H-NOX Influences Biofilm Formation, Central Metabolism, and Quorum Sensing in Paracoccus denitrificans

Abstract

The transition from planktonic to biofilm growth in bacteria is often accompanied by greater resistance to antibiotics and other stressors, as well as distinct alterations in physical traits, genetic activity, and metabolic restructuring. In many species, the heme nitric oxide/oxygen binding proteins (H-NOX) play an important role in this process, although the signaling mechanisms and pathways in which they participate are quite diverse and largely unknown. In Paracoccus denitrificans, deletion of the hnox gene results in a severe biofilm-deficient phenotype. Quantitative proteomics was used to assemble a comprehensive data set of P. denitrificans proteins showing altered abundance of those involved in several important metabolic pathways. Further, decreased levels of pyruvate and elevated levels of C₁₆ homoserine lactone were detected for the Δhnox strain, associating the biofilm deficiency with altered central carbon metabolism and quorum sensing, respectively. These results expand our knowledge of the important role of H-NOX signaling in biofilm formation.

Introduction

Paracoccus denitrificans is a well-studied, nonswimming, nonmotile Gram-negative bacterium belonging to the class Alphaproteobacteria.^1,2 It is often used as a model organism for studies of denitrification, respiratory chains, and polyhydroxyalkanoate production due to its metabolic versatility.^3,4 These features make biofilms formed by P. denitrificans species particularly significant in the context of bioremediation, synthetic chemistry, and wastewater treatment applications.⁵⁻⁹ In contrast to the complex, multilayered biofilms observed in other species like Pseudomonas aeruginosa, P. denitrificans exhibits a distinctive capability to form a nearly monolayer biofilm.¹⁰ Biofilms are defined by the existence of compact, multicellular communities that are enclosed inside a matrix consisting of extracellular polymeric components, including polysaccharides, proteins, and nucleic acids.¹¹ The process of biofilm development is commonly utilized by bacteria as a means of enhancing their survival capabilities, providing cells with increased physiological resilience against various stressors.¹² Notably, biofilms have been associated with many infections within the human body.^13,14 Hence, acquiring a mechanistic comprehension of their genesis across different species holds practical significance.

Biofilm formation in P. denitrificans is regulated by several factors. First, the process requires secretion of biofilm-associated protein A (BapA) through a type I secretion system, BapBCD.^10,15 This surface-associated protein of more than 200 kDa is comprised almost entirely of Asp-Ala peptide repeats thought to bind calcium, increase cell surface hydrophobicity,^10,16 and mediate surface attachment. Biofilm formation also appears to be regulated through a quorum sensing circuit homologous to the LuxI/R system, with Paracoccus enzymes PdeI and PdeR acting as acylhomoserine lactone (AHL) synthase and transcription factor/response regulator, respectively.¹⁷ Both pdeI deletion⁴ and pdeR overexpression¹⁸ were shown to promote biofilm formation. The former result suggests that AHL signaling inhibits biofilm formation, although the latter result suggests some additional complexity to this system. Finally, the secondary messenger cyclic diguanosine monophosphate (c-di-GMP) may influence biofilm formation as it does in numerous other bacterial species.¹⁹ C-di-GMP is produced through the enzymatic activity of diguanylate cyclase (DGC) enzymes. The genome of P. denitrificans encodes two annotated DGC homologues, DgcA and DgcB. Deletion of one or both dgc genes leads to a hyperbiofilm phenotype,¹⁵ suggesting that c-di-GMP is inhibitory to biofilm formation in this organism. This phenomenon is atypical, as heightened concentrations of c-di-GMP generally facilitate the process of biofilm development.¹⁹

A gene encoding a heme nitric oxide/oxygen binding protein (H-NOX) is adjacent to the dgcA gene. H-NOXs are hemoprotein sensors found in bacteria that play a crucial role in facilitating communal behaviors, such as the production of biofilms, in response to low levels of nitric oxide (NO).^20,21 The production of NO through the denitrification pathway has been observed to stimulate the formation of biofilms.²¹ Indeed, NO produced from the denitrification pathway appears to induce biofilm formation in P. denitrificans, and a hnox deletion strain showed decreased biofilm formation compared to WT.¹⁵ This suggests a model whereby H-NOX inhibits DgcA in a NO-dependent manner, resulting in decreased c-di-GMP production and an increase in biofilm formation in response to NO. However, H-NOX from P. denitrificans is one of a large group of H-NOX proteins that lacks the histidine conserved in other homologues to coordinate the heme iron. Thus, it is unclear how Pd H-NOX might respond to NO or other environmental signals. Further, the nature of H-NOX signaling in this species and how this influences biofilm formation remain unexplored.

The current study examines the function of Pd H-NOX in the context of biofilm formation in P. dentrificans. While recombinant Pd H-NOX is folded and stable, it does not show significant binding affinity toward either heme or zinc as expected from its primary sequence. Deletion of the hnox gene in P. dentrificans results in a severe biofilm-deficient phenotype, consistent with what has been observed previously.¹⁵ A comprehensive proteomics study identified alterations in a total of 104 proteins associated with metabolism, biofilm, denitrification, and ATP synthesis in the Δhnox strain. To the best of our knowledge, this study represents the most extensive analysis on proteomics in P. denitrificans conducted thus far, providing insight into the H-NOX regulatory pathway and its influence on energy production and biofilm formation.

Experimental Procedures

Experimental Design and Statistical Rationale

For microplate cell growth, crystal violet assays, pyruvate and acylhomoserine lactone quantitation, and qRT-PCR experiments, three biological replicates (n = 3) were used. For qRT-PCR and pyruvate quantitation, three technical replicates were also run for a total of nine measurements.

