Biosynthesis of astaxanthin in tobacco leaves by transplastomic engineering

Authors

  • Tomohisa Hasunuma,

    1. Research Institute of Innovative Technology for the Earth (RITE), 9-2 Kizugawadai, Kizugawa-shi, Kyoto 619-0292, Japan
    Search for more papers by this author
    • Present address: Department of Chemical Science and Engineering, Faculty of Engineering, Kobe University, 1-1 Rokkodai-cho, Nada-ku, Kobe, 657-8501, Japan.

  • Shin-Ichi Miyazawa,

    1. Research Institute of Innovative Technology for the Earth (RITE), 9-2 Kizugawadai, Kizugawa-shi, Kyoto 619-0292, Japan
    Search for more papers by this author
  • Satomi Yoshimura,

    1. Research Institute of Innovative Technology for the Earth (RITE), 9-2 Kizugawadai, Kizugawa-shi, Kyoto 619-0292, Japan
    Search for more papers by this author
  • Yuki Shinzaki,

    1. Research Institute of Innovative Technology for the Earth (RITE), 9-2 Kizugawadai, Kizugawa-shi, Kyoto 619-0292, Japan
    Search for more papers by this author
  • Ken-Ichi Tomizawa,

    1. Research Institute of Innovative Technology for the Earth (RITE), 9-2 Kizugawadai, Kizugawa-shi, Kyoto 619-0292, Japan
    Search for more papers by this author
  • Kazutoshi Shindo,

    1. Department of Food and Nutrition, Japan Women’s University, 2-8-1 Mejirodai, Bunkyo-ku, Tokyo, 112-8681, Japan
    Search for more papers by this author
  • Seon-Kang Choi,

    1. Marine Biotechnology Institute, Heita, Kamaishi-shi, Iwate 026-0001, Japan
    Search for more papers by this author
  • Norihiko Misawa,

    1. Marine Biotechnology Institute, Heita, Kamaishi-shi, Iwate 026-0001, Japan
    Search for more papers by this author
  • Chikahiro Miyake

    Corresponding author
    1. Research Institute of Innovative Technology for the Earth (RITE), 9-2 Kizugawadai, Kizugawa-shi, Kyoto 619-0292, Japan
      *(fax +81-78-803-5851; email: cmiyake@hawk.kobe-u.ac.jp).
    Search for more papers by this author
    • Present address: Department of Biological and Environmental Science, Faculty of Agriculture, Kobe University, 1-1 Rokkodai-cho, Nada-ku, Kobe, 657-8501, Japan.


*(fax +81-78-803-5851; email: cmiyake@hawk.kobe-u.ac.jp).

Summary

The natural pigment astaxanthin has attracted much attention because of its beneficial effects on human health, despite its expensive market price. In order to produce astaxanthin, transgenic plants have so far been generated through conventional genetic engineering of Agrobacterium-mediated gene transfer. The results of trials have revealed that the method is far from practicable because of low yields, i.e. instead of astaxanthin, large quantities of the astaxanthin intermediates, including ketocarotenoids, accumulated in the transgenic plants. In the present study, we have overcome this problem, and have succeeded in producing more than 0.5% (dry weight) astaxanthin (more than 70% of total caroteniods) in tobacco leaves, which turns their green color to reddish brown, by expressing both genes encoding CrtW (β-carotene ketolase) and CrtZ (β-carotene hydroxylase) from a marine bacterium Brevundimonas sp., strain SD212, in the chloroplasts. Moreover, the total carotenoid content in the transplastomic tobacco plants was 2.1-fold higher than that of wild-type tobacco. The tobacco transformants also synthesized a novel carotenoid 4-ketoantheraxanthin. There was no significant difference in the size of the aerial part of the plant between the transformants and wild-type plants at the final stage of their growth. The photosynthesis rate of the transformants was also found to be similar to that of wild-type plants under ambient CO2 concentrations of 1500 μmol photons m−2 s−1 light intensity.

Introduction

Carotenoids are the most widespread group of pigments found in nature. Over 750 different carotenoids are known to exist in bacteria, fungi, plants and animals (Britton et al., 2004). In general, carotenoids have high antioxidative activities (Stahl and Sies, 2003). They are fundamental components in our diet, and are considered to play an important role in human health. One of the most widely known carotenoids is β-carotene, which serves as a dietary precursor of vitamin A. Carotenoids are synthesized in all photosynthetic organisms, and in some bacteria and fungi. As animals are unable to synthesize carotenoids de novo, they must obtain them by dietary means. Industrial uses of carotenoids include nutritious supplements, food additives and colorants in cosmetics. Among the commercially available carotenoids, one of the most valuable is astaxanthin [(3S,3′S)–3,3′-hydroxy-β,β-carotene-4,4′-dione]. It is one of the most commonly found pigments in marine animal tissues. Astaxanthin provides the characteristic red or pinky color to salmon, trout and shrimp. This pigment has been industrially exploited as a feed dye, particularly as a feed supplement in poultry farming and aquaculture. Sales of astaxanthin as a pigmentation source in salmon aquaculture in the United States amount to about $200 million per year (Lorenz and Cysewski, 2000). The diverse biological functions of astaxanthin include involvement in the antioxidation of low-density lipoprotein (Iwamoto et al., 2000), anticancer activities (Tanaka et al., 1994), singlet oxygen-quenching activity (Tatsuzawa et al., 2000) and enhancement of immune responses (Jyonouchi et al., 1995). Therefore, its use in the pharmaceutical, food and feed industries is expected to increase dramatically in the near future if we succeed in the development of genetic engineering tools for the mass production of astaxanthin at a reasonable cost.

Currently, the majority of astaxanthin produced for commercial use is chemically synthesized. However, synthetic astaxanthin contains the stereoisomer by-products (3S,3′R) and (3R,3′R), in addition to the naturally occurring (3S,3S′). The presence of the by-products may have an inhibitory effect on the bioactivity of natural astaxanthin. Additionally, chemically synthesized astaxanthin may be contaminated with other reaction by-products or intermediates. Thus, its commercial applications are usually restricted to its use as a feed supplement in aquaculture. Astaxanthin can also be produced biologically, as several microorganisms are able to produce it at high levels. For example, the green alga Haematococcus pluvialis, which produces astaxanthin at levels representing 4–5% of its dry weight, is used for the commercial production of this pigment as a functional food supplement for human consumption. However, this organism requires high light intensities, which increase the production costs, and its slow growth rate in mild culture conditions increases the risks of contamination. Another astaxanthin-producing microorganism is the yeast Xanthophyllomyces dendrorhous, which can produce up to 0.5% of its dry weight of astaxanthin, the configuration of which is, exceptionally, 3R,3′R (Johnson and An, 1991).

