Stimulation of Vibrio vulnificus Pyruvate Kinase in the Presence of Glucose to Cope With H2O2 Stress Generated by Its Competitors

The bacterial phosphoenolpyruvate (PEP):carbohydrate phosphotransferase system (PTS) regulates a variety of cellular processes in addition to catalyzing the coupled transport and phosphorylation of carbohydrates. We recently reported that, in the presence of glucose, HPr of the PTS is dephosphorylated and interacts with pyruvate kinase A (PykA) catalyzing the conversion of PEP to pyruvate in Vibrio vulnificus. Here, we show that this interaction enables V. vulnificus to survive H2O2 stress by increasing pyruvate production. A pykA deletion mutant was more susceptible to H2O2 stress than wild-type V. vulnificus without any decrease in the expression level of catalase, and this sensitivity was rescued by the addition of pyruvate. The H2O2 sensitivity difference between wild-type and pykA mutant strains becomes more apparent in the presence of glucose. Fungi isolated from the natural habitat of V. vulnificus retarded the growth of the pykA mutant more severely than the wild-type strain in the presence of glucose by glucose oxidase-dependent generation of H2O2. These data suggest that V. vulnificus has evolved to resist the killing action of its fungal competitors by increasing pyruvate production in the presence of glucose.


INTRODUCTION
Microorganisms usually have complex ecological interactions with many different species in natural environments. To ensure their survival and prosperity in a wide variety of ecological conditions, microbes need to decide whether to cooperate or compete with other species for limited nutritional resources and defend themselves from potential competitors and predators. Therefore, they possess multiple regulatory systems to sense and adapt to the constantly changing environment. One bacterial sensory system that plays an important part in monitoring nutritional states and modulating rapid physiological adjustment to environmental changes is the phosphoenolpyruvate (PEP):carbohydrate phosphotransferase system (PTS) (Deutscher et al., 2006). This system uses PEP as the energy source for the concomitant transport and phosphorylation of its carbohydrate substrates in a process termed group translocation (Kundig et al., 1964;Roseman, 1969).
The PTS is composed of two general proteins, enzyme I (EI) and HPr, which are used for most PTS carbohydrates, and the carbohydrate-specific permeases, commonly referred to as the enzyme II (EII) complexes (Meadow et al., 1990;Postma et al., 1993). In the presence of a PTS carbohydrate, a phosphoryl group is sequentially transferred from PEP to EI, HPr, the EII complex and finally to the incoming carbohydrate. Therefore, phosphorylation of the PTS components increases in the absence and decreases in the presence of a PTS carbohydrate substrate such as glucose. In this way, the PTS carbohydrate transport system monitors nutritional changes in the environment and regulates a variety of metabolic processes through the phosphorylation state-dependent protein-protein interactions of the components involved (Deutscher et al., 2014;Park et al., 2015).
We recently reported that, in the presence of glucose, HPr of the PTS is dephosphorylated and interacts with pyruvate kinase A (PykA), catalyzing the conversion of PEP to pyruvate in Vibrio vulnificus (Kim et al., 2015). In V. vulnificus, there are two pyruvate kinase isozymes, PykF and PykA, and PykF is the major pyruvate kinase. Only PykA, but not PykF interacted with HPr in V. vulnificus. While we could not make a pykF deletion mutant even after repeated trials, pykA was dispensable under normal growth conditions in V. vulnificus (Kim et al., 2015). Interestingly, although PykA and HPr of Escherichia coli, belonging to the same γ-proteobacterial group as V. vulnificus, share 71 and 76% amino acid sequence identities, respectively, with their orthologs in V. vulnificus, the regulatory interaction of HPr with PykA is not observed in E. coli. This observation prompted the question of why the two very close species should have developed different regulatory mechanisms for the same enzyme. The primary habitat of E. coli is the lower intestine of warm-blooded animals (Whittam, 1989), whereas V. vulnificus is usually present in coastal marine environments (Bhadury et al., 2011;Horseman and Surani, 2011). Antagonistic interactions between bacteria and fungi in competing for a common substrate such as glucose have been documented in many habitats including an aquatic environment (Mille-Lindblom et al., 2006;Arvanitis and Mylonakis, 2015). Here, we show that fungi isolated from the natural habitat of V. vulnificus retarded the growth of the pykA mutant more severely than the wild-type strain in the presence of glucose by glucose oxidase-dependent generation of H 2 O 2 . Interestingly, the HPr-PykA interaction enables V. vulnificus to survive H 2 O 2 stress by increasing pyruvate production in the presence of glucose. These data suggest that V. vulnificus has evolved to resist the killing action of its competitors by increasing pyruvate production in the presence of glucose.

