The bacterial strains and plasmids used in this study are listed in Supplementary Table S1. All B. glumae and E. coli strains were aerobically grown in LB broth [0.1% tryptone, 0.5% yeast extract, and 0.5% NaCl (w/v); USB, Cleveland, OH, United States] or in LB broth buffered with 100mM 4-(2-hydroxyethyl) piperazin-1 ethanesulfonic acid (HEPES; Sigma-Aldrich, St. Louis, MO, United States) to maintain external pH at pH 7 at 37°C. When necessary, the LB broth was supplemented with 1μM C8-HSL, 10μg/ml gentamicin, and 50μM nigericin for growth experiments. For the alkaline-stress experiments, the wild-type B. glumae BGR1 and E. coli DH5α strains were diluted to an optical density (OD) of 0.05 based on absorbance at 600nm (A₆₀₀) and grown overnight for subculture. The cultures were washed twice with LB broth and transferred to LB broth buffered with 100mMN-(1,1-dimethyl-2-hydroxyethyl)-3-amino-2-hydroxypropanesulfonic acid (AMPSO) at pH 8, 9, or 10 (Sigma-Aldrich).

Wild-type B. glumae BGR1 and the QS mutants BGS2 (BGR1 tofI::Ω) and BGS9 (BGR1 qsmR::Ω) were grown at 37°C for 20, 24, and 28h in LB broth. To induce alkaline stress, wild-type BGR1 was transferred to LB broth buffered with AMPSO (pH 9) for 1, 4, and 8h. These cells were then harvested via centrifugation at 10,000×g for 10min and examined using transmission electron microscopy (TEM) as previously described (Kang and Hwang, 2018). Electron micrographs were acquired using a LIBRA 120 energy-filtration microscope (Carl Zeiss, Oberkochen, Germany) at 100kV.

Burkholderia glumae and E. coli DH5α cells were collected and stained with the following fluorescent dyes: LIVE/DEAD BacLight™, 40,6-diamidino-2-phenylindole dihydrochloride (DAPI), SYTO™ RNASelect™ Green Fluorescent cell stain, and CellROX™ Deep Red Reagent.

The numbers of live and dead bacteria were determined using a LIVE/DEAD BacLight™ bacterial viability kit (Life Technologies, Carlsbad, CA, United States). The cells were stained with a mixture of SYTO9 and propidium iodide (1:1) for 20min in the dark at room temperature (RT). DNA was stained with DAPI (Life Technologies) at a final concentration of 20ng/ml in the dark for 20min at RT, and RNA was visualized using the SYTO™ RNASelect™ Green Fluorescent cell stain (Life Technologies) at a final concentration 500nM in the dark for 20min at 37°C. ROS levels were quantified after staining with CellROX™ Deep Red Reagent (Life Technologies) as previously described with some modifications (Grinberg et al., 2012, 2013). CellROX™ reagent at a final concentration of 5μM was added to B. glumae or E. coli DH5α, and the cells were incubated in the dark for 30min at 37°C. All stained cells were washed twice and resuspended in phosphate-buffered saline (PBS), with the exception of cells stained with SYTO™ RNASelect™, which were washed with LB broth. Fluorescence images were immediately obtained using an Olympus BX53 fluorescent microscope (Olympus, Tokyo, Japan).

To examine membrane disturbance, B. glumae and E. coli DH5α cells were harvested via centrifugation and resuspended in PBS. Subsequently, the cells were treated with 5μM 3,3′-dipropylthiadicarbocyanine iodide [DiSC₃(5); Life Technologies] and 19μM bis-(1,3-dibutylbarbituric acid) trimethine oxonol [DiBAC₄(3); Life Technologies] for 20min in the dark at RT. DiSC₃(5) levels were measured using a SpectraMax M2e microplate reader (Molecular Devices, San Jose, CA, United States), and DiBAC₄(3) signals were observed under an Olympus BX53 fluorescent microscope (Olympus). In addition, 50μM nigericin was added to wild-type BGR1 to induce factitious membrane depolarization.

