Aconserved salt-stress co-expression network in japonica rice reveals partially dissociated hub gene connectivity and transcriptional induction

Abstract

Introduction Soil salinity severely compromises rice (Oryza sativa L.) growth and grain yield by disrupting ionic and osmotic homeostasis. Although transcriptome profiling has advanced understanding of salt-stress responses in rice, the conservation of co-expression networks across genetic backgrounds and the relationship between network connectivity and transcriptional induction remain poorly characterized. Methods We performed integrated RNA sequencing and weighted gene co-expression network analysis (WGCNA) on three japonica rice varieties under control (0 mM), moderate (85.6 mM), and severe (136.9 mM) NaCl stress. Results All three varieties exhibited basal salt tolerance. Differential expression analysis identified 2,067 and 4,900 non-redundant differentially expressed genes (DEGs) under moderate and severe stress, with 205 and 1,093 DEGs shared across the three varieties, respectively. WGCNA identified 32 co-expression modules. The brown module showed the strongest dose-dependent response to salt stress and was enriched in phenylpropanoid biosynthesis, plant hormone signal transduction, and MAPK signaling; its eigengene was highly concordant across varieties (mean r = 0.987), consistent with a shared leaf-level transcriptional program. Across brown-module genes, intramodular connectivity and salt-induced fold change were positively but only moderately correlated (Pearson r = 0.377): OsDREB1A was the most connected gene but not the most strongly induced, whereas LOC_Os10g35080 was the most strongly induced yet ranked 14th in connectivity. In the 15-node hyperosmotic salinity response subnetwork, no single hub dominated, and OsDREB1A occupied a peripheral position. Discussion Connectivity and induction thus reflect complementary aspects of the salt-stress response, and candidate hub genes for molecular breeding should be prioritized by both criteria jointly, followed by validation across diverse genetic backgrounds.