Isopentenyl/dimethylallyl diphosphate isomerase (IPI) catalyzes the interconversion of isopentenyl diphosphate (IPP) and dimethylallyl diphosphate (DMAPP), which are the universal C5 units of isoprenoids. In plants, IPP and DMAPP are synthesized via the cytosolic mevalonate (MVA) and plastidic methylerythritol phosphate (MEP) pathways, respectively. However, the role of IPI in each pathway and in plant development is unknown due to a lack of genetic studies using IPI-defective mutants. Here, we show that the atipi1atipi2 double mutant, which is defective in two Arabidopsis IPI isozymes, exhibits dwarfism and male sterility under long-day conditions and decreased pigmentation under continuous light, whereas the atipi1 and atipi2 single mutants are phenotypically normal. We also show that the sterol and ubiquinone levels in the double mutant are <50% of those in wild-type plants, and that the male-sterile phenotype is chemically complemented by squalene, a sterol precursor. In vivo isotope labeling experiments using the atipi1atipi2 double mutant revealed a decrease in the incorporation of MVA (in its lactone form) into sterols, with no decrease in the incorporation of MEP pathway intermediates into tocopherol. These results demonstrate a critical role for IPI in isoprenoid biosynthesis via the MVA pathway, and they imply that IPI is essential for the maintenance of appropriate levels of IPP and DMAPP in different subcellular compartments in plants.
Isoprenoids, the largest group of natural products in living organisms, are derived from a basic five-carbon unit, isopentenyl diphosphate (IPP), and its allyl isomer dimethylallyl diphosphate (DMAPP) (Sacchettini and Poulter 1997). IPP is sequentially condensed to DMAPP to yield the short-chain isoprenoid precursors geranyl diphosphate, farnesyl diphosphate (FPP) and geranylgeranyl diphosphate (GGPP), which are further metabolized to monoterpenes (C10), sesquiterpenes (C15) and diterpenes (C20). IPP isomerase (IPI; EC 22.214.171.124) catalyzes the interconversion of IPP to DMAPP, which is an essential starter moiety for the condensation reactions. IPIs are classified into two types: type I and type II. Type I IPIs have been identified in various organisms, including humans (Xuan et al. 1994), Saccharomyces cerevisiae (Mayer et al. 1992), Escherichia coli (Hahn et al. 1999) and Rhodobacter capsulatus (Hahn et al. 1996), whereas the type II enzymes that have been identified in archaea and some bacteria are FMN and NAD(P)H dependent (Kaneda et al. 2001). To date, no type II IPI has been identified in the plant kingdom.
Most plants have two type I IPI isozymes with distinct subcellular localizations, as in the case of FPP and GGPP synthases in Arabidopsis (Cunillera et al. 1997, Okada et al. 2000). In tobacco, IPIs have been found to be localized in the cytosol and plastids through the use of a green fluorescent protein (GFP)-fused protein (Nakamura et al. 2001), while in castor beans, mitochondrial and proplastidial IPIs have been detected by biochemical approaches (Green et al. 1975). Like other plant species, Arabidopsis also has two IPI genes, AtIPI1 and AtIPI2, which were isolated by complementation of the yeast idi mutant (Campbell et al. 1998). Both genes encode proteins with N-terminal extensions that may function as translocation signals for entry into plastids (Campbell et al. 1998, Cunningham and Gantt 2000); however, the subcellular localization of these enzymes has yet to be reported. IPI transcription is altered by various environmental stimuli; for example, tobacco IPI gene expression is responsive to strong light and cold stress (Nakamura et al. 2001), while green algae IPI is expressed predominantly under intense light during the red-cyst stage (Sun et al. 1998). Thus, IPI is thought to catalyze a regulatory step in isoprenoid biosynthesis.
IPP biosynthesis occurs via two pathways in living organisms: the mevalonate (MVA) pathway, which was first reported in yeast and mammals (Mayer et al. 1992), and the methylerythritol phosphate (MEP) pathway, which was recently identified in eubacteria and plants (Mayer et al. 1992, Lichtenthaler 1999, Rohmer 1999). As shown in Fig. 1, plants biosynthesize IPP and DMAPP via the MVA pathway in the cytoplasm and the MEP pathway in plastids (Lichtenthaler 1999, Rohmer 1999). Plants defective in MEP pathway enzymes display an albino phenotype (Estevez et al. 2000, Estevez et al. 2001, Okada et al. 2002, Guevara-García et al. 2005, Hsieh et al. 2005), whereas a block in the MVA pathway results in dwarfism, early senescence and male sterility (Suzuki et al. 2004). These observations suggest that the MVA and MEP pathways are metabolically separated and provide IPP and DMAPP for the biosynthesis of distinct sets of isoprenoids. However, evidence suggests the exchange of a small fraction of isoprenoid precursors, such as IPP, DMAPP and other prenyl diphosphates, between the cytosolic MVA pathway and the plastidial MEP pathway (Kasahara et al. 2002, Bick et al. 2003, Laule et al. 2003).