Growth Conditions
Vibrio vulnificus strains were cultured in Luria-Bertani medium containing 2.5% NaCl (LBS) or M9 minimal medium containing 0.2% casamino acids and 2.5% NaCl (M9S) at 30 • C. All E. coli strains were grown in LB medium at 37 • C. Fungal strains were cultured in potato dextrose agar (PDA) plates or M9S medium at 30 • C. Details of strain and plasmid constructions are provided in the Table 1.

Induction of VBNC State
Vibrio vulnificus cells in late log phase were harvested by centrifugation at 10,000 × g for 10 min and washed twice with artificial sea water [ASW; 60 mM NaCl, 20 mM MgSO 4 ·7H 2 O, 20 mM KCl, 2 mM CaCl 2 ·2H 2 O, 50 mM Tris-HCl (pH 8.0)]. Cells were diluted with ASW to a cell density of 10 8 CFU/ml, transferred to sterile microcentrifuge tubes, and incubated at 4 • C in the dark without shaking. Then the total cell number, viability, and culturability were assessed from each tube 0, 2, 4,

Detection of Catalase by Native Gel Electrophoresis
Vibrio vulnificus strains were cultured in LBS medium with or without 0.8 mM H 2 O 2 . At the exponential growth phase (OD 600 ∼ 0.5), cells were collected by centrifugation and the cell pellet washed with 50 mM potassium phosphate buffer (pH 7.0). Cells were then disrupted by passing twice through a French pressure cell at 10,000 psi and centrifuged at 5,000 × g for 20 min at 4 • C. The amount of protein in a cell lysate was determined by the Bradford assay by using bovine serum albumin as the standard. After separation on a 10% non-denaturing poly-acrylamide gel, the locations of catalase were visualized by staining the gel with a solution containing 2% K 3 Fe(CN) 6 and 2% FeCl 3 (Wayne and Diaz, 1986).

Determination of Glucose Oxidase Activities
Aspergillus strains were cultured in M9S media containing 0.2% glucose for 2 days. The fungal cells were then harvested by filtration and the cell pellet was washed twice with sodium acetate buffer (pH 5.5). Cells were then disrupted by sonication and passed twice through a French pressure cell at 10,000 psi, and centrifuged at 5,000 × g for 20 min at 4 • C. The amount of protein in a cell lysate was determined by the Bradford assay by using bovine serum albumin as the standard. For in-gel glucose oxidase assays, proteins in the cell extracts were separated on a 16.5% non-denaturing poly-acrylamide gel containing entrapped peroxidase [5 ml separating gel solution containing 19.2 purpurogallin units of horseradish peroxidase (HRP, Sigma-Aldrich)], and the locations of glucose oxidase were visualized by staining the gel with a solution containing 30 mM glucose and 17 mM ortho-phenylenediamine (Mateescu et al., 2012). We also determined glucose oxidase activities from fungal cell extracts at 30 • C spectrophotometrically by monitoring the change in A 460 due to oxidation of o-dianisidine by HRP, using 8.3 as the molar extinction coefficient (Kelley and Reddy, 1986). The reaction mixture consisted of sodium acetate buffer (50 mM, pH 5.5), 50 µM o-dianisidine, 1.6% glucose, 6 units of HRP, and 250 µl of crude extracts in a total volume of 1 ml. The purified glucose oxidase of A. niger (Sigma-Aldrich) was used as a positive control. The reaction mixture was bubbled with oxygen for 5 min before addition of glucose oxidase.

