Introduction

Non-inheritable antibiotic resistance or phenotypic resistance represent a serious therapeutic challenge for bacterial infectious diseases. Phenotypic resistance does not involve genetic mutations, instead, this type of resistance are transient. Thus, bacteria resume to normal growth after removing the antibiotics. Two extensively studied phenotypic resistance includes biofilm and bacterial persisters (Corona & Martinez, 2013). Biofilms have complex structures that impede the diffusion of antibiotics and contain elements sequestrating antibiotics to prevent their further actions (Ciofu et al, 2022). Moreover, biofilm-forming bacteria and persisters themselves have distinct metabolic state that also significantly negatively influence their susceptibility to antibiotics (Yan & Bassler, 2019). These two types of phenotypic resistance have common phenomena that their growth are retarded or even stopped in the presence of antibiotics (Corona & Martinez, 2013). However, factors promoting bacterial phenotypic resistance that is still proliferate in the presence of antibiotics are poorly defined.

There is a growing body of evidence that metal ions have diverse impact on chemical, physical and physiological phenomena that play roles in development of antibiotic resistance (Booth et al, 2011; Lu et al, 2020; Poole, 2017). One mechanism could involve genetic elements that harbor genes conferring resistance to metals and antibiotics (Poole, 2017) . On the other hand, recent reports suggest that metal cations directly hinder (or enhance) the activity of specific antibiotic drugs (Zhang et al, 2014) and that the metabolic environment impacts bacterial sensitivity to antibiotics (Jiang et al, 2023; Lee & Collins, 2012; Peng et al, 2015; Zhang et al, 2020; Zhao et al, 2021). Light metal ions, such as magnesium, sodium, and potassium, serve as cofactors to many enzymes (Du et al, 2016), some of which could influence drug efficacy. Heavy metal ions, including Cu²⁺ and Zn²⁺, confer resistance to antibiotics (Yazdankhah et al, 2014; Zhang et al, 2018). Recent reports suggest that sodium negatively regulates redox states to promote antibiotic resistance in Vibrio alginolyticus (Yang et al, 2018), while actively growing Bacillus subtilis cope with ribosome-targeting antibiotics based on ion flux modulation (Lee et al, 2019). But whether such ions can promote phenotypic resistance is less explored.

Results

Mg2+ promotes resistance to antibiotics

Magnesium is rich in marine and marine aquaculture, where antibiotics are commonly used. We directly investigate the effect of Mg²⁺ on balofloxacin (BLFX) susceptibility in two strains of Vibrios, V. alginolyticus ATCC33787 and V. parahaemolyticus VP01, which were isolated from marine aquaculture. In the initial experiment, the ionic environment of marine water was simulated by adding 39 mM NaCl, 35 mM Mg₂SO₄, 7 mM KCl, and 7 mM CaCl₂ to LB medium (Table EV1), to create a medium which we designated “artificial seawater” (ASWT). The minimum inhibitory concentrations (MIC) for BLFX in ATCC33787 and VP01 were 54 μg/mL and 25 μg/mL in ASWT and 1 μg/mL and 3 μg/mL in LB medium, respectively (Fig 1A). The role of exogenous NaCl in antibiotic resistance was investigated in a previous study (Yang et al., 2018); and the MIC for BLFX was the same in LB medium or M9 medium plus 7 mM KCl or 7 mM CaCl₂ (Fig EV1). However, the MIC for BLFX was higher in M9 medium plus Mg₂SO₄ or MgCl₂ (Fig 1B) and the MIC for BLFX increased with increasing concentration of MgCl₂. Specifically, addition of 50 mM or 200 mM MgCl₂ increased the MIC for BLFX 200- or 1600-fold, respectively (Fig 1B). Using the Oxford cup method, the zone of inhibition increased with increasing MgCl₂ from 15.6 mM – 62.4 mM (Fig 1D). Exogenous MgCl₂ also increased MICs for other quinolones (e.g. nalidixic acid, levofloxacin, ciprofloxacin, ofloxacin, and moxifloxacin (Fig 1E) and non-quinolone antibiotics including antibacterial peptides (colistin), macrolides (roxithromycin), tetracyclines (oxytetraycline), β-lactams (ceftriaxone, ceftazidime), and aminoglycosides (amikacine, kanamycin, and gentamicin) (Fig 1F). Such non- specific resistance promoting effects indicate that Mg²⁺ promotes phenotypic resistance.

Figure 1.

