A membrane protein of the rice pathogen Burkholderia glumae required for oxalic acid secretion and quorum sensing

Abstract Bacterial panicle blight is caused by Burkholderia glumae and results in damage to rice crops worldwide. Virulence of B. glumae requires quorum sensing (QS)‐dependent synthesis and export of toxoflavin, responsible for much of the damage to rice. The DedA family is a conserved membrane protein family found in all bacterial species. B. glumae possesses a member of the DedA family, named DbcA, which we previously showed is required for toxoflavin secretion and virulence in a rice model of infection. B. glumae secretes oxalic acid as a “common good” in a QS‐dependent manner to combat toxic alkalinization of the growth medium during the stationary phase. Here, we show that B. glumae ΔdbcA fails to secrete oxalic acid, leading to alkaline toxicity and sensitivity to divalent cations, suggesting a role for DbcA in oxalic acid secretion. B. glumae ΔdbcA accumulated less acyl‐homoserine lactone (AHL) QS signalling molecules as the bacteria entered the stationary phase, probably due to nonenzymatic inactivation of AHL at alkaline pH. Transcription of toxoflavin and oxalic acid operons was down‐regulated in ΔdbcA. Alteration of the proton motive force with sodium bicarbonate also reduced oxalic acid secretion and expression of QS‐dependent genes. Overall, the data show that DbcA is required for oxalic acid secretion in a proton motive force‐dependent manner, which is critical for QS of B. glumae. Moreover, this study supports the idea that sodium bicarbonate may serve as a chemical for treatment of bacterial panicle blight.

needs to increase by 26% by 2035 to meet the rice demand for Earth's growing population (White et al., 2020). While important research has been conducted to understand the virulence of B. glumae (Chen et al., 2012;Kim et al., 2004;Lelis et al., 2019), no chemical treatment has been adopted to control BPB in rice. Some rice varieties have shown reduced sensitivity to BPB; however, no rice variety has shown complete resistance to BPB (Ham et al., 2011). Therefore, there is an urgent need to find a solution to control BPB in rice to protect global rice production.
The occurrence of BPB due to B. glumae infection is a multifactorial process that includes several virulence factors, including toxoflavin (Ham et al., 2011). Toxoflavin, a 7-azapteridine antibiotic, is the major virulence factor of B. glumae and mainly responsible for the symptoms of BPB in the rice plant. Toxoflavin acts as an electron carrier between oxygen and NADH and is able to produce hydrogen peroxide, increasing the levels of reactive oxygen species, leading to toxicity to the plant (Latuasan & Berends, 1961;Park et al., 2019).
B. glumae synthesizes and transports toxoflavin by activating the transcription of the tox operons in a process controlled by quorum sensing (QS) (Chen et al., 2012;Kim et al., 2004;Suzuki et al., 2004).
The TofR-C8-HSL complex activates the expression of the ToxJ regulatory protein, which in turn activates the expression of ToxR, a LysR-type transcriptional regulator. ToxR binds to the promoters of the toxABCDE and toxFGHI operons and activates the transcription of toxoflavin biosynthesis and transporter genes. The TofR-C8-HSL complex also regulates protease activity, flagellum biogenesis, and flagellar motility in B. glumae (Ham et al., 2011). Interference with QS is a promising approach to treat or prevent plant diseases caused by bacteria (Helman & Chernin, 2015).
Oxalic acid is a well-known metabolite produced by bacteria, fungi, plants, and animals (Nakata, 2011). The functional role of oxalic acid is species-specific. In bacteria and fungi, oxalic acid plays several important roles, contributing to metal tolerance, nutrient acquisition, and virulence (Gadd, 1999;Hamel et al., 1999;Munir et al., 2001). Production of oxalic acid by B. glumae is regulated by QS and required to avoid alkaline toxicity (Goo et al., 2012).
Production of ammonia in nutrient-rich medium as a by-product of the metabolism of amino acids causes alkalinization of the culture medium and toxicity to B. glumae (Goo et al., 2012;Nam et al., 2021).
Oxalic acid acidifies the culture medium and reverses the alkaline pH toxicity. Acidification of culture medium during bacterial growth is also important to protect acyl-homoserine lactones (AHLs) from nonenzymatic inactivation, which occurs rapidly at alkaline pH, and thus for QS and virulence (Byers et al., 2002;Le Guillouzer et al., 2020;Yates et al., 2002). To date, no oxalic acid efflux transporter has been identified in B. glumae.
These pH changes are probably important during infection as well. Plant pathogens and symbionts replicate in a space outside the plasma membrane of plant cells termed the apoplast (Denny, 1995;Kang et al., 2008), where they must interact with aspects of plant immunity. In the early stages of bacterial infection, plants respond by secreting a number of metabolites resulting in alkalinization of the apoplastic space (Geilfus, 2017;Nachin & Barras, 2000;O'Leary et al., 2016), as well as increased levels of divalent cations including Ca 2+ and Mg 2+ (Fones & Preston, 2013;O'Leary et al., 2016).
However, little has been reported on the immune responses of rice to infection by B. glumae, and the functions of these apoplastic changes in plant defence are poorly understood.
The DedA membrane protein superfamily is found in nearly all living species. DedA proteins may function as membrane transporters Hama et al., 2022). Our laboratory has characterized DedA proteins in several bacterial species, including Escherichia coli, Burkholderia thailandensis, and B. glumae (Iqbal et al., 2021;Panta et al., 2019). Simultaneous deletion of the E. coli DedA family genes yqjA and yghB (encoding proteins with c.60% amino acid identity) results in altered proton motive force (PMF), cell division defects, induction of envelope stress responses, and sensitivity to elevated temperature, alkaline pH, antibiotics, and biocides (Kumar & Doerrler, 2015;Sikdar et al., 2013;Sikdar & Doerrler, 2010;Thompkins et al., 2008). Our laboratory and others have found that DedA family proteins are required for polymyxin and/or cationic antimicrobial peptide (CAMP) resistance of Salmonella enterica (Shi et al., 2004), Neisseria meningitidis (Tzeng et al., 2005), E.