Detection of H 2 O 2 Level
The H 2 O 2 levels were determined using the hydrogen peroxide assay kit (BioVision), according to the manufacturer's instructions. Briefly, control media or culture filtrates of fungal strains were mixed with a 50 µl reaction mixture containing assay buffer, OxiRed probe solution, and HRP solution, and incubated at room temperature for 10 min. Concentrations of H 2 O 2 were determined with a microplate reader at 570 nm.

Vibrio vulnificus PykA Activity Confers Resistance to H 2 O 2
It is known that, like many other Gram-negative bacteria, V. vulnificus is induced into a viable but non-culturable (VBNC) state by incubation at low temperatures (Oliver, 2010). Several studies have provided evidence for the involvement of reactive oxygen species (ROS) in the VBNC state of V. vulnificus by showing that a significant portion of the VBNC population of V. vulnificus can be resuscitated if a H 2 O 2 -scavenging agent such as catalase or pyruvate is introduced to the culture medium (Bogosian et al., 2000;Kong et al., 2004). While catalase decomposes H 2 O 2 into oxygen and water through its enzymatic activity (Loew, 1900), pyruvate reacts non-enzymatically with H 2 O 2 to yield carbon dioxide, water, and acetic acid (Bunton, 1949). Because pyruvate is a product of the reaction catalyzed by pyruvate kinases, we assumed that the PykA activity could be related to an adaptive defense against H 2 O 2 stress in V. vulnificus.
To test whether there were any differences in the development of the VBNC state between the wild-type and pykA mutant strain, we determined the viability and culturability of each strain at low temperature (Figure 1). The VBNC state was induced in the artificial sea water by incubation at 4 • C. As expected, the pykA mutant strain lost the ability to form colonies at a somewhat faster rate than the wild-type strain ( Figure 1B). These results suggested that the development of the VBNC state of V. vulnificus could be influenced by the activity of PykA. , viable cell counts of the wild-type V. vulnificus CMCP6 culture; , viable cell counts of the pykA mutant culture; , plate counts of the wild-type V. vulnificus CMCP6 culture; , plate counts of the pykA mutant culture. The results are presented as the mean ± standard deviations of three independent measurements. Statistical significance for the VBNC cells on Day 11 was determined by Student's t-test ( * P < 0.05, n = 3).
To study whether the effect of PykA on the VBNC state is related to the synthesis of pyruvate, a H 2 O 2 -scavenging agent, we determined the growth curves of wild-type and pykA mutant strain in the presence of 0.8 mM H 2 O 2 . While the two strains did not show a significant difference in growth in LB supplemented with 2.5% NaCl (LBS medium) (Supplementary Figure S1), the pykA mutant exhibited a significantly retarded growth compared to the wild-type in the presence of 0.8 mM H 2 O 2 (Figure 2A). The episomal expression of PykA from pRK-PKA could rescue the H 2 O 2 sensitivity of the pykA mutant, indicating that the increased sensitivity of the mutant to H 2 O 2 was a direct reflection of the loss of PykA.
The other H 2 O 2 scavenging agent, catalase, is a common enzyme found in nearly all living organisms exposed to oxygen and protects cells from oxidative damage (Loew, 1900). This raises the possibility that the decreased H 2 O 2 resistance of the pykA mutant could be due to a decreased catalase expression or activity in V. vulnificus cells. Therefore, we compared the relative catalase activities between the wild type, the pykA mutant, and a complemented strain (pykA/pRK-PKA) in the presence and absence of H 2 O 2 ( Figure 2B). While the addition of H 2 O 2 to the culture medium resulted in a significant induction of catalase in all strains tested, no significant difference could be detected in the catalase expression levels and activities among the three strains ( Figure 2B).