Extracellular Mg²⁺, intracellular Mg²⁺ and chelation of antibiotic

Previous reports suggest that chelation of antibiotics by magnesium ions inhibits antibiotic uptake (Deitchman et al, 2018; Lunestad & Goksøyr, 1990). To explore this possibility, the zone of inhibition for balofloxacin was measured with various concentrations of MgCl₂, ranging from 0 mM to 200 mM. Marine water usually contains 50 mM MgCl₂. Lower concentrations of MgCl₂, (0.78, 3.125 or 12.5 mM) did not alter the zone of inhibition, while higher concentration including 50 and 200 mM MgCl₂ decreased the zone of inhibition (Fig 2A). Correlation of the MgCl₂ concentration-dependent resistance suggests that chelation influences bacterial survival by 25.5% and 27.7% in 50 and 200 mM MgCl₂, respectively, but not at lower MgCl₂ concentrations (Fig 2B & 2C). Furthermore, intracellular balofloxacin decreased with increasing concentration of MgCl₂ (Fig 2D), while exogenous Mg²⁺ increased intracellular Mg²⁺ level in a dose-dependent manner, where 50 and 200 mM MgCl₂ increased 1.21 and 1.31 mmol/L, respectively (Fig 2E). We also examined rate of efflux and expression of LPS, which contribute to antibiotic resistance (Kobylka et al, 2020; Song et al, 2020). The results show that expression of TolC/tolC and LPS was unaffected by Mg²⁺ (Fig 2F and 2G). Loss of waaF, lpxA, and lpxC, three key genes involved in LPS biosynthesis, did not influence sensitivity/resistance to balofloxacin (Fig 2H). These data demonstrate that influx of Mg²⁺ plays roles in balofloxacin resistance.

Figure 2.

MgCl₂ affects bacterial metabolism

Although it is well established that Mg²⁺ regulates the activity of many metabolic enzymes and strongly influences bacterial metabolism(Garfinkel & Garfinkel, 1985), the mechanism of Mg²⁺ in promoting antibiotic resistance is less explored. Thus, V. alginolyticus was cultured in M9 medium in the presence of various amounts of MgCl₂ (0 mM, 0.78 mM, 3.125 mM, 12.5 mM, 50 mM, or 200 mM), analyzed by GC-MS, and identified 54 metabolites. Five biological and two technical replicates were included for each treatment (Fig EV2). Rank-sum test and a permutation test were used to identify differentially expressed metabolites. Forty-one metabolites were identified (p < 0.05) (Fig 3A), which were presented as a heatmap and Z-value in Fig EV3. Orthogonal partial least square discriminant analysis (OPLS-DA) was conducted that separated the six treatments into three groups: Group 1 included 0 mM, 0.78 mM, and 3.125 mM MgCl₂; Group 2 was a singlet of 12.5 mM; Group 3 included 50 mM and 200 mM MgCl₂. Component t[1] distinguishes Group 1 from Groups 2 and 3; component t[2] distinguishes Group 3 from Groups 1 and 2 (Fig 3B). Discriminating variables are shown as S-plot, where cut-off values are ≥ 0.05 absolute value of covariance p and 0.5 for correlation p(corr). Six biomarkers/metabolites were selected from component t[1] and p[1]. More specifically, the abundance of cadaverine, urea, palmitic acid, aminoethanol, and fumaric acid were increased, but pyroglutamic acid and glutamic acid were decreased (Fig 3C; Fig EV4). Pathway enrichment analysis suggests that twelve pathways are involved. Notably, one of the pathways is biosynthesis of unsaturated fatty acids (terminology in the software. Actually, it is biosynthesis of fatty acids) (Fig 3D). Interestingly, the abundance of palmitic acid and stearic acid were increased in an Mg²⁺ dose-dependent manner (Fig 3E; Fig EV5). Moreover, palmitic acid was a crucial biomarker (Fig 3C). Exogenous palmitic acid increased bacterial resistance to balofloxacin (Fig 3F). These results suggest that fatty acid metabolism may be critical to the resistance.

Figure 3.

Mg²⁺ regulates fatty acid biosynthesis

Fatty acids include saturated and unsaturated fatty acids whose biosynthesis shared common biosynthetic pathways from acetyl-CoA to enoyl-ACP, sequentially mediated by fabD or fabH, fabB/F, fabG and fabA/Z. Enoyl-ACP is the direct source to produce unsaturated fatty acids or is metabolized to Acyl-ACP via fabV to produce saturated fatty acids. Unsaturated fatty acids produce saturated fatty acids via tesA/B and yciF. qRT-PCR was used to quantify the expression of involved genes in bacteria treated with different concentration of MgCl₂. The expression of 18 genes was increased when treated with 50 or 200 mM MgCl₂ (Fig 4A). The increased expression of FabF was confirmed by Western blot (Fig 4B). Expression of tesA, tesB, and yciA, which convert unsaturated to saturated fatty acids, was increased (Fig 4C). The expression of fabA that promotes enoyl-ACP and biosynthesis of unsaturated fatty acids was decreased (Fig 4D), which was confirmed at protein level with Western blot (Fig 4B). Oxazole-2-amine and triclosan inhibit synthesis of fatty acids and decrease bacterial survival in the presence of balofloxacin (Fig 4E). These results indicate that Mg²⁺ inhibits synthesis of unsaturated fatty acids and promotes synthesis of saturated fatty acids.

Figure 4

We also quantified the gene expression of fatty acid degradation (Fig 4F). Interestingly, the expression of genes involved in unsaturated fatty acid degradation was decreased in an MgCl₂ dose-dependent manner (Fig 4G), while the expression of genes for exogenous fatty acid degradation was increased. Western blot data confirmed that the abundance of FadL decreased with increasing MgCl₂ (Fig 4H). FadH degrades unsaturated fatty acids, and influences the ratio of unsaturated to saturated fatty acids. In general, these results suggest that MgCl₂ inhibits degradation of fatty acids.