B. glumae
DbcA displays approximately 73% amino acid identity with B. thailandensis DbcA. We showed that deletion of dbcA causes sensitivity to colistin, decreased toxoflavin production, and loss of virulence (Iqbal et al., 2021). We could replicate these effects on toxin production and loss of virulence with sodium bicarbonate, which dissipates the ΔpH component of the PMF (Farha et al., 2020), and proposed this as a chemical intervention for BPB. In the present study, we investigate whether DbcA is required to maintain proper QS in B. glumae. We report that B. glumae ΔdbcA does not acidify the growth medium due to impaired oxalic acid production. As a result, the culture medium pH of ΔdbcA becomes alkaline during the stationary phase and this mutant fails to accumulate AHL and carry out QS signalling. Exposure of B. glumae wild type to sodium bicarbonate causes similar effects. These data collectively show that DbcA is required for QS of B. glumae.
2 | RE SULTS 2.1 | B. glumae ΔdbcA and ΔobcAB are unable to acidify the growth medium during the stationary phase The obcAB operon is responsible for oxalic acid biosynthesis in B.
glumae and is needed for acidification of culture medium during the stationary phase (Nakata & He, 2010). We observed that B. glumae 336gr-1 ΔdbcA displayed a partial growth defect ( Figure 1a) and did not acidify the culture medium at the stationary phase ( Figure 1b).
We hypothesized that this may be due to a defect in oxalic acid production and secretion. To test this, a B. glumae ΔobcAB strain was created that synthesizes no oxalic acid.
We measured the growth of B. glumae wild type, ΔdbcA, and ΔobcAB in LB medium buffered with 70 mM Tris at pH 7.0 ( Figure 1a) while monitoring the pH of the medium (Figure 1b). We note that while we used LB medium buffered to pH 7.0 with 70 mM bis-Tris propane (BTP) in our previous study (Iqbal et al., 2021), we found that BTP does not allow wild-type B. glumae to acidify the culture medium, probably due to its wide range of buffering capacity (pH 6.3 to 9.5) ( Figure S1). We therefore used LB medium buffered to pH 7.0 with 70 mM Tris for this study. The pK a of Tris allows B. glumae to produce its natural phenotype (acidification of culture medium) during the stationary phase of growth. During the first 24 h of growth the culture medium pH of wild-type B. glumae decreased from neutral to acidic, while the medium pH of ΔdbcA and ΔobcAB rose from neutral to alkaline (Figure 1b). B. glumae ΔdbcA and ΔobcAB also showed similar levels of growth and culture medium pH in unbuffered LB medium ( Figure S2). Because the cell number shown in Figure 1a was obtained using a spectrophotometer, the growth of all strains was confirmed using plate counts (Figure 1c), which showed that only ΔobcAB lost viability at the stationary phase, while ΔdbcA maintained viability. We conclude that both DbcA and ObcAB are needed for acidification of the culture medium during growth of B. glumae. B. glumae ΔobcAB was significantly less virulent than the wild type based on an onion scale assay (Iqbal et al., 2021;Figure S3), and expression of a cloned copy of obcAB restored the culture medium acidification phenotype of the mutant strain ( Figure S4).

