Betalain is one of four major plant pigments, which shares some features with another major pigment anthocyanin. However, no plant was found to biosynthesize both pigments. Previous studies have reported that anthocyanin biosynthesis is regulated by post-transcriptional gene silencing (PTGS) in some plants, but the importance of PTGS in betalain biosynthesis remained unclear. In this study, we report the occurrence of PTGS in betalain biosynthesis in bougainvillea (Bougainvillea peruviana) ‘Thimma’. ‘Thimma’ produces three different color bracts on the same plant, pink, white and pink-white bicolor. This unstable pigmentation phenotype resembles the unstable anthocyanin pigmentation associated with PTGS, thus PTGS in the betalain biosynthetic pathway was expected. To test this hypothesis, we performed pigment analysis, gene expression analysis, small RNA analysis and transient overexpression analysis. We demonstrated that PTGS of BpCYP76AD1, one of the betalain biosynthetic enzyme genes, is responsible for the loss of betalain biosynthesis in ‘Thimma’. Genomic background or DNA methylation in the BpCYP76AD1 genome could not explain the induction of PTGS implying other locus regulates the unstable pigmentation. Our data indicates naturally occurring PTGS contributes to color pattern diversification not only in anthocyanin biosynthesis but also in betalain biosynthesis.

Here, RNA-seq data (5.8 Gb for a pink bract and 5.7 Gb for a white bract, respectively) were assembled by Trinity. This data was used for identification of betalain-related genes and for small RNA mapping analysis.

RNA was extracted from around 1.5 cm length bracts (pink and white) of 'Thimma' using Sepasol RNA I Super G (Nacalai Tesque, Kyoto, Japan). Total RNA was sequenced using a Illumina Hiseq2500 (Illumina Inc., San Diego, CA, USA) with 101-bp paired end (5.8 Gb for a pink bract and 5.7 Gb for a white bract, respectively), and the sequenced data were assembled by Trinity, and finally 156,790 contigs were obtained. This data was used for identification of betalain-related genes and for small RNA mapping analysis.

This is FASTA format file for assembled transcript sequence data. RNA was extracted from around 1.5 cm length bracts (pink and white) of 'Thimma' using Sepasol RNA I Super G (Nacalai Tesque, Kyoto, Japan). Total RNA was sequenced using a Illumina Hiseq2500 (Illumina Inc., San Diego, CA, USA) with 101-bp paired end (5.8 Gb for a pink bract and 5.7 Gb for a white bract, respectively), and sequenced data were assembled by Trinity, and finally 156,790 contigs were obtained.