The plant phenotypic parameter data are shown Figure 1 . The impact of phosphate as factor was significant overall, and the total fresh biomass, as well as shoot and root fresh weight were all increased at high Pi, as compared to low Pi. A significantly higher biomass (total fresh weight and shoot fresh weight) is observed for plants inoculated with RiQS81-Pi⁺ inoculum compared to those inoculated with the RiQS81-Pi^- one. The highest values of total biomass were obtained under high Pi with non-inoculated plants and plants inoculated with RiQS81-Pi⁺. Under high Pi, application of RiQS81-Pi^- R. irregulare significantly reduced the plant growth parameters compared to the non-inoculated plants. Under low P, and as compared to the non-inoculated plants, inoculation with RiQS81-Pi^- strain significantly reduced shoot biomass, but without significant effect on root and total biomass, whereas the RiQS81-Pi⁺ significantly promoted shoot and total biomass. In addition, shoot Pi content ( Supplementary Figure S2 ) was only affected by Pi concentration, and not by mycorrhizal status, indicating that the nutrient use efficiency of the plant was not affected by either fungal inoculum.

Data obtained after the non-vital black ink staining are shown Figure 2 and Supplementary Figure S3 . The impact of phosphate was significant overall, with a decrease of most of fungal parameter values under high Pi conditions. The impact of AMF inoculum type was also significant overall, with differences due to the Pi levels. Under low Pi, plants inoculated with RiQS81-Pi⁺ presented a (i) significant increase of the arbuscule parameters (A^r%, a^r% and A^r%/V^r% ratio); (ii) significant reduction of the vesicle parameters (v%, V%, V^r% and v^r%) and (iii) no significant influence for all other investigated fungal parameters. Under high Pi, as compared to the RiQS81-Pi^- inoculum, the RiQS81-Pi⁺ (i) significantly increased F%, M%, m%, A%, V%, H%, A^r%, V^r%, H^r%, a^r%, h^r%, and Arum/Paris ratio, (ii) significantly decreased the h% and (iii) had no significant effect on all other fungal parameters studied.

The data obtained from the vital alkaline phosphatase (ALP) staining are shown in Figure 3 and Supplementary Figure S3 . The effect of phosphate as factor was not significant overall, but with specific fungal patterns between the two R. irregulare Pi phenotypes. The F%, M%, m% and A% values were significantly higher under high Pi than under low Pi for the RiQS81-Pi^-. For the RiQS81-Pi⁺ fungi, the M%, m% and A% values were significantly lower under high Pi than under low Pi. The impact of Pi phenotypes was significant with two contrasting situations. Under low Pi, inoculation with RiQS81-Pi⁺ inoculum was associated with significantly higher colonization for all fungal parameters assessed. Under high Pi, RiQS81-Pi⁺ inoculation (i) led to a significantly higher F% value, and (ii) has no significant influence for all other fungal parameters investigated. Furthermore, we noticed that vesicles were dark stained under high Pi but not, or less so under low Pi ( Supplementary Figure S4 ), and their abundance was significantly higher within the Pi⁺ phenotype under low Pi conditions. Finally, the vesicle abundance increased significantly with increasing Pi concentration for the Pi^- phenotype, but not for the Pi⁺ one. The percentages of fungal structure stained by ALP within the ink-stained part ( Supplementary Figure S5 ) showed similar profiles for all fungal parameters except for F%: the highest values were obtained under high Pi RiQS81-Pi^- (strongly and significantly higher than under low Pi), while the values obtained from RiQS81- Pi⁺ were intermediate and without significant differences between low and high Pi. Significant differences were observed between RiQS81-Pi^- and RiQS81-Pi⁺, the latter being associated with significantly higher values under low Pi for F%, M% and v%, but significantly lower values under high Pi for M%, m%, V% and A%).

The transcript levels of melon genes involved in respiratory and fermentative pathways were investigated ( Figure 4 ). The gene encoding the melon alternative oxidase CmAOX was repressed by Pi, while the genes involved in the COX pathway (CmCOX5b1, CmCOX5b2 and CmCytC) were induced by Pi. CmCytC and CmCOX5b2 showed a similar expression profile, with a higher expression in the inoculated samples under low Pi, as well as a higher transcription level in the plants inoculated with the RiQS81-Pi⁺ strain under this condition. The fungal Pi phenotype status of the inoculum played no role in the transcript accumulation of these genes in high Pi conditions. The lactate dehydrogenase CmLDH was constitutively expressed in our experimental conditions.

