Abstract Plants, arbuscular mycorrhizal (AM) fungi and soil bacteria coexist stably in ecosystems. Resource availability can affect the mutualistic relationship between plants and AM fungi, but it is unclear how it affects the reciprocal cooperation between AM fungi and bacteria. Here, we used an in vitro culture system containing a source of organic phosphorus (P) to explore the early-stage reciprocal cooperation between the AM fungus Rhizophagus irregularis and the phosphate-solubilizing bacterium (PSB) Rahnella aquatilis under two different levels of inorganic P. Under low inorganic P availability, the reciprocal cooperation between the AM fungus and the PSB, materialized by carbon (C)-P exchange, was strengthened. This was evidenced at transcriptional level by the activation of multiple C metabolism and P transport and regulation pathways of both partners. Conversely, under high inorganic P availability, the exchange of P for C with plants was slowed down. This was accompanied by the activation of only P transport and regulation pathways and less C metabolism pathway, weakening the reciprocal cooperation with the PSB. In conclusion, the availability of inorganic P can mediate the reciprocal cooperation between AM fungi and bacteria, which could also extend to host plants since they are inseparable in ecosystems. Subject terms: Microbiology, Physiology, Fungi __________________________________________________________________ Reciprocal cooperation at the early stage between an arbuscular mycorrhizal fungus and a phosphate-solubilizing bacterium is mediated by external phosphorus availability. Introduction Most natural environments are home to an abundant diversity of microbial species competing with each other for space and resources. They generally display highly social behavior, using chemical signals or shared low-cost resources as a means of communication, cooperation, and competition within the complex network of the ecosystem^[28]1,[29]2. Cooperation is the most effective way of maintaining community stability and improving productivity among conspecifics or heterospecifics. It refers to the interaction in which one individual provides fitness benefits to other individuals and obtains corresponding rewards in exchange^[30]3,[31]4. Cooperation occurs in the world of plants and microbes, where plants feed the complex and diverse soil microbiota through exudates, while the soil microbiota extends the habitat range and metabolic capacity of plants^[32]5,[33]6. An emblematic example of such cooperation is the symbiosis established between arbuscular mycorrhizal (AM) fungi and over two-thirds of plant species. These fungal symbionts increase the access of host plants to minerals (especially phosphorus—P and nitrogen—N), in exchange of carbon (C) resources that are vital for their life cycle^[34]7,[35]8. At more complex levels, AM fungi engage in reciprocal cooperation with soil bacteria, facilitated by their extensive extraradical hyphae and highly branched hyphal networks. These bacteria inhabit the hyphosphere and feed on hyphal exudates and, in exchange, increase the accessibility of minerals for AM fungi and improve their overall fitness^[36]9,[37]10. The cooperation between plants, AM fungi, and soil bacteria has been studied at genetic and evolutionary levels. Up to now, all the sequenced AM fungi (e.g., Rhizophagus irregularis, Rhizophagus clarus, and Gigaspora margarita) have genomes lacking key enzymes for de novo fatty acid biosynthesis and organic substrates degradation^[38]11–[39]13. Therefore, plant-derived fatty acids and sugars support the abundant development of AM fungal extraradical hyphae and the production of spores, enabling extensive exploration of soil nutrients (e.g., P, N), with a beneficial return to the host plants^[40]14,[41]15. However, AM fungi are unable to fully exploit polymer organic matters in soil because they lack most genes encoding carbohydrate-active enzyme (CAZyme)^[42]16. This has led AM fungi to cooperate specifically with hyphospheric bacteria harboring a large number of genes encoding CAZymes in their genome^[43]17,[44]18 rather than cooperating randomly with soil bacteria^[45]19,[46]20. Both plants and AM fungi can improve reciprocal cooperation by rewarding their most compatible partners with increased resource allocation, a process often regulated by host-derived C resources and external nutrients availability^[47]21,[48]22. For example, high P availability tends to limit root colonization by AM fungi and reduces the transfer of C from plants to AM fungi^[49]23–[50]25. Conversely, under low P conditions, plants can boost arbuscule development and derive most benefits from AM fungi^[51]26. If the P demand of plants is low, AM fungi can store the absorbed P within their hyphal network until root demand escalates, thereby ensuring a greater C return^[52]27,[53]28. When extraradical hyphae of AM fungi proliferate in a P-limited environment, they tend to supply a greater proportion of P to plants, as P transport can enable plants to donate more C for fungal growth and metabolism^[54]28,[55]29. In this scenario, the energetic cost of acquiring P from AM fungi may be increased, necessitating the recruitment of soil bacteria to aid in enhancing P accessibility and, in turn, offering C-rich metabolic compounds to reward these bacteria^[56]30. Among the many soil bacteria, phosphate solubilizing bacteria (PSB) are presumably the most reliable associates of AM fungi. They efficiently colonize the surface of extraradical hyphae and promote the mineralization of organic P, which is omnipresent in the soil, into inorganic P that is easily assimilated by AM fungi^[57]31,[58]32. Recently, a study found that AM fungi can exploit C sources to rapidly transport PSB to organic P patches via an extensive hyphal network, thereby facilitating the establishment of cooperative relationships^[59]31. AM fungi and bacteria have broader ecological and agricultural significance, for example, accelerating nutrient cycling, influencing soil C storage and promoting crop production and ecosystem stability^[60]33–[61]35. However, it remains unclear how the availability of external P affects the reciprocal cooperation and synergistic mechanisms between AM fungi and PSB and how these two organisms adapt their behavior in response to changes in P availability in the environment. In the present study, we used an in vitro culture system that comprises a single AM fungus, R. irregularis, and a single PSB, Rahnella aquatilis, to explore biological interactions in the hyphosphere and provide insights into microbial-driven nutrient cycling and potential plant production. The one-to-one model is particularly well-suited for elucidating the potential mechanisms underlying reciprocal cooperation between AM fungi and bacteria, as it allows the harvest of large quantities of hyphae without any contamination. More specifically, we conducted different experiments and transcriptome analysis to address the hypothesis that the cooperation between R. irregularis and R. aquatilis is mediated by the availability of inorganic P: (1) Under the scenario of low inorganic P availabililty, R. irregularis enhances metabolic activity in R. aquatilis and activate the P turnover pathway, thereby accelerating the mineralization of organic P and improving P availability for R. irregularis. (2) Under the scenario of high inorganic P availabililty, R. irregularis directly absorbs P from the environment and stores the excess P in search