3.1 Determination of inducing effect from W. anomalus Y-5
In order to determine whether inducing effect required live W. anomalus Y-5, the supernatant, heat-killed yeast, and live cell from W. anomalus Y-5 were separately added to 100 mL MRS broth inoculated with 1% overnight-activated L. paraplantarum RX-8, then assayed plantaricin production after growing at 37℃ for 24 h. The results of plantaricin production showed that only co-cultures with live W. anomalus Y-5 could increase the yield of plantaricin and retain the inducing effect (Fig. 1-A). Furthermore, to determine if the inducing effect was dependent on cell-to-cell direct contact between W. anomalus Y-5 and L. paraplantarum RX-8, the two microorganisms were cultured in an indirect co-culture system containing a membrane that is impermeable to bacterial cells but permeable to smaller molecules. It was found that plantaricin production was increased in the contact-independent co-culture system (Fig. 1-B). This suggested that the inducing effect may not require cell-to-cell contact and due to a secreted factor or metabolite, but the secreted factor or metabolite needed continual production by living bacteria. In a recent study, bacteriocin-producing strain L. plantarum DC400 indirectly co-cultured with two strains (Lactobacillus sanfranciscensis DPPMA174 and Lactobacillus rossiae A7) could increase the bacteriocin activity [10]. As well as the co-culture system of L. acidophilus N2 and L. delbrueckii subsp. lactis ATCC 4797 enhanced the bacteriocin production from non-detectable levels (< 100 AU/mL) to 3200 ~ 6400 AU/mL, and the key inducing compound was a protein with a molecular size of approximately 58 kDa [19]. Therefore, it can be speculated that one or more key inducing substances may be present in the culture system of W. anomalus Y-5 to promote efficient synthesis of plantaricin RX-8.
3.2 Transcriptomic and proteomic analyses of W. anomalus Y-5 in co-culture
To identify the DEGs and DEPs of W. anomalus Y-5 inducing L. paraplantarum RX-8 on bacteriocin production under co-culture, the cells of W. anomalus Y-5 in co-culture and mono-culture were harvested at 24 h for transcriptomic and proteomic analyses. Considering the large transcriptomic data, the threshold for considering genes as differentially expressed was established to fold change ≥ 2 and p < 0.05. Finally, the transcription of 1968 genes were detected, including 994 up-regulated genes and 974 down-regulated genes (Fig. 2-A). In addition, TMT-based quantitative proteomics was further performed to identify the DEPs of W. anomalus Y-5 in co-culture and mono-culture. Compared to the mono-culture group, 645 DEPs (fold change ≥ 1.3, p < 0.05) were identified, including 418 up-regulated proteins and 227 down-regulated proteins (Fig. 2-B). PCA performed between samples in co-culture (RY) and mono-culture (Y) conditions demonstrated that RY and Y were significantly different (Fig. <link rid="fig2">2</link>-C and 2-D). For each culture condition, all the samples showed a remarkable similarity. As shown in the Venn diagrams, the 1968 DEGs and 645 DEPs were subsequently applied to functional analysis in the functional enrichment of the KEGG pathway (Fig. 2-E). After KEGG analysis, these pathways were remarkably up-regulated under co-culture, which included the nitrogen metabolism, TCA cycle, and MAPK signalling pathway. Other pathways were significantly down-regulated under co-culture, which included fatty acid metabolism and biosynthesis of unsaturated fatty acids (Fig. 3).
