Species-dependent partitioning of C and N stable isotopes between arbuscular mycorrhizal fungi and their C3 and C4 hosts

• Natural 13C and 15N abundances of arbuscular mycorrhizal fungi vary among strains.
• Fungal δ13C are lower than plant δ13C for C4, but not always for C3 hosts.
• Isotopic partitioning between partners does not drive abundances in plant and fungi.
• Inoculation decreases shoot δ13C of C3 but not C4 host, due to improved water status.

Natural 13C and 15N abundances of mycorrhizal fungi are increasingly used in ecology but reference data on arbuscular mycorrhizal fungi (AMF) are scarce. In experiments with nine phylogenetically dispersed AMF strains inoculated on leek (C3 plant) and sorghum (C4) in pot cultures, we measured the 13C/12C and 15N/14N ratios in shoots, roots, AMF spores as well as carbon isotope signature of the C16:1ω5 fatty acid (FA), which is diagnostic for AMF. Spore δ13C values varied among AMF strains on any given host. They were significantly lower than shoot δ13C for sorghum (−2.5‰ on average) while for leek, no clear C isotope partitioning between spores and host shoots was observed. The FA C16:1ω5 fatty acids were more 13C-depleted than spores, without correlation with spore δ13C values. For both, sorghum and leek, spore δ15N was higher (+1–2‰ on average) than for shoots. We found no evidence that isotopic partitioning between the partners drives 13C and 15N abundances in plant–AMF symbiosis. Mycorrhizal roots displayed relatively high δ13C typical for heterotrophic organs, and not a mix between AMF and plant signatures. Interestingly, inoculation slightly decreased shoot δ13C on leek but not on sorghum, as compared with non-mycorrhizal plants, suggesting that AMF improved the plant's water status, a parameter affecting the δ13C of C3 but not C4 plants. Phylogenetically closer AMF displayed more similar spore δ13C and induced similar 13C and 15N abundances on leek shoots, but this observation was not confirmed for sorghum. Plant and AMF isotopic abundances hardly correlated with other parameters related to plant development, mineral nutrition or root mycorrhizal colonisation, and these correlations were never consistent between sorghum and leek. Thus, isotopic abundances in plant–AMF symbiosis were rather constrained by each AMF/plant interaction. Nevertheless, our data provide a valuable reference for future investigations of AMF communities and AM symbiosis in situ.