2014 Volume 83 Issue 2 Pages 156-162
We compared anthocyanidin composition patterns and the expression of anthocyanin biosynthesis genes in Rhododendron kiusianum, R. kaempferi, and their natural hybrids from the Kirishima mountain mass. Compared with the habitat, phenotypic characteristics including tree height and flower color did not change in the transplanted individuals after cutting propagation. According to flower color measurements, R. kiusianum and R. kaempferi belonged to purple and red series, respectively, and their natural hybrids belonged to either the red or purple series. HPLC analysis showed that the petals of most R. kiusianum and natural hybrids contained both cyanidin and delphinidin series pigments, while the petals of R. kaempferi contained only cyanidin series pigments. However, one R. kiusianum individual in the purple series contained only cyanidin series pigments and one natural hybrid individual in the red series contained both cyanidin and delphinidin series pigments. These individuals were thought to be influenced by co-pigmentation or a lack thereof, respectively. All samples expressed F3′H, DFR, and ANS genes in real-time quantitative RT-PCR, and the F3′5′H gene was always expressed in samples containing delphinidin series pigments. These results suggest that the expression of F3′5′H is essential for R. kiusianun and its natural hybrids to produce delphinidin series pigments. This study showed that interspecific hybridization between wild species with purple series flowers and wild species with red series flowers varied the pigment composition and anthocyanin biosynthesis-related gene expression in the natural hybrids, suggesting that it caused the flower color variation in the wild populations in the Kirishima mountains.
Rhododendron kiusianum Makino is endemic to the upper areas of the Kirishima Mountains in Kyushu, Japan, and R. kaempferi Planch. is the most commonly occurring species at the bottom of these mountains. Putative natural hybrid populations of the two species are found in the intermediate region and have many flower colors and morphological variations (Sakata et al., 1993). These species are thought to be one of the ancestral species related to cultivar groups such as the Edo-Kirishima and Kurume azaleas (Kunishige and Kobayashi, 1980; Yamazaki, 1996).
PCR-RFLP analysis of chloroplast DNA can detect specific bands for R. kiusianum and R. kaempferi in the 16S rDNA region. A population of interspecific hybrids with flower color and morphological variations was shown to comprise individuals that had banding patterns of either R. kiusianum or R. kaempferi. However, some individuals in the populations of R. kiusianum that were morphologically indistinguishable from R. kiusianum had the cpDNA pattern of R. kaempferi, suggesting that cytoplasmic introgression had occurred from R. kaempferi (Kobayashi et al., 2000, 2007). In addition, wild azalea populations in the Kirishima mountains were divided into three clusters by cluster analysis of morphological characteristics, and each cluster coincided with the chloroplast DNA patterns of R. kiusianum, R. kaempferi, and their natural hybrids (Kobayashi et al., 2003).
R. kiusianum has small purple series flowers containing anthocyanins of the cyanidin and delphinidin series, whereas R. kaempferi has large red series flowers with dark blotches containing only cyanidin series anthocyanins. However, individuals on the slopes of the mountain mass show a wide variation in phenotypic characteristics, with flower colors ranging from red to purple series. According to pigmentation characteristics, the lower the altitude, the more the individuals resemble R. kaempferi, while the higher the altitude, the more individuals resemble R. kiusianum (Sakata et al., 1993). The variations in flower color and pigmentation characteristics among the flowers of R. kiusianum and R. kaempferi endemic to the mountain masses in Kyushu were the same as in the Kirishima mountain mass (Miyajima et al., 1995). Moreover, when individuals were classified into wild species and cultivars, the wild species were divided into two groups. One group was convergently distributed in the red flower group, which contained R. kaempferi and had only cyanidin series pigments. The other group was convergently distributed in the purple flower group, which contained R. kiusianum and had both cyanidin and delphinidin series pigments. Cultivars were distributed within and between these two groups (Mizuta et al., 2009).
In a study of genes related to pigment synthesis in evergreen azaleas, the partial- and full-length sequences of eight structural genes involved in the flavonoid biosynthetic pathway, including the flavonoid 3′-hydroxylase gene (F3′H), flavonoid 3′,5′-hydroxylase gene (F3′5′H), dihydroflavonol 4-reductase gene (DFR), and anthocyanidin synthase gene (ANS), were isolated from floral tissues of R. × pulchrum Sweet ‘Oomurasaki’ and their expression was investigated in the different developmental stages of the corolla (Nakatsuka et al., 2008). Moreover, pigment and gene expression analyses of R. × pulchrum ‘Oomurasaki’ and its flower color sport were performed. Low transcript levels of F3′5′H in the red flower sport resulted in no accumulation of delphinidin series pigments, and consequently a flower color change between ‘Oomurasaki’ and its sport (Mizuta et al., 2010). In this study, we investigated flower color, anthocyanidin composition, and anthocyanin biosynthetic gene expression using R. kiusianum, R. kaempferi and natural hybrids from the Kirishima Mountains sampled by Kobayashi (1997) and Kobayashi et al. (2000, 2003, 2007), to clarify the correlation between flower color variation and the expression of related anthocyanin synthesis genes.