It has been reported that the S. cerevisiae IPI gene, IDI1, is an essential single-copy gene, and a disruption of the IDI gene results in a lethal phenotype (Mayer et al. 1992). Based on knowledge from other kingdoms, the conversion of IPP to DMAPP by IPI in plants is thought to be necessary for isoprenoid biosynthesis via the cytosolic MVA pathway (Heintz et al. 1972, Nes and Venkatramesh 1999, Lange and Ghassemian 2003). In contrast, DMAPP is synthesized directly along with IPP without IPI activity by 1-hydroxy-2-methyl-2-(E)-butenyl 4-diphosphate (HMBPP) synthase via the MEP pathway in plastids (Lichtenthaler 1999, Rohmer 1999, Rohdich et al. 2002). However, silencing of the tobacco IPI leads to a depletion of photosynthetic pigments, suggesting that reduced IPI activity affects isoprenoid biosynthesis in the plastids of tobacco leaves (although the effect on the cytosolic MVA pathway was not investigated) (Page et al. 2004). Given that no IPI-deficient plant data have been published, the contribution of IPI to isoprenoid biosynthesis via each pathway and in plant development is unknown.
In this study, we investigated the role of IPI in isoprenoid biosynthesis throughout the life cycle of Arabidopsis by analyzing T-DNA insertion mutants for IPI genes. To predict the roles of these genes in the MVA and MEP pathways, we examined the subcellular localization of the IPIs using GFP as a reporter, and we used tracer experiments to determine the metabolic flux through the MVA and MEP pathways in our ipi mutant. Based on our results, we discuss how IPI contributes to isoprenoid biosynthesis in Arabidopsis.
Organ-specific expression of AtIPI1 and AtIPI2
Previously, it was reported that two IPI homolog genes of Arabidopsis were expressed at higher levels in roots than in aerial parts (Campbell et al. 1998). To investigate organ-specific expression of two Arabidopsis IPI genes, AtIPI1 (At5g16440, IDI1/IPP1) and AtIPI2 (At3g02780, IPP2), at both mRNA and protein levels in more detail, we carried out real-time quantitative reverse transcription–PCR (QRT–PCR) and immunoblot analysis. QRT–PCR analysis showed that the AtIPI1 mRNA was mainly expressed in floral clusters, and that the AtIPI2 mRNA was predominantly expressed in roots and floral clusters (Fig. 2A). Immunoblot analysis using anti-AtIPI2 polyclonal antibodies, which reacts with both AtIPI1 and AtIPI2 proteins, showed that AtIPI1 and AtIPI2 proteins were most abundant in flowers and roots, respectively, but that both proteins were detectable in every organ examined in these experiments (Fig. 2B). These results indicated that the AtIPI genes were highly expressed in flowers and roots.
Subcellular localizations of the IPIs in Arabidopsis
In Nicotiana tabacum, two IPIs were separately localized in plastids (chloroplasts) and the cytosol, suggesting that IPIs function in both MVA and MEP pathways (Nakamura et al. 2001). Although Arabidopsis AtIPI1 and AtIPI2 have a putative organelle targeting sequence, experimental evidence for their subcellular localizations has not been demonstrated. To address this question, the subcellular localization of each AtIPI1 and AtIPI2 protein was investigated using GFP as a reporter. Stable transformants carrying IPI–GFP fusion constructs driven by the cauliflower mosaic virus (CaMV) 35S promoter were generated using RT–PCR-amplified cDNA fragments of the AtIPI1 and AtIPI2 genes (deposited in the RIKEN Arabidopsis full-length cDNA database, RAFL; http://rarge.gsc.riken.go.jp), and the plants were analyzed by confocal fluorescence microscopy. The GFP signal of AtIPI1–GFP was mostly cytosolic (Fig. 3A), whereas that of AtIPI2–GFP was mainly detected as small particles, which are most probably mitochondria (Fig. 3D). Staining with DiOC6 (3,3′-dihexyloxacarbocyanine iodide) indicated that there is no visible difference in the pattern of fluorescence between the stained cells of the wild-type plant and the AtIPI2-GFP transgenic plant, supporting the mitochondrial localization of the AtIPI2–GFP protein (Supplementary Fig. S1). Concerning AtIPI1, an extended mRNA of AtIPI1 was found that contains an alternative start codon 4 bp upstream of the 5′ end of the RAFL cDNA clone (RAFL04-20-N19), and which encodes a likely transit peptide for plastid localization (Fig. 4). In addition, several expressed sequence tag (EST) sequences that contain this alternative start codon of the AtIPI1 gene have been deposited on the database, although they are short fragments with truncated 3′ ends (see SeqViewer in The Arabidopsis Information Resource, TAIR, www.arabidopsis.org/servlets/s; locus ID, At5g16440.1). Therefore, we designated the version of AtIPI1 with the extended 5′-untranslated region (UTR) as AtIPI1L, and confirmed the localization. AtIPI1L–GFP was predominantly plastid localized, suggesting that AtIPI1 localizes to plastids only when AtIPI1L is fully transcribed (Fig. 3G, H). Because of this complexity, we further examined the subcellular localization of AtIPIs using transgenic plants that produce an IPI–GFP fusion protein under the control of the native promoter. Plasmid constructs used for these experiments were designated as AtIPI1g-GFP and AtIPI2g-GFP, which contain the genomic fragments of the loci of IPIs fused to the GFP gene. Whereas AtIPI2g–GFP was detected in the mitochondria of shoots and roots (Fig. 3E, F), AtIPI1g–GFP was mainly observed in the cytosol, but not in the plastid in shoots and roots (Fig. 3B, C). These observations suggest that IPIs exist mainly in the cytosol and mitochondria in Arabidopsis.
