PLASTOCHRON3/GOLIATH encodes a glutamate carboxypeptidase required for proper development in rice

Authors


*(fax +81 3 5841 5063; e-mail anagato@mail.ecc.u-tokyo.ac.jp).

Summary

Most aerial parts of the plant body are products of the continuous activity of the shoot apical meristem (SAM). Leaves are the major component of the aerial plant body, and their temporal and spatial distribution mainly determines shoot architecture. Here we report the identification of the rice gene PLASTOCHRON3 (PLA3)/GOLIATH (GO) that regulates various developmental processes including the rate of leaf initiation (the plastochron). PLA3/GO encodes a glutamate carboxypeptidase, which is thought to catabolize small acidic peptides and produce small signaling molecules. pla3 exhibits similar phenotypes to pla1 and pla2– a shortened plastochron, precocious leaf maturation and rachis branch-to-shoot conversion in the reproductive phase. However, in contrast to pla1 and pla2, pla3 showed pleiotropic phenotypes including enlarged embryo, seed vivipary, defects in SAM maintenance and aberrant leaf morphology. Consistent with these pleiotropic phenotypes, PLA3 is expressed in the whole plant body, and is involved in plant hormone homeostasis. Double mutant analysis revealed that PLA1, PLA2 and PLA3 are regulated independently but function redundantly. Our results suggest that PLA3 modulates various signaling pathways associated with a number of developmental processes.

Introduction

Shoot architecture in higher plants is specified by the distribution of leaves and branches along the shoot axis. Because branches are formed at the axils of leaves, the control of leaf production is of primary importance for the establishment of shoot architecture. In higher plants, leaves are initiated from the flank of the shoot apical meristem (SAM). Thus, genetic dissection of the pattern of leaf initiation is necessary for understanding how shoot architecture is established.

Plant hormones are thought to be important for the control of leaf initiation. Analysis of the auxin transport-deficient mutant pin-formed1 (pin1) revealed that a new leaf primordium initiates at a position where the local auxin concentration becomes maximum, and a localized application of auxin to the SAM causes ectopic leaf initiation at that position (Reinhardt et al., 2000, 2003). Thus, auxin is regarded as an inducer of leaf initiation. The newly initiated leaf primordium acts as a sink for auxin and interferes with initiation of new leaf primordia in its vicinity. Cytokinin is also involved in the regulation of leaf initiation via SAM homeostasis. Maize abphyl1 (abph1) mutation drastically changes the spatial distribution of leaves (phyllotaxy) from 1/2 alternate to decussate phyllotaxy accompanying enlargement of the SAM (Jackson and Hake, 1999). ABPH1 encodes an A-type response regulator, which is thought to act as a negative regulator of cytokinin signaling (Giulini et al., 2004).

Although phyllotaxy has been extensively studied, less attention has been paid to the effect of the leaf initiation rate (the plastochron). Several mutants with shortened or prolonged plastochrons have been reported (Itoh et al., 2000; Ikeda et al., 2005, 2007)(Chaudhury et al., 1993; Reed et al., 1993; Prigge and Wagner, 2001; Kwon et al., 2005; Chuck et al., 2007). In these studies, the rate of cell division is thought to be crucial for regulation of the plastochron. In pla1, pla2 and amp1, cell divisions in the SAM occur more frequently than in wild type. Furthermore, constitutive expression of cyclin D in tobacco (Nicotinum tabacum) causes accelerated cell divisions and a shortened plastochron (Cockcroft et al., 2000). On the contrary, a cytokinin receptor triple mutant, cre1/ahk4 ahk2 ahk3, showed a prolonged plastochron due to a reduced frequency of cell division in the SAM (Higuchi et al., 2004; Nishimura et al., 2004).

In Arabidopsis and maize (Zea mays), one of the microRNAs, miR156, is involved in plastochron regulation, and its overexpression causes a shortened plastochron (Schwab et al., 2005; Wu and Poethig, 2006; Chuck et al., 2007). miR156 targets members of the SQUAMOSA PROMOTER BINDING PROTEIN LIKE (SPL) family. Furthermore, miR156 is also involved in phase change. Overexpression of miR156 prolongs the juvenile phase and delays flowering in Arabidopsis, and transforms tassels into vegetative shoots in maize (Schwab et al., 2005; Wu and Poethig, 2006; Chuck et al., 2007). However, most studies have been focused on traits other than the plastochron.

In rice (Oryza sativa), pla1 and pla2 mutants show a short plastochron and small precocious leaves. In addition, they exhibit heterochrony, that is, ectopic expression of a vegetative program in the reproductive phase. Since PLA1 and PLA2 genes are expressed in young leaf primordia but not in the SAM, the primary function of PLA1 and PLA2 is thought to be suppression of precocious leaf maturation (Miyoshi et al., 2004; Kawakatsu et al., 2006). Based on these analyses, we propose a model in which the plastochron is determined by signals from immature leaf primordia that act non-cell-autonomously in the SAM to inhibit the initiation of new leaves. However, independent transcriptional regulation of PLA1 and PLA2 obscures the cues for identifying the signal that regulates the plastochron.

Here, we report a novel rice mutant, plastochron 3 (pla3) that shows a short plastochron. Our analysis revealed that an independently identified mutant, goliath (go), is allelic to pla3. Phenotypes of pla3/go are partially similar to those of pla1 and pla2 mutants with a shortened plastochron due to precocious leaf maturation and the conversion of inflorescence branches to vegetative shoots. In addition, pla3 showed other phenotypes that are not observed in pla1 or pla2. PLA3/GO encodes a glutamate carboxypeptidase and its molecular function is deduced to catabolize various small acidic peptides and release small signaling molecules. Double mutant analyses revealed that all PLA genes are involved redundantly in meristem maintenance.