To confirm the effects of PykA activity and pyruvate on resistance to H 2 O 2 stress, the H 2 O 2 sensitivity assays were performed on LBS medium containing different combinations of H 2 O 2 and pyruvate ( Figure 2C). In agreement with observations in Figure 2A, the pykA mutant exhibited a significantly decreased survival on the same medium but supplemented with 0.25 mM H 2 O 2 . Intriguingly, the addition of exogenous pyruvate fully protected the cells from H 2 O 2 killing. On the contrary, there was no significant difference in H 2 O 2 sensitivity between the wild-type E. coli MG1655 and its otherwise isogenic pykA mutant strain (Supplementary Figure S2). These results indicate that the role of PykA in H 2 O 2 resistance is not conserved among bacteria: PykA-dependent production of pyruvate seems to play an important role in H 2 O 2 scavenging in V. vulnificus, whereas catalase is known to be the primary scavenger in E. coli when H 2 O 2 levels are high (Seaver and Imlay, 2001a,b).

HPr Increases H 2 O 2 Resistance by Stimulating PykA Activity in the Presence of Glucose
We have recently shown that the presence of glucose in culture medium increases the dephosphorylated form of HPr, which tightly interacts with PykA, stimulating its activity by decreasing the Km for PEP in V. vulnificus (Kim et al., 2015). To examine whether the stimulation of pyruvate production by dephosphorylated HPr affects H 2 O 2 resistance in vivo, we constructed an expression vector for a nonphosphorylatable (His15 to Ala) mutant of V. vulnificus HPr (pRK-H15A) and a vector coexpressing PykA and HPr(H15A) of V. vulnificus (pRK-PKA&H15A in Table 1). As shown in Figure 3A, the pykA mutant strain carrying pRK-PKA was more resistant to H 2 O 2 than the mutant carrying pRK-H15A, whereas transformation of the mutant with pRK-PKA&H15A further increased H 2 O 2 resistance compared with the strain carrying pRK-PKA, indicating that HPr(H15A) confers increased resistance to H 2 O 2 stress in the presence of PykA. It should be noted that the expression of HPr(H15A) alone had little effect on H 2 O 2 resistance of the pykA mutant (Supplementary Figure S3), suggesting that the effect of HPr(H15A) on H 2 O 2 resistance is mediated by the stimulation of PykA.
The interaction with and activation of PykA by HPr is observed in V. vulnificus but not in E. coli and this species specificity of the HPr-PykA interaction is determined by the C-terminal domain of PykA (Kim et al., 2015). To examine whether H 2 O 2 resistance of the pykA mutant harboring pRK-PKA&H15A was due to the specific interaction of HPr(H15A) with PykA, we constructed an expression vector for vePykA (pRK-vePKA) in which the C-terminal domain (amino acids 334 to 480) was replaced with that of E. coli PykA and a vector coexpressing vePykA and HPr(H15A) (pRK-vePKA&H15A). The pykA mutant transformed with pRK-vePKA was more resistant to H 2 O 2 than the mutant carrying pRK415 or pRK-H15A (Supplementary Figures S3, S4). However, the mutant carrying pRK-vePKA&H15A exhibited little difference in H 2 O 2 resistance compared with the strain carrying pRK-vePKA (compare Figure 3A and Supplementary Figure S4), implying that the interaction and stimulation of PykA by dephospho-HPr is important for the increased H 2 O 2 resistance of V. vulnificus.