To examine further the plastid localization of the IPIs, chloroplast-enriched fractions prepared from the leaves of wild-type plants were subjected to immunoblot analysis with anti-IPI2 antibodies. We could verify the purity of the chloroplast fraction by confirming the presence and absence of the plastid- and cytosol-localized isoforms of glutamine synthetase (GS1, GS2), respectively, but no enrichment of IPI protein in the chloroplast fraction was observed (Fig. 3I). In contrast, deoxyxylulose reductoisomerase (DXR) of the MEP pathway (used as a positive control) was detected in the purified chloroplast fraction, consistent with the result reported by Carretero-Paulet et al. (2002). Based on these results, we conclude that in Arabidopsis, IPIs are not localized in plastids, at least not in photosynthetic organs and roots, unlike IPIs in tobacco and the Cinchona tree (Ramos-Valdivia et al. 1997, Nakamura et al. 2001).
IPI activity in the atipi1atipi2 double mutant
Loss-of-function mutants of AtIPI1 and AtIPI2 were used to investigate the contribution of IPIs to the production of isoprenoids via the MEP pathway in plastids and the MVA pathway in the cytosol. T-DNA-tagged mutants in AtIPI1 and AtIPI2, generated at the Salk Institute (SALK_006330 for AtIPI1) and the Torrey Mesa Research Institute (TMRI; SAIL_604 for AtIPI2), were obtained from the ABRC (Arabidopsis Biological Resource Center). Homozygous plants were then selected based on their resistance to antibiotic markers, and insertion of the T-DNA was confirmed by PCR using gene-specific primers and the T-DNA border primer. T-DNA insertions were identified at the junction between the third exon and the third intron in atipi1, and in the middle of the first exon in atipi2 (Fig. 5A).
mRNA expression of full-length AtIPI1 or AtIPI2 was determined by RT–PCR. As shown in Fig. 5B, amplification of the full-length genes failed to detect AtIPI1 or AtIPI2 mRNA expression. We also evaluated mRNA levels of AtIPI1 or AtIPI2 in each single mutant by QRT–PCR of the 3′-UTR of each gene, and found that AtIPI1 expression was 10 times lower in the atipi1 mutant than in wild-type plants, while AtIPI2 expression was about one-third that in wild-type plants (data not shown). These results suggested that no functional AtIPI1 and AtIPI2 mRNA is produced in the atipi1 and atipi2 mutant, respectively.
Immunoblot analysis using anti-AtIPI2 polyclonal antibodies was also employed to confirm the expression of each IPI in the mutant. Two bands corresponding to AtIPI1 and AtIPI2 were detected in wild-type plants, whereas only one of the two bands was present in each single mutant (the lower band in atipi1 and the upper band in atipi2; Fig. 5C). We generated an atipi1atipi2 double mutant by crossing the atipi1 and atipi2 single mutants as described in Materials and Methods, and confirmed that neither band was detected in the atipi1atipi2 double mutant. Using anti-DXR polyclonal antibodies, we confirmed that DXR expression was similar in all samples (Fig. 5C, lower panel). To verify further the lack of IPI in our mutants, we measured IPI activity using protein extracts of the ipi mutants. The ipi double mutant exhibited only 7% of the IPI activity in wild-type plants, whereas the single mutants exhibited 67 and 81% of the level of IPI activity in wild-type plants, respectively (Fig. 5D). Although IPI activity was substantially reduced in the double mutant (7% of the wild-type level), the presence of this small amount of IPI activity is significant relative to a negative control, which is the reaction using boiled protein extracts. In contrast to reduced IPI activity in the ipi mutants, similar levels of prenyltransferase activity were detected in these samples (Fig. 5E). Taken together, these results demonstrate that IPI activity is severely reduced in the atipi1atipi2 mutant. However, a small amount of functional IPI(s) is likely to be produced in the double mutant.
The pleiotropic phenotypes of the atipi1atipi2 double mutant
The atipi1 and atipi2 single mutants grown for 2 weeks on soil under continuous light were indistinguishable from wild-type plants in appearance (Fig. 6A), and did not show any visible phenotype throughout their life cycle (data not shown). In contrast, the atipi1atipi2 double mutant showed pleiotropic phenotypes when grown under the same conditions. The 2-week-old atipi1atipi2 double mutant was relatively small with reduced pigmentation (Fig. 6A), and variegated leaves emerged upon bolting (Fig. 6B). Small and curly rosette leaves were observed throughout development in the double mutant. Silique development was also arrested in the ipi double mutant (Fig. 6C), and the defect appeared to be caused by male sterility, because the mutant was able to be fertilized using wild-type pollen (data not shown). Five-week-old atipi1atipi2 double mutants exhibited wilting and dwarf phenotypes, but still had green leaves compared with wild-type plants (Fig. 6C). In addition, the pale green phenotype of the double mutant appeared only when the plants were grown under continuous light, but not under long-day (LD) conditions (Fig. 6D, E). In fact, when grown under continuous light, the amount of Chl a and lutein in the double mutant was about 80% that in wild-type plants, whereas the atipi1 and atipi2 single mutants showed almost no decrease in pigmentation during the vegetative phase (Table 1). Other phenotypes of the ipi double mutant were commonly observed under both continuous light and LD conditions.
|Wild type||356.2 ± 75.9||3.9 ± 0.2||1,026.3 ± 54.5||76.8 ± 12.9|
|atipi1||278.5 ± 15.6||3.9 ± 0.0||997.2 ± 111.8||67.9 ± 9.5|
|atipi2||232.7 ± 8.4||3.3 ± 0.0||955.2 ± 144.2||66.8 ± 9.2|
|atipi1atipi2||136.6 ± 16.5||1.8 ± 0.1||819.3 ± 67.3||54.0 ± 5.7|
|Wild type||356.2 ± 75.9||3.9 ± 0.2||1,026.3 ± 54.5||76.8 ± 12.9|
|atipi1||278.5 ± 15.6||3.9 ± 0.0||997.2 ± 111.8||67.9 ± 9.5|
|atipi2||232.7 ± 8.4||3.3 ± 0.0||955.2 ± 144.2||66.8 ± 9.2|
|atipi1atipi2||136.6 ± 16.5||1.8 ± 0.1||819.3 ± 67.3||54.0 ± 5.7|
a,cSeven-day-old seedlings grown under continuous light were harvested and analyzed for sterol and pigment contents using GC-MS.
bTwo-week-old seedlings grown under continuous light were harvested and the level of ubiquinone-9 was measured by HPLC. Values are the average of triplicate experiments ± SD.