Results

Identification of new shortened plastochron mutants

Two recessive mutations, pla3-1 and pla3-2, caused pleiotropic phenotypes including a shortened plastochron (Figures 1a and S1). Independent screening of rice seeds with enlarged embryos identified two go mutants (go-1 and go-2). Since independent cloning of pla3 and go revealed that the both mutations are found in the same gene, go-1 and go-2 were redesignated pla3-3 and pla3-4, respectively. About half of the pla3-1 seedlings died within 2 weeks of germination (see below), and the others survived and formed small leaves in a shortened plastochron. Most of the pla3-2 seedlings died within 2 weeks of germination, and the surviving pla3-2 plants showed almost the same phenotype as pla3-1 plants. The plastochron of pla3-1was 1.6 days, shorter than those of wild type, pla1-4 and pla2-1 (4.1, 2.1 and 1.9 days, respectively). Seeds from pla3-3 and pla3-4 could not germinate.

Figure 1.

 Seedlings and embryos of wild type and pla mutants.
(a) Seedlings of wild type, pla1-4, pla2-1 and pla3-1 at 3 weeks after germination.
(b, d, f, i) Wild type.
(c, e, g, h, j, k) pla3.
(b, d) Mature seeds. The embryo of pla3 is markedly larger than that of wild type.
(d–g) Longitudinal sections of mature embryos.
(f, g) Close-up views of the area within the red square in (d and e), respectively. In the wild type, three foliage leaves are formed, in contrast, four foliage leaves are present in pla3. Parts are numbered in order of initiation.
(i, j) Top views of a dormant embryo of wild type (i) and a viviparous one of pla3 (j).
(k) Viviparous seed of pla3. Bar = 5 cm in (a), 2.5 mm in (b and c), 200 μm in (d–g), 500 μm in (h), 1 mm in (i and j) and 2 mm in (k).

Phenotypes of pla3

All the surviving pla3 plants showed a shortened plastochron and conversion of panicle primary branches to vegetative shoots, as did pla1 and pla2 plants (Miyoshi et al., 2004; Kawakatsu et al., 2006). However, additional abnormal phenotypes not seen in pla1 and pla2 were also observed.

Embryo

Mature embryos of pla3/go mutants were significantly larger than those of wild type (Figure 1b–e). All of the embryonic organs were enlarged, including scutellum, plumule, radicle and epiblast (Figure 1d,e). The patterning and positioning of the organs were not disturbed. In pla3-3 (go-1) and pla3-4 (go-2), two radicles were occasionally formed (Figure 1h). In addition, four foliage leaves were differentiated in pla3 embryos in contrast to three leaves in wild-type embryos (Figure 1f,g). The overproduction of leaves could be due to a reduced dormancy, because pla3 seeds frequently showed vivipary (Figure 1i–k).

Vegetative phase

In the vegetative phase, abnormalities were mainly observed in shoots. Although roots were morphologically normal, gravitropism was slightly reduced (Figure S2). Since the short plastochron in pla1 and pla2 was correlated with enhanced cell divisions and enlarged SAMs (Miyoshi et al., 2004; Kawakatsu et al., 2006), we examined cell divisions in pla3 plants surviving more than a month by in situ hybridization using histone H4 specifically expressed in the S-phase of the cell cycle. In median longitudinal sections of pla3 SAM, many more histone H4 signals were detected than in the wild type (5.2 ± 2.0 in pla3-1, n = 9 versus 0.7 ± 0.3 in wild type, = 10) (Figure 2a,b). This value was larger than those in pla1 and pla2 (1.9 ± 0.7 in pla1-4, = 10 and 3.8 ± 1.7 in pla2-1, = 10) (Kawakatsu et al., 2006). Since the cell size of pla3-1 plants was not significantly altered (23.7 ± 6.2 μm in pla3-1 mature leaves versus 20.5 ± 4.7 in wild-type mature leaves), accelerated cell divisions are thought to directly contribute to the shortened plastochron in pla3. Simultaneous measurement of the shape and size of the SAM revealed that the pla3 SAM is larger than the wild-type SAM (Table 1).

Figure 2.

 Vegetative phenotypes of wild type and pla3.
(a, e, l, n) Wild type.
(b–d), (f), (g–k), (m), (o–q) pla3-1.
(a, b) Expression of histone H4 in the shoot apex.
(c) Lethal pla3-1seedling at 2 weeks after germination.
(d) Longitudinal section of the shoot apex in (c). A filamentous leaf primordium (arrow) is present in place of the shoot apical meristem (SAM).
(e, f) Expression of OSH1 in the shoot apex. OSH1 is down-regulated in the P0 region.
(g, h) Elongated SAM (arrowheads).
(i) Closely initiated leaves fused into the X-shaped leaf and sharing one midvein (arrow).
(j) Cleared image of SAM division. Two SAMs are present (arrowheads).
(k) Scanning electron microscopy image of a divided SAM. One SAM grows robustly (arrow), but the other grows weakly (arrowhead).
(l, m) Cross-section of shoot apex. Asterisks indicate differentiating procambial strands. Overlaps of two leaf margins (arrowheads) are observed at the P2 stage in wild type and the P3 stage in pla3.
(n, o) Longitudinal sections of the shoot apex. A ligule primordium protrusion is formed at the P3 leaf primordium (arrows).
(p) Leaf with jagged margin.
(q) Scanning electron microscopy image of jagged leaf margin. The leaf margin is not torn physically. Bars = 50 μm in (a, b, e, f), 500 μm in (c), 100 μm in (d, l–o), 120 μm in (g), 5 mm in (h), 250 μm in (i, q), 200μm in (j), 1 mm in (k) and 2 mm in (p).

