In angiosperms, chlorophyll biosynthesis is light dependent. A key factor in this process is protochlorophyllide oxidoreductase (POR), which requires light to catalyze the reduction of protochlorophyllide to chlorophyllide. It is believed that this protein originated from an ancient cyanobacterial enzyme that was introduced into proto-plant cells during the primary symbiosis. Here we report that PORs from the cyanobacteria Gloeobacter violaceus PCC7421 and Synechocystis sp. PCC6803 function in plastids. First, we found that the G. violaceus POR shows a higher affinity to its substrate protochlorophyllide than the Synechocystis POR but a similar affinity to plant PORs. Secondly, the reduced size of prolamellar bodies caused by a knockdown mutation of one of the POR genes, PORA, in Arabidopsis could be complemented by heterologous expression of the cyanobacterial PORs. Photoactive protochlorophyllide in the etioplasts of the complementing lines, however, was retained at a low level as in the parent PORA knockdown mutant, indicating that the observed formation of prolamellar bodies was irrelevant to the assembly of photoactive protochlorophyllide. This work reveals a new view on the formation of prolamellar bodies and provides new clues about the function of POR in the etioplast–chloroplast transition.
Chlorophyll (Chl) is the most abundant natural pigment on Earth and is necessary for photosynthetic organisms to capture sunlight and convert it to photochemical energy. The coordinated synthesis of Chl and its binding proteins is thus critically important for all living organisms. Within the Chl biosynthetic pathway, reduction of the fourth ring of the Mg-tetrapyrrole intermediate, protochlorophyllide (Pchlide), is one of the most important regulatory steps and has been extensively characterized (Lebedev and Timko, 1998; Masuda and Takamiya, 2004; Heyes and Hunter, 2005; Rudiger, 2006; Belyaeva and Litvin, 2007). Interest in this step is due to the dependence of angiosperms on light for the catalytic reaction. When angiosperm seedlings grow in the dark, such as in the soil, colorless plastids, called etioplasts, accumulate a large amount of Pchlide and are incapable of photosynthesis. Once the dark-grown seedlings are exposed to sunlight, they start to synthesize Chl and develop chloroplasts to achieve photosynthesis. A key regulatory factor in this process is Pchlide oxidoreductase (POR), which requires light energy to catalyze the reduction of Pchlide to chlorophyllide. In the dark, the POR of angiosperms is accumulated with its substrates, Pchlide and NADPH, to form aggregates called prolamellar bodies in etioplasts; these aggregates are broken down upon light illumination in parallel with Chl synthesis. Thus, the light-dependent reduction of Pchlide by POR is the first regulatory step in the overall greening processes in angiosperms.
The model plant Arabidopsis has three PORs (PORA, PORB and PORC) (Armstrong et al., 1995; Oosawa et al., 2000; Su et al., 2001). PORA (At5g54190) is expressed only in etiolated tissue in the dark, and is rapidly degraded in the light. PORB (At4g27440) is expressed concomitantly with PORA but remains stable in the light. PORC (At1g03630) is present in the leaves of light-grown plants. Thus, PORA and PORB are responsible for forming prolamellar bodies in etioplasts. Recently, PORA and PORB have been reported to have specific functions in the formation of prolamellar bodies in barley (Reinbothe et al., 1999). Specifically, PORA and PORB form supramolecular light-harvesting components designated light-harvesting POR Pchlide (LHPP) complexes in the prolamellar bodies; Pchlide b bound to PORA functions as a sensitizer and transfers light energy to Pchlide a bound to PORB for photoreduction of Pchlide a by PORB. The LHPP complexes have been speculated to serve as the central determinant of the prolamellar bodies in angiosperms (Reinbothe et al., 1999). However, since conflicting results about the accumulation of Pchlide b (Scheumann et al., 1999) as well as a species-dependent copy number of POR homologs (Oosawa et al., 2000; Su et al., 2001) have been reported, the LHPP model is still the subject of some controversy (Armstrong et al., 2000; Reinbothe et al., 2003). Characterization of PORs from organisms that do not synthesize Chl b, such as those of some cyanobacteria, seems a useful approach for gaining further insights into the LHPP model.