of more plant C return, while weakening cooperation with the PSB. (3) Under low P availability, the exchange of C and P between R. irregularis and R. aquatilis is accelerated, while under high P availability, this process is slowed down. Results The interaction between R. irregularis and R. aquatilis contributes to the mineralization of organic P We first quantified the concentrations of inorganic P remaining and organic P consumed in the liquid medium of the hyphal compartment (HC), 24 h after introduction of R. aquatilis. The inorganic P concentrations remaining in the P^Low and P^High RI^− RA^+ treatments were equivalent to the initial concentrations of inorganic P (Fig. [62]1a). In contrast, the concentrations of inorganic P in the P^Low or P^High RI^+ RA^− treatments were significantly lower as compared to the initial concentrations as well as to the two other treatments (Fig. [63]1a). In the RI^+ RA^+ treatment, whether under the P^Low or P^High availability, the concentrations of inorganic P remaining in the liquid medium of the HC was intermediate to the RI^+ RA^− and RI^− RA^+ treatments (Fig. [64]1a). Whatever the availability of inorganic P in the HC, the consumption of organic P was significantly higher in the RI^+ RA^+ treatment compared to the two other treatments (Fig. [65]1b). In addition, the organic P in the P^Low RI^+ RA^+ treatment was almost entirely consumed, an amount that was twice as much as that in the P^High RI^+ RA^+ treatment (Fig. [66]1b). Fig. 1. Inorganic P concentration, organic P consumption, ACP and ALP production. [67]Fig. 1 [68]Open in a new tab Concentrations of a inorganic P remaining and b organic P consumed in the MSR liquid medium of the HC under low (35 μM KH[2]PO[4]) and high (700 μM KH[2]PO[4]) P availabilities. The horizontal dashed lines represent the initial concentrations of inorganic and organic P after volume increase due to the introduction of organic P and bacterial suspension. c Acid and d alkaline phosphatase activity in the liquid medium of the HC under low and high P availability. AM fungus-induced e acid and f alkaline phosphatase production efficiency and g organic P mineralization efficiency under low and high P availabilities. The error bar represents the standard errors. For the same P availability, different capital letters indicate a significant difference (p < 0.05 by Tukey’s HSD test, n = 4 biologically independent sample) between the RI^− RA^+, RI^+ RA^− and RI^+ RA^+ treatments^. For the same RI^− RA^+, RI^+ RA^− or RI^+ RA^+ treatment^, the asterisks indicate a significant difference between low and high P availabilities (t-test, *p < 0.05; ***p < 0.001; n.s., p > 0.05, n = 4 biologically independent sample). Abbreviations: RI^− RA^+ only R. aquatilis cells present in the HC, RI^+ RA^−, only R. irregularis hyphae present in the HC, RI^+ RA^+ both R. irregularis hyphae and R. aquatilis cells present in the HC. We then measured the activity of acid phosphatase (ACP) and alkaline phosphatase (ALP) in the P^Low and P^High treatments. Whatever the concentration of inorganic P in the HC, a significant increase in ACP (Fig. [69]1c) and ALP (Fig. [70]1d) activities was observed in the RI^+ RA^+ treatment compared with the RI^+ RA^− and RI^− RA^+ treatments, while no difference was noticed between these two later treatments. In addition, the ACP and ALP activities in the RI^+ RA^+ treatment were significantly higher under P^Low availability compared to P^High availability (Fig. [71]1c, d). R. aquatilis is known to secrete ACP and ALP to mineralize phytate, a function that is not present in R. irregularis. Therefore, we calculated the efficiency of AM fungus-induced ACP and ALP production (EAP and ELP), and the efficiency of AM fungus-induced organic P mineralization (EPM). EAP and ELP were significantly higher at low P availability than at high P availability (Fig. [72]1e, f), which resulted in higher EPM (Fig. [73]1g). The exudates of R. irregularis stimulate the growth of R. aquatilis Increased phosphatase activity can potentially result from an increase in the population of R. aquatilis. In the absence of R. irregularis in the HC (i.e., RI^− RA^+ treatment), no difference was observed in the absolute abundance of R. aquatilis under P^Low and P^High availabilities. Conversely, in the presence of R. irregularis (i.e., RI^+ RA^+ treatment), a significantly higher absolute abundance of R. aquatilis was noticed at P^Low availability compared to P^High availability (Fig. [74]2a). This was consistent with the results obtained with the exudates of R. irregularis grown under the P^Low or P^High availabilities (Fig. [75]2b). The OD[600] of R. aquatilis increased rapidly from 2 to 24 h, followed by a slower rise from 26 to 48 h. Conversely, in the absence of exudates of R. irregularis and at both P^Low and P^High availabilities, the OD[600] of R. aquatilis increased slowly from 1 to 24 h, then remained unchanged from 26 to 48 h. The OD[600] of R. aquatilis in the 0.85% NaCl solution (i.e., the control) remained unchanged over the 48 h of observation (Fig. [76]2b). At each time-point, the OD[600] of R. aquatilis with the exudates of R. irregularis in the P^Low or P^High treatments were significantly greater than that in the other three treatments, with higher OD[600] of R. aquatilis at P^Low availability compared to P^High availability (Fig. [77]2b). Fig. 2. Growth of R. aquatilis in the presence or absence of R. irregularis and its hyphal exudates. [78]Fig. 2 [79]Open in a new tab a Absolute abundance of R. aquatilis 24 h after inoculation in the medium of HC under low (35 μM KH[2]PO[4]) or high (700 μM KH[2]PO[4]) inorganic P availability. The asterisks indicate a significant difference between the RI^+ RA^+ and RI^− RA^+ treatments under the same P availability, and between the RI^+ RA^+ treatments under the different P availabilities (t-test, ***p < 0.001, n = 4 biologically independent sample). b Effect of R. irregularis exudates collected from the HC under low or high inorganic P availability on the growth of R. aquatilis. The data represent optical density (OD) of the culture from 2 to 48 h minus the initial OD of the bacterial liquid cultures. The error bar represents the standard errors. Abbreviations: RI^− RA^+ only R. aquatilis cells present in the HC, RI^+ RA^+ both R. irregularis hyphae and R. aquatilis cells present in the HC, RI^− no R. irregularis hyphae present in the HC, RI^+ only R. irregularis hyphae present in the HC. R. irregularis enhances the C metabolism and cell motility of R. aquatilis C is an essential energy source for bacterial growth and motility. This is confirmed by the significant growth of R. aquatilis in the presence of R. irregularis exudates, whether grown in the P^Low or P^High treatments, compared to the absence of AM fungal exudates. This suggests that R. aquatilis has undergone an attenuation of C limitation. Therefore, we conducted an in-depth analysis of R. aquatilis gene expression and metabolic pathway change at the transcriptional level, in the presence or absence of R. irregularis grown under P^Low or P^High availabilities. A total of 1120 (565 upregulated and 555 downregulated) and 1446 (691 upregulated and 755 downregulated) genes were identified as differentially expressed (DEGs) (adjusted P values < 0.05) in the P^Low and P^High RI^+ RA^+ treatment vs RI^+ RA^− treatment, respectively (Fig. [80]S1a, b). The bubble chart depicted the top 20 functional categories that encompass the highest number of DEGs. The number