It is known that HPr is mostly dephosphorylated in the medium supplemented with glucose, whereas it is predominantly in the phosphorylated state in the medium supplemented with galactose in V. vulnificus (Kim et al., 2015). The effect of the carbohydrate type on H 2 O 2 sensitivity was investigated by spotting serially diluted cells onto agar plates with M9 medium supplemented with 0.2% casamino acids and 2.5% NaCl (M9S medium), containing different combinations of carbohydrate, H 2 O 2 , and pyruvate ( Figure 3B). The wild-type strain and the pykA mutant strain harboring the pRK-PKA plasmid were considerably less sensitive to H 2 O 2 in the presence of glucose than in the presence of galactose, whereas no effect of the FIGURE 3 | Effects of PykA and dephospho-HPr on resistance to H 2 O 2 stress. (A) The pykA mutant cells harboring the indicated plasmids were serially diluted fourfold from 10 8 cells/ml, and 2-µl aliquots were spotted onto LBS medium containing different combinations of H 2 O 2 and sodium pyruvate as indicated. pRK-PKA expresses V. vulnificus PykA from its own promoter and pRK-PKA&H15A expresses both PykA and HPr(H15A) (a phosphorylation site mutant mimicking the dephosphorylated form) from their own promoters. (B) The indicated strains were serially diluted fourfold from 10 8 cells/ml, and 2-µl aliquots were spotted onto M9S medium containing a carbohydrate, H 2 O 2 and sodium pyruvate either alone or in different combinations as indicated. It should be noted that M9S agar plates were supplemented with H 2 O 2 at half the concentration added to LBS plates because cells were more sensitive to H 2 O 2 in M9S plates than in LBS plates. (C) The indicated strains were serially diluted fourfold from 10 8 cells/ml, and 2-µl aliquots were spotted onto LBS medium containing 0.25 mM H 2 O 2 and 0.25 mM indicated supplements. Representative data from at least three independent experiments are presented.
carbohydrate type was seen in the pykA mutant. Because the catalase expression level was little affected by the presence of these carbohydrates (Supplementary Figure S5), this carbohydrate effect is likely to be mediated solely through the phosphorylation state-dependent stimulation of PykA activity by HPr. More specifically, dephosphorylated HPr interacts with and stimulates PykA so that an increase in PykA-mediated pyruvate production confers resistance to H 2 O 2 stress in the presence of glucose.
Since the first description of the chemical reaction between α-ketoacids and H 2 O 2 producing acetate and CO 2 in 1904 (Holleman, 1904), many studies have confirmed this nonenzymatic H 2 O 2 -scavenging reaction (Bunton, 1949;Desagher et al., 1997;Troxell et al., 2014;Lopalco et al., 2016). Because α-ketoacids such as pyruvate and oxaloacetate are known to play vital roles in central carbon metabolism as well as to serve as H 2 O 2 scavengers, it was necessary to determine whether the protective effect of pyruvate against H 2 O 2 resulted from its ability to react with H 2 O 2 or from its requisite role in central metabolism.
To verify this, we tested the H 2 O 2 sensitivity of the pykA mutant by spotting on LBS medium containing H 2 O 2 and an α-ketoacid (pyruvate or oxaloacetate) or an α-hydroxyacid (lactate or malate) ( Figure 3C). V. vulnificus has both lactate dehydrogenase and malate dehydrogenase, which catalyze the reversible oxidation of lactate and malate to pyruvate and oxaloacetate, respectively, coupled to the reduction of NAD + to NADH. Therefore, if the protective effect of pyruvate is due to an increased metabolic flux, the same concentration of lactate and malate should exert a higher, or at least similar, protective effect against H 2 O 2 than pyruvate and oxaloacetate, respectively. As shown in Figure 3C, However, only α-ketoacids, pyruvate, and oxaloacetate, could rescue the H 2 O 2 sensitivity of the pykA mutant. Furthermore, the addition of citrate to LBS medium did not rescue the pykA mutant cells from H 2 O 2 stress (Supplementary Figure S6). These data suggest that the protective effect of pyruvate is not due to an increased metabolic flux.