Proteome analyses were conducted on WT and Δhnox P. denitrificans PD1222 strains cultured in three (n = 3) biological replicates. These replicates were processed together, and the proteins were digested using trypsin, as described below. Peptides were labeled with the tandem mass tag (TMT) 10-plex isobaric label reagent set (Thermo) and then combined into a single multiplex sample group. The proteome samples were analyzed using liquid chromatography–tandem mass spectrometry (LC-MS/MS). All peptides and proteins that were uniquely identified, both quantified and nonquantified, are reported in Supplementary Table S1. Proteins with significantly differential abundance between strains were identified by applying a two-sided t test with a cutoff of log₂ fold change greater than 0.58 and a P value less than 0.05. These are reported in Supplementary Table S2. The sample preparation and data analysis sections provide a comprehensive discussion of the various sample preparation optimization experiments, growth conditions, MS parameters, and analysis parameters.  The data were analyzed using python, Origin, and R for various statistical tests, depending on the type of comparison being made. These tests included the two-tailed, unpaired Student’s t test for pairwise comparisons, the one-way ANOVA with Dunnett’s multiple comparisons test for samples grouped by one factor, and the two-way ANOVA with Tukey’s multiple comparisons test for samples grouped by two factors. The appropriate test was selected based on the specific analysis being performed. A significance level (P value ) of less than 0.05 was considered statistically significant.

Expression, Purification, and Characterization of Proteins

The entire hnox gene (Pden_3719) was amplified by PCR from P. denitrificans PD1222 genomic DNA. The PCR product was cloned into a pCDFDuet-1 vector (Novagen) using the NdeI restriction site using the Gibson cloning method.²² Plasmid was transformed into BL21 DE3 E. coli cells and cultured in LB medium with 50 μg/mL streptomycin at 37 °C and 220 rpm until reaching an optical density (OD) of 0.4 at 600 nm. To induce overexpression, IPTG was added to 0.4 mM, and the temperature was lowered to 18 °C while the cells were cultivated with shaking overnight. The cells were collected by centrifugation at 4000g for 30 min at 4 °C.

Cell pellets were resuspended in lysis buffer containing 50 mM HEPES pH 8, 300 mM NaCl, 5% glycerol and lysed using a homogenizer (Avestin, Inc.). Cell debris was removed by centrifugation at 25,000g for 30 min at 4 °C. Lysate was applied directly to a HiTrap butyl HP column (Cytiva) equilibrated with lysis buffer. Protein was eluted on a gradient of decreasing NaCl, followed by washing with deionized water. H-NOX eluted at the deionized water step and was dialyzed against 5 mM sodium phosphate pH 7.4, 0.5 mM DTT, and 20% glycerol at 4 °C overnight. Finally, the sample was concentrated and applied to a HiPrep Sephacryl S-100 HR column (Cytiva) equilibrated with dialysis buffer. At this point, the protein was highly pure, as judged by SDS-PAGE. Protein concentration was determined using an extinction coefficient at 280 nm of 21,684 M^–1 cm^–1, calculated as previously described.²³

Circular dichroism (CD) spectra were recorded at 20 °C using a Jasco-1500 spectropolarimeter with a temperature-regulated cuvette chamber. Protein was diluted to 5 μM in 5 mM K₂HPO₄ pH 7.4, in a 1 mm quartz cuvette. Spectra were acquired from 190 to 260 at 1 nm bandwidth, 2 s response time, 0.5 nm data pitch, and 10 nm/min scan speed. Each spectrum is the average of three accumulations and has been converted to mean residue ellipticity. Spectra were fitted using the Bestsel web server.²⁴ For thermal stability experiments, the ellipticity at 215 nm was monitored from 25 to 90 °C every 0.2 °C with a constant heating rate of 1.0 °C/min.

For metal content analysis, purified H-NOX protein was digested in 4 M HNO₃ overnight at 70 °C. Samples were further diluted in deionized water and analyzed on a 4300 DV inductively coupled plasma optical emission spectrometer (PerkinElmer Life Sciences), calibrated with a multielement standard. The wavelengths for measuring iron and zinc were 238.204 and 213.857 nm, respectively.

Heme Reconstitution

Purified protein was diluted to a final concentration of 11.2 μM in 5 mM sodium phosphate pH 7.4. A matching cuvette was loaded with an equivalent volume of buffer only, and both were titrated with 0.9 μL additions of hemin dissolved in 10 mM NaOH at a concentration of 2.61 mM. For experiments where reconstitution was followed by size exclusion chromatography (SEC), H-NOX was combined with hemin in phosphate buffer to a final concentration of 100 μM in both protein and hemin. This was incubated at room temperature for 1 h, followed by overnight incubation on ice. The sample was loaded onto a HiPrep Sephacryl S-100 HR column (Cytiva) equilibrated with 5 mM sodium phosphate pH 7.4, containing 10% glycerol.

P. denitrificans Growth, Survival, and Biofilm Assays

WT and Δhnox strains of P. denitrificans PD1222 were kindly provided by Dr. Stephen Spiro and described previously.¹⁵ Cultures were grown in LB media supplemented by 10 mM CaCl₂ or minimal media adapted from previous reports,²⁵ containing final concentrations of 38 mM MES pH 7.4, 19 mM KCl, 60 mM NH₄Cl, 126 μM EDTA, 236 μM CaCl₂, 764 μM MgSO₄, 50 mM succinate, 7.6 mM β-phosphoglycerate, 47 μM MnSO₄, 53 μM NaMoO₄, 1.6 μM CuCl₂, 20 μM Fe-citrate, and 50 μM ZnCl₂. NO was added using the NO-donor DETA NONOate at final concentrations of 2 and 5 mM, which correlate to steady-state NO concentrations in the low micromolar range (<8 and 20 μM) based on previous analysis. However, it should be noted that our experiment was performed at a lower temperature than previously used, which will decrease the decay rate of NONOate and, consequently, the NO concentration. Microplate assays for growth and biofilm were initiated from overnight cultures grown in LB at 30 °C and 200 rpm shaking. Each well contained 100 μL of media and was inoculated to an optical density at 600 nm (OD₆₀₀ = 0.05). For growth experiments in minimal media, starter culture cells were washed in minimal media twice before inoculation. Plates were shaken at 30 °C and imaged every hour at 600 nm in a BioTek Epoch2 microplate reader. After 24 h, cells from replicate wells were combined and diluted in LB media to create a dilution series for survival assays on LB agar plates. These were incubated for a further 24 h at 30 °C before imaging. For biofilm staining, microwell plates from growth experiments were washed three times with deionized water and allowed to air-dry. Once dry, 150 μL of 0.1% w/v crystal violet in water was added to each well, incubated 15 min, and discarded, and the plate was washed and dried as previously described. Finally, 200 mL of DMSO was added to each well and incubated 15 min prior to transferring 100 μL of each sample to a new plate for imaging. Absorbance values were collected at 570 nm and were corrected for dilution relative to the OD₆₀₀ values. All values are reported as recorded in microplates, where the approximate path length is 0.2 cm.

