Background

Chloroplasts are essential for plant development and growth, through manipulating the fixation of CO2 and biosynthesis of carbon skeletons and other physiological processes (Jarvis and López-Juez 2013). Accumulating evidence has shown the importance of chloroplast biogenesis and development during germination for plant vitality, seed set and growth (Lopez-Juez and Pyke 2005; Pogson and Albrecht 2011). Chloroplast biogenesis is initiated from proplastids through endosymbiosis from an ancestor of extant cyanobacteria and is dependent on the coordinated expression of genes encoded in both nuclear and plastid genomes (Lopez-Juez and Pyke 2005; Kessler and Schnell 2009). The development of chloroplasts differs between organs and species depending on the specialization of tissues and stage of development. For example, distinct phenotypes between cotyledons and true leaves were observed in variegated (var) and snowy cotyledon (sco) mutants in Arabidopsis, respectively (Albrecht et al. 2008; Liu et al. 2010; Zagari et al. 2017), i.e. chlorotic true leaves but green cotyledons in the var. and chlorotic or bleached cotyledons but green true leaves in the sco mutants. It is evident that nucleus-encoded polymerases (NEPs) and plastid-encoded polymerases (PEPs) involved in gene transcription, RNA maturation, protein translation and modification have great effect on the biogenesis of chloroplasts (Pogson and Albrecht, 2011; Yu et al. 2014).

Isoprenoids play essential roles in plant growth and development. In higher plants, biosynthesis of the basic isoprenoid units occurs by two different pathways: the mevalonate (MVA) pathway and the methyl-D-erythritol-4-phosphate (MEP) pathway. For decades, it was thought that the MVA pathway was solely responsible for isoprenoid biosynthesis. However, the alternative MEP pathway was discovered recently in bacteria, green algae, and higher plants (Rohmer et al. 1993; Cunningham Jr. et al. 2000; Rodriguez-Concepcion and Boronat, 2002; Rohmer, 2003). The MVA pathway occurs in the cytosol and is responsible for the synthesis of sterols, certain sesquiterpenes, and the side chain of ubiquinone (Disch et al. 1998). By contrast, the MEP pathway operates in the plastids and is involved in providing the precursors for monoterpenes, isoprene, chlorophylls, carotenoids, tocopherols, taxadiene, gibberellins, and abscisic acid (Zeidler et al. 1997; Eisenreich et al. 1998; Lichtenthaler, 1999). In recent years, the entire MEP pathway and almost all the enzymes involved have been identified in Escherichia coli. The activity of the corresponding enzyme and the roles they played in the pathway have been widely demonstrated (Lois et al. 1998; Herz et al. 2000; Hecht et al. 2001; Adam et al. 2002). The genes encoding each enzyme in the MEP pathway are highly conserved. Genes encoding enzymes involved in the MEP pathway are important for chlorophyll and carotenoid biosynthesis in plants. Knockout mutations of several genes in this pathway in Arabidopsis, Zea may and Nicotiana benthamiana display the albino phenotype (Budziszewski et al. 2001; Gutierrez-Nava et al. 2004; Page et al. 2004; Hsieh and Goodman, 2005; Lu et al. 2012). However, the genes involved in the MEP pathway in rice has not been reported yet and needs further study.

In this study, we obtained a rice mutant lethal albinic seedling 1 (las1) by the CRISPR/Cas9 system. Identification of CRISPR/Cas9 sites revealed sequence differences in the gene encoding a 4-hydroxy-3-methylbut-2-enyl diphosphate reductase, which is involved in the isoprenoids biosynthesis. The mutant seedlings were albinic and eventually died from lacking of chlorophyll. Additionally, the transcript levels of genes associated with chlorophyll biosynthesis, and those associated with chloroplast biogenesis were severely affected in the las1 mutant. These findings implicate that rice LAS1 plays an important role in chloroplast biogenesis during early leaf development.