Bacterial mutants deleted for a glycolytic gene usually show a severe growth defect in LB and minimal medium due to limited metabolic capabilities (Irani and Maitra, 1977;Commichau et al., 2013). However, there was no difference in growth between the wild type and the pykA mutant strain in LBS (Supplementary Figure S1). These results suggest that the V. vulnificus pykA mutant does not have any significant metabolic perturbation. We have previously shown that the intracellular pyruvate concentration drastically increased after the addition of glucose in the wild-type strain. Notably, the intracellular pyruvate concentration also increased in pykA mutant cells after the addition of glucose, but to a lesser extent. In the presence of galactose, the pyruvate concentration was similar or slightly higher in wild-type cells compared to the pykA mutant (Kim et al., 2015). Furthermore, there was no significant difference in the acetate concentrations between the two strains in LBS medium and M9S medium supplemented with glucose (Supplementary Figure S7). These data also suggest that the phenotype of the pykA mutant is not due to the limited metabolic flux.

PykA Confers Resistance to H 2 O 2 Stress Caused by Fungal Neighbors
Although the pykA mutant showed a higher rate of VBNC cell formation compared to the wild-type strain, this phenotype does not seem to be dependent on the presence of glucose, since V. vulnificus enters the VBNC state in response to a temperature downshift and cells in the VBNC state have very low metabolic activity (Ayrapetyan et al., 2015). Therefore, we assumed that PykA may play other glucose-dependent role(s) in the natural habitat of V. vulnificus. Since the regulatory interaction of HPr with PykA in the presence of glucose was not observed in E. coli, we questioned why V. vulnificus, but not E. coli, cells need a higher PykA activity in the presence of glucose. The primary habitat of E. coli is the lower intestine of warm-blooded animals (Whittam, 1989), whereas V. vulnificus is usually present in coastal marine environments (Bhadury et al., 2011;Horseman and Surani, 2011). Thus, we assumed that the biochemical and physiological differences between the two species may have arisen from their ecological differences. Recent studies have demonstrated the coexistence of several fungal species such as Aspergillus, Penicillium, Fusarium, and Rhodotorula with many bacteria including V. vulnificus in the gut of oysters and other shellfishes (Borzykh and Zvereva, 2012;Odu et al., 2012;Froelich and Oliver, 2013;Amadi, 2016;Chen et al., 2016). We could also confirm the coexistence of fungi and Vibrio species in the digestive gland of oysters (Supplementary Figure S8). High concentrations of glycogen in the stomach, digestive gland, and other organs of most FIGURE 5 | Effect of PykA activity on the survival of V. vulnificus against H 2 O 2 stress generated by its fungal neighbors in the presence of glucose. (A-D) Wild-type V. vulnificus CMCP6 (squares), the pykA mutant (triangles) and the complemented strain (circles) were inoculated in (A) fresh M9S medium containing 0.2% glucose, (B) the cell-free culture filtrate of A. fumigatus grown in M9S medium containing 0.2% glucose, (C) the cell-free culture filtrate of A. welwitschiae grown in M9S medium containing 0.2% glucose, or (D) the cell-free culture filtrate of A. welwitschiae grown in M9S medium containing 0.2% glucose and 1 mM sodium pyruvate. Bacterial growths were then monitored by measuring the optical density at 600 nm and presented as the mean ± standard deviations (n = 3). (E) The indicated strains of V. vulnificus were cultured alone (1-3) or with A. welwitschiae (Aw, 4-6) in M9S medium containing 0.2% glucose. The overnight cultures were serially diluted fourfold and 1.5-µl aliquots were spotted onto TCBS (thiosulfate-citrate-bile salts-sucrose) medium containing 2 µg/ml tetracycline. (F) The H 2 O 2 levels in culture media from (E) were determined using the hydrogen peroxide assay kit (BioVision). The results are presented as the mean ± standard deviations (n = 3).
