INTRODUCTION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Tocopherols are important lipophilic antioxidants that are synthesized by photosynthetic organisms. These include higher plants and certain eukaryotic algae where they are synthesized in the plastids, as well as photosynthetic prokaryotes such as blue-green algae. The four major forms of tocopherols, , , , and , differ in the position and number of methyl groups (Fig. 1). The predominant form in the leaves of higher plants is -tocopherol, whereas in seeds -tocopherol is often the major isoform (Tan, 1989; Demurin et al., 1996). Tocopherols predominantly function as antioxidants in vivo in photosynthetic organisms and in animals, as well as in isolated compounds such as oils. The antioxidant properties of tocopherols derive from their ability to quench free radicals and different tocopherols may be optimal as antioxidants for different biological systems. For human and animal utility, -tocopherol has the highest vitamin E activity and has been implicated in a variety of health areas, including possible benefits in preventing cardiovascular disease, certain cancers, and cataract formation (DellaPenna, 1999; Bramley et al., 2000). The amounts of vitamin E needed to achieve these effects are often quite high, 100 to 400 International Units (I.U.) and even up to 800 I.U. compared with the recommended daily allowance of 40 I.U. In fats and oils, tocopherols protect unsaturated fatty acids from oxidation. In these systems, -tocopherol appears to have the greater utility (Parkhurst et al., 1968; Chow and Draper, 1974; Gottstein and Grosch, 1990). In fact, tocopherols are often included in processed oils to help stabilize the fatty acids. For human health as well as food and feed utility, it is desirable to have plants with increased tocopherol content along with those where the tocopherol composition is customized.

View larger version (17K): [in this window] [in a new window]   Figure 1.   Tocol and tocopherol structure.

Tocopherols contain an aromatic head group, which is derived from homogentisic acid (HGA) and a hydrocarbon portion, which arises from phytyldiphosphate (phytyl-DP). HGA is derived from the shikimic acid pathway and phytyl-DP is generated from the condensation of four isoprenoid units. The isoprenoid contribution to tocopherol biosynthesis is thought to come primarily from the plastidal methyl-erythritol phosphate pathway, and not the cytosolic mevalonic acid pathway (Arigoni et al., 1997; Lichtenthaler et al., 1997). The condensation of HGA and phytyl-DP to form 2-methyl-6-phytylplastoquinol, the first committed step in tocopherol biosynthesis, is a prenyltransferase reaction that is performed by a homogentisate phytyltransferase (HPT; Fig. 2). Subsequent cyclization and methylation reactions (Soll et al., 1980; Fiedler et al., 1982; Marshall et al., 1985) result in the formation of the four major tocopherols (Fig. 1). The enzymatic reactions in tocopherol biosynthesis were identified 15 to 20 years ago (Soll et al., 1980; Schultz-Siebert et al., 1987), but cloning of the genes encoding these enzymes has only occurred in the last few years.

View larger version (18K): [in this window] [in a new window]   Figure 2.   Tocopherol biosynthetic pathway.

Tocopherol biosynthesis takes place in the plastid and the enzymes are associated with the chloroplast envelope (Soll et al., 1980, 1985). The membrane association of the enzymes has made purification difficult (Soll et al., 1980, 1985; Camara and d'Harlingue, 1985). With the advent of genomics and the availability of complete genome sequences of a number of organisms, including Synechocystis sp. PCC 6803 and Arabidopsis, it has become possible to use bioinformatics techniques to identify and clone additional genes in the tocopherol pathway.

The first enzyme cloned in the tocopherol pathway, -tocopherol methyl transferase (-TMT), was identified in Synechocystis sp. PCC 6803 and Arabidopsis using bioinformatics (Shintani and DellaPenna, 1998). In that study, the Arabidopsis -TMT was shown to alter seed tocopherol composition when overexpressed in Arabidopsis. Tocopherol, normally the predominant tocopherol isomer in Arabidopsis seeds, was almost completely converted to -tocopherol.

HPT catalyzes the first committed reaction in the tocopherol pathway, and was unidentified previously. Concomitant with this study, slr1736 was found to encode a HPT in Synechocystis sp. PCC 6803 (DellaPenna et al., 2000; Savidge et al., 2000; Collakova and DellaPenna, 2001; Schledz et al., 2001) and the Arabidopsis HTP was identified (DellaPenna et al., 2000; Savidge et al., 2000; Collakova and DellaPenna, 2001).

There are prenyltransferases that condense prenyl groups with allylic chains and those that condense prenyl chains with aromatic groups. The prenyltransferases that catalyze sequential condensations of isopentenylpyrophosphate with allylic chains share common features, including Asp-rich motifs, and lead to the formation of compounds with two isoprenoid units, such as geranylpyrophosphate, or to much longer molecules, such as rubber, which contains greater than 1,000 isoprenoid units (Chen et al., 1994; Ogura et al., 1997). Prenyltransferases that catalyze condensations with nonisoprenoid groups have an Asp-rich motif (Saiki et al., 1993) distinct from that of the allylic class (Ashby and Edwards, 1990; Carottoli et al., 1991; Marrero et al., 1992), and include UbiA, which attaches a prenyl group to 4-hydroxybenzoic acid, and chlorophyll synthase, which attaches a prenyl group to chlorophyllide (Melzer, 1994; Oster and Rudiger, 1997; Oster et al., 1997).

