摘要： Synthetic nitrogen fertilizers sustain modern rice yields but cause substantial environmental impacts due to low nitrogen-use efficiency. Flooded paddy soils harbor diverse diazotrophs capable of biological nitrogen fixation (BNF), yet BNF is highly variable across soils, management, and rice genotypes. Here, we synthesize recent evidence on how rice can be bred to more consistently recruit and stimulate N₂-fixing microbiomes. We first summarize major diazotroph taxa and niches in paddy ecosystems, highlighting the emerging contribution of iron-reducing bacteria at root iron-plaque interfaces and the principal environmental ‘gates’ on BNF, including flooding regime, bioavailable Fe phases, pH, carbon quality, and mineral-N inputs. We then integrate findings showing that rice genetic variation shapes diazotroph assembly and activity through four root-controlled levers: (i) root system architecture that positions rhizodeposition along redox gradients, (ii) exudate quantity and chemistry (notably flavonoids and low-molecular-weight organic acids) that fuel and signal to diazotrophs, (iii) aerenchyma-mediated radial O₂ loss that creates oxic-anoxic microsites, and (iv) iron plaque formation that couples Fe-C-N cycling and provides a scaffold for N2-fixing communities. Finally, we translate these mechanisms into a breeding roadmap, proposing a BNF-supportive ideotype, candidate loci/genes from GWAS/QTL and wild introgressions, and a validation-to-deployment pipeline combining gene editing, near-isogenic resources, marker-assisted/genomic selection, and multi-environment field testing under low-N management. We also discuss phenotyping bottlenecks, deployment constraints, and priorities for pairing BNF-supportive alleles with compatible microbiomes and agronomy to reduce fertilizer demand while maintaining yield.