Table 1.   Phenotypes of wild type, and pla mutants in vegetative phase (±SD)
TraitWild typepla1-4pla2-1pla3-1
  1. aSAM width was measured at the base of the SAM above the insertion of early stage P1 primordium in median longitudinal section.

  2. bSAM height was measured as the length between the base and top of the SAM in median longitudinal sections.

Plastochron4.1 ± 0.02.1 ± 0.11.9 ± 0.01.6 ± 0.0
SAM widtha (A) (μm)62.9 ± 5.873.5 ± 3.177.0 ± 8.877.3 ± 8.3
SAM heightb (B) (μm)34.1 ± 3.040.6 ± 3.243.1 ± 8.248.3 ± 6.0
SAM shape (B/A)0.54 ± 0.040.55 ± 0.050.56 ± 0.070.63 ± 0.07

It was frequently observed that pla3 seedlings produced several filamentous leaves, and died within 2 weeks of germination (Figure 2c). The last filamentous leaf occupied the position of the SAM, indicating that the SAM was consumed by leaf primordia (Figure 2d). Consistent with this situation, the SAM showed reduced OSH1 expression that represents indeterminate cells in the SAM (Figure 2e,f). Elongation of the SAM was also observed (Figure 2g). In an extreme case, a SAM-like structure protruded from the leaf sheath (Figure 2h).

More than half of the pla3 seedlings (12 out of 19 seedlings) produced twin shoots of equal size. These shoots had the same number of leaves, and the transverse sections revealed that the outermost leaves of the two shoots were fused on their abaxial sides (Figure 2i). Thus, these twin shoots could be derived from the division of the SAM. We obtained a longitudinal section that showed two SAMs aligned side by side (Figure 2j). The two sister SAMs did not always grow synchronously. Sometimes one SAM developed vigorously but the growth of the sister SAM was suppressed (Figure 2k). These results indicate that the activity of the pla3 SAM varies considerably and that PLA3 is involved in maintenance of the SAM.

pla3 had significantly smaller leaves than the wild type (Figure 1a). To confirm that the size reduction of pla3 leaves was caused by precocious maturation as in pla1 and pla2 (Kawakatsu et al., 2006), we examined the developmental course of pla3 leaves using several marker events including down-regulation of OSH1 at P0, overlapping of two leaf margins at P2, vascular bundle formation at P2, ligule formation at P3 and the growth period of blade and sheath (P4–P6). To represent the stages of leaf development we used a plastochron numbering (PN) system; P1 represents the youngest primordium, P2 the next youngest one, and so on (Itoh et al., 2005). P0 is defined as the founder cell stage, which would soon appear as a primordium bulge on the SAM. The developmental events in pla3 leaves occurred almost at the same plastochron number as in the wild type (Figure 2e,f,l–o). Because period of each plastochron in pla3 was truncated to less than half of the wild-type one, pla3 leaves are thought to maturate precociously. Sometimes, leaf margins of pla3 were jagged, as if they were peeled (Figure 2p). The scanning electron microscope (SEM) analysis revealed that leaf blade expansion was not synchronized along proximodistal direction (Figure 2q).

Reproductive phase

In rice, the transition between the vegetative and reproductive phases is accompanied by substantial elongation of the upper four or five internodes. pla3 showed severe dwarfism (<10% of wild-type plant height), but the number of elongated internodes was increased to eight except for the uppermost one (Figure 3a,b). Non-elongation of the uppermost internode and the increase in the number of elongated internodes were commonly observed in pla1, pla2 and pla3 (Figure 3a,b).

Figure 3.

 Reproductive phenotypes of wild type, pla1, pla2 and pla3.
(a) Elongation pattern of internodes. The culm of the wild-type plant is longer than in pla mutants. Arrowheads indicate nodes.
(b) Schematic representation of the elongation patterns of internodes in wild type, pla1-4, pla2-1, pla3-1.
(c) Panicle phenotypes of pla mutants, in which vegetative shoots emerge instead of primary branches. Arrows indicate elongated bracts.
(d, e) Scanning electron microscope images of primary branch primordia in wild type (d) and pla3-1 (e). Bars = 10 cm in (a), 5 cm in (c), 100 μm in (d) and 120 μm in (e).

At the heading stage, pla3 mutants produced ectopic shoots instead of panicle, as did pla1 and pla2 (Figure 3c). The number of ectopic shoots in pla3-1 (4.3; = 10) was larger than in pla1-4 and pla2-1 (3.2, 2.5, respectively; = 10). In pla3, after the formation of the flag leaf, primary branch primordia were formed normally in 2/5 spiral phyllotaxy, but they developed as vegetative shoots (Figure 3d,e). This conversion was always associated with over-growth of the bracts that were normally aborted without elongation (Figure 3c). These reproductive phenotypes commonly observed in pla1, pla2 and pla3 suggest that vegetative and reproductive programs are co-expressed in pla mutants.

