Enhancement of soil nitrogen (N) cycling by grazing has been observed in many grassland ecosystems. However, whether grazing affects the activity only of the key microbial functional groups driving soil N dynamics or also affects the size (cell number) and/or composition of these groups remains largely unknown. We studied the enzyme activity, size, and composition of five soil microbial communities (total microbial and total bacterial communities, and three functional groups driving N dynamics: nitrifiers, denitrifiers, and free N2 fixers) in grassland sites experiencing contrasting sheep grazing regimes (one light grazing [LG] site and one intensive grazing [IG] site) at two topographical locations. Enzyme activity was determined by potential carbon mineralization, nitrification, denitrification, and N2 fixation assays. The size of each community (except N2 fixers) was measured by the most‐probable‐number technique. The composition of the total soil microbial community was characterized by phospholipid fatty acid analysis (PLFA), and the genetic structure of the total bacterial community was assessed by ribosomal intergenic spacer analysis. The genetic structures of the ammonia‐oxidizing, nitrate‐reducing, and N2‐fixing communities were characterized by polymerase chain reaction and restriction fragment length polymorphism (PCR‐RFLP) or by polymerase chain reaction and denaturing gradient gel electrophoresis (PCR‐DGGE) targeting group‐specific genes. Greater enzyme activities, particularly for nitrification, were observed in IG than in LG sites at both topographical locations. The numbers of heterotrophs, nitrifiers, and denitrifiers were higher in IG than in LG sites at both topographical locations. The amplitude of changes in community size was higher than that of community enzyme activity. Phospholipid and nucleic acid analyses showed that the composition/structure of all the communities, except nitrate reducers, differed between IG and LG sites at both locations. For each community, changes in activity were correlated with changes in the occurrence of a few individual PLFAs or DNA fragments. Our results thus indicate that grazing enhances the activity of soil microbial communities but also concurrently induces changes in the size and composition/structure of these communities on the sites studied. Although the generality of our conclusions should be tested in other systems, these results are of major importance for predicting the effects of future disturbances or changed grazing regimes on the functioning of grazed ecosystems.

Management by combined grazing and mowing events is commonly used in grasslands, which influences the activity and composition of soil bacterial communities. Whether observed effects are mediated by management-induced disturbances, or indirectly by changes in the identity of major plant species, is still unknown. To address this issue, we quantified substrate-induced respiration (SIR), and the nitrification, denitrification and free-living N(2)-fixation enzyme activities below grass tufts of three major plant species (Holcus lanatus, Arrhenatherum elatius and Dactylis glomerata) in extensively or intensively managed grasslands. The genetic structures of eubacterial, ammonia oxidizing, nitrate reducing, and free-living N(2)-fixing communities were also characterized by ribosomal intergenic spacer analysis, and denaturing gradient gel electrophoresis (DGGE) or restriction fragment length polymorphism (RFLP) targeting group-specific genes. SIR was not influenced by management and plant species, whereas denitrification enzyme activity was influenced only by plant species, and management-plant species interactions were observed for fixation and nitrification enzyme activities. Changes in nitrification enzyme activity were likely largely explained by the observed changes in ammonium concentration, whereas N availability was not a major factor explaining changes in denitrification and fixation enzyme activities. The structures of eubacterial and free-living N(2)-fixing communities were essentially controlled by management, whereas the diversity of nitrate reducers and ammonia oxidizers depended on both management and plant species. For each functional group, changes in enzyme activity were not correlated or were weakly correlated to overall changes in genetic structure, but around 60% of activity variance was correlated to changes in five RFLP or DGGE bands. Although our conclusions should be tested for other ecosystems and seasons, these results show that predicting microbial changes induced by management in grasslands requires consideration of management-plant species interactions.

No abstract