It is generally accepted that the POR protein was introduced into plant cells during the primary symbiosis of an ancestral cyanobacterium (Suzuki and Bauer, 1995; Masuda and Takamiya, 2004); among prokaryotic phototrophs, POR orthologs have been found only in cyanobacteria, suggesting that POR function was established before the initiation of eukaryotic photosynthesis. However, because cyanobacteria use another distinct enzyme, the so-called light-independent POR catalyzing the reduction of Pchlide in the dark (Fujita, 1996; Armstrong, 1998), prolamellar bodies cannot be formed in cyanobacteria. Thus, it is unknown whether the capability of POR to form prolamellar bodies with its substrates is inherent to the protein, irrespective of the species of photosynthetic organism.
In this study, we performed complementation analysis of an Arabidopsis PORA knockdown mutant with cyanobacterial PORs. We used PORs from two cyanobacteria. One was from Gloeobacter violaceus PCC7421, which has been suggested to retain the ancestral properties of cyanobacteria, since on a phylogenetic tree based on the 16S rRNA sequences, it branches off at the earliest stage within the cyanobacterial linage (Nelissen et al., 1995). The other POR was from Synechocystis sp. PCC6803, which has been extensively characterized in vivo and in vitro (He et al., 1998; Heyes et al., 2000, 2002, 2003a,b, 2006). Our results indicate that the cyanobacterial PORs can compensate for the PORA knockdown mutation of Arabidopsis. The complementing plants were also used to address the functional implications of POR for chloroplast biogenesis. Evolutionary aspects of POR are also discussed.
A candidate for an Arabidopsis PORA knockdown mutant, designated porA-1, was obtained from the Salk Institute Genomic Analysis Laboratory collection (SALK_036137); it carried a T-DNA insertion in the PORA promoter region (Figure 1a). Sequencing analysis indicated that the T-DNA was inserted 291 bp upstream from the start codon (data not shown). Transcript analysis revealed that PORA mRNA in porA-1 was decreased to ∼60% of that in the wild-type Columbia line (Col) (Figure 1b). Western blot analysis indicated that PORA plus PORB levels in etiolated seedlings of porA-1 were ∼70% of that of wild type (Figure 1c). The size of prolamellar bodies in porA-1 was substantially smaller than that of wild type as evidenced by electron micrographs (Figures 2 and S1). The porA-1 line did not show any obvious phenotype under normal growth conditions (data not shown) except for a delay in the greening of etiolated seedlings. Specifically, when dark-grown seedlings were illuminated by white light for 20 h, Chl accumulation in porA-1 was ∼60% of that in wild type (Figure 3).
We next investigated whether the observed porA-1 phenotype could be complemented by heterologous expression of cyanobacterial PORs. Figure 4(a) shows a schematic description of the DNA constructs used to produce potentially complementing plants. The chloroplast transit peptide of PORA and a FLAG-tag were fused with both G. violaceus and Synechocystis PORs at the N-terminus and C-terminus, respectively. As a control, the FLAG-tag was also fused with Arabidopsis PORA at the C-terminus. These recombinant proteins could be expressed under the control of the cauliflower mosaic virus 35S promoter. These constructs were separately introduced into the porA-1 mutant via Agrobacterium tumefaciens-mediated transformation, and the plants obtained expressing PORA, G. violaceus POR and Synechocystis POR were designated APO, GPO and SPO, respectively. We selected two independent SPO lines, SPO-1 and SPO-2, showing different expression levels of the recombinant protein (Figure 4b). Western blot analysis with an anti-FLAG antibody indicated that all recombinant proteins were successfully expressed, although expression levels were different. The expression levels of FLAG-tagged POR in GPO, SPO-1 and SPO-2 were estimated to be ∼80, ∼200 and ∼400% of that in APO, respectively (Figure 4b). The estimated molecular weight of each protein was in good agreement with those calculated from the deduced amino acid sequences excluding the transit peptide, suggesting that the fusion proteins were targeted into plastids. The total PORA transcripts in APO were about seven-fold higher, and the level of PORA plus PORB was about twofold higher than in wild type (Figure 1b,c), indicating that APO is actually in a PORA over-expressing line.