For determination of biofilm formation in the presence of exogenous pyruvate, WT and Δhnox were grown in 10 mL cultures of LB supplemented with 10 mM CaCl₂, 20 mM HEPES pH 7.2, and 0–10 mM sodium pyruvate. The cultures were grown until reaching an OD₆₀₀ of around 0.6, and surface-attached biomass was quantified by using crystal violet staining. The tubes were rinsed with deionized water and then stained with a 0.1% w/v crystal violet solution. The tubes were shaken for 15 min and rinsed with deionized water, and then 5 mL of DMSO was added to dissolve the stain. The ratio of A₅₇₀ (crystal violet-stained biomass) to OD₆₀₀ was used to provide an estimate of relative biofilm formation.

P. denitrificans Proteomics Sample Preparation

WT and ΔhnoxP. denitrificans PD1222 were inoculated from overnight LB starter cultures to OD₆₀₀ = 0.05 in 10 mL of LB supplemented with 10 mM CaCl₂ and grown with gentle shaking at 50 rpm and 30 °C for 24 h. The culture underwent centrifugation at 4000g, and the cell pellets were then washed and resuspended in RIPA buffer (Pierce) with protease and phosphatase inhibitors at a ratio of 1:100. The samples were subjected to sonication and centrifugation at 14,000g to get the supernatants. Protein extraction was performed using 10 mM tris(2-carboxyethyl)phosphine hydrochloride (TCEP), followed by alkylation with iodoacetamide for 60 min in the dark. Subsequently, a chloroform–methanol extraction procedure was conducted, followed by centrifugation at 14,000g. The resulting aqueous layer was subsequently separated and eliminated. Following the addition of an additional 400 μL of methanol, the sample underwent vortexing and centrifugation at 20,000g for 5 min. Next, the supernatant was extracted, and the sample was subjected to air-drying. The sample was resuspended in 100 mM triethylammonium bicarbonate and digested with trypsin (Promega). The desalting procedure was carried out using Sep-Pak C18 plates (Waters). The peptides were labeled using a tandem mass tag 10-plex isobaric label reagent set (Thermo). They were then separated into 46 fractions using an Acquity BEH C18 column from Waters in conjunction with UltiMate 3000 UHPLC equipment (Thermo). The 18 superfractions were subdivided, and peptides were isolated via elution using a gradient technique over a duration of 75 min. The peptides were ionized by using electrospray and analyzed by using an Orbitrap Eclipse Tribrid mass spectrometer (Thermo). The mass spectrometry data was acquired at a resolution of 120,000 throughout a spectral range spanning from 375 to 1500 m/z. The acquisition of MS/MS data was performed by utilizing the ion trap analyzer operating in centroid mode and normal mass range. Subsequently, CID activation was conducted with a normalized collision energy of 31.0. In this study, the method of synchronous precursor selection was employed to opt for a maximum of 10 MS/MS precursors for HCD activation. The HCD activation was carried out with a normalized collision energy of 55.0. Following this, the acquisition of MS3 reporter ion data was conducted using the Fourier transform mass spectrometry (FTMS) analyzer in profile mode, with a resolution of 50,000 across a mass-to-charge ratio (m/z) range of 100–500.

Data Analysis

P. denitrificans cells were grown to the exponential phase and then collected to study the impact of hnox deletion on the global proteome. The tandem mass tags (TMT) method was employed to label the extracted proteins, which were subsequently analyzed using liquid chromatography–tandem mass spectrometry (LC-MS/MS). The protein identification and quantification of MS3 reporter ions were performed using MaxQuant software (Version 2.2.0.0).²⁶ The current analysis employed the UniprotKB Paracoccus denitrificans (UP000000361) database. The tolerance for the parent ion was set to 3 ppm, the tolerance for the fragment ion was set to 0.5 Da, and the tolerance for the reporter ion was set to 0.003 Da. We set oxidation of methionine and acetylation of the N-terminus of the protein as variable modifications, while carbamidomethylation of cysteines was chosen as the fixed modification. Peptide validation and protein identification by MS/MS analysis were performed by employing the Scaffold Q+S software. The false discovery rate for peptide-spectrum match and protein level was set to 1%. Additionally, a minimum of two identified peptides was selected for protein identification. The Protein Prophet algorithm was utilized to assign protein probabilities, as detailed in Nesvizhskii’s work from 2003.²⁷ The intensity levels of the TMT MS3 reporter ions for proteins and their normalization were evaluated using ProteinNorm.²⁸ The data normalization was executed by the VSN method.²⁹ The statistical study utilized the Linear Models for Microarray Data (limma) technique and incorporated Empirical Bayes (eBayes) for standard error estimates.³⁰ Python was utilized for data visualization. A two-tailed t test was used to identify significant candidates with a P value less than 0.05 and a log2 fold change greater than 0.58.