Based on the genome sequence of the melon, we identified seven Pi transporter genes belonging to the Pht1 family in melon, designated CmPT1;1 to CmPT1;7, respectively. Phylogenetic analysis revealed that the CmPT1;5 gene clustered with the AM-specific Pi transporters previously characterized in numerous plant species ( Supplementary Figure S6 ). The expression levels of Pht1 genes were studied in melon roots under high or low Pi availability, either in the absence of AMF or colonized by RiQS81-Pi^- or RiQS81-Pi⁺ R. irregulare ( Figure 5 ; Supplementary Figure S7 ). CmPT1;2 and CmPT1;7 were very weakly expressed in the root samples, but strongly expressed in other parts of the plant ( Supplementary Figure S8 ) and were not further considered in the analysis. All other Pi transporters were repressed by high Pi and most of them were significantly upregulated by inoculation, except CmPT1;6. CmPT1;5 was strongly upregulated in the presence of AMF, as expected from its similarity to previously characterized AMF-inducible Pi transporter genes. CmPT1;1, CmPT1;3 and CmPT1;4 showed a similar profile: in LP conditions, all three were induced by AMF, and in HP, the repression by the Pi was less strong in the presence of the RiQS81-Pi⁺ strain. Similar to all known AMF-inducible Pi transporters, the expression levels were correlated with the mycorrhizal colonization of the roots, suggesting that the difference between RiQS81-Pi⁺ and RiQS81-Pi^- is most likely due to the different colonization of the roots.

Melon orthologs of genes involved in fatty acid biosynthesis and transport previously described in other species were identified and named as follows: CmFatM (XM_008468942.3) encoding a thioesterase; CmRAM2 (XM_008445223.3) encoding a glycerol-3-P acyltransferase; CmSTR1 (XM_008447787.3) and CmSTR2 (XM_051088815.1) encoding two half size ATP-binding cassette transporters of subfamily ABCG ( Supplementary Figures S9 , S10 ). All these genes were strongly upregulated by the presence of AM fungi in roots. In mycorrhizal plants, they were all upregulated in RiQS81-Pi⁺ under both low and high Pi conditions ( Figure 6 ). For CmSTR1 and CmSTR2, the fungal Pi⁺ phenotype allowed to restore the same induction level in HP as in LP, although the mycorrhizal colonization was lower.

Certain Blue Copper-Binding Proteins (BCP) were previously identified as being encoded by AMF-inducible genes (Parádi et al., 2010). Their role remains unknown, but melon, like Medicago, possesses several AMF-specific BCPs. In this study, CmBCP1a (XM_008438947.3) and CmBCP1b (XM_008439498.1) were identified and investigated. Both were strongly induced by AMF, and as for the genes involved in fatty acid metabolism, the fungal Pi⁺ phenotype restored the induction levels of LP in HP ( Figure 6 ).

As for the host plant, fungal genes involved in respiration and nutrient uptake were studied, as well as some previously described genes involved in the Pi signaling pathway.

Genes encoding the alternative oxidase, RiAOX, and the lactate deshydrogenase gene, RiLDH, were both down regulated by high Pi fertilization, but no impact of the R. irregulare Pi phenotype was observed ( Figure 7 ). For the COX pathway, RiCytC was induced by Pi for the RiQS81-Pi^- fungus, but also by the fungal Pi phenotype under LP conditions. The expression of the sugar transporter gene RiMST showed a slight induction in the LP RiQS81-Pi⁺ condition, but the expression was mainly constitutive.

To determine the role of fungal Pi transporters in the mycorrhizal association between Pi^- and Pi⁺ phenotypes, the expression of nine previously described fungal Pi transporters (Walder et al., 2016; Xie et al., 2022) was investigated in melon roots ( Figure 8 ; Supplementary Figure S11 ). RiPT3 and RiPT4 presented a too high sequence similarity to design specific primers, and the expression of these two genes was very low to undetectable in our conditions; therefore, they were not considered for further analysis. RiPT1 and RiPHO1 were the most expressed Pi transporter genes within all conditions ( Supplementary Figure S11 ). The transcript levels of RiPT1 and RiSYG1 were not affected by Pi fertilization or the R. irregulare phenotype. RiPT6 and RiPT7 were only affected by the Pi level and were both repressed in HP conditions. RiPT2 and RiPT5 presented a similar profile, with a higher expression level only in the LP RiQS81-Pi⁺ condition. RiPHO1 was repressed in the LP RiQS81-Pi^- condition.