We examined several properties of the complemented lines. The first was prolamellar body formation. As shown in Figures 2 and S1, the size of the prolamellar bodies in APO was similar to those observed in the wild type, indicating that exogenous expression of FLAG-tagged PORA restored the function lost by the porA-1 knockdown mutation. More importantly, GPO and SPO-1 contained prolamellar bodies that were consistently larger than those found in the porA-1 line, indicating that over-expression of cyanobacterial PORs led to the formation of prolamellar bodies similar to those observed in wild type (Figures 2 and S1). The prolamellar bodies found in SPO-1 were substantially larger than those in APO and GPO. In fact, the expression level of FLAG-tagged POR in SPO-1 was much higher than its expression in APO or GPO (Figure 4b), supporting the idea that the cyanobacterial PORs contributed to the formation of prolamellar bodies in porA-1.
We next tested the greening of etiolated seedlings of these complementing lines. As shown in Figure 3, etiolated seedlings of APO accumulated almost the same amount of Chl upon light illumination as that observed in the wild type. This suggests that the observed deficiency of Chl accumulation in porA-1 was due to a low expression level of PORA. On the other hand, the degree of complementation by heterologous expression of cyanobacterial PORs differed in each line. Specifically, the GPO line had a restored Chl accumulation (Figure 3). However, one of the Synechocystis POR expression lines (SPO-1) had accumulated less Chl than that in the wild type, despite the elevated expression of FLAG-tagged POR (Figure 4b). These results indicate that G. violaceus POR fully complemented the Chl accumulation deficiency caused by the porA-1 knockdown mutation; however, Synechocystis POR partially complemented it. The difference in accumulation between the plants expressing the two different cyanobacterial genes, however, was not significant. It is notable that another Synechocystis POR expressing line (SPO-2) showing a much higher POR expression level (Figure 4b) exhibited accumulation of Chl as in the wild type, indicating that Synechocystis POR could complement the deficiency in Chl accumulation when it was highly expressed.
The results obtained above suggested that there are differences in the enzymatic properties between the two cyanobacterial PORs. Because the substrate affinity of PORs seems to be important for the formation of prolamellar bodies and chloroplast biogenesis (Reinbothe et al., 1999), we next characterized the in vitro enzymatic properties of the cyanobacterial PORs as well as Arabidopsis PORA and PORB by determining their kinetic parameters. The purified Arabidopsis PORA and PORB expressed in Escherichia coli possessed a light-dependent catalytic activity of Pchlide reduction (Figure S2). The calculated Km values (±SE) of PORA for Pchlide and NADPH were 0.15 ± 0.10 μm and 59.5 ± 3.2 μm, respectively (Figure S3a,e); those of PORB were 0.39 ± 0.15 μm and 26.8 ± 5.0 μm, respectively (Figure S3b,f). It may be notable that PORB activity, but not PORA activity, tended to be inhibited at higher Pchlide concentration with 50% inhibition at ∼15 μm Pchlide (Figure S3b). A similar Pchlide-dependent substrate inhibition was reported for oat POR purified from etioplasts (Klement et al., 1999), although it is not clear at present how this is achieved. The calculated Km values (±SE) of His-tagged G. violaceus POR for Pchlide and NADPH were 0.50 ± 0.15 μm and 19.6 ± 3.7 μm, respectively (Figure S3c,g); those of Synechocystis POR were 10.83 ± 3.49 μm and 7.5 ± 1.5 μm, respectively (Figure S3d,h). A similar Km value for Pchlide was also observed for Strep-tagged G. violaceus POR (∼0.72 μm; data not shown). Thus, G. violaceus POR exhibited ∼20-fold higher affinity to Pchlide than Synechocystis POR, although they showed similar affinity to NADPH; a relatively lower affinity to Pchlide in Synechocystis POR has been previously reported (Heyes et al., 2000). Table 1 summarizes the kinetic parameters of PORs from various sources. The Km value for Pchlide of G. violaceus POR (0.50 μm) was similar to those of plant PORs (∼0.15 to 0.47 μm). On the other hand, the Km values for NADPH did not differ much between these PORs.