Cuttings of evergreen azaleas, R. kiusianum, R. kaempferi, and their natural hybrids, were collected from the Kirishima mountain mass in 1993 and 1994 and propagated. These individuals were transplanted to the experimental field at Shimane University in 2004. For molecular analysis, each petal before flowering (candle stage) was frozen in liquid nitrogen and stored at −80°C until RNA extraction. For pigment analysis, the petal tissues were boiled in water at 100°C for 5 s and dried for 20 h at 40°C. These samples were subsequently stored in a desiccator at 4°C until HPLC analysis was performed. Fresh petals of all samples were collected for these analyses (excluding petals with blotches) in 2009.
Investigation of phenotypic characteristicsFlower and calyx size (diameter, width of corolla lobe, length of corolla tube, and length of calyx) were measured using five flowers per individual in 2009. Tree height was also recorded in 2010.
The color of the petals was recorded by photographs (Fig. 1) and by the RHSCC (Royal Horticultural Society Colour Chart), and were measured for lightness (L*) and two chromatic components a* and b* using a Color Reader (CR-10; Konica Minolta Sensing Inc., Tokyo, Japan) in 2009. Refer to Mizuta et al. (2009) for details.
Photographs of wild azalea flowers from individuals on the Kirishima mountains used in this study. The numbers correspond to those in Table 1.
The procedures used for pigment extraction and the analytical condition of HPLC used to investigate anthocyanidin composition were the same as previously described (Mizuta et al., 2009). The HPLC system used LC solution (on an LC workstation), an SPD-M20A UV-Vis photodiode array detector, an LC-20AD liquid chromatograph, and a CTO-20A column oven (Shimadzu Corp., Kyoto, Japan).
The same HPLC system was used for the analysis of flavonols and monitored at 360 nm. Existence of flavonols in each individual could be checked briefly although the above condition had been optimized for anthocyanidin analysis.
RNA extraction and cDNA synthesisTotal RNA was extracted using an RNeasy Plant Mini Kit (Qiagen GmbH, Hilden, Germany). To avoid DNA contamination, the total RNA from each petal was treated with RNase-free DNaseI (Takara Bio Inc., Shiga, Japan). PCR amplification was performed on the DNaseI-treated RNA using primers for the actin gene to confirm that the DNA was digested completely. The total RNA (1 μg) treated with DNaseI was reverse-transcribed by oligo (dT) and ReverTra Ace reverse transcriptase (Toyobo Co. Ltd., Osaka, Japan) according to the manufacturer’s instructions. The synthesized first-strand cDNA was then used for expression analysis.
Expression analysis by qRT-PCRPCR amplification was performed with gene-specific primers for F3′H (GenBank/EMBL/DDBJ Accession no: AB289597), F3′5′H (AB289598), DFR (AB289595), ANS (AB289596), and actin (AB610421) cDNAs as described by Nakatsuka et al. (2008). To investigate the expression of genes involved in anthocyanin biosynthesis, the cDNA was amplified with a Thermal Cycler Dice Real Time System and SYBR Premix Ex Taq II (Takara Bio Inc.). Amplification of actin cDNA under identical conditions was also conducted as an internal control to normalize the levels of cDNA. Thermal cycling conditions were 10 s at 95°C, followed by 40 cycles of 5 s at 95°C and 30 s at 60°C. The cycle threshold (Ct) value for each PCR reaction was calculated. After completion of the amplification steps, the melting curve was determined for each analysis, followed by electrophoresis of the PCR products to confirm the specific amplification of each real-time quantitative RT-PCR (qRT-PCR).
The mean values for flower diameter, length of the corolla tube, width of the corolla lobe, and length of the calyx of R. kaempferi were almost twice those of R. kiusianum (Table 1; Fig. 2). Flower diameter and the width of the corolla lobe of the natural hybrids (but not length of corolla tube and length of calyx) showed intermediate values between those of both species (Table 1; Fig. 2A). For all flower morphologies, the largest variation among individuals was in the natural hybrids compared to the two species. There was no difference in the number of stamens between R. kiusianum, R. kaempferi, and the natural hybrids (data not shown).