The first committed step in tocopherol biosynthesis is catalyzed by an aromatic prenyltransferase that transfers a phytyl chain to HGA. Assuming that structural features are shared among aromatic prenyltransferases, bioinformatics techniques were used to identify candidate genes encoding HPT in both Synechocystis sp. PCC 6803 and Arabidopsis, which were then characterized by biochemical and genetic studies. Furthermore, the current study provides evidence that seed-specific expression of HPT1 increases tocopherol levels 2-fold in Arabidopsis seed, an important first step in increasing tocopherol levels in the feed and food supply.

	RESULTS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Position-Specific Iterative (PSI)-BLAST Analysis

The HPT is thought to catalyze both the condensation of phytyl-DP with HGA and the decarboxylation of HGA, resulting in the formation of 2-methyl-6-phytylplastoquinol (Fig. 2). To identify the gene encoding this enzyme, a PSI-BLAST profile (Altschul et al., 1997) was generated using the Escherichia coli 4-hydroxybenzoate-octaprenyltransferase (ubiA, Gen-Bank accession no. 1790473) amino acid sequence as a query. UbiA was chosen for profile generation because the reaction catalyzed by this enzyme, the prenylation of 4-hydroxybenzoic acid to form 3-octaprenyl-4-hydroxy-benzoic acid, closely resembles that of the HGA phytyltransferase. The PSI-BLAST profile was used to search the Synechocystis sp. PCC 6803 genome (available at CyanoBase, http://www.kazusa.or.jp/cyano/index.html) for genesthat may encode a prenyltransferase belonging to the aromatic type. The search resulted in the identification of five candidate genes: slr1736, slr0926, sll1899, slr0056, and slr1518 (Table I).

A parallel PSI-BLAST search was performed on the Arabidopsis public database (TBLASTN) and resulted in the identification of four putative prenyltransferase genes that appeared to be potential orthologs of genes identified in the Synechocystis sp. PCC 6803 search (Table I). These sequences were originally designated ATPT2 (Arabidopsis prenyltransferase 2), ATPT3, ATPT4, and ATPT12 (GenBank accession nos. 3004556, 4454035, 3341672, and 2129675, respectively). However, after functional characterization, ATPT2 was renamed as HPT1 and will be referred to as such for the remainder of the paper. Although Slr1736 and its putative ortholog, HPT1, are annotated as unknown proteins, they do share low levels of similarity to chlorophyll synthase (ATPT12; Gaubier et al., 1995). Chlorophyll synthase catalyzes a reaction similar to that of the HPT, adding phytyl-DP to chlorophyllide. After obtaining 5' sequences of the four putative prenyltransferases sequences by RACE, only HPT1 and ATPT12 were strongly predicted to encode plastid targeted proteins (PSORT). Based on predicted targeting information and the observation that ATPT12 is likely chlorophyll synthase, HPT1 and the putative Synechocystis sp. PCC 6803 ortholog, slr1736, were the strongest candidates to encode HPT.

To determine if HPT1 and slr1736 encode HPT, isolation and functional testing of both genes were pursued. Primers for slr1736 were designed based on the published sequence (CyanoBase, http://www.kazusa.or.jp/cyanobase/). The sequence of the PCR product was verified and the product was cloned into appropriate vectors for functional testing. The full-length Arabidopsis HPT1 cDNA was isolated (GenBank accession no. AY089963) by performing 5' and 3' RACE based on a partial expressed sequence tag sequence. HPT1 encodes a 44-kD protein with 393 amino acids. Comparison of the full-length clone to the public predicted protein from Arabidopsis genomic sequence (Arabidopsis public database) revealed that the public predicted sequence lacked 110 amino acids in the amino terminus and 17 amino acids from the carboxy terminus. This inaccurate prediction is likely attributable to the fact that HPT1 contains 13 exons, most of which are only 100 to 130 bp long. The public predicted protein did not contain a plastid targeting sequence, and had only low levels of similarity to another prenyltransferase, chlorophyll synthase, making it difficult to infer function until the full-length sequence was obtained. Both Slr1736 and HPT1 have predicted transmembrane domains, based on Kyte-Doolittle hydropathy plotting (Kyte and Doolittle, 1982), suggesting that they are membrane proteins. Using ChloroP (Emanuelsson et al., 1999; http://www.cbs.dtu.dk/services/ChloroP/), the putative plastid transit peptide cleavage site of HPT1 is between amino acids 36 and 37; however, an alignment of HPT1 and Slr1736 (Fig. 3A) suggested that the transit peptide could be considerably longer. ClustalW was used to align Slr1736 and HPT1, revealing 37% identity overall, and several regions with higher levels of identity. Site-directed mutagenesis of another prenyltransferase of the aromatic type, heme-O-synthase (Saiki et al., 1993), resulted in the identification of a putative catalytic domain. The region corresponding to the putative catalytic domain was compared among Slr1736, HPT1, other Arabidopsis putative prenyltransferases, as well as UbiA, which was used as the query in the original PSI-BLAST search (Fig. 3B). Amino acid residues identified as essential for catalytic activity in heme-O-synthase are conserved among these proteins, suggesting that this region may also play a role in catalysis.