Table 1. Kinetic parameters of protochlorophyllide oxidoreductase (PORs)
To gain more information relating to the structure of the prolamellar bodies, we next obtained a low-temperature fluorescence spectrum to determine the amount of both photoactive and non-photoactive Pchlide in etiolated seedlings of each line in situ. Typical fluorescence emission spectra are shown in Figure S4. Previous spectroscopic studies have established that there are two in vivo forms of Pchlide (Franck and Strzalka, 1992; Sundqvist and Dahlin, 1997). One pigment form, Pchlide-F655, which is designated by its fluorescence maximum at low temperature (−196°C), is photoactive and directly photoreducible by a flashlight. The other form, designated Pchlide-F632, is a non-photoactive Pchlide, which, because of a wide bandwidth, is probably a mixture of several distinct Pchlide forms. Pchlide-F655 is considered to arise from aggregation of Pchlide:NADPH:POR ternary complexes within prolamellar bodies. In fact, PORA or PORB over-expressing lines, containing larger prolamellar bodies than that in the wild type, show a higher Pchlide-F655/Pchlide-F632 ratio than observed in the wild type (Sperling et al., 1998). As shown in Table 2, the relative ratio of Pchlide-F655/Pchide-F632 was higher in APO and lower in porA-1 than in the wild type, although the levels of total Pchlide were almost identical in the lines. These results were consistent with a sum of expression levels of the PORA plus PORB in each line (Figure 1c). On the other hand, the Pchlide-F655/Pchlide-F632 ratio in GPO and SPO-1 lines was not much different from that of the parent line porA-1. The total Pchlide levels in the complementing lines were not much different from those of Col and porA-1 (Table 2), indicating that photoactive Pchlide-F655 levels in GPO and SPO-1 remained low, as in porA-1. These results indicate that cyanobacterial PORs expressed in etioplasts did not contribute to the formation of photoactive Pchlide-F655.
Table 2. Ratio of photoactive protochlorophyllide (Pchlide) F-655 and non-photoactive Pchlide F-632 molecules, and total Pchlide content in etiolated seedlings
Total Pchlide (nanogram per eight seedlings)
Photoactive (F655) and non-photoactive (F632) Pchlide molecules were estimated by in situ low-temperature fluorescence emission intensity (440 nm excitation) at 655 and 632 nm, respectively, of dark-grown cotyledons (Figure S4). Relative values were calculated with the wild type set to 1.00. The values of total Pchlide are the means ± SD of at least three replications.
8.27 ± 0.63
8.72 ± 0.21
9.36 ± 0.49
8.25 ± 0.58
8.91 ± 0.58
After giving flashes, fluorescence from photoactive Pchlide-F655 disappeared and fluorescence from chlorophyllide appeared at 690 nm (Figure S4, broken lines). Thus, the amounts of converted chlorophyllide were proportional to the amounts of Pchlide-F655 in each line. No residual fluorescence at 655 nm was observed in any of the samples investigated in this study. The relative fluorescence ratio of Pchlide-F655/Pchlide-F632 represented the in vivo amount of photoactive Pchlide-F655.
We have shown that heterologous expression of cyanobacterial PORs can compensate for the reduced size of the prolamellar body (Figures 2 and S1) and deficient Chl accumulation (Figure 3) of the Arabidopsis PORA knockdown mutant. The results show that prokaryotic PORs function in plastids and retain the ability to form prolamellar bodies. It is notable that the size of the observed prolamellar bodies in the complementing lines correlated well with the expression levels of the exogenous PORs. Recombinant plants expressing FLAG-tagged PORA (APO) or G. violaceus POR (GPO) showed similar levels of POR expression (Figure 4b), and contained moderately sized prolamellar bodies; on the other hand, the SPO-1 line, showing higher expression of FLAG-tagged POR than APO and GPO, contained larger prolamellar bodies (Figures 2 and S1). These results indicate that cyanobacterial PORs retain the ability to form prolamellar bodies as efficiently as native PORA/PORB.
Several lines of evidence have suggested that Arabidopsis PORA and PORB are in an aggregated form with NADPH and the photoactive Pchlide-F655 in prolamellar bodies (Griffiths, 1978; Oliver and Griffiths, 1982; Lebedev et al., 1995; McEwen et al., 1996). The over-expression of FLAG-tagged PORA in transformed porA-1 (APO) resulted in the restoration of prolamellar body size (Figures 2 and S1) as well as Pchlide-F655 content (Figure S4 and Table 2), supporting previous data. However, the GPO and SPO-1 lines retained a low level of Pchlide-F655 (Figure S4 and Table 2), indicating that the prolamellar bodies formed with the cyanobacterial PORs did not accumulate an amount of photoactive Pchlide additional to that in the porA-1 mutant. Apparent sizes of prolamellar bodies were not necessarily correlated with the amounts of photoactive Pchlide. These results suggest that aggregation of POR in prolamellar bodies could partly be achieved with non-photoactive Pchlide-F632; this was the case at least in the complementing lines.
