Jasmonate regulates juvenile-to-adult phase transition in rice

After germination, higher plants continuously produce vegetative organs, such as leaves, stems and roots, before the onset of reproductive development. The vegetative period can be divided into juvenile and adult stages, which are distinguished by changes in many physiological and morphological traits (reviewed by Poethig, 2013). It is believed that plants acquire reproductive competence during the adult stage of vegetative development.

The juvenile-to-adult phase transition is controlled by two major regulators: microRNA156 (miR156) and their target transcription factors, SQUAMOSA promoter-binding protein-like genes (SPLs). miR156, one of the most abundant microRNAs, is conserved among all land plants, including mosses (Axtell and Bowman, 2008). Recent studies have shown that miR156 accumulation reaches the highest level in seedlings, then decreases gradually as the plant develops, resulting in a gradual increase in target SPL transcripts in a number of herbaceous and woody species (Wu and Poethig, 2006; Chuck et al., 2007; Wang et al., 2009, 2011; Wu et al., 2009; Tanaka et al., 2011; Yoshikawa et al., 2013; Hudson et al., 2014). Plants constitutively overexpressing miR156 share common phenotypes in several species, for example, a prolonged juvenile phase, accelerated leaf initiation and delayed flowering (Schwab et al., 2005; Wu and Poethig, 2006; Xie et al., 2006, 2012; Zhang et al., 2011; Chuck et al., 2011). Conversely, plants with endogenous miR156 function disrupted by mimicry targets or by T-DNA insertions exhibit opposite phenotypes, e.g. precocious adult traits (Franco-Zorrilla et al., 2007; Wu et al., 2009; Todesco et al., 2010; Yang et al., 2013; Yu et al., 2013). These results indicate that the appropriate miR156 level is necessary and sufficient for maintenance of the juvenile phase. Genetic studies have also shown that mutants precociously exhibiting adult characteristics have low miR156 expression, whereas mutants with an elongated juvenile phase overexpress miR156 (Smith et al., 2009; Li et al., 2012; Chuck et al., 2007; Tanaka et al., 2011). Such studies support the importance of miR156 levels in maintaining the juvenile phase. miR156 induces mRNA cleavage and/or inhibits the translation of target SPLs, which are crucial transcription factors for phase transition (Wu and Poethig, 2006; Yang et al., 2012). Although miR156 appears to be a major regulator of SPL expression and translation, recent studies have provided evidence that the expression or activity of SPLs is also regulated in a miR156-independent manner in response to environmental cues or the phytohormone gibberellin (Wang et al., 2009; Jung et al., 2012; Yu et al., 2012).

In addition to miR156, miR172 and its target genes are thought to play important roles in juvenile-to-adult phase transition. Unlike miR156, miR172 accumulation is low in the juvenile phase and then gradually increases. miR172 acts as a repressor of APETALA2-LIKE (AP2-like) transcription factors to promote adult epidermal identity, floral development and flowering (Aukerman and Sakai, 2003; Chen, 2004; Chuck et al., 2007; Jung et al., 2007; Lauter et al., 2005; Wu et al., 2009). Plants overexpressing miR156 have decreased miR172 expression, whereas plants with reduced expression of miR156 have increased levels of miR172 (Wu et al., 2009). In Arabidopsis, overexpression of miR156-resistant SPL9 caused an accumulation of miR172, which is a direct target of SPL9, and early flowering (Wang et al., 2009; Wu et al., 2009).

The hormone gibberellic acid (GA) is involved in vegetative phase change and regulates flowering in some plants. For example, in ivy, when the apical bud is removed from an adult branch, the subtending axillary buds will grow into typical adult branches. However, if GA is applied to the axillary buds, they will develop into juvenile branches (Rogler and Hackett, 1975; Poethig, 2013). It has been also reported that GA-deficient mutants exhibit delayed transition to the adult phase in Arabidopsis and maize (Telfer et al., 1997; Telfer and Poethig, 1998; Evans and Poethig, 1995). Several recent studies have reported that certain DELLA proteins, which act as transcriptional repressors of the GA response, interact with and repress SPL protein activity (Yu et al., 2012). Furthermore, the GA-mediated degradation of DELLA under long-day conditions allows the SPL-mediated activation of miR172 and floral identity genes, such as FT and MADS box genes, resulting in floral transition (Yu et al., 2012; Galvão et al., 2012). miR156 expression is influenced by auxin and ethylene, and the stability of DELLAs is also regulated by several other phytohormones, including auxin, ethylene and abscisic acid (Vanstraelen and Benková, 2012; Marin et al., 2010; Zuo et al., 2012). Thus, these phytohormones might also contribute to the regulation of miR156 and SPLs.

