Abnormal flowers are often induced by infection of certain plant pathogens, e.g. phytoplasma, but the molecular mechanisms underlying these malformations have remained poorly understood. Here, we show that infection with OY-W phytoplasma (Candidatus Phytoplasma asteris, onion yellows phytoplasma strain, line OY-W) affects the expression of the floral homeotic genes of petunia plants in an organ-specific manner. Upon infection with OY-W phytoplasma, floral morphological changes, including conversion to leaf-like structures, were observed in sepals, petals and pistils, but not in stamens. As the expression levels of homeotic genes differ greatly between floral organs, we examined the expression levels of homeotic genes in each floral organ infected by OY-W phytoplasma, compared with healthy plants. The expression levels of several homeotic genes required for organ development, such as PFG, PhGLO1 and FBP7, were significantly downregulated by the phytoplasma infection in floral organs, except the stamens, suggesting that the unique morphological changes caused by the phytoplasma infection might result from the significant decrease in expression of some crucial homeotic genes. Moreover, the expression levels of TER, ALF and DOT genes, which are known to participate in floral meristem identity, were significantly downregulated in the phytoplasma-infected petunia meristems, implying that phytoplasma would affect an upstream signaling pathway of floral meristem identity. Our results suggest that phytoplasma infection may have complex effects on floral development, resulting in the unique phenotypes that were clearly distinct from the mutant flower phenotypes produced by the knock-out or the overexpression of certain homeotic genes.
Phytoplasmas (class Mollicutes, genus Phytoplasma) are plant pathogenic bacteria that cause devastating damage to more than 700 plant species worldwide (Lee et al., 2000; Bertaccini, 2007). Phytoplasmas are obligate parasites, whose habitat is restricted to phloem tissue in their host plant (Christensen et al., 2005; Hogenhout et al., 2008). The infected plants show a wide variety of symptoms, including witches’ bloom, stunting and generalized yellowing. Particularly in floral organs, phytoplasma infection often induces unique morphological changes, such as phyllody (metamorphosis of the floral organs to leaf-like structures), virescence (green coloration of petals) or proliferation (vegetative growth where floral organs should develop). The phyllody and virescence symptoms were considered attractive and valuable about 1000 years ago, and the Chinese imperial court received a special annual tribute composed of green blossoms (Strauss, 2009). Moreover, even now, hydrangea plants with these symptoms are sold in markets in Japan. Clarifying the molecular mechanisms of these symptoms is very interesting and important for the ornamental plant industry. From a biological and evolutional viewpoint, these unique symptoms are thought related to the extension of phloem tissue in which the phytoplasmas are localized. For example, Japanese hydrangea phyllody (JHP) phytoplasmas were also observed in phloem tissues of the leaf-like pistil apex deformed by infection, whereas it is well known that phytoplasmas do not exist in the shoot apical meristem (Arashida et al., 2008). However, the molecular mechanisms of such malformations in floral organs induced by phytoplasma infection remain unclear.
A typical angiosperm flower consists of four organs: sepals, petals, stamens and pistils (including carpels and ovules). Genetic and molecular studies on floral organ development have been performed mostly in Arabidopsis thaliana, Antirrhinum majus and Petunia hybrida (Robles and Pelaz, 2005). These studies revealed that floral organ identity is specified by the combinatorial expression of floral homeotic genes encoding transcription factors. An ABC model of flower development was originally formulated to explain the specification of floral organ identity through a combination of floral homeotic genes (Bowman et al., 1991; Coen and Meyerowitz, 1991). Subsequently, by the addition of class-D and -E genes (Angenent and Colombo, 1996; Pelaz et al., 2000; Honma and Goto, 2001), the ABC model was extended to the ABCDE model (also called the ‘floral quartet model’; Theissen and Saedler, 2001). This revised model provides a molecular mechanism for the specification of floral organs: the development of each floral organ is determined by the interaction between members of different classes of transcription factors. Namely, class-A and -E genes are required to specify sepals; similarly, classes A + B + E specify petals, classes B + C + E specify stamens, classes C + E specify carpels and classes D + E specify ovules (Theissen and Melzer, 2007).