To investigate the effect of external phosphate concentrations and fungal Pi phenotype on the transcript levels of AM fungal genes involved in Pi uptake and metabolism, the expression profiles of six genes of the PHO pathway (Zhou et al., 2021; Xie et al., 2022) of R. irregulare in mycorrhizal C. melo roots were assessed ( Figure 9 ). Our data show (i) a response to high Pi in Pi^- phenotype in RiKCS1 (down-regulation), RiADO1, RiPHO81 (up-regulation), (ii) a response to high Pi in Pi⁺ phenotype in RiPHO2, RiPHO85 and RiKCS (down-regulation) and (iii) significant differences between the two R. irregulare phenotypes for RiPHO80, RiPHO81, and RiADO1 occurring only under low Pi conditions.

In conclusion, for each of the partners of the symbiosis, the Pi adaptation process influenced the accumulation of specific transcripts. In the fungus, we can distinguish between genes that respond only to Pi, those that are upregulated in RiQS81-Pi⁺ in low Pi and those whose expression is lower only in RiQS81-Pi^- in low Pi ( Supplementary Figure S12A ). Similarly, in the plant, the expression data can be clustered into genes responding only to Pi, those that respond to inoculation with the expression following the colonization rate, and those for which inoculation with the Pi+ phenotype restored an expression similar to LP in HP, regardless of the mycorrhization rate ( Supplementary Figure S12B ). Several significant correlations between phenotypic patterns and transcriptomic patterns were also obtained ( Supplementary Figures S13 - S16 ).

PT1 is the most studied Pi transporter in AMF. Its regulation by Pi remains uncertain, with conflicting results. Xie et al. (2016) and Zhang et al. (2023) reported that PT1 was repressed by Pi in Gigapora margarita and R. irregulare, whereas Fiorilli et al. (2013) found no Pi regulation of this gene in R. irregulare. In our study, RiPT1 transcript levels were not affected by Pi, but a significant and positive response to Pi adaptation was observed. RiPT1 has been proposed to be mainly responsible for the Pi uptake by the extraradical mycelium (Xie et al., 2022). However, the strong and constitutive expression of RiPT1 independent of the Pi concentration in our study highlights the importance of this transporter also at the root level.

The available data on other AMF PTs are limited, sometimes contradictory, and their role in the symbiosis remains to be defined. RiPT5 was shown by Walder et al. (2016) as the only Pi transporter gene in their study, which presented a trend of correlation between mycorrhizal Pi acquisition in plants and its level, defining it as the most promising candidate involved in intraradical Pi transfer to PAS. In more recent work, PT6, as PT1, was proposed to be mainly responsible for Pi uptake by the ERM, while the SPX domain-containing transporters, such as PT7 and PHO1, may be involved in Pi export at the symbiotic interface (Ezawa and Saito, 2018; Xie et al., 2022). The transcription of PT7 was shown to be Pi concentration-independent, but necessary for arbuscule development and symbiotic Pi delivery under medium to low Pi concentrations. Indeed, inactivation of this gene resulted in premature arbuscule degradation, and RiPT7 was hypothesized to play an important role in R. irregulare development within roots at late stage, with Pi delivered by this transporter acting as a signal for continued AMF symbiont development in roots (Xie et al., 2022). In our study, RiPT6 was the most strongly correlated with ink fungal parameters, although it was only regulated by Pi levels and was repressed by HP. Interestingly, transcripts of RiPHO1, RiCOX, RiMST and several fungal ALP parameters (especially the relative arbuscule abundance) are all positively and significantly correlated, suggesting a higher metabolic and exchange activity in the Pi⁺ phenotype. The sugar transporter gene RiMST2 has been shown to be expressed in the IRM and its down-regulation resulted in reduced root colonization (Helber et al., 2011), suggesting a role in sugar translocation from the plant to the fungus. These observations under our experimental conditions would be more consistent with a role for RiPHO1 in the exchange between the host and the fungal partner.

RiKCS1, RiADO1, RiPHO80, RiPHO81, RiPHO85 and RiPHO2 were selected to gain insight into the possible modulation of the fungal PHO pathway between the adapted and non-adapted fungus. Few data are available on these genes in AMF. In Gigaspora margarita, pho81 transcripts were induced under low Pi and repressed under high Pi, whereas the expression of pho80 and pho85 was not affected by ambient Pi levels, suggesting a primary sensing by Pho81 whose expression levels are largely responsive to external Pi levels (Xie et al., 2016). Our data show a different pattern of gene regulation, partially consistent with the Pi response, with differences between the two fungal Pi phenotypes. RiPHO2, RiPHO85, and RiKCS1 were the only P-responsive genes, whereas RiADO1, RiPHO81, and RiPHO80 responded to the Pi adaptation process under LP conditions. These observations again question the physiological status of the LP RiQS81-Pi^- plants and point to possibly less active fungal structures in this condition or a modulation of Pi sensing in the RiQS81-Pi⁺ strain, linked to a higher capacity to colonize the plants.