Cloning of the PLA3/GO gene

We isolated the PLA3/GO gene by positional cloning using an F2 population between PLA3/pla3-1 and cv. Kasalath (ssp. indica). The PLA3 gene was mapped at around 147 cM near the marker R2628 on chromosome 3. Since one homolog of the Arabidopsis AMP1 gene encoding glutamate carboxypeptidase, Os03g0790600, was closely linked to this region, we examined the Os03g0790600 genomic sequence in wild type and mutants. We detected the following point mutations in Os03g0790600: a G-to-A base change causing a Ser-to-Asn amino acid substitution in the second exon in pla3-1, and a G-to-A base change causing a Gly-to-Glu amino acid substitution in the second exon in pla3-2 (Figure 4a). In an independent cloning effort using go alleles, we detected a G-to-A single base change at the first base of the first intron of the gene causing mis-splicing in go-1 (=pla3-3) and a 29-nucleotide deletion in the first exon, causing a frameshift and introducing a premature stop codon in go-2 (=pla3-4) (Figure 4a). Because the introduction of 9.9-kb genomic fragment containing the candidate gene rescued a pla3-1 homozygous plant (Figure 4b), we concluded that PLA3 is located on the Os03g0790600 locus and encodes glutamate carboxypeptidase.

Figure 4.

 Structure of the PLA3 gene.
(a) Exon/intron structure of the PLA3 gene. Ten boxes indicate exons.The protease associated (PA) domain, the M28 peptidase domain and the transferrin receptor-like dimerization (TFR) domain are indicated by black, dark grey and light grey, respectively. Locations of the four pla3 mutations are indicated.
(b) Complementation of the pla3 mutation. Adult plants of wild type (left), pla3-1 (center) and pla3-1 rescued by introducing a genomic clone of PLA3 (right).
(c) Phylogenetic tree of glutamate carboxypeptidases. Numbers at each branch point indicate bootstrap values (= 1000).
(d, e) Complementation of Arabidopsis amp1 by PLA3/GO: (d) amp1 seedling, (e) amp1 seedling rescued by introducing a rice PLA3/GO cDNA.

We found that the Os03g0790600 sequence deposited in the Rice Annotation Project (RAP) database (http://rapdb.dna.affrc.go.jp/) is truncated at the 5′ end of the first exon. Our full-length PLA3 cDNA sequence revealed that the predicted PLA3 protein consists of 749 amino acids and belongs to glutamate carboxypeptidase II (GPCII), which is similar to the human N-acetyl alpha-linked acidic dipeptidase (NAALADase) (Pangalos et al., 1999). Like GPCIIs, the PLA3 protein contains a protease-associated (PA) domain, a M28 peptidase domain and a transferrin receptor-like dimerization (TFR) domain from N-terminal to C-terminal (Figures 4c and S3). Since five zinc-binding residues and a glutamate, possibly acting as a nucleophile in catalysis, are also conserved in PLA3, PLA3 is expected to have GPCII activity (Figure S3).

Rice, maize and Arabidopsis GPCIIs were classified into two subfamilies. One comprises PLA3, VP8 and AMP1, and the other consists of the several other GPCIIs (Figure 4c). Since the rice PLA3/GO cDNA complemented an amp1 mutant phenotype regarding leaf number (Figure 4d,e) and silique length (wild-type silique = 1.4 ± 0.15 cm; complemented plants silique = 1.45 ± 0.09 cm; and amp1 silique = 0.69 ± 0.05 cm long; mean ± SD, = 40), we concluded that the rice PLA3/GO and Arabidopsis AMP1 are orthologous.

Expression pattern of PLA3

To further understand the function of PLA3 we examined its expression pattern. First, we performed semi-quantitative RT-PCR to reveal the organ-specific expression pattern. Total RNA was isolated from embryos at 10 days after pollination (DAP), from vegetative shoot apices at 3 weeks after germination, and from leaf blades, roots, inflorescence apices and spikelets. PLA3 was expressed uniformly in all organs examined (Figure 5), in contrast to PLA1 and PLA2 that are expressed specifically in young leaf primordia and the inflorescence apex (Miyoshi et al., 2004; Kawakatsu et al., 2006).

Next, we analyzed the detailed expression pattern of PLA3 by in situ hybridization with digoxigenin (DIG)-labeled antisense RNA PLA3 probe. PLA3 transcripts were detected throughout the whole plant body except the endosperm, suggesting that PLA3 was expressed ubiquitously (Figure S4a–d). In a control experiment we could not detect any expression with sense RNA PLA3 probe (Figure S4e). We detected strong PLA3 expression in the epidermal layer of the scutellum of 10-DAP embryos and in spikelet meristems (Figure S4a and d).

Genetic interaction between PLA genes

To examine whether PLA genes are involved in the regulation of other PLA genes, we compared the expression of PLA genes in the shoot apices of the three mutants by in situ hybridization or semi-quantitative RT-PCR. We have already reported that PLA1 expression in pla2 and PLA2 expression in pla1 were not altered (Kawakatsu et al., 2006). Also, PLA1 expression was not altered in the pla3-1mutant (Figure 6a,b). Expression levels of PLA2 in pla3-1 and PLA3 in pla1-4 and pla2-1 were normal (Figure 6c). Accordingly, transcription of PLA genes could be independently regulated.

Figure 5.

 Expression pattern of PLA3.
A RT-PCR analysis of the PLA3 mRNA expression in different tissues. RNA was isolated from an embryo at 10 days after pollination (DAP; EM), a vegetative shoot apex (VS), an inflorescence apex at stage In5 (IA), a leaf blade (LB), a root (R) and a flower (F).

Next, to examine genetic interaction we generated double mutants. Both pla3-1 pla1-4 and pla2-1 pla3-2 double mutants rarely germinated, and viable plants showed similar phenotypes; severely retarded growth, defective SAM maintenance and occasional SAM elongation (Figure 6d–f). The seedling phenotypes of the double mutants were comparable to the severe phenotype of pla3 seedlings. Thus, PLA1, PLA2 and PLA3 are regulated independently but function redundantly.