Metabolic engineering in higher plants by using astaxanthin biosynthesis genes is potentially one of the most powerful tools to bioproduce astaxanthin, as plants have the ability to accumulate carotenoids in the thylakoid membranes and in the lipid globules within the plastid at very high concentrations (Mann et al., 2000). Astaxanthin can be synthesized with the introduction of keto- and hydroxyl-moieties at the 4,4′ and 3,3′ positions of the β-ionone rings of β-carotene. This can be achieved by two enzymes, a β-carotene ketolase (4,4′-oxygenase; CrtW or BKT, and CrtO) and a β-carotene hydroxylase (3,3′-hydroxylase; CrtZ or BHY), by way of the eight intermediates of ketolated and/or hydroxylated carotenoids (Figure 1). Plants exhibitβ-carotene hydroxylase activity, but apart from a very few exceptions, such as Adonis flowers (Cunningham and Gantt, 2005), they are devoid of β-carotene ketolase. Therefore, the functional expression of a heterologous β-carotene ketolase gene in plants is required for the production of astaxanthin. In a first attempt at expressing a β-carotene ketolase gene in plants, the H. pluvialis bkt gene that was linked to a transit peptide coding sequence under the tomato phytoene desaturase (PDS) promoter was introduced into the tobacco nuclear genome by Agrobacterium-mediated gene transfer (Mann et al., 2000). The transgenic tobacco (Nicotiana tabacum) plants accumulated astaxanthin in the nectary tissue, as indicated by a color change from yellow to red, whereas only trace levels of ketocarotenoids were found in the leaves (Mann et al., 2000). As the nectary is a small tissue encircling the base of the ovary in the flowers, the production level of astaxanthin per plant was extremely low. Later, Ralley et al. (2004) integrated both the crtW and crtZ genes from a marine bacterium Paracoccus sp., strain N81106, into the tobacco nuclear genome, under the control of the CaMV 35S promoter, to enhance the metabolic flux from β-carotene to astaxanthin. This resulted in the accumulation of ketocarotenoids in the leaves as well as in the nectaries. However, ketocarotenoid content in the leaves was less than 5% of the total carotenoids. Moreover, the most abundant ketocarotenoid in the leaves was echinenone, not astaxanthin. Gerjets et al. (2007) recently reported that although cyanobacterial crtO and Erwinia uredovora crtZ genes were co-expressed under the control of the CaMV 35S promoter in transgenic N. tabacum, the astaxanthin content remained below the detectable level. These results show that efficient conversion of β-carotene to astaxanthin by way of the ketocarotenoid intermediates appears to be very important for the high accumulation of astaxanthin in plants.

Figure 1.

 Schematic representation of a potential carotenoid production pathway, resulting from Brevundimonas crtZ and crtW insertions into the tobacco plastid genome. Carotenoids found in transplastomic tobacco plants are highlighted in bold typeface, and eight intermediates of ketolated and/or hydroxylated carotenoids in the synthesis of astaxanthin are underlined. Enzymes are indicated by their gene assignment symbols: CrtL-e, ε-ring hydroxylase; CrtW, β-carotene ketolase; CrtZ, β-carotene hydroxylase; LCY-b, lycopene β-cyclase; LCY-e, lycopene ε-cyclase; NXS, neoxanthin synthase; PSY, phytoene synthase; VDE, violaxanthin de-epoxidase; ZEP, zeaxanthin epoxidase.

β-Carotene ketolase genes have been isolated from the marine bacteria, cyanobacteria and green algae that produce ketocarotenoids including astaxanthin. Cyanobacterial CrtW and bacterial CrtO enzymes were shown to poorly catalyze the conversion of zeaxanthin to astaxanthin via adonixanthin, because it is difficult to accept a 3-hydroxy-β-ionone ring as a substrate (Choi et al., 2007; Tsuchiya et al., 2005). Likewise, CrtW enzymes from the marine bacteria Paracoccus sp. N81106 and Paracoccus sp. PC1 have also low conversion efficiencies of adonixanthin to astaxanthin (Choi et al., 2005; Fraser et al., 1997;Misawa et al., 1995). Recently, novel crtW and crtZ genes were isolated from a marine bacterium Brevundimonas sp., strain SD212 (Nishida et al., 2005). Escherichia coli complementation studies showed that CrtW from Brevundimonas sp. SD212 carried out the efficient conversion of adonixanthin to astaxanthin in vivo, and resulted in the accumulation of higher levels of astaxanthin in E. coli (Choi et al., 2005). These studies also demonstrated that CrtZ from the same strain has the most efficient conversion activity in the pathway from β-carotene to astaxanthin among all CrtZ proteins examined, including those of Paracoccus sp. N81106, Paracoccus sp. PC1, Pantoea ananatis (formerly known as Erwinia uredovora), Flavobacterium sp. P99-3 and Thermus thermophilus HB27 P450 (CYP175A1) (Choi et al., 2006). Therefore, the expression of both the crtW and crtZ genes from Brevundimonas sp. SD212 in plants is expected to achieve the most efficient synthesis of astaxanthin.

Plastids of higher plants have their own genome that can be genetically engineered by the insertion of foreign genes. Plastid transformation offers several advantages over nuclear transformation, although the technology is presently applicable to only a few crops (reviewed in Verma and Daniell, 2007; Maliga, 2004). The expression of a transgene in plastids eliminates the need for a transit peptide required for the import of proteins synthesized in the cytosol into plastids (Misawa et al., 1993). Moreover, following translocation across the plastid membranes, the transit peptide needs to be proteolytically removed to yield a mature functional protein. However, nuclear transformants can potentially accumulate unprocessed proteins that are not functional, and that can interfere with the expected results (Jayaraj et al., 2007). These problems can be circumvented by using plastid transformation. In addition, because the plastid genome is highly polyploid (500–10 000 copies of genome per cell) (Bendich, 1987), the expression level of a transgene is likely to be high because it is present in high copy number, and thus has the potential to confer higher levels of protein accumulation compared with nuclear transgene expression (Miyake et al., 2006; Viitannen et al., 2004). Thus, plastid transformation appears to be the method of choice in engineering metabolic pathways that are localized in the plastid, such as the carotenoid biosynthetic pathways. In the present study, we successfully generated transplastomic tobacco plants that expressed the genes encoding Brevundimonas sp. CrtW and CrtZ to produce high levels of astaxanthin in their leaves.

