assays for the degradation of pGpG were set up in 100-µL aliquots of reaction buffer. Assays contained 100 ng of clean purified Orn-His×6 and a final concentration of 100 μM c-di-GMP or pGpG. Reactions were quenched by the addition of 4.2 µL of 0.5 M EDTA. The quenched reaction was centrifuged at 20,000 × g for 5 min at 4 °C to remove precipitated protein. The supernatant was transferred to a new tube and sent for nucleotide analysis by LC-MS/MS.

Assays for PDE Activity in Cell Lysates.

Appropriate P. aeruginosa strains were grown in 5 mL BM2 at 37 °C to an OD₆₀₀ of ∼0.6. Cells were harvested by centrifugation at 1,700 × g for 5 min at 4 °C. Cell pellets were washed once with 5 mL PBS and then collected again by centrifugation. Subsequently, the supernatant was discarded, and pellets were suspended in 800 μL PDE assay buffer, which contained 50 mM Tris⋅HCl (pH 8.0), 10 mM MgCl₂, 250 mM NaCl, 5 mM β-mercaptoethanol, 1 mM PMSF, and cOmplete EDTA-free protease inhibitor mixture (Roche). The cells were flash frozen in liquid nitrogen and stored at −80 °C. The next day, cells were thawed on ice and lysed using high-frequency sonication. Afterward, c-di-GMP was added to the lysates at a concentration of ∼67 nM, and pGpG was added to a portion of these at an initial concentration of either 2 or 20 μM. The cell lysates were then divided into 100-μL aliquots. Aliquots of these lysates were incubated at 30 °C for 30, 60, 120, and 180 min for PA2133 and for 60 min for PvrR and then heat-inactivated at 95 °C for 10 min. Finally, the lysates were centrifuged at 20,000 × g for 20 min, and supernatants were transferred to new tubes. c-di-GMP was measured in each sample by LC-MS/MS. All assays were performed in triplicate.

MANT-c-di-GMP Assays for PDE Activity of Purified RocR.

Time-dependent kinetics of RocR-His×6 activity were tracked using MANT-conjugated c-di-GMP (19). Assays were conducted in 100-µL reactions with 1 μM RocR-His×6 in a reaction buffer containing 50 mM Tris⋅HCl (pH 8.0), 250 mM NaCl, 25 mM KCl, 5 mM MgCl₂, and 1 mM DTT. The reaction was started by the addition of 0.4 μM MANT-c-di-GMP with or without 0.4 mM pGpG and 1 μM Orn. Fluorescence was monitored in real time using a Synergy 4 microplate reader (BioTek) with an excitation wavelength of 355 nm and emission wavelength of 440 nm.

LC-MS/MS Assays for PDE Activity of Purified RocR.

Quantitative assays for the degradation of c-di-GMP were set up in 100-µL aliquots of reaction buffer (described above). Assays contained 1 μM of clean, purified RocR-His×6 and a final concentration of 100 nM c-di-GMP with or without 80 μM pGpG. Reactions were quenched by the addition of 4.2 µL of 0.5 M EDTA and then were heat inactivated for 10 min at 95 °C. The quenched reaction was centrifuged at 20,000 × g for 20 min at 4 °C to remove precipitated protein. The supernatant was transferred to a new tube and sent for nucleotide analysis by LC-MS/MS.

Bacterial Species Surveys to Identify orn Orthologs.

The distribution of orn gene orthologs in bacteria was determined from NCBI tBLASTn (29) searches of 3,109 completely sequenced bacterial genomes using the P. aeruginosa PAO1 Orn protein (NP_253638) as the query sequence. Orn-based searches were performed according to phylum and class to ensure complete search saturation (ranging from 135% to 153% database coverage based on the ratio of nonredundant to redundant hits) of the NCBI genome database. All orn homologs identified from these surveys were collected based on e-value cutoffs at or below 1 × 10⁻⁵, and redundant sequences were eliminated by sequence identity and genome accession numbers resulting in a total of 568 bacterial species with at least one orn ortholog.

Bacterial genome psi-BLAST searches for EAL and HD-GYP domain-containing genes were performed using the search methods and conditions previously described (1). All bacterial species with predicted domains were combined with species possessing Orn orthologs. The co-occurrence of these domains was determined using pairwise comparisons between the Orn dataset and each domain dataset.

Phylogenetic Analysis of Orn Proteins.

The 569 Orn protein sequences were aligned with COBALT (30). This dataset was reduced to reflect key taxonomic differences between the identified classes using BLASTClust (29), resulting in a final set of 170 Orn protein sequences. These sequence alignments were further refined using Jalview Version 2 (31) to remove gaps and unalignable regions within the Orn protein dataset. Subsequently, phylogenetic analyses was performed using MrBayes 3 (32). The Bayesian inference of the Orn dataset was completed using 10 million generations, sampled at every 1,000th generation with a 25% burn-in setting and a mixed amino acid model. This analysis reached convergence after 6 million generations where the SD of the split frequencies reached a value of 0.005. Bayesian trees were generated using Archaeopteyrx (33), and all trees generated by this program included confidence values of ≥60% at significant branch nodes.