Genetic and biochemical basis for alternative routes of tocotrienol biosynthesis for enhanced vitamin E antioxidant production

Authors

  • Chunyu Zhang,

    1. National Key Laboratory of Crop Genetic Improvement and College of Plant Science and Technology, Huazhong Agricultural University, Wuhan, China
    2. Department of Biochemistry and Center for Plant Science Innovation, University of Nebraska, Lincoln, NE, USA
    Search for more papers by this author
  • Rebecca E. Cahoon,

    1. Department of Biochemistry and Center for Plant Science Innovation, University of Nebraska, Lincoln, NE, USA
    Search for more papers by this author
  • Sarah C. Hunter,

    1. Plant Genetics Research Unit, Donald Danforth Plant Science Center, United States Department of Agriculture-Agricultural Research Service, Saint Louis, MO, USA
    Search for more papers by this author
  • Ming Chen,

    1. Department of Biochemistry and Center for Plant Science Innovation, University of Nebraska, Lincoln, NE, USA
    Search for more papers by this author
  • Jixiang Han,

    1. Donald Danforth Plant Science Center, Saint Louis, MO, USA
    Search for more papers by this author
  • Edgar B. Cahoon

    Corresponding author
    1. Department of Biochemistry and Center for Plant Science Innovation, University of Nebraska, Lincoln, NE, USA
    • National Key Laboratory of Crop Genetic Improvement and College of Plant Science and Technology, Huazhong Agricultural University, Wuhan, China
    Search for more papers by this author

For correspondence (e-mail ecahoon2@unl.edu).

Summary

Vitamin E tocotrienol synthesis in monocots requires homogentisate geranylgeranyl transferase (HGGT), which catalyzes the condensation of homogentisate and the unsaturated C20 isoprenoid geranylgeranyl diphosphate (GGDP). By contrast, vitamin E tocopherol synthesis is mediated by homogentisate phytyltransferase (HPT), which condenses homogentisate and the saturated C20 isoprenoid phytyl diphosphate (PDP). An HGGT-independent pathway for tocotrienol synthesis has also been shown to occur by de-regulation of homogentisate synthesis. In this paper, the basis for this pathway and its impact on vitamin E production when combined with HGGT are explored. An Arabidopsis line was initially developed that accumulates tocotrienols and homogentisate by co-expression of Arabidopsis hydroxyphenylpyruvate dioxygenase (HPPD) and Escherichia coli bi-functional chorismate mutase/prephenate dehydrogenase (TyrA). When crossed into the vte2–1 HPT null mutant, tocotrienol production was lost, indicating that HPT catalyzes tocotrienol synthesis in HPPD/TyrA-expressing plants by atypical use of GGDP as a substrate. Consistent with this, recombinant Arabidopsis HPT preferentially catalyzed in vitro production of the tocotrienol precursor geranylgeranyl benzoquinol only when presented with high molar ratios of GGDP:PDP. In addition, tocotrienol levels were highest in early growth stages in HPPD/TyrA lines, but decreased strongly relative to tocopherols during later growth stages when PDP is known to accumulate. Collectively, these results indicate that HPPD/TyrA-induced tocotrienol production requires HPT and occurs upon enrichment of GGDP relative to PDP in prenyl diphosphate pools. Finally, combined expression of HPPD/TyrA and HGGT in Arabidopsis leaves and seeds resulted in large additive increases in vitamin E production, indicating that homogentisate concentrations limit HGGT-catalyzed tocotrienol synthesis.

Introduction

Two forms of vitamin E tocochromanols occur in plants: tocopherols and tocotrienols. Tocopherols, which are found widely in plants, contain a polar tocochromanol head group linked to a saturated hydrocarbon tail that is derived from the C20 isoprenoid intermediate phytyl diphosphate (PDP) (Kamal-Eldin and Appelqvist, 1996) (Figure 1). Tocotrienols are the major form of vitamin E in the seed endosperm of most monocots and some dicots, such as members of the Apiaceae family (Horvath et al., 2006; Falk and Munné-Bosch, 2010). Tocotrienols have chemical structures that are similar to tocopherols, except that the hydrocarbon tail contains three trans double bonds arising from the C20 isoprenoid intermediate geranylgeranyl diphosphate (GGDP). Tocopherols and tocotrienols are lipid-soluble antioxidants that have diverse chemical and biological properties (Sen et al., 2006; Hunter and Cahoon, 2007; Falk and Munné-Bosch, 2010; Prasad, 2011).

Figure 1.