In rice, a number of traits are altered during the juvenile-to-adult phase transition, including the size and shape of leaf blades, the size and shape of the shoot apical meristem (SAM), the presence of midribs, plastochron, photosynthetic rate, and miR156 and miR172 expression levels (Asai et al., 2002; Itoh et al., 2005; Tanaka et al., 2011). Based on these observations, it is proposed that the juvenile stage comprises the first and second leaf stages, the juvenile-to-adult transition stage spans the third to fifth leaf stages, and afterwards the rice plant is considered an adult. To date, two mutants have been reported that are defective in the juvenile-to-adult phase of vegetative development in rice: mori1 and peter pan syndrome (pps). The mori1 mutant keeps producing second-leaf-like organs over the course of its lifetime, whereas the pps mutant has an extended juvenile period but shows early flowering (Asai et al., 2002; Tanaka et al., 2011). Both of these mutants show a prolonged juvenile period, but rice mutants with truncated juvenile phases have not been reported. In the present study, we describe the isolation and cloning of the precocious (pre) rice mutant, which exhibits long leaves, a shortened juvenile phase and early flowering. We show that the mutated gene (hereafter termed ‘PRE’) encodes an allene oxide synthase (OsAOS1), a key enzyme in jasmonate biosynthesis. This finding suggests that jasmonate plays a substantial role in the juvenile-to-adult transition in rice.

While screening leaf developmental mutants, we isolated a mutant exhibiting long leaves (Fig. 1A) and designated it precocious (pre), because the pre plant seemed to reach the adult stage precociously. Because most of the pre homozygous mutants withered before flowering under flooding conditions in a paddy field, pre mutant and wild-type plants were grown in MS medium or in pots under non-flooding conditions (see Materials and Methods). The leaf blades of pre were longer than those of the wild type (cv. Nipponbare) from the second to the twelfth leaves (Fig. 1B), resulting in taller plants (Fig. 1A). The leaf blade width of pre from the second to the ninth leaves was slightly greater than that of the wild type (Fig. 1C). To elucidate the cause of the elongated leaf blades in the pre mutant, we measured the length and width of epidermal cells in the second leaf blades. In the pre mutant, epidermal cell width was comparable to that in the wild type, whereas the length was approximately 20% greater (Fig. S1). We then estimated the number of epidermal cells in the second leaf blades by dividing the blade length and width by epidermal cell length and width, respectively, and revealed that the epidermal cell number in pre was greater by approximately 70% in the direction of length and 25% in the direction of width (Fig. 1D). These results indicate that the greater leaf size in the pre mutant is mainly caused by the increase in cell number, although cell elongation is also involved.

The ‘long leaf’ phenotype was more conspicuous in the second to the fourth leaves than in the upper leaves (Fig. 1E). This suggests that juvenile and juvenile-to-adult transition leaves (second-fourth leaves) may have precociously acquired adult characteristics in pre plants, resulting in a shortened juvenile phase. We next examined juvenility-related characteristics in the pre mutant. The midrib is a clear morphological marker distinguishing juvenile and adult leaves; it is almost absent in juvenile leaves (second leaf) but present in approximately 50% and over 70% of the blade length in the third leaf and adult leaves, respectively, in wild type (Asai et al., 2002; Tanaka et al., 2011; Fig. 1F; Fig. S2). In pre plants, midribs in the second and third leaf blades were formed in approximately 22% and 60% of the total leaf blade length, respectively (Fig. 1F; Fig. S2). These values were greater than those of the wild type (∼7% and 50%, respectively). This result supports the notion that pre second and third leaves precociously express adult characteristics.