3.2.1 Glucose signaling pathway of W. anomalus Y-5 in co-culture
Compared with W. anomalus Y-5 in mono-culture, W. anomalus Y-5 in co-culture suffered multiple stress including acid stress, osmotic pressure and absence of nutrients, which led to W. anomalus Y-5 growth slowly and cell number decline [14]. In the acid environment, the protein kinase A (PKA) pathway in yeast regulates cell death [20]. And the PKA pathway was rapidly triggered by glucose or other essential nutrients (nitrogen, phosphate and sulfate, etc.) through cAMP in appropriately starved yeast [21]. The cAMP was produced by adenylate cyclase (AC) which was encoded by the gene CYR1 [22]. And the cAMP was hydrolyzed by Pde2 cAMP phosphodiesterase [23]. The activity of AC was controlled by two glucose-sensing systems. On the one hand, the G-protein-coupled receptor (GPCR) system, consisting of Gpr1 and Gpa2 proteins, is responsible for sensing extracellular glucose and activating cAMP synthesis [24]. Glucose binded to the Gpr1 receptor, which activated cAMP synthesis through the Gpa2 protein [25]. Under co-culture conditions, the protein expression and transcriptional expression of Gpa2 were up-regulated by 2.455- and 1.304-fold, respectively. On the other hand, AC was dependent on Ras proteins in yeast cells, and the Cdc25, Sdc25, Ira1 and Ira2 controlled the activity of the Ras proteins [26]. The Ira proteins could inhibit the activity of the Ras proteins, then avoid an apoptotic death of cells through ROS-producing mitochondria which were induced by acid stress [27]. Under co-culture, the expression of Ira1 protein up-regulated 1.4-fold. It was suggested that the IRA1 negatively regulated Ras-PKA to increase the resistance of yeasts to acidic conditions.
Meanwhile, PKA directly phosphorylates cytosolic enzymes against adverse conditions. Trehalose is a storage carbohydrate and stress protectant in yeast and other fungi [26]. It was found that trehalase encoded by gene NTH2 in W. anomalus Y-5 under co-culture was up-regulated by 2.49-fold. The up-regulation of hexokinase II (encoded by Hxk2) and trehalose-6-phosphate phosphatase (encoded by Tps2) indicated that the increase of uptake on glucose and the glucose was converted into trehalase in W. anomalus Y-5 under co-culture. Furthermore, gene hxk2p played an important role in acid tolerance, protein Hxk2p could avoid acid-induced apoptosis and cell death by preventing the entry of Ras2p in the mitochondria [28]. Therefore, the up-regulation of gene Hxk2 under co-culture is helpful to yeast survive in acid conditions.
3.2.2 Amino acid biosynthesis of W. anomalus Y-5 in co-culture
Glutamate (Glu) and glutamine (Gln) are major precursors for amino acid biosynthesis in yeast [29]. Owing to the competition of W. anomalus Y-5 and L. paraplantarum RX-8, W. anomalus Y-5 in co-culture suffered poor nitrogen conditions. Protein Gap1 took charge of amino acid uptake and its activity is regulated by the available nitrogen source via nitrogen catabolite repression (NCR) [30]. Under preferred nitrogen sources, the amino acid transport activity of Gap1p is low while in the presence of poor nitrogen sources Gap1 activity is high [29]. In the co-culture system, the protein expression and transcriptional expression of Gap1 were up-regulated by 1.32- and 16.73-fold, respectively. Moreover, the decrease in intracellular concentrations of Glu and Gln would influence the expression of NCR genes. The TORC1 was not stimulated any longer, which led to active Tap42 protein phosphatase complexes being released into the cytosol [26]. Tap42 reduced phosphorylation of Ure2 and subsequent stimulation of NCR gene expression [26]. Under co-culture conditions, the transcriptional expression of Tap42 was up-regulated by 1.27-fold, and the Ure2 protein was up-regulated by 1.592-fold. Tap42 also dephosphorylated Mks1, which then complexes with Rtg2. This allows Rtg1, 3 nuclear localization resulting in stimulation of the expression of retrograde (RTG) genes, sustaining amino acid biosynthesis through the synthesis of glutamate and glutamine [31]. Specifically, glutamate dehydrogenase synthesized glutamate from α-ketoglutarate and ammonia as precursors. And glutamine synthetase Gln1 catalyzes ammonia and glutamate to synthesize glutamine [30]. Therefore, the concentration of α-ketoglutarate was important to the synthesis of Glu and Gln. The RTG signaling pathway could increase the expression of enzymes used for the synthesis of α-ketoglutarate from oxaloacetate in the TCA cycle. Under co-culture conditions, the protein expression of Rtg2 in W. anomalus Y-5 was up-regulated by 2.81-fold, the protein involved in Glu and Gln synthesis, Cdh1 and Gln1, were up-regulated by 1.59- and 1.31-fold, respectively. It was indicated that protein Rtg2 induced RTG signaling pathway promoting Glu and Gln synthesis to satisfy the biosynthesis of amino acid and protein in yeast. In addition, the production of nisin was increased when L. lactis subsp. lactis DY13 co-cultured with L. monocytogenes Scott A and the glutamate decarboxylase (encoded by gadD2) also showed a greater expression under co-culture conditions [32]. It was found that gadD2 was up-regulated 12.81-fold in W. anomalus Y-5 co-cultured with L. paraplantarum RX-8. And it was observed that transcriptional expression of glutamate decarboxylase increased by 2.02-fold [14]. Thus, it was supposed that Glu and Gln were produced by W. anomalus Y-5 and consumed by L. paraplantarum RX-8, which have a key inducing effect on bacteriocin production.