bivalves including oysters indicate that bivalves may be a good source of glucose for both bacteria and fungi (de Zwaan, 1983;Tikunov et al., 2010). Antagonistic interactions between bacteria and fungi in competing for a common substrate have been documented in many habitats including an aquatic environment (Mille-Lindblom et al., 2006;Arvanitis and Mylonakis, 2015). Interestingly, many fungi have been shown to naturally produce glucose oxidase (GOX), which displays antibacterial activity through the production of H 2 O 2 in the presence of glucose (Wong et al., 2008). Therefore, we sought to determine whether the glucose-dependent activation of PykA in V. vulnificus has anything to do with the glucose-dependent H 2 O 2 production by its fungal neighbors. Because the V. vulnificus CMCP6 strain was originally isolated from a patient in the west coast of Korea (Kim et al., 2003), we tested the GOX activity from the two fungal strains belonging to the Aspergillus genus isolated from the west coast of Korea (Lee et al., 2016) (Figure 4A). Both zymographic analyses ( Figure 4B) and coupled enzyme assays with peroxidase and o-dianisidine ( Figure 4C) revealed that A. welwitschiae produced a significant amount of GOX when grown in the presence of glucose, whereas little GOX was detected in the crude extract of A. fumigatus cells. In accordance with this observation, little H 2 O 2 was accumulated in the culture medium of A. fumigatus regardless of the carbohydrate source, whereas a measurable amount of H 2 O 2 was accumulated in the culture supernatant of A. welwitschiae grown in M9S medium containing glucose, but not in the presence of galactose ( Figure 4D). These data suggest that some fungal neighbors of V. vulnificus can produce a significant amount of H 2 O 2 through the GOX activity to outcompete surrounding bacterial cells when glucose is available.
To test whether or not a fungus can indeed kill neighboring V. vulnificus cells by generating H 2 O 2 in the presence of glucose, A. fumigatus and A. welwitschiae were inoculated to M9S medium supplemented with glucose. After overnight incubation, the culture medium was cleared from the fungal cells and mycelia, and then inoculated with wild-type, pykA mutant, and the pykA mutant carrying pRK-PKA. As expected from H 2 O 2 sensitivity analyses above, the three strains did not exhibit any growth defects in fresh M9S/glucose medium ( Figure 5A) and in the cellfree culture filtrate of the A. fumigatus in M9S/glucose medium ( Figure 5B). However, the pykA mutant displayed severe growth retardation when inoculated in the filtrate of the A. welwitschiae culture in M9S/glucose medium (Figure 5C), but this growth retardation was not observed in the cell-free filtrate of the A. welwitschiae culture in M9S/glucose medium supplemented with pyruvate ( Figure 5D). The filtrate of the A. welwitschiae culture in M9S/galactose medium did not inhibit growth of the three strains (Supplementary Figure S9). Together, these data indicate that a fungus expressing an active form of GOX can kill V. vulnificus cells by generating H 2 O 2 in the presence of glucose and PykA plays an important role in the protection of V. vulnificus cells from H 2 O 2 stress.
To further confirm the role of PykA in response to H 2 O 2 stress caused by fungal competitors, the wild-type, pykA mutant, and the complemented strain were co-inoculated with A. welwitschiae in M9S/glucose medium. After co-culture of V. vulnificus strains with the fungus for 24 h, the cultures were serially diluted and then spotted onto selective medium for V. vulnificus (Figure 5E). When co-cultured with A. welwitschiae in M9S/glucose medium, the pykA mutant barely grew, whereas obvious growth was detected in the other two strains. Intriguingly, the H 2 O 2 level in the medium of the co-culture of the fungi and the pykA mutant was significantly higher than those from the co-cultures of the fungi and the other two strains (Figure 5F). Based on these data, we assumed that V. vulnificus resists the H 2 O 2 -mediated killing activity of its fungal competitors by increasing PykA-mediated pyruvate production in the presence of glucose.