Other morphological and physiological traits that differ between the juvenile and adult phases in rice (Itoh et al., 2005) also supported this idea. The wild-type SAM gradually becomes enlarged during plant development. We compared the SAM width at the second to fifth leaf stages between the pre mutant and the wild type. The width of the SAM in pre plants was almost equal to that in the wild type during the second leaf stage; however, the width was consistently larger than that in the wild type after the third leaf stage (Fig. 1G). The physiological traits associated with the phase transition showed the same tendency. Generally, in wild-type plants, the photosynthetic rate is low in juvenile leaves (the second leaves) and then increases gradually in juvenile-to-adult transition leaves (the third to fifth leaves) (Asai et al., 2002; Tanaka et al., 2011). In the wild type, the photosynthetic rate reaches its peak during the fourth leaf stage (Fig. 1H). By contrast, the pre mutant reached its maximum photosynthetic rate as early as the third leaf stage (Fig. 1H). We measured the plastochron (the period between the initiation of two successive leaves), which was shorter in the juvenile phase than in the adult phase. After the third week post-sowing, pre plants consistently produced fewer leaves than did the wild type by approximately one leaf throughout the vegetative phase (Fig. 1I). This indicated that pre mutants had the same plastochron as the wild type during the adult vegetative phase and that the reduction in leaf production rate took place transiently after the first week post-sowing. All of the data described above support the idea that the mutation in PRE accelerates the appearance of adult traits in leaf length, midrib formation, SAM size, photosynthesis rate, and plastochron.

We next examined the reproductive phenotypes of the pre mutant. The pre mutant flowered approximately 5 days earlier than did the wild type (Fig. 1J). At the heading, the total number of leaves in wild type was 15 on average, and that in the pre mutant was 14.3 (n≥12). Although the juvenile-to-adult phase transition occurred early in the pre mutant, the duration of the adult phase should be unaffected. In wild-type flowers (florets), swelling of lodicules causes flower opening, and their withering causes flower closing soon after flowering (Fig. 1K). By contrast, the flowers of pre plants remained open after flowering (Fig. 1L) owing to their non-withering lodicules (Fig. 1M). Furthermore, pre flowers rarely set seeds. Thus, PRE is also associated with the function of lodicules and seed formation.

Recently, miR156 and miR172 have been reported as molecular markers for the juvenile-to-adult phase transition in several plant species: the expression level of miR156 is high during the juvenile phase and downregulated during the adult phase, whereas miR172 expression follows the opposite pattern (Aukerman and Sakai, 2003; Chuck et al., 2007; Lauter et al., 2005; Tanaka et al., 2011; Wu and Poethig, 2006; Yoshikawa et al., 2013). Similarly, in rice, the expression level of miR156 was highest in the second leaf and then declined, whereas miR172 expression was repressed in the second leaf but increased with development (as previously reported; see Tanaka et al., 2011; Fig. 2). The expression of miR156 in pre plants was, however, consistently low from the second leaf onwards, and miR172 expression was high as early as the third leaf (Fig. 2). Precocious downregulation of miR156 and upregulation of miR172 in pre plants during the early stages, together with the results of morphological and physiological analyses, signifies an accelerated juvenile-to-adult phase transition.

We employed a map-based cloning approach to isolate the PRE gene from F2 plants derived from the cross between pre heterozygotes and Kasalath (indica variety). The PRE locus was mapped to the long arm of chromosome 3 around 143 cM (Fig. 3A). We compared the nucleotide sequences of 25 candidate genes in this region between pre and the wild type and found only one difference, a 33-nucleotide deletion in pre (Fig. 3B). This deletion influenced two transcripts, Os03t0767000-01 and Os03t0766900-02 (Fig. 3B). Os03t0767000-01 protein derived from the Os03g0767000 locus was truncated by 11 amino acids, whereas 33 bp were deleted from the 3′-untranslated region of Os03t0766900-02. Semi-quantitative reverse transcription (RT)-PCR analysis revealed that the expression levels of these genes in the pre mutant showed no striking change (Fig. 3C), suggesting that this deletion might have little impact on the stability of these transcripts. As the pre mutation seemed to impair Os03g0767000 protein activity, we introduced a 6.0-kb genomic DNA fragment containing the entire Os03g0767000 open reading frame (ORF) without the Os03g0766900 ORF into the pre mutant. The transformed plants complemented pre phenotypes in leaf blades and flowers (Fig. 3D,E). Thus, Os03g0767000 is the gene responsible for the pre phenotype.

