All experiments were performed in biosafety level 3 (BSL-3) facilities within the Regional Biocontainment Laboratory at Colorado State University with approvals from the Center for Disease Control and Prevention and the Colorado State University Institutional Biosafety Committee. B. pseudomallei strain 1026b, a clinical isolate from a human case of septicemic melioidosis in Thailand (Deshazer et al., 1997), was used in this study. Transposon (T24) mutant strains used in this study were generated during the production of a comprehensive two-allele sequence defined transposon mutant library (manuscript in preparation). T24 is a Tn5-derived transposon containing a kanamycin resistance selection marker approved for use in B. pseudomallei that was constructed in Colin Manoil’s laboratory at the University of Washington (manuscript in preparation). Briefly, the two-allele library was assembled via random transposon insertion into gene loci spanning both chromosomes of B. pseudomallei 1026b, resulting in insertions into 4,931 of 6,070 (81.2%) total predicted genes. In total, the library contains two copies of 4,415 genetic mutants that were re-sequenced. All transposon insertion mutants were verified by additional sequencing for these studies. Bacterial strains were grown in Lysogeny Broth (LB) media consisting of 10 g/L tryptone (Fisher Scientific), 5 g/L yeast extract (Becton, Dickinson and Company), and 5 g/L NaCl (Fisher Scientific) as previously described (Plumley et al., 2017). For experiments involving media with sodium nitrate or sodium nitrite, LB was supplemented with a final concentration of 10 mM sodium nitrate (Sigma) or 10 mM sodium nitrite (Matheson, Coleman & Bell) unless otherwise stated. Transposon mutants were grown in LB with 300 μg/mL kanamycin (Gold Biotechnology).

A combination of open-access genomic databases were used to assess the comparative identity of genetic orthologs in closely related Burkholderia spp. as well as P. aeruginosa strain PAO1. Predicted nucleotide sequences in the B. pseudomallei 1026b (tax id: 884204) genome for chromosomes I and II (accession numbers NC_017831.1 and NC_017832.1, respectively), were acquired from the NCBI GenBank database using the BLASTN program under default parameters to identify sequences similar to the PAO1 reference genome (GenBank assembly accession: GCA_000414275.1). Functional nitrogen metabolism pathways were evaluated using the Kyoto Encyclopedia of Genes and Genomes (KEGG) (Ogata et al., 1999). Further comparative genomic analyses, as well as gene annotations and locus positions were deduced using the Burkholderia Genome Database (Winsor et al., 2008) in conjunction with the Pseudomonas Genome Database (Winsor et al., 2015). Sequence comparison among nitrogen metabolism genes was analyzed with the following: B. pseudomallei K96243 (tax id: 272560), B. thailandensis E264 (tax id: 271848), and P. aeruginosa PAO1 (tax id: 208964). Percent identity between genetic orthologs was calculated using the Multiple Sequence Comparison by Log-Expectation (MUSCLE) tool provided by the European Bioinformatics Institute (EMBL-EBI), which calculates a percent identity matrix using Clustal 2.1 (Edgar, 2004). Gene names and functional descriptions were not entirely available for B. pseudomallei 1026b, and in such cases, were assigned based on consensus among the better annotated species detailed above.

Sequence analyses for the predicted nitrogen metabolism clusters encoded on chromosome I and II from the sequenced genome of B. pseudomallei 1026b (tax id: 884204) were visualized using open-source bioinformatics platforms and genomic sequence databases. We used BLASTN (BLAST, NCBI) under default parameters in conjunction with the Burkholderia Orthologous Groups cataloging system from the Burkholderia Genome Database¹ to identify homologous regions. Sequences for B. pseudomallei 1026b chromosome I (accession number: NC_017831.1) and chromosome II (accession number: NC_017832.1) were downloaded from the GenBank database (NCBI, NIH). Sequence fragments were extracted with Geneious version 7.1.7² (Kearse et al., 2012) and visualized with EasyFig version 2.23 (Sullivan et al., 2011) using Python version 2.7³. To calculate homology between input sequences, command line parameters for EasyFig included a minimum identity cut-off for BLASTN of 0.60 and a threshold E-value of 1E-3.