| B. glumae
ΔdbcA is sensitive to divalent cations and resistance can be restored with external sodium oxalate or acidic pH In a previous study, we showed that B. thailandensis ΔdbcA is sensitive to the divalent cations Ca 2+ and Mg 2+ .
We were interested in measuring the cation sensitivity of B. glumae ΔdbcA because divalent cations have been reported to be part of the plant immune response to invading bacterial pathogens (Fones & Preston, 2013;O'Leary et al., 2016). We screened B. glumae ΔdbcA sensitivity against several monovalent (Na + and K + ), divalent (Ca 2+ , Mg 2+ , and Mn 2+ ), and trivalent cations (Al 3+ and Fe 3+ ). We found that It has been shown that oxalic acid is required for aluminium tolerance in Pseudomonas fluorescens and transformation of toxic metals in mining sites by the fungus Beauveria caledonica (Fomina et al., 2005;Hamel et al., 1999). We tested whether supplementation of external oxalate in LB agar can reverse the divalent cation and colistin sensitivity of B. glumae ΔdbcA. Oxalic acid is a strong organic acid with pK a1 1.25 and pK a2 4.27 (Palmieri et al., 2019). To exclude the pH effect, we added oxalate in the form of sodium oxalate, which does not change the pH of the growth medium. We found that F I G U R E 1 Growth and culture medium pH of Burkholderia glumae wild type (336gr-1), ΔdbcA, and ΔobcAB. (a) Growth of B. glumae strains in LB broth buffered to pH 7.0 with 70 mM Tris measured using a spectrophotometer. Equal numbers of cells (5 × 10 7 ) were inoculated into 250-mL culture flasks containing 40 mL of indicated growth medium and grown at 37°C with shaking. (b) At 6-h intervals, a portion of the bacterial culture was aseptically removed to measure the medium pH. (c) The viable cell number of B. glumae, ΔdbcA, and ΔobcAB. Aliquots were taken at the indicated time points, serially diluted, and plated on LB agar plates containing 10 μg/mL nitrofurantoin. Colonies were counted after 48 h at 37°C. The data are presented as mean ± standard deviation (SD). Each experiment was repeated three times with three independent biological replicates.  Divalent cation sensitivity on solid medium. Ten-fold serially diluted cells of B. glumae and ΔdbcA transformed with control vector (vec) and pSC301 (dbcA) were spotted and grown on LB agar containing 100 μg/mL trimethoprim. For determination of cation sensitivity, plates were supplemented with either CaCl 2 , MgSO 4 , or MnCl 2 at the indicated concentrations. For determination of colistin sensitivity, plates were supplemented with either 0 or 100 μg/mL colistin. Sodium oxalate (Na 2 C 2 O 4 ) was added to plates at a concentration of 50 mM to test cation and colistin sensitivity in the presence of external oxalate. LB medium pH was set to 5.5 with hydrochloric acid to test the cation and colistin sensitivity in acidic pH. PC, positive control; NC, negative control. Each experiment was repeated three times with three independent biological replicates. Representative plates are shown.

F I G U R E 3
Oxalic acid levels and acyl-homoserine lactone (AHL) accumulation during growth of Burkholderia glumae wild type (336gr-1), ΔdbcA, and ΔobcAB. (a) Oxalic acid production in LB broth buffered to 7.0 with 70 mM Tris. Inset bar graph shows oxalic acid levels at 6 h. Equal numbers of cells (5 × 10 7 ) were inoculated into either unbuffered or buffered LB broth and grown at 37°C with shaking. Culture supernatants of B. glumae strains were collected by centrifugation at the indicated time points and the oxalic acid level was measured. (b) AHL quantification from culture supernatant of indicated strains grown in buffered LB broth based on β-galactosidase activity. Representative wells are shown. N-octanoyl homoserine (C8-HSL, 10 μM) was added to the positive control, while no C8-HSL was added to the negative control. The data are presented as mean ± standard deviation (SD). Each experiment was repeated three times with three independent biological replicates. Asterisks indicate a statistically significant difference between B. glumae and ΔdbcA. *p < 0.05, **p < 0.01, ***p < 0.001; ns, not significant.