Os03g0767000 consists of one exon comprising 1539 nucleotides and encodes a protein of 513 amino acid residues with three conserved domains: an I-helix, a heme-binding site and a chloroplast transit peptide (Haga and Iino, 2004). These domains are the consensus sequences for ALLENE OXIDE SYNTHASE (AOS), a key enzyme for the biosynthesis of jasmonic acid (JA), which catalyzes the conversion of 13-HPOT [13S-hydroperoxy-(9Z,11E,15)-octadecatrienoic acid] to 12,13-EOT [(9Z,13S,15Z)-12,13-oxido-9,11,15-octadecatrienoic acid]. The amino acid sequences encoded by the deleted 33-bp region in the pre mutant were widely conserved among several plant species (Fig. S3A). Four AOS genes were predicted in the rice genome, and PRE encodes OsAOS1 (Haga and Iino, 2004; Fig. S3B). OsAOS1 is expressed predominantly in seedlings, whereas the other three paralogs are expressed mainly in roots at a low level compared with OsAOS1 (based on data retrieved from RiceXPro: http://ricexpro.dna.affrc.go.jp) (Haga and Iino, 2004; Sato et al., 2011). Furthermore, endogenous JA content was significantly lower in pre plants (Fig. 3F). Therefore, PRE/OsAOS1 plays essential roles in JA biosynthesis and the juvenile-to-adult phase transition.

We performed qRT-PCR analysis to elucidate the patterns of PRE expression. Expression of PRE was high in ovaries at 2 days after pollination (DAP) and in mature leaf sheaths and blades (Fig. S4). To characterize the spatial expression patterns of PRE in detail, we carried out in situ hybridization. In the shoot apex, signals were not detected in SAM or P1 leaf primordium, but strong signals were detected in stem and P3 to P5 leaf sheaths (Fig. 4A) and in the boundary region between the leaf sheaths and blades of P2 and P3 leaf primordia (Fig. 4B). Strong expression was also detected in the bundle sheath extension of P3 and P4 leaf sheaths and vasculature (Fig. 4C,D) and in the ground meristem adjacent to the peripheral cylinder of vascular bundles from which the crown root initiated (Fig. 4E). In young inflorescences, PRE mRNA was observed in the meristem and vasculature (Fig. 4F). All floral organs, including glumes, seemed to express PRE, but signals were particularly strong in the stamens and vasculature (Fig. 4G). In 2 DAP ovaries, signals were detected in the embryo, endosperm and vasculature of the ovary wall (Fig. 4H). When in situ hybridization was performed using the PRE sense probe, signals were detected in a patchy pattern in leaf primordia (Fig. S5). This expression pattern would reflect that of Os03t0766900-02, because the PRE ORF overlapped with the 3′-UTR of the Os03t0766900-02 transcript in the reverse direction (Fig. 3B).

We examined whether the application of exogenous methyl jasmonate (MeJA) could counteract the pre defects. In the second leaf of the wild type, growth of both leaf blades and sheaths was suppressed in a dose-responsive manner (Fig. 5A). A marked dose-responsive reduction occurred in pre leaves, and the second leaf blade length of the pre mutant became nearly equal to that of the wild type when more than 10⁻⁸ M MeJA was applied (Fig. 5A,B). In addition, wild-type seedlings grown on medium containing MeJA exhibited short leaf blades and a decline in the rate of midrib formation in third leaf blade (Fig. S6). This result is opposite to that in the pre mutant grown on JA-free conditions, suggesting that exogenous JA extends the period of the juvenile phase.

Next, we examined the GA response in pre mutants. GA is reported to promote the juvenile-to-adult phase transition; the GA-deficient mutant ga1-3 failed to transit to the adult phase in Arabidopsis, and the dwarf mutant in maize showed a prolonged juvenile phase owing to low GA content (Evans and Poethig, 1995; Telfer et al., 1997). It has also been reported that the GA signaling repressor DELLA binds directly to the JA signaling repressor JAZ, signifying an interaction between GA and JA (Hou et al., 2010). To evaluate GA sensitivity, we examined the effect of GA application on the growth of second leaf blades. Leaf blade growth was more enhanced in pre plants than in the wild type when more than 10⁻⁷ M GA was applied (Fig. 5C). Thus, the precocious juvenile-to-adult phase transition in pre mutants might be caused by a reduction in JA content and enhanced sensitivity to GA.