FabR, FadR, ArcA, and cAMP/CRP are transcriptional factors in regulating fatty acid biosynthesis. FabR inhibits biosynthesis of unsaturated fatty acids; FadR promotes biosynthesis and inhibits fatty acids degradation; ArcA and cAMP/CRP inhibit and promote fatty acid degradation, respectively (Feng & Cronan, 2010; Fujita et al, 2007). The expression of fabR and fadR was increased, arcA was decreased. The expression of N646_1004 did not change with increasing MgCl₂, while expression of N646_1885 increased only in the presence of 200 mM MgCl₂ (Fig 4I). Western blot data confirmed the decreased expression of FadR with increasing MgCl₂ concentration (Fig 4J). Viability of ΔfadR and ΔarcA was decreased in the presence of MgCl₂, balofloxacin or both in a dose-dependent manner (Fig 4K). Thus, FadR and ArcA could play roles in bacterial resistance to balofloxacin, presumably by a mechanism that is linked to the abundance of saturated/unsaturated fatty acids.

Furthermore, to investigate whether Mg²⁺ functions through FadR and ArcA, expression of 8 genes of fatty acid biosynthetic genes (accA, accC, fabD, fabH, fabR, fabG, fabV, and yciA) and 5 genes of fatty acid degradation (fadL, fadD, fadB, and fadA) were quantified in wildtype, ΔfadR and ΔarcA bacterial strains. As expected, the expression of fatty acid biosynthetic genes was decreased in ΔfadR cells (Fig 4L). However, higher expression of fadD, fadB, and fadA (elevated fadA only for ΔfadR) was detected in ΔfadR or ΔacrA strains, but the expression of these genes was negatively regulated by 200 mM MgCl₂ (Fig 4M-O). These results indicate that Mg²⁺ promotes biosynthesis of fatty acids by positively regulating fadR, and inhibits degradation of fatty acids through fadR, arcA and their downstream targets.

MgCl₂ affects 16-carbon and 18-carbon fatty acid metabolism

To investigate the effect of Mg²⁺ on biosynthesis of saturated and unsaturated fatty acids, LC-MS was used to quantify 16-carbon and 18-carbon fatty acids, the main precursors for lipid biosynthesis (Zhang & Rock, 2008b). The abundance of saturated fatty acids, palmitic acid (C16:0), was increased, while the abundance of five unsaturated fatty acids, palmitoleic acid (C16:1), linoelaidic aid (C18:2), linoleic acid (C18:2), alpha-linoleic acid (C18:3), stearidonic acid (C18:4) was decreased with increasing MgCl₂ concentration (Fig 5A). The abundance of stearic acid (C16:0) and vaccenic acid (C18:1) was unaffected. Total saturated fatty acids increased and totalunsaturated fatty acids decreased with increasing Mg²⁺ (Fig 5B). These results support the conclusion that Mg²⁺ upregulates biosynthesis of saturated fatty acids but downregulates biosynthesis of unsaturated fatty acids.

Figure 5.

Furthermore, direct exposure to palmitic acid inhibited balofloxacin-mediated killing (Fig 3F) but linolenic acid promoted balofloxacin-mediated killing in a dose-dependent manner (Fig 5C). Increasing exogenous palmitic acid and linolenic acid increased intracellular palmitic acid and linolenic acid, respectively (Fig 5D). When cells were co-treated with palmitic acid, linolenic acid, and balofloxacin, the effects of the two fatty acids appeared to be antagonistic as demonstrated by Bliss model (Fig 5E & 5F). These results indicate that saturated and unsaturated fatty acids have antagonistic effects on balofloxacin resistance.

Mg²⁺ promotes phospholipid biosynthesis

Fatty acids are biosynthetic precursors of lysophosphatidic acid (LPA) and phosphatidic acid (PA), two key components of the cell membrane (Zhang & Rock, 2008a). Thus, LC-MS was used to profile the effect of Mg²⁺ on membrane lipid composition (Fig EV6A). Increasing concentration of Mg²⁺ increased the percentage of lipids increased from 61% to 67%, and saturated fatty acids from 24% to 26%, but decreased the percentage of unsaturated fatty acids from 15% to 8% (Fig EV6B). The abundance of 11 lipids, 32 lipids, and 53 were increased and of 26 lipids, 52 lipids, and 107 lipids were decreased in 3.125 mM, 50 mM, and 200 mM MgCl₂, respectively (Fig 6C). Saturated fatty acid and lipid increased and unsaturated fatty acid decreased in an Mg²⁺ dose-dependent manner (Fig EV6D). A total of 52 lipids were quantified, including 17 high abundance lipids (Fig EV6E). Phosphatidylethanolamine (PE) and phosphatidylglycerol (PG) were the first and second highest lipids in abundance (Fig EV6E) and PE decreased while PG increased with increasing Mg²⁺ (Figs 6A, 6B). Similar effect was observed for unsaturated derivatives of PE and PG (Fig 6C). Principal component analysis (PCA) showed that component t [1] differentiates 0 and 3.125 mM Mg²⁺ from 50 and 200 mM Mg²⁺, while component t [2] separates 50 mM Mg²⁺ from 0 and 200 mM Mg²⁺ and variants in 3.125 mM of Mg²⁺ treatment (Fig 6D).

