Nonallelic homologous recombination events responsible for copy number variation within an RNA silencing locus

The structure of chalcone synthase (CHS) gene repeats in different alleles of the I (inhibitor) locus in soybean spawns endogenous RNA interference (RNAi) that leads to phenotypic change in seed coat color of this major agronomic crop. Here, we examined CHS gene copy number by digital PCR and single nucleotide polymorphisms (SNPs) through whole genome resequencing of 15 cultivars that varied in alleles of the I locus (I, i , i , and i) that control the pattern distribution of pigments in the seed coats. Lines homozygous for the i allele had the highest copy number followed by the I and i cultivars which were more related to each other than to the lines with i alleles. Some of the recessive i alleles were spontaneous mutations, and each revealed a loss of copy number by digital PCR relative to the parent varieties. Amplicon sequencing and whole genome resequencing determined that the breakpoints of several i to i mutations resulted from nonallelic homologous recombination (NAHR) events between CHS genes located in segmental duplications leading to large 138-kilobase deletions that erase the structure generating the CHS siRNAs along with eight other non-CHS genes. Functional hybrid CHS genes (designated CHS5:1) were formed in the process and represent rare examples of NAHR in higher plants that have been captured by examining spontaneous mutational events in isogenic mutant lines.