The pre mutant transiently exhibited a prolonged plastochron during the juvenile-to-adult transition stage (Fig. 1I). This phenotype contrasts with those of pla1 and pla2 mutants, which have short plastochrons (Itoh et al., 1998; Miyoshi et al., 2004; Kawakatsu et al., 2006). PLA1 and PLA2 encode a cytochrome P450 protein, CYP78A11, and an RNA-binding protein, respectively (Miyoshi et al., 2004; Kawakatsu et al., 2006). Hence, it is possible that overexpression of PLA1 and/or PLA2 may partly recreate the pre phenotype. qRT-PCR analysis revealed that PLA1 expression was upregulated in pre plants, whereas that of PLA2 was not affected (Fig. 6A). PLA1 expression in pre plants did not alter its specific localization (Fig. S7). Next, we generated the PLA1-High Copy (PLA1-HC) line by introducing a 6.0-kb PLA1 genomic fragment into the wild type, in which PLA1 expression was increased more than fivefold (Fig. S8). PLA1-HC plants exhibited lengthened plastochrons and long leaves (Fig. 6B-D). These results suggest that the pre phenotype is partly mediated by enhanced PLA1 expression.

To investigate the relationship between PRE and PLA1 further, we generated the pre pla1 double mutant (Fig. 6E) and measured plastochron and leaf blade length each week (Fig. 6F,G). As shown in Fig. 1I, leaf number in 2- to 4-week-old plants was lower in the pre mutant than in the wild type. The pre pla1 double mutant showed short plastochrons, comparable to those of pla1 plants (Fig. 6F). Leaf blades of pre pla1 double mutants, however, were significantly longer than those of pla1 mutants at all leaf positions examined, although much shorter than those of pre plants (Fig. 6G). These results suggest that PRE might be involved in plastochron regulation via PLA1, but that PRE and PLA1 act independently in the regulation of leaf blade length.

We also generated the pre pla2 double mutant (Fig. 6H). Unlike pre pla1 plants, leaf number in 2- to 5-week-old pre pla2 plants was significantly lower than that of pla2 plants, although it was significantly higher than that of wild type (Fig. 6I). The leaf blade of pre pla2 plants was also significantly longer than that of pla2 plants (Fig. 6J). Thus, PRE and PLA2 may function independently in both plastochron and leaf blade length regulation.

In this study, we revealed that the pre mutant exhibited long leaves and a precocious transition from juvenile to adult owing to a loss-of-function mutation in the allene oxide synthase gene OsAOS1. To date, JA and its various metabolites regulate plant responses to abiotic and biotic stresses, as well as plant growth and developmental processes, including senescence, tendril coiling, leaf abscission, tuber formation and inflorescence and flower development (reviewed by Santino et al., 2013; Yuan and Zhang, 2015). The pre mutant exhibited not only defects related to the juvenile-to-adult phase change but also withering in flooding conditions and defects in flower closing and fertilization. Withering in flooding conditions might reflect diminished tolerance against abiotic stress. Sterile phenotype is consistent with other reports in which most of the mutants defective in JA biosynthesis show male sterility (Ishiguro et al., 2001; von Malek et al., 2002; Park et al., 2002; Caldelari et al., 2011; Yuan and Zhang, 2015).

JA controls leaf growth by negatively affecting cell cycle progression and endoreduplication in a COI1-dependent manner in Arabidopsis (Noir et al., 2013). In rice, it has been reported that a cpm2 (coleoptile photomorphogenesis2)/hebiba mutant defective in allene oxide cyclase (AOC) and an eg1 (extra glume1) mutant defective in a lipase involved in JA biosynthesis have longer leaves (Riemann et al., 2003, 2013; Cai et al., 2014). Similarly, the longer leaf phenotype of the pre mutant was caused by increased cellular proliferation rather than cellular expansion (Fig. 1D). Thus, JA-related mutants exhibit long leaves in both dicots and monocots, although this phenotype has not been discussed in relation to the juvenile-to-adult phase transition.