With advances in the understanding of the molecular genetic mechanisms underlying floral organ identity, recent studies have focused on the relationships between the phytoplasma infection and the expression of floral homeotic genes. Pracros et al. (2006) reported that stolbur phytoplasma infection affected the expression of some floral development genes in tomato plants that showed flower malformations. Moreover, in the JHP phytoplasma-infected hydrangea, the expression levels of class-A, -B and -C genes were suppressed only in the early stage of floral organ development (Kitamura et al., 2009), implying that the temporary suppression of these genes may be related to abnormal floral organ morphogenesis of hydrangea, such as the homeotic conversion of sepals or pistils to leaves. However, these expression analyses of floral development genes do not explain all flower malformations caused by phytoplasma infection. For example, although the expression levels of class-A, -B and -C genes were suppressed in the phytoplasma-infected hydrangea, no structural change was observed in petals or stamens (Kitamura et al., 2009). As the expression levels of homeotic genes differ greatly between floral organs (Soltis et al., 2007), expression analyses using the whole flower meristem may obscure floral organ-specific changes in gene expression.
Here, we divided phytoplasma-infected petunia flower meristems into four floral organs, and examined the expression patterns of floral homeotic genes of class A–E in each of these four organs. The petunia plant has been well studied as a model for the determination of floral organ identity, and has an adequate flower size for experimentation. We show the downregulation of floral homeotic genes in the morphologically changed organs in phytoplasma-infected plants. Moreover, the effects of phytoplasma infection on floral development are discussed.
Disease symptoms of OY-W phytoplasma in petunia
Petunia plants infected with OY-W phytoplasma (Candidatus Phytoplasma asteris, onion yellows phytoplasma strain, line OY-W) showed stunting, witches’ brooms or flower malformations (Figure 1). Because a petunia flower is a cymose inflorescence, the apical inflorescence meristem (IM) develops into the floral meristem (FM) and the lateral IM, resulting in the formation of zigzagged stems, as seen in healthy plants. However, OY-W phytoplasma infection initiated defective IMs, and flowers even developed from non-apical meristems, leading to bushy plants (Figure 1).
To further investigate the flower malformations, we observed each floral organ in detail. All floral organs showed dwarf phenotypes, and the organs except stamens showed severe abnormalities (Figure 2a–d,p). Whereas healthy sepals have an elongated shape, infected sepals have a rounded shape like leaves (Figure 2a). OY-W phytoplasma-infected petals exhibited hypopigmentation and green coloration along the midveins (Figure 2b,j–l). The most severe malformation was observed in pistils, such as defects in carpel fusion (Figure 2h), formation of trichome (Figure 2f) and loss of the style (Figure 2d). In addition, normal ovule development was totally disrupted in the OY-W-infected ovary, and some leaf-like structures occurred in place of normal ovules (Figure 2m–o). In several infected plants with severe symptoms, all floral organs, except the stamens, appeared to be further converted to leaf-like structures (Figure 2p,q): in particular, pistils were almost entirely converted to leaves, which is a typical ‘proliferation’ symptom.
In our study, we used healthy and infected petunias under controlled conditions, as described in Figure S1. Under these conditions, the times of flowering initiation and opening of the flowers from germination were the same in healthy and phytoplasma-infected petunias. The formation of the flower meristems was initiated approximately 8 weeks after germination in both healthy and phytoplasma-infected petunias, and no delay occurred in the development of infected flowers. Under these conditions, very mild phenotypic differences were observed between healthy and infected plants at around the 8th week. After the 10th week, the first clear phenotypic differences were detected, e.g. yellowing, a bushy phenotype caused by increasing shoot numbers (witches’ broom), dwarfism and flower malformation. Moreover, if insect feeding started before 4 weeks after germination (see the schedule in Figure S1), most of the phytoplasma-infected petunias exhibited flowerless symptoms (Figure S2), and several plants died from the damage caused by insect feeding.
Expression patterns of ABCDE genes in each floral organ
Real-time RT-PCR analyses indicated that the expression levels of most homeotic genes were significantly changed upon infection with OY-W, whereas the expression of all homeotic genes in healthy plants agreed with the reported results for wild-type petunia (Figures 3 and S3). Expression levels of two class-A genes, PFG and PhAp2A, were significantly reduced in sepals (to 19 and 20%, respectively). Expression levels of two class-B genes, PhGLO1 and PhGLO2, were reduced in petals (to 76 and 53%, respectively). Although PhDEF expression in petals was normal, that in stamens was significantly upregulated (4.5-fold). Expression levels of two class-C genes, pMADS3 and FBP6, were upregulated in pistils (to 3.3- and 1.5-fold, respectively). On the other hand, expression levels of two class-D genes, FBP7 and FBP11, were markedly downregulated in pistils (to 19 and 55%, respectively). Expression levels of two class-E genes, FBP2 and FBP5, were upregulated in stamens (by 8.5- and 8.2-fold, respectively) (Figure 3). Taken together, these results indicate that the phytoplasma infection induced similar effects for the same class genes: the expression of class-A genes was reduced in sepals, the expression of class-B GLO/PI-lineage genes was reduced in petals, the expression of class-C genes was increased in pistils, the expression of class-D genes was reduced in pistils and the expression of class-E genes was increased in stamens.