View larger version (38K): [in this window] [in a new window]   Figure 3.   Protein alignment of HPT1 and Slr1736 and conserved putative catalytic domain. A, Alignment of HPT1 and Slr1736 using the ClustalW algorithm. B, Alignment of putative catalytic domain of Slr1736 and Arabidopsis putative prenyltransferases identified using UbiA as a query in a PSI-BLAST search. Gray shading indicates identity. Underlined sequence corresponds to the putative catalytic domain. Asterisks indicate amino acids required for catalytic activity in heme-O-synthase, which are conserved in HPT1. The putative chloroplast target peptide processing site is indicated by an arrowhead.

Synechocystis sp. PCC 6803 slr1736 Null Mutant and Complementation with HPT1

To determine if slr1736 plays a role in tocopherol biosynthesis, a null mutant in Synechocystis sp. PCC 6803 was generated in the slr1736 ORF via insertion of the nptI gene. A confirmed mutant strain was assayed for tocopherol content and composition using HPLC. No tocopherols were detected in the null strain (Fig. 4), suggesting ORF slr1736 is essential for tocopherol biosynthesis. A complementation experiment was performed with the Synechocystis sp. PCC 6803 slr1736 null mutant to determine if HPT1 is the ortholog of slr1736. A vector, pMON21690, containing HPT1 under control of the Tac (Russell and Bennet, 1982) promoter, was transformed into the slr1736 null strain and grown under standard conditions for 5 d. HPLC analysis demonstrated that tocopherol biosynthesis was restored in this strain (Fig. 4C), thus confirming that HPT1 and slr1736 encode proteins of similar function.

View larger version (20K): [in this window] [in a new window]   Figure 4.   HPLC chromatographic analysis of ethanol extracts of wild-type Synechocystis sp. PCC 6803, slr1736 null mutant, and slr1736 null mutant complemented with HPT1. A, Wild-type Synechocystis sp. PCC 6803 transformed with control vector. Peak 4 corresponds to -tocopherol at 4.6 min. B, slr1736 null mutant lacking tocopherol peak at 4.6 min. C, slr1736 null mutant complemented with Arabidopsis HPT1 with tocopherol signal apparent at 4.6 min. A compound eluting at 8.5 min represents the tocol internal standard (peak 5). Peak 1 corresponds to the solvent front, and peaks 2 and 3 are two unknown compounds.

Phytyltransferase Activity of the slr1736 Gene Product

The reaction carried out by HPT results in the formation of 2-methyl-6-phytylplastoquinol from the condensation of HGA and phytyl-DP (Fig. 2). To confirm that the slr1736 gene product catalyzed this reaction, slr1736 was expressed in the Baculovirus Expression System (Invitrogen, Carlsbad, CA) and assayed for activity (Table II). The membrane fraction from slr1736 expressing Sf9 cells was able to catalyze the phytylation of HGA to generate 2-methyl-6-phytylplastoquinol, whereas membrane fractions from Sf9 control cells showed no conversion (Table II). In addition, HPT activity was detected in membrane fractions from Synechocystis sp. PCC 6803 wild type, but not in the null strain (Table II; Fig. 5). The ability of the slr1736 gene product to catalyze this reaction, combined with the genetic data described above, provides strong evidence that slr1736 encodes a phytyltransferase involved in tocopherol biosynthesis. Membrane fractions from spinach chloroplasts were used as a positive control in the phytyltransferase assay.

View larger version (18K): [in this window] [in a new window]   Figure 5.   HPT activity of Synechocystis sp. PCC 6803 wild-type and slr1736 null mutant membrane fractions. A, Formation of [3H]2-methyl-6-phytylplastiquinone in membrane fractions of Synechocystis sp. PCC 6803 wild-type cells from [3H]HGA and phytyl-DP. B, Lack of accumulation of [3H]2-methyl-6-phytylplastiquinone formation by membrane fractions from Synechocystis sp. PCC 6803 slr1736 null mutants.