Results

Characterization of the las1 Mutant in Rice

We isolated a rice mutant by the CRISPR/Cas9 system, which exhibited defects in leaf color, termed lethal albinic seedling 1(las1). The las1 mutant exhibited the albino phenotype and eventually died (Fig. 1a; Additional file 1: Figure S1). Consistent with their phenotype, the chlorophyll contents were significantly reduced in las1 compared with that of the wild type (Fig. 1b). Sequence confirmation indicated that the phenotype of las1 resulted from the LOC_Os03g52170 mutation encoding a 4-hydroxy-3-methylbut-2-enyl diphosphate reductase, OsHMBPP (Fig. 1c). We identified five homozygous mutants and the phenotype of the heterozygous mutants were similar to that of WT (Additional file 1: Figure S1). Protein sequence alignments of the homozygous mutants and the wild type protein revealed that three mutants showed premature translational stops, and the line las1–1 encodes the least amino acids. So we chose the line las1–1 to represent the LAS1 mutant for further study. These data suggested that LAS1/OsHMBPP is essential for early chloroplast development in rice.

Fig. 1
figure 1

The phenotype and pigment contents of wild type and las1. a Morphology of wild type (WT) and las1 seedling (Bar = 5 cm). b Pigment contents of WT and las1 seedlings. Values represent the mean ± SD of three biological replicates. **p < 0.01. c Sequence confirmation of the LAS1 mutation. The 23-nt target sequence of the Cas9/sgRNA complex is underlined in red

Full size image

Chloroplast Ultrastructure of the las1 Mutant

Chloroplast development depends on the successive synthesis and integration of Chl into photosynthetic complexes (Wang & Grimm, 2015). To further investigate the albino leaf phenotype of las1, we observed the chloroplast structure of las1 and wild type in leaves by TEM. The chloroplasts of the WT leaves were normal in shape, containing large starch granules and normal thylakoid membranes with stacks of grana (Fig. 2a-b). In contrast, chloroplasts were severely disrupted in the las1 leaves (Fig. 2c-d). Our results indicated that the albino leaf phenotype of las1 resulted from the abnormal development of chloroplasts.

Fig. 2
figure 2

TEM analysis of chloroplast ultrastructure in wild type and the las1 mutant. a-b Chloroplast structure in WT cell. c-d Chloroplast structure in las1 cell. Scale bars 1 μm for (b) and (d); 5 μm for (a) and (c)

Full size image

Expression Pattern of OsHMBPP

According to the Rice eFP Browser (http://bar.utoronto.ca/efprice/cgi-bin/efpWeb.cgi), we found that OsHMBPP was expressed in all tissues, especially in leaves (Additional file 1: Figure S2). To verify these data, we examined the expression level of LAS1 in different organs of the wild type by qRT-PCR. OsHMBPP was highly expressed in the mature leaves and also detected in the sheaths, seeds, roots and panicles (Fig. 3a). Our results demonstrated that OsHMBPP is constitutively expressed in various tissues and mainly functions in the green tissues.

Fig. 3
figure 3

Expression analysis of OsHMBPP. a Expression of OsHMBPP in roots, sheaths, leaves, panicles, and seeds of wild-type plants. b Expression pattern of OsHMBPP in a 14-h light/ 10-h dark cycle. Values represent the mean ± SD of three biological replicates

Full size image

To further investigate whether the expression of OsHMBPP is modulated by photoperiod, we compared the day/night expression patterns during the 14-h light/10-h dark photoperiod conditions. Test samples were taken every 4 h. qRT-PCR showed the expression of OsHMBPP oscillated during a light/dark cycle and the highest transcript level was detected during light period (Fig. 3b).

LAS1/OsHMBPP Encodes a Conserved Chloroplast-Targeted Protein

LAS1/OsHMBPP encodes a HMBPP reductase, which is involved in catalyzing 2C-methyl-D-erythritol4-phosphate (MEP) converted to IPP and DMAPP in the nonMVA pathway in Arabidopsis thaliana and maize, and affect the development of chloroplasts (Hsieh and Goodman, 2005; Lu et al. 2012). Therefore, we measured the contents of IPP and DMAPP between WT and las1. IPP and DMAPP contents in the las1 seedlings were significantly reduced, indicating that the nonMVA pathway in las1 was impaired (Additional file 1: Figure S3). Based on amino acid sequence alignment, OsHMBPP shares high identity with its orthologs in other species, including Sorghum bicolor (89.7%), Arabidopsis (73.8%), and Zea mays (90.1%) (Fig. 4b). Given that OsHMBPP and Arabidopsis, HDR, share 73.8% of amino acid identity and that structural predictions of both proteins revealed three-dimensional similarities by I-TASSER algorithm (Fig. 4a). ChloroP analysis revealed that the OsHMBPP protein contains a predicted chloroplast transit peptide of 31 amino acid residues at its N-terminal (Additional file 1: Table S1). To further determine its subcellular localization, a transient expression system was performed in rice protoplasts. As expected, the fusion protein OsHMBPP-GFP was clearly co-localized with Chl auto-fluorescence (Fig. 5), indicating OsHMBPP was localized in chloroplasts.

