Single-nucleus long-read RNA sequencing reveals isoform usage shifts in mouse acute kidney injury

Introduction Single-cell and single-nucleus short-read RNA sequencing has elucidated transcriptomic changes across kidney cell types at the gene expression level. However, its limited read length precludes reliable isoform identification, hindering the characterization of individual protein-coding isoforms and non-protein-coding isoforms. We combined long-read sequencing with single-cell and single-nucleus RNA sequencing technologies to investigate isoform regulation in healthy and acutely injured mouse kidneys. Methods Single-cell and single-nucleus long-read RNA sequencing datasets were generated from kidneys of three healthy adult male C57BL/6J mice. An additional three adult male C57BL/6J mice underwent bilateral renal ischemia-reperfusion injury, and kidneys were subjected to single-nucleus long-read RNA sequencing. Results A total of 1.3 billion reads were generated, with a median read length of 757-884 bases across samples. Single-nucleus libraries yielded a smaller proportion of reads assigned to isoforms than single-cell libraries, reflecting their larger proportion of intronic reads. Nevertheless, single-nucleus libraries detected more distinct isoforms and greater isoform expression diversity, partly owing to the near absence of mitochondrial gene-derived transcripts. Single-nucleus analysis of healthy and injured kidneys revealed that ischemia-reperfusion injury downregulated oxidative phosphorylation-associated genes in proximal tubule cells both at the gene expression level and through shifts in isoform composition. Beyond oxidative phosphorylation-associated genes, Fut9 and Prrc2b were identified as genes exhibiting prominent shifts in isoform usage after ischemia-reperfusion injury. Conclusion Single-nucleus long-read RNA sequencing detected injury-induced shifts in isoform usage in mouse kidney, adding a layer of transcriptomic information that is largely inaccessible to conventional gene-level short-read analysis.