Nutrient cross-feeding and TOR signaling propelled the consortium assembly and biomass growth
G. lucidum and L. plantarum have significantly distinct physiological and nutrient requirements for growth. We started by designing suitable physiological conditions and a medium that could support the growth of both strains. Our result showed that YPDA (pH 6.8) could support the growth of both strains at an incubation temperature of 28°C (Fig. 1a). Conversely, other synthetic media that we assessed (nutrient agar (NA), potato dextrose agar (PDA), MRS) couldn’t simultaneously promote the growth of the two strains. We observed that the interaction between the strains in the coculture plate did not reduce the colony size of the L. plantarum nor affected the morphological appearance of the G. lucidum mycelia relative to the monoculture plates. Although the spatiotemporal expansion of G. lucidum was faster on the monoculture plate within the first seven days of the culture than on the coculture plate, competitive exclusion of L. plantarum by the faster-growing G. lucidum was not observed in the coculture. Notably, there was no inhibitory zone between the strains at contact points; instead, the two strains steadily grew and overlapped each other on the coculture plate before the 14th day, which hints at a cooperative and not antagonistic relationship.
Subsequently, we co-cultured the strains in an okara (soybean pulp) based liquid medium. Interestingly, the cells quickly formed a pellet-like microcosm (spiky and fluffy cotton ball) that progressively grew bigger with increased incubation time (Fig. 1b). The population dynamics provided valuable information that helped to understand the inter-species interactions in the consortium. The population growth of the cultures was quantified by measuring the accumulated biomass weight. The dry cell weight of the G. lucidum monoculture (GLM14) was 6.9 g/L, the L. plantarum monoculture (LBM14) 2.1 g/L, and the G. lucidum/L. plantarum coculture (GLC14) 11.6 g/L. The cell population of the GLC increased continuously, while that of the monocultures remained the same after 8 days of incubation (Fig. 1c).
Similarly, while the Brix values of the monocultures declined significantly after day 8 of the incubation, the Brix value of the coculture remained steady at 4.5% from its initial 5%. There was a sharp decrease in the pH of the GLM (5.33 to 3.19 ) and GLC (5.36 to 2.94) between days 1 to 8 of the incubation. Conversely, after day 8 of the fermentation, the pH of the GLM rose from 3.19 to 3.96 and the GLC from 2.94 to 3.51 on day 14. At the same time, the pH of the LBM decreased continuously from 5.35 to 3.31 (Fig. 1e). The rise in the pH after day 8 of the culture indicates that G. lucidum has acid production regulatory mechanisms. The major organic acids produced in the coculture were citric, pipecolic, gluconic, and kynurenic acid (Fig. 1f). G. lucidum produced citric, pipecolic, and kynurenic acids, while L. plantarum produced gluconic and kynurenic acids.
Nutrient competition affects transient and long-term population dynamics and the overall stability of a microbial consortium. The okara medium provided multiple carbon sources for the strains, including sucrose, D-maltose, D-psicose, α-D-glucose, L-fucose, sorbitol, L-gulose, and D-fructose (Fig. 1g). D-Fructose was the predominant sugar, while sucrose was the least abundant sugar in the medium. At the early stage of the monocultures (GLM7 and LBM7), when D-glucose concentration was low, the G. lucidum and the L. plantaruum utilized D-fructose, L-fucose, and L-gulose as their primary energy sources. At the same time, sucrose, D-maltose, and other complex polysaccharides in the medium were hydrolyzed to D-glucose, sorbitol, and D-psicose. Meanwhile, G. lucidum switched to sorbitol as its primary carbon source when its concentration became the most abundant (8.30 g/L) in the GLM. This metabolic switch resulted in the accumulation of D-fructose, D-glucose, L-fucose, and L-gulose at the later stage (GLM14) of the culture. A similar metabolic transition was also observed in the LBM when the sorbitol concentration rose from 4.25 to 9.08 g/L in the medium; however, there was no concurrent accumulation of D-glucose and L-gulose at LBM14. In contrast, to the GLM and LBM, the GLC simultaneously utilized all the available sugars (sucrose, D-maltose, α-D-glucose, L-fucose, sorbitol, L-gulose, and D-fructose) except for D-psicose as its energy source at the early stage of the consortium assembly (GLC7) (Fig. 1f). But, after the formation of a stable consortium, the cells intuitively toggle switched to D-glucose as its primary carbon source.