Morphological characteristics of evergreen azaleas derived from the Kirishima mountains.
The relationship between flower diameter and width of the corolla lobe (A) or length of the corolla tube and length of the calyx (B) in wild azalea populations derived from the Kirishima mountains.
In terms of plant height, R. kaempferi (excluding T-600-2 and Ki-440-3) and the natural hybrids were not less than 125.0 cm, whereas individuals of R. kiusianum were not more than 82.0 cm. Individuals of T-600-2 and Ki-440-3 were smaller in height because they were planted in pots (Table 1).
Flower color analysisThe RHSCC numbers of R. kiusianum, except for the white flowers, ranged from 73C (Red-purple group) to 78C (Purple group) in the purple series, and the RHSCC numbers of R. kaempferi ranged from 39A (Red group) to 47C (Red group) in the red series. The RHSCC numbers of the natural hybrids ranged from 39A (Red group) to 74C (Red-purple group) in the red and purple series. Hence, the flower color of the natural hybrids covered a wide range (Table 2).
RHS colour chart codes, CIEL*a*b* color coordinates, percentage of anthocyanidin compositions and presence of flavonol in wild evergreen azalea petals derived from the Kirishima mountains.
Distributions based on two chromatic components, a* and b* values, of R. kiusianum and R. kaempferi showed a convergent distribution. The natural hybrids were distributed over a wide range within the b* value range of these species (Table 2; Fig. 3).
Distribution of flower color based on CIEL*a*b* coordinates for wild azaleas derived from the Kirishima mountains.
According to HPLC analysis, the petals of R. kiusianum (except T-1210-3) contained red cyanidin series pigments (cyanidin and peonidin) and blue delphinidin series pigments (delphinidin, petunidin, and malvidin), whereas all petals of R. kaempferi contained only cyanidin series pigments (Table 2).
Among the natural hybrid individuals in the purple and red series, the petals of T-800-15 contained only cyanidin series pigments, while the others contained both cyanidin and delphinidin series pigments. Interestingly, cyanidins and delphinidins were also detected in the white petals of R. kiusianum (Table 2).
All samples of R. kiusianum contained flavonol pigments. In the natural hybrids, petals of S-940-10 and T-800-10 contained flavonols, whereas T-800-13 and T-800-15 contained no flavonols. Flavonol pigments of all R. kaempferi were absent (Table 2).
Anthocyanin synthesis-related gene expression analysesThe expressions of four genes (F3′H, F3′5′H, DFR, and ANS) in the petals of the azalea plants were investigated by qRT-PCR. F3′H, which is involved in the synthesis of cyanidin series pigments, was expressed in all individuals (Fig. 4). F3′5′H, which is involved in the synthesis of delphinidin series pigments, was expressed in R. kiusianum except for T-1210-3 (Sample No. 1). Only one natural hybrid, T-800-15 (No. 9), did not express F3′5′H. R. kaempferi T-600-8 (No. 11) and Ki-440-3 (No. 14) did not express F3′5′H, whereas R. kaempferi of T-600-2 (No. 12) and Ki-520-8 (No. 13) did (Fig. 4). DFR and ANS were expressed in all individuals, as was F3′H (Fig. 4).
Relative expression of flavonoid biosynthesis genes in the petals of wild azaleas derived from the Kirishima mountains by qRT-PCR analysis. Sample numbers correspond to those in Table 1, 2 and Figure 1. Quantitative RT-PCR amplification of actin was used to normalize the gene expression under identical conditions. Shaded and white columns indicate colored and white flower samples, respectively. The vertical bars represent SE of the means of three replications.
F3′H was expressed at a lower level than F3′5′H, DFR, and ANS. R. kiusianum with white flowers (Nos. 5, 6) expressed all four genes analyzed, and there was no difference in the gene expression trends compared with other colored individuals (Fig. 4).
The height of R. kaempferi individuals was equivalent to the natural hybrids, except for T-600-2 and Ki-440-3, which were planted in pots. The R. kiusianum plants were comparatively shorter. These results followed the same trends as the research data of Kobayashi (1997) and Kobayashi et al. (2000). The heights of the cultivated azaleas after cutting propagation were similar to those of their wild counterparts in the Kirishima mountains habitat.
As described by Sakata et al. (1993) and Kobayashi et al. (2000), each natural hybrid individual showed an intermediate phenotype between R. kiusianum and R. kaempferi. According to flower color measurements using the RHSCC, compared with the individual data of Kobayashi (1997), the data in this study showed no change in color category. This confirmed that the flower traits of the azaleas were mainly controlled by genetic factors.