The pre mutant showed abnormal morphological and physiological phenotypes other than long leaves, all of which indicated that juvenile (second leaf) and juvenile-to-adult transition (third to fifth leaves) leaves appear more adult than those of the wild type. The expression of two microRNAs (miR156 and miR172), known as molecular markers for the juvenile-to-adult phase transition (Poethig, 2013), supports the hypothesis that pre mutants underwent a precocious phase transition. Other JA-related mutants of rice (cpm1, cpm2 and eg1) also have long leaves (Haga and Iino, 2004; Riemann et al., 2003; Cai et al., 2014), but phase transition-related traits have not been examined. Based on phenotypic similarities, these mutations may also affect the juvenile-to-adult phase transition.

Based on the results of clonal analysis and shoot apex culture of maize, Orkwiszewski and Poethig (2000) argued that vegetative phase identity is determined after leaf initiation, and factors driving phase transition do not originate in the SAM. PPS is also expressed strongly in the fourth and fifth leaves, but not in the SAM (Tanaka et al., 2011). The present results, which revealed strong expression of PRE (OsAOS1) in the leaf sheath, stem and young leaf primordia, but not in the SAM (Fig. 4; Fig. S4), support the hypothesis that the initial step of the vegetative phase transition occurs in leaves.

Several mutants with truncated juvenile phases have been reported in some plant species. The zip (zippy) mutant in Arabidopsis produced first and second leaves with adult characteristics, i.e. serrate and narrow leaves with abaxial trichomes (Hunter et al., 2003). Also, the zip mutant showed an accumulation of SPL3 transcripts, which is one of the targets of miR156 (Wu and Poethig, 2006). The maize glossy15 mutant showed an accelerated phase transition only in leaf epidermal cells (epicuticular wax composition, cell wall characteristics and the existence of differentiated epidermal macrohairs and bulliform cells), but other traits involved in phase transition appeared normal (Moose and Sisco, 1994). Thus, among heterochronic mutants, some may affect only specific organs/tissues. The pre mutant displayed abnormalities in all traits examined, indicating an upstream role for PRE and/or JA in the juvenile-to-adult phase transition cascade.

The heading date of pre plants was approximately 5 days earlier than that of the wild type (Fig. 1J). Considering that the pre juvenile period was truncated by one plastochron (∼5 days), the duration of the adult period is considered to be unaffected. A similar phenomenon is observed in Arabidopsis hasty, in which both juvenile-to-adult phase transition and floral induction were accelerated (Bollman et al., 2003). Conversely, the rice pps mutant exhibited a retarded vegetative phase transition but early heading, resulting in a truncated adult phase (Tanaka et al., 2011). Because PPS is an Arabidopsis COP1 ortholog that has pleiotropic functions on development, PPS function in reproductive transition should be independent of that in vegetative transition. Considering cases of various heterochronic mutants, the timing of juvenile-to-adult phase transitions is not expected to affect the timing of the reproductive phase transition directly.

The phytohormone GA is a positive factor for vegetative phase transition in several plant species (Evans and Poethig, 1995; Telfer et al., 1997; Tanaka et al., 2011). The present results indicate that JA is a negative factor for the vegetative phase transition. In Arabidopsis, GA and JA are antagonistic in developmental events such as hypocotyl elongation, root growth, flowering and defense responses via an interaction between DELLA and JAZ proteins (Navarro et al., 2008; Hou et al., 2010; Wild et al., 2012). Furthermore, in rice, GA application caused leaf expansion whereas JA application reduced leaf size (Fig. 5A), suggesting that the antagonistic relationship between JA and GA (and also between DELLA and JAZ) is retained in rice.

pre leaf blades showed hypersensitivity to GA when more than 10⁻⁷ M GA₃ was used (Fig. 5C), raising the possibility that the change in susceptibility to GA in the pre mutant influenced the timing of the juvenile-to-adult phase transition. It is interesting to consider whether enhanced tissue-specific GA responsiveness associated with reduced JA content, or JA signaling directly, is responsible for the precocious transition to adult phase.