An RNeasy Mini kit (Qiagen, Venlo, Netherland) was used to isolate total RNA from wild-type B. glumae BGR1 and QS mutants in LB broth cultured at 37°C for 10h. Total RNA was isolated from wild-type BGR1 after 4h of alkaline stress at pH 9 or at 8h after the addition of 50μM nigericin. RNase-free DNaseI treatment of the isolated RNA for 30min at 37°C, reverse transcription to cDNA at 42°C for 1h, and reverse-transcription PCR (RT-PCR) were performed as described previously (Jang et al., 2014). The primer pairs used for quantitative reverse-transcription PCR (qRT-PCR) are listed in Supplementary Table S2. qRT-PCR analysis was performed to compare transcriptional levels using SsoFast EvaGreen Supermix (Bio-Rad, Hercules, CA, United States) and the CFX96 Real-Time PCR system (Bio-Rad). The thermal cycling parameters were as follows: 95°C for 30s, followed by 35cycles of 95°C for 5s and 60°C for 5s. All reactions were run in triplicate from three independent cultures with three technical repeats per dose, and the 16S rRNA gene was used for data normalization.

For the quantification of DNA damage, we evaluated the levels of damage to DNA gyrase A (BGLU_1G08730) relative to that in undamaged controls as described previously (Patenge, 2017). Amplifications were normalized by DNA gyrase A copy number by concurrently amplifying a “short-range” PCR fragment. For the DNA gyrase A product, we selected the following primer sequences: long forward, 5′-CGAACTGAACAACGACTGGA-3′; short forward, 5′-AGCTGCAGGACACGTTCG-3′; and long and short reverse, 5′-GCATCTCCTTCAGGTTCAGC-3′. The lesion frequency per amplicon was calculated assuming a random distribution of lesions as described previously (Patenge, 2017).

To determine the protein expression levels in wild-type BGR1 and QS mutants, GFP-tagged Hu-beta and green fluorescent protein (GFP)-tagged HNH endonuclease were constructed as described previously (Jang et al., 2014). Hu-beta and HNH endonuclease genes with the promoter region and lacking the stop codon were cloned into pJW23 containing an enhanced GFP (eGFP) gene and subcloned into pUC18R6KT miniTn7T-Tc. Subsequently, Mini-Tn7 expressing Hu-beta and HNH endonuclease with eGFP was integrated into the glmS gene of B. glumae (Choi et al., 2006). GFP fluorescence was measured using a PerkinElmer Victor X3 microplate reader (PerkinElmer, Waltham, MA, United States).

To measure the internal pH, we constructed the plasmid pPHI4 containing the trc promoter and ratiometric pHluorin. For plasmid construction, the ratiometric pHluorin gene (GenBank accession no. AF058694.2) was synthesized. Overnight bacterial cultures were aerobically grown in LB broth at 37°C and diluted to an OD of 0.05 based on A₆₀₀. After 14h of subculture, bacteria in 2ml of each culture were harvested via centrifugation (10,000×g for 2min), washed twice with 1ml of Dulbecco’s phosphate-buffered saline (DPBS; 150mM NaCl, 3mM KCl, 1mM KH₂PO₄, 6mM Na₂HPO₄, 0.5mM MgCl₂, and 1mM CaCl₂) as previously described with some modifications (Reifenrath and Boles, 2018), and resuspended in 1ml of DPBS. To induce alkaline stress, wild-type B. glumae BGR1 and E. coli DH5α were placed in DPBS at pH 9.

