Rhizosphere Biochemical and Microbial Community Reorganisation Under Consecutive Multi-Year Penicillin Application in a Citrus Orchard

Abstract Background Antibiotic use in citrus production may alter non-target rhizosphere processes, but replicated field evidence remains limited. We evaluated whether documented penicillin application histories were associated with coordinated soil biochemical and bacterial and fungal community patterns in a commercial Wogan mandarin orchard. Results A blocked retrospective comparison included three spatial orchard blocks and four documented application-history groups (no recorded application or application recorded in one, two, or three years), with one plot per group in each block. Rhizosphere soil from three focal trees per plot was pooled, yielding 12 plot-level samples from 36 trees. Nine soil biochemical variables were integrated with near-full-length bacterial 16S rRNA gene and fungal ITS amplicon profiles. All nine reported soil variables differed among groups after block adjustment. Observed bacterial richness and Shannon diversity showed strong non-linear group-associated patterns, whereas fungal alpha-diversity evidence was less robust across parametric and exact Friedman sensitivity analyses. Block-restricted PERMANOVA detected bacterial and fungal Bray-Curtis separation, but multivariate dispersion also differed in both domains; centroid and dispersion effects were therefore interpreted jointly. Taxonomic and environmental associations were exploratory, and no FAPROTAX-predicted function remained significant after false-discovery-rate correction. Conclusions Documented penicillin application history was associated with coordinated, non-linear rhizosphere biochemical and microbial patterns. Because the histories were retrospective and residue concentrations and several soil and disease covariates were unavailable, the findings support field associations and testable hypotheses rather than a direct cumulative penicillin effect.