Plant hormones in pla mutants

Cytokinin (CK) is a classical plant hormone that influences various developmental processes, including the promotion of cell division and leaf senescence (Sakakibara, 2006). The Arabidopsis amp1 mutant was reported to contain a significantly larger amount of CK than wild type (Chaudhury et al., 1993). We measured CK levels in shoot apices including the SAM and P0–P5 leaf primordia at 3 weeks after germination. We measured nucleobases (i.e. tZ, trans-zeatin, and iP, isopentenyladenin), which are active CKs, and their nucleosides (i.e. tZ riboside, tZR, and iP riboside, iPR) and nucleotides (i.e. tZR-5′-monophosphate, tZRMP, and iPR-5′-monophosphate, iPRMP). In pla1, tZ, tZR and tZRMP were increased (Figure 7a). In pla2, tZR and iPR was increased and tZRMP and iPRMP were decreased (Figure 7a). In pla3, tZ, tZR, iP and iPR were increased, but tZRMP, iPRMP were decreased (Figure 7a). Total CKs including active CKs were increased in pla1 and pla3, but not in pla2. However, the extent of the increase in pla mutants was much smaller than in amp1.

Figure 7.

 Hormone contents and responsiveness to ABA of the wild type and pla mutants.
(a) Hormone contents in wild type and pla1-4, pla2-1 and pla3-1. Active cytokinins (CKs) are increased in pla1-4 and pla3-1. The ABA levels of pla mutants were significantly lower than that of wild type.
(b) Effects of ABA application on seed germination. In the wild type, ABA treatment strongly lowered the seed germination rate and retarded seedling growth in a concentration-dependent manner. In the bulk of wild type and the pla3-1 heterozygote the germination rate was lower than pla3, but higher than wild type. In pla3-1 homozygote, germination was not inhibited by ABA treatment, although growth was retarded.

We also measured the ABA content in seedlings of wild type and pla mutants, because pla3 seeds were frequently viviparous. The amounts of ABA in all the three pla mutants were significantly lower than in wild type (Figure 7a). This is consistent with vivipary of pla3. Then, we examined the sensitivity to ABA in pla3-1. The application of ABA to wild-type seeds lowered the germination rate and strongly retarded seedling growth (Figure 6b). In contrast, the germination rate of pla3 homozygotes was not affected, even in 30 μm ABA, and retardation of seedling growth was milder than that in wild type (Figure 7b). Thus, pla3 is less sensitive to ABA treatment. When non-mutant seeds, the bulk of heterozygous (pla3-1/+) and wild-type seeds (+/+) were treated with ABA, the germination rate and suppression of seedling growth were lower than in pla3, but higher than in wild type (Figure 7b), suggesting dose dependency of PLA3.

Figure 6.

 Interrelationships among PLA genes.
(a, b) In situ localization of PLA1 transcripts: (a) wild type. (b) pla3-1. (c) A RT-PCR analysis of PLA2 and PLA3 expression in wild type and pla mutants.
(d) Seedling of double mutants. From left to right: pla3-1, pla1-4 pla3-1 and pla2-1 pla3-2. The pla3-1 seedling on the left shows a relatively severe phenotype among pla3 seedlings.
(e) Longitudinal section of a pla1-4 pla3-1 shoot apex showing an aborting shoot apical meristem (SAM).
(f) Longitudinal section of a pla2-1 pla3-2 shoot apex showing an elongated SAM. Bars = 100 μm in (a and b), 1 cm in (d), 100 μm in (e and f).

The indole-3-acetic acid (IAA) content was reduced in pla2-1 and pla3-1 (Figure 7a) to about two-thirds of wild type. This result is consistent with the observation that pla3-1 roots showed reduced gravitropism (Figure S2).

As shown above, contents of three hormones (CK, ABA and IAA) in pla mutants significantly deviated from those of wild type. Thus, hormonal homeostasis would be disturbed in pla1, pla2 and pla3 mutants.

Discussion

PLA3/GO encodes a glutamate carboxypeptidase II and is an ortholog of Arabidopsis AMP1 and maize VP8

In this study, we identified the PLA3 gene, a new locus that regulates the leaf initiation rate in rice. The PLA3 encodes a protein similar to a glutamate carboxypeptidase II (GPCII), a member of the M28 peptidase family of metalloproteases. In mammals, this protein has been named differently according to its functions – prostate-specific membrane antigen (PSMA), N-acetylated alpha-linked acidic dipeptidase I (NAALADase I) or folate hydrolase FOLH1(Zhou et al., 2005). Prostate-specific membrane antigen is named because of its strong expression in the prostate, although its function is unknown. In the central nervous system, NAALADase I and its homolog NAALDase II hydrolyze the small but very abundant neuropeptide N-acetyl-l-aspartate-l-glutamate (NAAG), releasing C-terminal glutamate (Pangalos et al., 1999). The NAALADases also possess a dipeptidyl peptidase IV activity to hydrolyze Gly-Pro 7-amido-4-methylcoumarin, involved in general metabolic pathways (Pangalos et al., 1999). In the proximal small intestine, FOLH1 removes gamma-linked glutamates from poly-gamma-glutamated folate (Pinto et al., 1996; Tiffany et al., 1999). These facts indicate that members of mammalian GPCIIs are involved in various metabolic and signaling pathways by catabolizing small peptides (Zhou et al., 2005).