RT-qPCR Validation

WT and Δhnox cells were grown in LB media supplemented with 10 mM CaCl₂ to an OD₆₀₀ of 0.7. 4 mL of 5% (v/v) phenol in ethanol was added to 10 mL cells, chilled on ice for 30 min, and centrifuged, and the cell pellets were stored at −80 °C. RNA extraction was done using a PureLink RNA Mini Kit (Invitrogen). DNA contamination was removed using an on-column DNase digestion protocol (Invitrogen). RNA concentration and purity were determined spectrophotometrically using a NanoDrop 2000 spectrophotometer (Thermo). cDNA was synthesized from 1 μg of pure RNA in 21 μL reaction volume using a SuperScriptIII first strand synthesis kit (Invitrogen). cDNA was diluted with nuclease-free water to a final concentration of 10 ng/μL and used for real-time PCR reactions. The primers were designed to amplify 100–150 base pairs (bp) of target genes with an average T_m ≈ 55 °C, and the final concentration used was 0.4 μM. Quantification of amplified PCR product using PowerUp SYBR Green PCR Master Mix (Applied Biosystems) was monitored by a CFX-Connect real-time system (Bio-Rad). The relative expression of genes was normalized to dnaN, a housekeeping gene encoding a sigma factor previously used for real-time PCR experiments in P. denitrificans.³¹

Quantitation of Pyruvate

P. denitrificans WT and Δhnox were grown in LB media supplemented with 10 mM CaCl₂ to an OD₆₀₀ of about 0.6. Cells were collected by centrifugation, washed with 2 mL of ice-cold PBS, resuspended in 1X pyruvate assay buffer (Cayman Chemicals), and disrupted by sonication on ice. Protein concentration was determined by a Bradford assay and used to normalize samples. Deproteinization and neutralization were carried out according to the manufacturer’s protocol. The samples were evaporated to dryness by speedvac and resuspended with 100 μL of deionized water. The fluorescence assay was carried out according to the manufacturer’s protocol using an Eclipse fluorescence spectrophotometer (Varian).

Quantitation of Hexadecanoyl Homoserine Lactone

P. denitrificans WT and Δhnox cells were grown in 15 mL LB cultures with shaking at 30 °C and 200 rpm until the optical density for each reached approximately 1.5, 4.0, or 6.0 at 600 nm. Cells were removed by centrifugation at 3500 rpm for 30 min at 4 °C, and 14 mL of media was extracted three times with equal volumes of ethyl acetate containing 0.1% formic acid. Solvent was dried by rotary evaporation, and samples were resuspended with 500 μL of acetonitrile containing 0.1% formic acid, centrifuged at 20,000g for 20 min, and filtered through 0.22 μm centrifugal filter units (Millipore). Standard solutions of N-hexadecanoyl-l-homoserine lactone (Cayman Chemical) were dissolved in the same solvent and used to determine retention times, parent and fragment ion masses, and to generate a calibration curve for quantitation. 5 μL of standards and samples were injected onto an Acclaim 120 C18 2.1 × 100 mm reverse phase column (3 mm particle size, ThermoFisher) at 80% buffer B (acetonitrile/0.1% formic acid) and 20% buffer A (water/0.1% formic acid) and 30 °C at a flow rate of 0.6 mL/min. This concentration was maintained for 1 min, brought to 100% B at 1.5 min, maintained at 100% B until 5.5 min, and returned to the starting conditions. LC effluent was delivered to a Shimadzu LC-MS 8050 triple-quadrupole mass spectrometer operating in positive ion mode. For multiple reaction monitoring (MRM), precursor ion m/z was set to 340.30, and two product ions were monitored at 102.20 and 238.00.

Results

Recombinant P. denitrificans H-NOX Contains No Heme or Zinc

Untagged Pd H-NOX was expressed in E. coli and purified to homogeneity using a combination of hydrophobic interaction and size exclusion chromatography. The resultant protein was completely colorless and exhibited no visible absorption features (Figure 1A). Both iron and zinc were undetectable by ICP-OES. For some preparations, the heme biosynthetic precursor 5-aminolevulinic acid (δ-ALA) was included to ensure heme availability, but this had no impact on the color or spectrum of the purified protein. The CD spectrum was similar to that obtained for Vc H-NOX³² and indicated that the protein was folded (Figure 1B). Thermal denaturation appears to occur with a Tm of approximately 50 °C (Figure 1C). Altogether, the data indicate that recombinant Pd H-NOX is folded and stable as isolated and contains neither heme nor zinc to any appreciable extent.

Figure 1.

Characterization of purified recombinant Pd H-NOX. (A) Absorbance spectrum. (B) CD spectrum. Data are indicated by circles and a secondary structure fit generated by Bestsel^33,24 is indicated by a solid line. Fit parameters are α helix = 22.0%, antiparallel sheet = 28.8%, turns = 12.1%, and others = 37.2%. (C) Thermal denaturation followed the CD signal at 208.5 nm.

It has been observed that some heme proteins recombinantly expressed in E. coli are purified without heme but can be reconstituted with heme in vitro as is the case for the dissimilative nitrate respiration regulator (DNR) from Pseudomonas aeruginosa.³⁴ Additions of hemin to buffer (Figure 2A, red traces) or to Pd H-NOX (black traces) revealed subtle differences in absorption, particularly at low hemin concentrations (Figure 2B), which may be indicative of heme binding. However, after excess hemin was removed by size exclusion chromatography, only weak heme signals were observed around 350 and 400 nm, suggestive of a heterogeneous heme environment likely dominated by nonspecific surface binding and/or hydrophobic interactions.³⁵

Figure 2.

Heme reconstitution of Pd and H-NOX. (A) Hemin was titrated into buffer (red traces) or Pd H-NOX (black traces) at 0.2 molar equivalents per addition. (B) Comparison of 0.2 molar equivalents of hemin in buffer (red) or Pd H-NOX (black). (C) Pd H-NOX after incubation with 1 molar equivalent of hemin and size exclusion chromatography.

H-NOX Is Important for Biofilm Formation in P. denitrificans

We set out to assess the role of hnox in P. denitrificans growth, biofilm formation, and survival by using a microplate crystal violet assay and dilution plating. WT and Δhnox strains were grown in either rich media (LB) or minimal media in the presence or absence of the NO donor DETA NONOate at 2 mM. Growth as assessed by OD₆₀₀ was not significantly different among any strain or condition after 24 h of shaking (Figure 3A,B), which is consistent with viability assays indicating similar survival (Figure 3D). However, biofilm formation was dramatically attenuated in the Δhnox strain under all of the conditions tested (Figure 3C,E). No significant differences were observed for biofilm formation as a consequence of NO treatment, suggesting that the H-NOX pathway promoting biofilm formation was constitutively active under these conditions. This is consistent with a previous report indicating that 5 mM added sodium nitrate had very little impact on WT biofilm formation compared with the ∼0.5 mM nitrate found in rich media preparations.¹⁵ Our work does differ in that P. denitrificans biofilms were detected in a 96-well microplate format whereas the previous work required the increased surface area of larger Petri dishes. Another report quantified P. denitrificans biofilms in 24-well plates.¹⁰ Clearly, biofilm growth is highly sensitive to culture conditions, including culture vessel surface chemistry, oxygen availability, stirring speed, etc. As the proteomics experiments described below would require a greater biomass than is available from microplate assays, we next set out to determine whether larger cultures grown in 50 mL centrifuge tubes would recapitulate the observed phenotypes. Indeed, WT cells grown under these conditions exhibited clear CV staining while very little staining was observed for the Δhnox mutant (Figure 3E).