of DEGs within 14 and 10 functional categories constituted over 40% of the entire pathway at P^Low and P^High availability, respectively (Fig. [81]S2a, b). Among these functional categories, eleven exhibited enrichments at P^Low availability (adjusted P values < 0.05). Specifically, seven were associated with carbohydrate metabolism, one with amino acid metabolism, two with cell motility, and another one with membrane transport (Fig. [82]3a, Supplementary data [83]1). In contrast, only six functional categories were found to be enriched (adjusted P values < 0.05) at P^High availability. Four were associated with carbohydrate metabolism, one with amino acid metabolism, and one with cell motility. Furthermore, all these enriched functional categories exhibited a tendency towards up-regulation (Fig. [84]3b, Supplementary data [85]1). Fig. 3. Functional enrichment and pathway map of C metabolism gene expression in R. aquatilis. [86]Fig. 3 [87]Open in a new tab Enrichment of functional categories (P < 0.05) among differentially expressed genes in R. aquatilis in the RI^+ RA^+ treatment vs RI^− RA^+ treatment under a low (35 μM KH[2]PO[4]) and b high (700 μM KH[2]PO[4]) P availabilities. Each section of the circle plot represents an enrichment pathway. Inner bar plot indicates the log[10] of the adjusted P-value for each enrichment pathway and the color shows the z-score value, indicating global upregulation (the value > 0) of genes within each pathway. Detailed log[2] fold-change (log[2]FC) value of each gene within each pathway is plotted as a dot plot in the outer circle (upregulated in red and downregulated in blue). c Transcriptional responses of key genes in R. aquatilis pathways including sugar transport and metabolism, TCA cycle, chemotaxis, two-component system, and flagellar synthesis and regulation under low and high P availabilities. The heat map shows the log[2] of transcripts per kilobase million (TPM) calculated based on the initial reads count. The dashed lines represent simplified pathways. The asterisks are used to indicate a significant difference in gene expression level between (2) P^Low RI^+ RA^+ and (1) P^Low RI^− RA^+ treatments; between (4) P^High RI^+ RA^+ and (3) P^High RI^− RA^+ treatments; and between (2) P^Low RI^+ RA^+ and (4) P^High RI^+ RA^+ treatments (DESeq2 test, *p < 0.05; **p < 0.01; ***p < 0.001, n = 4 biologically independent sample). Abbreviations: RI^- RA^+ only R. aquatilis cells present in the HC, RI^+ RA^− only R. irregularis hyphae present in the HC, RI^+ RA^+ both R. irregularis hyphae and R. aquatilis cells present in the HC. Based on these results, we generated a pathway heatmap depicting gene expression for sugar transport, C metabolism, and bacterial motility, aiming for more precise insights into key metabolic processes of R. aquatilis. The expression of several sugar transporter genes, specifically those encoding for trehalose (treB), glucose (crr), maltose/glucose (malX), and fructose (fruB) transport, was significantly upregulated in the P^Low and P^High RI^+ RA^+ treatments (Fig. [88]3c, Supplementary data [89]2). The sugar assimilated by R. aquatilis was first metabolized via the glycolysis pathway, and then, following several conversion steps, produced acetyl-CoA, which subsequently entered the tricarboxylic acid (TCA) cycle to produce energy. Among them, the expression of multiple genes associated with the glycolysis pathway, such as glucokinase (glk), fructokinase, 6-phosphofructokinase, and fructose-bisphosphate aldolase (FBA), were significantly upregulated in the P^Low and P^High RI^+/RA^+ treatments. R. irregularis specifically stimulated genes encoding key enzymes in the TCA cycle at P^Low availability (Fig. [90]3c, Supplementary data [91]2). This includes bifunctional aconitate hydratase 2/2-methylisocitrate dehydratase (ancB), E1 and E2 components of 2-oxoglutarate dehydrogenase (OGDH and DLST), and alpha subunit of succinyl-CoA synthetase. In addition, some genes involved in sugar metabolism and TCA cycle, such as trehalose 6-phosphate phosphatase (otsB), isocitrate dehydrogenase (IDH1) and succinate dehydrogenase flavoprotein subunit (frdA), whose expression was stimulated by R. irregularis, showed higher expression levels at P^Low availability compared to P^High availability. The chemotactic two-component system and flagellar assembly were pivotal in governing the motility of R. aquatilis. The expression of numerous genes within these pathways, such as cheA, cheB, cheV, cheY, flhD, flgN, motA, and motB, was significantly upregulated in the P^Low or P^High RI^+/RA^+ treatments. Particularly, many genes of these pathways exhibited higher expression levels at P^Low availability compared to P^High availability (Fig. [92]3c, Supplementary data [93]2). The presence of R. irregularis enhances the P turnover function in R. aquatilis To explore whether R. aquatilis has an important contribution to phosphatase activity in liquid medium, we analyzed at the transcriptional level the genes associated with P turnover and calculated their variance-stabilized count abundance (Fig. [94]4a, Supplementary data [95]3). The abundance of six genes inducing P mineralization, eight involved in P transportation, and four responsible for P regulation were significantly increased in the P^Low RI^+ RA^+ treatment vs RI^+ RA^− treatment. In contrast, only five genes involved in P mineralization, four in P transport, and two in P regulation showed a significantly increased abundance in the P^High RI^+ RA^+ treatment vs RI^+ RA^- treatment. Furthermore, R. aquatilis reduced the variance stabilized count abundance of most P solubilization genes at both P^Low and P^High availability (Fig. [96]4a). The regulation genes (i.e., phoB, phoP, phoQ, and phoR) of the two-component system were responsible for sensing nutrient changes in the environment and the stimulation of AM fungi, thereby activating or inhibiting the expression of other P turnover functional genes. The variance stabilized count abundance of P regulation genes was significantly positively correlated with genes involved in P mineralization (R^2 = 0.57, P = 0.0008) and in P transport (R^2 = 0.40, P = 0.0090), but negatively correlated with genes involved in P solubilization (R^2 = 0.41, P = 0.0072) (Fig. [97]4b). Fig. 4. Expression levels of genes involved in P cycling function in R. aquatilis. [98]Fig. 4 [99]Open in a new tab a Heat map showing variance stabilized counts of P functional genes in R. aquatilis: P solubilization, mineralization, transport, and regulation pathways. The eight lines above represent variance-stabilized counts of related genes in the RI^− RA^+ and RI^+ RA^+ treatments under low (35 μM KH[2]PO[4]) P availability. The eight lines below represent variance-stabilized counts of related genes in the RI^− RA^+ and RI^+ RA^+ treatments under high (700 μM KH[2]PO[4]) P availability. The red circles and blue circles indicate that variance-stabilized counts were significantly increased and decreased in the RI^+ RA^+ treatment compared with the RI^− RA^+ treatment. b Spearman correlation between P regulation genes and P solubilization genes, mineralization genes, and transport genes based on the calculation of variance stabilized counts. c Transcriptional responses of genes in R. aquatilis P solubilization, mineralization, transport, and regulation pathways under low and high P availability. The genes with higher expression level in the P^Low RI^+ RA^+ treatment than in the P^High RI^+ RA^+ treatment were selected