Transcripts of two ent-copalyl diphosphate synthase genes differentially localize in rice plants according to their distinct biological roles

Highlight Expression of the diterpene synthase gene for gibberellin biosynthesis occurs in tissues different from those in which its isoform for phytoalexin biosynthesis is expressed, reflecting their distinct biological roles in rice.


Introduction
Diterpenoids are a class of terpenoids mainly derived from the C20 prenyl substrate geranylgeranyl diphosphate (GGDP). The linear substrate GGDP is converted into various cyclic hydrocarbons by specific diterpene cyclases. The carbon skeletons are successively chemically modified by enzymes including P450 monooxygenases, dehydrogenases, methyltransferases, glucosyl transferases, and others. Gibberellins (GAs) are labdane-related diterpene phytohormones that regulate various aspects of plant growth, such as germination, stem elongation, and flowering (Yamaguchi, 2008;Hedden and Thomas, 2012). GAs are biosynthesized from the intermediate tetracyclic hydrocarbon ent-kaurene, which is converted from GGDP by two-step cyclization (Fig. 1). The two steps are catalysed by two distinct diterpene cyclases, ent-copalyl diphosphate (ent-CDP) synthase and ent-kaurene synthase. Other GA biosynthetic genes have been identified and characterized in detail (Yamaguchi, 2008;Hedden and Thomas, 2012).
A difference in enzymatic properties between recombinant OsCPS1 and OsCPS2 was previously shown (Hayashi et al., 2008). However, these studies did not clearly explain the nonredundant function of OsCPS1 and OsCPS2. Therefore, herein the localization of these transcripts in rice plants and the subcellular localization of their translated products were compared in order to account for their non-redundancy. Consequently, it was found that transcripts of OsCPS1 and OsCPS2 are differentially localized in rice plants according to their biological roles, and a complementation experiment using an OsCPS1 mutant by ectopic expression of OsCPS2 was performed to verify the conclusion genetically.

Plant materials
Rice (Oryza sativa L. cv. Nipponbare) seedlings were grown at 25 °C under continuous light conditions in a growth chamber until the third-leaf stage, and the upper and basal 2 cm regions were excised from the second-leaf sheath and used for quantitative analyses of transcripts. The oscps1-1 mutant (Sakamoto et al., 2004) is a Tos17inserted mutant NE3024 (Nipponbare background; Supplementary  Fig. S1 available at JXB online), and its M 1 seeds were purchased from the Rice Genome Resource Center, National Institute of Agrobiological Sciences. A heterozygous M 1 plant was used for transformation.

Laser microdissection
The upper parts of the second-leaf sheaths of third-leaf stage rice seedlings were fixed in ethanol:acetic acid (3:1, v/v). Paraffin embedding and laser microdissection were performed as previously described (Takahashi et al., 2010). In brief, 16 μm thin sections were cut from paraffin blocks and mounted on PEN membrane class slides (Life Technologies Corporation, CA, USA) for laser microdissection. The vascular bundle-rich and epidermis-rich tissues were collected from the leaf sheath sections using a Veritas Laser Microdissection System LCC1704 (Life Technologies Corporation). In addition, mesophyll-rich tissues were collected.

RNA extraction and quantitative reverse transcription-PCR (qRT-PCR)
Total RNA was extracted and purified from the leaf sheath samples using an RNAqueous kit (Invitrogen, Carlsbad, CA, USA),  and cDNA templates were synthesized from total RNA using a QuantiTect Reverse Transcription kit (Qiagen KK, Tokyo, Japan). Total RNA was extracted from tissue samples prepared by laser microdissection, using a PicoPure RNA Isolation kit (Life Technologies Corporation) according to the manufacturer's instructions. The quantity of RNA was determined by the Quant-iT RiboGreen RNA Assay Kit (Life Technologies Corporation). RNA integrity was assessed using an Agilent 2100 Bioanalyzer with an Agilent RNA 6000 Nano kit (Agilent Technologies, CA, USA). cDNA templates were synthesized and amplified using a WT-Ovation RNA Amplification System version 1.0 (NuGEN Technologies, CA, USA). qRT-PCR was performed using a Thermal Cycler Dice Real Time System TP800 (Takara Bio, Otsu, Japan) and SYBR Premix Ex Taq Perfect Real Time version 2 (Takara Bio), as previously described . The concentration of each transcript was normalized to 18S rRNA. Nucleotide sequences of the primers used are listed in Supplementary Table S1 at JXB online. The primer set for OsCPS1 can amplify a fragment derived from the wild-type OsCPS1 mRNA but not a fragment from Tos17-inserted OsCPS1 mRNA ( Supplementary Fig. S1).