The reason why the cyanobacterial PORs could not assemble Pchlide-F655 is open to speculation. One possibility is that post-translational modifications of POR apo-protein are necessary for assembly of Pchlide-F655, and such a reaction could not be achieved by cyanobacterial enzymes even when localized in etioplasts. If this is the case, cytokinins may be involved in the reaction since they were previously shown to have important roles in regulating the formation of prolamellar bodies irrespective of the levels of POR protein (Seyedi et al., 1999, 2001). Another possibility is that the LHPP model, based on barley etioplasts (Reinbothe et al., 1999), is applicable to Arabidopsis etioplasts; PORA binds Pchlide b in LHPP complexes in wild-type plants, but the cyanobacterial PORs could not bind Pchlide b to form LHPP complexes in the complementing lines. If this is the case, the formation of LHPP complexes, but not the formation of prolamellar bodies, is coupled with assembly of photoactive Pchlide-F655. However, fluorescence spectra did not indicate the presence of Pchlide b in any samples used in this study. In addition, Arabidopsis PORA has a chlorophyllide synthesis activity with a higher affinity for Pchlide than that of PORB (Table 1). Taking these results together, the process of POR expression and the assembly of prolamellar bodies might be separated, or an additional factor(s) is required for formation of prolamellar bodies. This has led to a new view about the formation of prolamellar bodies in etioplasts.
The kinetic parameters of G. violaceus POR resemble those of plant PORs rather than that of Synechocystis POR (Table 1). A previous phylogenetic analysis showed that G. violaceus POR is closely related to plant-type PORs (Masuda and Takamiya, 2004), suggesting that an ancestral plant POR might already have had a high affinity for Pchlide. The GPO line showed higher accumulation of Chl than in SPO-1 upon illumination (Figure 3), although both lines retained low levels of Pchlide-F655 (Table 2). These results indicate that the high affinity for Pchlide in the G. violaceus POR (Km = ∼0.5 μm) is important for efficient accumulation of Chl during the etioplast–chloroplast transition, irrespective of the level of the photoactive Pchlide-F655 in prolamellar bodies. Assembly of photoactive Pchlide-F655 by plant-type POR may be important to protect host cells against Pchlide-induced photo-oxidative damage, since Pchlide-F632 is a potent photosensitizer (Runge et al., 1996). Clearly, further characterization of the cyanobacterial PORs expressed in planta will be useful for gaining insight into the function and evolution of POR as well as the mechanism of formation of prolamellar bodies, which can now be addressed using the transgenic plants obtained in this study.
Plant materials and growth conditions
All plants used in this study were the Columbia ecotype of Arabidopsis thaliana. Surface-sterilized seeds were plated on 0.8% (w/v) agar solidified Murashige–Skoog medium. Plates were placed in the dark at 4°C for 3 days prior to receiving 6 h white light illumination to synchronize germination. Then, plants were grown at 22°C with white light (25 μmol photons m−2 sec−1) or in the dark.
Construction of transgenic plants
The primers used are shown in Table S1. A cDNA clone for Arabidopsis PORA was kindly provided by the RIKEN Bioresource Center, Ibaraki, Japan. The DNA constructs for recombinant genes composed of PORA transit peptide and cyanobacterial PORs were achieved by a double PCR strategy (Wang and Malcolm, 1999). A coding region for the transit peptide of PORA was amplified from the cDNA clone with a forward primer AttB-Arabi-vivo-F and a reverse primer Arabi-Glr2486-Rv (for G. violaceus POR) or Arabi-Slr0506-Rv (for Synechocystis POR). Coding regions for cyanobacterial PORs were also amplified from genomic DNA with two combinations of primers; they are a forward primer Arabi-Glr2486-Fw and a reverse primer Glr2486-R-FLAG-Sm (for G. violaceus POR), and a forward primer Arabi-Slr0506-Fw and a reverse primer Slr0506-R-FLAG-Sm (for Synechocystis POR). The first PCR fragments for PORA transit peptide and each POR coding region were mixed and then the second PCR was achieved with AttB-Arabi-vivo-F and AttB-FLAG-R primers. A coding region for Arabidopsis PORA was amplified with AttB-Arabi-vivo-F and Arabi-vivo-R-FLAG. The three final PCR fragments obtained were separately cloned into pDONR/ZeO vector by use of the Gateway system (Invitrogen, http://www.invitrogen.com/). After checking the correct sequences of the inserted DNA, the fragments were cloned into pGWB2 vector (kindly provided by Dr T. Nakagawa, Shimane University, Matsue. Japan). The resulting constructs (Figure 4a) were introduced into Arabidopsis porA-1 via an Agrobacterium tumefaciens-mediated transformation method.