Homogentisate biosynthesis and strategies for bypassing regulatory points to enhance vitamin E tocochromanol production. Homogentisate is derived from shikimate via a pathway that is highly regulated by tyrosine concentrations. Regulatory points may be bypassed by transgenic expression of microbial enzymes such as E. coli bi-functional chorismate mutase/prephenate dehydrogenase (TyrA) and S. cerevisiae prephenate dehydrogenase (TYR1). Further enhancement of homogentisate production may be achieved by over-expression of the last enzyme in the pathway: hydroxyphenylpyruvate dioxygenase (HPPD). De-regulated homogentisate synthesis results in enhanced vitamin E tocochromanol synthesis and homogentisate geranylgeranyl transferase (HGGT)-independent tocotrienol production (as indicated by red arrows). The studies described here examine whether homogentisate phytyltransfersas (HPT) or a related prenyltransferase such as homogentisate solanesyltransferase (HST) is recruited for tocotrienol production in response to de-regulated homogentisate synthesis, and whether the HGGT pathway and the HGGT-independent pathway for tocotrienol synthesis may be combined for enhanced vitamin E tocochromanol accumulation.

Tocopherols and tocotrienols are synthesized in the plastids of plant cells (Falk and Munné-Bosch, 2010; Yang et al., 2011). The committed step in the synthesis of tocopherols is condensation of homogentisate from the shikimate pathway with PDP, and is catalyzed by homogentisate phytyltransferase (HPT) (Collakova and DellaPenna, 2001; Savidge et al., 2002). The remainder of the tocopherol biosynthetic pathway involves a tocochromanol ring cyclization and head group methylation steps that are also shared with the tocotrienol biosynthetic pathway (Soll et al., 1980; Cahoon et al., 2003; Horvath et al., 2006; Yang et al., 2011). The synthesis of tocotrienols differs from that of tocopherols in the use of GGDP, rather than PDP, as the isoprenoid substrate for the initial condensation reaction. We have previously shown that, in monocot endosperm, this reaction is catalyzed by the enzyme homogentisate geranylgeranyl transferase (HGGT), which is related to HPT but has distinct substrate specificity (Cahoon et al., 2003; Yang et al., 2011). Recombinant barley HGGT was found to be approximately six times more active with GGDP than with PDP, while Arabidopsis HPT was nine times more active with PDP than with GGDP (Yang et al., 2011). HGGT has also been shown to be a useful enzyme for biotechnological enhancement of vitamin E antioxidant levels in plants. Expression of barley HGGT in Arabidopsis leaves, soybean seeds and maize embryos resulted in six- to tenfold increases in total vitamin E tocochromanols, primarily in the form of tocotrienols (Cahoon et al., 2003, 2006a; Meyer, 2011).

In addition to studies performed with monocot HGGT (Cahoon et al., 2003, 2006a; Meyer, 2011), considerable research has been directed at increasing vitamin E content in plants. One approach has been to enhance expression of HPT, which has typically resulted in only modest increases in vitamin E content, particularly in oilseed crops such as soybean and canola (Savidge et al., 2002; Collakova and DellaPenna, 2003; Karunanandaa et al., 2005; Lee et al., 2007; Seo et al., 2011). A likely explanation for this result is that PDP pools limit the production of tocopherols. It is known that the primary route of de novo PDP synthesis in Arabidopsis does not occur by direct desaturation of GGDP by geranylgeranyl reductase (Valentin et al., 2006), but by a more intricate route that involves desaturation of geranylgeranyl bound to chlorophyll (Keller et al., 1998; Ischebeck et al., 2006; Valentin et al., 2006). The resulting phytol moiety is then released from chlorophyll and converted to the diphosphate form by sequential kinase reactions (Valentin et al., 2006).

Another possible limitation for enhancement of vitamin E production in plants is the pool size of the homogentisate precursor of the tocochromanol head group. Attempts to increase vitamin E accumulation in transgenic plants initially focused on up-regulation of hydroxyphenylpyruvate dioxygenase (HPPD), which catalyzes the final step in homogentisate synthesis (Figure 1) (Jefford and Cadby, 1981; Lindstedt and Rundgren, 1982). Over-expression of HPPD in a variety of plant species and organs resulted in small to moderate increases in total vitamin E concentrations (Tsegaye et al., 2002; Falk et al., 2003; Rippert et al., 2004; Karunanandaa et al., 2005).

A metabolic constraint for synthesis of homogentisate via the shikimate pathway is the high degree of metabolic regulation exerted by tyrosine in reactions that lead to synthesis of the hydroxyphenylpyruvate (HPP) precursor (Figure 1). Tyrosine, the immediate precursor of HPP, is known to strongly feedback-inhibit arogenate dehydrogenase and prephenate dehydrogenase (Figure 1) (Eberhard et al., 1996; Rippert and Matringe, 2002; Less et al., 2010; Tzin and Galili, 2010a,b). Yeast and bacterial enzymes have been introduced into plants that bypass these regulated steps in HPP biosynthesis in order to increase homogentisate production (Herbers, 2003; Rippert et al., 2004; Karunanandaa et al., 2005; Tzin et al., 2009). These enzymes include the yeast prephenate dehydrogenase, which converts prephenate directly to HPP, and the bi-functional chorismate mutase/prephenate dehydrogenase encoded by bacterial TyrA that directly converts chorismate to HPP (Herbers, 2003; Rippert et al., 2004; Karunanandaa et al., 2005). Co-over-expression of yeast prephenate dehydrogenase and Arabidopsis HPPD in tobacco resulted in a tenfold increase in leaf tocochromanols (Rippert et al., 2004). Unexpectedly, the increase in tocochromanols was largely due to tocotrienol rather than tocopherol production. The unexpected accumulation of tocotrienols in response to de-regulated homogentisate production via expression of microbial enzymes has been replicated in several other studies (Herbers, 2003; Karunanandaa et al., 2005; Tzin et al., 2009), most notably in seeds of canola and soybean engineered to express the bacterial TyrA and HPPD genes (Karunanandaa et al., 2005).