To construct a calibration curve to measure the internal pH in all strains, harvested cells were equilibrated in DPBS adjusted to different pH values (pH 5, 6, 7, 8, and 9) in the presence of 40mM potassium benzoate and 40mM methylamine hydrochloride before detection of fluorescence intensity (Martinez et al., 2012; Reyes-Fernández and Schuldiner, 2020). Benzoic acid (permeant acid) and methylamine hydrochloride (permeant base) were used to collapse the pH difference across the membrane and equialize the external and internal pH. Aliquots of 200μl of each cell suspension were distributed into a 96-well plate, and the fluorescence produced was measured by scanning a range of excitation wavelengths (330–482nm) with emission at 510nm using a SpectraMax M2e microplate reader (Molecular Devices). Because the excitation peaks of dual wavelength in B. glumae and E. coli DH5a were different, the 370/470-nm fluorescence excitation ratios (R_370/470) of B. glumae and R_410/470 values of E. coli DH5α were determined, respectively. Finally, the calibration curves were constructed using designated excitation ratio, R_370/470 or R_410/470, and the internal pH was calculated as described previously (Martinez et al., 2012; Crauwels et al., 2018; Arce-Rodríguez et al., 2019; Reyes-Fernández and Schuldiner, 2020).

DNA cloning, restriction mapping, and gel electrophoresis were used to introduce nhaA into E. coli (Sambrook et al., 1989). We constructed pNha13, in which the native promoter and nhaA of E. coli (GenBank accession no. YP_001729002.1) were expressed. For plasmid construction, the coding region of the native promoter and nhaA was amplified with the primers NhaA_F (5′-CCAAGAGCTCCTATCTGCCGTTCAGCTAATGC-3′) and NhaA_R (5′-ACCGTGGGCCCCGTGTCA-3′′) using the genomic DNA of E. coli DH5α. The amplified PCR product containing SacI and KpnI recognition sites was inserted into pLAFR6, resulting in pNha13 (Supplementary Tables S1, S2).

To test the constitutive expression of genes encoding putative Na⁺/H⁺ antiporter genes from B. glumae, the promoter region was replaced with the trc promoter. Genes annotated as related to the Na⁺/H⁺ antiporter of B. glumae were amplified with the primers listed in Supplementary Table S2. The amplified PCR product was digested with NdeI and BamHI and inserted into pBP1 such that the translation initiation codon of the putative Na⁺/H⁺ antiporter gene was fused to the trc promoter. The DNA fragments carrying the putative Na⁺/H⁺ antiporter genes (BGLU_1G32530, 1G09320, 2G00450, and 2G17030) with trc promoter obtained from pNha4, pNha5, pNha22, and pNha23 via digestion performed with SacI and KpnI, SacI and BamHI, or XbaI and HindIII, were cloned into pLAFR6 to produce pNha7, pNha8, pNha24, and pNha25. Furthermore, pLAFR6 containing E. coli nhaA with trc promoter was designated pTrc-nhaA (Supplementary Table S1).

All statistical analyses and ANOVA testing followed by Tukey’s honest significance difference post hoc analysis were performed using IBM SPSS Statistics software (ver. 20 x86-x64; IBM Corp., Armonk, NY, United States).

To determine whether cell death in QS mutants at alkaline environmental pH is similar to the known apoptosis-like cell death in other bacteria, we first observed the cellular contents of two QS mutants, BGS2 (BGR1 tofI::Ω) and BGS9 (BGR1 qsmR::Ω), using TEM during the stationary phase. TEM images showed that the cellular materials of the QS mutants were less negatively stained with uranyl acetate than those of wild-type BGR1 (Figure 1). Uranyl acetate, which produces negative stain contrast by interaction with lipids, proteins, and nucleic acid phosphate groups, was used as a stain and fixative for electron microscopy. Cytoplasmic membranes were intact at the initial stage of cell death in the QS mutants; these mutants appeared as ghost-like cells at 24h after incubation in LB broth at 37°C (Figure 1).