Overnight cultures were grown in LB medium (kanamycin as indicated for transposon mutants) and sub-cultured 1:50 in 2 mL LB with or without 10 mM sodium nitrate. Pellicles were grown in 17 mm × 100 mm polystyrene tubes (#1485-2810, USA Scientific) at 37°C for 24 h before incubation at room temperature for 14 days without disturbance then photographed.

Overnight cultures were grown in LB and diluted to OD₆₀₀ of 0.1 in LB with or without 10 mM sodium nitrate. Experiments using sodium nitrite were performed in the same manner as those using sodium nitrate. Gradient concentrations {0, 1, 2, 3, 5, 8, 10, 20, 50, and 100 mM} of sodium nitrate, sodium nitrite, or sodium chloride in LB were evaluated as indicated. One hundred microliters was added, in replicates of six, to individual wells of a 96-well polystyrene plate (Nunc^TM Microwell^TM 96-well microplates #243656, Thermo Scientific) and incubated at 37°C for 24 h. Static biofilms were processed as previously described (Plumley et al., 2017). Briefly, bacterial supernatant was removed from the biofilms and wells were washed once with PBS to remove non-adherent cells. The wells were stained with 0.05% crystal violet (Sigma–Aldrich, St. Louis, MO, United States) and incubated at room temperature for 15 min. Stain was removed and wells were washed again with PBS. Adherent crystal violet was solubilized by addition of 95% EtOH and incubated for 30 min. The solubilized dye was transferred to a new 96-well polystyrene plate and the absorbance was measured at OD₆₀₀ on a Synergy HT plate reader (BioTek Instruments, Winooski, VT, United States). Significance was calculated with an unpaired Student’s t-test using the Holm-Sidak method to account for multiple comparisons.

Complementation constructs were designed for conditional expression of re-integrated genes via IPTG induction from the tac promoter, following select-agent-compliant methods for genetic modification in B. pseudomallei (Choi et al., 2008; Plumley et al., 2017). The full-length sequences of Bp1026b_I1013 (narL) and Bp1026b_I1014 (narX) were amplified using the following PCR primer pairs: (5′-NNNCCCGGGAGGAGGAT ATTCATGACCATACGGGTACTGTT-3′) and (5′-NNNAAGC TTTTATGCCTCGGCCGGATGCG-3′) for Bp1026b_I1013, and (5′-NNNCCCGGGAGGAGGATATTCATGGCTCCCGCC CTCCCCGA-3′) and (5′-NNNAAGCTTCTATGCCGCCTGTC GCGCGT-3′) for Bp1026b_I1014. Cloned genes were ligated into pUC18T-mini-Tn7T-Km-LAC, which uses the tac promoter from Escherichia coli to drive gene expression (Rholl et al., 2011). 5 mM IPTG was used to induce expression from the tac promoter. 5 mM NaNO₃ was used to stimulate biofilm inhibition.

Overnight cultures were grown in LB and diluted to OD₆₀₀ of 0.1 in LB supplemented with increasing concentrations of sodium nitrate or sodium chloride as indicated. Sodium chloride was supplemented in addition to the original 170 mM NaCl concentration in LB. Each experimental growth condition was repeated in six replicates. One hundred microliters of each culture was added to a 96-well polystyrene plate (Nunc^TM Microwell^TM 96-well microplates #243656, Thermo Scientific) in triplicate. Plates were incubated at 37°C shaking aerobically at 250 rpm for 48 h. Absorbance (OD₆₀₀) was measured hourly on a Synergy HT plate reader (BioTek Instruments).

