Previous studies have concentrated on the characterization of red colour in the fruit of Actinidia species [33, 34]. In those studies three cyanidin-based and two delphinidin-based anthocyanins were identified in fruit of six Actinidia species. While the cyanidin compounds were found in all the species examined, only two taxa, A. melanandra and A. arguta var. purpurea, contained the delphinidin-based compounds: delphinidin 3-O-(xylosyl)galactoside and delphinidin 3-O-galactoside. They were not found in either A. chinensis or A. eriantha. In our study we found that both delphinidin compounds were present in the petals in a proportion of the progeny genotypes within three of the four families resulting from a backcross between A. eriantha and A. chinensis var. chinensis. Cyanidin and delphinidin are produced through different branches of the anthocyanin pathways, from the action of either the flavonoid 3'-hydroxylase (F3'H) or flavonoid 3'5'-hydroxylase (F3'5'H), so our results suggest that both pathways are activated in the flowers of the Actinidia backcross population studied here .
Anthocyanin colour perception in plant tissues is affected by the ability of the anthocyanins to undergo changes in chemical form resulting from their interactions with metal ions, other anthocyanin molecules (self-association), or other unrelated compounds (co-pigmentation). These associations may cause a change in hue, or an increase in colour intensity [35, 36]. A wide range of potential co-pigment compounds have been reported, but by far the most common are the colourless flavonol and flavone types of flavonoids. Flavonols were found in the petal samples examined here, with three of the four flavonols identified, quercetin-rutinoside, quercetin-glucoside and kaempferol-glucoside, being present in all samples. While kaempferol-rutinoside was present in 60 of the 134 samples, it was only found in small concentrations, and its presence or absence did not correlate with the perceived hue of the petals. Neither was the presence or absence of delphinidin compounds aligned with perceived hue. This can be seen from the two genotypes 21-07-11a and 21-09-13d, where delphinidin was not present in either genotype but petals of 21-09-13d appeared to have a blue tinge (Figure 2). Therefore we conclude that the hue variation observed in the different genotypes cannot be explained by the combinations of anthocyanin or flavonol types recorded. The differences in depth of colour seen in the petals corresponded to the total anthocyanin concentration.
This study indicates that anthocyanin and flavonol biosynthesis in Actinidia is a complex and highly regulated process. The flavonol kaempferol and the anthocyanin pelargonidin share the same dihydrokaempferol (DHK) precursor. However, while kaempferol was present in the petals, the DHK was not apparently utilised by dihydroflavonol 4-reductase (DFR) and anthocyanidin synthase (ANS) to produce pelargonidin, as this compound was not detected. Rather, the biosynthetic route to anthocyanins proceeded by the action of F3'H and F3'5'H on the precursors prior to their conversion by DFR and ANS to cyanidin (3'4'-hydroxylation) and delphinidin (3'4'5'-hydroxylation). The presence of F3'5'H activity, however, did not result in the production of the flavonol myricetin, which also has 3'4'5'-hydroxylation. The apparent channelling of substrate into flavonols and anthocyanins with varying patterns of hydroxylation, despite the presence of the F3'H and F3'5'H, has also been observed in other species . It could reflect different developmental timings for the production of flavonols and anthocyanins, metabolic channelling within enzyme complexes, or particular substrate specificity of biosynthetic enzymes such as DFR. The DFR of genera such as Petunia and Cymbidium have been shown to have only weak activity with DHK, resulting in the near absence of pelargonidin-based anthocyanins in these genera .
In general, flavonoids with a free hydroxyl group at the C-3 position of the heterocyclic ring are unstable under physiological conditions, and are therefore typically found as their glycosylated forms. Unlike the case in many plants, differential glycosylation occurs at the C-3 position for flavonols and anthocyanins of Actinidia. The flavonols identified were glucosides and rutinosides (resulting from addition of rhamnose to the glucose), while the main anthocyanins were galactosides and xylosyl-galactosides. Only low concentrations of anthocyanin-glucosides were found, and anthocyanin-rutinosides were not detected.
Three R2R3-MYB genes have been identified for Actinidia that show the conserved amino acid sequences of activators of anthocyanin biosynthesis – MYB10, MYB110a and MYB110b. MYB110a is thought likely to be the major gene defining petal colour as it was expressed at high levels in red but not white petals, and MYB110b expression was not detected at all. Expression for MYB110a (and to a lesser extent, MYB10) was highest before full bloom, and declined as the flower expanded. This pattern occurs with anthocyanin-related MYBs controlling other floral phenotypes, eg. petunia . The white-petalled phenotype of flowers was complemented through introduction of MYB110a, confirming its identity as an anthocyanin regulator. Furthermore, specific gene markers were obtained that linked different alleles of MYB110a to the presence of white versus red petals. The results show that MYB110a is the agent of specificity of anthocyanin expression in the flower petals while MYB110b is probably a tightly linked, but not expressed, relative of MYB110a. Fruit flesh and flower ovary and filament colour must be under the control of other R2R3-MYB gene family members, or alleles, as these did not segregate with MYB110a. Future work will concentrate on identifying fruit-specific R2R3-MYB markers for use in marker assisted breeding.
In transient assays, adding exogenous bHLH (AtTT8) had surprisingly little effect on the activation of the F3GT1 and AtDFR promoters by MYB110a. However, it is unlikely that MYB110a functions without a bHLH. Instead MYB110a may be more efficient at interacting with tobacco bHLHs already present in the transiently expressing cell, and this could be determined using hairpins to tobacco endogenous bHLH genes. Although the bHLH is thought to be the main component that links with the different WDR, R2R3- and R3-MYB proteins in the MBW regulatory complex, variation in bHLH expression is unlikely to provide for tissue-specific anthocyanin production. The results in our study of Actinidia support the proposal that variation in R2R3-MYB activator function and expression is the key determinant of spatial and temporal patterning of anthocyanin production in most plant species [15, 16].