Computational spatial transcriptomic re-analysis reveals biofilm-depth-dependent compartmentalization of SOS-associated and DNA-repair transcriptional programs in Escherichia coli.

Biofilms are spatially structured microbial communities in which local physiological conditions vary across depth, potentially shaping DNA stress responses, repair pathways, and genome-plasticity-associated cellular states. However, the spatial organization of these transcriptional programs within individual bacterial biofilms remains poorly understood. Here, we re-analyzed public RAINBOW-seq spatial transcriptomic data from Escherichia coli BW25113 biofilms (GSE197541) using a curated 73-gene panel encompassing SOS response, stress response, biofilm matrix, conjugation- and competence-associated proxy genes, housekeeping controls, and metabolic reference genes. Because sequencing depth was spatially biased across biofilm depth, all analyses were performed using TMM-normalized logCPM expression values. Twenty genes displayed significant spatial gradients after false-discovery-rate correction. SOS-associated genes did not behave as a single spatially uniform module. Instead, DNA repair and damage-tolerance genes, including uvrA, uvrB, dinB, and sbmC, were enriched toward the biofilm interior, whereas ruvC, encoding a Holliday junction resolvase, was enriched toward the periphery. Module-level analyses confirmed this spatial dichotomy, which remained robust in leave-one-biofilm-out analyses. Beyond SOS-associated functions, broader categories also exhibited depth-dependent polarization. General stress genes were predominantly interior enriched, whereas several metabolic genes were peripheral enriched. Candidate housekeeping genes were not uniformly spatially invariant, with rpoD and gapA displaying significant but opposing spatial gradients. These findings demonstrate that DNA repair-, stress-response-, and genome-plasticity-associated transcriptional states are spatially compartmentalized within E. coli biofilms rather than being uniformly distributed throughout the community. The results also highlight the importance of validating reference genes in spatial transcriptomic studies of structured microbial systems.