Figure 3.

Growth, survival, and biofilm formation in P. denitrificans. WT (blue) and Δhnox (orange) strains were grown in LB (A) or minimal media (B) in the absence (squares) or presence (circles) of 2 mM DETA NONOate. After 24 h, biofilm formation was determined by CV staining (C). Results are from three replicate experiments with error bars reporting standard deviation. After 24 h of growth, a dilution series of cultures grown in LB (D, left) or minimal media (D, right) were plated onto LB agar. LB cultures of 10 mL were also grown in 50 mL tubes for 24 h and stained with CV (E). Bars represent mean (±SEM) of relative abundance of CV staining signal from at least three (n = 3) independent experiments/cultures. **P < 0.01 and ***P < 0.001 compared to control by two-way ANOVA with Tukey’s multiple comparisons test (C) or Student’s t test for pairwise comparisons (E).

A TMT-Based Proteome Study of ΔhnoxP. denitrificans

To provide an in-depth investigation of H-NOX signaling in P. denitrificans, a tandem mass tag (TMT) proteomics approach was utilized to assess the proportional presence of proteins in both WT and Δhnox samples. The Δhnox and WT samples were analyzed in triplicate (n = 3), with all measurements completed within a measurement duration of less than 12 h. Through the adoption of standard methodologies,²⁸ a comprehensive analysis led to the identification of a total of 2,694 proteins along with 21,122 unique peptide sequences, with an average rate of 22.4 protein quantifications per minute for every individual sample (Table S1). To evaluate the distribution of protein intensity across all samples, a robust analysis was performed, which highlighted a persistent pattern, as clearly depicted in a violin plot (Figure 4A). Moreover, an average identification of 2,690 proteins in each sample indicates a significant degree of consistency and reliability with comparable experimental configurations (Figure S1). Notably, we discovered a total of 462 proteins that were uniquely identified in the current study (Figure 4B), highlighting the improved proteome coverage of this data in comparison to previous studies.³⁶⁻³⁸ The impact of the hnox deletion on a broader spectrum was assessed by principal component analysis (PCA) on both WT and ΔhnoxP. denitrificans. The first two components (PC1 and PC2) of the PCA yielded a cumulative variance of 78.2% for the proteome data set. Though the PCA plot exhibited a consistent pattern throughout several biological repetitions as shown in Figure 4C, a clear differentiation between WT and Δhnox samples demonstrated a notable impact of the H-NOX mechanism on the global proteome. To understand the biological association between specific genes in Δhnox and WT P. denitrificans, hierarchical clustering has been implemented that reveals an apparent difference between the Δhnox and WT samples (Figure 4D). Moreover, the Pearson correlation coefficients display a robust correlation among biological replicates, with values exceeding 0.9 as illustrated in the scatter plots (Figure S2).

Figure 4.

Proteome analysis of P. denitrificans. (A) Protein intensity distributions in Δhnox and WT P. denitrificans by violin plots. (B) Principal component analysis compares global protein expression patterns between WT and Δhnox P. denitrificans samples. (C) Venn diagram of protein identification counts from several investigations, including the present study. (D) Hierarchical clustering of protein abundance in WT and Δhnox strains. An elevated expression is shown in orange, while decreased expression is shown in turquoise.

To gain a deeper understanding of the H-NOX mechanism, a systematic evaluation was performed on proteins exhibiting significant variations in expression levels between the Δhnox and WT samples. Specifically, proteins with a fold change of at least 1.5 and a p value ≤0.05 were selected for further analysis. As shown in Figure 5A, a comprehensive analysis revealed a total of 104 significantly regulated proteins, with 55 being down-regulated and 49 being upregulated in the Δhnox strain relative to WT (Table S2).

Figure 5.

Functional study of differentially expressed proteins. (A) The volcano plot illustrates the differential expression of proteins in Δhnox samples compared to WT. Yellow indicates upregulated proteins, whereas green represents downregulated candidates. (B) Real-time qPCR assays to evaluate the expression levels of specific target genes in WT (dark blue) and Δhnox (gray) strains of P. denitrificans. The relative expression level values were determined by combining the technical replicates (n = 3) and biological replicates (n = 4) for each candidate. Error bars indicate the mean ± SEM (****p ≤ 0.0001), and “ns” stands for not significant as determined by a Student’s t test. (C) The table displays a comparative analysis of the expression of selected significant candidates at both the proteome and transcriptome levels. The final column includes the primer sequences that were employed. (D, E) GO analysis of Δhnox in P. denitrificans. The bar chart illustrates the biological process of 11 down-regulated proteins that engaged in cellular nitrogen compound biosynthesis process, and the cellular compartments indicate the association of 8 proteins from periplasmic space.

Next, we validated the expression levels of 4 selected proteins using qRT-PCR (Figure 5B,C). Relative expression was normalized to dnaN (pden0970),³¹ a housekeeping gene whose expression was constitutive at the protein level (A1B5X4, see Table S1). Pden_4985, Pden_4983, and Pden_1517 are annotated as the α/β hydrolase fold enzyme (A/B), pyruvate dehydrogenase E1 subunit (PDH), and XRE family transcriptional regulator (TR), respectively. All are downregulated comparably at both the protein and transcript levels (Figure 5C). Since BapA is not annotated in the Uniprot database,¹⁵ the initial proteome analysis did not identify any peptides from this protein. Searching the data using the predicted sequence revealed the presence of 6 unique BapA peptides, but there was no significant difference in expression level between WT and Δhnox samples (Table S3). This was confirmed by qRT-PCR, which showed no significant differences in the transcript levels between strains.