| B. glumae DbcA is required for oxalic acid production
B. glumae produces ammonia due to metabolism of amino acids in LB culture medium, creating alkaline pH toxicity (Goo et al., 2012;Nam et al., 2021), which is prevented by secretion of oxalic acid (Goo et al., 2012). Because B. glumae ΔdbcA showed defects in acidification of growth medium during the stationary phase ( Figure 1b) and resistance to divalent cations could be restored by addition of external oxalate (Figure 2b), we directly measured oxalic acid production of B. glumae strains. We found that wild-type B. glumae consistently produced significantly higher levels of oxalic acid, sufficient to acidify the culture medium, during its growth from the exponential to the stationary phase compared to ΔdbcA (Figure 3a). At 6 h, when no difference in growth was observed between the two strains ( Figure 1a), B. glumae ΔdbcA produced significantly lower levels of oxalate compared to the wild type (Figure 3a, inset). The ΔobcAB oxalate-deficient mutant, which was used as a negative control in the experiments, produced no detectable levels of oxalate.

| B. glumae DbcA is required to preserve AHL molecules at the stationary phase
It has been reported that the QS signalling molecules AHLs undergo inactivation via lactonolysis at alkaline pH (Byers et al., 2002;Le Guillouzer et al., 2020;Yates et al., 2002). Because B. glumae ΔdbcA cannot acidify growth medium during growth due to a defect in oxalic acid secretion, we asked whether B. glumae ΔdbcA is deficient in AHL accumulation, which could influence QS-dependent gene expression. We measured the relative levels of AHLs in growth medium with the β-galactosidase-based biosensor strain Agrobacterium tumefaciens KYC5. B. glumae wild type and ΔdbcA produced roughly the same amounts of AHLs at 6 h during the exponential phase ΔdbcA as well as the oxalic acid operon. However, it remains to be determined how QS regulates oxalic acid production in B. glumae.
We then determined if the ΔdbcA mutation affects the expression of the QS genes tofI and tofR, encoding QS signalling proteins, and qsmR, toxJ, and toxR, encoding regulatory proteins. We found that ΔdbcA mutation did not affect the expression of qsmR, toxJ, or toxR, but expression of tofI and tofR was up-regulated ( Figure 4b).
This important result suggests that the reduction in AHL levels during the stationary phase is solely due to alkalinization of the medium and is not due to a defect in AHL production, because the F I G U R E 4 Expression of toxA, toxH, and obcA is down-regulated in Burkholderia glumae ΔdbcA. (a) Relative normalized expression levels of toxA, toxH, and obcA in B. glumae wild type (336gr-1) and ΔdbcA. (b) Relative normalized expression levels of qsmR, tofI, tofR, toxJ, and toxR in B. glumae and ΔdbcA. The data are presented as mean ± standard deviation (SD). Each experiment was repeated three times with three independent biological replicates. The statistical significance of differences between B. glumae wild type and ΔdbcA was calculated using the unpaired Student's t test. *p < 0.05, **p < 0.01, ***p < 0.001; ns, not significant. mutant strain is able to induce expression of genes involved in AHL synthesis.

| Treatment with sodium bicarbonate prevents oxalic acid production
Previously, we showed that B. glumae DbcA is required to maintain normal PMF (Iqbal et al., 2021). Sodium bicarbonate (NaHCO 3 ) dissipates the ΔpH component of PMF at physiological concentrations (Farha et al., 2018(Farha et al., , 2020Rose et al., 2020). Therefore, we asked whether treatment of B. glumae with 5 mM NaHCO 3 , which does not by itself affect the pH of the culture medium, could reduce oxalic acid production and cause alkalinization of the culture medium. First, we analysed the growth and culture medium pH of B. glumae wild type grown with either 0 or 5 mM NaHCO 3 in LB medium buffered with 70 mM Tris (pH 7.0). We found that B. glumae wild type grown with 5 mM NaHCO 3 did not acidify the growth medium and showed a slight growth defect in the stationary phase compared to the wild-type B. glumae grown without NaHCO 3 (Figure 5a,b). This pattern is like that observed with ΔdbcA when grown in buffered LB medium (Figure 1b).
We measured oxalic acid production in wild-type B. glumae grown with 0 and 5 mM NaHCO 3 . To exclude the effect of cell number on the assay, the oxalic acid production level was measured at the exponential (10 h) and the stationary phase (24 h). We found that wildtype B. glumae grown with 5 mM NaHCO 3 produced significantly less oxalic acid at 10 h (Figure 5c), and little difference in growth was found (Figure 5a). B. glumae wild type grown with NaHCO 3 also produced significantly less oxalic acid at 24 h compared to the strain grown without NaHCO 3 (Figure 5c). We then tested the effect of NaHCO 3 on AHL production during the stationary phase and found that B. glumae wild type grown with NaHCO 3 was compromised for accumulation of AHL at 24 h (Figure 5d,e). However, we did not find a significant difference in AHL levels at 10 h (Figure 5d,e), when the culture medium pH of both strains is near neutral (Figure 5d,e). We also measured the expression levels of toxA, toxH, and obcA for B.
glumae wild type grown with or without NaHCO 3 . We found that expression of these genes was significantly down-regulated in wildtype B. glumae grown with NaHCO 3 (Figure 5f). These results indicate that oxalic acid is probably secreted in a PMF-dependent manner.
Disruption of the PMF with NaHCO 3 causes alkalinization of the culture medium, inactivation of AHLs, and down-regulation of virulence genes (toxA and toxH) and oxalate biosynthesis genes (obcAB).
F I G U R E 5 Sodium bicarbonate (NaHCO 3 ) reduces oxalic acid production in Burkholderia glumae. (a, b) Growth and culture medium pH of B. glumae in buffered LB broth with or without 5 mM NaHCO 3 . Equal numbers of cells (5 × 10 7 ) were inoculated into culture flasks containing buffered LB medium supplemented with either 0 or 5 mM NaHCO 3 . Bacterial cultures were grown at 37°C with shaking for 48 h. (c) Oxalic acid measurement of B. glumae grown in buffered LB broth with or without 5 mM NaHCO 3 . (d) Acyl-homoserine (AHL) quantification from culture supernatant of B. glumae grown in buffered LB broth with or without 5 mM NaHCO 3 based on β-galactosidase activity. (e) Representative wells are shown. N-octanoyl homoserine (C8-HSL, 10 μM) was added to the positive control, while no C8-HSL was added to the negative control. (f) Expression levels of toxA, toxH, and obcA in B. glumae grown in buffered LB broth with or without NaHCO 3 . The data are presented as mean ± standard deviation (SD). Each experiment was repeated three times with three independent biological replicates. *p < 0.05, **p < 0.01, ***p < 0.001; ns, not significant.

