Rhizosphere Osmoprotectant Metabolite Accumulation Following Inoculation with a Synthetic Bacillus–Pseudomonas Consortium Associates with Nitrogen Cycling and Heat-Stress resilience Indicators in Tomato

Rising global temperatures threaten tomato productivity, demanding sustainable strategies for thermotolerance without increasing agrochemical inputs. Plant-growth-promoting rhizobacteria (PGPR) can alleviate abiotic stress, but single strains often perform inconsistently due to limited functional repertoires. Rationally designed synthetic consortia combining complementary traits offer promise, yet experimental evidence for coordinated rhizosphere osmoprotectant accumulation remains scarce. We assembled a binary consortium of proline- and auxin-producing Bacillus sp. (B. amyloliquefaciens/velezensis operational group) BaC21 and phosphate-solubilizing, trehalose- and glycine-betaine-accumulating Pseudomonas fluorescens DS17R, previously characterized for biocontrol activity, to test whether complementary in vitro traits associate with enhanced rhizosphere osmoprotectant pools and nitrogen cycling under heat stress (38°C) in a tomato microcosm. The consortium sustained rhizosphere colonization (primarily reflecting the more heat-tolerant BaC21) and was associated with significantly higher soil available nitrogen (86% increase) and ammonium nitrogen (70 % increase) than the control, with elevated urease and dehydrogenase activity. Rhizosphere polyamines, soluble sugars, free amino acids, and proline were highest under consortium treatment, consistent with strain profiles, though our design cannot partition bacterial versus plant contributions. These patterns co-occurred with better-preserved chlorophyll (∼2.7 mg g⁻¹ FW), higher relative water content (∼85 %), reduced water loss, and greater shoot/root biomass than single strain treatments. Multivariate analysis revealed a consortium-linked phenotypic syndrome associating osmolyte accumulation with plant growth and soil function, but these hypothesis-generating associations require targeted mechanistic and source partitioning validation.