A number of studies have shown that AM colonization of plants down-regulates the expression of the DPU Pi transporters (Rausch et al., 2001; Burleigh, 2002; Glassop et al., 2005; Requena, 2005; Nagy et al., 2006) while other studies demonstrated different results. Chen et al. (2007) showed a mycorrhiza-induced down-regulation of Pht1;1 and Pht1;2 expression under low-Pi conditions, but Pi-regulation of Pht1;2 under high-Pi conditions for pepper, eggplant, and tobacco. In tomato, Nagy et al. (2009) found no change in the transcript abundance for SlPT1 and SlPT2 upon root colonization. Our data are consistent with these latter observations, as presence of AMF did not induce downregulation of melon Pht1 transporters in any of our conditions. All melon Pht1 transporters expressed in the roots were downregulated by Pi, and all except CmPT1;6 were significantly upregulated by the presence of AMF, with CmPT1;5 expressed exclusively in AM roots. CmPT1;1, CmPT1;3, CmPT1;4 and CmPT1;5 were up-regulated in plants inoculated with RiQS81-Pi⁺ in HP conditions, in direct correlation with the mycorrhizal colonization parameters ( Supplementary Figure S15 ).

Although one cannot necessarily deduce the enzymatic activities from RNA accumulation for the genes involved in respiratory pathways (Sluse and Jarmuszkiewicz, 1998), the results show, both in the plant and the fungus, a down regulation of genes involved in the AOX pathway associated with an upregulation of genes involved in the COX pathway in high Pi conditions, which was expected (Bedini et al., 2018). Consistently, strong correlations were obtained between the mitochondrial respiration pathway components with plant biomass ( Supplementary Figures S12 , S13 ). In agreement with previous reports (Del-Saz et al., 2017), upregulation of plant genes involved in the COX pathway (CmCytC and CmCOX5b2) were induced by the presence of AMF in low Pi conditions, with prominent response with the Pi⁺ phenotype. This upregulation is reliable with positive MGR obtained with RiQS81-Pi⁺, but not with the negative MGR observed with RiQS81-Pi^-. This would be coherent with Smith and Smith (2011), who argued that a negative MGR does not translate to a parasitic fungal behavior since a transfer of Pi and carbon still occur.

In the fungus, significant differences were also observed between the two Pi phenotypes, mainly under low Pi condition (RiCOX5b and RiCytC). This could indicate, in relation with the ALP straining results, a higher metabolic activity of Pi⁺ phenotype. We observed a positive correlation between the ALP stained fungal structures and the transcript abundance of fungal genes as RiCOX, RiMST and RiPHO1 ( Supplementary Figure S12 ), reinforcing the hypothesis that the ALP activity could be related with fungal metabolic process linking sugar uptake, Pi transport and mitochondrial activity (Kizawa et al., 2016; Ortíz et al., 2020). A link between ALP parameters and RiCOX transcripts makes sense, since it was reported that the inhibition of the COX pathway by KCN suppress also alkaline phosphatase activity (Van Belle, 1972; Ezawa et al., 1995). On the other side, plant parameters are positively correlated with RiCytC transcripts. This would be consistent with a nutrient delivery from fungus to plant, as an active process requiring ATP provided by the COX pathway. In conclusion, mitochondria are major actors in mycorrhizal association and play a role in fungal tolerance to Pi.

It was previously hypothesized that the LDH could possibly be related to AOX and fermentation in AMF, to sustain energy demand (Bedini et al., 2018). This could be partially comforted by our results where the expressions patterns of RiAOX and RiLDH were similar, as well as CmAOX and RiLDH, but no correlation was found between CmAOX and CmLDH. Moreover, we noticed that both RiAOX and RiLDH were significantly and positively correlated with fungal ink parameters, and notably with the relative arbuscule and vesicle abundance, but negatively correlated with the proportion of ALP-stained structures within ink. It would therefore indicate that the fungal structures which were not stained by ALP were also viable, with a metabolism associated with the AOX pathway. This might translate to a state of dormancy notably for the vesicles. Therefore, the divergent proportion of ALP under low Pi condition between both R. irregulare phenotypes would be related to a differential physiological time, Pi^- being more advanced in the life stage than Pi⁺ and inversely under high Pi.