The induction of tocotrienol synthesis in response to de-regulated homogentisate production occurs in the absence of a functional HGGT in dicots, including Arabidopsis, tobacco and soybean (Herbers, 2003; Rippert et al., 2004; Karunanandaa et al., 2005). How this occurs is unclear. Tocotrienol formation requires the condensation of GGDP and homogentisate. Although the Arabidopsis HPT displays low in vitro activity with GGDP (Yang et al., 2011), over-expression of HPT from Arabidopsis and other plants has not been previously shown to result in tocotrienol synthesis (Savidge et al., 2002; Collakova and DellaPenna, 2003; Lee et al., 2007; Seo et al., 2011). This contrasts with the monocot HGGT, which functions primarily to generate tocotrienols but can also produce tocopherols in planta (Cahoon et al., 2003; Yang et al., 2011). The production of tocopherols by monocot HGGT activity corresponds to developmental stages at which PDP pool sizes are known to be more enriched (Yang et al., 2011), such as during senescence, when chlorophyll degradation releases phytol that is converted to PDP (Ischebeck et al., 2006; Valentin et al., 2006). It is not clear whether HPT or a related prenyltransferase is recruited for tocotrienol synthesis upon de-regulation of homogentisate synthesis. It has been previously reported that the Arabidopsis HPT is 9–12 times more active with PDP than with GGDP (Sadre et al., 2006; Yang et al., 2011), but has a similar apparent Km for the two substrates (Sadre et al., 2006). As such, HPT recruitment for tocotrienol synthesis requires a large increase in pools of GGDP relative to PDP. Alternatively, a gene related to VTE2 (encoding HPT) (referred to as VTE2–2, At3 g11945) has been shown to encode a prenyltransferase that is involved in homogentisate condensation with the C45 isoprenoid solanesyl diphosphate in biosynthesis of plastoquinone–9 (Sadre et al., 2006; Venkatesh et al., 2006; Tian et al., 2007). This enzyme, designated homogentisate solanesyl diphosphate transferase (HST), was also found to have in vitro activity with GGDP and PDP, and was at least five times more active with GGDP than with PDP (Sadre et al., 2006).

In this paper, we address the question of whether HPT or a related prenyltransferase such as HST is associated with the production of tocotrienols upon de-regulation of homogentisate biosynthesis in a manner related to GGDP and PDP pool sizes? We also explore the hypothesis that homogentisate is a limiting substrate for the synthesis of tocochromanols in plants, and that, when combined with HGGT, de-regulated homogentisate production leads to high levels of tocotrienol accumulation.

Results

Strong up-regulation of homogentisate biosynthesis alters plant growth and tocochromanol compositions and concentrations

Experiments were performed to determine whether the activity of homogentisate phytyltransferase (HPT) or an alternative prenyltransferase such as homogentisate solanesyltransferase (HST) is responsible for tocotrienol biosynthesis induced by strong up-regulation of flux through the shikimate pathway (Figure 1). To initiate these studies, Arabidopsis Col–0 lines were engineered to constitutively over-express Arabidopsis hydroxyphenylpyruvate dioxygenase (HPPD) and the E. coli bi-functional chorismate mutase/prephenate dehydrogenase (TyrA) fused with a plastid targeting sequence. Lines were generated using two approaches: (i) genetic crossing of a confirmed HPPD over-expression line with a confirmed TyrA over-expression line, and (ii) co-expression of HPPD and TyrA by transformation with a single binary vector containing transgenes for both polypeptides. Using the first approach, lines were obtained that express HPPD or TyrA under the control of strong constitutive promoters. Enhanced expression of HPPD and expression of TyrA in these lines was confirmed by Western blotting (Figure 2a). HPPD and TyrA over-expression lines accumulating the highest tocochromanol levels were selected for crossing based on levels of the corresponding polypeptides from Western blots. These studies were facilitated by the presence of a DsRed selection marker in the HPPD over-expression lines. By crossing HPPD over-expression lines with TyrA over-expression lines, it was possible to visually identify seeds from F1 progeny of successful crosses by fluorescence conferred by the DsRed marker. The presence of DsRed also enabled monitoring of hybrid seed development, and, as shown in Figure 2(b), no seed abortion was detected in progeny from the crosses. In addition, DsRed-positive seeds obtained from the crosses were viable. However, the resulting plants were significantly reduced in size relative to HPPD and TyrA over-expression parental lines (Figures 2c and 3a). Co-expression of HPPD and TyrA transgenes on a single T–DNA gave rise to plants that were morphologically similar to plants derived from crosses of HPPD and TyrA over-expression lines (Figures 2c and 3a).