Western blot analysis
Plants were homogenized in Buffer A [20 mm 2-amino-2-(hydroxymethyl)-1,3-propanediol (TRIS)–HCl, pH 7.5, and 10 mm NaCl], the homogenate was centrifuged at 5000 g for 10 min, and the supernatant was used for the analysis. An equal amount of total protein (∼5 μg) was separated on a SDS–PAGE gel and electroblotted onto polyvinylidene fluoride (PVDF) membrane. The bands of immunoreactive against anti-wheat POR antibody (kindly provided by Dr H. Aronsson, Leicester University, Leicester, UK) (Greby et al., 1989) and anti-FLAG M2 antibody (Sigma-Aldrich, http://www.sigmaaldrich.com/) were detected using the ECL system (GE Healthcare, http://www.gehealthcare.com/).
Determination of Pchlide and Chl accumulation in seedlings
For determination of Chl accumulation, seedlings grown for 4 days in the dark and 20 h in white light were harvested and homogenized in 80% acetone, followed by centrifugation at 5000 g to remove debris. The Chl content was determined as described (Arnon, 1949) with at least four replications. For the determination of total Pchlide, seedlings grown for 4 days in darkness were homogenized in 80% acetone, followed by centrifugation to remove debris. The Pchlide content was determined as described (Masuda et al., 2003) with at least three replications.
Total RNA was extracted from 4-day-old dark-grown seedlings using the SV Total RNA Isolation System (Promega, http://www.promega.com/). Using the RNA as a template, cDNA was synthesized by TaKaRa RNA PCR Kit Ver.3.0 (TaKaRa, http://www.takara-bio.com/). Quantitative real-time PCR using SYBR Premix Ex Taq (TaKaRa) was employed to determine transcript levels using first-strand cDNA as a template on a Thermal Cycler Dice Real Time System (TaKaRa). The primer pairs PORA-F and PORA-R, and ACT-F and ACT-R (Table S1) were used for specific amplification of PORA and actin8 mRNAs, respectively. For normalization across samples, the actin8 gene was used as an internal standard.
Determination of the kinetic parameters of PORs
Arabidopsis PORA and PORB were expressed in E. coli using a T7 RNA polymerase-based over-expression system with a self-cleavable intein/chitin tag (New England Biolabs, http://www.neb.com/). For this construction, cDNA for putative mature forms of PORA and PORB, which lack a 69- (for PORA) or 66-amino-acid (for PORB) N-terminal predicted transit peptide (Figure 4a), were amplified by PCR using two combinations of primer pairs, PORA-F-NdeI and PORA-R-EcoRI, and PORB-F-NdeI and PORB-R-EcoRI (Table S1), respectively. The amplified fragments were separately cloned into NdeI–EcoRI-cut pTYB12 (New England Biolabs). After checking the correct sequences of the inserted DNA fragments, these plasmids were used to transform E. coli strain BL21(DE3) (Merck, http://www.merck.com/), and recombinant PORA and PORB were over-expressed by induction with 1 mm isopropyl-β-d-thiogalactopyranoside at 16°C for 16 h in Luria broth medium. Recombinant PORA and PORB were purified from the harvested cells with a chitin bead column (New England Biolabs) as described previously (Masuda et al., 2004). Specifically, harvested cells suspended in medium containing 0.5 m NaCl, 0.2 mm EDTA and 20 mm TRIS–HCl (pH 8.0) were ruptured, and the soluble fraction was collected by centrifugation at 28,000 g for 30 min. The soluble fraction was applied to a chitin affinity column (New England Biolabs), and the column was washed with ∼20-fold column volumes of the cell-suspended buffer. The chitin/intein tag was removed by incubation with the same buffer containing 100 mm dithiothreitol at 4°C for 16 h. The purified PORA and PORB contain only three amino acids (Ala, Gly and Met) at the N-terminus with no tag. The isolated proteins were more than 95% pure as based on Coomassie blue staining of an SDS–PAGE gel (Figure S2a). Purification of G. violaceus and Synechocystis PORs were as described previously (Ikeda et al., 2008). Briefly, the cyanobacterial PORs were over-expressed in E. coli as N-terminal His-tagged proteins, and purified using His-Bind resin (Merck).
