Genetic control of anthocyanin pigmentation in Antirrhinum flowers

Abstract

The genus Antirrhinum (commonly known as snapdragons) contains more than twentyfive recognised species. The genus has been divided into three morphological subsections: Antirrhinum, Streptosepalum and Kickxiella (Rothmaler, 1956). One of the major characteristics distinguishing the three subsections is flower colour. Most species in subsection Antirrhinum have dark pink or yellow flowers, Kickxiella species are white or pale pink and Streptosepalum species have yellow or pale pink flowers. All Antirrhinum species can be crossed to produce fertile hybrids which allow the genes that underlie their differences to be identified.

I used quantitative trait locus (QTL) analysis on hybrids of A. majus (dark magenta flowers) and A. charidemi (pale-pink flowers) to map genomic regions underlying differences in flower colour. This identified two major-effect loci, in Linkage Group 3 (LG3) and LG7, that explained most of the differences between these species. I used near-isogenic lines (NILs) to further test involvement of two candidate genes - Rosea (Ros) in LG3, which encodes a regulator of the anthocyanin biosynthesis pathway (ABP) and Incolorata (Inc) in LG7 which encodes a rate-limiting enzyme of the ABP. In both cases, the A. majus allele increased pigmentation. Sequence differences between Ros alleles of A. majus, A. charidemi and A. molle (a Kickxiella species with white flowers) suggest that A. molle carries a ros loss-of-function mutation and that a transposon insertion in the ROS promoter might contribute to differences in expression between A. majus and A. charidemi. Ros genotypes were found to be strongly correlated with pigmentation in the corolla tube in A. majus x A. charidemi hybrids, and to a lesser extent with corolla lobe pigmentation, although NILs suggested that ROS did not correspond to the major-effect QTL indentified in LG3. I also mapped a minor-effect QTL for tube pigmentation to a region of LG4 containing the ABP structural gene Candica. Analysis of NILs revealed that Inc was not the second major-effect QTL mapped to LG7, although sequence differences were detected between Inc alleles of A. majus and A. charidemi. I was further able to narrow down the region containing the second LG7 major-effect QTL to an interval of 11 cM, between two molecular markers, which could be used to determine the likely QTL genotypes of segregating NILs. Surprisingly, several ABP genes, particularly Nivea, Inc and Pallida, were expressed at higher levels in pale flowers that were homozygous for the A. chardemi QTL allele than in their dark flowered siblings that carried an A. majus allele. This suggests that ABP genes might be up-regulated in pale flowers as part of a negative feedback mechanism. Two potential roles of the LG7 QTL are considered 1) its requirement for anthocyanin modification or transport to the vacuole, so that a build-up of cytosolic anthocyanins or their break-down products in pale flowers increases structural gene expression but cannot compensate for the overall reduction in anthocyanin, or 2) a role in promoting production of flavonols at the expense of anthocyanins.