Nucleotide extraction methods using formic acid were adapted from Massie et al. (2012). Overnight cultures of B. pseudomallei 1026b were grown in LB at 37°C shaking (250 rpm). Cultures were diluted 1:50 in 4 mL of M9 Salts minimal medium with or without 10 mM sodium nitrate supplemented and grown statically at 37°C for 18 h in six-well Costar polystyrene plates (#3516, Corning). Cells from 1 mL culture including the pellicle biofilm were collected and resuspended in 100 μL of cold 40% (vol/vol) acetonitrile/40% (vol/vol) methanol/20% (vol/vol) LC-MS-grade water/1% formic acid. An internal control of 2-chloro-adenosine-5′-O-monophosphate (2-Cl-5′-AMP, Axxora, LLC) was used at a final concentration of 100 nM (Irie and Parsek, 2014). The buffer, cells, and matrix were incubated at 65°C for 10 min and then immediately transferred to -20°C for 30 min to ensure complete lysis. Samples were then centrifuged at 16,000 × g at 4°C for 5 min and the supernatant fraction containing c-di-GMP was collected and neutralized with 20 μL 1 M KOH, followed by another spin at 16,000 × g at 4°C for 5 min. A calibration curve using known standards of chemically synthesized c-di-GMP (Axxora, LLC) and 2-Cl-5′-AMP was generated and extracted in identical conditions to experimental samples. Absolute quantification of c-di-GMP was calculated using the linear regression equation generated from known standards and peak areas normalized to the internal standard. Analysis was performed on an Acquity M-Class UPLC (Waters) coupled to a Xevo TQ-S triple quadrupole mass spectrometer (Waters) for LC-MS/MS. A Waters Atlantis dC18 stationary phase column (300 μm × 150 mm, 3.0 μm) was used for chromatographic separations. Protein pellets from the initial cell collection were analyzed for total protein concentration using the Pierce 660 nm Protein Assay (Thermo Scientific) for normalization of the absolute concentration of c-di-GMP. Nucleotide extraction experiments were repeated on two different days using four biological replicate cultures with three technical replicates each. Statistical significance was determined using an unpaired Student’s t-test in GraphPad Prism.

Burkholderia pseudomallei 1026b was grown statically in M9 salts minimal media with or without 10 mM sodium nitrate supplemented at 37°C in six-well Costar polystyrene plates (Corning). After 18 h of static growth, 1 mL of culture was collected and cells were centrifuged at 12,000 × g for 2 min at 4°C. Bacterial pellets were resuspended in 350 μL RNAprotect Bacteria Reagent (Qiagen) before centrifuging at 5,000 × g for 10 min at 4°C. Bacterial pellets were resuspended in 1 mL QIAzol Lysis Reagent (Qiagen), incubated for 5 min at room temperature, and transferred to screwcap tubes containing 250 μL of 0.1 mm glass disruption beads (RPI Corp.) on ice. Cells were lysed using a TissueLyser II homogenizer (Qiagen) using three rounds of 60 s pulses at 30 Hz with 60 s on ice between each lysis round. After lysis, the cell mixture was incubated at room temperature for 5 min before adding 200 μL chloroform and vortexing for 5 s. Samples were incubated again at room temperature for 5 min before centrifugation at 13,000 × g for 10 min. The aqueous phase was collected and RNA was purified with an RNeasy Mini Kit (Qiagen) using the protocol recommended by the manufacturer. RNA samples were treated with two rounds of TURBO DNase (Ambion) before cDNA was synthesized using the Transcriptor First Strand cDNA Synthesis Kit (Roche). qRT-PCR was performed on a LightCycler 480 System with SYBR Green I Master (Roche) using 10 ng total cDNA and the following cycling conditions: pre-incubation at 95°C for 10 min; three-step amplification at 95°C for 10 s, 55°C for 15 s, 72°C for 15 s; melting curve at 95°C for 5 s, 65°C for 1 min, and 97°C for continuous acquisitions per 5°C. Primers were designed using the PrimerQuest tool (IDT DNA Technologies). Previously published 23S rRNA primers were used for normalization (Mima and Schweizer, 2010). All sets of primers were validated for amplification efficiency in a standard curve. The primer set used for cdpA was as follows: forward (5′-AAGCTGGCTGGAGCAAA-3′) and reverse (5′-GCAGATAGTCGCGGTGATAA-3′). The primer set used for Bp1026b_II2523 was as follows: forward (5′-GCAAGATCGAGAGCGTGTT-3′) and reverse (5′-GTGATCGAAGCGGAACAGATAC-3′). The primer set used for Bp1026b_ II0885 was as follows: forward (5′-CTGCACCTACCGCTTCTT-3′) and reverse (5′-AAGCACAGCGAGAAGTAGTC-3′). The primer set used for Bp1026b_I3148 was as follows: forward (5′-CGCTGCATCTGGAGAACTT-3′) and reverse (5′-CCTGAAACGGATCGGTGAG-3′). Transcript abundance was measured using the Pfaffl method (Pfaffl, 2001), taking into account primer amplification efficiency, including three independent biological samples performed in technical triplicates.