To confirm whether these phenotypes are truly due to the lack of functional AtIPI1 and AtIPI2, a genetic complementation test was performed. AtIPI1 cDNA fused to the GFP gene driven by the CaMV 35S promoter was introduced into the atipi1 homozygous–atipi2 heterozygous mutant. Gentamycin-resistant T1 transformants were collected and their GFP signals from the transgene expression were checked by fluorescent microscopy (data not shown). The ipi double homozygous mutant plants harboring the AtIPI1-GFP transgene, which had been confirmed in its homozygosity by genomic PCR, were phenotypically indistinguishable from wild-type plants (Supplementary Fig. S2). These results indicate that IPIs of Arabidopsis are functionally redundant, and they are necessary for normal development under continuous light and LD conditions. The results also suggest that a defect in IPI activity conditionally affects the production of chloroplastidic pigments via the MEP pathway under the stressful circumstance where lighting is continuously provided to the plant.
Complementation of the male-sterile phenotype in the atipi1atipi2 double mutant
Suzuki et al. (2004) reported that male sterility in the hmg1-1 mutant, which is defective in 3-hydroxy-3-methylglutalyl coenzyme A (HMG-CoA) reductase activity in the MVA pathway, can be complemented by the exogenous application of squalene. Using the same experimental strategy, we investigated whether the male sterility of the atipi1atipi2 mutant was caused by the depletion of sterols produced in the MVA pathway. Treatment of flower buds of atipi1atipi2 plants with squalene overcame the male sterility of the mutants, enabling the production of mature seeds (Fig. 6F, G). In fact, in 7-day-old atipi1atipi2 seedlings, the level of sitosterol, which is widely distributed throughout the plant kingdom, was about 38% that in wild-type seedlings (Table 1). Thus, the male sterility of the mutant is likely to be caused by reduced sterol content, as was the case for hmg1-1. In addition, the atipi1atipi2 mutant showed a decrease in the level of ubiquinone-9 to 46% of that of the wild-type plant (Table 1), which is produced in mitochondria via the MVA pathway, indicating that IPIs significantly affect isoprenoid biosynthesis through the cytosolic MVA pathway.
Incorporation of [1-13C]DX and [2-13C]MVL into isoprenoids in the ipi double mutant
The subcellular localization of the GFP fusion proteins and the phenotypes of the ipi mutants suggested that Arabidopsis IPIs play a significant role in the production of MVA pathway-derived isoprenoids, but contribute little to the production of MEP pathway-derived isoprenoids under LD conditions. To examine the effect of the atipi mutation on the MVA and MEP pathways directly, we carried out feeding of [2-13C]mevalonolactone (MVL) or [1-13C]1-deoxy-d-xylulose (DX) to either wild-type or atipi1atipi2 seedlings. We fed [2-13C]MVL in the presence of mevastatin, a specific inhibitor of HMG-CoA reductase, and [1-13C]DX in the presence of ketoclomazone, an inhibitor of 1-deoxyxylulose 5-phosphate synthase, for efficient labeling of the metabolites by blocking the endogenous MVA or MEP pathway (Fig. 7A). We analyzed the level of 13C in campesterol and α-tocopherol as representative metabolites of the MVA and MEP pathways, respectively, using gas chromatrography–mass spectrometry (GC-MS; Fig. 7A).
GC-MS spectra of a trimethylsilyl (TMS) derivative of authentic non-labeled campesterol showed a molecular ion peak at m/z 472 (68%), with isotopomer ions at m/z 473 (28%) and 474 (4%; Fig. 7B, yellow bars). In the [2-13C]MVL-fed wild-type plants, campesterol was efficiently labeled with 13C atoms, and the GC-MS spectra showed a molecular ion at m/z 477 (60%), corresponding to five 13C atoms per molecule (Fig. 7B, green bars). In contrast, the spectra of a TMS derivative of campesterol from the [2-13C]MVL-fed ipi double mutant plants involved an ion cluster consisting of m/z 474 (12%), 475 (24%), 476 (30%), 477 (25%) and 478 (9%; Fig. 7B, red bars). These results indicate that the incorporation of [2-13C]MVL into campesterol was significantly decreased in the ipi double mutant.
The full-scan GC-MS spectral data for a TMS derivative of non-labeled α-tocopherol showed a molecular ion at m/z 502 (67%), with isotopomer ions at m/z 503 (27%) and 504 (6%; Fig. 7C, yellow bars). A TMS derivative of α-tocopherol from [1-13C]DX-fed wild-type seedlings showed a molecular ion peak at m/z 506 (60%), suggesting that α-tocopherol was efficiently labeled with four 13C atoms (Fig. 7A). We next analyzed α-tocopherol from [1-13C]DX-fed ipi double mutant plants and found that its TMS derivative also possessed a molecular ion peak at m/z 506 (58%), with isotopomer ion peaks at m/z 505 (15%) and 507 (24%; Fig. 7C, red bars). These results indicate that incorporation of [1-13C]DX into α-tocopherol via the MEP pathway was not significantly different between the wild-type and ipi double mutant plants. Altogether, our results support the idea that IPI contributes to isoprenoid biosynthesis mainly through the MVA pathway in the cytosol in Arabidopsis.