Fig. 4
figure 4

Protein structure of OsHMBPP and sequence alignment of multiple amino acids from different organisms. a Structure prediction of the HMBPP reductase IspH (Escherichia coli), HDR (Arabidopsis), HDR(Zea mays) and OsHMBPP (Oryza sativa) by I-TASSER algorithm. b Amino acid sequence alignment of the 3 types of OsHMBPP homologs. Amino acids that were fully or partially conserved are shaded blue and green, respectively

Full size image
Fig. 5
figure 5

Subcellular location of the OsHMBPP protein. GFP signals of OsHMBPP-GFP fusion protein was located in the chloroplasts by transient expression analyses in rice protoplasts. Green fluorescence shows GFP, red fluorescence shows chloroplast autofluorescence, and yellow fluorescence shows the merged fluorescence

Full size image

Altered Expression of Chlorophyll Biosynthesis and Plastid-Encoded Genes

We examined the expression levels of nuclear genes associated with Chl biosynthesis. qRT-PCR analyses showed that some Chl biosynthesis related genes were significantly down-regulated in the las1 mutant (Fig. 6b). To investigate whether the LAS1 mutation affects transcription by PEP and NEP, we examined the transcript abundance of various plastidic genes in the las1 mutant by qRT-PCR. Plastid genes consist of three types of genes, which are transcribed by NEP, PEP, and both NEP and PEP, respectively (Yu et al. 2014). In las1, the expression levels of class III genes (e.g rpoA, rpoB, rpoC) were strongly increased, while class I genes (e.g psaA, psbA, rbcL) decreased (Fig. 6a). These results suggest that las1 was defective in PEP activity and influenced the optimal expression of plastidic genes in rice seedlings.

Fig. 6
figure 6

Expression analysis of genes involved in chloroplast development and chlorophyll biosynthesis. a Expression analysis of chloroplast development related genes in wild-type and las1. b Expression analysis of chlorophyll biosynthesis related genes in wild-type and las1. The relative expression level of each gene in WT plants was set to 1.0. Ubiquitin gene was used as a reference. Error bars indicate SD (n = 3)

Full size image

Mutation in LAS1 Affects Four Plastdic Editing Sites and Chloroplast Ribosome Biogenesis

The translation deficiency in las1 suggests that LAS1 might be involved in posttranscriptional RNA metabolism processes. We next detected whether LAS1 was involved in the chloroplast RNA editing and RNA splicing. There are 21 chloroplast RNA editing sites in the 12 chloroplast genes (Corneille et al. 2000). Therefore, we analysed all 21 sites in the chloroplast cDNA in wild type and las1. We found that four C sites of chloroplast genes in las1 did not edit normally. The editing of rpl2-C2, ndhG-C(− 11), and atpA-C1159 were reduced in las1, but the editing of rpoB-C467 was increased (Fig. 7). There was no difference between the wild type and las1 in the rest 17 RNA editing sites (Additional file 1: Figure S4). There are 18 introns in the rice chloroplast genome (Kaminaka et al. 1999). Furthermore, we carried out RT-PCR with primers flanking the introns and compared the PCR products length between wild type and las1 by agarose gel. Splicing of the 17 chloroplast transcripts were not significantly impeded in las1 (Additional file 1: Figure S5). These data suggest that LAS1 affects the RNA editing efficiency of chloroplast genes, but not function in chloroplast RNA splicing.

Fig. 7
figure 7

The OsHMBPP protein is required for RNA editing at multiple sites

Full size image

Abnormal chloroplast development is often associated with inhibition of chloroplast ribosome synthesis (Ge et al. 2017; Cao et al. 2019), which is composed of 30 S small subunit and 50 S large subunit. We further examined whether the activity of chloroplast ribosome changed in las1. qRT-PCR analysis indicated that the expression levels of 16 S and 23 S were decreased whereas 18 S and 25 S were increased in las1, compared with the wild type (Fig. 8). The result indicated that las1 was defective in chloroplast ribosome biogenesis.