To further our understanding of H-NOX’s role in P. denitrificans, proteins that exhibited differential expression (fold change ≤1.5, p value ≤0.5) were chosen for further examination by gene ontology (GO) analysis. Surprisingly, we could not detect any significant enrichment of GO terms with the upregulated candidates. In contrast, the downregulated proteins showed a notable enrichment of terms related to biological processes involved in the cellular nitrogen compound biosynthetic process, ATP metabolic process, organic cyclic compound biosynthetic process, and heterocycle biosynthetic process. Additionally, an analysis of the cellular compartments suggested the segregation of eight candidates within the periplasmic space.

hnox Deletion Associates with Pyruvate Depletion in P. denitrificans

Proteins expressed from an apparent operon (Pden_4980–Pden_4985) are all strongly downregulated in the Δhnox strain (Figure 5A, Table S2). Among these, several are annotated as E1 or E2 subunits of oxoacid dehydrogenase complexes, which include pyruvate dehydrogenase. This is noteworthy as pyruvate metabolism has been shown to influence biofilm formation and dispersal in other species.^39,40 Thus, we set out to evaluate the status of intracellular pyruvate in WT and Δhnox strains using a commercially available fluorescence assay. The results indicate a statistically significant decrease of approximately 20% in pyruvate levels in the Δhnox strain relative to WT (Figure 6A) consistent with the hypothesis that H-NOX signaling influences pyruvate metabolism in P. denitrificans.

Figure 6.

Impact of hnox deletion on pyruvate and quorum sensing. (A) Pyruvate quantitation in Δhnox relative to WT P. denitrificans. (B) Crystal violet staining of WT and ΔhnoxP. denitrificans cultures grown in LB media with or without pyruvate supplementation. Relative biofilm formation is expressed as a ratio of A₅₇₀/OD₆₀₀. The relative expression levels were calculated by pooling the technical replicates (n = 3) and biological replicates (n = 3) for each candidate. Error bars indicate mean ± SEM (*p ≤ 0.05, **p ≤ 0.01). (C) Hexadecanoyl homoserine lactone quantification in Δhnox and WT P. denitrificans. Concentration is reported in nM in culture media, as extrapolated from LC-MS results and assuming complete extraction. Optical density ranges of cultures used in this experiment are reported for each set of samples. Error bars indicate mean ± SEM (B) or mean ± SD (C) for 3 biological replicates. *p ≤ 0.05, **p ≤ 0.01, and ***p ≤ 0.001 compared to control by Student’s t test for pairwise comparisons (A) or by two-way ANOVA with Tukey’s multiple comparisons test (B, C).

Pyruvate Supplementation Increases Biofilm Formation in P. denitrificans

As a reduction in pyruvate levels seems to correlate with a decreased level of biofilm formation, we sought to determine whether pyruvate supplementation of the media might reverse this phenotype. WT and ΔhnoxP. denitrificans were grown in LB media as previously or supplemented with 5–10 mM pyruvate. While the total growth rate in liquid culture was not influenced by pyruvate supplementation in either strain (Figure S3), biofilm formation increased significantly in both WT and ΔhnoxP. denitrificans at 10 mM pyruvate (Figure 6B). Pyruvate supplementation resulted in a nearly 2-fold increase of biofilm in ΔhnoxP. denitrificans, yet it is still well below WT levels. Thus, pyruvate levels appear to correlate with biofilm in P. denitrificans but other mechanisms must also be operating to suppress biofilm formation in the Δhnox strain.

Hexadecanoyl Homoserine Lactone Synthesis Is Increased in ΔhnoxP. denitrificans

Hexadecanoyl homoserine lactone (C₁₆-HSL) is a quorum sensing signal synthesized by P. denitrificans that acts as a negative regulator of biofilm formation.^1,2,8 We quantified C₁₆-HSL in WT and ΔhnoxP. denitrificans at several cell densities using a mass spectrometry-based protocol derived from a previous study⁴ (Figure 6C). Consistent with a function as an autoinducer, C₁₆-HSL concentrations increase with increasing cell density in both strains. However, at high cell densities, we observe a significant increase in C₁₆-HSL production in the mutant strain relative to the WT.

Discussion

Recent studies have shed light on the H-NOX-mediated mechanisms in bacterial communal behaviors including motility, biofilm formation, host colonization, and quorum sensing.¹⁰⁻¹³ However, all of these revolve around NO-sensing by the H-NOX heme, whereas there is a large family of H-NOX proteins that lack the proximal heme-binding His residue. To the best of our knowledge, only one other study has investigated the role of an H-NOX member of this family. That work showed that hnox deletion in P. denitrificans resulted in a biofilm-deficient phenotype and that NO, endogenously derived through denitrification of nitrate found in LB media, likely stimulated biofilm formation.¹⁵ We further characterized this mutant, confirming that the Δhnox strain did not exhibit any significant deficiency in growth or survival but did exhibit a dramatic biofilm-deficient phenotype. To provide a comprehensive understanding of the influence of the hnox deletion in P. denitrificans on its biofilm generation and to identify its potential targets, we performed a comparative analysis of proteome data derived from WT and ΔhnoxP. denitrificans. Our proteome data set offers a comprehensive compilation of P. denitrificans proteins and identifies altered abundance of those associated with denitrification, ABC transporters, sulfur oxidation, TCA cycle, and ATP synthase, as discussed below.

NO Signaling and Denitrification

Denitrification is the process by which nitrate is reduced to dinitrogen and requires the enzymes nitrate reductase (Nar), nitrite reductase (Nir), nitric oxide reductase (Nor), and nitrous oxide reductase (Nos). Denitrifying bacteria such as P. denitrificans are subject to the endogenous formation of nitric oxide (NO) through the nitrite (NO₂^–) reduction activity of Nir. This was proposed to activate H-NOX signaling resulting in biofilm formation based on the observation that deletion of nir resulted in a biofilm deficient phenotype comparable to Δhnox.¹⁵ However, we see significant biofilm formation for cultures grown in minimal media that are free of nitrate and no response to exogenous NO for any strain tested. Further, recombinantly expressed Pd H-NOX exhibits proper folding and stability yet is purified without heme, although we cannot rule out the possibility that E. coli lacks the requisite machinery for cofactor insertion. Nonetheless, our data suggest that Pd H-NOX functions independently of NO signaling.