Although the biochemical function of plant GPCIIs has not been determined, the N-terminal membrane span and the zinc-binding and catalytic residues conserved between PLA3 and NAALDase indicate that plant GPCIIs retain similar biochemical function to those in mammals. Loss-of-function mutants of PLA3 show pleiotropic phenotypes, suggesting that they have diverse functions in various pathways, possibly via small signaling molecules. Arabidopsis AMP1 and maize Vp8 encode GPCIIs, involved in various developmental functions. Phylogenetic analysis, complementation of amp1 phenotypes by PLA3 cDNA, and high synteny between rice PLA3 and maize Vp8 (data not shown), indicate that PLA3, AMP1 and Vp8 are orthologous, and that their molecular functions would be conserved. Consistently, pla3, amp1 and vp8 show similar phenotypes, such as aberrant cell proliferation, a shortened plastochron, abnormal hormone homeostasis and vivipary. However, there are several distinct phenotypes in each mutant (see below). Thus, although plant GPC II-like proteins have well-conserved molecular and biological functions, they have acquired species-specific functions during evolution.

PLA3 regulates cell proliferation in various tissues

Several phenotypes of pla3 could be interpreted as results of aberrant cell proliferation. Interestingly, aberrant cell proliferation occurred in both directions, enhancement and reduction. Embryonic organs of pla3 were enlarged due to excess cell proliferation because cell sizes in each organ were comparable between wild type and pla3. The pla3 SAM showed two opposite phenotypes – elongation and bifurcation due to enhanced cell divisions in the central zone and rib zone, respectively (Figure 2g,h), and disappearance due to reduced cell divisions (Figure 2d). Abnormal cell proliferation was also seen in leaves, in which synchronous cell proliferation along the central–marginal axis was disturbed and the leaf margin became jagged (Figure 2i,j). These phenotypes suggest that PLA3 acts as a balancer of cell proliferation. Although amp1 and vp8 also show aberrant cell proliferation, abnormalities differ among the three mutants. pla3 showed enlarged embryos in contrast to the reduced embryo in vp8, and enlargement of the SAM was observed in pla3 and amp1 (Suzuki et al., 2008). These phenotypic differences suggest that PLA3, AMP1 and Vp8 have distinct roles in cell proliferation.

PLA genes regulate leaf maturation and SAM maintenance

Similar to pla1 and pla2, pla3 showed a shortened plastochron and small leaves due to precocious maturation, suggesting that PLA3 is also a regulator of leaf initiation and maturation. The similar pattern of internode elongation in pla1 and pla3 suggests a genetic association between PLA3 and PLA1 (Figure 3). Recently, the KLUH gene in Arabidopsis was revealed to be a cytochrome P450, CYP78A5, orthologous to PLA1/CYP78A11 (Anastasiou et al., 2007). Both pla1 and kluh mutants showed partially similar phenotypes such as a shortened plastochron(Miyoshi et al., 2004; Anastasiou et al., 2007). In the amp1 mutant, KLUH expression was up-regulated (Helliwell et al., 2001). Thus, it is possible that PLA3/AMP1 negatively regulates PLA1/KLUH activity or is involved in the production of their substrate. However, expression of PLA1 in the pla3 shoot apex was almost normal (Figure 6), indicating independent transcriptional regulation of the two genes in rice. Because pla3-1 pla1-4 and pla2-1 pla3-2 double mutant plants were lethal at the seedling stage, as in severe pla3 plants, the three genes act redundantly in SAM maintenance/leaf production. In any case, the three PLA genes and their homologs in maize and Arabidopsis are commonly involved in plastochron regulation whether they are expressed in the SAM or not.

From the analysis of pla1, pla2 and kluh it is postulated that they generate mobile signaling molecule that enable cross-talk between leaf primordia and the SAM (Anastasiou, et al., 2007). Although PLA3 is expressed in both leaf primordia and the SAM, PLA3 is also expected to be involved in this signaling pathway, because PLA3 probably functions to produce a signaling molecule.

Heterochronic nature of pla mutants

pla3 showed conversion of rachis branches to vegetative shoots, suggesting that PLA3 acts as a heterochronic gene to suppress the vegetative program in the reproductive phase. We previously suggested that the ectopic expression of the vegetative program is closely associated with overgrowth of bracts, which normally abort as vestiges in the wild type (Miyoshi et al., 2004; Kawakatsu et al., 2006). The overgrowth of bracts was also observed in pla3. Maize dominant heterochronic mutants Teopod1 (Tp1), Tp2 and Tp3 (Dudley and Poethig, 1991, 1993; Bassiri et al., 1992), which are thought to promote vegetative programs in inflorescence meristems, also showed bract outgrowth. Accordingly, outgrowth of bracts is a good marker to evaluate ectopic expression of the vegetative program in the reproductive phase. Although single vp8 mutation does not cause tassel-to-shoot conversion, combination of vp8 and Tp1 drastically enhances all heterochronic traits, indicating that Vp8 and Tp1 act redundantly (Evans and Poethig, 1995). Suzuki et al. (2008) suggest the existence of a dominant suppressor of the vp8 mutation. Tp1 may suppress such a dominant suppressor of the vp8 mutation. Corngrass1 (Cg1), showing similar but more severe phenotypes than Tp mutants, is caused by over-expression of miR156 (Chuck et al., 2007). The phenotypes of Cg1 are partially similar to that of pla3 in that leaf production is accelerated and vegetative traits are expressed in the reproductive phase. Over-expression of miR156 also causes a shortened plastochron in Arabidopsis (Schwab et al., 2005; Wu and Poethig, 2006). This effect is most prominent when miR156 is over-expressed in young leaf primordia, strongly suggesting an interaction between PLA genes and miR156/SPLs (Wang et al., 2008). In addition, The over-expresser of OsmiR156 in rice (Xie et al., 2006) resembles a weak pla1 allele (Itoh et al., 1998). However, the PLA1/KLUH pathway and the miR156/SPLs pathway are thought to be parallel (Wang et al., 2008), and the expression level of OsmiR156 was not significantly altered in pla seedlings (data not shown). Thus, at least in the vegetative phase, PLA genes are unlikely to act upstream of OsmiR156. It is yet possible that PLA genes regulate SPL genes independently of miR156, or they may act downstream of SPL genes.