Fig. 8
figure 8

The expression level of 16 S, 23 S, 18 S and 25 S in wild type and las1. Error bars represent the SD from three independent experiments

Full size image

LAS1/OsHMBPP Interacts with the MORF Family Protein in Yeasts

The multiple organellar RNA editing factor (MORF) family proteins are essential for the chloroplast/mitochondrion RNA editing in plants (Bentolila et al. 2012; Takenaka et al. 2012; Zhang et al. 2017). There are nine MORF proteins and seven MORF proteins in Arabidopsis and rice, respectively. To investigate whether LAS1/OsHMBPP interacted with the MORF family proteins, we used the yeast two hybrid to detect the interaction. Notably, the results showed that only Os09g33480 had an interaction with OsHMBPP (Fig. 9a-b). Furthermore, we performed BiFC assays using N. benthamiana cells and found that OsHMBPP-nYFP and Os09g33480-cYFP could interact with each other (Fig. 9c). Os09g33480 belongs to the Arabidopsis MORF8 branch. Os09g33480 may be targeted to both mitochondria and chloroplasts similar to Arabidopsis MORF8. Meanwhile, WSP1, Os06g02600, Os09g04670, Os08g04450, Os11g11020, and Os03g38490 did not interact with OsHMBPP (Fig. 9a-b). Based on the Rice eFP Browser, Os09g33480 is ubiquitously expressed in many rice organs, especially in seeds, panicles and leaves (Additional file 1: Figure S6).

Fig. 9
figure 9

The interaction assay between OsHMBPP and MORF families in rice. a-bOsHMBPP was fused to the pGBKT7 vector (OsHMBPP-BD). Seven MORFs were fused to the pGADT7 vector. DDO, control medium (SD–Leu/−Trp); QDO, selective medium (SD–Leu/−Trp/−His/−Ade). c Interaction between OsHMBPP and Os09g33480 in BiFC assays. Bars = 100 μm

Full size image

Expression Patterns of the MEP Pathway Genes in a Normal Day/Night Cycle

To further investigate whether expression of the MEP pathway genes is regulated by photoperiod, we compared the expression patterns of six genes during the 14-h light/10-h dark photoperiod conditions (Additional file 1: Figure S7). The expression patterns of CMK and DXS were similar during the 14-h-light/10-h-dark cycle. Peak levels of CMK and DXS were detected in the early period of the light cycle (8–12 h in the light). Oscillations in the expression of DXR, HDS, MCS, and CMS also occurred during the light/dark cycle. In contrast to CMK and DXS, peak levels of DXR and HDS expression appeared in the late period of the dark cycle (20-24 h). The expression of all MEP pathway genes is significantly repressed during the transition from light to dark (Additional file 1: Figure S7).

Discussion

Characterization of LAS1/OsHMBPP

Chloroplast defective development generally results in an albino or chlorotic leaf phenotype, which severely affects the biomass and even the survival of plants, particularly in crops. Isoprenoids play an essential role in chloroplast development. In this study, we isolated the las1 mutant, which displayed an albino leaf phenotype and was seedling lethal (Fig. 1). TEM observation demonstrated that the phenotype of las1 resulted from the defective development of thylakoids, eventually leading to the breakdown of entire chloroplasts (Fig. 2). LAS1 encodes 4-hydroxy-3-methylbut-2-enyl diphosphate reductase OsHMBPP in rice (Fig. 5).HMBPP functions in the isoprenoid biosynthesis via the MEP pathway (McAteer et al. 2001; Guevara-Garcia et al. 2005; Hsieh and Goodman, 2005). Isoprenoids are derived from a basic five-carbon unit, isopentenyl diphosphate (IPP) and allyl isomerdimethylallyl diphosphate (DMAPP). HMBPP is the rate limiting enzyme of this process, and regulates the biosynthesis of chlorophylls, ABA and GA. Mutation of HMBPP in maize and A. thaliana produced an albino phenotype with impaired chloroplasts containing large vesicles (Hsieh and Goodman, 2005; Lu et al. 2012). In rice, the OsHMBPP mutation resulted in a reduction in the content of IPP and DMAPP (Additional file 1: Figure S3), which led to decrease of the chlorophyll content and plant death. On the other hand, OsHMBPP interacted with the MORF protein Os09g33480(Fig. 9), and affected the RNA editing of chloroplast genes. These results indicate that the function of HMBPP is very conserved in different species, and is essential for the early chloroplast development in rice.