To further highlight similarities between cell death in the QS mutants and the known apoptosis-like cell death in other bacteria, we employed cell biological tools. Live and dead cell staining with the LIVE/DEAD BacLight™ kit revealed that the vast majority of QS mutant cells were dead (red) after 24h (Figure 2A). DAPI staining showed that DNA in the QS mutants was disintegrated, unlike the condensed forms observed in wild-type BGR1, and RNASelect™ staining indicated that RNA appeared to be degraded in the QS mutants compared to wild-type BGR1 based on fluorescence intensity (Figure 2A). Furthermore, membrane depolarization was observed in the QS mutants upon DiBAC₄(3) staining (Figure 2A). Excessive generation of ROS was detected in the QS mutants when cells in the stationary phase were stained with CellROX™ Deep Red Reagent (Figures 2A,B). However, when BGS2 (BGR1 tofI::Ω) was cultured in LB broth supplemented with 1μM C8-HSL or 100mM HEPES (pH 7), all cell-death phenotypes disappeared, and the phenotype was restored to that of wild-type BGR1 (Figure 2).

Because cell death in the QS mutants was due to the alkaline environmental pH, we examined whether artificially amended alkaline conditions induced the same phenomenon in wild-type BGR1. The phenotype of wild-type BGR1 exposed to pH 9 was similar to that of dead QS mutants in the stationary phase in the TEM images (Figure 3A); consequently, these cells exhibited a decreased population density (Figure 3B). Cell biological assays of wild-type BGR1 challenged with alkaline stress also produced results similar to those obtained for the QS mutants during the stationary phase, suggesting DNA disintegration and RNA degradation (Figure 4A). The occurrence rates of ROS and membrane depolarization in wild-type BGR1 were higher under alkaline stress than under normal conditions (Figures 4A,B). Monitoring of the membrane charge-sensitive dye DiSC₃(5) verified the occurrence of membrane depolarization in wild-type BGR1 under alkaline stress (Figure 4C). Escherichia coli DH5α cells exposed to pH 9 did not exhibit an apoptosis phenotype (Figure 4). Moreover, wild-type BGR1 treated with gentamicin, a bacterial antibiotic that binds the 30S subunit of the bacterial ribosome and disturbs protein synthesis by misreading mRNA, did not exhibit an apoptosis-like cell-death phenotype despite dead cells being observed with LIVE/DEAD staining (Figures 4A,B). These results suggest that the cell death mechanism induced by alkaline stress in B. glumae was different than that attributable to gentamicin.

To observe whether DNA damage similar to that reported in apoptosis-like cell death in other bacteria occurred during cell death in QS mutants, we evaluated the extent of DNA damage in wild-type BGR1 and the QS mutants BGS2 (BGR1 tofI::Ω) and BGS9 (BGR1 qsmR::Ω) using qRT-PCR. Quantitative analysis of nucleic acid damage in wild-type BGR1 and the QS mutants showed lesion rates suggesting fivefold higher damage per 10kb DNA in the QS mutants than in wild-type BGR1 after 24h (Figure 5A). We also showed that artificially adjusted alkaline conditions caused increased DNA damage, fivefold that observed in wild-type BGR1 under non-alkaline conditions (Figure 5B).

In addition to the roles of ROS in DNA damage, we also investigated whether nucleotide-degrading enzymes were involved in DNA damage in the QS mutants. Based on our previous RNA sequencing results from comparative studies on the gene expression profiles of wild-type BGR1 and the QS mutants (Supplementary Table S3; Goo et al., 2012), three genes encoding nucleic acid-degradation enzymes, HNH endonuclease (BGLU_1G15690), oligoribonuclease (BGLU_1G09120), and ribonuclease E (BGLU_1G09400), and a gene encoding the chromosomal binding protein Hu-beta (BGLU_1G13530) were chosen for further analysis. The qRT-PCR results showed that the expression of these genes was higher in the QS mutants than in wild-type BGR1 after 10h (Figure 6A). However, when both the QS mutants and wild-type BGR1 were grown in LB broth supplemented with 100mM HEPES (pH 7), the expression levels of the four genes were similar (Figure 6A). The increase in expression of the four genes was similarly reproduced in wild-type BGR1 exposed to alkaline stress (Figure 6B, upper panel). In addition, measurements of fluorescence intensity of GFP-tagged HNH endonuclease and Hu-beta signals confirmed that the expression of HNH endonuclease and Hu-beta was increased both in the QS mutants and wild-type BGR1 under alkaline stress (Supplementary Figure S1).