and log[2] of transcripts per kilobase million (TPM) were calculated based on the initial reads count. The asterisks indicate a significant difference in gene expression level between P^Low RI^+ RA^+ and P^Low RI^− RA^+ treatments; between P^High RI^+ RA^+ and P^High RI^− RA^+ treatments and between P^Low RI^+ RA^+ and P^High RI^+ RA^+ treatments (DESeq2 test, *p < 0.05; **p < 0.01; ***p < 0.001; n.s., p > 0.05, n = 4 biologically independent sample). Abbreviations: RI^− RA^+ only R. aquatilis cells present in the HC, RI^+ RA^− only R. irregularis hyphae present in the HC, RI^+ RA^+ both R. irregularis hyphae and R. aquatilis cells present in the HC. The EAP, ELP, and EPM were significantly greater at P^Low availability compared to P^High availability, potentially attributable to the increased expression of related genes regulating or encoding phosphatase proteins. The expression of eleven genes involved in P mineralization, transport and regulation were significantly higher in the P^Low RI^+ RA^+ treatment (Fig. [100]4c). Among them, four genes encode phosphatase proteins and four regulate phosphatase synthesis and transport, namely inositol phosphatase gene phy, acid phosphatase gene phoA, alkaline phosphatase genes phoN1 and phoN2, P regulon response regulator phoB, P sensor histidine kinase phoQ, P regulon sensor histidine kinase phoR, and transcriptional regulator phoU (Supplementary data [101]3). Low inorganic P availability stimulates R. irregularis to absorb and transport P to the host plant while enhancing its C and energy metabolism The sustained growth of R. aquatilis by R. irregularis at P^Low availability, can be attributed to an increase in C metabolism, deriving from altered P needs of R. irregularis. Therefore, we conducted a thorough analysis of R. irregularis gene expression patterns at the transcriptional level. In the presence of R. aquatilis, R. irregularis exhibited 449 (373 upregulated and 76 downregulated) and 347 (255 upregulated and 92 downregulated) DEGs (adjusted P values < 0.05) in the P^Low and P^High RI^+ RA^+ treatment vs RI^− RA^+ treatment, respectively (Fig. [102]S3a, b). Seven functional categories were found to be enriched (adjusted P values < 0.05) at P^Low availability. Specifically, two were associated with carbohydrate metabolism, three with lipid metabolism, two were involved in amino acid metabolism, and another one was involved in transport and catabolism. Conversely, only one functional category associated with amino acid metabolism was enriched (adjusted P values < 0.05) at P^High availability (Fig. [103]S4a, b, Supplementary data [104]4). AM fungi can change P transfer to host plants based on inorganic P availability in the environment. The expression of several P-absorbing protein genes (i.e., pho84 and pho89) located on the membrane of extraradical hyphae and P transporter genes (i.e., vtc1 and vtc4) located on the membrane of vacuoles were significantly upregulated in the P^Low or P^High RI^+ RA^+ treatment, and pho84, pho89, and vtc1 genes had higher expression levels at P^Low availability compared to P^High availability (Fig. [105]5a, Supplementary data [106]5). Once entering the vacuole, P was converted into Poly-P and subsequently transferred with arginine to intraradical hyphae, where it underwent further hydrolysis and transport to the host. The expression of multiple genes, such as argG, argH, argI, ypq2, ppn, and pho91, associated with arginine synthesis, Poly-P degradation and the subsequent transport of P was only induced in the P^Low RI^+ RA^+ treatment. In addition, at P^Low availability, the positive regulator pho4 in the PHO regulatory system was stimulated, while the negative regulator pho85 was inhibited (Fig. [107]5a, Supplementary data [108]5). Fig. 5. Pathway map of C and P metabolism gene expression in R. irregularis. [109]Fig. 5 [110]Open in a new tab Transcriptional responses of genes in R. irregularis a P transport and regulation and arginine biosynthesis and b fatty acid elongation and degradation, energy metabolism, TCA cycle, gluconeogenesis and glycolysis pathways, under low (35 μM KH[2]PO[4]) and high (700 μM KH[2]PO[4]) P availabilities. The heat map shows the log[2] of transcripts per kilobase million (TPM) calculated based on the initial reads count. The dashed lines represent simplified pathways. The asterisks indicate a significant difference in gene expression level between (2) P^Low RI^+ RA^+ and (1) P^Low RI^+ RA^− treatments; between (4) P^High RI^+ RA^+ and (3) P^High RI^+ RA^− treatments; and between (2) P^Low RI^+ RA^+ and (4) P^High RI^+ RA^+ treatments (DESeq2 test, *p < 0.05; **p < 0.01; ***p < 0.001, n = 4 biologically independent sample). Abbreviations: RI^− RA^+ only R. aquatilis cells present in the HC, RI^+ RA^− only R. irregularis hyphae present in the HC, RI^+ RA^+ both R. irregularis hyphae and R. aquatilis cells present in the HC. AM fungi transfer P to the host plant in return for fatty acids, which may result in enhanced C and energy metabolism, thereby strengthening the reciprocal cooperation between AM fungi and bacteria. The expression of multiple genes involved in R. irregularis fatty acid elongation and degradation, TCA cycle and gluconeogenesis/glycolysis was significantly upregulated in the P^Low RI^+ RA^+ treatment, mainly including fatty acid elongase 2, long chain-3-oxoacyl-CoA reductase (HSD17), long chain-3-hydroxyacyl-CoA dehydratase, acetyl-CoA C-acetyltransferase (ACAT), citrate synthase, OGDH, enolase, phosphoenolpyruvate carboxykinase (pckA), FBA, fructose-1,6-bisphosphatase I (FBP), glk. In contrast, the expression of only a few genes, such as HSD17, ACAT, and IDH1, was stimulated in the P^High RI^+ RA^+ treatment (Fig. [111]5b, Supplementary data [112]5). Furthermore, several genes, specifically ATPeF0B, ATPeF1G, and ATPeV0D, which are crucial in energy metabolism, had their expression exclusively induced in the P^Low RI^+ RA^+ treatment (Fig. [113]5b, Supplementary data [114]5). Low inorganic P accelerates C-P reciprocity between R. irregularis and R. aquatilis To further explore the cooperation strategies employed by R. irregularis and R. aquatilis at varying P availability, we calculated the P uptake ratio (PUR) of R. irregularis and the C acquisition ratio (CAR) of R. aquatilis to characterize the C-P exchange rate. The results showed that under P^Low availability, R. irregularis had a PUR 4.78 times higher than that under P^High availability (Fig. [115]6a). Under P^Low availability, R. aquatilis also acquired C more efficiently (i.e., 1.91 times higher) compared to P^High availability (Fig. [116]6b). Fig. 6. R. aquatilis C acquisition ratio and R. irregularis P uptake ratio. [117]Fig. 6 [118]Open in a new tab a R. aquatilis CAR and b R. irregularis PUR under low (35 μM KH[2]PO[4]) and high (700 μM KH[2]PO[4]) P availabilities. The asterisks indicate a significant difference between low and high P availabilities (t-test, ***p < 0.001, n = 4 biologically independent sample). Discussion Plants, AM fungi, and bacteria cooperate intimately by metabolic exchanges or cross-feeding to ensure the stable development of all the players^[119]36. While it is well established that resource availability affects the cooperative relationship between plants and AM fungi^[120]28,[121]29,[122]37, the way in which such cooperation is regulated between AM fungi and bacteria in the hyphosphere has not yet been fully elucidated. Here, we studied at the ecological and molecular levels how the availability of