While the effectors of biofilm dispersal and inhibition are extensive and diverse, nitrogenous compounds have been shown to modulate biofilm formation in both Gram-negative and Gram-positive bacteria (Schlag et al., 2007; Van Alst et al., 2007; Zemke et al., 2014). Sodium nitrate (NaNO₃), a naturally occurring salt compound and a synthetic agricultural additive, inhibits biofilm formation in a dose-dependent manner (Figure 1A). Likewise, sodium nitrite also inhibits biofilm formation following a similar trend, although the effect is not as robust initially (Figure 1A). To assess the impact of nitrate/nitrite on biofilm formation, wild-type B. pseudomallei 1026b was grown statically for 24 h with increasing concentrations of sodium nitrate or sodium nitrite (Figure 1A). Biofilm inhibition effects were first noted at 1 mM while the most significant biofilm inhibition occurred with the addition of 10 mM sodium nitrate or 10 mM sodium nitrite. There were no additional inhibitory effects on biofilm formation with treatment concentrations greater than 10 mM NaNO₃. As such, concentrations of 10 mM NaNO₃ were used for all subsequent experiments. To evaluate the effects of sodium concentration on biofilm formation, experiments with only sodium chloride (NaCl) indicated that biofilm inhibition was not significantly altered during biofilm formation under the same conditions (Figure 1A). Viability was also tested by measuring growth kinetics in media supplemented with NaCl (Figure 1B) and NaNO₃ (Figure 1C). Growth inhibition was observed only when cells were grown in media supplemented with 100 mM of NaNO₃ and 100 mM of NaCl, concentrations that are well beyond those used in this study. No significant differences in growth dynamics were observed at 10 mM NaNO₃ suggesting that nitrate dosing at this concentration does not affect cell viability. These results suggest that nitrate sensing and metabolism, and not sodium, is responsible for biofilm inhibition in B. pseudomallei 1026b and the observed inhibition is not the result of decreased cellular viability.

Bioinformatics analyses identified 25 B. pseudomallei 1026b genes predicted to be involved in nitrogen metabolism (Table 1), of which 21 were included in this study based on the availability from a two-allele sequence defined transposon library. The genes included in this study were selected based on sequence homology among closely related Burkholderia spp. to P. aeruginosa PAO1, which has an extensively annotated reference genome that has been biochemically validated for the nitrogen metabolism genes. Within the Proteobacteria phylum, the β-proteobacteria, which contains the genus Burkholderia, is closely related to the genus Pseudomonas, which is a γ-proteobacteria (Estrada-De Los Santos et al., 2013), thus inter-genus comparisons described in this study will provide insights into the functions among these conserved nitrogen metabolism proteins. While B. pseudomallei 1026b has many genes predicted to encode enzymes in the denitrification pathway, there is a need for more comprehensive functional annotation in this relatively new clinically isolated strain. Therefore, nucleotide sequence identity was analyzed across three Burkholderia species in relation to P. aeruginosa PAO1 (Table 1 and Supplementary Table 1).

Orthologs of the denitrification enzymes in B. pseudomallei 1026b presented here share 99–100% nucleotide identity to the fully sequenced genome of B. pseudomallei K96243, 90–97% nucleotide identity to B. thailandensis E264, and 60–78% nucleotide identity to the more distant relative P. aeruginosa PAO1 (Supplementary Table 1). This underscores the highly conserved nature of the genes constituting the essential denitrification pathway among these soil-dwelling bacteria. Interestingly, bioinformatics analysis revealed duplications of specific genes implicated in the denitrification pathway in B. pseudomallei 1026b, which is consistent with the predicted genetic supplementation and metabolic redundancy of B. pseudomallei (Holden et al., 2004). The major nitrate reductases are encoded on chromosome I at Bp1026b_I1015 – Bp1026b_I1018 and on chromosome II at Bp1026b_II1222 – Bp1026b_II1225, and share sequence homology of 69–79% (Supplementary Table 2). Additionally, the major nitrite reductase is encoded on chromosome I at Bp1026b_I2984 – Bp1026b_I2985 and on chromosome II at Bp1026b_II1317 – Bp1026b_II1318, sharing only 65% nucleotide identity (Supplementary Table 2).