To investigate whether membrane depolarization occurs during apoptosis-like cell death in QS mutants, we evaluated membrane depolarization in QS mutant and wild-type BGR1 cells undergoing apoptosis-like cell death following alkaline stress. When 50μM nigericin was added to wild-type BGR1 to induce membrane depolarization, the cell death patterns were similar to those observed with alkaline stress-induced cell death (Figures 4A,B). Addition of the membrane depolarization-inducing agent nigericin to wild-type BGR1 caused a significant increase in the expression of the HNH endonuclease gene compared to untreated wild-type BGR1 (Figure 6B, bottom panel). However, the expression of the genes encoding oligoribonuclease, ribonuclease E, and Hu-beta in wild-type BGR1 was not significantly affected by the addition of 50μM nigericin, unlike with alkaline stress (Figure 6B, bottom panel).

Because alkaline stress-induced cell death was correlated with membrane depolarization, cellular pH in B. glumae was examined under alkaline environmental conditions using the pH-sensitive ratiometric probe pHluorin. The fluorescent excitation spectra following emission at 510nm showed clear differences across various pH ranges between wild-type BGR1 and BGS2 (BGR1 tofI::Ω; Figure 7A). The standard calibration curve and internal pH for wild-type BGR1 and BGS2 (BGR1 tofI::Ω) were obtained by calculating R_370/470 values (Figure 7B). The external pH for wild-type BGR1 and BGS2 (BGR1 tofI::Ω) was 5.2 and 8.5, respectively, as the cultures entered the stationary phase (Figure 7C). In wild-type BGR1, the internal pH was maintained at 6.6 even though the external pH of the culture fluid was acidic at pH 5.2, whereas the internal pH of BGS2 (BGR1 tofI::Ω) reached 8.8 (Figure 7C). Using the same method, we determined the internal pH of B. glumae BGR1 and E. coli DH5α following exposure to pH 9 (Figure 8). The fluorescent excitation spectra following emission at 510nm were obtained, and the standard calibration curve was calculated for measuring internal pH (Figures 8A,B). The internal pH of B. glumae, which was maintained at pH 6.5 before exposure to pH 9, rapidly increased to 8.8 after exposure to pH 9 and did not recover, whereas the internal pH of E. coli remained under 8 and did not fluctuate significantly after the shift to pH 9 (Figure 8C).

Because the maintenance of a homeostatic neutral cellular pH in alkali-tolerant bacteria generally relies on Na⁺/H⁺ antiporter function, we tested two scenarios. First, we tested whether the constitutive expression of genes annotated as Na⁺/H⁺ antiporters alleviated alkaline toxicity in B. glumae. Second, we heterologously expressed genes involved in maintaining cellular pH homeostasis in other bacteria to determine whether their heterologous expression would affect the viability of B. glumae under toxic alkaline conditions. Four putative Na⁺/H⁺ antiporter genes in B. glumae exhibited low homology with Na⁺/H⁺ antiporter genes of known alkali-resistant bacteria (Supplementary Table S4). The constitutive expression of each gene in wild-type BGR1 under the trc promoter did not rescue alkali toxicity (Supplementary Table S4). On the other hand, when the Na⁺/H⁺ antiporter gene nhaA derived from E. coli was expressed from its native promoter in B. glumae, the population density of wild-type BGR1 with the nhaA gene in pLAFR6, named pNha13, declined less during 8h of exposure to pH 9 compared to the density of wild-type BGR1 with the empty vector pLAFR6 (Figure 9). When nhaA was constitutively expressed with the trc promoter, the population density of wild-type BGR1 carrying pTrc-nhaA indicated similar viability to wild-type BGR1 carrying pNha13; however, both could not completely overcome the alkaline toxicity, and cell death was observed after 24h of alkaline stress (Figure 9).