β-Glucuronidase (GUS) assay
The GUS cDNA and NOS terminator originating in the pBI221 vector were inserted into the binary vector pZH2B (Kuroda et al., 2010) using BamHI and EcoRI sites (pZH2B-GUS-Nos; Supplementary  Fig. S3 at JXB online). The genomic DNA fragment of 2.1 kb 5' upstream of the OsCPS1 ATG start site plus the coding sequence in the second exon of OsCPS1 (OsCPS1p) was amplified by PCR using KOD Plus version 2 and primer set IF-AscI-CPS1p-F/IF-BamHI-CPS1p-R (Supplementary Table S3), and directly subcloned into pZH2B-GUS-Nos, which was digested with AscI and BamHI, using an In-Fusion HD cloning kit ( Supplementary Fig. S3). The plasmid pZH2B-OsCPS1p::GUS was introduced into Nipponbare rice cells through Agrobacterium transfection following the method previously described (Kuroda et al., 2010). Transgenic rice was grown and used for GUS assay. GUS assay was performed as previously described (Yamaguchi et al., 2001), using an Axioplan 2 microscope system (Zeiss Japan, Tokyo, Japan).

Complementation experiments
The NOS terminator originating in the pBI221 vector was inserted into pZH2B using SacI and EcoRI sites (pZH2B-Nos; Supplementary Fig. S3 at JXB online). The full-length open reading frame (ORF) cDNA of OsCPS2 was amplified by RT-PCR using the primer set KpnI-CPS2-F and KpnI-CPS2-R (Supplementary  Table S) and subcloned into the pZH2B-Nos vector using the KpnI site ( Supplementary Fig. S3). The OsCPS1p fragment was cleaved from the pZH2B-OsCPS1p::GUS plasmid and ligated into pZH2B-OsCPS2 using AscI and SmaI sites (pZH2B-OsCPS1p::OsCPS2; Supplementary Fig. S3). A heterozygous M 1 plant of oscps1-1 was selected as a transformation host by PCR genotyping using MightyAmp DNA polymerase (Takara Bio) and the primer sets CPS1-WT-F/CPS1-WT-R (~230 bp) for the wild-type gene and Tos17-F/CPS1-WT-R (~500 bp) for the mutant gene ( Supplementary  Fig. S1). pZH2B-OsCPS1p::OsCPS2 was introduced into the heterozygous rice plant by Agrobacterium infection (Kuroda et al., 2010), and T 1 seeds of 12 transgenic lines were obtained. The T 1 seedlings were grown and subjected to genotyping. The host OsCPS1 genotype was confirmed by PCR using the above primer sets, and the transgene OsCPS2 cDNA was detected by PCR using the primer set CPS2-QRT-F/CPS2-QRT-R (Supplementary Table  S1), ~250 bp from the endogenous gene (with intron) and ~170 bp from the transgene cDNA (no intron). The targeted-genotype plants (homozygous oscps1-1 mutants with OsCPS2 cDNA and segregated wild-type plants without OsCPS2 cDNA) were grown and T 2 seeds were obtained for further analyses. The upper and basal parts of T 2 seedlings were collected and used for gene expression analysis, after genotyping of each seedling by PCR, as described above.

Subcellular localization of OsCPS1 and OsCPS2
Diterpene cyclases, including ent-CDP synthase and entkaurene synthase, in higher plants have been considered to localize in the plastid (Toyomasu and Sassa, 2010). In general, transit peptides for plastid targeting are present at the N-termini of the plant diterpene cyclases (Toyomasu and Sassa, 2010), and Arabidopsis CPS, which is responsible for GA biosynthesis, has been shown to be transported into the plastid (Sun and Kamiya, 1994). OsCPS1 (867 amino acids, LOC_Os02g17780) and OsCPS2 (800 amino acids, LOC_Os02g36210) also have transit peptide-like sequences at their N-termini (Otomo et al., 2004;Prisic et al., 2004). To verify the plastid localization of OsCPS1 and OsCPS2, the N-terminal transit peptide-like sequences OsCPS1-N153 and OsCPS2-N108 ( Supplementary Fig. S2A at JXB online) were fused to GFP and expressed in rice mesophyll cells. GFP fluorescence in plastids indicated that both OsCPS1-N153-GFP and OsCPS2-N108-GFP were transported into the plastids (Fig. 2). These results suggest that OsCPS2 as well as GA biosynthetic OsCPS1 localize in the rice cell plastid.