inorganic P in the environment impacts the mechanisms of cooperation at the early stage between an AM fungus and a PSB. The results confirmed our hypothesis that low inorganic P availability (35 μM KH[2]PO[4]) stimulates the reciprocal cooperation between the AM fungus and the bacterium, accompanied by an increased acquisition of C by the bacterium and an enhanced uptake of P by the AM fungus. Conversely, under high inorganic P availability (700 μM KH[2]PO[4]), the AM fungus directly absorbs and stores excess P, slowing down the exchange of C with plants and weakening the cooperation with the PSB. These findings contribute to elucidating the mechanisms underlying cross-kingdom interactions facilitated by resource availability. It may establish a foundation for comprehensively understanding microbial-mediated soil nutrient cycling, enhancing plant productivity, and optimizing P fertilizer application strategies to safeguard soil health. Low inorganic P availability can strengthen cooperation between the AM fungus and the PSB, based on C and P exchange In previous studies^[123]30–[124]32,[125]38, the existence of reciprocal cooperation between the AM fungus R. irregularis and the PSB R. aquatilis, was demonstrated based on the exchange of C and P. This cooperation may result from the functional complementarity of the metabolism of the two partners as well as the scarcity of resources available in the soil. In the absence of exploitable C sources, bacteria typically enter a dormant state, a temporary adaptation that slows down metabolism to reduce energy expenditure and sustains viability during a prolonged period of C limitation^[126]39,[127]40. In contrast, a dense network of hyphae is known to release significant amounts of C-containing compounds, which are mainly composed of low molecular weight sugars, amino acids, and carboxylates (see Table [128]S1). Sugars, in particular glucose, fructose, and trehalose, have been shown to promote the growth of soil bacteria, so that they emerge from dormancy and develop rapidly^[129]9. This is consistent with our findings showing that the expression of genes encoding glucose, fructose, and trehalose transporters on the cell membrane of R. aquatilis was significantly upregulated in the presence of R. irregularis. In addition, the growth of the bacterial population was increased in the presence of either the AM fungus or only its hyphal exudates: the increase was 21% and 75% respectively (Fig. [130]2). Cooperation requires both parties to engage in an exchange of resources aimed at achieving greater benefits at minimum cost^[131]3,[132]41. The AM fungus provides the PSB with easily accessible sources of C and, in exchange, the PSB provides the AM fungus with more important sources of inorganic P^[133]30,[134]42. This is further supported by the increase in phosphatases secreted by R. aquatilis in the presence of R. irregularis, which accelerates the mineralization of organic P, thus increasing the availability of inorganic P for the AM fungus (Fig. [135]1). This scenario is, however, dependent on the availability of P in the environment. Under a high inorganic P scenario, we observed that R. irregularis was able to directly assimilate inorganic P from the medium via P transporters (e.g., Pho80, Pho84, and Pho89) that belong to the phosphate transporter 1 family, located on the plasma membranes of extraradical hyphae^[136]43–[137]46. This process may consume less energy and costs than cooperating with the PSB for the acquisition of inorganic P from organic sources. Conversely, under low inorganic P scenario, the AM fungus was more likely to cooperate with the PSB to increase the metabolic exchange of C and P, improving its overall ability to adapt to the environment (Fig. [138]1 and Fig. [139]2). At the transcriptional level, we confirmed that, as well as acting as a source of C to stimulate PSB growth and activate its glycolysis, TCA cycle and other C metabolic pathways, hyphal exudates may also contain signaling molecules and chemotactic substances that induce flagellum synthesis and cell motility in the bacterium, further promoting the formation of reciprocal cooperation (Fig. [140]3). The two-component signaling system is a ubiquitous and conserved regulatory component in bacteria that can specifically recognize and sense external environmental signals and induce specific metabolic responses by activating or inhibiting gene expression in downstream pathways^[141]47. Higher gene expression levels of C metabolism and cell motility mediated by sensor and response regulatory receptor, as well as more enriched metabolic pathways, were observed in R. aquatilis under low versus high inorganic P availability (Fig. [142]3). Bacteria are the main drivers of the P cycle in the soil, facilitating the solubilization of non-soluble P and the mineralization of organic P, and participating in the transport and regulation of P^[143]48–[144]50, while AM fungi are mainly responsible for P uptake, storage, and transport, owing to their limited saprophytic capabilities^[145]16,[146]51. In the present study, we identified potential regulatory trade-offs of R. aquatilis in solubilizing non-soluble P and mineralizing organic P. Specifically, in the presence of organic P as the available P source in addition to inorganic P, R. aquatilis stimulated the organic P mineralization pathway at the same time suppressing the non-soluble P solubilization pathway via the P regulatory pathway (Fig. [147]4) as earlier observed^[148]47. This suggests that the PSB can accurately identify the form of P in the environment and make strategic adjustments to effectively cooperate with AM fungi. The exchange of C and P leads to the reciprocal cooperation between AM fungi and bacteria, which is a common mode of cooperation, widely documented in the literature^[149]20,[150]52–[151]54. In the hyphosphere, bacteria from multiple orders, such as Myxococcales, Betaproteobacteriales, Fibrobacterales, Cytophagales, and Chloroflexales, interact closely with AM fungi and form the hyphobiome^[152]52,[153]53. These bacteria thrive in the hyphosphere, dominate the soil bacterial community, and secrete phosphatases that enhances the availability of soil P to AM fungi^[154]20,[155]53. Inorganic P availability alters the P delivery and C acquisition of the AM fungus with the host to enhance or weaken cooperation with the PSB Upon uptake by the extraradical hyphae of AM fungi, inorganic P is first polymerized into Poly-P, then accumulated in tubular vacuoles or transported with arginine to the intraradical hyphae, where it is transferred from fungal cell to plant cell^[156]55. AM fungi are able to regulate P homeostasis and control P transfer to the host via the PHO regulatory pathway^[157]56,[158]57, depending on the nutritional status of the host and the availability of inorganic P in the environment, to maximize the return of C^[159]27,[160]28. In the present study, we found that high inorganic P availability allowed R. irregularis to directly absorb large amounts of P from the environment (Fig. [161]1). The AM fungus may accumulate the inorganic P rather than transport it to the host, as the expression of core regulatory factor Pho4, the hydrolase genes involved in the degradation of Poly-P, and multiple genes responsible for arginine synthesis were not increased in the presence or absence of R. aquatilis (Fig. [162]5). On the contrary, low inorganic P availability enhanced reciprocal cooperation between the AM fungus and the PSB, allowing