Beyond phosphate transport and signaling, other genes related to symbiotic function have been investigated, such as those involved in fatty acid biosynthesis and transport. Arbuscular mycorrhizal fungi (AMF) lack the ability to synthesize fatty acids due to the absence of cytosolic type I fatty acid synthase genes, suggesting their reliance on the plant for fatty acid delivery. This dependence is further supported by the discovery of an AM-specific lipid biosynthetic pathway, and the upregulation of lipid biosynthesis and transporter genes in colonized cortical cells during symbiosis (Zhang et al., 2010; Gutjahr et al., 2012; Tisserant et al., 2013; Bravo et al., 2017; Jiang et al., 2017; Keymer et al., 2017; Luginbuehl et al., 2017; Morin et al., 2019; Malar C et al., 2021). The melon orthologs of the gene encoding the thioesterase FatM, the GPAT RAM2, and a pair of ABCG half-transporters unique to mycorrhizal plants and necessary to support arbuscule formation, named STR1 and STR2 (Zhang et al., 2010; Gutjahr et al., 2012; Bravo et al., 2017; Jiang et al., 2017), were examined. All of these AM-inducible plant genes, and particularly CmRAM2, were positively correlated with several fungal (ink) parameters, especially arbuscle abundance ( Supplementary Figure S12 ). The transcript levels of some of these plant genes were significantly repressed by high Pi in the presence of the R. irregulare Pi^- phenotype, which could be consistent with lower colonization; however, this is not the case for the Pi⁺ phenotype. Remarkably, all these plant genes are significantly differentially regulated between the two R. irregulare phenotypes under high Pi condition, as well as CmFATM under low Pi condition.

While a link between fungal colonization (ALP), RiMST and RiCOX was observed, indicating a consistent flow from sugar uptake and assimilation from arbuscule to the mitochondria, the fate of lipid seems to correlate modestly to this metabolic path. It would be interesting to see in a future trial if a stronger correlation exists with the extraradical spore density.

Blue copper proteins (BCPs), characterized by their involvement in electron transfer reactions, have also been investigated in the context of AM roots. The AM-induced BCP family was shown as specifically and strongly Pi-regulated in arbuscule-containing regions of mycorrhizal roots, in the plasma membrane and the periarbuscular membrane around the arbuscule trunk (Pumplin and Harrison, 2009; Parádi et al., 2010). Although their precise role remains unclear, gene promoter activity studies suggest that BCPs may serve as mediators of electron transfer processes in arbuscule-containing cells and their vicinity (Hohnjec et al., 2005). We monitored the accumulation of transcripts encoding two melon orthologs of the mycorrhiza-responsive genes MtBCP1 (Parádi et al., 2010). Similar to the transcript levels found for genes involved in lipid metabolism and transport, the expression patterns of CmBCP1a and CmBCP2b cannot be interpreted easily, but the significant and higher levels of transcript in RiQS81-Pi⁺ as compared to RiQS81-Pi^- is a clear result from the adaptation process.

Taken together, the molecular data suggests that the Pi tolerance status of the fungus had an influence on the molecular dialogue between the two partners of the symbiosis. It is interesting to note that most of the transcriptional differences between the two Pi phenotypes are observed under low Pi condition on the fungal side but are observed under high Pi on the plant side. We noticed fewer significant changes and larger standard errors for the fungal transcript dataset. One of the probable biases is that we used RNA extracted from whole mycorrhized roots, which means a mixture of different structures (vesicles, hyphae, arbuscules), each of which probably has different regulatory profiles. For a clearer answer about the exchanges with the plant, it would be necessary to perform the transcriptional analyses in arbuscocytes only. It would also be interesting to see how the genes related to signaling are regulated in the ERM, and in kinetics through the lifespan of the symbiosis.