HPT1 Expression in Arabidopsis Seed

Because phytylation of HGA is the first committed step of tocopherol biosynthesis, it was hypothesized that expression of HPT1 in Arabidopsis may result in increased tocopherol levels. To test this hypothesis, we expressed HPT1 under the seed-specific napin promoter. Tocopherol analysis of pooled segregating T₂ seed from the pNapin::HPT1 sense lines indicated that expression of HPT1 resulted in up to a 60% increase in total seed tocopherol (Fig. 6). Statistical evaluation of tocopherol data from pNapin::HPT1 events compared with controls revealed that 33 of 36 independent events produced elevated tocopherol levels. The analysis of T₃ seed pools from three selected events (1,848, 1,860, and 1,863) demonstrated that the transgenic populations were distinct (P < 0.001) from wild-type and vector controls (Table III), further validating the increased tocopherol phenotype. Whereas homozygous lines, identified by kanamycin selection, generally produced the highest tocopherol content (up to a 2-fold increase), only events 1,848 and 1,863 showed significant differences in tocopherol content between homozygous and hemizygous in T₃ seed populations. Based on their kanamycin selection pattern, these three lines were also determined to have a single insert. Genomic PCR using gene-specific and napin promoter-specific primers demonstrated that T₃ plants from the three events, 1,848, 1,860, and 1,863, all contain the HPT1 transgene (data not shown).

View larger version (65K): [in this window] [in a new window]   Figure 6.   Tocopherol content of T2 seed from pNapin::HPT1 Arabidopsis plants. A, Total tocopherol levels (ng mg1 seed) in individual pools of segregating T2 seed derived from 36 independent transgenic events containing the pCGN10822 construct (pNapin::HPT1) compared with vector (VC) and wild-type (WT) control populations. B, Total tocopherol levels (ng mg1 seed) in individual pools of T2 seed derived from 86 independent transgenic events harboring the pCGN10803 construct (e35S::HPT1antisense) are compared with control populations. Error bars on control samples represent the 95% confidence interval with the sample size indicated as n. The gray bar in the background includes the 95% confidence interval of both controls. Seed tocopherol levels of wild-type (), vector control (), and T2 transgenic lines ().

To determine if HPT1 is required for tocopherol accumulation in Arabidopsis seed, an antisense expression construct under control of the enhanced cauliflower mosaic virus e35S promoter was tested. Tocopherol levels were assayed in pooled segregating T₂ seed from 88 independent transformation events. Seed tocopherol levels of 19 events fell outside of the lower limit of the 95% confidence interval for the controls, indicating that HPT1 is necessary for tocopherol biosynthesis. Two events (1,393 and 1,401) with significantly reduced tocopherol levels were carried forward to the T₃ generation, revealing up to a 10-fold decrease in total tocopherols in some individual T₃ pools (data not shown). Statistical analysis of T₃ seed pools from these lines demonstrated that the transgenic populations were distinct from wild-type and vector controls, further validating the reduced tocopherol phenotype (Table III). Genomic PCR using gene-specific and e35S promoter-specific primers demonstrated that T₃ plants from the two events, 1,393 and 1,401, both contain the HPT1 transgene in the antisense orientation (data not shown). Because of the low HPT enzyme activity observed in chloroplast preparations (Table II), enzyme assays were not performed on HPT1-expressing or HPT1-antisense seed samples. The whole plant phenotype of T₂ antisense lines did not differ substantially from the wild type. These combined sense and antisense data further confirm that HPT1 encodes an HGA phytyltransferase, and show that it is possible to alter tocopherol levels in seed using this gene.

	DISCUSSION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Vitamin E is comprised of a mixture of various tocopherols, with -tocopherol being the most bioactive (Sheppard and Pennington, 1993). Tocopherols are naturally occurring micronutrients, produced in plants and cyanobacteria. Many studies have demonstrated that the antioxidant activity of these molecules has the potential to positively impact human and animal health. Therefore, tocopherols are valuable micronutrients and there is consequently interest in developing plants that produce high levels of natural tocopherols. One possible strategy to elevate tocopherol levels is to increase flux through the pathway by overexpressing the enzyme that catalyzes the first committed step in tocopherol biosynthesis, the HPT.

Evidence that slr1736 is involved in tocopherol biosynthesis in Synechocystis sp. PCC 6803 has been reported recently (Schledz et al., 2001). In that study, an slr1736 deletion mutant produces reduced levels of tocopherol, and based on a colorimetric assay, accumulates HGA. Other prenyllipids including phylloquinones, plastoquinones, and carotenoids are not affected in this mutant. However, slr1736 occurs in an operon-like structure with a downstream gene (slr1737). Even though the amino acid sequence of slr1736 has similarity to known prenyltransferases, complementation analysis of these genes in the deletion mutant, or an enzyme assay of a recombinant expressed gene, is necessary to verify function and that loss of Slr1736 activity is responsible for the mutant phenotype. Here, the isolation and characterization of HPT from Synechocystis sp. PCC 6803 (slr1736) and Arabidopsis (HPT1) is described. HPT activity was demonstrated for slr1736 in the baculovirus system and HPT1 was shown to complement the slr1736 null mutant. Further, it was shown that expression of the Arabidopsis gene, HPT1, under a seed-specific promoter led to increased levels of seed tocopherols.