Expression analyses by qRT-PCR and in situ hybridization indicated that PRE quantitatively, but not spatially, regulated PLA1 expression, resulting in the prolonged plastochron in pre (Fig. 6F). However, the regulation of leaf blade length by PRE was independent of that by PLA1 (Fig. 6G). Based on these results, the phenotype of pre pla1 double mutants may be interpreted as follows: both pla1 and pre pla1 plants exhibited similar plastochrons owing to the loss of function of PLA1, but leaf blade length was greater for pre pla1 than for pla1 plants, because PRE and/or JA contents regulate plastochron via PLA1 function during early stage and leaf blade length independently of PLA1.

The relationship between PRE and PLA2 differs from that between PRE and PLA1. Both plastochrons and lead blades of pre pla2 plants were longer than those of pla2 plants, suggesting that PRE is independent of the PLA2 pathway. In accordance with this result, PLA1 was upregulated in pre mutants, but PLA2 was not (Fig. 6A). Kawakatsu et al. (2006) reported that PLA1 and PLA2 act in independent pathways regulating leaf maturation, plastochrons and the duration of the vegetative phase. In addition, GA promoted PLA1 and PLA2 expression, and the GA inhibitor uniconazole suppressed their expression (Mimura et al., 2012).

As described above, the relationships among PRE, PLA1, PLA2, JA and GA are highly complex. PRE-mediated JA content and GA play substantial roles in the vegetative phase transition. Although the roles of PLA1 and PLA2 in phase transition have not been described explicitly, the regulation of PLA1 expression by PRE suggests that PLA1 is a phase transition regulator.

In the present study, we examined the pre mutant in detail and revealed that it underwent a precocious phase transition. Identification of the causal gene revealed a substantial role for JA as a negative regulator of the vegetative phase transition in rice. Recently, Beydler et al. have reported that genes related to stress response pathways, such as jasmonic acid, are upregulated in juvenile leaf primordia and exogenous JA application delays the emergence of adult traits in maize (Beydler et al., 2016). Additional studies will be required to determine how JA has an impact on vegetative phase transition and to examine the relationship between the stress response and developmental phase change.

The pre mutant was obtained from the population of Tos17 insertion lines of cv. Nipponbare (Hirochika, 2001; Hirochika et al., 1996). pla1-4 and pla2-1 mutants were reported in previous works (Miyoshi et al., 2004; Kawakatsu et al., 2006). These mutations were maintained in a heterozygous state. For double mutant analysis, wild type, each single mutant, pre pla1-4 and pre pla2-1 double mutants generated from double heterozygous plants were identified by PCR-based genotyping. In a paddy field, most of the pre homozygous plants died before heading. For that reason, pre mutant and wild-type plants were grown in MS medium (Murashige and Skoog, 1962) containing 3% sucrose and 1% agar at 28°C under continuous light conditions or in pots under non-flooding conditions. Transgenic plants were grown in a biohazard greenhouse at 30°C in the daytime and 25°C at night.

For paraffin-embedded sections, samples were fixed in formaldehyde-acetic acid solution (formaldehyde:glacial acetic acid:ethanol=2:1:17) for 24 h at 4°C. They were dehydrated in a graded ethanol series, and embedded in Paraplast plus (McCormick Scientific). Microtome sections (8 μm thick) were stained with Delafield's Hematoxylin and observed with a light microscope.

For measuring photosynthetic rate, fully expanded second to sixth leaf blades of the wild type and pre mutant were used. Rates of photosynthetic CO₂ assimilation were measured using a portable gas exchange system (LI-6400; Li-Cor). Measurements were replicated for three plants. Light was provided by an LED source (red/blue, 6400-02 LED source; Li-Cor). During the measurement of photosynthetic CO₂ assimilation rates, the photon flux density was 1300 μmol/m²s, leaf temperature was 25°C, and the reference CO₂ concentration was 370 ml/l.