Figure 6

S-plot analysis showed that upregulated lipids include behenic acid, PG[16:1(9Z)/16:1(9Z)], PG[18:2 (9Z, 12Z)/16:0], PA [14:0/22:6 (4Z, 7Z, 10Z, 13Z, 16Z, 19Z)], lysoPE(16:0/0:0) and downregulated lipids include linoelaidic acid, palmitoleic acid, PE-NMe(24:0/24:0), PE[16:0/14:1(9Z)], lysoPA(i-12:0/0:0) (Fig 6E). Among them, the abundance of PG[16:1(9Z)/16:1(9Z)], PG[18:2(9Z,12Z)/16:0], and behenic acid was increased, while that of linoelaidic acid, palmitoleic acid and PE[16:0/14:1(9Z)] were decreased with increasing Mg²⁺ (Fig 6F). Therefore, altered phospholipid abundance may play a role in the effect of Mg²⁺ on resistance to balofloxacin.

PE and PG are the two end products of phospholipid metabolism (Fig 6G). Expression of most genes in the phospholipid biosynthetic pathway, including psd, pldB, and glpQ, was increased in the presence of 200 mM MgCl₂, but not pssA and etuB, etuC, encoding the first enzyme and the last enzyme in the pathway, respectively. In addition, the expression of gpsA, encoding GpsA that transforms sn-glycerol-3P to glycerone-P, was increased in an independent of Mg²⁺ dose (Fig 6H). Phosphatidylglycerol phosphate synthase (PGS) encoded by pgsA and phosphatidylserine synthase (PSS) encoded by pssA are crucial enzymes for biosynthesis of PG and PE, respectively. Activities of PGS and PSS were elevated and reduced, respectively, by the increasing concentration of Mg²⁺ (Fig 6I). The effect of Mg²⁺ on the activity of PSS was confirmed with recombinant PSS (Fig 6J). The deletion of plsB and pgpA, the first and last genes for PG biosynthesis, respectively, lowered cell viability in the presence of balofloxacin (we failed to obtain Δpsd) (Figs 6K-I). These results indicate that phospholipids may play a role in the effect of Mg²⁺ on resistance to balofloxacin; more specifically, PE may inhibit and PG may promote resistance to balofloxacin.

Mg²⁺ regulates membrane polarization, permeability and fluidity to confer balofloxacin resistance

Membrane potential and membrane polarization influence membrane permeability and uptake of antibiotics (Lee et al., 2019; Peng et al., 2015). Thus, we speculated that exogenous Mg²⁺ reduces membrane polarization. Here, we show that 12.5-200 mM MgCl₂ promotes membrane depolarization in a dose-dependent manner (Fig 7A). The effect of Mg²⁺ on proton motive force (PMF) displayed MgCl₂ dose- and time- dependent manner (Fig 7B). Specifically, membrane depolarization decreased over time without MgCl₂. However, depolarization gradually increased for 2 h, reached a plateau for 3 - 5 h and decreased from 5 to 10 h when supplemented with 0.78 or 3.125 mM MgCl₂. The extent of maximum depolarization and the length of time at higher level of depolarization also increased by MgCl₂ (Figs 7C and 7D). These results suggest that Mg²⁺ -dependent changes in membrane depolarization may influence bacterial resistance to antibiotics.

Figure 7.

Moreover, efficacy of antibiotics is strongly influenced by bacterial membrane permeability and fluidity (Saeloh et al, 2018; Zhao et al., 2021). Hence, membrane permeability and fluidity were investigated with fluorescence microscopy to visualize V. alginolyticus cells in the presence of FM5-95 (red dye) and various concentrations of MgCl₂. Mg²⁺ decreased red staining, and no staining was observed in the presence of 200 mM Mg²⁺ (Fig 7D). Consistently, the uptake of SYTO9 dye decreased with increasing MgCl₂, indicating bacterial membrane permeability decreased in an Mg²⁺dose-dependent manner (Fig 7E). These results indicate that exogenous Mg²⁺ decreased bacterial membrane permeability.

Furthermore, exogenous palmitic acid shifted the fluorescence signal peaks to the left in an MgCl₂- and palmitic acid concentration-dependent manner (Fig 7F). In contrast, exogenous linolenic acid shifted the peak to the right in a dose-dependent manner (Fig 7G). These results support the conclusion that exogenous palmitic acid and linolenic acid decrease or increase membrane permeability, respectively. In ΔfadR cells, membrane permeability increased and then slowly decreased with increasing Mg²⁺, remaining higher than the control (Fig 7H). However, ΔarcA cells displayed reduced membrane permeability in the presence of low but not high MgCl₂ (Fig 7I). ΔpgpA cells exhibited higher membrane permeability than control cells at all MgCl₂ concentrations, but the effect was smaller at higher MgCl₂ (Fig 7J). These data are consistent with the viability of these mutants as described above (Figs 4L and 6L).

Relative membrane permeability appeared to correlate with relative intracellular balofloxacin and antibiotic efficacy. Exogenous palmitic acid and linolenic acid reduced and promoted uptake of balofloxacin 159- and 18-fold, respectively (Fig 7K). Loss of fadR and pgpA increased intracellular balofloxacin (Fig 7K). These data suggest that bacterial membrane permeability is a critical factor in the efficacy and uptake of balofloxacin, in the presence or absence of exogenous MgCl₂.