Figure 2.

Impact of transgenic TyrA and HPPD expression on development of seeds and the morphology of plants from genetic crosses. (a) Co-expression of TyrA and HPPD in parents (Wt/TyrA, Wt/HPPD) and F1 (F1, Wt/H+T) and F2 (F2, Wt/H+T) progeny of crosses was confirmed by Western blot analyses. Purified recombinant proteins (PP) for TyrA and HPPD are shown as controls for Western blot analyses. (b) Seed set was not affected in crosses as shown by examination by white and fluorescent light of developing seeds from progeny resulting from crossing an HPPD transgenic line that includes a DsRed marker with the TyrA transgenic line. (c) F1 plants from the cross described in (b) were significantly reduced in size relative to the parental lines and wild-type Arabidopsis.

Figure 3.

Impact of transgenic expression and co-expression of TyrA and HPPD on plant morphology and homogentisate and tocochromanol production. (a) Plants over-expressing HPPD were indistinguishable from wild-type plants, but TyrA-over-expressing plants had more rounded leaves than wild-type plants. Co-expression of TyrA and HPPD resulted in a reduced rosette size. (b) Over-expression of HPPD does not result in detectable increases in total vitamin E tocochromanols in leaves of 3-week-old plants, whereas significant increases were detected in plants expressing TyrA and co-expressing TyrA and HPPD transgenes. Tocotrienol and homogentisate accumulation were also detected in leaves of these plants, and homogentisate accumulation increased with TyrA and HPPD co-expression. Values are means ± SE (= 3).

Metabolite analyses of leaves from HPPD-over-expressing transgenic lines revealed no significant increase in total tocochromanol and no detectable tocotrienols (Figure 3b). However, leaves from TyrA-over-expressing lines accumulated tocotrienols at levels up to 20–25% of the total tocochromanol content, and overall tocochromanol concentrations in leaves from these plants were two- to threefold higher compared to wild-type control and HPPD over-expression lines (Figure 3b). Tocotrienols constituted approximately 33% of the overall tocochromanol content of leaves from lines with TyrA and HPPD transgenes linked on a single T–DNA (Figure 3b), and approximately 50% in leaves from crossed HPPD and TyrA-over-expressing plants (Figure S1). Overall increases in tocochromanol concentrations in leaves from these lines were approximately the same as in leaves of plants expressing TyrA alone. In addition, no accumulation of free homogentisate was detectable in leaves of wild-type and HPPD over-expression lines (Figure 3b). By contrast, free homogentisate accumulation was detected in leaves from TyrA-over-expressing lines, and these concentrations were increased three- to fivefold by co-expression of TyrA and HPPD on a single T-DNA (Figure 3b).

A functional HPT is required for tocotrienol synthesis induced by de-regulated homogentisate production

Using lines that produce tocotrienols via strong up-regulation of homogentisate synthesis, experiments were performed to determine whether tocotrienol production is the result of HPT or an alternative prenyltransferase activity. For these studies, three independent lines engineered for TyrA and HPPD co-expression from a single T–DNA (male parent) were crossed with the Arabidopsis vte2–1 mutant (female parent), which lacks a functional HPT and produces no detectable tocochromanols (Figures 4 and 5c). Plants from F1 seeds of this cross displayed the same tocochromanol profiles as the male parents (Figure 5d). In the F2 generation, plants expressing the DsRed marker and with morphology similar to the male parent were selected for genetic background identification using a SNP marker that enabled differentiation of plants with wild-type and non-functional mutant copies of the HPT gene (Figure 4c). By combining the DsRed marker linked to the TyrA and HPPD transgenes and the HPT SNP marker, lines with enhanced homogentisate production in backgrounds that are heterozygous or homozygous for the vte2–1 (hpt) mutant were readily identified. Morphological differences among plants over-expressing HPPD/TyrA in either the wild-type, vte2–1 heterozygous or vte2–1 homozygous backgrounds were undetectable. However, HPLC analyses indicated that tocochromanols were absent from the leaves in plants co-expressing TyrA and HPPD transgenes in the homozygous vte2–1 background (Figure 5e). These results provide unequivocal genetic evidence that a functional HPT is required for tocotrienol synthesis in plants with de-regulated homogentisate production.

Figure 4.

Introduction of HPPD/TyrA transgenes into the vte2-1 HPT-null mutant. (a,b) Morphologies of vte2-1 (left), Wt/HPPD + TyrA (right) and F1 hybrids (middle) at the vegetative (a) and reproductive stages (b). (c) Genotyping of genetic backgrounds in the F2 generation from crosses of vte2-1 and Wt/HPPD + TyrA using a SNP marker specific to the Arabidopsis HPT locus. F2 individuals expressing HPPD and TyrA transgenes in the vte2-1 background were confirmed using a SacI site that is specific to the vte2-1 mutation. As described in Experimental procedures, generation of a lower-molecular-weight band following SacI digestion of a partial genomic HPT PCR product is indicative of the vte2-1 background.

Figure 5.