3.2.3 MAPK pathway and glycerol catabolism of W. anomalus Y-5 in co-culture
To rapidly respond the environmental changes, yeast developed a series of signaling pathways for sensing and responding appropriately to various stimuli [33]. One important pathway for eliciting these responses is the three-tiered cascade of protein kinases known as the mitogen-activated protein kinase (MAPK) pathway. Under co-culture conditions, many genes and proteins related to the MAPK pathway in W. anomalus Y-5 have changed at different levels. In this context, the most significant change was high osmolarity glycerol mitogen-activated protein kinases (HOG-MAPK). The function of the HOG-MAPK pathway in yeast is to combat external hypertonic stress by adding the synthesis of glycerol and trehalose [34]. Under co-culture conditions, the transcriptional and protein expression of Hog1 in W. anomalus Y-5 were up-regulated 1.29 and 1.609-fold, respectively. And the gene of ptc2 in W. anomalus Y-5 was up-regulated 1.105-fold. Protein Sln1 and Sho1 were the upstream signal sensors of the tertiary kinase cascade system and were responsible for the perception of signals in the environment [33]. The serine-threonine phosphatases PTC (Ptc1, Ptc2 and Ptc3) took charge of the negative regulation of Hog1 to avoid excessive activation of Hog1 from causing cell death.
The HOG-MAPK pathway was reported to increase glycerol biosynthesis in response to hypertonic and acid stress [35]. Glycerol is produced in two steps, the glycerol-3-phosphate dehydrogenase (encoded by GPD) catalyzed glycolytic dihydroxyacetone phosphate to L-glycerol 3-phosphate (G3P), then G3P was dephosphorylated to glycerol by a glycerol 3-phosphatase (encoded by GPP) [36]. The expression of GPD and GPP was highly dependent on the HOG-MAPK pathway [34]. For the glycerol catabolic pathway, GDH oxidated glycerol to dihydroxyacetone (DHA), which was subsequently phosphorylated to DHAP via DHA kinase (DAK) [36]. Meanwhile, DAK reduced NAD to NADH. Under co-culture conditions, the gene GPD, GPP and DAK1 were up-regulated 18.957-, 5.789-, and 2.333-fold, respectively. The up-regulation of glycerol biosynthesis and metabolism could generate more NADPH for combating reactive oxygen species, and glycerol catabolism in Saccharomyces cerevisiae BY4741 was activated upon acid stress [34].
3.2.4 Antioxidant system and cell cycle of W. anomalus Y-5 in co-culture
The oxidative stress response in yeast was induced by the acid environment, and NADPH was used for the detoxification of oxidized compounds [34]. It was observed that several genes involved in oxidative stress were up-regulated, cytosolic catalase T (CTT1) show 1.522-fold up-regulation and gene PRX1 was up-regulated 9.658-fold. Protein CTT1 and thioredoxin peroxidase (encoded by PRX1) could protect the yeast cells from oxidative damage by H2O2 [34]. It has been reported that the expression of gene CTT1 in S. cerevisiae was controlled by protein Ras2, AC, and cAMP-dependent protein kinases which sensed essential nutrients [37]. Gene GRE2 and GRE3 encoded NADPH-dependent aldose reductases which detoxified the harmful oxidizing compound methylglyoxal [38]. The expression of Gre2 and GRE3 was dependent on the Hog1 under osmotic, ionic and oxidative stress [39]. Under co-culture, the gene GRE3 in W. anomalus Y-5 was up-regulated 6.177-fold. Similarly, GRE3 was up-regulated 2-fold in Saccharomyces cerevisiae JP1 under acid stress, and its paralog GRE2 show up-regulated in Saccharomyces cerevisiae BY4741[40–41]. The yAP-1 and Skn7 transcriptional regulators regulated the expression of thioredoxin-related genes, including TRX2, TRR1, TSA1, AHP1 and YDR453c. Under co-culture conditions, the transcriptional and protein expression of Skn7 in W. anomalus Y-5 were up-regulated 1.514 and 2.16-fold, respectively. Tsa1 showed peroxidase activity towards hydrogen peroxide and alkyl hydroperoxides, and it was 9.042 up-regulated in W. anomalus Y-5 under co-culture.