Discussion

The present study explores the effect of Mg²⁺ on phenotypic in V. alginolyticus. The results suggest that exogenous Mg²⁺ promotes biosynthesis of saturated fatty acids and inhibits biosynthesis of unsaturated fatty acids, while it also upregulates the activity of PGS and the abundance of PG, and downregulates the activity of PSS and abundance of PE. As a result, membrane permeability and antibiotic uptake decrease (Fig 7L), enabling phenotypic resistance. The results are consistent with the observed impact of exogenous Mg²⁺ on bacterial survival in the presence of balofloxacin and other antibiotics, and it is well recognized that membrane permeability critically impacts drug efficacy. Another pathway that influences drug uptake involves glutamine-inosine transport mediated by CpxA/R-OmpF (Zhao et al., 2021). The results also show that exogenous fatty acids influence membrane permeability and bacterial survival in the presence of balofloxacin.

Mg²⁺ is the most abundant divalent cation in the cell (Pohland & Schneider, 2019; Pontes et al, 2015) and it plays many essential roles. For example, Mg²⁺ stabilizes macromolecular complexes and membranes, binds cytoplasmic nucleic acids and nucleotides, interacts with phospholipid head groups and cell surface molecules, and is an essential cofactor in many enzymatic reactions (Groisman et al, 2013). A novel observation of the present study is that exogenous Mg²⁺ influences the abundance of palmitic acid and linolenic acid and that it upregulates PGS and downregulates PSS. These findings extend our understanding of the complexity of the biological functions of the divalent magnesium cation.

Recent studies suggest a link between fatty acid biosynthesis and antibiotic resistance in E. tarda (Su et al, 2021). The present study shows that Mg²⁺ has opposite effects on the abundance of saturated and unsaturated fatty acids, stimulating biosynthesis of saturated fatty acids at moderately high levels, while inhibiting biosynthesis of unsaturated fatty acids at ≥ 3 mM Mg²⁺. Therefore, ratio of saturated and unsaturated fatty acids should be a key clue to understand antibiotic resistance.

In previous studies of gram-positive and gran-negative bacteria (Kumariya et al, 2015; Said et al, 1987), it was shown that 10 - 20 mM Mg²⁺ disrupts Staphylococcus aureus membranes and kills stationary-phase S. aureus cells, but it does influence survival of Escherichia coli and Bacillus subtilis (Xie & Yang, 2016). Low concentrations of Mg²⁺ (≤ 10 mM) induce PmrAB-dependent modification of lipid A in wild-type E. coli (Herrera et al, 2010). A putative zwitterionic amino-containing phospholipid increased significantly, whereas amounts of phosphatidylglycerol and cardiolipin decreased in two mutants of Enterococcus faecium resistant to mundticin KS (Sakayori et al, 2003). However, Mg²⁺ induces antibiotic resistance through regulating the activity of PGS and PSS is not reported. More importantly, the regulation targets the two enzymes in a reverse manner to enlarge the change in the ratio between PE and PG for recognition of antibiotic resistance, which is a previously unknown mechanism mediated by Mg²⁺. Previous studies on membrane permeability also show that PhoP/Q is activated by low magnesium, and that inhibition or inactivation of PhoP/Q stimulates uptake of β-lactam antibiotics in Stenotrophomonas maltophilia (Huang et al, 2021). The effects of the outer membrane permeabilizers, polymyxin B nonapeptide and EDTA, are completely abolished by 3 mM Mg²⁺ (Kwon & Lu, 2006). In response to Mg²⁺ limited growth, enteric Gram-negative bacteria show higher acylation of lipid A, which alters membrane permeability and reduces uptake of cationic antimicrobial peptides (Guo et al, 1998). Mg²⁺ reverts the high sensitivity of rough mutants of Salmonella typhimurium, S. minnesota, and Escherichia coli 08 (i.e. with defects in the carbohydrate core of the lipopolysaccharide) to several antibiotics (Stan-Lotter et al, 1979). Mg²⁺ (1 mM) totally inhibited aminoglycoside-mediated outer membrane permeabilization in Pseudomonas aeruginosa (Hancock et al, 1981). However, mechanisms underlying Mg²⁺ regulation to antibiotic resistance are largely unknown. The present study reveals that Mg²⁺ plays the crucial role in the membrane permeability responsible for antibiotic resistance via modulating PE and PG in phospholipid metabolism.

In summary, the present study shows that Mg²⁺ modulates the phenotypic resistance of V. alginolyticus to balofloxacin and potentially other quinolones and other classes of antibiotics by a mechanism that involves altered membrane permeability and reduced antibiotic uptake. These findings inform the ongoing and future search for ways to improve the therapeutic effects of antibiotics to magnetism-induced phenotypic resistance.

Materials and Methods

Bacterial strains and culture

V. alginolyticus ATCC33787 and V. parahaemolyticus 01 are from our laboratory collection (Jiang et al, 2022; Kou et al, 2022). They were cultured in 0.5% yeast broth (HuanKai Microbial, Guangdong, China) (pH 7.0) with 3% NaCl at 30 °C overnight. The cultures were diluted 1:50 (v/v) and grown in fresh 0.5% yeast broth supplemented with desired concentrations of MgCl₂ at 30 °C until OD600 = 0.6. Bacteria were harvested by centrifugation at 8,000 rpm for 3 min and resuspended in corresponding concentrations of MgCl₂ to 0.6 of OD600.