Tocochromanol composition of leaves from expression of TyrA and HPPD in the vte2-1 background. The panels show chromatograms from HPLC analysis of tocotrienols (T3) and tocopherols (T), as detected by fluorescence emission, in 4-week-old plants of the designated genotype: (a) wild-type plants (Wt), (b) wild-type plants co-over-expressing HPPD and TyrA transgenes (Wt/HPPD + TyrA), (c) non-transformed vte2-1 plants (vte2-1), (d) F1 plants from crosses between vte2-1 and Wt/HPPD + TyrA (vte2-1 × Wt/HPPD + TyrA), (e) vte2-1/HPPD + TyrA plants selected by genotyping as shown in Figure 4c (vte2-1/HPPD + TyrA), and (f) tocotrienol and tocopherol standards (α, γ, β and δ forms). Note: the γ and β forms of tocotrienols and tocopherols are not resolved by this HPLC method.

HPT can produce tocotrienol precursors in the presence of high ratios of GGDP to PDP

It has previously been shown that the Arabidopsis HPT is 9–12 times more active with PDP to produce the tocopherol precursor 2methyl-6phytylbenzoquinol, than with GGDP to produce the tocotrienol precursor 2methyl-6geranylgeranylbenzoquinol (Sadre et al., 2006; Yang et al., 2011). To explore the ability of this enzyme to produce tocotrienol precursors, competition assays were performed with recombinant Arabidopsis HPT in reactions containing varying ratios of PDP:GGDP: 10 μm PDP:1 μm GGDP, 1 μm PDP:1 μm GGDP and 1 μm PDP:10 μm GGDP. The activity assays used microsomal extracts from insect Sf-9 cells expressing the mature Arabidopsis HPT, and measured the incorporation of 14Chomogentisate into the tocopherol and tocotrienol precursors, which were resolved from assay extracts by reversed-phase TLC. Through this approach, HPT was found to produce approximately 20-fold and approximately 10-fold more of the tocopherol precursor than the tocotrienol precursor in assays with 10 μm PDP:1 μm GGDP and 1 μm PDP:1 μm GGDP, respectively (Figure 6). However, when the prenyl diphosphate ratio was shifted to 1 μm PDP:10 μm GGDP, an almost 1:1 ratio of 2methyl-6phytylbenzoquinol:2methyl-6geranylgeranylbenzoquinol was produced (Figure 6). These findings are consistent with the genetic evidence implicating HPT in tocotrienol synthesis in response to enhanced homogentisate production, but also indicate that the tocotrienol biosynthetic ability of HPT is dependent on high ratios of GGDP:PDP in planta.

Figure 6.

Substrate competition assays with recombinant Arabidopsis HPT. Enzyme assays were performed using differing molar ratios of PDP and GGDP. Reactions were performed using 10 μm GGDP and 1 μm PDP (10:1), 1 μm GGDP and 1 μm PDP (1:1), and 1 μm GGDP and 10 μm PDP (1:10). Assays were performed with reaction conditions in the linear activity range, and geranylgeranyl and phytyl benzoquinol reactions products were resolved by reversed-phase TLC and measured by scintillation counting. Activity was measured from three independent reactions with each substrate ratio (mean ± SD).

Tocotrienol content varies with developmental stage in transgenic lines with de-regulated homogentisate synthesis

In our previous studies with monocot HGGT, we demonstrated that the ability of this enzyme to generate tocotrienols or tocopherols in the Arabidopsis vte2–1 background was dependent on the age of plants, reflective of the pool sizes of GGDP and PDP (Yang et al., 2011). In the present study, leaves from wild-type and plants from five independent lines co-expressing HPPD/TyrA transgenes were analyzed at 3, 5 and 7 weeks after planting for tocochromanol content and composition (Figure 7). At 3 weeks of age, nearly half of the tocochromanols in the transgenic plants were present as tocotrienols (Figure 7). As the plants increased in age, relative and absolute concentrations of tocotrienols in leaves of the transgenic plants decreased. By 7 weeks, only 5–10% of the total tocochromanols in leaves of the HPPD/TyrA-over-expressing plants was present as tocotrienols, with the remaining 90–95% of tocochromanols present as tocopherols (Figure 7). Overall concentrations of tocochromanols in the leaves of these transgenic plants showed little change with age, whereas tocochromanol concentrations increased almost threefold in leaves of wild-type plants between 3 and 7 weeks of age. As described previously (Ischebeck et al., 2006; Yang et al., 2011), the overall increase in tocopherols in leaves of wild-type plants and the large enrichment in tocopherols in leaves of the HPPD/TyrA-over-expressing plants at 7 weeks of age may be attributed to enhanced concentrations of PDP from increased chlorophyll turnover as plants approach senescence. These in planta data are consistent with the dependence of tocopherol or tocotrienol precursor synthesis via HPT on the relative pool sizes of GGDP and PDP, as described above.

Figure 7.

Variation in tocotrienol content with plant age in lines co-expressing TyrA and HPPD transgenes. Measurements of total tocotrienol and tocopherol content in leaves of 3-, 5- and 7-week-old plants are shown. Values are means ± SE (= 3–5).