Expression patterns of OsCPS1 and OsCPS2 in the second-leaf sheath of rice seedlings
The data suggested the same subcellular localization of OsCPS1 and OsCPS2 in the rice cell. Next, expression analysis of OsCPS1 and OsCPS2 was performed to compare the localization of these transcripts in rice plants. The second-leaf sheath grows sensitively in response to exogenously applied GA; therefore, it has been used in bioassays to assess GA activity (Murakami, 1968;Nishijima and Katsura, 1989). The second-leaf sheath was used here as material for qRT-PCR analysis. The upper and basal 2 cm were excised from secondleaf sheaths of third-leaf stage rice seedlings (Fig. 3A) and used for RNA extraction. The qRT-PCR analysis showed that the OsCPS2 transcript level was much lower than that of OsCPS1 in the basal part including the meristem tissues, whereas OsCPS2 expression was almost similar to OsCPS1 expression in the upper part (Fig. 3B). Furthermore, the transcript contents were quantified in separate tissues from the upper part of the second-leaf sheath in which the transcript contents of the two OsCPS genes were almost similar. Vascular bundle-rich, epidermis-rich, and residual tissues (mesophyll-rich tissues) were excised from leaf sheath slices, collected by laser microdissection (Fig. 4A), and used for RNA extraction. qRT-PCR was performed using amplified cDNAs derived from RNA obtained as a template.
OsCPS1 transcripts were significantly more abundant in vascular bundle-rich tissues than in other tissues, whereas OsCPS2 transcript levels were more abundant in epidermis-rich and mesophyll-rich tissues than in vascular bundle-rich tissues (Fig. 4B). Expression of the GA biosynthetic gene AtCPS, which encodes ent-CDP synthase in Arabidopsis, is observed around vascular tissues (Silverstone et al., 1997;Yamaguchi et al., 2001), similar to OsCPS1, but unlike OsCPS2.
Complementation of the severe dwarf phenotype of the OsCPS1 mutant by OsCPS2 expression under the OsCPS1 promoter A complementation experiment with a loss-of-function OsCPS1 mutant by OsCPS2 under the regulation of the OsCPS1 promoter was performed to confirm genetically that the OsCPS1 and OsCPS2 expression sites are different in rice. First a suitable DNA region of the OsCPS1 promoter wase identified using a GUS reporter gene. A genomic DNA fragment of 2.1 kb 5′ upstream of the OsCPS1 ATG start site plus the coding sequence in the second exon of OsCPS1 (OsCPS1p) was amplified using PCR and fused to GUS (Fig. 5) in the pZH2B binary vector (Kuroda et al., 2010). It has been shown previously that the first 1-2 introns of AtCPS are required for proper expression in Arabidopsis (Silverstone et al., 1997). GUS staining was observed in the vascular bundle of the second-leaf sheath of rice seedlings ( Fig. 5; Supplementary Fig. S4 at JXB online). The close correspondence of GUS staining to the accumulation pattern of OsCPS1 transcripts by tissue-specific qRT-PCR (Fig. 4B)