Measurement of the minimum inhibitory concentration (MIC) by microtitre- dilution-method

MIC using microtitre-dilution-method was performed as previously described (Zhang et al, 2019). In brief, the overnight bacterial cultures in 3% NaCl were diluted at 1:100 (v/v) in fresh 0.5% yeast broth supplemented with desired concentrations of MgCl₂, cultured at 30 °C and collected when the cells arrived at an OD600 of 0.6 in medium without MgCl₂. Then, 1 X 10⁵ CFU cells were dispensed into each well of a 96-well microtiter polystyrene tray after which a series of 2-fold dilutions of antibiotic was added. Following incubation at 30 °C for 16 h, the MIC was defined as the lowest antibiotic concentration that inhibited visible growth. Three biological repeats were carried out.

Measurement of MIC by Oxford method

Determination of MIC by Oxford cup was performed as previously described (Liu et al, 2015). Bacterial cells cultured overnight in yeast medium were diluted at 1:100, shaken at 30 ⁰C and 200 rpm until the OD600 nm reached at 0.6. Aliquot of 100 μL cells were spread on 0.5% yeast solid medium containing 3% NaCl and 0, 0.625, 1.25, 2.5, 5, 10, 20, or 40 mM of MgCl₂. Oxford cups were placed on the solid medium, and various amounts of balofloxacin (0, 0.39, 0.78, 1.56, 3.125, 6.25, and 12.5 μg) were added. After culturing at 30 °C for 12 h, diameter of the inhibition zone was measured and the inhibition area was calculated.

Measurement of intracellular balofloxacin

Measurement of intracellular balofloxacin was performed by LC-MS analysis and plate-counting assay. For LC-MS analysis, ATCC33787 cultured in the desired concentrations of MgCl₂ were collected and adjusted to OD 0.6. Aliquot of 50 mL bacterial cells were added into a 250 mL Erlenmeyer flask and then 300 μL of 10 mg/mL balofloxacin (final concentration is 60 μg/mL). After being cultured for 6 h at 30 °C, these bacterial cells were collected and adjusted to OD 1.0 with 3% NaCl containing corresponding concentrations of MgCl₂. Aliquot of 30 mL bacteria were collected and washed three times using mobile phase (acetonitrile: double distilled water containing 0.1mol/l formic acid = 35:65) for LC-MS detection. The bacterial cells were added 1 mL mobile phase, crushed in ice water bath for 10 min (crush for 2S, pause for 3S at 35% power). Following by centrifugation, supernatants were collected, filtered, and then measured by using liquid chromatography for balofloxacin. Different concentrations of balofloxacin were used for a standard curve. For plate-counting assay, ATCC33787 cells cultured overnight were diluted at 1:100, shaken at 30 ⁰C and 200 rpm until the OD600 nm reached at 0.6. The cells were diluted at 1:100 and spread on LB agar to preparing ATCC33787 plates. Meanwhile, ATCC33787 cells were cultured in medium with desired MgCl₂ and at 30 °C and 200 rpm for 6 h. These cells were collected, washed and sonicated. The sonicated solution was added to Oxford cups with the ATCC33787 plates. After culturing at 30 °C for 12 h, diameter of the inhibition zone was measured. A gradient of balofloxacin was used as a standard curve for drug quantification.

Measurement of intracellular magnesium

Measurements of intracellular Mg²⁺ concentration were carried out as previously described with a modification (Yang et al., 2018). In brief, V. alginolyticus cultured in 0%, 0.78%, 3.125%, 12.5%, 50%, and 200% MgCl₂ were collected, washed, and resuspended in the same concentrations of MgCl₂ to an OD600 of 1.0. Aliquots of 20 mL of bacterial suspensions were centrifuged and washed by ddH2O once. The resulting cells were weighted, which was designated wet weight. These samples were freeze-dried overnight, which was designated dry weight. Intracellular water volume was calculated using the following formula: W × (1-W/D-0.23) as described previously (Unemoto et al, 1973). Two hundred microliters of concentrated nitric acid was added to the dried cells and then heated in 75 °C for 20 min. After diluting the samples 20- fold, they were analyzed for Mg²⁺ concentration by inductively coupled plasma mass spectrometry (ICP-MS) (iCAP 6500, Thermo Fisher). The intracellular Mg²⁺ concentration was calculated according to the volume of intracellular water.

Western blotting

Western blotting was carried out as described previously (Yao et al, 2019). Bacterial protein samples were prepared by ultrasound treatment, resolved on a 12% SDS-PAGE and transferred to nitrocellulose membranes (GE Healthcare Life Sciences). The membranes were incubated with 1:100 of the primary mouse antibodies, followed by goat anti-mouse secondary antibodies conjugated with horseradish peroxidase. Band intensities were detected by using a chemiluminescence imaging analysis system, Tanon-5200.