The yeast cell was forced to enter the quiescent G0 state in glucose, nitrogen or phosphorous starvation conditions [42]. The genes TRX1 and TRX2 encoding cytoplasmic thioredoxins were important to sulfate assimilation. And the deletion of TRX1 and TRX2 could prolong S phase and shorten G1 interval in the cell cycle. Under co-culture conditions, the gene of Trx1 in W. anomalus Y-5 was up-regulated 2.012-fold in co-culture. The HOG-MAPK pathway was reported to repress mating in cell wall-damaged yeasts during acid stress [34]. It could be proposed that W. anomalus Y-5 in co-culture enter the quiescent G0 state to reduce energy consumption and avoid DNA replication or chromosome segregation errors. It was helpful to acquire adequate time for proper adaptation and successful passage to the next phase [42]. Under co-culture conditions, the gene related to cell cycle progression had been down-regulated.
3.4 The change of exo-metabolome on plantaricin production in co-culture
According to the results of transcriptome and proteome, we found that the potential key inducing metabolites major concentrate on the amino acids and central carbon metabolism (CCM). Then, the concentrations of amino acids and compounds in CCM from the supernatant of L. paraplantarum RX-8 / W. anomalus Y-5 in mono-culture and co-culture at 24 h using targeted exo-metabolome to pinpoint exchanged metabolites. Metabolites with p < 0.05 and VIP > 1 were considered differetial expression metabolites (DEMs) between mono-culture and co-culture, and 37 metabolites were identified as DEMs between RX-8 mono and co-culture, 47 metabolites were identified as DEMs between Y-5 mono and co-culture. The mass spectrometry approach revealed the change of secreted/uptaken metabolites among RX-8, Y-5 and co-culture. Metabolites that accumulate with the growth of bacteriocin-inducing strain (W. anomalus Y-5) and become depleted during the plantaricin production of bacteriocin-producing strain (L. paraplantarum RX-8), thus exhibiting down-regulated trends in co-culture compared with Y-5 and up-regulated trends in co-culture compared with RX-8 (red cluster in Fig. 6), were selected as candidates mediating inducing effect during the plantaricin production in co-culture. This cluster showed many candidates exchanged compounds, including glutamine, L-aspartate, uracil, guanosine, inosine, fumaric acid, adenine, and pyruvic acid. Glutamine and L-aspartate were observed both in the targeted exo-metabolome and predicted modeling. Metabolites in central carbon metabolism (CCM) flow mainly from L. paraplantarum RX-8 to W. anomalus Y-5. Among them, aspartic acid undergoes dehydrogenation to produce fumaric acid, which in turn provides intermediates for the TCA cycle [43]. Growth of L. plantarum depends on several available TCA cycle intermediates, organic acids and amino acids, such as pyruvic and fumaric acid [44]. In the co-culture of Saccharomyces cerevisiae and L. plantarum, the metabolites from S. cerevisiae could promote the growth of L. plantarum, and 2-oxoglutarate, glutamine and threonine were important cross-feeding metabolites [45]. Overall, amino acids were major agents of co-culture-inducing bacteriocin production.