In this study, we analyzed phenotypic and metabolic features of Arabidopsis mutants with severely reduced IPI activity. The atipi1atipi2 double mutant exhibited pleiotropic phenotypes, including dwarfism, male sterility and reduced leaf pigmentation, and it contained only 38% of the sterols and 46% of ubiquinone present in wild-type plants, suggesting a critical role for IPI in the MVA pathway. This conclusion was further supported by our in vivo labeling experiments, in which the incorporation rate of [13C]MVA (fed as MVL) into sterols was significantly reduced in the atipi1atipi2 double mutant. In addition, the dwarf and male-sterile phenotypes of the atipi1atipi2 double mutant are similar to those of the hmg1-1 mutant, which is defective in HMG-CoA reductase of the MVA pathway. hmg1-1 and wild-type plants treated with mevastatin, a specific inhibitor of HMG-CoA reductase, have shortened roots. However, our ipi double mutant did not exhibit a change in root length (data not shown). The hmg1-1 mutant is likely to have a severe depletion of IPP (and DMAPP), whereas IPP synthesis in the ipi double mutant should be intact. Detailed metabolite profiling of the hmg1-1 and ipi mutant plants (especially in roots) may reveal how the difference in the IPP synthesis ability affects accumulation of various terpenoids and whether it would cause the phenotypic difference in the root. Regarding the hmg1-1 mutant, the HMG2 gene, which is a homolog of the HMG1 gene, could not fully compensate the defect of the HMG1 gene; the HMG2 gene is expressed only in meristematic and floral tissues, which may be the reason for the incomplete complementation of a reduced root length phenotype observed in the hmg1-1 mutant by the HMG2 gene (Enjuto et al. 1994).
The only previously reported loss-of-function analysis for IPI was performed in tobacco using a viral-mediated transient assay system. Page et al. (2004) found that Nicotiana benthamiana leaves in which IPI expression was down-regulated by tobacco rattle virus-mediated gene silencing exhibited a mottled white–pale green phenotype and an 80% reduction in the level of chlorophyll compared with control leaves. They concluded that although not absolutely required, IPI plays a significant role in plastidic isoprenoid biosynthesis in higher plants. In our study, the atipi1atipi2 double mutant conditionally showed a 20% decrease in chlorophyll and carotenoids compared with control plants under continuous light. Thus, a pale green phenotype seems to be a common feature of plants defective in IPI activity, although the extent of variegation varies significantly among plant species. However, the appearance of a pale green phenotype was largely absent from the Arabidopsis ipi double mutant under LD conditions, indicating that IPI activity is necessary for the production of photosynthetic pigments via the MEP pathway under conditions that necessitate an increased isoprenoid production. Although the pale green phenotype of the atipi1atipi2 double mutant suggests a role for IPIs in the MEP pathway, our results from 13C labeling experiments (performed under continuous light) indicated that the MEP pathway was not significantly affected in the double mutant, at least for the production of α-tocopherol (Fig. 7C). Therefore, we speculate that the pigmentation phenotype of the double mutant plants might be attributed to a secondary metabolic effect. Page et al. (2004) reported that IPI-repressed tobacco accumulated isopentenyl monophosphate and isopentenol in variegated tobacco leaves, presumably due to IPP accumulation (and endogenous phosphatase activity). We found that Arabidopsis seedlings grown on medium containing 0.8 mM isopentenol exhibited a severe albino phenotype and died at the seedling stage (data not shown). Additional studies are required to understand the mechanism by which the pale green phenotype is produced.
Our 13C labeling experiments show that the incorporation rate of [13C]MVL into campesterol is significantly reduced in the the atipi1atipi2 double mutant (Fig. 7B). However, the same experiments also indicate that the cytosolic MVA pathway is still functional for the production of sterols in the double mutant (Fig. 7B). Thus, the MVA pathway is likely to contribute, at least in part, to the accumulation of substantial levels of sitosterol and ubiquinone-9 in the double mutant (38 and 46% of wild-type levels, respectively; Table 1). Future studies are necessary to determine whether this is fully attributable to the leaky nature of the atipi1 and/or atipi2 mutation (Fig. 5) or if any other unknown mechanism is also responsible for the conversion of IPP to DMAPP in the cytosolic MVA pathway in the ipi double mutant.
In contrast to the atipi1atipi2 double mutant, neither the atipi1 nor the atipi2 single mutant showed any visible phenotype. This observation is interesting given that AtIPI1 is localized in the cytoplasm, whereas AtIPI2 is targeted to mitochondria. It is also intriguing to note that no drastic decrease was detected in the level of ubiquinone-9 in the mitochondria of the atipi2 single mutant, despite the loss of mitochondrial IPI. These observations indicate that cytoplasmic AtIPI1 is sufficient for the synthesis of isoprenoids in mitochondria, and they suggest that IPP, DMAPP and other prenyl diphosphates such as FPP are able to move between the mitochondria and the cytosol. The mitochondrial uptake of an isoprenoid precursor was previously reported by Lutke-Brinkhaus et al. (1984); the exogenously applied [1-14C]IPP is successfully incorporated into the polyprenyl chain of the ubiquinone by penetrating membranes of intact mitochondria.