Metabolomics analysis

Metabolomics analysis was performed by GC-MS as described previously (Yang et al., 2018). Briefly, ATCC33787 were cultured in 0.5% yeast broth with desired MgCl₂. Equivalent numbers of cells were quenched with 60% (v/v) cold methanol (Sigma) and then centrifuged at 8,000 rpm at 4 °C for 5 min. One milliliter of cold methanol was used to extract metabolites. To do this, the samples were sonicated for 5 min at a 10- Wpower setting using the Ultrasonic Processor (JY92-IIDN, Scientz, China), followed by centrifugation at 12,000 rpm in 4 °C for 10 min. Supernatants were collected and 10 µL ribitol (0.1 mg per mL, Sigma-Aldrich, USA) was added into each sample as an internal quantitative standard. The supernatants were concentrated for metabolite derivatization and then used for GC-MS analysis. Every experiment was repeated by five biological replicates. GC-MS detection and spectral processing for GC-MS were carried out using the Agilent 7890A GC equipped with an Agilent 5975C VL MSD detector (Agilent Technologies, USA) as described previously (Unemoto et al., 1973; Yang et al., 2018). Statistical difference was obtained by Kruskal Wallis test and Manne Whitney test using SPSS 13.0 and a p value < 0.01 was considered significant.

Hierarchical clustering was completed in the R platform (https://cran.r-project.org/) with the function “heatmap. 2” of “gplots library”. Z score analysis was used to scale each metabolite. Multivariate statistical analysis included principal component analysis (PCA) and orthogonal partial least square discriminant analysis (OPLS-DA) implemented with SIMCA 12.0 (Umetrics, Umeå, Sweden). Control scaling was selected prior to fitting. All variables were mean centered and scaled to pareto variance of each variable. PCA was used to reduce the high dimension of the data set. Differential metabolites to their respective biochemical pathways were outlined in the MetaboAnalyst 3.0 (http://www.metaboanalyst.ca/). Pathways were enriched by raw p value < 0.05.

qRT-PCR

Quantitative real time polymerase chain reaction (qRT-PCR) was carried out as described previously (Yang et al, 2020). Total RNA was extracted from V. alginolyticus using TRIZOL regent (Invitrogen Life Technologies) according to the protocol. RNA Electrophoresis was carried out in 1% (w/v) agarose gels to identify quality of the extracted RNA. By using a PrimeScriptTM RT reagent Kit with gDNA eraser (Takara, Japan), reverse transcription-PCR was carried out on 1 μg of total RNA and primers are listed in Table EV2. qRT-PCR was performed in 384-well plates with a total volume of 10 μL and the reaction mixtures were run on a LightCycler 480 system (Roche, Germany). Data are shown as the relative mRNA expression compared to 0% MgCl₂ test with the endogenous reference 16S rRNA gene.

Antibiotic bactericidal assay

Antibiotic bactericidal assay was performed as described previously with a modification (Li et al, 2016; Zhao et al., 2021). The cultured bacteria of ATCC33787 and its mutant strains were transferred to fresh yeast medium at a dilution of 1:1000 and dispensed into test tubes and then the indicated concentrations of MgCl₂ and balofloxacin were added. If desired, 2 mM 2-aminooxazole and 1 μg/mL triclosan were or a metabolite complemented. These mixtures were incubated at 30 °C and 200 rpm for 6 h. Cells were collected. To determine CFU per mL, 100-μL samples were 10-fold serially diluted with 900 μL M9 buffer and an aliquot of 5 μL of each dilution was spotted onto the LB agar plates and cultured at 30 °C for 8 h.

Construction of gene-deleted mutants

Construction of gene-deleted mutants was carried out as described previously (Kuang et al, 2021). Primers were designed as shown in Table EV3 using CE Design V1.03 software. To construct gene-deleted mutants, upstream and downstream 500-bp fragments were amplified from the genome using two pairs of primers (primers P1 and P2, primers P3 and P4), and then merged into a 1,000-bp fragment by overlap PCR using a pair of primers (primers P1 and P4). After the fragments were digested by XbaI, they were ligated into the pDS132 vector digested by the same enzymes, transformed into MC1061 competent cells. Conversion products were coated on LB plate containing 25 μg/mL chloramphenicol and cultured at 37 ℃ overnight. The plasmids from colony growing on the plate were identified by PCR using a pair of primers (primers P1 and P4) and sequenced. The sequenced plasmids were transformed into MFD λ pir competent cells as donor. MFD λpir and recipient bacterium ATCC 33787 were cultured to an optical density (OD) of 1.0 and then mixed at a ratio of 4:1. After centrifugation, the pellets were resuspended with 50 μL LB medium including DAP (100 μg/mL), dropped onto sterilized filter paper on LB medium including DAP (100 μg/mL), and cultured for 16-18 h at 37 °C.

All bacteria rinsed from the filter paper with LB medium were smeared onto the LB plate with chloramphenicol (25 μg/mL) and Ampicillin (100 μg/mL). After cultured for 16-18 h at 37 °C, bacteria were screened by LB plate with above two antibiotics. The bacteria were identified by plasmid PCR using a pair of primers P1 and P4 and sequencing and then were cultured and smeared onto the LB plates with 20% sucrose. The clones were cultured and smeared onto the LB plates with 20% sucrose again. The clones, which did not grow on the LB plates with chloramphenicol but grew on the LB plates with 20% sucrose, were identified by PCR using primer P7P8 (primer P7 is set at about 250 bp upstream of P1, and primer P8 is set at about 250 bp downstream of P4), P4P7, and P5P6 (for amplification of the target gene).