| External C8-HSL rescues oxalic acid production in B. glumae ΔdbcA
It has been reported that alkalinization of the culture medium in B. glumae BGR1 QS mutants, ΔtofI and ΔqsmR, can be reversed by addition of external C8-HSL in growth medium (Goo et al., 2012).
Because we showed that B. glumae ΔdbcA was deficient for AHL accumulation during the stationary phase, we tested if addition of external C8-HSL reverses the culture medium alkalinization. We analysed the growth and culture medium pH of B. glumae wild type and ΔdbcA grown with or without a physiologically relevant concentration of 5 μM C8-HSL in buffered LB medium. We found that addition of C8-HSL completely reversed culture medium alkalinization and that B. glumae ΔdbcA displayed a growth rate similar to that of the wild type (Figure 6a,b). Addition of C8-HSL did not affect the growth and culture medium pH of the wild type (Figure 6a,b). We measured oxalic acid production in B. glumae wild type and ΔdbcA grown with or without C8-HSL and found that its addition completely restored oxalic acid production in B. glumae ΔdbcA (Figure 6c). C8-HSL also provided a partial recovery of toxoflavin production by B. glumae ΔdbcA (Figure 6d). These data collectively show that a defect in QS is linked to each of these phenotypes of ΔdbcA and there may exist additional contributors in the case of toxoflavin production.
We then measured the expression levels of toxA, toxH, and obcA in B. glumae wild type and ΔdbcA grown with or without C8-HSL. We found that expression of obcA was significantly increased in B. glumae ΔdbcA grown with C8-HSL (Figure 6e). This result indicated that external C8-HSL induces expression of the obcAB operon, which in turn produces more oxalic acid to reverse the culture medium alkalinization of B. glumae ΔdbcA (Figure 6b), consistent with the observation that oxalate production is regulated in a QS-dependent manner (Goo et al., 2012). We also found that the expression levels of toxA and toxH were increased in B. glumae ΔdbcA grown with C8-HSL, albeit not to the levels observed in wild-type cells (Figure 6e). This result is consistent with the partial complementation of toxoflavin production in B. glumae ΔdbcA grown with C8-HSL (Figure 6d).