Our results indicate that neither AtIPI1 nor AtIPI2 is mainly localized to plastids. These results do not agree with a previous report by Nakamura et al. (2001) who showed that IPI1 and IPI2 from N. tabacum were localized in the cytosol and plastids, respectively. Other localization studies of plant IPIs also suggested that IPIs are localized in various subcellular compartments and that the localization patterns vary among different plant species (Ramos-Valdivia et al. 1997). The biological meaning of this species dependency is currently unknown, but the different localizations of the Arabidopsis and tobacco IPIs may reflect different contributions of each IPI to the biosynthesis of isoprenoids via the MEP pathway in plastids; the effect of the reduced level of IPI activity on pigmentation was much greater in tobacco (80% decrease) than in the Arabidopsis (20% decrease in the atipi1atipi2 double mutant) under continuous light.
In summary, our study shows that IPI proteins are mainly localized in the cytosol and mitochondria in Arabidopsis, and that they play important roles in the production of isoprenoids via the MVA pathway. We also found that the effect of IPI deficiency on the MEP pathway in plastids was conditional: it was observed under continuous light, but not under LD conditions. Further analysis using Arabidopsis mutants involved in either the MEP or MVA pathway in combination with the ipi mutants presented here may help explain how early isoprenoid precursors are produced and move into distinct organelles where isoprenoid biosynthesis is individually controlled.
Materials and Methods
Plant material and growth conditions
Arabidopsis thaliana ecotype Columbia (Col-0) was used in this study. Seeds for T-DNA insertion mutants (Salk_006330 for AtIPI1, SAIL_604 for AtIPI2) were obtained from the ABRC (Columbus, OH, USA) and from the TMRI (San Diego, CA, USA). Sterilized seeds were imbibed on Murashige–Skoog (MS) medium (pH 5.7) containing 1% sucrose and 0.8% agar, placed in a 4°C chamber for 1 d. All plants used in this study were allowed to germinate and grow at 22°C under continuous cool-white illumination. Photoperiodical illumination in LD conditions (16 h light/8 h dark) was only used in the experiment comparing the phenotypes of ipi mutants. Seedlings were used directly or transplanted in ProMix (Premier Horticulture Inc., Red Hill, PA, USA) and grown under the same conditions for experiments.
Generation of the atipi1atipi2 double mutant
To generate an atipi1atipi2 double mutant, the homozygous single mutants were crossed using the atipi2 mutant (BASTA resistant) as a pollen donor and the atipi1 mutant (kanamycin resistant) as a recipient. In the F1 population, BASTA-resistant plants were selected to rule out unexpectedly self-pollinated seeds. After selecting atipi1 homozygous plants that exhibited BASTA resistance, atipi1atipi2 double homozygous plants were selected in the next generation. The homozygosity of the atipi1atipi2 double mutant was confirmed by genomic PCR with a primer set encompassing T-DNA insertion sites: for atipi1, IPP1TRS 5′-TCGATGAGCTCGGTATTGTAGC-3′ and IPP1ANT 5′-AGGAATAGATTTTGTTGCCCAG-3′; for atipi2, IPP2SXb 5′-ATGTCTGCTTCTTCTTTATTTAATCTC-3′ and IPP2ANT2 5′-GTAAGTGTCTCACATATCCC-3′.
For complementation analysis, the AtIPI1L-GFP fragment obtained from pGWB5-IPI1L (see below; plasmid construction) in pPZPY122 (Yamamoto et al. 1998) was used to transform atipi1 homozygous–atipi2 heterozygous plants by Agrobacterium tumefaciens-mediated transformation using the floral dip method (Clough and Bent 1998), and gentamycin-resistant T1 transformants were self-pollinated to select atipi1atipi2 double homozygotes in the T2 generation. The homozygosity of the ipi double mutant was confirmed by genomic PCR as described above.
RNA isolation and quantitative RT–PCR
Total RNA was extracted from separated organs (flower, stem, cauline leaves, rosette leaves and root) in 4-week-old plants using an RNeasy plant total RNA isolation kit (Qiagen, Valencia, CA, USA). A 1 μg aliquot of total RNAs treated with RQ1 RNase-free DNase (Promega, Madison, WI, USA) was used for first-strand cDNA synthesis with poly(T) oligo primer using a SuperScriptII reverse transcriptase (Gibco-BRL, Cleveland, OH, USA). Quantitative real-time PCR using PowerSYBR Green PCR Master Mix (Applied Biosystems, Foster City, CA, USA) was employed to determine transcript levels using the first-strand cDNA as a template on a sequence detector system (model 7700; Applied Biosystems). The primers AtIPI1Q-S 5′-TGGGATCATGTTGAGAAAGGAAC-3′ and AtIPI1Q-A 5′-GTTGCCCAGTTTTGTCTGTAATCA-3′, AtIPI2Q-S 5′-GAGAAAGGAACTTTGGTTGAAGC-3′ and AtIPI2Q-A 5′-GTTTTGTAAGTGTCTCACATATCCC-3′, all of which can specifically amplify the 3′-UTRs of each AtIPI1 and AtIPI2 gene, were used for QRT–PCR. To calculate the mRNA level of the AtIPI1 and AtIPI2 genes, the copy numbers of the AtIPI1 and AtIPI2 mRNAs were determined by generating a standard curve using a series of known concentrations of the target sequence. For normalization across samples, the Arabidopsis ACT2 gene was amplified with the primers AtACT2-S 5′-CAGTGGTCGTACAACCGGTATTG-3′ and AtACT2-A 5′-TGGTGAACATGTAACCTCTCTCTGTAA-3′, and used as an internal standard.