R. irregularis to efficiently acquire P from the environment and transfer it to the host when resources were scarce. This process was accompanied by increased gene expression of AM fungal C and energy metabolism, suggesting that the AM fungus may obtain more C returns (Fig. [163]5). In ecosystems, plants, AM fungi and their associated bacteria can drive top-down C flows as well as bottom-up nutrient flows^[164]36. The accelerated P transfer and enhanced C metabolism of R. irregularis under low inorganic P availability are likely to furnish more C resources for R. aquatilis and induce this PSB to accelerate the mineralization of organic P (Figs. [165]2, [166]4, and [167]5). This ultimately improves the PUR of the AM fungus and the CAR of the PSB, forming a cross-kingdom closed loop of C and P exchange (Fig. [168]6). Unlike fatty acids delivered by plants to AM fungi, sugars are generally thought to be passively exuded by AM fungi and acquired by PSB (Table [169]S1), which may represent the surplus of nutrients for AM fungi^[170]58,[171]59. However, in the present study, we found that the continuous stimulation of PSB growth by hyphal exudates under low inorganic P availability is attributed to the transfer of P from AM fungi to the host, thereby enhancing C transfer from plant to fungus and increasing energy metabolism in the AM fungus (Figs. [172]2, [173]5). This finding suggests that the exudates released by the hyphae may not be a purely passive process, but a conscious effort by AM fungi and presumably the host to quest reciprocal cooperation with PSB, which support the reciprocity theory that both partners need to proactively provide resources^[174]60,[175]61. What are the significances of cooperation between AM fungi and PSBs for agricultural management? Increasing the yield of food systems is highly reliant on mineral fertilizers, whose production is energy-intensive, contributes massively to global greenhouse gas emissions, and, in the case of P, is dependent on finite resources^[176]62. In farmland soil, many crops have low P fertilizer use efficiency, resulting in low recovery of applied fertilizer^[177]63. Nowadays, there is a growing interest to reduce our agricultural footprint and reliance on chemical fertilizers and pesticides through the use of microbial inoculants^[178]64. AM fungi and their associated bacteria, among which the PSBs, are an important component of agricultural systems. They modify soil structure, mineralize organic matter, and improve the uptake and transfer of nutrients and water to the plants, thereby increasing crop productivity^[179]36,[180]65. They represent potential tools in increasing the sustainability of our food production systems. However, the results are often lagging behind the expectations for many reasons attributed to the microorganisms, as well as the environmental conditions and agricultural practices. For instance, under intensive agricultural systems, a reduction in the diversity of these fungi and bacteria and an inhibition of their ecological functions have been observed and attributed partly to the wide application and accumulation of P fertilizers^[181]66,[182]67. Consequently, exploiting beneficial soil microbial communities to increase food production while minimizing the use of mineral P fertilizer and preserving non-renewable soil resources is an essential objective of sustainable agricultural development. There is thus a crucial need to better understand how AM fungi and their associated PSBs cooperate under conditions of low to high P concentration. For instance, a recent study on maize has reported that in the presence of a well-established AM fungal community, the application of P fertilizer can be reduced by 30 kg per hectare^[183]68. Maize can fully exploit the potential of AM fungi to obtain P under limited soil P supply conditions^[184]25. In the present study, we demonstrated that cooperation between the PSB R. aquatilis and the AM fungus R. irregularis accelerated the mineralization of organic P when inorganic P availability was low, a process that may give an indication of the reuse of organic P in the soil. Therefore, optimizing the application of mineral fertilizers to increase crop yields based on the mutualistic cooperation strategies between PSBs and AM fungi, as well as AM fungi and plants, may represent a promising strategy for achieving sustainable agricultural development^[185]34,[186]69. Conclusion Under low inorganic P availability, the reciprocal cooperation between the two partners, materialized by C-P exchange, was strengthened, accompanied by the activation of multiple C metabolism and P transport and regulation pathways. Conversely, under high inorganic P availability, the exchange process and metabolism involved were less solicited, and only glycolysis and phosphatase synthesis pathways of the PSB as well as P transport and regulatory pathway of the AM fungus were activated (Fig. [187]7). Therefore, P availability serves as a mediator for the reciprocal cooperation between AM fungi and soil bacteria. However, the existing understanding is confined solely to reciprocal cooperation between two individual species, and the extent to which plant C allocation in roots can facilitate mutualistic interactions between multiple AM fungi and soil bacteria remains an unresolved question. Fig. 7. Schematic representation of reciprocal cooperation based on C and P exchange between R. irregularis and R. aquatilis. [188]Fig. 7 [189]Open in a new tab Schematic representation of reciprocal cooperation based on C and P exchange between the arbuscular mycorrhizal fungus R. irregularis and the phosphate solubilizing bacterium R. aquatilis under under a low (35 μM KH[2]PO[4]) and b high (700 μM KH[2]PO[4]) P availabilities. A series of pathways for C metabolism and transport (green arrows) as well as P metabolism and transport (purple arrow) in the AM fungus and the bacterium are described. Pathways with enrichment and upregulated are highlighted in red. Pase, phosphatase; TCA cycle, tricarboxylic acid cycle. Although root organ cultures (ROC) is a well-established and widely recognized model system, with nutritional, metabolic, and resource transfer characteristics close to those of whole plant systems^[190]19,[191]21,[192]27, it presents some constraints to fully understand the dynamics of nutrient exchange that occur within ecosystems. In the invisible environment of the soil, multiple fungi and bacteria coexist in micro-ecological niches created by plants, maintaining a microbial network of interactions and interdependencies^[193]70. The role of individual species in the community and the cooperative mechanism between species need to be verified using more realistic potted and field systems. Consequently, future studies should focus on extending the influence of external resources on the strategies and mechanisms of cooperation between two individual species to the whole plant-AM fungi-bacteria community, which is necessary if researchers are to explore the complex ecological and evolutionary implications of plant-microbe social relationships. Materials and methods Biological material The AM fungus used was R. irregularis (Błaszk., Wubet, Renker & Buscot) C. Walker & A. Schüßler as “irregulare” MUCL 43194. It was isolated from the roots of Fraxinus americana L. (American ash) in Pont Rouge, Québec, Canada^[194]71 and maintained in bi-compartmented Petri plates on ROC of Daucus carota L. clone DC2 on the Modified Strullu-Romand (MSR) medium^[195]72. The PSB used was R. aquatilis HX2. It was isolated from the rhizosphere soil of a vineyard located in Beijing, China^[196]73. The plant Medicago truncatula L. (barrel clover) Gaertn. cv. Jemalong A17 was used for experiments. In vitro culture system design The AM fungus was grown with ROC of D. carota or M. truncatula in bi-compartmented Petri plates, which is the most reliable method for generating substantial quantities of extraradical hyphae without any microbial contaminant. Briefly, in one compartment (i.e., the root compartment—RC), a carrot or barrel clover root was associated with the AM fungus, while in the other compartment (i.e., the hyphal compartment—HC) only the extraradical hyphae of the AM fungus were allowed to proliferate (Fig. [197]S5a, b). 