| DISCUSS ION
Previously, we showed that DbcA is required for colistin resistance, toxoflavin production, and virulence of B. glumae. Chemical alteration of the PMF by NaHCO 3 treatment can also cause loss of virulence of B. glumae (Iqbal et al., 2021). We proposed NaHCO 3 as a potential chemical agent for BPB intervention in rice. In this study, we examined the impact of the ΔdbcA mutation and NaHCO 3 treatment on QS, the master regulator of virulence in B. glumae (Chen et al., 2012;Kim et al., 2004Kim et al., , 2007Lelis et al., 2019;Peng et al., 2020). We show that DbcA is required for oxalic acid production, growth medium acidification at the stationary phase, accumulation of AHL, and transcription of QS-dependent genes (Figures 1, 3, 4, and 6d). We show that alteration of the PMF in B. glumae with NaHCO 3 can also reduce oxalic acid production and cause alkalinization of the culture medium, which in turn results in reduced AHL levels in the stationary phase ( Figure 5). We also show that addition of external C8-AHL can restore the oxalic acid production and medium acidification phenotype in B. glumae ΔdbcA (Figure 6). We show for the first time that B. glumae DbcA is required for maintenance of proper QS via its necessity for oxalic acid secretion.
Burkholderia and other bacterial species use amino acids as a major carbon source in rich LB medium and produce ammonia due to deamination of amino acids (Goo et al., 2012). Production of ammonia increases the pH of the culture medium, causing alkaline pH toxicity to the bacterial cell. Burkholderia species produce and secrete oxalic acid to neutralize the ammonia-mediated alkaline pH toxicity (Goo et al., 2012;Nam et al., 2021). It has been reported that B. glumae BGR1 QS (ΔqsmR and ΔtofI) and oxalate (ΔobcA and ΔobcB) mutants display a "massive population crash" when the culture medium pH exceeds 8.0 (Goo et al., 2012). We created a B. glumae ΔobcAB mutant to compare the growth of an oxalate-deficient mutant with that of ΔdbcA and analyse the loss of viability during the stationary phase. We found that B. glumae ΔobcAB underwent such a population crash during the stationary phase, but this was not observed for ΔdbcA although the growth medium of both strains underwent alkalinization to a similar extent. It is possible that the low amount of oxalic acid produced by ΔdbcA (Figure 3a) allows better population survival into the stationary phase.
Oxalic acid is a strong metal chelator that can form an oxalatemetal complex (Fomina et al., 2005;Palmieri et al., 2019) and may therefore chelate divalent cations (Ca 2+ , Mg 2+ , and Mn 2+ ). As a result, B. glumae ΔdbcA showed resistance to divalent cations when external oxalate was provided, even when delivered in the form of a sodium salt, suggesting a direct role for oxalate in reducing sensitivity to divalent cations. Among all metals, the role of Ca 2+ in the innate immune responses of plants is well understood (Fones & Preston, 2013;Gao et al., 2021). The plant apoplast is a dynamic compartment containing water, nutrients, sugars, and organic acids (Sattelmacher, 2001). The apoplast is surrounded by cell walls and can support the growth of pathogenic bacteria (O'Leary et al., 2016).
The plant cell wall contains Ca 2+ , which acts as a secondary intracellular messenger (Nishad et al., 2020). The increased concentration of cytosolic Ca 2+ triggers several pathogen-mediated immune responses, including accumulation of H 2 O 2 and generation of oxidative burst at the infection site (Grant et al., 2000). Plants can also alkalize the apoplastic pH in response to pathogen invasion in a response mediated by plant peptide-receptor complexes (Liu et al., 2022). In this context, it is plausible that B. glumae may suppress the pathogenmediated plant innate immunity by secreting oxalic acid to chelate apoplastic Ca 2+ and acidify the apoplast.
No oxalic acid efflux transporter has been identified in B. glumae.
However, the anaerobic gram-negative bacterium Oxalobacter formigenes encodes an oxalate:formate antiporter that imports oxalate in exchange for formate (Hirai & Subramaniam, 2004). A secondary oxalate efflux transporter (FpOAR) has been identified in the fungus Fomitopsis palustris (Watanabe et al., 2010). FpOAR is a PMFdependent oxalate efflux transporter and displays no similarity with other known oxalate transporters. The export activity of FpOAR can be significantly inhibited by abolishing either ΔΨ or the ΔpH component of the PMF (Watanabe et al., 2010). It is possible that B. glumae ΔdbcA initially secretes less oxalic acid due to a compromised PMF (Iqbal et al., 2021). Alternatively, DbcA may be directly involved in oxalic acid secretion.
We used NaHCO 3 to verify that PMF is required for oxalic acid production. Sodium bicarbonate is a common buffer and can dissipate the ΔpH component of the PMF at physiological concentrations and modify bacterial sensitivity to several types of antibiotics (Farha et al., 2018(Farha et al., , 2020Rose et al., 2020). Previously, we showed that NaHCO 3 alters the PMF by partially increasing ΔΨ in B. glumae wild type, chemically replicating the ΔdbcA phenotype (Iqbal et al., 2021).
We tested whether alteration of the PMF with NaHCO 3 can reduce oxalic acid production and found a significant reduction (Figure 5c).
Due to impaired oxalic acid production, B. glumae wild type grown with NaHCO 3 could not acidify the culture medium and was deficient in AHL production during the stationary phase (Figure 5b,d).
This result suggests that alteration of the PMF in B. glumae can reduce oxalic acid production, creating alkaline conditions, resulting in degradation of AHLs and down-regulation of toxoflavin and oxalic acid production. B. glumae DbcA plays an important role in maintaining normal oxalic acid production and QS. All these effects are reversed by addition of C8-HSL to the culture medium ( Figure 6).