To construct GFP-fused AtIPI1 and AtIPI2 genes, the full-length cDNAs were amplified by RT–PCR. The full-length Arabidopsis AtIPI1, AtIPI1L and AtIPI2 cDNAs were amplified from the first-strand cDNA by consecutive PCRs using the primers AtIPI1-S 5′-CACCTACTGCTTCACTATTTAGCT-3′ or AtIPI1L-S 5′-CACCATGTCTACTGCTTCACTA-3′ and AtIPI1-A 5′-GAGCTTGTGAATGGTTTTCATG-3′ (for AtIPI1 or AtIPI1L), and AtIPI2-S 5′-CACCATGTCTGCTTCTTCTTTATTTAATCTC-3′ and AtIPI2-A 5′-GAGTTTGTGGATGGTTTTCATG-3′ (for AtIPI2). These primers were also used in the RT–PCR analysis of full-length IPI gene expression. Genomic fragments of AtIPI1 and AtIPI2 genes including promoter regions were amplified by PCR with genomic DNA prepared from 12-day-old Arabidopsis seedlings using the primers of AtIPI1g-S 5′-CACCGATCCCTAGTCTCTTGACG-3′ or AtIPI2g-S 5′-CACCAAAACCGCAGGCTGCTA-3′ in combination with the antisense primer AtIPI1-A (for the AtIPI1 genomic fragment) or AtIPI2-A (for the AtIPI2 genomic fragment). The amplified fragments were directly cloned into pENTR/D/TOPO vector (Invitrogen, Carlsbad, CA, USA) and sequenced. The resulting plasmids were designated as pEN-IPI1, pEN-IPI1L, pEN-IPI2, pEN-IPI1g and pEN-IPI2g, respectively. These plasmids were subjected to LR reaction using Gateway technology (Invitrogen) to yield GFP-fused constructs using binary vector pGWB4 or pGWB5. The resulting plasmids designated as pGWB5-IPI1, pGWB5-IPI1L, pGWB5-IPI2, pGWB4-IPI1g and pGWB4-IPI2g were used for transformation of Arabidopsis by the Agrobacterium-mediated floral dipping method (Clough and Bent 1998). Transgenes introduced from pGWB4-IPI1g and pGWB4-IPI2g were driven by the promoters of the AtIPI1 and AtIPI2 gene, and transgenes from three other constructs using the pGWB5 vector were constitutively driven by the CaMV 35S promoter.
For preparing anti-AtIPI2 polyclonal antibodies, recombinant AtIPI2 protein fused with a hexa-histidine tag was purified from E. coli cells transformed with plasmid pQE31-IPI2 using metal chelating chromatography with Ni-NTA agarose (Qiagen). Polyclonal antibodies for AtIPI2 were raised in a rabbit injected with purified recombinant His-AtIPI2 protein (Qiagen). For immunoblot analysis, crude protein extracts from wild-type and ipi mutant Arabidopsis tissues were obtained by grinding in ice-cold homogenization buffer [0.1 M Tris–HCl (pH 7.5), 1 mM EDTA, 1 mM phenylmethylsulfonyl fluoride (PMSF), 1% (v/v) proteinase inhibitor cocktail (Sigma, St Louis, MO, USA)]. Protein concentration was determined using a protein reagent (Bio-Rad Laboratories, Hercules, CA, USA) according to the manufacturer's instructions. Proteins were subjected to SDS–PAGE on 12% (w/v) polyacrylamide gels and either stained with Coomassie blue or transferred to a Hybond P membrane (Amersham, Piscataway, NJ, USA). Arabidopsis AtIPIs, DXR and GSs were detected with the anti-AtIPI2 serum [dilution 1 : 1,000 (v/v)], anti-DXR serum [dilution 1 : 1,000 (v/v)] and anti-GS serum [dilution 1 : 1,000 (v/v)], respectively, as the primary antibody, and anti-rabbit immunoglobulin horseradish peroxidase conjugate [Amersham, dilution 1 : 25,000 (v/v)] as the secondary antibody. Chemiluminescent detection was carried out with the ECL plus system (Amersham), following the recommendations of the supplier.
Metabolite extraction and analyses
For ubiquinone analyses, 2-week-old plants (∼1 g) were homogenized with a mortar and pestle in 100% MeOH (15 ml). A 15 ml aliquot of Tris/NaCl buffer [50 mM Tris–HCl (pH 7.5), 1 M NaCl] was added to the MeOH extract, and then the MeOH–Tris/NaCl mixture was partitioned against chloroform (10 ml) three times. The combined chloroform fraction was evaporated to dryness, and dissolved in 5 ml of n-hexane. The n-hexane fraction was subjected to SiO2 gel column chromatography (column size, 5 × 2 cm) and eluted with 15 ml of 95, 94 and 85% n-hexane/ether. The 85% n-hexane/ether elute was evaporated to dryness, and dissolved in 1 ml of 100% EtOH. Then 0.1 ml of the 100% EtOH fraction was analyzed by HPLC with ethanol as the solvent (Crane and Barr 1971).