A similar self-regulated metabolic switch was observed in the metabolism of the amino acid. At the early stage of the consortium assembly (GLC7), G. lucidum synthesized L-isoleucine, tryptophan, and L-glutamic acid to support L. plantarum growth, while L. plantarum, in turn, synthesized L-tyrosine and L-proline for the G. lucidum resulting in multiple amino acids cross-feeding (Table S1). However, after forming a stable consortium (GLC14) and a significant build-up of essential amino acids, the cells synchronized and intuitively switched to separate amino acids as their nitrogen sources, thereby reducing nutrient competition and stabilizing community life in the niche.
Target Of Rapamycin (TOR) is a highly conserved nutrient-sensitive protein kinase that has been extensively reported to control cells’ physiology and growth in response to nutrient availability (Loewith & Hall, 2011). We probed the possible influence of TOR signaling in the consortium’s autonomous regulation of carbon and amino acid metabolism. The TOR region of the G. lucidum was silenced using the RNA interference (RNAi) technique and subsequently cocultured with the L. plantarum. The mycelium growth of the TOR-silenced G. lucidum strain was slower than the wild-type. The biomass accumulation by the silenced strain in both the monoculture and coculture was significantly lower than in the wild-type (Fig S1). More importantly, the TOR-silenced strain could not establish a similar spatiotemporal alignment that promoted the assembly and the sustenance of the wild-type strain/ L. plantarum consortium. In contrast to the wild-type, the TOR-silenced G. lucidum strain did not demonstrate the capacity to toggle-switches carbon and nitrogen sources that enabled the wild-type strain self-regulate sugar and amino acid metabolism (Tables S2 & S3). It utilized and exhausted its preferred carbon source and the most abundant amino acids in the medium and failed to transit to other available sugars and amino acids, leading to carbon and nitrogen dearth and cell death.
Electron shuttling molecules and extracellular electron transfer (EET) drive the consortium robustness
Besides the intuitive toggle-switching of carbon and nitrogen sources and multiple metabolites cross-feeding, electron shuttling molecules, and extracellular electron transfer played a crucial role in sustaining the metabolic function of the cells in the consortium. We proposed a schematic flow of nutrients and electrons within the consortium, as presented in Fig. 2a, as driving forces that propelled the consortium assembly and stimulated biomass growth. FDA hydrolysis capacity was used to assess the levels of metabolic activities of the cells during monoculture and coculture. The metabolic activities of the monoculture declined from day 8, while the coculture had a steady increase in the hydrolysis of the colorless FDA into the green fluorescein (Fig. 2b). The intracellular NAD+/NADH, NADP+/NADPH, and ADP+/ATP ratios corroborate the FDA hydrolysis data and prove that the cells had higher metabolic activities during coculture than in the monocultures (Fig. 1c, 1d & 1d). They increased sharply at the early stage of the cultures and remained relatively steady after day 8 in the coculture but drastically declined in monocultures after day 8. The intracellular/extracellular H + ratio shows a similar trend to the pH and indicates that the coculture had formed an intrinsic regulatory framework that ensured a balance of H+int./H2ext. levels in the system (Fig. 1e). It further hints the possible transmembrane protons transfer by the cells’ NADH dehydrogenase and the consequent role of H2 in energy distribution within the consortium. H2, a reduced energy carrier produced via the cells’ metabolic activities, has been reported to contribute to electron transfer between redox couples (Naradasu, Long, Okamoto, & Miran, 2020).
To further understand the thermodynamic driving force that propelled the higher metabolic activity, the extracellular electron transfer system of the cells and the extracellular NAD, NADH, NADPH, FMN, and riboflavin were determined. Electroactive microbes transport electrons from their intracellular electron transport chain to an external electron acceptor and accept electrons from extracellular electron donors via extracellular electron transfer processes (Singh, Chaudhary, Yadav, & Patil, 2022). NAD, NADH, NADPH, FMN, and riboflavin can be transported across plasma membranes of cells and function as extracellular redox molecules (Ying, 2006). The electron transmission system activity (ETSA) correlated with the FDA hydrolysis and the intracellular NAD+/NADH, NADP+/NADPH, and ADP+/ATP ratios. The coculture had the highest ELSA. It increased gradually throughout the incubation, while monocultures declined from day 8 (Fig. 2d). The higher metabolic activities were triggered by increased electron flux, NAD+, NADPH, and ATP regeneration in the cocultured cells. The results indicate that the cells utilized different carbon molecules generated during the fermentation, such as glucose, citrate, acetate, formate, succinate, and fumarate, as electron donors/acceptors. The intracellular NAD+/NADH and NADP+/NADPH provided high-energy electrons for ATP regeneration in the cells, while extracellular NAD, NADPH, FMN and riboflavin served as redox shuttles that facilitated electron transfer between the cells and the external electron acceptors/donors (indirect electron transfer (IET)) (Table S3).
