Medicago arborea can be used for re-vegetation purposes under semiarid conditions. These woody legumes have the ability to form an association with arbuscular mycorrhizal (AM) fungi and rhizobial bacteria, which can be maximised by microorganisms producing certain stimulating metabolites acting as plant growth promoting rhizobacteria (PGPR). The effects of single and combined inoculations using microorganisms with different and interactive metabolic capacities, namely three Glomus species, two Rhizobium meliloti strains (a wild type, WT and its genetically modified derivative GM) and a plant growth promoting rhizobacterium, (PGPR), were evaluated. All three inoculated AM fungi affected Medicago growth in different ways. Differences were maintained when soil was co-inoculated with each of the rhizobial strains (WT or GM) and the PGPR. Mycorrhizal fungi were effective in all cases, but the PGPR only affected plant growth specific microbial situations. PGPR increased growth of G. mosseae-colonised plants associated with Rhizobium WT strain by 36% and those infected by G. deserticola when associated with the rhizobial GM strain by 40%. The most efficient microbial treatments involved mycorrhizal inoculation, which was an indication of the AM dependency of this plant species. Moreover, PGPR inoculation was only effective when associated with specific mycorrhizal endophytes (G. mosseae plus WT and G. deserticola plus GM rhizobial strain). The reduced root/shoot (R/S) ratio resulting from PGPR inoculation, was an indication of more effective root function in treated plants. AM colonisation and nodule formation were unaffected by the type of AM fungus or bacteria (rhizobial strain and/or PGPR). AM from natural soil were less infective and effective than those from the collection. The results supported the existence of selective microbial interactions affecting plant performance. The indigenous AM fungi appeared to be ineffective and M. arborea behaved as though it was highly dependent on AM colonisation, which implied that it must have a mycorrhizal association to reach maximum growth in the stressed conditions tested. Optimum growth of mycorrhizal M. arborea plants was associated with specific microbial groups, accounting for a 355% increase in growth over nodulated control plants. The beneficial effect of PGPR in increasing the growth of a woody legume, such as M. arborea under stress, was only observed with co-inoculation of specific AM endophytes. As a result of the interaction, only shoot biomass was enhanced, but not as a consequence of enhancing of the colonising abilities of the endophytes. The growth stimulation, occurring as a consequence of selected microbial groups, may be critical and decisive for the successful establishment of plants under Mediterranean climatic and soil conditions.