Bioinformatics analyses also identified putative nitrogen metabolism genes with no clear annotations across Burkholderia spp. Therefore, Bp1026b_I1019, Bp1026b_I1020, Bp1026b_II1319, and Bp1026b_II1540 were assigned genetic annotations based on the P. aeruginosa PAO1 reference genome. Two genes were identified as assimilatory nitrate reductases (Bp1026b_I2986 and Bp1026b_II1316) based on sequence homology to PAO1; however, no annotated homologs were available throughout Burkholderia spp., thus the annotation was assigned to be consistent with P. aeruginosa PAO1. Lastly, Bp1026b_II1580, which is predicted to encode an anaerobically induced outer membrane nitrite reductase, was specific to Burkholderia spp. with no representative homolog in P. aeruginosa.

Based on the information available from the previously described open-source genome databases and the genetic functional pathway predictions provided by the KEGG database, we mapped and compared the genetic arrangement of the predicted nitrogen metabolism enzymes across both chromosomes (Figures 2A,B). The narX/narL predicted two-component signaling system lies downstream and adjacent to the primary nitrate reductase on chromosome I, which is flanked by the cytoplasmic membrane-associated nitrate/nitrite transporters (Figure 2A). Another gene cluster of interest for these studies is the assimilatory nitrite reductase on chromosome I, with the adjacent molybdopterin oxidoreductase nasA, which is predicted to function as an assimilatory nitrate reductase (Ogawa et al., 1995) (Figure 2B). Interestingly, the nitrate metabolism gene clusters found on chromosome I are replicated in similar configurations on chromosome II. The relatively equal distribution of denitrification-associated loci across both B. pseudomallei chromosomes is not surprising considering that chromosome II is not an accessory genome and also encodes for essential and central metabolism genes (Holden et al., 2004); however, the predicted nitrate metabolism loci on chromosome II are not directly homologous to those on chromosome I. This disparity in sequence identity among similarly annotated gene loci with parallel genetic arrangements indicates that the nitrate metabolism clusters are not direct paralogs and may be required for specific metabolic functions based on extracellular cues.

We hypothesized that pellicle biofilms cannot form in the presence of exogenous nitrate under the conditions tested. Pellicle biofilms have been historically well characterized in the Gram-positive soil-dwelling bacterium Bacillus subtilis (Armitano et al., 2014), as well as pathogenic Gram-negative species such as P. aeruginosa (Cole et al., 1989) and Burkholderia cenocepacia (Fazli et al., 2013). B. pseudomallei forms pellicle biofilms in static cultures in vitro and at the surface-liquid interface near the root zone of plants. Wild-type and transposon insertional mutants of genes in the denitrification pathway were grown statically with or without sodium nitrate to assess the impact of nitrate on pellicle formation. The wild type did not form a pellicle biofilm in the presence of 10 mM NaNO₃, while a robust pellicle biofilm was observed in LB without nitrate (Figure 3). Furthermore, pellicle formation was inhibited by sodium nitrate in 16 of 21 transposon insertional mutants (Supplementary Figure 1). Interestingly, five transposon insertional mutants in the denitrification pathway were resistant to inhibition of pellicle formation by exogenous nitrate (Figure 3 and Supplementary Figure 1). Transposon insertions in narL, a DNA-binding response regulator, and narX, a nitrate sensor protein, which comprise a two-component signaling system regulated by nitrate, were resistant to pellicle inhibition by nitrate. Similarly, transposon mutants with insertions in narG-1 and narH-1, which are predicted to be the major nitrate reductase alpha and beta subunits were also insensitive to inhibition of pellicle formation by nitrate. Interestingly, transposon mutations into narG-1 and narH-1 did not confer resistance to biofilm inhibition by sodium nitrite (Supplementary Figure 2), indicating that nitrite may inhibit biofilm formation independently of the major nitrate reductase. It is curious that only the cytoplasmic subunits of the major nitrate reductase were hit during our transposon screen for the biofilm inhibitory phenotype via sodium nitrate, an observation that may provide insights into the respiratory electron transfer pathway in B. pseudomallei. Additionally, it is worth noting that the membrane-anchored narI-1 gamma subunit of the nitrate reductase was not available in our transposon mutant library, and as such, its relation to biofilm dynamics remains uncharacterized. We identified a fifth transposon mutant with an insertion in narK-1, encoding a predicted nitrate/nitrite transporter that was also insensitive to pellicle biofilm inhibition by nitrate. These results suggest that this genetic quintet, Bp1026b_I1013 (narL), Bp1026b_I1014 (narX), Bp1026b_I1017 (narH-1), Bp1026b_I1018 (narG-1), Bp1026b_I1020 (nark-1), contributes to the regulation and formation of B. pseudomallei biofilms when exogenous nitrate is encountered.

