Response of soil microbial communities to contrasted histories of phosphorus fertilisation in pastures

Abstract

Several microbial taxa that potentially affect soil phosphorus (P) biogeochemical cycling and plant nutrition have been identified, in particular some members of the Actinobacteria, Pseudomonas, and arbuscular mycorrhizal fungi (AMF) have been found to solubilise P and enhance P uptake. This study investigated the diversity of selected key taxa within a pastoral agricultural system, to address the hypothesis that if these taxa have an intrinsic association to soil P, long-term P-manipulation should present a selective pressure. We defined this selective pressure as a shift in community assemblage, based on either PCR-DGGE banding patterns or taxon-specific quantitative polymerase chain reaction (qPCR) (mycorrhizae). The work was conducted on a long-term fertiliser field trial site where soil P status had been altered for ∼40 years, with total P ranging from 464 to 1102mgkg−1 soil. The structure of both Actinobacteria and Pseudomonas communities was found to differ according to fertiliser treatment (P<0.05). However, only the Actinobacteria were associated with the underlying P fertility gradient (soil C:P ratio; ρ=0.481; P=0.019); no significant links between the Pseudomonas community and soil physicochemical properties were found. As the structure of the general bacterial or α-Proteobacterial communities did not vary between treatments (P>0.05), the hypothesis that Actinobacteria communities are intrinsically linked to soil P cycling is supported. Similarly, mycorrhizal community structure was strongly influenced by soil P status (P=0.002), with particularly strong differences between the control treatments (no P fertiliser) and fertilised plots, meaning the addition of fertiliser P to soil per se was more important than quantity. P-fertilizer application increased the diversity of AMF. The AMF community of the control plot was dominated by Glomus intraradices and Glomus claroideum; these taxa may have had the highest fitness under the low P conditions. The total fungal community structure, assessed by denaturing gradient gel electrophoresis (DGGE), also varied over the fertiliser trial (P=0.001), however the variation was not correlated with soil P status (biota and/or environment matching; BIO-ENV test). Our results expand our knowledge on the microbiology supporting soil P cycling in pasture soils. Furthermore, we demonstrate that communities of Actinobacteria and AMF, proposed to be involved in supplying P to pasture plants, are linked to the underlying soil P status.