Changes in expression levels of floral identity genes
To examine whether phytoplasma infection influences early stage meristems, the expression levels of floral meristem identity (FMI) genes that restrict transcription of the ABCDE homeotic genes were examined in flower meristems of OY-W-infected and healthy petunia. TERMINATOR (TER), ABERRANT LEAF AND FLOWER (ALF) and DOUBLE TOP (DOT) genes, which are homologs of Arabidopsis WUSCHEL (WUS), LEAFY (LFY) and UNUSUAL FLORAL ORGANS (UFO), respectively, were used in this study (Stuurman et al., 2002; Souer et al., 2008). Quantitative analyses revealed that expression levels of all these genes, TER, ALF and DOT, were significantly downregulated to 25, 20 and 35%, respectively, in OY-W-infected flower meristems (Figure 4).
Homeotic genes are involved in phytoplasma symptoms
In this study, we clarified that OY-W phytoplasma infection affected the expression of the floral homeotic genes in petunia plants. Moreover, our study also revealed that in the floral organ exhibiting morphological changes, the expression level of some homeotic genes that were required for organ development was significantly downregulated (Figure 5). For example, as class-A and -E genes are known to be involved in the morphogenesis of sepals, in the malformed sepals of the OY-W-infected plant, the expression of class-A genes was significantly downregulated (Figure 5). Similarly, the expression levels of class-B (GLO/PI lineage) and class-D genes were significantly downregulated in petals and pistils, respectively (Figure 5). According to the floral quartet model, the identity of the individual organ is determined by the binding of floral homeotic protein complexes to promoter regions of target genes (Theissen and Saedler, 2001). When expression of one transcription factor was decreased, it was thought that normal development of floral organs could not occur (Pelaz et al., 2000; Theissen and Saedler, 2001). Therefore, it is suggested that the unique morphological changes in OY-W-infected flowers could result from a significant decrease in expression of one or more crucial homeotic genes. Although it has been reported that the phytoplasma infection affected the expression of some floral development genes in the whole flower meristem of Solanum lycopersicum (tomato) and hydrangea plants (Pracros et al., 2006; Kitamura et al., 2009), it has remained unclear how the floral homeotic genes in each floral organ are regulated by phytoplasma infection. Transcriptional regulation during phytoplasma infection may be obscured by expression analyses using the whole flower meristem, because the expression levels of homeotic genes differ greatly between floral organs (Soltis et al., 2007). In fact, we did an experiment to examine the expression levels of ABC genes in entire flower meristems, and as a result, no genes showed any significant differences between healthy and infected plants (Figure S4). In this study, because the expression levels of homeotic genes were examined in each floral organ, detailed expression changes of homeotic genes could be examined, and the causes of morphological changes resulting from OY-W phytoplasma infection could be explained by changes in homeotic gene expression.
In OY-W-infected stamens, the expression levels of class-E genes, which are required for the development of stamens, were much higher than those in healthy stamens. However, no morphological change was observed in OY-W-infected stamens. According to the floral quartet model, as multiple homeotic proteins form a complex and function as a transcription factor (Theissen and Saedler, 2001), the overexpression of one of the homeotic genes would not affect the development of floral organs. In fact, it was previously reported that the overexpression of homeotic genes involved in the morphogenesis of petals (PI, AP3, AP1 or SEP3) did not change the petal phenotype in Arabidopsis, whereas it converted leaves into petals (Honma and Goto, 2001). Similarly, in the OY-W-infected stamens, the overexpression of class-E genes might not affect morphological development.
Relationship between the green-petal symptoms and expression of class-B genes
In petunia, the class-B mutant phglo1 exhibits greenish petals (Vandenbussche et al., 2004). The morphologies of infected petals observed in Figure 2 were very similar to those of the phglo1 mutant. In addition, PhGLO1 expression in OY-W-infected petals was significantly downregulated (Figures 3 and S3). Taken together, these results suggest that the green coloration of infected petals may be associated with decreased expression of PhGLO1. Previously, the decreased expression of class-B genes was observed in both stolbur phytoplasma-infected tomato and JHP phytoplasma-infected hydrangea, and these two phytoplasmas caused phyllody or virescence symptoms in petals (Pracros et al., 2006; Kitamura et al., 2009). Thus, phytoplasma symptoms in petals could be commonly associated with the downregulation of class-B genes.