B. glumae regulates its virulence factors in a QS-dependent man-
ner (Chen et al., 2012;Kim et al., 2004). We tested whether alkalinization of the culture medium can affect QS in B. glumae ΔdbcA.
While B. glumae wild type and ΔdbcA accumulated similar levels of AHLs in their early phases of growth, B. glumae ΔdbcA accumulated a significantly lower level of AHLs when the culture medium pH became alkaline in the stationary phase (Figure 3b). Because the stability of AHLs is highly dependent on the pH of the culture medium (Byers et al., 2002;Le Guillouzer et al., 2020;Yates et al., 2002), it is likely that the reduced AHL levels measured during the stationary phase are due to the alkaline pH of the medium, and this in turn is responsible, at least in part, for down-regulation of the tox and obc operons in B. glumae ΔdbcA.
Our results indicate that both DbcA and QS are required for oxalic acid production and growth of B. glumae. From our results, we conclude that B. glumae DbcA is required to establish a synergistic link between the PMF and QS in which both are presumably dependent upon each other for the regulation of toxoflavin F I G U R E 6 N-octanoyl homoserine (C8-HSL) restores oxalic acid production in Burkholderia glumae ΔdbcA. (a, b) Growth and culture medium pH of B. glumae wild type (336gr-1) and ΔdbcA in buffered LB broth with or without 5 μM C8-HSL. Equal numbers of cells (5 × 10 7 ) were inoculated in a culture flask containing buffered LB broth and grown at 37°C with shaking. (c) Oxalic acid measurement of B. glumae and ΔdbcA grown in buffered LB broth with or without C8-HSL. (d) Toxoflavin production by B. glumae and ΔdbcA grown in buffered LB broth with or without C8-HSL. (e) Expression levels of toxA and toxH, and obcA in B. glumae and ΔdbcA grown in buffered LB broth with or without C8-HSL. The data is presented as mean ± standard deviation (SD). Each experiment was repeated three times with three independent biological replicates. *p < 0.05, **p < 0.01, ***p < 0.001; ns, not significant. production and virulence (Figure 7). Evidence for a direct role for DbcA in oxalic acid secretion awaits further structural and biochemical studies.

| Analysis of bacterial growth and culture medium pH
B. glumae strains were directly streaked onto LB agar plates containing no antibiotics from −80°C freezer stocks. Plates were incubated for 36 or 48 h. Bacterial colonies were transferred from LB agar plates with sterile loops and suspended in 1 mL sterile LB medium.
Equal numbers of cells (5 × 10 7 ) were inoculated into 250-mL conical flasks containing 40 mL fresh medium without antibiotics and grown for up to 48 h at 37°C with shaking. At 6-h intervals, aliquots of bacterial cultures were aseptically removed from flasks to measure the bacterial cell number using a Bio-Rad Smartspec Plus spectrophotometer and pH using a standard pH meter.

| Transformation and complementation analysis
Heat shock was used for transformation of E. coli unless otherwise stated (Froger & Hall, 2007). Biparental conjugation was used for transformation of B. glumae as previously described (Iqbal et al., 2021;López et al., 2009

F I G U R E 7
Inactivation of dbcA or treatment with NaHCO 3 results in a series of events leading to loss of toxoflavin production and virulence of Burkholderia glumae. Secretion of oxalic acid lowers the pH of the bacterial environment, which prevents nonenzymatic degradation of quorum sensing (QS) AHL signalling molecules (Yates et al., 2002). Plants respond to bacterial infection by producing metabolites that cause alkalinization of the apoplastic space (Geilfus, 2017;Nachin & Barras, 2000;O'Leary et al., 2016). QS activates expression of the tox operons required for virulence and the obc operon for oxalic acid synthesis (Kim et al., 2004;Nakata & He, 2010). Reduction of oxalic acid secretion by B. glumae ΔdbcA or exposure to NaHCO 3 prevents acidification, interfering with QS and tox expression, which in turn reduces the virulence of B. glumae (Goo et al., 2012). Loss of QS also represses the expression of the obc operon (Goo et al., 2017) and further reduces oxalic acid production and potentially amplifies the alkaline pH conditions.

