Asymmetric Roles of Ryegrass and Earthworms in Shaping Microbial Community Composition and Nutrient-Acquiring Enzyme Activities in Petroleum-Contaminated Soil

Petroleum contamination of the fragile loess soils of Northern Shaanxi, China, calls for low-cost biological remediation strategies. Combining plants with soil macrofauna is a promising option, but it remains unclear whether the two agents act on the same or on different components of the soil microbial system. In a 150-day soil-column experiment on a moderately contaminated loess soil (initial total petroleum hydrocarbon, TPH, 1.88 g kg−1), we crossed ryegrass (Lolium perenne) and earthworms (Metaphire tschiliensis) in a 2 × 2 factorial design and measured bacterial and fungal community composition, five carbon (C)-, nitrogen (N)-, and phosphorus (P)-acquiring enzyme activities and ecoenzymatic stoichiometry. Ryegrass was the sole significant driver of hydrocarbon removal (p < 0.001; TPH reduced by 26.6% under the combined treatment), whereas earthworms had no significant main effect (p = 0.168). The two agents acted on distinct targets: earthworms restructured the bacterial community (PERMANOVA R2 = 0.253, p < 0.001) without changing α-diversity and strongly raised NO3−-N and N-acquiring NAG activity, whereas ryegrass shaped the fungal community (R2 = 0.105, p = 0.036) and increased C-acquiring enzyme activities. A significant ryegrass × earthworm interaction emerged for β-1,4-glucosidase (p = 0.017) and for microbial C limitation (p = 0.011), which is consistent with the additional rhizosphere C demand being expressed only after earthworms had relieved N shortage. The relative abundance of the hydrocarbon-degrading guild was unrelated to residual TPH (r = 0.19, p = 0.424), indicating that the taxonomic abundance of potential degraders did not predict hydrocarbon removal. Vector angles (51.7–53.7°) fell between the conventional 45° and the recently proposed 55° threshold; available N:P ratios, substrate stoichiometry and the enzymatic response to earthworm-derived N are consistent with N rather than P limitation and suggest that the 45° threshold overestimates P limitation in alkaline soils. Within the conditions tested, plant establishment appears to be the priority for remediating petroleum-contaminated loess, and the combination of plants, earthworms and moderate N-containing organic amendments is a promising strategy that now requires direct experimental evaluation.