Loss-of-function mutants of orthologous plastochron genes differ in heterochronic phenotypes. For example, rice pla1 and pla3 show not only delayed vegetative-to-reproductive transition but also conversion of primary rachis branches into vegetative shoots. Their Arabidopsis orthologs kluh and amp1, however, show only slightly early flowering. The maize te1 mutant orthologous to pla2 does not exhibit the conversion of tassel branches to vegetative shoots (Veit et al., 1998; Kawakatsu et al., 2006). In other words, so far, a prolonged vegetative program in reproductive phase is observed only in rice. Thus, rice PLA genes would have acquired a novel heterochronic function concerning the termination of the vegetative program during evolution.

PLA genes are involved in plant hormone homeostasis

In the Arabidopsis amp1 mutant, CK content is markedly increased by seven-fold (Chaudhury et al., 1993), although other hormone contents were not measured. Similarly in the pla3 mutant, CK content was increased but the extent was not as large as in amp1. In pla mutants, hormone contents, including CK, were greatly disturbed (Figure 7b). In pla1 and pla3, the level of CK was increased, and the auxin content was reduced in pla2 and pla3. The ABA content was markedly decreased in all three pla mutants. Since deduced substrates for PLA3 are small acidic peptides and one possible substrate of PLA1 is lauric acid (Miyoshi et al., 2004), the simplest hypotheses that PLA3 and PLA1 are directly involved in biosynthesis or degradation of a specific hormone could be ruled out. Thus, it is natural to think that mutations in three PLA genes would have disturbed hormone homeostasis. Among the hormones examined, ABA content was decreased significantly in all mutants. Moreover, the three genes are associated with the leaf maturation rate. These results prompted us to assume that precocious leaf maturation affects ABA content (biosynthesis or signaling).

Several phenotypes of pla3 are apparently related to hormonal disturbance. pla3 showed vivipary and reduced ABA sensitivity upon germination. Although it is unknown whether the ABA level is decreased in pla3 embryos, the vivipary accompanying excess leaf production would reflect decreased ABA sensitivity. PLA3 was strongly expressed in the epidermal layer of the scutellum, where many maturation-related genes such as OsVP1 and OsEM are expressed (Miyoshi et al., 2002). This suggests that PLA3 is required for ABA signaling in the scutellar epithelium. This speculation is consistent with the decreased VP1expression in vp8 (Suzuki et al., 2008) in maize.

Recently, antagonism between AMP1 and MONOPTEROS (MP) on meristem-niche-associated auxin signaling was reported (Vidaurre et al., 2007). Consistent with this, the auxin level was decreased and root gravitropism was reduced in pla3. Thus. PLA3 may be involved also in auxin signaling. It is well known that the size of the SAM is regulated by CKs. A loss-of-function mutant of rice LOG encoding a CK-activating enzyme has small SAMs (Kurakawa et al., 2007), and the maize abphyl encoding one of the A-type response regulators has an enlarged SAM (Giulini et al., 2004). Thus the enlarged SAM of pla3 is probably associated with increased CK content.

In this paper we have characterized phenotypes of pla3 and identified the PLA3 gene. PLA3 regulates the leaf maturation rate, the plastochron, SAM maintenance and phase change, as do PLA1 and PLA2. Furthermore PLA3 has a more striking role in fundamental processes such as cell proliferation and hormone homeostasis. Recent studies have revealed the importance of small peptides in plant development (Matsubayashi and Sakagami, 1996; Ito et al., 2006; Kondo et al., 2006). Pleiotropic functions of PLA3 and its deduced molecular function strongly suggest that PLA3 is one of the key players generating the small peptide-like molecules essential for proper homeostasis. Further studies will reveal possible substrates for the PLA proteins and greatly expand our understanding of plant development.

Experimental procedures

Plant materials

We identified two single-gene recessive mutants that showed a short plastochron, pla3-1, pla3-2 and one mutant with an enlarged embryo, pla3-3 (go-1), from an M2 population of rice (Oryza sativa L.) cv. Kinmaze mutagenized with N-methyl-N-nitrosourea. We identified another mutant with an enlarged embryo, pla3-4 (go-2), from the Taichung65 population that was created by regenerating plants from tissue culture. We also used pla1-4 and pla2-1 mutants derived from cv. Taichung 65. pla1-4 and pla2-1 are the most severe alleles among four pla1 alleles and two pla2 alleles, respectively. Mutant and wild-type plants were grown in pots or in paddy fields under natural conditions for plastochron measurement and analyses of reproductive phenotypes. Developmental staging was followed as previously described (Ikeda et al., 2004; Itoh et al., 2005). For early vegetative phase analyses, mutant and wild-type seeds were grown on Murashige and Skoog (MS) medium supplemented with 3% sucrose and 1% agar at pH 5.8 in a plant box at 28°C (Murashige and Skoog, 1962). Transgenic plants were grown in a biohazard greenhouse at 30°C in the day and 25°C at night. Double mutants were generated by crossing a PLA3/pla3-1 heterozygote with a PLA1/pla1-4 heterozygote and a PLA3/pla3-2 heterozygote with a PLA2/pla2-1 heterozygote. For analysis of ABA responsiveness, seeds were imbibed in 0, 5, 10, 20, 30 μm ABA.