Biofilm inhibition by nitrate was assessed in the same array of transposon insertion mutants using the quantitative biofilm growth method (O’Toole and Kolter, 1998) to calculate results from static biofilm assays. In conjunction with the direct observation of pellicle formation, the static biofilm assay provides quantitative analysis on the biofilm-forming capabilities for each strain in this study. Biofilm formation was significantly abolished with nitrate treatment in 16 of the 21 transposon mutants; however, biofilm formation of transposon mutants in narL, narX, narG-1, narH-1, and narK-1 was not significantly altered in the presence of nitrate (Figure 4A). These results corroborated the qualitative observations of the pellicle biofilm assay, as the same five mutants were similarly resistant to the biofilm inhibitory effects of nitrate in both assays. Planktonic growth for narL, narX, narG-1, narH-1, and narK-1 was similar when compared to wild-type B. pseudomallei 1026b (Supplementary Figure 3).

To further assess the effects of nitrate metabolism genes on biofilm inhibition, we calculated the percent of relative biofilm inhibition for all mutant strains in this study (Figure 4B). While wild-type biofilm formation was reduced by nearly 80% when sodium nitrate is encountered, transposon insertions in narL, narX, narG-1, and narH-1 did not respond to the inhibitory effects of treatment with sodium nitrate. The decrease in the percentage of relative biofilm formation for the transposon-inactivated narK-1 mutant (Figure 4C) indicates that this mutant is affected by nitrate more so than the narL, narX, narG-1, and narH-1 mutants, and is also reflective of a fundamental biofilm-forming defect in the narK-1 mutant. Taken together, these results provide quantitative measurements of biofilm inhibition and formation in the presence of exogenous nitrate, and support our conclusions regarding the role of these functionally inactivated genetic loci as observed in the pellicle biofilm assay.

To confirm that the transposon mutant insertions identified in the T24-mutagenesis screen were causal to the observed phenotypes, we complemented I1013::T24 (narL) and I1014::T24 (narX) with full-length genes which were conditionally expressed with IPTG induction. The resulting constructs, Bp1026b_I1013::T24 P_tac-I1013 (narL) and Bp1026b_I1014::T24 P_tac-I1014 (narX) produced similar amounts of biofilm as the wild-type empty-vector (EV) control strain in the absence of NaNO₃ (Figure 5). Conditional expression of narL in Bp1026b_I1013::T24 P_tac-I1013 and narX in Bp1026b_I1014::T24 P_tac-I1014 resulted in an approximate 50% reduction in biofilm formation with the addition of 5 mM NaNO₃ that was equivalent to the wild-type EV control strain (Figure 5). These data further support the hypothesis that the narX/narL two-component signaling system, which is predicted to respond to nitrate sensing, is implicated in biofilm growth dynamics in B. pseudomallei.

Biofilm formation and dispersal mechanisms are ultimately dependent on the level of intracellular c-di-GMP in many pathogenic bacteria (Tamayo et al., 2007), thus we hypothesized that biofilm inhibition via exogenous nitrate sensing regulates c-di-GMP concentrations in B. pseudomallei 1026b. We analyzed the intracellular concentration of c-di-GMP via liquid chromatography tandem-mass spectrometry (LC-MS/MS) using a triple quadrupole mass spectrometer. The intracellular concentration of c-di-GMP in wild-type bacteria grown statically in minimal media with 10 mM sodium nitrate was reduced by 38% as compared to biofilms cultivated statically in the absence of nitrate (Figure 6A). For static biofilms cultivated without sodium nitrate, we observed an average concentration of 80.46 nmol c-di-GMP per mg of total protein and for biofilms cultivated in the presence of 10 mM nitrate the average concentration was 50.19 nmol c-di-GMP per mg total protein (Figure 6A). These results suggest that nitrate metabolism regulates intracellular c-di-GMP levels in B. pseudomallei 1026b biofilms.