Results

Generation of transplastomic tobacco

The sequences of the crtZ and crtW genes from Brevundimonas sp. have a high GC content (70%) compared with tobacco chloroplast. As heterologous gene expression in plastids can be limited by differences in codon usage (Maliga, 2003), we used crtZ and crtW genes, the codon usage of which was changed to a dicotyledonous plant type, according to that of rape (Brassica napus). The plastid transformation vectors were designed to insert the foreign genes between rbcL and accD through homologous recombination (Figure 2a). Both crtZ and crtW were arranged as an operon under the control of a single tobacco-derived promoter, Prrn. Such an arrangement is possible because plastids transcribe polycistronic mRNA. Two kinds of vectors that contained different permutations of the crtZcrtW tandem, as well as vectors with the crtZ or crtW alone, were used for transformation. The correct integration of the transgenes into the plastid genomes was verified by Sourthern-blot analysis. When hybridized to the rbcL probe, an EcoRV digestion of the crtZ/crtW-cointegrated, crtW-integrated, crtZ-integrated and wild-type plastid genomes produced a 10.0-, 9.5-, 9.2- and 7.1-kb fragment, respectively (Figure 2b). Using gene-specific probes, both control and wild-type plants did not show any fragments. These results indicated that the transgenes were inserted into the expected region between rbcL and accD. The expression of the transgene mRNAs was assessed by real-time RT-PCR (Figure 2c). All transformants accumulated transgene-derived mRNA and, as expected, no expression was observed in wild-type plants.

Figure 2.

 Molecular characterization of transplastomic plants. (a) Schematic representation of the plastid genome region of wild-type plants and transformants that harbor the transformation vectors pLD200-ZW, pLD200-WZ, pLD200-W and pLD200-Z. The aadA gene confers spectinomycin resistance for the selection of transformed shoots. Prrn, 16S rRNA promoter; TpsbA, psbA terminator; Trps16, rps16 terminator. (b) Southern blot analysis of plastid transformants. Total DNA was digested with EcoRV and probed with rbcL, crtZ and crtW. ZW-2, ZW-9, W-6, Z-1 and WZ-4 represent five individual transgenic lines; WT, wild type. (c) Relative expression levels of crtW and crtZ in independent transgenic tobacco lines, as determined by real-time PCR. mRNA abundance was normalized to actin mRNA expression. The ratio of transgene: actin expression of the transformed line ZW-2 was arbitrarily set to 1. (d) Relative expression levels of isoprenoid biosynthetic genes (dxs, ipi, psy, pds and crtL-1) in wild-type, ZW-2 and ZW-9 plants. The ratio of transgene:actin expression of wild-type plants was arbitrarily set to 1. The data represent average values from the analysis of between six and eight different tobacco plants. Error bars indicate ±SEM.

Pigment analysis

A dramatic change of color was observed in the transplastomic plants (ZW-2) that expressed both Brevundimonas sp. SD212 crtZ and crtW (Figure 3a). The color of their leaves and stems was reddish brown, and the stigmas and corollas were pink, instead of their normal green color (Figure 3b). The measured concentrations of carotenoids in leaves of the transplastomic and wild-type plants are summarized in Table 1. In the transplastomic plants, where crtZ is upstream of crtW, ZW-9 and ZW-2, the major carotenoid was an unesterified form of astaxanthin (74 and 71% of the total carotenoids, respectively). ZW-9 plants accumulated 5.44 mg g−1 (dry weight; DW) of astaxanthin in their leaves. The crtW-expressing line (W-6) and the crtW/crtZ-coexpressing line (WZ-4), where crtW is upstream of crtZ, accumulated 1.88 and 3.29 mg g−1 (DW) astaxanthin, respectively. The transgenic Z-1 line that expressed only crtZ did not produce astaxanthin or any other ketocarotenoids. The astaxanthin-producing lines synthesized a number of other ketocarotenoids, such as fritschiellaxanthin (4-ketolutein), adonixanthin, adonirubin, 3′-hydroxyechinenone and canthaxanthin. Moreover, a novel carotenoid, 4-ketoantheraxanthin was identified in these plants. The structural determination is reported in Shindo et al., 2008. The ZW-2 and ZW-9 lines included carotenoids that are ubiquitously synthesized in wild-type plants, such as β-carotene, zeaxanthin and lutein, but, in very small quantities. Neoxanthin and antheraxanthin were not detected in these two transgenic lines. The violaxanthin content in Z-1 plants was 3.7-fold higher than in wild-type plants.

Figure 3.

 Color changes in the aerial part (a) and flowers (b) of transplastomic tobacco expressing Brevundimonas CrtZ and CrtW. Arrows indicate the corolla (yellow) and the stigma (blue).

Table 1.   Pigment analysis in leaves of wild-type and transplastomic tobacco plants
 Pigment (mg g−1 dry weight)
Wild-typeZW-2ZW-9W-6Z-1WZ-4
  1. The total carotenoid content was calculated as the sum of the content of each carotenoid. Pigment levels were analyzed by one-way anova (Sokal and Rohlf, 1995). A post hoc Tukey’s Honestly Significant Difference (HSD) test was carried out on the grouped means. Values are the averages from measurements of three or four different tobacco plants, ±SEM. Values followed by the same letter are not significantly different (> 0.05). DW, dry weight of tissues; n.d., not detected.