Histological analysis

Shoot apices were fixed in FAA [formaldehyde:glacial acetic acid:ethanol (1:1:18)] for 24 h at 4°C and then dehydrated in a graded ethanol series. Dehydrated samples in 100% ethanol were replaced with xylene and embedded in Paraplast plus (McCormick Scientific, http://www.mccormickscientific.com/). Microtome sections (8 μm thick) were stained with Delafield’s hematoxylin and observed with a light microscope.

For SEM analysis, dehydrated samples in 100% ethanol were infiltrated with 3-methyl-butyl-acetate, critical point dried, sputter coated with platinum and observed under a SEM (S-4000; Hitachi, http://www.hitachi.com/) at an accelerating voltage of 10 kV.

Map-based cloning and phylogenetic analysis

F2 populations of PLA3/pla3-1 (ssp. japonica) and cv. Kasalath (ssp. indica) were used as mapping populations. Using sequence-tagged sites, and cleaved-amplified polymorphic sequence markers derived from database of the Rice Genome Project (RGP; http://rgp.dna.affrc.go.jp), the PLA3 locus was roughly mapped at 147 cM on chromosome 3. For complementation tests, the 9.9-kb XhoI–SacI fragment, including the PLA3 candidate and 2.0 kb directly upstream of the initiation codon, was cloned into a binary vector and introduced into pla3-1 homozygotes by the Agrobacterium tumefaciens-mediated transformation method (Hiei et al., 1994).

Multiple sequence alignments were performed using clustalx and boxshade (http://www.ch.embnet.org/software/BOX_form.html) programs (Thompson et al., 1997). The phylogenetic tree was constructed based on full-length amino acid sequences by the neighbor-joining method using clustalx and treeview programs (Page, 1996; Thompson et al., 1997). The numbers at the branching points indicate the times that each branch topology was found during bootstrap analysis (= 1000).

Complementation of the amp1 mutation

Arabidopsis amp1-1 mutant seeds (stock no. CS8324) were obtained from the Arabidopsis Biological Resource Center (ABRC). A binary vector comprising a 35S promoter linked to the rice PLA3/GO open reading frame, followed by the phaseolin terminator region, was constructed and introduced into amp1-1 homozygous plants using A. tumefaciens. Complemented plants were confirmed to have both amp1-1 mutation and integrated PLA3/GO gene by sequencing.

Semi-quantitative RT-PCR

Total RNA was extracted from 100 mg of tissue (embryo, roots, shoot apices, leaves, inflorescence apices and flowers) using TRIzol reagent (Invitrogen, http://www.invitrogen.com/). After RNase-free DNase I (Takara, http://www.takara-bio.com/) treatment, 2.5 μg of RNA was reverse transcribed using oligo(dT) primer and SuperScript III (Invitrogen). The primers used for amplification and cycles were: 5′-TCCATCTTGGCATCTCTCAG-3′ and 5′-GTACCCGCATCAGGCATCTG-3′ for ACTIN (25 cycles), 5′-ACAAGGCGTTCCACAAGCAACC-3′ and 5′-GGCGCTGTCATGAGCTCCTG-3′ for PLA2 (30 cycles), 5′-TGGTTCTACCCCCCAGTTGG-3′ and 5′-TTCAACAATCGCCGCCTCATC-3′ for PLA3 (30 cycles).

In situ hybridization

Paraffin sections were prepared as described above except that 8-μm thick microtome sections were applied to slide glasses coated with aminopropylsilane (Matsunami Glass, http://www.matsunami-glass.co.jp/english/company_e/enkaku_e.html). Digoxigenin-labeled antisense and sense probes were prepared from the full-length cDNAs of histone H4, PLA1 and PLA3. In situ hybridization and immunological detection of the hybridization signals were performed as described by Kouchi and Hata (1993), except that hybridizations were performed at 55°C for PLA1 because of the high GC content of the gene.

Hormone measurement

Sampling of about 100 mg of fresh leaves from more than five seedlings was repeated three times for pla1-4, pla2-1, pla3-1 and wild type, respectively. Extraction and determination of CKs, ABA and IAA for each sample were performed using a liquid chromatography-tandem mass chromatography system (model 2695/Quattro Ultima Pt; Waters, http://www.waters.com/) as described previously (Nakagawa et al., 2005). The CK nucleotides in pla3 were determined for only two samples because of a mechanical accident, but the data were quite similar between the two replicates.

Accession numbers

PLA3 (AB447403), PLA1 (Q7Y1V5), PLA2 (AB244276), Os01g0740500 (BAF06116), Os01g0740600 (BAF06117). Os01g0743300 (BAF06131), AMP1 (ABQ85084), At4g07670 (CAB81137), At5g19740 (AAP37682), Vp8 (ACA62934), NAALADase I (AAC53423), NAALADase II (Q9Y3Q0), and NAALADase L (Q9UQQ1).

Acknowledgements

We thank Dr Yutaka Sato (Nagoya University) for technical assistance on microRNA analysis. We also thank N. Washizu, K. Ichikawa, R. Soga and K. Yatsuda for their assistance in cultivating rice plants at the Experimental Farm of the University of Tokyo. This work is supported in part by a Grant-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology of Japan (20248001 to Yasuo Nagato, and 20780001 and 20061005 to Jun-Ichi Itoh).

Ancillary