For α-tocopherol and campesterol analyses, whole seedlings (∼1 g) were pulverized with a mortar and pestle under liquid N2, and a fraction of powdered tissues (∼50 mg) was extracted with 80% acetone/H2O (250 μl). After centrifugation at 15,000×g for 5 min, the organic layer was transferred and diluted with H2O (800 μl), re-extracted with dichloromethane (200 μl; ×2), and the organic solvent was evaporated in vacuo. After chemical modification with N-methyl-N-trimethylsilyltrifluoroacetamide, the TMS derivative of α-tocopherol was analyzed by GC-MS. GC-MS analysis was performed on a GC-mate II mass spectrometer (JEOL, Tokyo, Japan) connected to an Agilent 6890 series GC system with a 15 m × 0.25 mm capillary column DB-1 (0.25 μm film thickness, J&W Scientific), with helium as the carrier gas. The column temperature program was 80°C for 1 min, to 245°C at 30°C min−1, to 280°C at 5°C min−1 and then 280°C for 1 min. The mass spectra were obtained at 70 eV at 250°C.
For the quantitative analysis of sitosterol, 7-day-old seedlings (∼30 mg) were placed in a 1.5 ml centrifuge tube and ground with a hand-held homogenizer in 80% acetone (250 μl) involving D6-cholesterol (Sigma-Aldrich, USA) as internal standard. After extraction, sitosterol was derivatized and analyzed by GC-MS following the same protocol used for campesterol analysis. The endogenous level of sitosterol was determined as the ratio of the peak areas of the molecular ion for the endogenous sitosterol and for the internal standard.
For lutein analysis, 7-day-old seedlings (∼30 mg) were placed in a 1.5 ml centrifuge tube and ground with a hand-held homogenizer in 80% acetone (200 μl) involving apocarotenal as internal standard. Ethyl acetate (120 μl) was added and the mixture was further homogenized. H2O (150 μl) was added, and the mixture was vortexed, then centrifuged at 15,000×g for 5 min. The organic layer was transferred to a new 1.5 ml centrifuge tube and the solvent was evaporated in vacuo. The extract was resuspended in dimethylsulfoxide (DMSO) and analyzed by HPLC. HPLC analysis was performed using an HPLC system (Waters 600 controller) on a reverse-phase ODS column (Waters symmetry shield C18, 15 cm × 4.6 mm) with a 30 min gradient of ethyl acetate (0–100%) in acetonitrile : H2O (9 : 1, v/v) at a flow rate of 1 ml min−1. Lutein was detected at 432 nm by a photodiode array detector (Waters 2996). The level of lutein was determined as the ratio of the peak areas for the lutein and for the internal standard.
Chlorophylls were extracted from 0.5 g of seedlings and quantified according to the method of Lichtenthaler (1987).
Chemicals for tracer experiment
[1-13C]DX (95% labeled) was synthesized as previously reported (Kasahara et al. 2004). dl-MVL and [2-13C]MVL (99% labeled) were purchased from Aldrich (St Louis, MO, USA), and mevastatin was from Sigma.
Feeding of [1-13C]DX to 5-ketoclomazone-treated plants
Wild-type and the ipi double mutant seedlings were grown for 5 d on MS agar medium, then transferred to MS liquid medium (15ml). Immediately after the transfer, 5-ketoclomazone (final concentration 1 μM) and [1-13C]DX (filter-sterilized H2O solution) were added aseptically to the liquid culture. Ten days after cultivation under continuous light, seedlings were harvested, frozen by liquid N2, and stored at −80°C until use. α-Tocopherol was analyzed as descried above.
Feeding of [2-13C]MVL to mevastatin-treated plants
Wild-type and the ipi double mutant seedlings were prepared as mentioned above. Immediately after the transfer to the MS liquid medium, mevastatin (MeOH solution, final concentration 10 μM) and [2-13C]MVL (filter-sterilized H2O solution) were added aseptically to the liquid medium, and the plants were grown for10 d. Campesterol was analyzed as descried above.
IPP isomerase and prenyltransferase assays
Crude enzymes prepared from wild-type and ipi mutants seedlings, which had also been used in immunoblot analysis, were subjected to IPI enzyme assay as reported by Kaneda et al. (2001). Prenyltransferase activitiy in the crude enzymes was also measured using [1-14C]IPP (2.22 GBq/mmol; Amersham International, Buckinghamshire, UK) and DMAPP (Sigma) as substrates, and prenyl diphosphates produced in the reaction were detected as their alcohol forms by reversed phase thin-layer chromatography (TLC) as reported previously (Okada et al. 2000).
Supplementary material mentioned in the article is available to online subscribers at the journal website www.pcp.oxfordjournals.org.
A Grant-in-Aid for Young Scientists (B) (No. 15780068 to K.O.); a grant for Basic Science Research Projects from the Sumitomo Foundation (No, 020452 to K.O.); the Program for the Promotion of Basic Research Activities for Innovative Biosciences (PROBRAIN) to H.Y.
We thank Dr. Tsuyoshi Nakagawa (Shimane University) for providing the pGWB series vectors, Dr. Tadao Asami (University of Tokyo) for providing 5-ketoclomazone, Dr. Hitoshi Sakakibara (RIKEN) for providing anti-GS antibodies, and Dr. Tomohisa Kuzuyama (University of Tokyo) for critical reading of the manuscript. We also thank the ABRC and TMRI for providing ipi mutants.
Arabidopsis Biological Resource Center
cauliflower mosaic virus
gas chromatography–mass spectrometry
green fluorescent protein
3-hydroxy-3-methylglutalyl coenzyme A
isopentenyl/dimethylallyl diphosphate isomerase
quantitative reverse transcription–PCR
Torrey Mesa Research Institute