Total carotenoid3.51 ± 0.71a7.20 ± 0.20c7.38 ± 0.68c3.84 ± 0.15ab3.99 ± 0.29ab5.72 ± 0.47bc
Total ketocarotenoidn.d.7.05 ± 0.20ab7.29 ± 0.69b3.83 ± 0.15cn.d.5.70 ± 0.47a
β-Carotene1.39 ± 0.28a0.03 ± 0.02b0.01 ± 0.00bn.d.0.81 ± 0.16c0.01 ± 0.01b
Lutein1.11 ± 0.25a0.09 ± 0.02b0.07 ± 0.02b0.01 ± 0.00b0.96 ± 0.05a0.01 ± 0.00b
Zeaxanthin0.02 ± 0.02a0.01 ± 0.00a0.01 ± 0.01an.d.0.01 ± 0.02an.d.
Antheraxanthin0.03 ± 0.01n.d.n.d.n.d.n.d.n.d.
Violaxanthin0.42 ± 0.09a0.01 ± 0.00b0.01 ± 0.00bn.d.1.54 ± 0.16c0.01 ± 0.00b
Neoxanthin0.54 ± 0.12an.d.n.d.n.d.0.66 ± 0.01an.d.
Fritschiellaxanthinn.d.1.15 ± 0.14a1.12 ± 0.06a0.63 ± 0.13bn.d.2.00 ± 0.17c
Canthaxanthinn.d.0.08 ± 0.00a0.08 ± 0.04a0.14 ± 0.02an.d.n.d.
3′-Hydroxyechinenonen.d.0.04 ± 0.01a0.03 ± 0.02a0.34 ± 0.03bn.d.n.d.
Adonirubinn.d.0.11 ± 0.00a0.14 ± 0.06a0.82 ± 0.05bn.d.0.08 ± 0.00a
Adonixanthinn.d.0.15 ± 0.02a0.11 ± 0.01an.d.n.d.0.05 ± 0.05a
4-Ketoantheraxanthinn.d.0.40 ± 0.07a0.37 ± 0.02a0.03 ± 0.00bn.d.0.29 ± 0.02a
Astaxanthinn.d.5.13 ± 0.11a5.44 ± 0.75a1.88 ± 0.09bn.d.3.29 ± 0.56b
Total chrolophyll23.45 ± 5.05a15.82 ± 0.61abc14.08 ± 1.55abc9.82 ± 0.79c22.29 ± 1.26ab13.55 ± 0.37bc
Chlorophyll a17.74 ± 3.72a11.53 ± 0.38abc10.37 ± 1.09bc7.42 ± 0.63c16.65 ± 1.04ab9.56 ± 0.23bc
Chlorophyll b5.71 ± 1.33a4.29 ± 0.23ab3.72 ± 0.46ab2.40 ± 0.15b5.64 ± 0.22a3.99 ± 0.16ab

Expression of isoprenoid and carotenoid biosynthesis genes in transplastomic tobacco

The total carotenoid content in ZW-2 and ZW-9 was 2.1-fold higher than in wild-type plants (Table 1). As the carotenoid pool size is regulated mainly at the transcription level of rate-limiting enzymes present upstream of the isoprenoid and carotenoid biosynthetic pathways, such as 1-deoxy-d-xylulose 5-phosphate synthase (DXS) and phytoene synthase (PSY) (Römer and Fraser, 2005; Sandmann et al., 2006), the increase in total carotenoid is not easily explained by the enhancement of β-carotene-converting activity. To explore the possibility of alterations in the transcription of genes in the pathway leading to β-carotene, the mRNA levels of the genes encoding DXS, isopentenyl diphosphate isomerase (IPI), PSY, PDS and lycopene β-cyclase (CrtL-1) were determined by quantitative RT-PCR (Figure 2d). An increase in dxs and psy transcript levels was observed in the ZW-2 and ZW-9 lines, compared with wild-type plants, but the difference in these transcript levels was not statistically significant according to a one-way anova (P > 0.05).

Growth analysis

The overproduction of astaxanthin in plants led to drastic changes in their carotenoid composition (Table 1), which raised a concern that it may have a negative effect on plant photosynthetic growth. We measured the growth of the transformants under light conditions of 300 μmol photons m−2 s−1. As shown in Figure 4a, the transformants showed slower growth rates of their aerial parts compared with wild-type plants, but the final height of the transplastomic plants was not significantly different from that of wild-type plants at the stage of flower initiation. Also, there was no significant difference in the diameter of plants between the transplastomic and the wild-type plants at the flowering stage (Figure 4b).

Figure 4.

 Plant growth and photosynthesis. Changes in both height (a) and diameter (b) of WT (bsl00041), ZW-2 (□) and ZW-9 (Δ) plants grown under ambient CO2 and 300 μmol photons m−2 s−1 light conditions are plotted. The diameter is defined as the length of the projected plant silhouette on the ground. Light-response curves for CO2 assimilation rates (A) at ambient (370 μmol mol−1) CO2 conditions (c) and the effect of altered intracellular CO2 concentration (Ci) on the CO2 assimilation rate in the leaves of WT (bsl00041), ZW-2 (□) and ZW-9 (Δ) plants under 1000 μmol photons m−2 s−1 light conditions (d). Values are the average from measurements on three or four different tobacco plants, ±SEM. CO2 assimilation rates were analyzed by one-way anova (Sokal and Rohlf, 1995). A post hoc Tukey’s HSD test was carried out on the grouped means. Values followed by the same letter are not significantly different (> 0.05).

Photosynthesis analysis

Photosynthesis parameters for the leaves of the transplastomic plants grown under light conditions of 300 μmol photons m−2 s−1 were analyzed (Table 2). The chlorophyll contents of the ZW-2, ZW-9, W-6, Z-1 and WZ-4 lines were 52.0, 41.8, 36.2, 77.0 and 33.3% lower than that of wild-type plants, respectively. The ZW-2, ZW-9 and WZ-4 lines showed lower chlorophyll a/b ratios compared with wild-type plants, indicating that a change in the chlorophyll distribution ratio was caused by the expression of both crtZ and crtW. The light absorption efficiency (p) of leaves was 85.9, 82.9, 78.9 and 76.1% in wild-type, ZW-2, ZW-9 and WZ-4 plants, respectively. There was no significant difference in p between wild-type and ZW-2 plants, although the chlorophyll content of ZW-2 plants was almost reduced by half compared with wild-type plants. A parameter of chlorophyll fluorescence (Fv/Fm) in ZW-2 and ZW-9 plants was only 69% of that in wild-type plants, indicating that the efficiency of light-energy utilization for photosynthesis reaction in the transplastomic plants was lower than that in wild-type plants.

Table 2.   Photosynthesis parameters for wild-type and transplastomic tobacco plants
ParameterWild-typeZW-2ZW-9W-6Z-1WZ-4
  1. p, light absorption efficiency of the leaf. Parameters were analyzed by one-way anova (Sokal and Rohlf, 1995). A post hoc Tukey’s HSD test was carried out on the grouped means. Values are the averages from measurements of three or four different tobacco plants, ±SEM. Values followed by the same letter are not significantly different (> 0.05).