Combining HGGT and enhanced homogentisate production to understand constraints in tocochromanol biosynthesis

Experiments were performed to determine whether HGGT expression and strong enhancement of flux through the shikimate branch may serve as synergistic approaches for increased tocochromanol production. Initially studies were designed to monitor tocochromanol accumulation in progeny from crosses of Arabidopsis plants constitutively expressing barley HGGT with plants constitutively expressing TyrA and HPPD transgenes. Two independent lines for each parent were used for the crosses. Severe impairment of growth was observed in hybrid seedlings from these crosses, and plant growth did not progress beyond the seedling stage in the hybrid plants (Figure 8a–c). Metabolite analyses revealed as much as a five- to sevenfold increase in total tocochromanol concentrations in seedlings from the crosses relative to either the HGGT- or HPPD/TyrA-over-expressing parents, with 75–85% of the tocochromanols present as tocotrienols (Figure 8d).

Figure 8.

Constitutive co-expression of HGGT and HPPD/TyrA transgenes results in strongly enhanced vitamin E tocochromanol accumulation but severely impaired growth. (a) Measurements of total tocotrienols and vitamin E tocochromanols in leaves of 5–week-old plants from lines expressing HGGT and HPPD/TyrA transgenes and progeny of independent crosses of HGGT and HPPD/TyrA lines. (b) Seedlings of plants expressing barley HGGT (Wt/HGGT, left), HPPD/TyrA (Wt/HPPD + TyrA, right) and hybrids from crosses of HGGT and HPPD/TyrA lines (Wt/HGGT x Wt/HPPD + TyrA, middle). The white arrows indicate F1 hybrid seedlings from crossing experiments. (c) Vitamin E tocochromanol concentrations in seedlings from parents expressing HGGT (HGGT) and HPPD/TyrA (HPPD + TyrA) and F1 (HGGT x HPPD + TyraA) from two independent crosses.

As an alternative approach to avoid a negative impact on vegetative growth, experiments were performed using transgenes under the control of strong-seed specific promoters. For these studies, barley HGGT was linked to the Brassica napus napin promoter, and transgenes for TyrA under the control of the soybean glycinin1 promoter and HPPD under control of the soybean α′ subunit of the βconglycinin promoter were combined on a single binary vector. The napin:HGGT transgene was introduced into Arabidopsis Col0 by kanamycin selection, and the glycinin1:TyrA/βconglycinin:HPPD transgenes were introduced into Arabidopsis Col0 by DsRed selection. Of 34 independent events generated for the napin:HGGT construct, seeds from events with the highest tocochromanol concentrations contained 2129 μg/g seed weight of tocochromanols, an approximately 4.5-fold increase relative to wild-type seeds (Figure 9a and Figure S2). Of the total tocochromanols, tocotrienol concentrations in seeds of the napin:HGGT line with the highest amount of tocochromanols was 1467 μg/g seed weight. No tocotrienols were detected in wild-type Arabidopsis seeds. Co-expression of the HPPD/TyrA transgenes resulted in 817 μg/g seed weight of tocochromanols in the highest tocotrienol-accumulating line or a 1.8-fold increase in tocochromanols relative to wild-type seeds (Figure 9a and Figure S2). Seeds of this line also produced tocotrienols at concentrations of 103 μg/g seed weight. On the basis of the results of these experiments, two homozygous napin:HGGT lines HGGT5 and HGGT17 were re-transformed with the βconglycinin:HPPD/glycinin1:TyrA construct. The seeds of the HGGT5 lines transformed with the seed-specific HPPD/TyrA transgenes displayed 1.8- and 2.1-fold increases in total tocohromanols and tocotrienols, respectively, compared to seeds from the HGGT5 parent (Figure 9b and Figure S2). Similarly, seeds from the highest tocochromanol-accumulating HGGT17 transformants contained 2.2- and 2.7-fold increases in total tocochromanols and tocotrienols, respectively, compared to seeds of the HGGT17 parent (Figure 9c and Figure S2). Seeds from this line had tocochromanol concentrations of approximately 3600 μg/g seed weight, with almost 80% of this present as tocotrienols. This represented a 7.9-fold increase in total tocochromanols relative to wild-type seeds. The combination of seed-specific transgenes for HGGT/HPPD/TyrA also resulted in a marked increase in relative amounts of the least methylated tocochromanol species δ-tocotrienol and δ-tocopherol compared to seeds of the individual HGGT and HPPD/TyrA transformants (Figure 10). The accumulation of these species suggests that the enhanced production of tocochromanols in these lines exceeds the capacity of methyltransferases to convert tocochromanols to their more methylated forms.

Figure 9.