| Deletion of the obcAB operon from B. glumae 336gr-1
Deletion of the entire obcAB operon was performed using homologous recombination as previously described (Melanson et al., 2017). Oligonucleotide primers used for the deletion of obcAB are listed in Table S1. The GeneElute Bacterial Genomic DNA Extraction Kit (Sigma-Aldrich) was used to extract the genomic DNA from B. glumae 336gr-1. Q5 DNA polymerase (New England Biolabs) was used for PCR amplification. The QuickClean 5M PCR purification kit (GeneScript) was used to purify the PCR products.
The 392-bp upstream and 421-bp downstream regions of the obcAB operon were amplified from B. glumae 336gr-1 genomic DNA.
The 3′ end of the upstream fragment has 20 bp homology with the 5′ end of the downstream fragment. The upstream and downstream fragments were assembled by PCR to generate an obcAB deletion construct using the Oxalate-upNEW2FP and Oxalate-DWN2RP primers (Hilgarth & Lanigan, 2020 Figure S6).
To construct a complementation plasmid expressing obcAB, the operon was PCR-amplified from genomic DNA of B. glumae using primers obc_Fwr_NdeI and obc_Rv_HindIII (Table S1). The purified PCR product was treated with NdeI and HindIII and ligated into the corresponding restriction sites of the expression vector pSCrhaB2, resulting in pSC700 (Table 1).

| Susceptibility assays
Sensitivity was measured in liquid medium in 96-well plates or on solid medium using the broth microdilution method by spotting 5 μL of 10-fold serially diluted bacterial cells. The plates were incubated at 37°C, and bacterial growth was analysed after 48 h of incubation.

| Oxalic acid measurement
Oxalic acid measurement was performed using an oxalate colourimetric assay kit (Abcam) according to the manufacturer's protocol (Liu et al., 2021). Culture supernatants were diluted into 50 μL oxalate assay buffer. Two microlitres of oxalate converter was added and the tubes were incubated at 37°C in the dark. After 1 h, 50 μL of oxalate development master mix (46 μL oxalate development buffer, 2 μL oxalate enzyme mix, and 2 μL oxalate probe) was added to each tube and incubation was continued for 1 h at 37°C in the dark. Absorbance was measured at 450 nm. The oxalic acid concentration was calculated using an oxalate standard curve ( Figure S7).

| AHL quantification
Production of AHLs was determined using the β-galactosidasebased biosensor strain A. tumefaciens KYC55 (Zhu et al., 2003), which responds to the presence of AHLs by expressing lacZ (βgalactosidase; Barton et al., 2021;Gupta et al., 2017). Culture supernatants were collected by centrifugation at the indicated time points and passed through a 0.22μm filter. An equal number of KYC55 cells (5 × 10 7 cfu/mL) was added to culture tubes containing 5 mL AT broth supplemented with 40 μg/mL X-gal. Five microlitres of B. glumae supernatant was added and samples were incubated at 28°C for 6 h.
Control tubes were supplemented with 0 or 10 μM C8-HSL. The development of blue colour was read at 635 nm. The absorbance of the negative control was subtracted from the absorbance of each sample. Data were normalized to the OD 600 value of the tested culture.

| Toxoflavin measurement
Measurement of the toxoflavin level in culture media was performed as previously described (Iqbal et al., 2021).

| RNA isolation and RT-qPCR
Overnight cultures were diluted 1:100 in fresh LB medium without antibiotics and grown to an OD 600 value of approximately 0.6 at 37°C with shaking. Three millilitres of bacterial cultures was col- The Luna Universal One-Step RT-qPCR Kit (New England Biolabs) was used to perform qPCR. qPCR was performed in a 20-μL reaction mixture containing 1× Luna Universal One-Step reaction mix, 1× Luna WarmStart RT enzyme mix, 0.4 μM gene-specific forward and reverse primers, and 300 ng RNA. RT-qPCR was performed using an Applied Biosystems QuantStudio 6 Flex Real-Time PCR system using SYBR Green I dye with the following PCR conditions: reverse transcription at 55°C for 10 min and initial denaturation at 95°C for 1 min, followed by 40 cycles of denaturation at 95°C for 10 s and extension at 60°C for 1 min. A melt curve was produced for each run at a temperature range from 60°C to 95°C with 1°C increments. The comparative C t method (2 −ΔΔCt ) was used to calculate the fold change value of gene expression using the housekeeping gene gyrA as an internal reference (Lelis et al., 2019). Statistical analysis was performed using the unpaired Student's t test with GraphPad Prism 9.

| Statistical analysis
The data are presented in the graphs as mean ± standard deviation (SD). Each experiment was repeated three times with three independent biological replicates. Graphs were produced with GraphPad Prism v. 9.0 and statistical significance was calculated using the unpaired Student's t test.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.