Metal ions and metalloenzymes facilitated energy transduction in the consortium
To gain more insight into the morphological features of the cells during the monocultures and coculture, we obtained the micrograph image of the cells using a scanning electron microscope (SEM). The photomicrograph revealed apparent cell lysis of the G. lucidum monoculture during the late incubation period (GLM14). Conversely, the coculture appeared robust with intertwined structural alignment between the two strains (Fig. 3a). Furthermore, recent studies have indicated that microbial extracellular electron transfer may also occur through minerals that link spatially separated electron donors and acceptors (Gartman et al., 2017). We explored the possible relationships between the cells’ ETSA and the cultures’ mineral composition. We used Energy Dispersive X-ray Spectroscopy (EDS) to obtain a comprehensive profile of electron transduction metallic ions in the cells. The detected minerals were Fe, Cu, Zn, Mo, Mn, Co, Ca, Mg, Na, K, Ti, Cr, Li, and Cs. The % composition of these minerals varied at different stages of the cells’ growth in the GLM, LBM, and GLC (Fig. 3b). The result indicated that the G. lucidum and the L. plantarum uptake several minerals from the medium and might have extensively utilized them to facilitate extracellular electron transfer within the consortium.
To further explore the possible influence of these minerals in enhancing energy transduction and metabolic function of the cells within the consortium, we evaluated the activities of two metalloenzymes (NADH dehydrogenase/ oxidoreductase and sulfite oxidase SO) in the cultures. Fe and Mo were among the major metals detected in the cells (Fig. 3b). The cells had low NADH dehydrogenase and SO activities before being introduced to the medium containing the metal ions (day 1). The results further revealed that the monocultures had the highest NADH dehydrogenase and SO activities at GLM7 and LBM7, while their activities declined drastically at the late incubation stage (GLM14 and LBM14) (Fig. 3c & d). On the contrary, the actions of these enzymes remained relatively stable in the coculture at both early and late stages (GLC7 and GLC14). Notably, the functionality of these enzymes could not be sustained at the late stages of the monocultures despite the presence of metal ions. The observed cell lysis of GLM14 further corroborates the low NADH dehydrogenase and SO activities in the monocultures. We proposed that other factors, such as extracellular and intracellular accumulation of ROS, could have hampered the activities of these enzymes, culminating in low metabolic activity and cell death in the monoculture.
ROS-scavenging enzymes maintained the viability of the cells through intracellular ROS removal
Elevated intracellular ROS levels induce cell oxidative damage and death (R. Liu et al., 2022). We evaluated the ROS levels in the cultures using an H2O2 fluorescent probe (DCFH-DA) to understand its correlation with the culture’s metabolic and enzymatic activity levels. The monocultures produced higher ROS fluorescence intensity than the coculture, which implies higher ROS accumulation. More interestingly, while the fluorescence intensity of the monocultures increased during the incubation stage, the coculture instead showed lower fluorescence intensity which decreased within the incubation time (Fig. 4a & b). The results showed that the monocultures (GLM & LBM) accumulated higher ROS than the coculture (GLC). To further quantify the ROS levels in the culture, the concentration of generated H2O2 was measured. The quantified H2O2 levels correlated with the ROS fluorescence intensity levels in the cultures. The GLM accumulated the highest H2O2 level, followed by LBM, while the GLC had the lowest level of H2O2. We hypothesized that the ROS levels in the cultures might correlate with NADH oxidase, NADPH oxidase, and other antioxidant enzyme activities.
We next probed the possible correlation between the ROS levels in the cultures and catalase (CAT), superoxide dismutase (SOD), glutathione peroxidases (GPX), and ascorbate peroxidase (APX), NADH oxidase, NADPH oxidases activities. Notably, the GLM7 and LBM had higher NADH and NADPH oxidase activities than the coculture. Conversely, the NADH and NADPH oxidase activities declined dramatically at GLM14 and LBM14, while that of the coculture remained relatively stable throughout the coculture. Interestingly, the activities of the antioxidant enzymes followed a reverse trend as the NADH and NADPH oxidase in the cultures. The GLM had the lowest CAT, SOD, and GPX activities. However, the GLM had higher APX activity than the LBM. The GLC had the highest CAT, SOD, GPX, and APX activities (Fig. 1d, 1e, 1f, 1g, 1h, 1i).
