Abstract
We determined the molecular basis of three soybean lines that vary in seed coat color at the R locus which is thought to encode a MYB transcription factor. RM55-r(m) is homozygous for a mutable allele (r(m)) that specifies black and brown striped seeds; RM30-R* is a stable black revertant isoline derived from the mutable line; and RM38-r has brown seed coats due to a recessive r allele shown to translate a truncated MYB protein. Using long range PCR, 454 sequencing of amplicons, and whole genome re-sequencing, we determined that the variegated RM55-r(m) line had a 13 kb CACTA subfamily transposon insertion (designated TgmR*) at a position 110 bp from the beginning of Intron2 of the R locus, Glyma09g36983. Although the MYB encoded by R was expressed at only very low levels in older seed coats of the black revertant RM30-R* line, it upregulated expression of anthocyanidin synthase genes (ANS2, ANS3) to promote the synthesis of anthocyanins. Surprisingly, the RM30-R* revertant also carried the 13 kb TgmR* insertion in Intron2. Using RNA-Seq, we showed that intron splicing was accurate, albeit at lower levels, despite the presence of the 13 kb TgmR* element. As determined by whole genome methylation sequencing, we demonstrate that the TgmR* sequence was relatively more methylated in RM30-R* than in the mutable RM55-r(m) progenitor line. The stabilized and more methylated RM30-R* revertant line apparently lacks effective binding of a transposae to its subterminal repeats, thus allowing intron splicing to proceed resulting in sufficient MYB protein to stimulate anthocyanin production and thus black seed coats. In this regard, the TgmR* element in soybean resembles McClintock's Spm-suppressible and change-of-state alleles of maize. This comparison explains the opposite effects of the TgmR* element on intron splicing of the MYB gene in which it resides depending on the methylation state of the element.