LAS1 Encodes a Chloroplast-Targeted Protein and Regulates the Early Chloroplast Development in Rice

Our results showed that LAS1 was constitutively expressed, with high levels of transcript in the leaves, implying that LAS1 might play an essential role in leaf development (Fig. 3a). Subcellular localization in rice protoplasts demonstrated that the LAS1 protein is routed to the chloroplasts (Fig. 4). The chloroplast is a semi-autonomous organelle, which contains about 100 genes, although more than 3000 proteins function within it (Leister, 2003). Thus, nucleus-encoded factors play essential roles in regulating chloroplast development, which requires the coordinated expression of both nucleus-encoded and chloroplast-encoded genes. The LAS1 mutation disrupts the transcripts of plastid and nuclear genes associated with chloroplast development (Fig. 6). The transcript accumulation of both PEP-and NEP-dependent genes(atpB) and PEP-transcribed plastid genes (psaA, psbA, rbcL) were severely suppressed (Fig. 6), suggesting that the accumulation of transcripts for PEP components did not result in the formation of functional PEP due to the disruption of transcription/ translation apparatus.

OsHMBPP Is Required for Chloroplast Genes Transcript Editing

RNA editing is an important event in post-transcriptional RNA process, which alters RNA sequences by converting specific target cytidines to uridine in transcripts of plastids and mitochondria (Takenaka et al. 2013). Abolishing editing in plastidic genes can disturb plastidic functionality and inhibit rice normal development and growth (Tang et al. 2017; Zhang et al. 2017; Liu et al. 2018; Cui et al. 2019). For example, in the wsp1 mutant, editing at ndhD-878, rps14–80, ndhG(− 11), and rpoB-467, 545, 560 are abolished, resulting in the defective chloroplast development. Meanwhile, the mutation of pentatricopeptide repeat protein gene DUA1 reduced the editing efficiency of rps-182 and rpoB-545, 560, leading to the seedling lethality under cold stress. Our findings showed that the editing efficiency of four chloroplast RNA editing sites in las1 is remarkably altered, compared with the wild type (Fig. 7), which may be an explanation for the phenotype of las1. Edited rpoB codon 467 is located in the Dispensable Region I and the editing of the rpoB-467 site is non-essential (Corneille et al. 2000). In rice, the rpoB-467 site is partial editing. Therefore, the altered editing of the rpoB-467 site was not the cause of the phenotype of las1. The chloroplast and mitochondria RNA editing events may be an indication of evolutionary adaptation (Chu and Wei, 2019). The rpl2, rpoB and atpA gene sequences of the chloroplast genomes of different species were analyzed from the perspective of evolution. The rpoB-467 and atpA-1159 encode a Ser in monocots, but encode either a Ser or a Leu in eudicots (Additional file 1: Figure S8). We found an interesting phenomenon that rpl2–2 encodes Thr in monocots, but it encodes Met in eudicots (Additional file 1: Figure S8). These results showed that the rpl2–2 sites of monocots and eudicots were highly differentiated. The ‘C’ at rpl2–2 of monocots may have been acquired during the evolution.

Plant MORF proteins are essential for RNA editing and splicing, but they are rarely reported in rice. Recently, A MORF2-like protein, WSP1, was identified as a rice RNA editing/splicing factor (Zhang et al. 2017). The WSP1 mutation affects the RNA editing of ndhG and rpoB, suggesting that WSP1 and OsHMBPP might cooperate together for RNA editing. Therefore, we tested the interaction between OsHMBPP, WSP1 and other six MORF proteins. Y2H assays showed that OsHMBPP and WSP1 did not interact together in vivo (Fig. 9). However, we confirmed that OsHMBPP can interact with Os09g33480 by Y2H and BIFC assays (Fig. 9a-c), but not interact with the other six MORFs in yeasts (Fig. 9a). Our results suggested that OsHMBPP may act as an organellar RNA editing factor via an editosome coupled with Os09g33480.

Conclusions

We isolated a novel seedling-lethal albino rice mutant las1 by the CRISPR/Cas9 system. Mutation in the HMBPP reductase is responsible for the phenotype of las1. Our results demonstrate that LAS1/OsHMBPP is important for plastid gene expression and normal chloroplast development. Also, we found the plastid RNA editing in las1 was altered, and LAS1/OsHMBPP interacted with the MORF protein Os09g33480 in vivo. Further studies on identifying the mutant phenotype of other MEP pathway genes, and how the rice MEP pathway enzymes are regulated.