25 mL of sterilized (121 °C for 15 min) MSR medium solidified with 3 g L^−1 Phytagel (Gelzan^TM CM, G3251, CP Kelco, USA) was added in the RC. To facilitate the passage of extraradical hyphae from the RC to the HC, 4 mL of the same medium lacking sucrose, inorganic P, EDTA, calcium, and vitamins was added to create a slope from the top of the HC to the bottom of the HC next to the barrier. In addition, 10 mL of liquid MSR medium with the same composition as the slope (except for the absence of gelling agent) was added to the HC. A control treatment was also included under identical conditions, except that a 0.5 cm wide strip of MSR medium was excised from the RC next to the barrier to prevent the passage of hyphae into the HC. Experiment 1: impact of inorganic P availability on the mechanism of cooperation between R. irregularis and R. aquatilis The experiment was conducted in an in vitro bi-compartmented culture system on MSR medium. Briefly, ROCs of M. truncatula were associated with R. irregularis MUCL 43194 in the RC, while only the hyphae were allowed to develop on a slope in the HC. After hyphae had crossed the partition wall separating the RC from the HC and densely covered the slope, 10 mL of liquid MSR medium with the same composition as the slope (except for the absence of gelling agent) was added to the HC. After another two weeks, the HC was covered with numerous actively growing extraradical hyphae. At this point, the medium was removed and replaced with 10 mL of fresh liquid MSR medium containing two different concentrations of P. In half of the experimental systems, 35 μM KH[2]PO[4] was added, representing a low P concentration in soil solution and corresponding to the low inorganic P availability treatment, while in the other half, 700 μM KH[2]PO[4] was added, representing a high P concentration in soil solution and corresponding to the high inorganic P availability treatment^[198]21. Prior to this change, the HC of the experimental systems were gently rinsed three times with sterilized (121 °C for 15 min) deionized water. The inoculum of R. aquatilis was prepared as follows: The bacterium was cultured in liquid Luria–Bertani (LB) medium with shaking at 180 rpm at 28 °C until optical density (OD)[600] reached 0.4–0.6 (logarithmic phase). Then it was centrifuged at 5878 g for 10 min. The supernatant was removed, and the pellet was re-suspended and washed three times with inorganic P-free liquid MSR medium. The supernatant was subsequently adjusted to OD[600] = 0.6 and stored at 4 °C before use. One week later, the remaining liquid medium in the HC was collected and mixed homogeneously with 3 mL of R. aquatilis supernatant (concentration of ~10^8 CFUs mL^−1) and 1 mL of 280 μM phytate (Na-phytate, Sigma-Aldrich). The solution was then carefully returned to the HC of each experimental system using a micropipette. 24 h later, R. irregularis hyphae and R. aquatilis cells, as well as the remaining liquid medium, were collected. To satisfy the sequencing requirements for the AM fungus and PSB, each set of three experimental systems was pooled into one. To collect R. aquatilis cells, the culture medium was carefully agitated with a micropipette, and a 9 mL aliquot of medium containing cells was removed using a micropipette and immediately stored in liquid nitrogen for R. aquatilis absolute abundance and transcriptome analysis. Likewise, R. irregularis hyphae were carefully picked out with sterilized forceps, placed into a 2 mL tube, and immediately frozen in liquid nitrogen for transcriptome analysis. The remaining liquid medium was filtered through an Acrodisc Syringe Filter (0.2 μm Supor Membrane, Pall Corporation, Cornwall, UK) to eliminate bacterial cells, and stored at −20 °C for P concentration and phosphatase activity analysis. In total 6 treatments were considered, namely, the experimental systems with (1) only R. irregularis hyphae developing in the HC under high and low inorganic P availability (i.e., the P^High RI^+ RA^− and P^Low RI^+ RA^− treatments), (2) only R. aquatilis cells in the HC under high and low inorganic P availability (i.e., the P^High RI^− RA^+ and P^Low RI^− RA^+ treatments), and (3) both R. irregularis hyphae and R. aquatilis cells in the HC under high and low inorganic P availability (i.e., the P^High RI^+ RA^+ and P^Low RI^+ RA^+ treatments) (Fig. [199]S5a, b). Each treatment was replicated four times, resulting in a total of 72 experimental systems. RNA isolation, library preparation, and sequencing The total RNA was extracted from the frozen hyphae of R. irregularis and cells of R. aquatilis using the RNeasy Plant Mini Kit and RNeasy Mini Kit (QIAGEN GmbH, Hilden, Germany), respectively. The concentration and integrity of the isolated RNA were determined using NanoDrop-ND 1000 UV–Vis Spectrophotometer, after which the samples were shipped to Novogene Co., Ltd (Cambridge, UK) for library preparation and sequencing. For the AM fungus, Poly-T oligo-attached magnetic beads were used to isolate mRNA from the total RNA. The process involved synthesizing two strands of cDNA using random hexamer primers dUTPs. In contrast, for the PSB, rRNA was first removed from the total RNA, and then the remaining RNA was subjected to ethanol precipitation. During synthesis of the second cDNA strand, dTTPs were replaced by dUTPs in the reaction buffer. The directional library was prepared following end repair, A-tailing, adapter ligation, size selection, USER enzyme digestion, amplification, and purification. It was then assessed using Qubit and real-time q-PCR for quantification and a bioanalyzer for size distribution analysis. The quantified libraries were subsequently pooled and sequenced on Illumina platforms, in accordance with their effective concentrations and the desired data output. Bioinformatics analysis of R. irregularis and R. aquatilis at the transcriptional level The raw data (raw reads) were first processed by an internal Perl script. Clean data (clean reads) were obtained by removing reads containing adapter, reads containing ploy-N and low-quality reads. All subsequent analyses were based on these high-quality clean data. For the AM fungus, Hisat2 v2.0.5 was used to build the reference genome index and align the paired clean reads to the reference genome^[200]74. For PSB, building the reference genome index and aligning the clean reads to the reference genome were done using Bowtie2^[201]75. FeatureCounts v1.5.0-p3 was used to calculate the number of aligned reads for each gene. Subsequently, the transcripts per kilobase million (TPM) for each gene was calculated, considering the gene length and the count of reads mapped to that gene^[202]76. As a quantitative metric of gene expression, TPM is currently the most