p0.859 ± 0.014a0.829 ± 0.011abc0.789 ± 0.010bc0.769 ± 0.011bc0.834 ± 0.017ab0.761 ± 0.003c
Fv/Fm0.84 ± 0.02a0.58 ± 0.02b0.59 ± 0.01b0.59 ± 0.00b0.79 ± 0.01a0.50 ± 0.01c
Total chlorophyll (μmol m−2)558 ± 36a290 ± 31b233 ± 29b202 ± 20b430 ± 28c186 ± 21b
Chlorophyll a/b3.24 ± 0.16a2.74 ± 0.05bc2.85 ± 0.06bd3.14 ± 0.08ad3.00 ± 0.08bc2.44 ± 0.07c
Nitrogen (mmol m−2)89.4 ± 5.9ab85.8 ± 9.2ab69.1 ± 6.6ab101.5 ± 10.7a81.5 ± 6.1abd71.8 ± 9.1b
Rubisco (g mol N−1)20.1 ± 1.5a23.6 ± 1.5a22.3 ± 1.9a19.0 ± 3.3a24.8 ± 2.0a19.5 ± 2.3a

The light-intensity dependence of photosynthetic rate in transformants was determined under ambient CO2 concentration. Although ZW-2 and ZW-9 plants showed lower CO2 assimilation rates than wild-type plants at less than 1000 μmol photons m−2 s−1, there was no significant difference in the CO2 assimilation rate between transplastomic and wild-type plants at 1500 μmol photons m−2 s−1 (Figure 4c). In general, the CO2 assimilation rate in leaves is limited by either ribulose-1,5-bisphosphate (RuBP) carboxylase/oxygenase (Rubisco) activity or by the rate of regeneration of RuBP, the substrate of Rubisco. Based on the photosynthesis models of Farquhar et al. (1980) and Sharkey (1985), the determination of the dependence of CO2 assimilation rate on intracellular CO2 concentrations reveals the rate-limiting factor of photosynthesis. The assimilation rate of CO2 is limited by Rubisco activity under low CO2, and is limited by RuBP regeneration, which is dependent on the energy (ATP and NADPH) supplied by electron transport in photosystem complexes at above ambient CO2 concentrations. As shown in Figure 4d, ZW-2 and ZW-9 plants showed similar CO2 assimilation rates under low CO2 concentrations, compared with wild-type plants, indicating that the CO2 assimilation capacity was not impaired by the expression of crtW and crtZ. On the other hand, the CO2 assimilation rates of transformants are significantly lower than that of wild-type plants under high CO2 concentrations. These results indicate that the depression of the CO2 assimilation rate in CrtZ/CrtW-expressing plants (Figure 4c) depends on the decrease in their electron transport activities.

Discussion

In this study we have demonstrated that the genetic engineering of carotenoid biosynthesis in chloroplasts enables plants to produce large quantities of the highly valued red carotenoid, which changed the color of the leaves from green to reddish brown. Transplastomic tobaccos that expressed two genes encoding CrtW and CrtZ from Brevundimonas sp. SD212 accumulated large quantities of astaxanthin at concentrations of up to 5.44 mg g−1 DW, corresponding to 74% of the total carotenoids. This greatly exceeds the levels reached in previous reports (Tables 1 and 3). In the transplastomic ZW-9 plants, the total ketocarotenoid content reached 7.29 mg g−1 DW, corresponding to 99% of the total carotenoids (Table 1). Ralley et al. (2004) showed that the integration of Paracoccus sp. N81006 crtW and crtZ genes into the tobacco nuclear genome accumulated trace levels of astaxanthin and 0.6–0.8 mg g−1 DW of ketocarotenoid intermediates, which was one tenth of those found in ZW-9 plants. In a transgenic potato leaf expressing the crtO gene of Synechocystis sp. PCC6803 under the CaMV 35S promoter, the content of ketocarotenoids was only 0.2 mg g−1 DW including trace levels of astaxanthin (Gerjets and Sandmann, 2006). Complementation analyses using E. coli revealed that CrtW and CrtZ from Brevundimonas sp. SD212 have the highest conversion efficiency of β-carotene to astaxanthin by way of the ketocarotenoid intermediates, among various CrtW (or CrtO) and CrtZ (or CYP175A) proteins, respectively, including the corresponding proteins from Paracoccus sp. N81106 and Synechocystis sp. PCC6803 (Choi et al., 2005, 2006, 2007). It is therefore likely that higher levels of astaxanthin content in the transplastomic tobacco plants are attributable to the highest efficiency of astaxanthin synthesis by the Brevundimonas sp. SD212 CrtW and CrtZ. Recently, Jayaraj et al. (2007) expressed the H. pluvialis bkt gene linked to the pea Rubisco small subunit transit peptide sequence in carrot nuclear genome under the control of three different promoters (double CaMV 35S, Arabidopsis–ubiquitin or RolD from Agrobacterium rhizogenes), and showed that the transgenic carrot plants were able to accumulate larger quantities of astaxanthin in the leaves and roots. The level reached 15.4% of total carotenoids in the leaves, and 26.5% in the roots, although the level of astaxanthin in the leaves was only 0.0347 mg g−1 (fresh weight; FW). As plastid genetic engineering has the potential to confer high levels of protein accumulation in transformants (Viitannen et al., 2004), the overproduction of the ketocarotenoids, including astaxanthin, in transplastomic tobacco plants could be achieved by a greater enhancement of both CrtZ and CrtW activities. The overproduction could also be attained by the use of the most efficient crtZ and crtW genes encoding the Brevundimonas sp. SD212 corresponding proteins, the codon usage of which is optimized for expression in higher plants. In order to verify a decisive factor for the high astaxanthin yields shown in the present study, both the accumulation levels of the CrtZ and CrtW proteins, and the turnover rates of ketocarotenoids in the transformatns, will be examined in more detail in the near future, although such experiments do not seem easy to perform.

Table 3.   Comparison of astaxanthin production with previous reports
Transgenic plant (cultivar)Transgene (origin)RecombinationProduction tissueAstaxantin level (μg g−1 dry weight)Reference
  1. Asterisks indicate data for fresh weight.