De-regulation of homogentisate synthesis by seed-specific co-expression of HPPD/TyrA results in additive increases in tocotrienols and total vitamin E tocochromanols when combined with seed-specific expression of HGGT. (a) Tocochromanol content of seeds of wild-type plants (Wt) and lines of plants expressing HPPD/TyrA (Wt/HPPD + TyrA) and HGGT (Wt/HGGT). Values are means ± SE (= 34). (b,c) Tocochromanol content of seeds from homozygous lines expressing HGGT (Wt/HGGT5 or Wt/HGGT17; T5) and representative lines of Wt/HGGT5 or Wt/HGGT17 co-expressing HPPD/TyrA (T2 seeds). Values are means ± SE from measurement of seeds from 3–5 independent transgenic events or samples.

Figure 10.

Tocochromanol composition of Arabidopsis seeds from expression of HGGT and HPPD/TyrA transgenes and co-expressing HGGT and HPPD/TyrA transgenes under control of strong seed-specific promoters. The panels show chromatograms from HPLC analysis of tocotrienols and tocopherols with detection by fluorescence emission for seeds of (a) wild-type plants (Wt), (b) plants expressing HPPD and TyrA transgenes (Wt/HPPD + TyrA), (c) plants over-expressing HGGT (Wt/HGGT), (d) plants co-expressing HGGT and HPPD/TyrA transgenes (Wt/HGGT/HPPD + TyrA) and (e) tocotrienol and tocopherol standards (α, γ, β and δ forms). Note: the γ and β forms of tocotrienols and tocopherols are not resolved by this HPLC method. All analyses were performed using extracts from 5 mg seeds plus 1.5 μg of the internal standard 5′,7′-dimethyltocol (IS) (added before extraction), and the chromatograms shown represent equal analysis volumes.

Overall, the significant increases in tocochromanols and tocotrienols achieved by introduction of HPPD/TyrA transgenes into HGGT expression lines suggest that homogentisate is a limiting substrate for tocochromanol production in Arabidopsis leaves and seeds.

Discussion

Here we provide genetic and biochemical evidence that tocotrienol synthesis induced by strong up-regulation of homogentisate production is solely dependent on HPT, rather than related prenyltransferases such as homogentisate solanesyltransferase (HST or ‘VTE2-2′) (Sadre et al., 2006; Venkatesh et al., 2006). In this regard, Arabidopsis engineered for tocotrienol synthesis by co-expression of HPPD and TyrA transgenes ceases to produce all vitamin E tocochromanols when introduced into the HPT-null vte2-1 mutant. By in vitro assay of Arabidopsis HPT activity, we also show that this enzyme can use GGDP as a substrate to produce the tocotrienol precursor geranylgeranyl benzoquinol, but this activity is only favored in substrate competition assays when GGDP concentrations greatly exceed those of PDP, the typical in planta HPT substrate. In addition, it was observed that tocotrienol accumulation decreased during the late stages of growth in plants engineered for HPPD/TyrA co-expression, when in planta pools of PDP are known to increase (Ischebeck et al., 2006; Valentin et al., 2006). Overall, these findings indicate that strong up-regulation of homogentisate production shifts ratios of C20 prenyl diphosphate towards high levels of GGDP relative to PDP, such that HPT-mediated tocochromanol synthesis results in tocotrienol production. A likely metabolic scenario is that de-regulated homogentisate production depletes PDP pools such that the ratio of GGDP:PDP is significantly increased, leading to induction of HPT-mediated tocotrienol synthesis. However, direct measurement of plastidic GGDP and PDP pools is currently limited by lack of routine methods for analysis of these pools and for distinguishing plastidic and cytosolic pools. However, the observation that tocotrienol accumulation markedly decreases in HPPD/TyrA-over-expressing plants as they reach latter stages of maturity reflects the dynamic nature of GGDP and PDP pools during plant growth and its impact on the quantity and composition of the vitamin E tocochromanols that are produced by plant cells. This phenomenon was also previously observed for constitutive barley HGGT expression in the Arabidopsis vte2-1 mutant (Yang et al., 2011). As these plants reached maturity, HGGT-mediated tocopherol production sharply increased due to the apparent enrichment of PDP in prenyl diphosphate pools.