Using the UbiA protein sequence as a query in a PSI-BLAST search, several HTP candidates from Synechocystis sp. PCC 6803 and Arabidopsis were identified. Based on the level of similarity to the UbiA protein, efforts were focused on one Arabidopsis candidate, HPT1, and the corresponding Synechocystis sp. PCC 6803 ortholog, Slr1736, to test in functional and enzymatic assays. It is worthwhile to note that when UbiA was used as a query in a standard BLAST search (which performs only a single pass) of the Arabidopsis public database, the only significant hit was to the putative UbiA Arabidopsis ortholog, ATPT3 (8e³). In contrast, the use of a protein profile generated by the iterative PSI-BLAST program (fourth iteration) resulted in a score of 1e¹⁸ for HPT1 and 2e²⁹ for Slr1736. The increased sensitivity of this program was particularly important given that the public predicted protein of HPT1 lacked a significant portion of the amino terminus and lacked the last 17 amino acids at the carboxy terminus, making it more difficult to identify similarity to known proteins and targeting signal peptides.

The fact that Slr1736 and HPT1 share conserved amino acid residues known to be required for catalytic function in another aromatic prenyltransferase (Fig. 3) suggests that these residues may be involved in catalysis of the prenyltransfer reaction. These residues are also found in the other Arabidopsis prenyltransferases presented, as well as in UbiA, suggesting that these residues function in a broad array of prenyltransferase reactions.

Characterization of the tocopherol null mutant of slr1736 in Synechocystis sp. PCC 6803 and Arabidopsis HPT1 antisense lines with substantially reduced seed tocopherol levels confirmed that these genes are essential for tocopherol biosynthesis. Restoration of tocopherol biosynthesis in the Synechocystis sp. PCC 6803 slr1736 null mutant by HPT1 confirmed that these genes are orthologs. Interestingly, the slr1736 null strain was not compromised in its ability to grow, indicating that tocopherols are not required for the growth of this cyanobacterium under growth conditions used here. Whereas tocopherols were not detected in the slr1736 null mutant in this study, Schledz et al. (2001) report a 90% reduction in tocopherol levels. The basis for this discrepancy is unclear.

Molecular analysis of slr1736 and HPT1, described above, suggests that these genes encode HPTs. To demonstrate that these genes encode functional HPT, an enzyme assay was developed. HPT is thought to catalyze the phytyltransfer from phytyl-DP onto HGA and the decarboxylation reaction to yield 2-methyl-6-phytyl plastoquinol. Soll et al. (1980) determined that in spinach, the HPT activity is associated with membrane fractions. HPT1 and slr1736 are predicted to encode integral membrane proteins, agreeing with these observations. Enzyme assays of Synechocystis sp. PCC 6803 wild-type and slr1736 null mutant membrane fractions confirmed that slr1736 is essential for this activity (Table II). Expression of slr1736 in insect cells conferred HPT activity to the membrane fraction, further confirming that these genes encode for HPT activity.

In a previously published study of a screen for Arabidopsis carotenoid mutants (Norris et al., 1995), a putative mutant in HPT or a related step, pds2, was identified. The gene encoding PDS2 has not been cloned, but it maps to chromosome III, whereas HPT1 maps to the top of chromosome II. Thus, pds2 does not correspond to HPT1. There are a variety of functions that could be encoded by PDS2, including a regulatory factor or effector molecule. Further mapping or other studies will be needed to determine the sequence of PDS2, but the identification of at least two loci involved in controlling this step points to the diversity of regulation that the sequencing of the Arabidopsis genome will allow us to explore.

In plants, HPT1 was shown to be important for tocopherol accumulation based on the reduction of total tocopherols in e35S::HPT1 antisense T₃ seed pools (Table III). Total tocopherol levels as low as 50 ng mg¹ tissue were observed, which is 10-fold lower than in wild-type seed. Interestingly, the levels of -tocopherol were not significantly reduced in individual T₃ lines, whereas - and -isoforms are dramatically reduced in the most affected progeny (data not shown). These data suggest that there is still sufficient -tocopherol to saturate -TMT, such that wild-type levels of -tocopherol are still achieved.

Expression of HPT1 under the seed-specific napin promoter resulted in up to a 2-fold increase in tocopherols in T₃ homozygous seeds. These data further validate the role of HPT1 in tocopherol biosynthesis in Arabidopsis seed and demonstrate that tocopherol levels can be modified in seed. The majority of the increase in total tocopherols was caused by an increase in -tocopherol (data not shown). -Tocopherol did not increase, again indicating that -TMT is saturated at wild-type levels of -tocopherol.