However, our proteomics data do establish a link between H-NOX and denitrification. Specifically, maturation factors NorQ and NirE were significantly although modestly downregulated in the Δhnox strain (Table S2). Mutational studies conducted on the nor gene cluster have unveiled the critical role played by the norQ gene in the formation of a fully operational Nor through insertion of the nonheme iron cofactor Fe_b.⁴¹ The NirE enzyme is associated with heme d1 biosynthesis, and mutant forms of this enzyme lead to an inactive cd1-type nitrite reductase that lacks heme d1.^42,43 Thus, downregulation of NorQ and NirE protein expression in the Δhnox proteome indicates that normal H-NOX signaling may promote denitrification either directly or indirectly.

ATP Binding Cassette (ABC) Transporters

ABC transporters are conserved throughout all domains of life and function in the transport of a vast array of solutes across the cell membrane. These are minimally composed of two transmembrane permease domains/proteins and two cytoplasmic ATPase domains/proteins. Bacterial importers of this type also involve an extracellular solute binding protein (SBP) responsible for binding the relevant solute with high affinity and specificity and delivering it to the membrane permease for import. ABC transporters of carbohydrates are classified into two families, namely carbohydrate uptake transporter 1 (CUT1) and carbohydrate uptake transporter 2 (CUT2), depending on their domain organization and substrate specificity.⁴⁴ CUT1 family members typically transport di- or oligosaccharides, while CUT2 transporters are specific for monosaccharides. In ΔhnoxP. denitrificans, we observe significant downregulation for an entire operon of the CUT1 family (Pden_4439, Pden_4440, and Pden_4442) as well as several monosaccharide SBPs of the CUT2 family (Pden_4771, Pden_4845, and Pden_4851) and an amino acid SBP of the HAAT family (Pden_4797). This indicates a disruption of the organism’s ability to acquire important carbohydrate and amino acid nutrients, which may contribute to the biofilm deficient phenotype observed for the Δhnox strain.

Sulfur Oxidation

The bacterial Sox (sulfur oxidation) pathway, which operates within the periplasmic compartment, serves as a means of oxidizing thiosulfate in a variety of sulfur-oxidizing bacteria. This oxidation process yields electrons that are harnessed for the purpose of generating respiratory energy or facilitating carbon fixation.⁴⁵ The SoxYZ, SoxXA, SoxB, and SoxCD proteins of the Sox enzyme complex in Paracoccus pantotrophus participate to oxidize thiosulfate, hydrogen sulfide, sulfur, and sulfite, converting them into sulfate. It is worth noting that the overexpression of WT SoxS leads to heightened antibiotic resistance in E. coli. In addition, SoxS also regulates the production of biofilm in Salmonella typhimurium.⁴⁶ Furthermore, the metaproteogenomic analysis of microbial communities associated with chimneys revealed that the Sox pathway was among the most prominently expressed pathways for energy production.⁴⁷ We have identified seven proteins within the Sox enzyme complex that exhibit a strong protein–protein interconnection (Figure S4 A). Our proteome data indicate that several hits, namely SoxZ, SoxY, and SoxAX, exhibit a significant decrease in the Δhnox strain, while the other candidates exhibit a downward trend. Here again, we observe downregulation of a pathway that may be important for bacterial biofilms to generate metabolic energy.

Pyruvate Metabolism

Metabolic remodeling is thought to occur during the process of microbial biofilm formation, as exemplified by a metabolomics time course study showing alteration of a wide array of metabolites during biofilm formation in Candida albicans. These were associated with various processes, including the tricarboxylic acid (TCA) cycle, lipid biosynthesis, amino acid metabolism, glycolysis, and oxidative stress.⁴⁸ Pyruvate, the product of glycolysis, represents a hub of central carbon metabolism. It can be reduced during fermentation to lactate, oxidatively decarboxylated to acetyl-CoA for entry into TCA by the pyruvate dehydrogenase (PDH) complex, or combined with glyceraldehyde-3-phosphate to generate 1-deoxy-d-xylulose 5-phosphate, a key intermediate in the pentose phosphate pathway as well as isoprenoid biosynthesis. Pyruvate depletion leads to a reduction in the amount of P. aeruginosa biofilm,⁴⁹ suggesting that dispersion may be triggered by a disruption in the regulation of redox balance and cellular homeostasis. The reduced levels of pyruvate in ΔhnoxP. denitrificans may indicate H-NOX’s association with pyruvate in maintaining the redox balance necessary for coping with stressful conditions during biofilm production. Notably, the addition of pyruvate to the media partially mitigates the biofilm deficient phenotype in ΔhnoxP. denitrificans, suggesting that altered pyruvate metabolism occurs alongside other factors that promote biofilm dispersion.

Further, we see a downregulation in proteins annotated as E1 (Pden_4984 and Pden_4985) and E2 (Pden_4983) components of PDH. This is counterintuitive, given that downregulation of these proteins would presumably lead to an increase of the substrate pyruvate in the Δhnox strain. However, it is unknown whether Pden_4983–4985 encode a functional PDH. Several operons in P. denitrificans encode putative PDH genes, and other oxoacid dehydrogenases, such as α-ketoglutarate dehydrogenase of the TCA cycle, are closely related to PDH. To our knowledge, a PDH function for any such operon in P. denitrificans has not been experimentally validated. Further, it is possible that pyruvate levels may regulate protein expression or activity. Thus, we cannot say that PDH expression is downregulated as a direct result of the hnox deletion; we can only say that oxoacid metabolism is likely perturbed and that intracellular pyruvate levels are decreased. Nevertheless, our results, combined with those of previous studies, make it tempting to speculate that targeting pyruvate metabolism might serve as a potential approach to impede biofilm formation in multiple bacterial species.