On the other hand, a loss-of-function mutant of class E is also known to exhibit green-petal symptoms in Arabidopsis and petunia (Pelaz et al., 2000; Rijpkema et al., 2009). However, OY-W phytoplasma infection did not reduce class-E FBP2 gene expression in petals (Figure 3). In petunia, the class-E mutant fbp2 exhibited a progression of petal greening from the edges and midveins of petals (Vandenbussche et al., 2003), whereas the class-B mutant phglo1 exhibited a progression of petal greening from only the midveins of petals (Vandenbussche et al., 2004). In OY-W-infected flowers, the green coloration of petals progressed only from midveins (Figure 2), which was similar to the phenotype of the class-B mutant. Therefore, these data suggest that class-B genes are likely to be involved in petal virescence symptoms upon infection with OY-W phytoplasma.
Relationship between pistil symptoms and expression of class-C and -D genes
In the OY-W-infected petunia pistils (including carpels and ovules), the class-D genes that are required for ovule development were significantly decreased (Figure 3). This downregulation is in good agreement with the malformed ovule, which was replaced by leaf-like tissue (Figure 2). In contrast, although carpels were changed to leaf-like structures (Figure 2), the expression of class-C and -E genes, which are required for carpel development, was not decreased in the pistils (Figure 3). It has been reported that class-C genes (AG and SHP) and class-D genes have redundant roles in ovule identity (Pinyopich et al., 2003); however, the Arabidopsis class-D gene STK has not been demonstrated to participate in carpel development. Thus, the carpel phenotype could not be explained by known mechanisms in previous studies of mutant analysis. Our results implied that there might be other mechanisms involved in carpel development. We cannot, however, rule out the possibility of the downregulation of class-C genes in infected carpels, as in our study it is unclear whether the tissue with high expression level of class-C genes would be carpel or ovule.
Downregulation of genes involved in activating ABCDE genes
The expression levels of TER, ALF and DOT genes, which participate in floral meristem identity, were significantly downregulated in OY-W-infected petunia meristems (Figure 4). These results imply that phytoplasma infection affects an upstream signaling pathway of floral meristem identity. In fact, most petunias that had been infected with OY-W phytoplasma at a younger stage exhibit a defect in the transition from the IM to the FM (Figure S2), indicating that phytoplasma infection could affect the FM identity. It has been reported that co-expression of ALF and DOT activates expression of class-A, -B, -C, -D and -E genes in petunia (Souer et al., 2008). Accordingly, the downregulation of class-A and -C genes in infected petunia might be caused by the reduced expression of ALF and DOT. The upregulation of class-D and -E genes, however, could not be explained by reduced expression of ALF and DOT, suggesting that phytoplasma infection might have complex effects on floral development.
It was previously reported that the earlier activation of the AG ortholog could downregulate the WUS ortholog in stolbur phytoplasma-infected tomato (Pracros et al., 2006). Likewise, our results also showed an increased expression of pMADS3 and FBP6 (orthologs of AG) and a decreased expression of TER (ortholog of WUS) in OY-W-infected petunia. TER is thought to be critical for stem cell maintenance, and petunia ter mutants had very bushy phenotypes similar to an Arabidopsis wus mutant (Laux et al., 1996; Stuurman et al., 2002). The bushy phenotype was also observed in the phytoplasma-infected petunia (Figure 1), raising the possibility that this might be caused by the decreased expression of WUS.
Complex effects on floral development of phytoplasma infection
The overall phenotypes of phytoplasma-infected flowers, such as phyllody, virescence and proliferation, were clearly distinct from those of mutant flowers with knock-out or overexpression of certain homeotic genes, although both phenotypes were somewhat similar in some organs. For example, class-B mutants show homeotic conversion of petals to sepals, as in OY-W-infected petals, but they also exhibit homeotic conversion of stamens to carpels, unlike OY-W-infected stamens (Figure 2; Vandenbussche et al., 2004).
This difference could be attributed to the complex effects of phytoplasma infection on homeotic gene expression. In fact, the expression changes in the floral homeotic genes caused by OY-W infection varied among floral organs (Figure 5). For example, in a class-A gene mutant, the expression of class-A genes is uniformly lost in all organs, and class-C genes are expressed in sepals as a result of antagonistic interactions between class-A and -C genes (Drews et al., 1991; Tsuchimoto et al., 1993). On the other hand, in OY-W-infected petunia, although the expression of class-A genes was significantly decreased in sepals, the expression of class-C genes was increased only in pistils, but not in sepals. Taken together, our results suggest that phytoplasma infection has complex effects on floral development, resulting in unique morphological changes to floral organs.