OsCPS1-N153
OsCPS2-N108 GFP Bright field GFP Chlorophyll Marged suggested that OsCPS1p promoter activity was effective. When the OsCPS1p fragment is fused to the full-length OsCPS2 ORF cDNA in-frame, the resulting translated product is OsCPS2, which has the OsCPS1-N91 at its N-terminus ( Supplementary Fig. S2B). Both OsCPS1-N91 and a chimeric peptide OsCPS1-N91:OsCPS2-N108 led GFP to the plastid ( Supplementary Fig. S5), similar to OsCPS2-N108 (Fig. 2). These results suggest that a pseudo-mature form of OsCPS2 is transported into plastids when driven by OsCPS1p. The OsCPS1p fragment was accordingly fused to OsCPS2 ORF cDNA (Fig. 5) in the pZH2B vector and introduced into a heterozygous oscps1-1 mutant by Agrobacterium infection. oscps1-1 is a Tos17-inserted OsCPS1 mutant that displays a severe dwarf phenotype caused by GA deficiency (Sakamoto et al., 2004). Because the homozygous oscps1-1 mutant was incapable of dedifferentiation, a heterozygous mutant was transformed by Agrobacterium infection, and self-pollination of transgenic plants produced T 1 seeds. Transgenic homozygous oscps1-1 (complemented line), non-transgenic homozygous oscps1-1 (knockout line), and non-transgenic wild-type (segregated wild-type line) plants were observed by genotyping the T 1 seedlings (Fig. 6A). Four complemented line T 1 plants were identified, and displayed almost the same height as that of the segregated wild-type lines (Fig. 6B). OsCPS1 and OsCPS2 expression patterns were analysed in the second-leaf sheath to check successful ectopic expression of OsCPS2. qRT-PCR showed that OsCPS2 transcripts accumulated in basal parts of the second-leaf sheath of T 2 complemented line seedlings, similar to OsCPS1 transcripts in wild-type Nipponbare (Fig. 3B) and in segregated wildtype line seedlings, whereas wild-type OsCPS1 transcripts were not detected in complemented line plants (Fig. 7). These results indicate that OsCPS2 expression under the OsCPS1 promoter rescued the oscps1-1 mutant phenotype.