S. tuberosum (Mayan Gold)bkt (H. pluvialis)NuclearTuber13.9Morris et al., 2006
S. tuberosum (Baltica)crtO (Synechocystis 6803)NuclearTuber
Leaf
0.7
Detactable
Gerjets and Sandmann, 2006
N. tabacum (Samsun)crtW/crtZ (Paracoccus sp. N81108)NuclearLeaf, nectaryDetectableRalley et al., 2004
D. carotabkt (H. pluvialis)NuclearRoot
Leaf
91.6*
34.7*
Jayaraj et al., 2007
N. tabacum (Xanthi)crtW/crtZ (Brevundimonas sp.)PlastidLeaf5440This study

It was surprising that in transplastomic plants (ZW-2, ZW-9, W-6, and WZ-4), almost all of the native carotenoids, such as β-carotene and violaxanthin, neoxanthin and lutein, that are synthesized in wild-type leaves were replaced with ketocarotenoids, including a large proportion of astaxanthin, which are not naturally found in most of the higher plants (Table 1). The strong catalytic activity of CrtW in the chloroplasts seems to preferentially convert the native carotenoids into ketocarotenoids, e.g. zeaxanthin into astaxanthin via adonixanthin. A novel carotenoid 4-ketoantheraxanthin and a rare carotenoid fritschiellaxanthin (Matsuno, 1991) were also detected in the transformants. The strong activity of CrtW is considered to be responsible for the conversion of antheraxanthin and lutein into 4-ketoantheraxanthin and fritschiellaxanthin, respectively (Figure 1). Transplastomic plants that expressed both the crtW and crtZ genes (ZW-2, ZW-9 and WZ-4) showed higher astaxanthin content, compared with the plants that expressed only crtW, indicating that the crtZ gene contributes to the accumulation of astaxanthin. Overexpression of CrtZ in the Z-1 plants leads to an increase of the size of the xanthophyll cycle pool (reversible interconversion of violaxanthin and zeaxanthin), as has been shown in previous reports (Davison et al., 2002; Johnson et al., 2007). Hydroxylation reactions of β-carotene are very likely to be a rate-limiting step of xanthophyll biosynthesis. Therefore, an increase in the supply of zeaxanthin would confer an enhancement of astaxanthin production in the crtZ/crtW co-expressing plants.

Moreover, the total carotenoid content in the transplastomic ZW-2 and ZW-9 plants resulted in a 2.1-fold increase, compared with wild-type plants. It should be noted that the size of the carotenoid pool in leaves is considered to be controlled by the activity of PSY, which mediates the pathways upstream of the carotenoid biosynthesis pathway (Fray et al., 1995; Römer and Fraser, 2005; Sandmann et al., 2006) (Figure 1), e.g. in a transgenic potato tuber, the carotenoid epoxidation was suppressed, showing a 5.7-fold increase in total carotenoid content, and the psy mRNA levels were also considerably increased (Römer et al., 2002). DXS in the MEP (2-C-methyl-D-erythritol-4-phosphate) pathway that supplies the carotenoid precursor, isopentenyl diphosphate, is also known to be one of the rate-limiting enzymes for carotenoid production (Estévez et al., 2001). In the present study, quantitative RT-PCR analysis demonstrated an upward trend in the levels of the psy and dxs transcripts in the leaves of transformants, compared with those in wild-type plants (Figure 2d), although the difference was not statistically significant. Therefore, the overproduction of carotenoids may be achieved by the enhancement of the expression of psy and dxs genes concomitantly with the strong enhancement of the CrtZ/CrtW pathways. In a previous report (Fray et al., 1995), the constitutive expression of the psy gene in transgenic tomato plants caused dwarfism. Also, the expression level of DXS in transgenic Arabidopsis showed a negative correlation with the growth and germination rates (Estévez et al., 2001). On the other hand, no significant difference was observed in plant size between the transplastomic and wild-type tobacco plants. At present, the mechanism that causes dwarfism by the expression of the psy or dxs gene is still obscure, and is yet to be examined in more detail.

Carotenoids are essential components of the photosynthesis apparatus in plants, where they participate in both light harvesting, for energy transfer to the chlorophylls, and photoprotection, by quenching triplet-state chlorophyll molecules. Chloroplast is a major organelle that synthesizes carotenoids de novo, and contains β-carotene in substantial quantities (40% of total carotenoid). β-Carotene is a major substrate for astaxanthin synthesis in the transplastomic plants expressing the crtW gene. As carotenoids play such crucial roles, a dramatic change in carotenoid composition in leaf chloroplasts could potentially have a negative effect on plant photosynthetic growth. In this study, we have demonstrated that the final size of the transformants is similar to that of wild-type plants, although their growth rate is reduced (Figure 4a,b). On the basis of Farquhars’ photosynthesis models (Farquhar et al., 1980; Sharkey, 1985), the response of photosynthesis rate to intracellular CO2 concentrations in leaves (Figure 4d) indicates that the CO2 assimilation capacity in the transplastomic plants is similar to that in wild-type plants. This is also supported by the fact that there is no difference in Rubisco content between the transplastomic and the wild-type plants (Table 2). As shown in Figure 4c, the photosynthesis rate in the transformants is approximately 55% of that in wild-type plants under light conditions of 300 μmol photons m−2 s−1. By contrast, there is no significant difference in the photosynthesis rate between the transformants and the wild-type plants under light conditions of 1500 μmol photons m−2 s−1. The reduction in the quantum yield (Fv/Fm) of photosynthesis in the transplastomic plants (Table 2) would correlate with the suppression of the light harvesting function by the decrease in the chlorophyll content. Accordingly, an insufficiency of ambient light energy under low light conditions (300 μmol photons m−2 s−1) is likely to have led to the decrease in photosynthesis rate in the transplastomic leaves. On the other hand, with a sufficient supply of light energy to leaves under high light conditions (more than 1500 μmol photons m−2 s−1), the transplastomic plants would harvest enough light energy to show similar performance in photosynthesis rate to wild-type plants.

Therefore, on the assumption that the carbon atoms necessary for astaxanthin synthesis are supplied by photosynthesis, the transformants grown under natural sunlight (approximately 2000 μmol photons m−2 s−1) would produce more carotenoids, compared with transformants grown under low light intensity. This implies that an additional increase in astaxanthin content in the transplastomic plants could be achieved by improving the growth conditions. This would compensate for the growth delay of the transplastomic plants (Figure 4a,b). These results indicate that chloroplasts can accumulate large quantities of ‘unnatural’ carotenoids, specifically astaxanthin, without any significant damage to the plants.

Plastid transformation offers attractive features in plant genetic engineering. It has many advantages over nuclear genome transformation: high-level foreign protein accumulation, no need of a transit peptide, absence of gene silencing, convenient transgene stacking in operons and gene containment resulting from the strict maternal inheritance of the chloroplast genome. Tobacco has been the most widely exploited host for plastid transformation, because of its ease of use in genetic manipulations (Verma and Daniell, 2007). As 1 acre of tobacco can produce more than 40 metric tons of leaves per year (Verma and Daniell, 2007), this plant holds great promise as a host for the production of useful compounds. Furthermore, the platform of plastid transformation technology has been expanded to several vegetable crops, such as tomato (Ruf et al., 2001), carrot (Kumar et al., 2004) and lettuce (Kanamoto et al., 2006). The modification of carotenoid composition in crop plants is important for the improvement of nutritional value and for the enhancement of stress tolerance (Römer and Fraser, 2005). The present study highlights the utility of plastid transformation as an excellent tool to drastically alter carotenoid compositions and contents in crop plants.