The possibility to induce R. irregulare Pi⁺ phenotype after a short number of generations raises the question of whether high Pi-tolerant AMF lines exist naturally. High Pi levels resulting from long-term high Pi fertilization are common in many conventional field soils (Cheng et al., 2013). The number of propagules in such soils may be low or even absent in some cases (Mäder et al., 2000; Oehl et al., 2003; Roy et al., 2017). However, other practices challenge the sole effect of Pi in inhibiting the mycorrhizal colonization and soil richness, such as tillage (Jansa et al., 2002; Castillo et al., 2006; Avio et al., 2013; Verzeaux et al., 2017), lack of cover crops (Soti et al., 2016), selection of low mycorrhizal plant species (Martín-Robles et al., 2018), or application of pesticides (Jin et al., 2013; Riedo et al., 2021; Edlinger et al., 2022). Moreover, a lack of Pi inhibition on the fungal community under field conditions has also been observed in long- or short-term fertilization approaches (Peña Venegas et al., 2021), but to the best of our knowledge, investigations of possible high Pi-tolerant AMF species have never been carried out. Perhaps the elaboration of high-resolution markers specific to the Pi⁺ phenotype could help in the detection of Pi-tolerant lines in natural and managed ecosystems, and could improve our knowledge of AMF ecology. At least, our results could suggest that adaptation to higher Pi concentrations of R. irregulare could occur in managed conventional systems, and, if true, would give more weight to the negative influence of other factors.

Our data suggest that propagule abundance may not be the sole determinant for achieving beneficial effects on crops. Traditionally, the production of propagule-rich inocula has been pursued to increase the capacity for inoculating larger areas for commercial purposes and to meet quality criteria for scientific usage. An inoculum with fewer propagules would require higher doses to maintain a minimal but efficient number of propagules applied per plant. As a result, larger quantities would need to be produced, resulting in higher costs for customers and growers. However, the application of elevated Pi concentration during the first generation of ex vitro production of RiQS81-Pi⁺ significantly reduced its ability to produce propagules compared to the classical production route of RiQS81-Pi^-. Interestingly, the difference between the adapted and non-adapted R. irregulare gradually decreased over subsequent generations. This suggests that in the future, it may be possible to achieve a similar number of propagules between the two R. irregulare phenotypes during production under contrasting Pi fertilization, providing definitive evidence of short-term Pi adaptation.

Conversely, all other things being equal, it is possible that production systems that typically involve specific fertilization, especially low phosphate fertilization, may adapt mycorrhizal species to produce high amounts of propagules, but could potentially lead to reduced response or specific loss of native adaptive traits to a given soil ecosystem. Furthermore, previous studies have shown that the application of exogenous mycorrhizal strains can lead to the genetic replacement of native fungal species and result in limited plant responses (Calvente et al., 2004; Schwartz et al., 2006; Rowe et al., 2007; Estrada et al., 2013; Emam, 2016). It is possible that this limitation is due to the fertilization used during mycorrhizal production, which would make mycorrhizal strains or isolates less able to thrive in certain soil conditions or lead to incompatibilities between exogenous and native species (Bencherif et al., 2023). Our results open the possibility of producing innovative and generic mycorrhizal inocula that could be better adapted to specific fertilization regimes or elemental concentrations.

Building on the points raised in the previous paragraph, it is logical to address a potential concern about the design of inoculum production and its consequence for scientific purposes. The effects of high phosphate concentrations on mycorrhizal development and response have been extensively studied for decades. Numerous publications conducted under controlled conditions have consistently demonstrated the inhibition of mycorrhizal colonization by phosphate, often associated with negative MGR (Mosse, 1973; Thomson et al., 1986; Smith and Read, 2008; Nagy et al., 2009; Breuillin et al., 2010; Balzergue et al., 2011). Over time, these studies have delved deeper into the topic, using increasingly sophisticated tools such as transcriptomics, genetics, and mutant plants to gain a better understanding of the role of phosphate in endomycorrhizal symbioses. However, it is worth questioning whether we have generated artificial knowledge by using artificially domesticated fungal inocula. Our data challenge the notion that the phenomenon of phosphate inhibition is solely the result of a top-down mechanism in which the plant controls the fungus. Instead, we emphasize the importance of fungal respiration and phosphate metabolism, which play a role in the fungus’ own phenotypic behavior and may also have consequences for the plant response. It is important to note that the mycorrhizal inocula used in previous studies were primarily derived from pot culture systems typically produced by laboratories, companies, or glomeromycotan banks. These systems often include specific phosphate fertilization regimes as a widely used standard for optimal fungal propagation (Ijdo et al., 2011). The production of mycorrhizal fungi under such specific fertilization regimes over multiple generations could have profound effects on fungal physiology, the extent of which is largely unknown when these inocula are used for research purposes. This could potentially explain or contribute to discrepancies in the interpretation of the effects of compounds that influence mycorrhizal behavior, such as hormones or nutrients. Therefore, it will become critical to consider how a fungal inoculum used in research is produced, as it may have important implications for the results and interpretation.