3.5 Addition of metabolites at different times as induction in the bacteriocin production by L. paraplantarum RX-8
Following the indications of candidate exchanged metabolites by exo-metabolome, we added targeted metabolites at detected concentrations at different times in the L. paraplantarum RX-8 system. The plantaricin production of the above samples at 24 h was shown in Fig. 7. Apart from aspartic acid, other exchanged metabolites could induce plantaricin production in the proper times. It could be found that most targeted metabolites (except aspartic acid) showed inducing effects when they were added at 8 h. When glutamine was added at 8–20 h, L. paraplantarum RX-8 presented the higher antimicrobial activity compared to the bacteriocin production in mono-culture. Adding glutamine at 0 h and 4 h in the L. paraplantarum RX-8 system did not have an inducing effect on bacteriocin production. However, only adding pyruvic acid at 8 h in the L. paraplantarum RX-8 system had an inducing effect on bacteriocin production. Guanosine and inosine showed inducing effects when they were added after 4 h incubation of L. paraplantarum RX-8. It was earlier than other targeted metabolites. All the exchanged metabolites did not show inducing effects when they were added at 0 h in the L. paraplantarum RX-8 system. When the inducing compounds were added in early times, these metabolites may be depleted by L. paraplantarum RX-8 which led to the loss of induction. This could explain why directly adding the supernatant of W. anomalus Y-5 at 0 h failed to induce bacteriocin synthesis.
In the previous study, we found that 0.5 mg/mL glutamine could induce plantaricin production at 0 h, while 733.086 ng/mL glutamine also have an inducing effect on plantaricin production when it was added in the L. paraplantarum RX-8 system at 8 h in this research. It was suggested that the inducing effect was influenced by adding time and concentration, inducing substances should be added at a specific time to exert an inducing effect. Similarly, the production of plantaricin Q7 increased significantly when glutamine was added to L. plantarum Q7 culture system for 6 h, but glutamine had little effect on plantaricin Q7 yield when it was added at 2, 12, and 16 h [8]. It was reported that microbisporicin production, a potent type I lantibiotic, was dependent on the guanosine pentaphosphate (ppGpp) synthetase gene (relA) of Microbispora corallina [46]. It was helpful to generate ppGpp when adding guanosine to the culture system. The uracil from E. coli could increase the growth rate and production of antimicrobial substance of L. plantarum [47]. Pyruvate is considered to be one of the key intermediates in the Embden-Meyerhof-Parnas (EMP) pathway and the tricarboxylic acid (TCA) cycle, playing an important role in the synthesis of proteins, peptides and amino acids. Thus, some findings suggest that the addition of pyruvate may initiate the biosynthesis of bacteriocin Lac-B23 earlier or prolong the effective biosynthetic phase [48].
3.6 Effect of the adding time on inducing activity from the supernatant of W. anomalus Y-5
The effect of different concentrations of W. anomalus Y-5 supernatant and different addition times (0, 4, 8, 12, 16 and 20 h) on RX-8 induced activity was tested (Fig. 8). When we performed the induction with non-concentrated W. anomalus Y-5 supernatant, there was no significant increase in bacteriocin production throughout the incubation compared to the negative control (L. paraplantarum RX-8 monoculture medium), which is also in agreement with the results in 3.1, further validating the accuracy of the experiment. Next, we concentrated the yeast supernatant to 1/6 of the original volume to ensure the concentration of the inducing substances in it, and found that there was a significant increase in bacteriocin production when the concentrated W. anomalus Y-5 supernatant was added at 8, 12, and 16 hours, while the addition of W. anomalus Y-5 supernatant at 0, 4, and 20 hours had a negligible effect on bacteriocin production. This is the same trend as the induced metabolites in 3.5.
From the above results, it was found that the actual induction time of the supernatant of W. anomalus Y-5 was 8 h-16 h, which is roughly the same as the time for the individual inducing substances in 3.5 to act alone. It was found that Lactobacillus paracasei HD1-7 secreted a bacteriocin named Paracin1.7 during the culture process, and this bacteriocin would be synthesized in large quantities by the fermentation supernatant of yeast and bacteria, which showed a strong antibacterial effect. This bacteriocin inhibits the growth of some bacteria but not yeast [49]. It is with this in mind that our study was conducted. In contrast, the addition of Klebsiella weissii Br26 to fermentation supernatants at different growth stages resulted in a significant but not identical enhancement of its inhibitory activity, suggesting that the biosynthesis of bacteriocin can be self-induced in addition to receiving the influence of the time of addition [50]. Similarly, the addition of further cell-free supernatant of Bacillus subtilis B38 increased its own antimicrobial activity by 8-fold to 160 AU/mL [51]. Summarizing the results of the above studies, it could be inferred that when the concentration of the inducing substance in the inducing supernatant reaches a certain threshold, the inducing effect appears; while when the inducing supernatant is added at the wrong time, the inducing effect does not occur.