Total RNA was extracted using TRIzol reagent (Invitrogen) or RNeasy (Qiagen) according to the manufacturer's instructions. The reverse transcriptional reaction was performed using High Capacity cDNA Reverse Transcription Kits (Applied Biosystems) according to the manufacturer's instructions. To quantify gene expression, qRT-PCR analysis of PLA1, PLA2 and OsACTIN1 (ACT1) was performed using the TaqMan Fast Universal PCR Master Mix (Applied Biosystems) and FAM-labeled TaqMan probes for each gene, and qRT-PCR analysis of miR156 and miR172 was performed using TaqMan MicroRNA Assays (Applied Biosystems), on a StepOnePlus real-time PCR system (Applied Biosystems). The primers for ACT1 were 5′-GCTATGTACGTCGCCATCCA-3′ and 5′-GCTGACACCATCACCAGAGT-3′, and the TaqMan probe was 5′-CAATACCTGTGGTACGACCACTGGCATACAG-3′ including FAM dye at the 5′-end and BQH1 at the 3′-end. The primers and TaqMan probes for PLA1 and PLA2 were prepared as described by Mimura et al. (2012). For RT-PCR, RNA extraction was performed as described above. Reverse-transcribed cDNA was synthesized from 1 μg of total RNA using Superscript III reverse transcriptase (Invitrogen). The primers for RT-PCR are listed in Table S1.

To map the PRE locus, pre heterozygous plants (O. sativa L. ssp. japonica) were crossed with cv. Kasalath (ssp. indica), and siblings with longer leaf blades were used for calculating recombination between the mutation site and PCR-based polymorphic markers. The PRE locus was mapped onto the long arm of chromosome 3 using STS markers obtained from the rice genome database (http://rgp.dna.affrc.go.jp/E/publicdata/caps/index.html). Further fine-mapping limited the PRE locus around 143 cM between two markers (K0437 and K0011). Primers for fine-mapping of pre are listed in Table S1. To define the molecular lesion in the pre mutation, the candidate genes were amplified via PCR from both pre and wild-type genomic DNAs, and these PCR products were sequenced.

For the complementation test, a 6.0-kb PRE genomic fragment, including 2.0 kb upstream and 2.0 kb downstream, was ligated to an Agrobacterium binary vector, pPZP2H-lac. The PRE gene construct was introduced into Agrobacterium tumefaciens strain EHA105 and transformed into pre mutant calli by the Agrobacterium-mediated transformation method (Ozawa, 2012). As a control, the pPZP2H-lac empty vector was also transformed into pre mutant calli. For construction of the overexpressing PLA1 line, a pBGH/P450 vector including a PLA1 genomic fragment (Miyoshi et al., 2004) was introduced into wild type (cv. Taichung 65). As a control, the pPZP2H-lac empty vector was also transformed into wild type.

Samples were fixed in 4% (w/v) paraformaldehyde and 1% Triton X-100 in 0.1 M sodium phosphate buffer for 48 h at 4°C, dehydrated in a graded ethanol series, and embedded in Paraplast plus (McCormick Scientific). Paraffin sections (8 μm thick) were applied to microscope slides coated with MAS (Matsunami Glass). The PRE antisense probe was prepared from the coding regions without the poly(A) region. Digoxigenin (DIG)-labeled antisense and sense RNA probes were generated by in vitro transcription with T7 RNA polymerase and DIG-RNA labeling mix (Roche). In situ hybridization and immunological detection of the hybridization signals were carried out as described by Kouchi and Hata (1993).

Wild-type and mutant seeds were sterilized with 3% NaClO for 25 min, and washed four times in sterile distilled water. The seeds were then placed on MS medium containing various concentrations of MeJA (methyl jasmonate; WAKO) or GA₃ (gibberellin A₃; WAKO). Plants were grown in a growth chamber under continuous light at 28°C. After 11 days, the second or third leaf blade and sheath length were measured for more than five plants in each treatment.

Sampling of approximately 100 mg of fresh whole shoot samples from 7-day-old seedlings (same as the seedlings in Fig. 1A) was repeated three times for both pre mutants and the wild type. Extraction and determination of JA in each sample were performed using a liquid chromatography-tandem mass spectrometry system (AQUITY UPLC System/XEVO-TQS; Waters; http://www.waters.com; Kojima et al., 2009; Kojima and Sakakibara, 2012).

We thank H. Teshima, R. Soga and K.-i. Ichikawa (University of Tokyo, Japan) for their assistance in cultivating rice plants at the Experimental Farm of the University of Tokyo. We also thank Dr T. Kawakatsu and Dr Y. Sato (University of Tokyo, Japan) for the initial studies of the pre mutant.