Petunia hybrida, cultivar Vakara Rose and Blue (Sakata Seeds Ltd, http://www.sakata.com) was used in this study. Petunia plants were grown from seed and maintained under natural glasshouse conditions (at 25°C). A wild strain of onion yellows phytoplasma (OY-W) was isolated in Saga Prefecture, Japan (Shiomi et al., 1996). OY-W phytoplasma has been maintained in a plant host, Chrysanthemum coronarium (garland chrysanthemum), using a leafhopper vector, Macrosteles striifrons, as described previously (Oshima et al., 2001). The experimental design for phytoplasma infection is summarized in Figure S1. Four weeks after germination, all of the petunia plants were covered with plastic and mesh cages, and approximately 20 healthy or OY-W phytoplasma-infected leafhoppers were added to each cage. After 7 days, all of the leafhoppers were removed. Note that the ‘healthy petunias’ used in all of the qPCR experiments were petunia plants that were fed on by healthy leafhoppers for 1 week. About 1 month later, petunia plants with distinct symptoms were used for the subsequent experiments. Under these controlled conditions, the times of flower initiation and flower opening after germination were similar in both healthy and infected petunias. The flower meristems began to form ca. 8 weeks after germination, and the flowers opened ca. 10 weeks after germination in both healthy and infected petunias.
Pistils from floral buds and sepals were fixed for 4 h in the fixative solution FAA (formaldehyde, 4% w/v; ethanol, 50% v/v; acetic acid, 5% v/v) at 25°C. Every hour, the fixative was replaced and a vacuum (−0.9 bar) was applied for 15 min. Fixed tissues were embedded in Paraplast Plus (Sherwood Medical, http://www.sherwood-scientific.com), and sectioned to 12-μm thickness with a microtome (PR-50; Yamato Kohki, http://www.yamato-web.co.jp). For observation, sections were stained with 0.1% (w/v) safranin O (Nacalai Tesque, http://www.nacalai.co.jp).
Total RNA was extracted using an Isogen kit (Nippon Gene, http://nippongene.com), according to the manufacturer’s instructions. For analysis of homeotic gene expression levels in different flower organs (Figure 3), total RNAs of each floral organ were sampled separately from the flower buds 1 day before emergence, both in healthy and OY-W phytoplasma-infected plants. Such state of flowers was easily identified by visual inspection: the color of the flower buds was completely changed and the front edges of flower buds were slightly opened. RNA extraction was replicated three times for every floral organ. To examine the expression levels of TER, ALF and DOT genes (Figure 4), total RNAs of floral meristems were extracted from flower meristems before the differentiation of floral organs. These meristems developed at the branching regions (proximal region) of flower buds in which the floral organs have already differentiated. This extraction was replicated five times, independently.
Total isolated RNA was treated with RNase-free DNase I (TaKaRa, http://www.takara-bio.com), and 2 μg of total RNA was reverse transcribed using the High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems, http://www.appliedbiosystems.com), following the manufacturer’s protocols. A 1-μl aliquot of 50-fold diluted cDNA was used for real-time PCR using the Thermal Cycler Dice real-time PCR system (TaKaRa) with SYBR Premix ExTaq (TaKaRa). The primers used for real-time PCR are shown in Table 1. Real-time RT-PCR analyses were performed three times for each RNA sample extracted.
For more accurate and reliable normalization of gene expression data, we compared the expression levels of four commonly used housekeeping genes (PhGAPDH, PhACTIN, PhEF1a and PhHistonH4) between healthy and phytoplasma-infected flower meristems (Figure S5 and Table S1). Every flower meristem was sampled five times, independently, before the differentiation of floral organs. Gene expression levels of each gene were divided by the geometric mean of the expression levels of three genes (PhGAPDH, PhACTIN and PhEF1a), as described in Vandesompele et al. (2002). As a result, the expression levels of PhACTIN and PhHistonH4 varied between healthy and infected meristems, but PhGAPDH and PhEF1a showed only minor variations between healthy and infected meristems, and PhGAPDH showed a lower standard deviation than PhEF1a (Figure S5). Thus, in this study, the PhGAPDH gene was used for normalization.
The data were expressed as the means ± SDs obtained for at least three independent determinations. Statistical analysis was performed using the F-test, followed by Welch’s t-test because P < 0.05 by the F-test in all statistical analyses. P values < 0.05 were considered to be statistically significant.
We thank Dr Ryo Arashida (Tokyo, Japan) for his excellent technical advice. This work was supported by Grants-in-Aid for Scientific Research from the Japan Society for the Promotion of Science (category A of Scientific Research Grant 21248004), and by the Program for Promotion of Basic Research Activities for Innovative Bioscience (PROBRAIN).