Opening the black box: outcomes of interactions between arbuscular mycorrhizal (AM) and non-host genotypes of Medicago depend on fungal identity, interplay between P uptake pathways and external P supply

We investigated the physiology that underlies the influence of arbuscular mycorrhizal (AM) colonization on outcomes of interactions between plants. We grew Medicago truncatulaA17 and its AM-defective mutant dmi1 in intragenotypic (two plants per pot of the same genotype, x2) or intergenotypic (one plant of each genotype, 1+1) combinations, inoculated or not with Rhizophagus irregularis (formerly Glomus intraradices) or Gigaspora margarita. We measured plant growth, colonization, contributions of AM and direct P uptake pathways using P-32, and expression of plant P-i transporter genes at two levels of P supply. A17 (x2) responded positively to inoculation only at low P. The response was enhanced with 1+1 even at high P where colonization in A17 was reduced. With R.irregularisP uptake by the AM pathway was unaffected by P supply, whereas with G.margarita, the AM pathway was lower at high P, and direct uptake higher. Gene expression varied and was unrelated to P uptake through the two pathways. There was no evidence of plant control of P uptake via R.irregularis at high P but there was via G.margarita. Importantly, growth responses of plant genotypes grown alone did not predict outcomes of intergenotypic interactions. Ecological studies of effects of soil microorganisms on plant-plant interactions are frequently based on a black box' approach; underlying mechanisms are not investigated and interpretation of data is based on generalised concepts. We provide direct evidence relating to arbuscular mycorrhizal (AM) function in phosphorus (P) uptake and variations in effects of different AM fungi on plant interactions which should be taken into account in interpreting ecological data. Importantly, we found evidence of plant control of P uptake for only one of two AM fungi used, and growth responses to AM inoculation of plant genotypes (host and non-host) grown alone did not predict outcomes of intergenotypic interactions.