Metabolomics analysis reveals higher natural products biosynthesis in the consortium
UHPLC-MS/MS-based untargeted metabolomic analysis was used to comprehensively evaluate the chemical biosynthetic capabilities of the cultures. The UHPLC- MS/MS and HPLC-MS/MS metabolite quantification, the multivariate and KEGG analysis results revealed that the coculture triggered the biosynthesis of several important secondary metabolites more than the monocultures. The volcano plot and the KEGG analysis revealed interesting dynamics in metabolite biosynthesis at GLC7 and GLC14 (Fig. 5a & b). At GLC7, the coculture produced primary metabolites from amino acids and carbohydrate metabolism but switched to the biosynthesis of secondary metabolites at the later stage of the culture (GLC14). The most discriminant metabolites belong to the classes of different therapeutic and nutritional natural products, including triterpenoids, polyols, terpene glycosides, pteridines and derivatives, flavonoids, amino acids, peptides, and analogues, among others. The pharmacological and biotechnological applications of some of the metabolites have been demonstrated in several scientific reports. For instance, the antioxidative activity, neuroprotective potency, and natural anti-cancer effects of salvianolic acid G and cucurbitacin P have been documented (Ma, Tang, & Yi, 2019). Bilobalide is a bioactive natural product with established brain protective and neuroprotective effects and radical scavenging activity (Luo, 2001). Similarly, the anticlastogenic activity of asperulosidic acid has made it a first-choice drug against tumors in Chinese medicine (G. Liu et al., 2001), while the vast pharmacological applications of GA and EPS have been well documented (S. Chen et al., 2012; Gu et al., 2018; Zhou et al., 2014). The HPLC-MS/MS analysis corroborates the metabolomic data and further proved that the coculture stimulated the bioproduction of several natural products with vast therapeutic, nutritional and biotechnolical application (Table S5).
Signal transduction molecules and over-expression of target genes enhanced triterpenoids and exopolysaccharide biosynthesis
The metabolomics and HPLC-MS/MS analysis showed that the coculture produced various therapeutic natural products, including ganoderic acid B and ganoderic acid A (Table S5). To gain a general view of the levels of triterpenoids and exopolysaccharides produced, we quantified the total ganoderic acids (GA) and exopolysaccharides (EPS) produced in the cultures (Fig. 6a and b). The results are in concordance with the metabolomics and HPLC-MS/MS data. It showed that the coculture sample produced higher GA and EPS than the monocultures. The coculture produced 0.9 g/L of GA at GLC7 and 1.6 g/L at GLC14. The GLM produced only 0.4 g/L at GLM14, while the LBM produced no GA. A similar trend was observed for the EPS; the coculture produced 4.8 g/L at GLC14, while the monocultures produced 2.9 and 0.7 g/L for GLM14 and LBM14, respectively. To gain more insight into the intrinsic molecular process that propelled the triterpenoids and exopolysaccharides biosynthesis, we quantified the expression of genes associated with GA and EPS in the cultures. We found that the 3-hydroxy-3-methylglutaryl-CoA reductase (HMGR), squalene synthase (SQS), 2, 3-oxidosqualene-lanosterol cyclase (OSC) genes had higher expression levels in the coculture than in the GLM, with the HMGR gene being the most expressed gene at GLC14 (Fig. 6c). A similar trend was observed in the expression levels of putative EPS genes analyzed. The relative expression levels of genes encoding glucokinase (GL26783), α-phosphoglucomutase (GL24280), UDP-glucose-1-phosphate uridylyltransferase (GL25739), and 1,3-β-glucan synthase (GL24465) was higher in GLC7 and GLC14 than in the monoculture (Fig. 6d).
Further, several studies have reported that signaling molecules play an important role in GA biosynthesis in G. lucidum (Gu et al., 2018). We probe for the possible correlation between the small-signaling molecules (cyclic AMP, salicylic acid, Calcium ion, and pipecolic acid) and GA/EPS biosynthesis in the culture. The four signaling molecules stimulated GA and EPS biosynthesis in monocultures and coculture samples. However, the molecules induced higher GA and EPS production in the coculture than in the monoculture. Notably, the combined effect of the four signaling molecules further increased GA and EPS biosynthesis in the coculture than in the monoculture (Fig. 6e and f). The results further indicate that the addition of Ca2+ induced the highest GA biosynthesis in the coculture while cyclic AMP and salicylic acid produced the highest amount of EPS. The HMGR and the putative 1,3-β-glucan synthase (GL24465) coding sequence were amplified from the G. lucidum cDNA. The sequences were cloned into the pMDTM18-T vector and transformed into G.lucidum. The new G.lucidum-HMRG and GL24465-transformants were cocultured with the L. plantarum. The over-expression of the HMGR gene further increased GA biosynthesis by 3-folds in GLC7 and 5-folds in GLC14 compared to the wild-strain. Similarly, the over-expression of the putative 1,3-β-glucan synthase (GL24465) led to a 2.6-folds and 2.2-folds increase in EPS biosynthesis in GLC7 and GLC14, respectively, compared to the wild-strain.