The initial step in tocopherol biosynthesis is the condensation of HGA and phytyl-DP by HPT. It is often assumed that the first committed step in a pathway will be a regulated step. As a consequence, one might predict that enhanced expression of such an enzyme would lead to increased flux through a given pathway. An example of increased flux was demonstrated in another plastidial isoprenoid pathway, with the overexpression of phytoene synthase in canola (Brassica napus) seeds resulting in a 50-fold increase in carotenoid levels (Shewmaker et al., 1999). Expression of HPT1 under the napin promoter consistently resulted in elevated levels of total tocopherols in seed; however, the magnitude of increase was not above 2-fold in homozygous T₃ seed. In the high-carotenoid canola seeds, the 50-fold increase was possible only because carotenoids normally comprise such a small fraction of the total isoprenoid population. The overall increase in isoprenoid units was only 4-fold (Shewmaker et al., 1999). These data suggest that there may be a limit to the level that isoprenoids can be increased without increasing flux through the methyl-erythritol phosphate pathway. In fact, there are studies that demonstrate that geranylgeranylpyrophosphate, a phytol precursor, may be limiting in tocopherol biosynthesis (Furuya et al., 1987).

The antioxidant properties of tocopherols are strongly implicated in many aspects of human health, including heart disease, cancer, and inflammatory responses. Because of these demonstrated health benefits, there is interest in developing ways to increase the intake of natural tocopherols in human and animal diets. One strategy is to increase the levels of tocopherols in oilseed crops by engineering components of the tocopherol biosynthetic pathway. To engineer the pathway, it is first necessary to identify the genes that encode all of the enzymes of the pathway. In this study, the HPT responsible for the first committed step in tocopherol biosynthesis was identified from Arabidopsis and Synechocystis sp. PCC 6803. It was further demonstrated that seed-specific expression of the Arabidopsis HPT1 in Arabidopsis can elevate seed tocopherols 2-fold, a first step in engineering oilseeds for high levels of tocopherols.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Bacterial Strains and Growth Conditions

Cultures of Synechocystis sp. PCC 6803 were grown photoautotrophically in BG11 media (Sigma, St. Louis) at 30°C under a light intensity of 50 µE m² s¹, and 70% relative humidity. Growth media for null mutants was supplemented with 25 µg mL¹ of kanamycin. Plasmids were stabilized in Synechocystis sp. PCC 6803 by addition of 10 µg mL¹ gentamycin. For growth on solid media, BG11 was supplemented with 10 mM TES, pH 8.0, and 15 g L¹ agar.

Construction of slr1736 Deletion Plasmid and Synechocystis sp. PCC 6803 Transformation

Synthetic oligos A, B, C, and D were generated to amplify regions from the 5' and 3' ends of slr1736 (A, 5'-TAATGTGTACATTGTCGGCCTC-3'; B, 5'-GCAATGTAACATCAG-AGATTTTGAGACACAACGTGGCTTTCCACAATTCCCC-GCACCGTC-3'; C, 5'-GGTATGAGTCAGCAACACCTTC-TTCACGAGGCAGACCTCAGCGGAATTGGTTTAGGTTA-TC-3'; and D, 5'-AGGCTAATAAGCACAAATGGGA-3'). The underlined nucleotides indicate regions homologous to the nptI gene. The 5' ends of primers B and C contain a 40-bp region of DNA sequence that is complementary to the 5' and 3' sequence of the nptI gene from pUC4K (GenBank accession no. X06404), respectively. The 5' and the 3' fragments of slr1736 were PCR amplified and gel purified (spin columns, Qiagen Inc., Valencia, CA) separately. The nptI gene was obtained from pUC4K by HincII digest followed by gel purification. To insert the nptI gene into slr1736, the purified 5' and 3' slr1736 fragments were mixed in a 1:1 ratio with the purified nptI gene annealed and amplified for 40 cycles under the following conditions: 1 min of incubation at 94°C, 1 min at 55°C, and 1 min at 72°C (+5 s per cycle) using pfu polymerase (Stratagene, La Jolla, CA) in 100 µL of total reaction volume (Zhao and Arnold, 1997). A volume of 1 to 5 µL of this reaction was used as template DNA for a second amplification reaction using primers A and D, so that the resulting product contained 100 to 200 bp of the 5'-end of slr1736, nptI, and 100 to 200 bp of the 3' end of slr1736. This PCR product was then cloned into the vector pGEM-T easy (Promega, Madison, WI), resulting in pMON21681 and used for stable integration into the Synechocystis sp. PCC 6803 genome. Synechocystis sp. PCC 6803 transformations were performed as described in Porter (1988). Cells grown with 25 µg mL¹ kanamycin were harvested to verify successful gene disruption by PCR, and for tocopherol analysis.