LC-MS analysis for lipidomics

ATCC33787 were cultured in medium with desired concentrations of MgCl₂, washed by 3% NaCl with the desired concentrations of MgCl₂, and adjusted to OD 1.0. Aliquot of 30 mL the cultures was centrifuged each sample and bacterial cells were collected. Aliquot of 800 μL distilled water was added and treated for 5 min at 100 °C to inactivate phospholipase C. Protein concentration was determined with BCA kit. Sample solution with 3 mg protein was transferred into a 10 mL centrifuge tube, and supplemented with distilled water to 800 μL. Then all samples were processed as follows: 1 mL chloroform and 2 mL methanol (to make the volume ratio of chloroform: methanol: water = 1:2 : 0.8) were added and cholic acid was as an internal standard; These mixtures were vortexed for 2 min and then 1 mL chloroform was add and vortex for 30s; 1 mL 10% NaCl solution was add, vortexed for 30s, and placed at room temperature overnight. After the solution was layered, a 2.5 mL syringe was used to suck the lowest layer (chloroform layer) to a new EP tube. Solution of the chloroform layer was evaporated in the rotary evaporator for 2 h and then 1 mL mobile phase (50% and 50%B. A, 25 mM ammonium acetate / methanol (30 : 70); B, methanol) was added for analysis of lipids by LC-MS. Database (https://hmdb.ca) was used for lipid identification.

Measurement of enzyme activity

Enzyme activity was carried out as previously described (Jiang et al, 2020). Cells cultured in medium were collected and washed three times with saline. The bacterial cells were suspended in Tris–HCl (pH 7.4) and disrupted by sonic oscillation for 6 min (200 W total power with 35% output, 2 s pulse, 3 s pause over ice). After centrifugation, supernatants were collected. The protein concentration of the supernatant was determined using Bradford assay (Beyotime, P0009). Then, 200 μg proteins were used for determination of pyruvate dehydrogenase (PDH), α-ketoglutarate dehydrogenase (KGDH) and succinate dehydrogenase (SDH) activity.

Measurement of proton motive force

Measurement of the transmembrane voltage delta PSI (δPSI), which is the electrical component of the PMF, was performed as previously described (Cheng et al, 2018). Bacteria cultured in medium with desired concentrations of MgCl₂ were collected at centrifugation and labeled by DiO2(3). Approximately 1 × 10⁷ CFU were added to a flow cytometry tube containing 1 mL buffer with 10 μM DiOC2(3) (Sigma) and incubated in the dark for 30 min at 30 ⁰C. Samples were assayed with BD FACSCalibur flow cytometer with a 488 nm excitation wavelength. Gates for bacterial populations were based on the control population by using forward versus side scatter and red versus green emission. Size and membrane potential determined the intensity of Red (488 nm excitation, 610 nm emission) fluorescence. The diverse ratios of red and green indicated fluorescence intensity values of the gated populations. Computational formula of membrane potential:

Measurement of depolarization

The membrane potential was estimated by measuring the fluorescence of the potential- dependent probe DiBAC4(3) (Saint-Ruf et al, 2016). ATCC33787 were cultured in medium with desired concentrations of MgCl₂, collected and then adjusted to OD 0.6. Aliquot of 100 μL bacterial cells were diluted to 1 mL, and 2 μL of 5 mM DiBAC4(3) was added. After incubated for 15 min at 30 ℃ without light and vibration, these samples were filtered and detected by BD FACSCalibur.

Measurement of fluidity by fluorescence microscopy

ATCC33787 were cultured in medium with desired concentrations of MgCl₂, collected and then adjusted to OD 0.6. Aliquot of 100 μL bacteria cells of each sample were diluted to 1 mL and 10 μL (10 mg/mL) FM5-95 was added. After incubated for 20 min at 30 ℃ at vibration without light, the sample was centrifuged for 10 min at 12,000 rpm. The pellets were resuspended with 20 μL of 3% NaCI. Aliquot of 2 μL sample was dropped on the agarose slide, and take photos under the inverted fluorescence microscope.

Measurement of membrane permeability

Measurement of membrane permeability was carried out as described previously (Su et al., 2021). ATCC33787 were cultured in medium with desired concentrations of MgCl₂, collected and then adjusted to OD 0.6. Aliquot of 100 μL bacteria cells of each sample were diluted to 1 mL and 2 μL 10 mg/mL SYT09 was added. After incubated for 15 min at 30 ℃ at vibration without light, the mixtures were filtered and measured by BD FACSCalibur.

Acknowledgements

This work is sponsored by International Cooperation and Exchange program of National Natural Science Foundation of China (32273177) (to P.B.), Innovation Group Project of Southern Marine Science and Engineering Guangdong Laboratory (Zhuhai) 311020006 (to P.B.), Natural Science Foundation of Guangdong Province 2022A1515012079 (L. H.) and the Science and Technology Planning Project of Guangdong Province (No. 2023B1212060028).

Author Contributions

• Peng conceived the idea, designed the experiments and supervise the projects. H. Li and J. Yang, S.F. Kuang conducted the experiments, and interpreted the results.

Declaration of interest statement

Authors declare that they have no competing interests.