Mechanisms of complex programmed patterns of anthocyanin pigment formation in Antirrhinum majus : a thesis presented in partial fulfilment of the requirements for the degree of Doctor of Philosophy in Plant Molecular Biology at Massey University, Palmerston North New Zealand
Antirrhinum majus is a model plant used in flower pigmentation studies. Anthocyanin
pigment production is mainly controlled by regulation of transcription of the anthocyanin
biosynthetic genes. Two types of transcription factors, M Y B and bHLH, together with a
WD40 type co-regulator have been shown to regulate the transcription of the anthocyanin
biosynthetic genes. In antirrhinum, in addition to the wild type Rosea 1 phenotype, in
which pigmentation occurs throughout the inner and outer epidermis of the petal , other
complex pigmentation patterns are observed, such as anthocyanins being produced only
in the outer (abaxial) epidermis of both lobes and upper tube region of the dorsal petals
(roseadorsea phenotype). The major objective of this research project was to understand the
genetic regulatory system leading to the development of the two different floral
pigmentation patterns in antirrhinum as a means to understanding differential regulation
of gene expression in similar cells.
Promoter deletion analysis coupled with linker scanning mutagenesis identified
the - 1 62 bp to - 1 20 bp region of the Rosea l promoter as i mportant for the regulation of
the Rosea l gene. Four putative transcription factor-binding sites within this region : a Wbox,
a pyrimidine box, a DOF and a WRKY transcription factor binding site were shown
to be important for Rosea l gene regulation.
Promoter deletion analysis carried out on the rosea ldorsea promoter showed that the
proximal 1 87 bp deletion was, surprisingly, not responsible for the roseadorsea phenotype.
Cloning and characterisation of the Rosea l promoter sequence from various Antirrhinum
species and accessions verified this finding. The rosealdorsea promoter analysis also
indicated that - 1 5 1 bp of the promoter was sufficient for its expression as well as for the
maintenance of petal specific expression. The rosea ldorsea allele was also shown to
encode a functional protein .
In situ hybridisation analysis showed that Rosea l transcripts were present in the inner
and outer epidermis of the petal tissue of both wild type and roseadorsea petal tissue.
Vascular expression of the Roseal mRNA is indicative of regulation of this gene through
sugar or hormonal cues. However, rosea ldorsea transcript levels (in roseadorsea) were much
lower than Roseal (wild type). Lowered expression of rosea ldorsea transcripts may be
responsible for the overall weak pigmentation in the roseadorsea flowers. Analysis of the
intron sequences of the two alleles revealed that many sequence changes were present in
the intron 2 of rosea l dorsea. These changes may lead to instability or the lower expression
of the rosea l dorsea mRNA and may be responsible for the roseadorsea phenotype. Another
possibility is that a fourth Myb gene may be responsible for the roseadorsea phenotype.
The role of the Deficiens gene in direct regulation of Rosea l was analysed by RNAi and
bioinformatics-based methods. The presence of potential MADS box binding sites in the
intron 2 region of the Roseal allele indicated that Rosea l might be directly regulated by
Deficiens. Initial experiments using transient assays did not support this suggestion.
However, silencing of Deficiens in wild type antirrhinum buds led to the loss of
anthocyanin pigments in the petals. Further analysis of the RNAi tissue using SEM
revealed that the proper development of conical shaped epidermal cells was also affected .
The RNAi tissue also developed chlorophyll pigments underscoring the plasticity of petal
identity. This work demonstrated that proper expression of Deficiens is required
throughout flowering for anthocyanin pigment production as well as maintenance of petal
The current investigation revealed that the higher order regulation of the Rosea l alleles in
antirrhinum petals is much more complex than initially postulated.