In a simplified model of c-di-GMP regulation, a decrease in intracellular c-di-GMP levels suggests either decreased diguanylate cyclase activity or increased phosphodiesterase activity. We chose to evaluate the expression of previously described diguanylate cyclases and phosphodiesterases that are known to alter biofilm dynamics in B. pseudomallei (Lee et al., 2010; Plumley et al., 2017). C-di-GMP metabolism in bacteria, including B. pseudomallei 1026b, is classically mediated by diguanylate cyclases and phosphodiesterases that contain conserved GG(D/E)EF and EAL catalytic domains, respectively (Tamayo et al., 2007; Plumley et al., 2017). B. pseudomallei 1026b contains six predicted proteins with conserved EAL domains and five predicted GGDEF-EAL composite proteins (Plumley et al., 2017); however, only one of these predicted proteins, Bp1026b_I2284, has been shown to potentially function as a classical phosphodiesterase (Plumley et al., 2017). Bp1026b_I2284 (cdpA) is a previously described phosphodiesterase and known regulator of c-di-GMP levels in B. pseudomallei KHW (Lee et al., 2010). Recently, Bp1026b_I2284 (cdpA) has also been shown to contribute to motility and is therefore potentially implicated in reducing c-di-GMP levels in B. pseudomallei 1026b during planktonic growth (Plumley et al., 2017). Thus, we hypothesized that cdpA transcript expression is regulated during nitrate metabolism in B. pseudomallei 1026b, leading to a decrease in intracellular c-di-GMP and inhibition of biofilm formation.

During the transition between biofilm and planktonic states multiple phosphodiesterases and diguanylate cyclases may be expressed in a coordinated manner to alter the levels of c-di-GMP in response to stimulatory signals and conditions. To begin to address the complexity associated with c-di-GMP signaling during nitrate metabolism, we evaluated the transcription of previously identified phosphodiesterase and diguanylate cyclase genes (Plumley et al., 2017). Bp1026b_II0885, a composite protein that is similar to cdpA, was chosen for analysis based on the conservation of the EAL domain and GGDEF catalytic domains (Plumley et al., 2017). Bp1026b_I3148 was also chosen based on the high level of conservation of the EAL domain (Plumley et al., 2017). To evaluate the effects on diguanylate cyclase gene expression, we chose to analyze Bp1026b_II2523 based on its conserved GGEEF domain and its previously observed role in the thermoregulation of biofilm growth dynamics (Plumley et al., 2017).

To assess the impact of exogenous nitrate on the expression of c-di-GMP phosphodiesterases and diguanylate cyclases, wild-type B. pseudomallei 1026b was grown statically under biofilm-inducing conditions in minimal media with or without 10 mM NaNO₃. The expression of cdpA was significantly higher, representing a nearly twofold increase in cells grown in 10 mM NaNO₃-supplemented minimal media when normalized to 23S (Figure 6B). Interestingly, the expression values for the additional two transcripts that potentially encode c-di-GMP phosphodiesterases showed a similar trend to that of cdpA; however, these analyses lacked statistical significance (Figure 6B). The expression of Bp1026b_II2523, a putative diguanylate cyclase showed no significant response to nitrate, which indicates it may not be involved in the nitrate-dependent biofilm inhibition cascade under the conditions tested. Thus, while nitrate has a clear regulatory effect on cdpA, it is possible that other phosphodiesterases and diguanylate cyclases that remain to be tested also contribute to the intracellular concentration of c-di-GMP. Moreover we showed that cdpA is not required for nitrate- or nitrite-dependent biofilm inhibition (Supplementary Figure 4). Nonetheless, these results indicate that nitrate can potentially increase phosphodiesterase activity in B. pseudomallei 1026b and can explain the significant decrease in c-di-GMP concentration during growth under identical conditions. Taken together with the absolute quantification of c-di-GMP, these data support our hypothesis that exogenous nitrate inhibits B. pseudomallei 1026b biofilms by reducing the intracellular concentration of c-di-GMP.