Plant Vector Construction, Arabidopsis Transformation, and Plant Growth

HPT1 5' and 3' ends were extended using RACE (Advantage PCR kit, CLONTECH, Palo Alto, CA) from Arabidopsis accession No-O inflorescence and silique cDNA libraries using the Marathon cDNA Amplification kit (CLONTECH) plus primers 5'-CCCACCAGCAGCGGA-AACAAGAGAAGAACT-3' and 5'-GTTTCTGGCTTGGGTG-GATTGTTGGTTCAT-3'. The RACE products were sequenced and the information was used to amplify the complete coding region from the silique cDNA library using the primers 5'-GGATCCGCGGCCGCACA ATGGAGTCTC-TGCTCTCTAGTTCT-3' and 5'-GGATCCTGCAGGTCACT-TCAAAAAAGGTAACAGCAAGT-3'. BamHIand NotI sites flanked the 5' end of the gene and Sse8387I (Fisher Scientific, Pittsburgh) and BamHI sites flanked the 3' end of the gene. The PCR product was cloned into the pCR2.1 TA vector (Invitrogen), resulting in the formation of pCGN10817. The DNA sequence was confirmed by DNA sequencing. Subsequently, NotI- and Sse8387I-digested HPT1 was cloned into the binary expression vectors pCGN8643 (pNapin::HPT1_sense::napin 3'; Kridl et al., 1991) and pCGN8644 (e35S::HPT1_antisense::tml 3'; Chibbar et al., 1993), with final construct names pCGN10822 and pCGN10803, respectively. These vectors were electroporated into Agrobacterium tumefaciens strain ABI and grown under standard conditions (McBride et al., 1994), reconfirmed by restriction analysis, and transformed into Arabidopsis accession No-O using the dipping method (Clough and Bent, 1998). T₀ plants were grown in a growth chamber under 16 h of light, 19°C, and T₁ seeds were selected on germination plates with kanamycin (1× Murashige and Skoog salts, 10 g L¹ Suc, 100 mg L¹ myo-inositol, 1 mg L¹ thiamine-HCl, 0.5 mg L¹ pyridoxine-HCl, 0.5 mg L¹ nicotinic acid, 0.5 g L¹ MES, 100 mg L¹ carbenicillin, 50 mg L¹ kanamycin, and 20 mg L¹ benlate, pH 5.7) and resistant plants were grown at 22°C. The number of insertions was determined by plating segregating T₂ seed onto germination plates and scoring the number of germinating and non-germinating seeds.

Complementation of slr1736 Null Mutant Strain

For complementation of the Synechocystis sp. PCC 6803 slr1736 null mutant, mature HPT1 was cloned into pSL1211, a vector based on the broad host range plasmid RSF1010 (Ng et al., 2000). The mature HPT1 gene was amplified from the vector pCGN10817 by PCR using primers HPT1nco.pr (CCATGGATTCGAGTAAAGTTGTCGC) and HPT1r1.pr (GAATTCACTTCAAAAAAGGTAACAG). These primers were designed to remove 36 amino-terminal amino acids, which are predicted to serve a plastidial target sequence. In addition, these primers engineered an NcoI site at the new translational start codon and an EcoRI site at the 3' end of the gene. The PCR product was ligated into pGEM-T easy (Promega), resulting in the formation of pMON21689 and sequence confirmed. The NcoI/EcoRI fragment from pMON21689 was ligated with the EagI/EcoRI and EagI/NcoI fragments from pSL1211, resulting in the formation of pMON21690. The plasmid pMON21690 was introduced into the Synechocystis sp. PCC 6803 slr1736 null mutant via conjugation (Elhani and Wolk, 1988).

Baculovirus Expression Vectors

For confirmation of HPT activity of slr1736 and mature HPT1 expression products, both genes were cloned as EcoRI fragments into the Bac-to-Bac Baculovirus Expression Systems (Invitrogen). Integration into the bacmid, transformation, and gene expression was done according to the manufacturer's protocol.

HPT Assay

HPT was assayed using tritiated HGA (40 Ci mmol¹) and nonlabeled phytyl-DP as substrates. Tritiated HGA was obtained by bromination of unlabeled HGA at room temperature in the presence of acetic acid and subsequent tritiation in the presence of Pd-activated charcoal in ethanol (Koelsch, 1955). 3,4-[³H]HGA was stored in 0.1% (v/v) H₃PO₄. Phytyl-DP was synthesized as described by Joo et al. (1973). Standard compounds for 2-methyl-6-phytyl-plastoquinol and 2,3-dimethyl-5-phytylquinol were synthesized as described by Soll et al. (1980). The structures were verified by mass spectroscopy. Nonlabeled HGA was obtained from Sigma. -, -, -, -tocopherol, and tocol, were purchased from Matreya (Pleasant Gap, PA). For enzyme assays of Synechocystis sp. PCC 6803 strains, total membranes were isolated by a variation of the procedure of Zak et al. (1999), in which the chlorophyll content was adjusted to 0.1 to 0.5 mg chlorophyll mL¹, protease inhibitor cocktail (Roche, Basel) was added to the extraction buffer, and the ultracentrifugation time for isolation of total membranes was increased from 30 min to 1 h at 100,000g. Membrane fractions from insect cells were isolated as described in Cases et al. (1998). The HPT assay was validated using spinach (Spinacia oleracea) chloroplasts and Synechocystis sp. PCC 6803 membrane preparations as positive controls. For HPT assays using spinach chloroplasts, chloroplasts were isolated from 250 g of spinach leaves obtained from local markets as described by Douce and Joyard (1982).