DISCUSSION
We report here a new mechanism that V. vulnificus uses to resist the H 2 O 2 -mediated killing of its fungal neighbors. Our earlier work established that dephosphorylated HPr interacts with and stimulates the activity of PykA in the presence of glucose in V. vulnificus but this regulatory interaction does not occur in E. coli (Kim et al., 2015), even though both species belong to γ-proteobacteria. While the primary habitat of E. coli is the lower intestine of warm-blooded animals (Whittam, 1989), the FIGURE 6 | Model showing how V. vulnificus cells cope with H 2 O 2 stress in the presence of glucose. Some fungi are known to inhibit bacterial growth in the presence of glucose by generating H 2 O 2 through the reaction catalyzed by glucose oxidase. Meanwhile, HPr of the PEP-dependent carbohydrate-transporting phosphotransferase system is dephosphorylated in the presence of glucose, and dephosphorylated HPr stimulates pyruvate kinase A in V. vulnificus. The increased pyruvate production then enables this bacterium to cope with H 2 O 2 stress imposed by their fungal neighbors or other environmental factors. estuarine environment is the primary habitat of V. vulnificus. Therefore, we assumed that regulatory functions of the PTS in bacteria might have evolved as adaptations to environmental conditions.
Most bacteria have anti-oxidant defense systems to deal with oxidative stress by synthesizing catalase, which decomposes H 2 O 2 into water and oxygen (Loewen, 1997). While E. coli possesses periplasmic catalase HPI (KatG) and the cytoplasmic catalase HPII (KatE) (Switala et al., 1990), V. vulnificus has been reported to express only the katG gene (Park et al., 2004). In the previous study, Park et al. (2004) reported that V. vulnificus cells were generally more sensitive to H 2 O 2 than other enteric bacteria and also suggested that V. vulnificus may have mechanisms for oxidative stress response that are distinct from those found in E. coli (Park et al., 2004). For this reason, we speculated that V. vulnificus could use pyruvate as a defense mechanism to H 2 O 2 . Pyruvate is known to react with H 2 O 2 and decompose it into CO 2 and acetate to protect the cell from oxidative stress (Nath et al., 1994;Desagher et al., 1997;Giandomenico et al., 1997;Miwa et al., 2000). Therefore, V. vulnificus may increase pyruvate production via the HPr-PykA interaction to protect themselves from H 2 O 2 stress caused by their fungal neighbors in the presence of glucose (Figure 6). The conversion of PEP to pyruvate is catalyzed by two pyruvate kinases, PykF and PykA. PykF, whose activity is not regulated by HPr, is essential for normal growth of V. vulnificus. PykA was dispensable under normal growth conditions in V. vulnificus (Figure 5 and Supplementary Figures  S1, S9A), suggesting that the protective effect of pyruvate is not due to an increased metabolic flux. The stimulation of PykA activity by dephosphorylated HPr results in the additional production of pyruvate by converting an incoming glucose.
The critical role of pyruvate in protection from H 2 O 2 damages was also reported in other bacteria. Pyruvate protects pathogenic spirochetes, Borrelia burgdorferi and Leptospira interrogans from H 2 O 2 toxicity (Troxell et al., 2014). In particular, B. burgdorferi lacks genes encoding catalase and therefore is sensitive to a micromolar dose of H 2 O 2 generated by GOX. However, exogenously supplied pyruvate fully protected B. burgdorferi against H 2 O 2 killing. Under the H 2 O 2 -challenged environment, Pseudomonas fluorescens also increases pyruvate production, although the precise mechanism is still unclear (Bignucolo et al., 2013). In this study, we uncovered a sophisticated strategy which allows V. vulnificus to cope with the killing action of competitors. This strategy, involving the stimulation of pyruvate kinase in the presence of glucose to cope with H 2 O 2 stress, could be widespread in the aquatic environments, whereas H 2 O 2 can be generated from both biotic and photochemical reactions (Baltar et al., 2013).