Experimental procedures

Construction of plastid transformation plasmids

The crtW and crtZ genes, encoding the CrtW and CrtZ proteins from Brevundimonas sp. SD212, respectively, were chemically synthesized according to the codon usage of rape (GenBank accession no. AB377271 and AB377272, respectively). Two kinds of PCR fragments were amplified from the chemically synthesized crtZ gene using two PCR-primer sets: P1 (5′-CGGGATCCAAAGAGGAGAAATTACATATGGCTTGGCTTACTTGGATCGCTC-3′)/P2 (5′-GCTCTAGATTAAGCTCCGCTAGAAGAAGATCCCCTCTTTTG-3′) and P5 (5′-GCTCTAGAAAAGAGGAGAAATTACATATGGCTTGGCTTACTTGGATCGCTC-3′)/P6 (GGAATTCTCAAGCTCCGCTAGAAGAAGATCCCCTCTTTTG-3′). Similarly, chemically synthesized crtW was amplified using two PCR primer sets: P3 (5′-CGGGATCCAAAGAGGAGAAATTACATATGACTGCTGCTGTTGCTGAGCCTAG-3′)/P4 (5′-GCTCTAGATTAAGACTCTCCTCTCCAAAGTCTCCACCAAG-3′) and P7 (5′-GCTCTAGAAAAGAGGAGAAATTACATATGACTGCTGCTGTTGCTGAGCCTAG-3′)/P8 (GGAATTCTCAAGACTCTCCTCTCCAAAGTCTCCACCAAG-3′). The BamHI, XbaI and EcoRI sites are underlined. PCR-amplified crtZ (using P1 and P2) and crtW (using P1 and P2) were digested with BamHI and XbaI, and were then ligated into the BglII–XbaI site of pLD7-rrnP-MCS (GenBank accession no. AB375764), which contains the rrn promoter/5′-untranslated region (UTR), rps16 terminator and aminoglycoside 3′-adenyltransferase (aadA) gene under the control of the rrn promoter and psbA terminator, yielding plasmid pLD7-rrnP-crtZ and pLD7-rrnP-crtW. PCR-amplified crtZ (using P1 and P2) was digested with XbaI and EcoRI, and was then ligated into pLD7-rrnP-crtW, yielding plasmid pLD7-rrnP-crtW-crtZ. PCR-amplified crtW (using P7 and P8) was digested with XbaI and EcoRI, and was ligated into pLD7-rrnP-crtZ, yielding plasmid pLD7-rrnP-crtZ-crtW. The resultant plasmids (pLD7-rrnP-crtZ, pLD7-rrnP-crtW, pLD7-rrnP-crtZ-crtW and pLD7-rrnP-crtW-crtZ) were cut with NotI and SalI, and were introduced between the rbcL and accD sequences of similarly digested pLD200 (GenBank accession no. CS165378), yielding the final constructs, pLD200-Z, pLD200-W, pLD200-ZW and pLD200-WZ, that were used for plastid transformation.

Production and growth of transplastomic plants

Wild-type tobacco plants (N. tabacum L. cv. Xanthi) were grown on MS medium (Murashige and Skoog, 1962) supplemented with 0.2% gellan gum (Wako, http://www.wakousa.com), and aseptic leaves were subjected to particle bombardment for homologous recombination, as described previously (Svab and Maliga, 1993). Homoplastomic transgenic shoots were obtained by repeated shoot regeneration on RMOP medium (Svab and Maliga, 1993) containing 500 μg ml−1 spectinomycin dihydrochloride, and were then rooted on MS medium. The resultant T0 plants were transplanted into Metro-Mix350 (Sun Gro Horticulture, http://www.sungro.com) in pots (700 ml plant−1) and were cultured in a growth chamber at 25°C under a 16-h light/8-h dark regime at a photon flux density of 300 μmol m−2 s−1. Plants were fertilized with diluted (1/500) HYPONeX (Hyponex Japan, http://www.hyponex.co.jp) three times a week. Plants from the homoplastomic T1 generation were used in all of the experiments.

DNA and RNA analysis

Total plant DNA was isolated using the Qiagen DNeasy mini kit (http://www.qiagen.com). Plant DNA (2 μg) was digested with EcoRV, fractionated on a 0.8% agarose gel and then transferred onto a nylon membrane. The preparation of specific probes, hybridizations and washings were carried out with AlkPhos Direct (GE Healthcare, http://www.gehealthcare.com) following the manufacturer’s instructions. Chemiluminescent substrate, CDP-Star (GE Healthcare) was used for signal detection.

Total RNA was extracted from tobacco leaves using QuickGene-810 (Fujifilm, http://www.fujifilm.com) and the QuickGene RNA cultured cell kit (Fujifilm). For cDNA synthesis, the RNA was incubated with RNase-free DNase I (Promega, http://www.promega.com) and was reverse transcribed by ImProm-II Reverse Transcriptase (Promega) with random primers. Real-time quantitative PCR was performed on a GeneAmp 5700 (Applied Biosystems, http://www.appliedbiosystems.com) using a TaqMan PCR master mix. The relative levels of the amplified mRNA were evaluated according to the 2−ΔΔCt method (Winer et al., 1999) using the actin gene for normalization. A value, Ct, is calculated based on the time at which the reporter fluorescent emission increases beyond a threshold level. FAM Dye-labeled TaqMan probes and primers were designed, based on cDNA sequences from tobacco using primer express 1.5 software (Applied Biosystems) as follows: for crtW, 5′-TGTTCTTACTGCTCTTGTTCTTATCGCTCTTTTCG-3′ (probe) and 5′-AACTTACTTCGGATGGAGAGAGATG-3′/5′-GAAGGTTAGCAGGTCTAGCTCCAA-3′(primers); for crtZ, 5′-TTGGACTTGGAATCACCGCTTACGG-3′ (probe) and 5′-TGGCCTTGGGCTCTTCCT-3′/5′-TCCATCGTGGAAGAAGAAGTACAC-3′ (primers); for dxs (GenBank accession no. AJ291721), 5′-CGCTATGGGTGGTGGGACCGGT-3′ (probe) and 5′-AGATAAGGACATTGTTGCAATCCAT-3′/5′-GCGACGGTGGAATAGGTTCA-3′(primers); for ipi (GenBank accession no. Y06943), 5′-AGCTGCCTCTCTCCGCTGCTATTCCTC-3′ (probe) and 5′-TCCCCTCTTCTTAAAAGCCATAGA-3′/5′-GGCGTCACCCATTCTTGTAGA-3′ (primers); for psy (GenBank accession no. AX657549), 5′-ATTTGCTGGAAGAGTGACTGATAAGTGGAGGAA-3′ (probe) and 5′-GGGCTCTCCGACGAAGACA-3′/5′-CTCGCCCTCTGAATTTGTTTCT-3′ (primers); for pds (GenBank accession no. AJ616742), 5′-TGGCCTTTTTAGATGGTAACCCTCCTGAGA-3′ (probe) and 5′-CTTCAGGAGAAACATGGTTCAAAA-3′/5′-CAACAATCGGCATGCAAAGT-3′ (primers); for crtL-1 (GenBank accession no. X81787), 5′-CCATCGCCGAAATTGATATGGCCC-3′ (probe) and 5′-GGACTATCGGTTGTTTCAATCGA-3′/5′-TCATCCACCCAAACACCATAGTT-3′ (primers); and for actin (GenBank accession no. U60494), 5′-CTTCCCCATGCCATCTTGCGGTT-3′ (probe) and 5′-TCCCCATCTATGAGGGATATGC-3′/5′-ACCACGACCGGCAAGATC-3′ (primers).