The enzyme assay contained 60 µM [³H]HGA, which had been adjusted to a specific activity of 0.16 to 4 Ci mmol¹. In addition to HGA, the enzyme assay (final volume 1 mL) contained 50 mM Tris-HCl, pH 7.6, 4 mM MgCl₂, and 100 µM phytyl pyrophosphate. The reaction was initiated by addition of membrane fractions or chloroplast preparations and terminated by adding 2 mL of chloroform:methanol (1:2 [v/v]) after 2 h of incubation at 23°C. The extraction procedure was initiated by addition of tocol (2-5 µg L¹ final concentration), which served as an internal standard to monitor the extraction efficiency. Phase separation was achieved after supplementation with 2 mL of 0.9% (w/v) aqueous NaCl solution and vigorous shaking. This extraction procedure was repeated three times. The organic layer containing the prenylquinones was filtered (0.2-µm Gelman PTFE acrodisc, 13-mm syringe filters, Pall Gelman Laboratory Inc., Ann Arbor, MI), evaporated under N₂, and then resuspended in 100 µL of ethanol.

The reaction products were separated by isocratic normal-phase HPLC (90% [v/v] hexane and 10% [v/v] methyl-t-butyl ether), using a Zorbax silica column (Agilent Technologies, Atlanta), 4.6 × 250 mm (5 µm). Alternatively, samples were analyzed by isocratic reversed-phase HPLC (0.1% [v/v] H₃PO₄ in MeOH), using a Vydac 201HS54 C18 column (Western Analytical, Murrieta, CA), 4.6 × 250 mm, coupled with a C18 guard column (Alltech, Inc., Nicholasville KY). The amount of reaction products were calculated based on the specific radioactivity of the substrate, and adjusted according to the recovery based on the tocol standard. Tocol recovery was determined based on fluorescence measurement.

Tocopherol and Chlorophyll Analysis

Tocopherols were separated by normal phase HPLC eluting with a hexane (solvent A) methyl-t-butyl ether (solvent B) gradient (gradient conditions: 0-10 min, 90% [v/v] A and 10% [v/v] B; 11 min, 25% [v/v] A and 75% [v/v] B; and 12 min, 90% [v/v] A and 10% [v/v] B) using an injection volume of 20 µL, a flow rate of 1.5 mL min¹, and a run time of 12 min (40°C). Tocopherol concentration and composition was calculated based on standard curves for -, -, -, and -tocopherol using Chemstation software (Agilent Technologies). Synechocystis sp. PCC 6803 samples were harvested in late logarithmic growth phase by centrifugation. One gram of 0.1-mm microbeads (Biospecifics Technologies Corp., Lynbrook, NY) and 500 µL of 1% (w/v) pyrogallol (Sigma) in ethanol were added to a cell pellet from 1 mL of culture. Tocol was added as internal standard. The mixture was shaken for 1 min in a mini-Beadbeater (Biospecifics Technologies Corp.) on "fast" speed. For seed tocopherol determination, 10 mg of mature seed was added to 1 g of microbeads (Biospecifics Technologies Corp.) in a sterile microfuge tube to which 500 µL of 1% (w/v) pyrogallol (Sigma) in ethanol was added. The mixture was shaken for 3 min in a mini-Beadbeater (Biospecifics Technologies Corp.) on "fast" speed. The extract was filtered (0.2-µm Gelman PTFE acrodisc, 13-mm syringe filters, Pall Gelman Laboratory Inc.) into an autosampler tube. HPLC was performed on a Zorbax silica HPLC column, 4.6 × 250 mm (5 µm), with a fluorescent detection using a Hewlett-Packard HPLC (Agilent Technologies). Sample excitation was at 290 nm, and emission was monitored at 336 nm. Chlorophyll concentration was determined (Arnon, 1949).

Statistical evaluation of seed tocopherol data was performed with Excel 2000 (Microsoft Corp., Seattle). Because T₂ seed pools were derived from independent transformation events, they do not represent replicates, and cannot be regarded as a homogeneous population. Therefore, tocopherol data from single events were compared with the 95% confidence interval of control populations. Data points outside the 95% confidence interval were regarded as significantly different from the controls. In contrast, T₃ seed samples are homogeneous for the insertion locus of a given event. Therefore, T₃ populations were compared via a Student's t test to the control populations. Populations with a P value < 0.05 were regarded as significantly different.