Purification of Pchlide and activity measurements were performed as described (Ikeda et al., 2008). Briefly, Pchlide was isolated from 10-day dark-grown barley with acetone extraction followed by column chromatography. For the determination of enzymatic activities, purified POR was mixed with Pchlide and NADPH in assay buffer containing 0.08% Triton X-100, 0.08%β-mercaptoethanol and 60 mm TRIS–HCl (pH 7.5) at room temperature (∼23°C). Spectra were recorded using a Hitachi U-0080D spectrophotometer (http://www.hitachi.com/). Measurements were recorded pre- and post-illumination by a halogen lamp. The kinetic parameters for catalysis by PORs were determined by monitoring the initial rate of chlorophyllide production (change in absorbance at 670 nm) using the following equation:
where ν is the initial rate, [S] is the substrate concentration, Vmax is the initial rate achieved as [S] approaches infinity and Km is the apparent value of [S] giving Vmax/2. Data were fitted and apparent Km values (±SE) were calculated using the sigmaplot program (Systat Software, http://www.systat.com/). For more details on the kinetic analysis see the legend to Figure S3.
Transmission electron microscopy
For electron microscopy, etiolated seedlings were collected under safelight and fixed on ice with 2.5% (w/v) glutaraldehyde in 0.1 m cacodylate buffer (pH 7.2) overnight in the dark. The samples were washed several times with the same buffer, and post-fixed with 1% osmium tetroxide (w/v) for 2 h on ice. Then the samples were washed with the same buffer, dehydrated by graded acetone series and embedded in Spurr’s low-viscosity resin (Spurr, 1969). Thin sections were stained with uranyl acetate and lead citrate, and observed under a JEOL 1200 EX transmission electron microscope (http://www.jeol.com/).
In situ low-temperature fluorescence spectroscopy
During sample preparation, all operations were done in complete darkness. Seedlings grown in the dark for 4 days were harvested and immobilized on an acrylic plate. The acrylic plate was set in a custom-made acrylic cuvette and 15% (w/v) polyethylene glycol (with an average molecular weight of 3350; Sigma-Aldrich) solution was added to the cuvette to obtain homogeneous ice. The cuvette was immersed into liquid nitrogen, and the low-temperature fluorescence spectrum was measured with a Hitachi 850 spectrofluorometer with a homemade Dewar bottle (Mimuro et al., 1999). The spectral sensitivity of the apparatus was corrected using a sub-standard lamp (Hitachi) of known radiation profile. The excitation wavelength was 440 nm for all samples, and a cutoff filter (R-60, Toshiba, http://www.toshiba.co.jp/worldwide/) was used to protect scattering of the excitation light. After the measurement of dark-grown seedlings, the cuvette was warmed up to −20°C, then the sample was subjected to flash illumination (half duration = 4 μs, 1 Hz, 20 flashes) and immediately immersed into liquid nitrogen. The measurement after illumination was conducted under the same conditions as that before illumination. Baselines and scattering from the samples were subtracted by using signals from a solution that did not contain any samples.
We thank Dr H. Shimada (Hiroshima University, Higashi-Hiroshima, Japan) for helpful suggestions during the study, and Dr T. Nakagawa, (Shimane University, Matsue, Japan) and Dr H. Aronsson (Leicester University, Leicester, UK) for providing experimental materials. We also thank the RIKEN Bioresource Center, Ibaraki, Japan and the Arabidopsis Biological Resource Center, Ohio State University, Columbus, OH, USA for providing a cDNA clone and a T-DNA insertion line, respectively. This work was supported in part by the Research Foundation for Opto-Science and Technology to SM, by the Toyoaki Scholarship Foundation to YF, by a Grant-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology (MEXT) of Japan to SM, TM, YF, and HO, and by a Grant-in-Aid for the Creative Research from the Japanese Society for Promotion of Science (grant no. 17GS0314) to MM. JN is also grateful for a Fellowship from the JSPS for Japanese Junior Scientists (grant no. 19010733).