Methods

Plant Materials and Growth Conditions

The las1 mutant was obtained from the pool of the japonica cultivar Nipponbare by the CRISPR/Cas9 system. The CRISPR/Cas9 binary vectors were constructed as previously described (Lu et al. 2017). The seedlings were grown in a growth chamber or in the field. The growth chamber condition was set as follow:14-h light at 30 °C and 10-h dark at 25 °C.

Measurement of Pigment Contents

Pigments were extracted from fresh leaf samples (0.2 g) from wild-type and las1 plants that were cut into small segments and incubated with 5 ml of 95% ethanol in the dark for 48 h. The supernatant was measured with a dual-beam ultraviolet spectrophotometer (TU1901, Beijing) at 470, 645, and 663 nm. Chl a, Chl b, and carotenoid (Car) contents were measured according to the methods by Porra et al. (1989). Three biological replicates were analyzed.

Transmission Electron Microscopy (TEM)

For TEM analysis, transverse sections of leaf samples were taken from las1 and WT seedlings. Leaf samples were fixed in 2.5% glutaraldehyde and then in 1% OsO4. Samples were prepared as described previously (Gothandam et al. 2005), and were observed using a Jeol 1230 electron microscope (Tokyo, Japan).

Protein Structure and Sequence Alignment Analysis

The predicted full-length LAS1 protein sequence, with 459 amino acids, was obtained from RGAP (http://rice.plantbiology.msu.edu/). The sequences used in the alignment analysis were obtained by a BLASTP search using the LAS1 protein sequence as the query at the National Center for Biotechnology Information (NCBI, http://www.ncbi.nlm.nih.gov/). The full-length amino acid sequences were aligned using the DNAMAN.

Measurement of IPP and DMAPP Contents

IPP and DMAPP contents were determined by the specific ELISA Kit from the Sci-tech innovation company (Qingdao, China). Baed on the manufacturer’s instructions, each sample was assayed five times.

Subcellular Localization

To investigate the subcellular localization of LAS1, we amplified the full-length cDNA of LAS1 without the termination codon (Additional file 1: Table S2). The fragment was cloned into the GFP vector pAN580. The LAS1-GFP vector was transformed into rice protoplasts (Wang et al. 2016), and the transformed protoplasts were observed with a Zeiss LSM700 laser scanning confocal microscope.

RNA Extraction and qRT-PCR Analysis

Total RNA was extracted and purified from different tissues using the RNApure Plant Kit (CWBIO) according to the manufacturer’s instructions. First strand cDNAs were synthesized from 2 μg total RNA using a PrimeScript 1st Strand cDNA Synthesis Kit (Takara). qRT-PCR was performed with a CFX96 Touch Real-time PCR Detection System using the SYBR Premix Ex TaqTM kit (Takara). The rice UBQ gene was used as an internal control. All primers for qRT-PCR are obtained as reported previously (Wang et al. 2016; Ge et al. 2017; Liu et al. 2018). The data were expressed as the mean ± SD of three biological replicates.

RNA Editing Analysis

All 21 known rice chloroplast RNA editing sites were assayed as reported (Corneille et al. 2000). Specific cDNA fragments with the editing sites were amplified between wild type and las1, and then sequenced. The cDNA sequences were compared to identify C to T changes resulting from RNA editing by DNAMAN software. All primers for RNA editing analysis are obtained as reported previously (Wang et al. 2017).

Yeast Two-Hybrid and Transient Expression Assays

Full-length coding sequences of LAS1 was cloned into pGBKT7 vector, and seven MORF genes (Os06g02600, WSP1, Os03g38490, Os09g33480, Os09g04670, Os11g11020, Os08g04450) were cloned into pGADT7 vector, resulting in LAS1-BD and MORFs-AD respectively (Additional file 1: Table S2). The assay was performed following the manufacturer’s instructions (Clontech).

For Bimolecular Fluorescent Complimentary assay, OsHMBPP-nYFP, Os09g33480-cYFP and corresponding empty vectors were transformed into Agrobacterium strain EHA105 and transfected leaves of N. benthamiana (Additional file 1: Table S2). The fluorescence was observed with a confocal scanning microscopy (ZEISS).