Comparative transcriptome analysis reveals key insights into male sterility in Salvia miltiorrhiza Bunge

BACKGROUND: Large-scale heterosis breeding depends upon stable, inherited male sterility lines. We accidentally discovered a male sterility line (SW-S) in the F(1)progeny of a Salvia miltiorrhiza Bunge from Shandong, China (purple flowers) crossed with a S. miltiorrhiza f. alba from Sichuan, China (white flowers). We sought to provide insights into the pollen development for male sterility in S. miltiorrhiza. METHODS: The phenotypic and cytological features of the SW-S and fertile control SW-F were observed using scanning electron microscopy and paraffin sections to identify the key stage of male sterility. Transcriptome profiles were recorded for anthers at the tetrad stage of SW-S and SW-F using Illumina RNA-Seq. RESULTS: The paraffin sections showed that sterility mainly occurred at the tetrad stage of microspore development, during which the tapetum cells in the anther compartment completely fell off and gradually degraded in the sterile line. There was little-to-no callose deposited around the microspore cells. The tetrad microspore was shriveled and had abnormal morphology. Therefore, anthers at the tetrad stage of SW-S and fertile control SW-F were selected for comparative transcriptome analysis. In total, 266,722,270 clean reads were obtained from SW-S and SW-F, which contained 36,534 genes. There were 2,571 differentially expressed genes (DEGs) in SW-S and SW-F, of which 63.5% were downregulated. Gene Ontology (GO) enrichment analysis indicated that the differentially expressed genes were enriched in 56 functional groups (GO terms); of these, all DEGs involved in microgametogenesis and developmental maturation were downregulated in SW-S. These results were confirmed by quantitative RT-PCR. The two GO terms contained 18 DEGs, among which eight DEGs (namely: GPAT3, RHF1A, phosphatidylinositol, PFAS, MYB96, MYB78, Cals5, and LAT52) were related to gamete development. There were 10 DEGs related to development and maturation, among which three genes were directly related to pollen development (namely: ACT3, RPK2, and DRP1C). Therefore, we believe that these genes are directly or indirectly involved in the pollen abortion of SW-S. Our study provides insight into key genes related to sterility traits in S. miltiorrhiza, and the results can be further exploited in functional and mechanism studies.