used method for comparing gene expression levels across different groups^[203]77. The differential expression analysis was done comparing P^Low RI^+ RA^+ versus P^High RI^+ RA^+ treatments, P^Low RI^+ RA^+ versus P^Low RI^+ RA^− treatments, P^High RI^+ RA^+ versus P^High RI^+ RA^− treatments, P^Low RI^+ RA^+ versus P^Low RI^− RA^+ treatments, P^High RI^+ RA^+ versus P^High RI^− RA^+ treatments, using the DESeq2 R package. DESeq2 proposes statistical methods designed to identify DEGs in digital gene expression datasets, using a model based on the negative binomial distribution. The P-values obtained were subsequently adjusted using the Benjamini and Hochberg method to control the false discovery rate^[204]78. The genes identified by DESeq2 with an adjusted P-value of less than 0.05 were considered to be differentially expressed. The Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis of DEGs was conducted using the clusterProfiler R package. Pathways from KEGG with a corrected P-value of less than 0.05 were considered to be enriched by the DEGs^[205]79. The KEGG pathway for the non-model species R. irregularis was annotated by referencing the model fungus Saccharomyces cerevisiae (budding yeast). DNA isolation and real-time q-PCR analysis of the absolute abundance of R. aquatilis DNA was extracted from the frozen cells of R. aquatilis using the QlAprep Spin Miniprep Kit (QIAGEN GmbH, Hilden, Germany). Each sample was analyzed in triplicate via real-time q-PCR using a Roche LightCycler^® 96 System with R. aquatilis specific primers (atpD-F, 5’-GATGTGTCCGTTCGCTAA-3’ and atpD-R, 5’-GATAACCTACCGCTGATGG-3’). Each q-PCR reaction was conducted in a total volume of 20 μL, comprising 2 μL of DNA, 10 μL of Master mix, 0.5 μL of each primer, and 7 μL of ddH[2]O. The following thermal cycling conditions were applied: 95 °C for 300 s, 40 cycles at 95 °C for 10 s, 60 °C for 30 s and 72 °C for 45 s. The melting curve, baseline range, and cycle threshold (Ct) values were monitored and calculated. The number of copies was calculated using the Ct values, which are directly proportional to the biomass of R. aquatilis in the samples. This was done through the application of standard curves, which was established in triplicate using five serial 10-fold dilutions. The number of copies was then determined by identifying the initial copy number, factoring in the plasmid concentration and the aggregate number of base pairs (encompassing both vector and primer sequences). Following this, the DNA isolation efficiency was calculated by referencing the Ct values from the standard curves and normalizing the relative abundance of PSB to account for the DNA isolation efficiency. Quantification of organic and inorganic P concentrations and ACP and ALP activities in the HC Inorganic P concentration in the liquid medium of the HC was quantified using the molybdate-blue method^[206]80. Total P concentration was assessed through inductively coupled plasma atomic emission spectroscopy. Organic P (i.e., Na-phytate) concentration was calculated by subtracting the inorganic P concentration from the total P concentration. ACP and ALP activities in liquid medium (pKatal mL^−1) were measured according to the method of Neumann (2006)^[207]81. Evaluation of phosphatase production efficiency, organic P mineralization efficiency, C acquisition ratio by R. aquatilis, and P uptake ratio by R. irregularis EAP, ELP and EPM were calculated by Eqs. [208]1 and [209]2, respectively: [MATH: EAPELP=ZRI+RA+ ZRIRA+ :MATH] 1 [MATH: EPM=C< /mi>RI+RA +CRIRA+ :MATH] 2 (Z [RI+ RA+], enzyme activity detected in the RI^+ RA^+ treatment; Z [RI− RA+], enzyme activity detected in the RI^− RA^+ treatment; C [RI+ RA+], the consumption of organic P in the RI^+ RA^+ treatment; C [RI− RA+], the consumption of organic P in the RI^− RA^+ treatment) The R. aquatilis CAR and R. irregulars PUR were calculated by Eqs. [210]3 and [211]4, respectively: [MATH: PUR= CRI+ RA+C RIRA+CRI+RA :MATH] 3 [MATH: CAR=X< /mi>RI+RA +XR< /mi>IRA+XRI−< /mo>RA+ :MATH] 4 (C [RI+ RA+], the reduction of total P in the RI^+ RA^+ treatment; C [RI− RA+], the reduction of total P in the RI^− RA^+ treatment; C [RI+ RA−], the reduction of total P in the RI^+ RA^− treatment; X [RI+ RA+], bacterial absolute abundance in the RI^+ RA^+ treatment; X [RI− RA+], bacterial absolute abundance in the RI^− RA^+ treatment) Experiment 2: effects of AM fungal exudates, produced in the presence of low and high availability of P, on the growth of R. aquatilis Bi-compartmented experimental systems were prepared with ROC of M. truncatula associated with R. irregularis MUCL 43194 in the RC. After eight weeks, extraradical hyphae developed profusely in the solid slope and in the liquid MSR medium of the HC. At this point, the medium was removed and replaced with 10 mL of fresh liquid MSR medium containing 35 μM KH[2]PO[4] or 700 μM KH[2]PO[4] as described in Experiment 1. After one week, the hyphal exudates were collected. The inoculum of R. aquatilis was prepared as described in Experiment 1 with slight modifications: the LB medium containing R. aquatilis cells was centrifuged to eliminate the supernatant, and R. aquatilis cells were re-suspended threefold in a sterilized 0.85% NaCl solution. The supernatant was then diluted to OD[600] = 0.1. In a microwell plate, 500 μL of sterilized 0.85% NaCl solution was introduced into the peripheral wells to mitigate potential edge effects, and 450 μL of liquid medium collected from HC (i.e., P^High RI^+, P^High RI^-, P^Low RI^+ and P^Low RI^− treatments) or 450 μL of sterilized 0.85% NaCl solution (i.e., control treatment) were dispensed into the central wells of the microwell plate. Then, 50 μL of the prepared R. aquatilis suspension was introduced into the central wells and uniformly mixed with the liquid medium. Each treatment was replicated four times. The growth curve of PSB was monitored at 37 °C over a 48-h period and the OD[600] measurements of the liquid cultures were taken at 2-h intervals. Statistics and reproducibility Statistical analyses were conducted using SPSS v. 16.0 (SPSS Inc., Chicago, IL, USA). Tukey’s honest significant difference test (P-value < 0.05) was used to evaluate the significant differences among RI^+ RA^+, RI^− RA^+ and RI^+ RA^− treatments at the same P availability. In addition, a t-test (P-value < 0.05) was conducted to determine the significant differences between the low and high P treatments. To ensure the normal distribution of data, the TMPs of various genes were subjected to a log transformation. The DESeq2 R package was used to obtain variance-stabilized counts for genes associated with the P function by transforming read counts. Heatmaps were generated using the heatmap R package to depict the abundance of P cycle-related genes and the gene expression levels of multiple metabolic pathways. The Spearman method was used to assess the correlation between variance-stabilized count-based P regulation genes and genes involved in P solubilization, mineralization, and transport. Reporting summary Further information on research design is available in the [212]Nature Portfolio Reporting Summary linked to this article. Supplementary information [213]Supplementary material^ (905.7KB, pdf) [214]Supplementary data 1-5^ (29.2KB, xlsx) [215]Supplementary data 6^ (2.9MB, xlsx) [216]42003_2025_8501_MOESM4_ESM.pdf^ (41.8KB, pdf) Description of additional supplementary files [217]Reporting summary^ (2.9MB, pdf) [218]Transparent peer review file^ (943.8KB, pdf) Acknowledgements