Factorial Inoculation Reveals Composition-Dependent Effects of Microbial Formulations on Shuixian Tea Seedlings

Microbial inoculants can support sustainable crop production, but mixtures may show non-additive interactions, making formulation compatibility important to evaluate in tea. We tested three inoculant factors in a balanced 2 × 2 × 2 root-drench experiment with Shuixian tea seedlings: laboratory-cultured Burkholderia contaminans ZCC (Z), a Bacillus subtilis-based commercial formulation (B), and a Trichoderma harzianum-based commercial formulation (H). Plant physiology, rhizosphere soil properties, and bacterial 16S rRNA gene and fungal ITS1 communities were assessed across eight treatments. Plant responses were composition dependent: B and HB increased root activity by 68.2% and 60.8%, whereas ZHB reduced it by 56.8%. Soil responses were narrower; ZH and HB increased cellulase, Z increased total N, and ZB decreased total N. All seven factorial terms were significant in the eight-trait plant PERMANOVA, with Z × B explaining the most variation (R2 = 0.315), whereas only Z × B was significant for the 11-variable soil matrix (R2 = 0.101). For bacterial community structure, Z × B was the only factorial term consistently retained by the high-permutation Aitchison analysis and complementary robustness checks, and ANCOM-BC2 identified 23 robust genus–factorial-term associations. Fungal communities showed broad whole-community restructuring, but no named genus passed the FDR- and sensitivity-based differential-abundance criteria. Apparent plant–microbiome concordance under unrestricted permutations disappeared when permutations were restricted within treatments, indicating shared treatment structure rather than additional within-treatment coupling. Overall, formulation composition and interactions, rather than inoculant number alone, governed the observed responses, supporting factorial compatibility screening before field validation in tea.