ATP Synthase

The f₀f₁ ATP synthase is a multisubunit molecular machine that couples discharge of the electrochemical proton gradient generated through the electron transport chain to ATP synthesis. In our proteomics analysis, eight proteins of the ATP synthase complex have been identified with significant protein–protein interactions in P. denitrificans (Figure S4B). Three of these (AtpA, AtpG, and AtpB) experience a significant decrease, whereas the other proteins show a downward trend in the Δhnox relative to that of WT P. denitrificans. It was recently shown that ΔatpAStaphylococcus aureus exhibited altered biofilm morphology and decreased biofilm burden in a mouse infection model.⁵⁰ Thus, like pyruvate metabolism, ATP synthesis seems to be a conserved requirement for the optimal biofilm development across bacterial species.

Mechanisms of H-NOX Signaling

Collectively, the phenotypic and proteomic findings above indicate that H-NOX has a significant impact on energy metabolism and biofilm formation in P. denitrificans that appears to be independent of NO. In addition, H-NOX signaling appears to be manifested through both transcriptional and post-translational pathways. As the hnox gene is adjacent to one of two annotated DGC genes (dgcA), the most obvious mechanism of H-NOX signaling is through binding and alteration of DgcA activity, altering levels of intracellular c-di-GMP. This secondary messenger can bind allosterically to enzymes or to transcription factors, accounting for both altered enzyme activity and gene expression, respectively. Further, c-di-GMP and quorum-sensing circuits are tightly integrated and can regulate one another in various bacterial species.¹⁹ Unfortunately, we were unable to confirm this association in P. dentrificans because c-di-GMP levels were too low to reliably quantify in our hands. Nevertheless, there are some mechanistic insights that can be drawn from the data presented above.

Both the putative E1 and E2 oxoacid dehydrogenase subunits are regulated at the transcriptional level (Figure 5B). ABC transporters, SOX proteins, and ATP synthase proteins are also likely to be regulated at the transcriptional level. It is interesting to note that a putative transcription factor (Pden_1517) is also strongly downregulated in ΔhnoxP. denitrificans at both the protein and mRNA levels. The genome of P. denitrificans also includes several homologues of Crp family transcriptional regulators, which act as cyclic nucleotide monophosphate (e.g., cAMP) receptors. Substitution of Glu for a conserved Ser residue of Crp proteins provides specificity to c-di-GMP, as is observed for the c-di-GMP-responsive Clp transcription factors from several species.⁵¹ One of the P. denitrificans Crp homologues (Pden_1850) has an Asp residue at this position, making it an intriguing candidate as a c-di-GMP receptor. Relatively high sequence coverage (40%) of this protein in our proteomics data set (Table S1) suggests that it is expressed at a reasonably high level. We are currently in the process of optimizing c-di-GMP binding protein profiling experiments to validate this and other possible c-di-GMP effectors.

We also observe a significant increase in the level of C₁₆-HSL synthesis in the mutant strain at high cell densities, indicating that quorum sensing is affected. There is no evidence of an altered abundance of either PdeI or PdeR from the proteomics data, but these experiments were performed at low cell density before changes in C₁₆-HSL become apparent. The PdeI/R regulon was suggested to comprise nearly 1,000 genes by RNA-seq of ΔpdeI and ΔpdeR strains,¹⁷ whereas we see a much smaller set of differentially expressed proteins in Δhnox. This suggests that temporally resolved proteomics profiling will reveal a much larger set of proteins at high cell densities whose expression is influenced indirectly by H-NOX via its intersection with quorum sensing.

In conclusion, the phenotypic and proteomic data above provide a holistic view of changes accompanying hnox deletion and their impact on biofilm formation. Although our data illustrates perturbation in central carbon and energy metabolism, metabolomics studies will allow a precise description of how metabolic flux through various pathways is altered to promote or inhibit biofilm formation in this organism. These studies are underway in our laboratory. We also observe a tight interrelationship between H-NOX and quorum-sensing signaling pathways. Careful promoter analysis and characterization of direct H-NOX interaction partners should help us to tease apart transcriptional and post-transcriptional mechanisms of H-NOX signaling.

Data Availability Statement

Mass spectrometry data have been deposited to the ProteomeXchange Consortium via the MassIVE partner repository with the data set identifier MSV000094619 and may be accessed from https://massive.ucsd.edu/ProteoSAFe/user/summary.jsp?user=shariful007. The annotated MS/MS spectra can be accessed using the search key “qj7waplyqe” in https://msviewer.ucsf.edu/cgi-bin/mssearch.cgi?report_title=MS-Viewer&search_key=qj7waplyqe&search_name=msviewer. All other data are contained within the article or the Supporting Information.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.jproteome.4c00466.

• Table S1: List of identified proteins in P. denitrificans (XLSX)

• Table S2: List of differentially expressed proteins in Δhnox and WT P. denitrificans (XLSX)

• Table S3: List of BapA peptides in Δhnox and WT P. denitrificans (XLSX)

• Figure S1: Number of identified proteins in Δhnox and WT P. denitrificans. Figure S2: The multiplot scattering plot shows the Pearson’s Correlation across all the experiments. Figure S3: Growth curve under pyruvate-treated conditions. Figure S4: Protein–protein interaction networks. (PDF)

Author Contributions

M.S.I. prepared proteomics samples, analyzed and interpreted proteomics data, and was a major contributor in writing the manuscript. A.A. performed pyruvate and AHL quantitation experiments. C.C.L.-L. developed AHL quantitation methodology. F.S. expressed, purified, and characterized H-NOX protein. E.T.Y. conceptualized the study, performed growth and biofilm assays, and was a major contributor in writing the manuscript. All authors read and approved the final manuscript.

This study was funded by the National Science Foundation grant number 2103676. The funder played no role in study design, data collection, analysis and interpretation of data, or the writing of this manuscript. Proteomics data were collected at the IDeA National Resource for Quantitative Proteomics, which is funded through a grant from the National Institute of General Medical Sciences R24GM137786.

The authors declare no competing financial interest.