Mizuta et al. (2009) performed flower color measurements of various wild species and cultivars, and similar results were obtained in this study. The wild species were divided into two groups that showed convergent distributions in the areas of positive and negative b* values. The cultivars were distributed within the b* value range of these wild species. In this study, according to flower color measurements using the RHSCC and a color reader, the largest variation among individuals was in the natural hybrids, compared with both R. kiusianum (except for those with white flowers) and R. kaempferi. This result was in agreement with the previous report by Mizuta et al. (2009).
In terms of the correlation between flower color and pigment composition, R. kiusianum and natural hybrids with purple series flowers (except for T-1210-3) contained both cyanidin and delphinidin series pigments, whereas R. kaempferi and natural hybrids with red series flowers (except for T-800-15) contained only cyanidin series pigments. These results followed the same trends reported by Sakata et al. (1991, 1993). Moreover, we compared the anthocyanidin constitution and the expression pattern of the F3′5′H gene, which is related to the synthesis of delphinidin series pigments. We found that expression of the F3′5′H gene corresponded with the presence of delphinidin series pigments. These results suggested that F3′5′H gene expression was necessary for individuals to synthesize delphinidin series pigments. Mizuta et al. (2010) described how radically decreased transcript levels of the F3′5′H gene resulted in no accumulation of delphinidin series pigments, and caused a flower color change between ‘Oomurasaki’ and its red-flowered sport. F3′H is a key gene that induces cyanidin series pigments, much like the F3′5′H gene induces delphinidin series pigments. The reason why R. kaempferi T-600-2 and Ki-520-8 contain no delphinidin series pigments, despite expressing the F3′5′H gene, may be because of the low transcription rate of F3′5′H compared with the expression ratio between F3′H and F3′5′H in individuals with delphinidin series pigments (Table 2; Fig. 4). It was found that these individuals had the cpDNA pattern of R. kaempferi (Kobayashi et al., 2003). However, it is necessary to analyze the DNA structure of the F3′5′H gene in R. kiusianum, R. kaempferi, and these individuals, including speculation of the possibility of introgressive hybridization between R. kiusianum expressing the F3′5′H gene and R. kaempferi in the past, as Kobayashi et al. (2000, 2007) described.
Although the individual T-1210-3 did not express F3′5′H and contained no delphinidin series pigments, it belonged to the red-purple group of the purple series. The b* value of this individual was −10.2, which indicated blue. In addition, the natural hybrid T-800-13 expressed the F3′5′H gene and contained delphinidin series pigments but belonged to the red group. The b* value of this individual was +9.0, which indicated yellow. These samples showed no correspondence between flower color and anthocyanidin pigment constitution (Table 2). In azalea, it is known that co-pigmentation between anthocyanins and flavonols has a blueing effect (Asen et al., 1971; De Loose, 1978; Mizuta et al., 2009; Sakata et al., 1993; Umeki and Inazu, 1989). We simply checked the existence of flavonol pigments in each individual. Although the petals of T-1210-3 lacked blue delphinidin series pigments, they contained flavonols. This suggests that T-1210-3 has a co-pigmentation effect because of flavonols. Conversely, T-800-13 was considered not to have a co-pigmentation effect because it lacks flavonols; De Loose (1978) reported that the carmine red flower color of the R. simsii Planch. cultivar was produced by co-pigmentation between cyanidin glycosides and flavonol glycosides or might also occur when delphinidin glycosides occurred in the absence of flavonol glycosides (pH of juice squeezed from petals was less than 4).
In our previous study, the petals of white-flowered samples, including R. ripense Makino, R. kaempferi, and Kurume azalea, showed peaks in anthocyanidin analysis although none was detected in anthocyanin analysis, showing that white flowers of evergreen azalea contain the precursors to anthocyanin (Mizuta et al., 2009). Bate-Smith (1953) reported that an extract of white flowers of Camellia japonica changed to red after heating with hydrochloric acid, which suggested that leuco-anthocyanin was present in the white flowers. Here, peaks were detected in anthocyanidin analysis of the white flowers of R. kiusianum (T-1450-2 and T-1100-1) (Table 2). This result showed that the white flowers of R. kiusianum also contain anthocyanin precursors.
This study showed that interspecific hybridization between wild species with purple series flowers and wild species with red series flowers resulted in variations of pigment composition and anthocyanin synthesis-related gene expression. Our results suggest that this variation caused flower color variations in wild populations, such as in the Kirishima mountain mass.