Discussion
Phytohormone GAs are biosynthesized from GGDP, a common precursor of diterpenoids, through several steps catalysed by various biosynthetic enzymes including diterpene cyclases in plastids, P450 monooxygenases in the endoplasmic reticulum, and 2-oxoglutarate-dependent dioxygenases in the cytoplasm (Yamaguchi, 2008;Hedden and Thomas, 2012). The sites of expression of genes encoding the soluble dioxygenases GA 20-oxidase and GA 3-oxidase, which are responsible for direct synthesis of physiologically active GAs, have been characterized in rice previously (Kaneko et al., 2003), whereas the sites of expression of diterpene cyclase genes, including OsCPS1 and OsKS1, involved in the initial step of bioactive GA biosynthesis have not been elucidated. The diterpene cyclases responsible for biosynthesis of entkaurene, a GA biosynthetic intermediate, have been identified not only in plants but also in bacteria (Morrone et al., 2009) and fungi (Kawaide et al., 1997;Toyomasu et al., 2000), both of which produce GAs as specialized metabolites. Although two distinct cyclases successively convert GGDP into entkaurene via ent-CDP in higher plants, one peptide bi-functional diterpene cyclase produces ent-kaurene from GGDP in bryophytes (Hayashi et al., 2006;Kawaide et al., 2011). Coniferous gymnosperms also have ent-CDP synthase and ent-kaurene synthase for GA biosynthesis, while they have bi-functional labdane-related diterpene cyclases involved in specialized metabolites, resin acids produced via (+)-CDP (Keeling et al., 2010;Zerbe et al., 2012). Synthesis of ent-CDP is the first branch point of GA biosynthesis from GGDP, but ent-CDP and ent-kaurene may not be specific intermediates of only GAs. For example, Stevia rebaudiana produces steviol glycosides via ent-kaurene. In S. rebaudiana, ent-kaurene is synthesized by highly accumulated ent-CDP synthase and ent-kaurene synthase, both of which are responsible for GA biosynthesis, in leaf parenchyma (Richman et al., 1999). In addition, rice produces labdane-related phytoalexins converted from ent-CDP; phytocassanes A-E and oryzalexins A-F. Rice has an ent-CDP synthase, OsCPS2, specific for phytoalexin biosynthesis as well as OsCPS1 specific for GA biosynthesis (Otomo et al., 2004;Prisic et al., 2004), whereas S. rebaudiana uses ent-CDP synthase for GA biosynthesis to synthesize steviol glycosides.
Here it is shown that transcript levels of OsCPS2 were drastically lower than those of OsCPS1 in the basal parts of rice seedling second-leaf sheath. The basal parts include the meristematic tissues necessary for rice growth. It has been shown that growth signals mediated by GAs engage in antagonistic cross-talk with defence signals mediated by jasmonic acid via the DELLA-JAZ interaction (Hou et al., 2013). DELLA and JAZ are key repressors in GA and jasmonic acid signalling, respectively. Therefore, expression of the defence gene OsCPS2 might be suppressed in the basal parts with high growth activity, although it has been unclear whether crosstalk of these hormones is involved in the OsCPS2 regulation in the basal parts of rice seedlings. Furthermore, qRT-PCR analysis using separate tissues prepared by laser microdissection indicated that OsCPS1 transcripts mainly accumulated in vascular bundle-rich tissues, whereas OsCPS2 transcripts were mainly observed in epidermis-rich tissues. Thus, the different localization of OsCPS1 and OsCPS2 transcripts in rice plants was found.On the other hand, it was suggested that both OsCPS2 and OsCPS1 function in the plastid. It was previously shown that GAs are biosynthesized mainly from GGDP that is derived through the methylerythritol 4-phosphate (MEP) pathway in the plastid (Kasahara et al., 2002). Transcript levels of several genes in rice responsible for the MEP pathway are drastically up-regulated after elicitor treatment, suggesting that the MEP pathway also participates in diterpene phytoalexin biosynthesis in rice (Okada et al., 2007). Therefore, it is reasonable that OsCPS2 functions in the plastid like GA biosynthetic OsCPS1. Although recombinant OsCPS2 converts GGDP to ent-CDP in vitro, OsCPS2 cannot rescue loss-of-function OsCPS1 mutants under control of the native promoter (Sakamoto et al., 2004). In the present study, it was shown that OsCPS2 under regulation by the OsCPS1 promoter complemented the OsCPS1 mutant phenotype. The qRT-PCR results showed differences in tissue-specific expression of the two CPS genes. These results strongly suggest that proper tissue-specific expression of ent-CDP synthase genes is critical for GA biosynthesis. Results of a GUS reporter gene assay in germinating Arabidopsis seeds suggest that AtCPS transcripts for GA biosynthesis localize in vascular tissues, although no signal was detected by in situ hybridization (Silverstone et al., 1997). Another study showed that transcripts of dioxygenases, responsible for later steps of GA biosynthesis, localize mainly in the cortex and endodermis, and promoter-swapping experiments indicate that intercellular transport of the biosynthetic intermediate ent-kaurene is required to produce bioactive GAs (Yamaguchi et al., 2001). An SrCPS signal involved in stevioside biosynthesis in S. rebaudiana was detected in mesophyll by in situ hybridization, whereas its signal for GA biosynthesis in vascular tissues was not detected using this method (Richman et al., 1999), similar to AtCPS (Silverstone et al., 1997). It has been suggested that a spatially different localization of CPS transcripts separates stevioside biosynthesis from GA biosynthesis in S. rebaudiana. In any case, CPS expression patterns associated with GA biosynthesis are considered well conserved among higher plant species. However, it is reasonable that transcription of stress-inducible OsCPS2 mainly occurs near the epidermis, producing phytoalexins in response to environmental stress.  results confirm that the first step in GA biosynthesis occurs in the vascular tissues for effective production of bioactive GAs. The monocots wheat and maize not only have the ent-CDP synthase genes TaCPS3 and An1, which are responsible for GA biosynthesis, but also have stress-inducible ent-CDP synthase genes TaCPS1 and An2 (Harris et al., 2007;Wu et al., 2012). Similar to rice, TaCPS1 and An2 may be expressed in different tissues from those of the GA biosynthetic genes TaCPS3 and An1.
It was shown herein that OsCPS1 and OsCPS2 transcripts were differentially localized in rice plants according to their distinct biological roles, one for growth and the other for defence, although these translated products have the same enzymatic activity in vitro. It is concluded that OsCPS2 contributes little to GA biosynthesis and cannot prevent the GA-deficient dwarf phenotype of loss-of-function OsCPS1 mutants because of different localization of OsCPS2 transcripts from OsCPS1 transcripts.

Supplementary data
Supplementary data are available at JXB online. Figure S1. Primer design for transcript analyses and genotyping Figure S2. Transit peptide-like sequences of OsCPSs Figure S3. Construction of plasmids for introducing the transgene Figure S4. GUS staining of rice seedlings. Figure S5. Subcellular localization of GFP fused to OsCPS1-N91 and OsCPS1-N91:OsCPS2-N108 at its N-terminus Table S1. Sequences of primers used for qRT-PCR. Table S2. Sequences of primers used for GFP experiments. Table S3. Sequences of primers used for complementation experiments.