Pigment analysis

Pigment levels were determined on the same leaves as were used for the photosynthetic measurements. Leaf discs (15 mm in diameter) were punched from the leaf, immediately frozen in liquid nitrogen and ground to a fine powder. Pigments were extracted from the powder as described previously (Fraser et al., 2000). Pigments were analyzed with Acquity UPLC system (Waters, http://www.waters.com) using the photodiode array detector. Data was collected and analyzed using masslynx software (Waters). Pigments were separated using an ACQUITY BEH Shield RP18 (Waters) column (2.1 × 150 mm, 1.7-μm particle size) operating at 35°C, and were eluted using 30% solvent A (methanol:water, 50:50, v/v) for the first 3.5 min, followed by a 4.5-min linear gradient to 100% solvent B (acetonitrile), which continued isocratically until the end of the 11-min separation. A flow rate of 0.6 ml min−1 was used. Pigments were identified by their retention time and absorption spectra, and were quantified by integrating peak areas. We converted peak areas to molar concentrations by comparison with authentic standards that were either purchased from Wako or purified from transgenic tobacco plants expressing crtW and crtZ.

Fritschiellaxanthin (4-ketolutein) and 4-ketoantheraxanthin were purified and identified as follows. Freeze-dried tobacco leaves (44.28 g) were ground in a mill. The resulting powder was extracted twice with 1 l each of CH2Cl2:MeOH (1:1). The extracts were combined and concentrated into a small volume in vacuo, and were partitioned with EtOAc/H2O without adjusting the pH. The EtOAc layer was evaporated to dryness and subjected to silica gel chromatography, using hexane:EtOAc (1:1). The red-colored fractions were collected and concentrated to dryness to produce a red oil (150.3 mg). This red oil was subjected to preparative Octa Desyl Silyl (ODS) HPLC (Senshu Pak PEGASIL ODS column, 20 × 250 mm; Senshu Scientific, http://www.ssc-jp.com), and was separated with CH3CN:CH2Cl2 (8:2) as a solvent (flow rate 8.0 ml min−1). Three red components could be separated by this chromatography. The main red component (eluted at 11.0 min, 15.9 mg) was identified to be astaxanthin by direct comparison with the authentic standard. The red compound that eluted at 12.1 min (6.7 mg) was identified to be fritschiellaxanthin (Matsuno and Ookubo, 1982) by the 1H-NMR and HRESI-MS spectral data. The red compound that eluted at 9.8 min (6.5 mg) was further purified by preparative silicagel HPLC (YMC-Pack SIL column, 20 × 250 mm; YMC, http://www.ymc.co.jp) using hexane:acetone (7:3) to give a pure novel carotenoid 4-ketoantheraxanthin (3.2 mg). The structure of 4-ketoantheraxanthin was determined by HRESI-MS and NMR analyses, and their data are reported elsewhere (Shindo et al., 2008). Their absolute configuration was deduced from that of lutein and antheraxanthin, which seem to be the substrates for fritschiellaxanthin and 4-ketoantheraxanthin, respectively.

Spectral data for the novel 4-ketoantheraxanthin are as follows: UPLC-PDA, retention time 1.9 min, λmax 456 nm; HREESI-MS, m/z 621.39235 [C40H54O4Na, calculated for 621.39198; the intense (M + Na)+ peak in CH2Cl2 in the positive-ion mode] (Shindo et al., 2008).

Photosynthesis analysis

CO2 fixation (gas exchange) and chlorophyll fluorescence were measured simultaneously using an LI-6400 (LI-COR, http://www.licor.com) after incubation of tobacco plants in the dark for more than 60 min (Miyake et al., 2005). All measurements were repeated at least three times using three or four different plants. The temperature of the leaf chamber was adjusted to 25°C. The mixture of gases was saturated with water vapor at 16°C. The ratio of light absorbed by chloroplasts in tobacco leaves, p, was determined with an LI-250A Light Meter and an 1800-12S External Integrating Sphere (LI-COR). p was calculated as: p = 1 – reflectance – transmittance (Miyake et al., 2005).

Measurements of Rubisco and leaf nitrogen

The levels of Rubisco and total leaf nitrogen were determined on the same leaves as were used for the gas exchange analysis. After the photosynthetic measurements, leaf discs (15 mm in diameter) were obtained from the leaves, and were immediately frozen in liquid nitrogen. The leaf discs were homogenized, using a chilled mortar and pestle, in 1 ml of 50 mm potassium phosphate buffer (pH 7.5) containing 1 mm ascorbate, 1 mm EDTA and protease inhibitor cocktail. The homogenate was centrifuged at 15 000 g for 10 min at 4°C. The quantity of Rubisco protein in the supernatant was determined spectrophotometrically by formamide extraction of the Coomassie brilliant blue R-250 stained protein after SDS-PAGE, as described previously (Miyake et al., 2005). The quantity of leaf nitrogen was determined by the method of Miyake et al. (2005), with a DK20/26 decomposer connected with SMS scrubber and JP recirculating pump (ACTAC, http://actac.co.jp) and a SuperKjel 1200/1250 System (ACTAC). The leaf disc was subjected to Kjeldahl digestion.

Acknowledgement

This work was supported by the New Energy and Industrial Technology Development Organization (NEDO), Japan.

Ancillary