All previous attempts to enhance vitamin E production in numerous plant organs, including leaves, fruits and seeds, by over-expression of plant HPTs alone have failed to yield substantial production of tocotrienols (Savidge et al., 2002; Collakova and DellaPenna, 2003; Karunanandaa et al., 2005; Lee et al., 2007; Seo et al., 2011). Instead, these studies have typically reported modest increases in tocopherol concentrations. These findings indicate that, in the absence of de-regulated homogentisate production, PDP pool sizes are sufficiently in excess of GGDP pools, such that tocopherols are the principal form of tocochromanol produced as a result of HPT over-expression. The metabolic capacity to maintain PDP concentrations to support high levels of tocopherol synthesis is probably limited due to the more intricate pathway for PDP production via geranylgeranyl reductase activity, principally on geranylgeranyl chlorophyll substrates (Valentin et al., 2006). In contrast, tocotrienol production may be conferred to a wide variety of plant cells and organs by transgenic expression of monocot HGGT, which, in the case of barley HGGT, is approximately sevenfold more active with GGDP than with PDP (Yang et al., 2011). For example, CaMV 35S promoter-mediated expression of barley HGGT increased total vitamin E tocochromanols in leaves of wild-type Arabidopsis by as much as 10–15-fold, primarily as tocotrienols (Cahoon et al., 2003). The large increase in tocotrienols is somewhat surprising given that PDP pool sizes probably exceed GGDP pool sizes in Arabidopsis leaves. High levels of tocotrienol accumulation were also reported in soybean seeds showing seed-specific expression of HGGT (but not HPT) (Cahoon et al., 2006a; Meyer, 2011). A possible explanation for these observations is that GGDP concentrations regulate plastid isoprenoid flux, given the central importance of this metabolite in the synthesis of compounds including not only vitamin E tocochromanols, but also carotenoids, gibberellins, abscisic acid and chlorophyll (Lichtenthaler, 1999). As such, depletion of GGDP for tocotrienol synthesis may cause up-regulation of flux in the isoprenoid pathway to generate substrate for high levels of tocotrienol production in transgenic plants. To examine this scenario, we performed preliminary quantitative PCR experiments using lines constitutively expressing barley HGGT and HPPD/TyrA to measure expression levels of genes for 1-deoxy-d-xylulose-5-phosphate synthase (DXS) and 1-deoxy-d-xylulose-5-phosphate reductoisomerase (DXR), which catalyze initial steps in the plastid isoprenoid pathway (Schwender et al., 1999; Estevez et al., 2000, 2001; Carretero-Paulet et al., 2002). No up-regulation of the DXS and DXR genes was detected in leaves of these lines (Figure S3), indicating that, if regulation is mediated at these steps, it occurs through post-translational mechanisms.

Transgenic expression of HPPD/TyrA for increased homogentisate production and HGGT for increased homogentisate prenylation have proven to be effective independent approaches for enhanced vitamin E tocochromanol synthesis (Karunanandaa et al., 2005). To more completely understand metabolic bottlenecks that limit vitamin E tocochromanol production, we combined the two approaches using constitutive and seed-specific expression of transgenes. For both promoter types, combined expression of HGGT and HPPD/TyrA yielded strong enhancements in total tocochromanol production, largely as tocotrienols, that were greater than achieved with expression of the individual transgenes. In the case of seed-specific expression studies, two homozygous HGGT-expressing Arabidopsis lines were re-transformed with HPPD/TyrA transgenes. Seeds obtained from these lines had vitamin E tocochromanol concentrations as much as twofold higher than in the HGGT-over-expressing background. The vitamin E tocochromanol concentrations in the highest producing events were >3000 μg/g seed weight, which was approximately 3.5-fold higher than previously reported in studies with seeds of canola engineered to express HPPD/TyrA and the Synechocystis HPT (Karunanandaa et al., 2005). A comparison of our results with those from previous metabolic engineering studies of Arabidopsis and canola (Savidge et al., 2002; Karunanandaa et al., 2005) is summarized in Figure S4. These findings indicate that homogentisate availability limits HGGT-mediated tocochromanol production. Although this is of biotechnological significance, the regulation of homogentisate production may also contribute to the mediation of tocotrienol synthesis in monocot endosperm, where tocotrienol concentrations are considerably lower than those achieved with transgenic HGGT expression, despite relatively high expression of HGGT. For example, vitamin E tocochromanol concentrations in barley seeds are <100 μg/g seed weight (Falk et al., 2004), compared to concentrations of >1500 μg/g seed weight in Arabidopsis seeds expressing barley HGGT. Conversely, our findings indicate that tocotrienol production via de-regulated homogentisate synthesis is limited by the substrate properties of the downstream prenyltransferase: PDP-preferring HPT restricts tocotrienol production in response to de-regulated homogentisate synthesis, whereas GGDP-preferring HGGT allows high levels of tocotrienol synthesis in response to de-regulated homogentisate synthesis.

Overall, the results reported highlight the complexities of metabolic regulation that arise at the junction of the shikimate and isoprenoid pathways for vitamin E tocochromanol synthesis, and indicate how the dynamics of GGDP and PDP pool sizes may over-ride the substrate specificities of prenyltransferases to determine the composition of vitamin E tocochromanols formed. It is still unclear whether pool sizes of C20 prenyl diphosphate, particularly GGDP, regulate flux through the isoprenoid pathway, and if so, what the mechanism of such regulation is. The results also suggest that maintenance of high PDP concentrations in organs with de-regulated homogentisate synthesis limits production of the tocopherol form of vitamin E, even in the presence of HPT over-expression (Karunanandaa et al., 2005). This metabolic bottleneck may be addressed by introduction of a PDP biosynthetic route that directly converts GGDP to PDP, bypassing the existing chlorophyll-linked pathway for PDP synthesis in Arabidopsis (Valentin et al., 2006).

Experimental Procedures

Plant material and growth conditions

Arabidopsis thaliana Columbia-0 ecotype was used for all studies. Arabidopsis vte2-1 seeds (Sattler et al., 2004) were provided by Dean DellaPenna (Department of Biochemistry, Michigan State University, East Lansing, Michigan). Seeds were sown directly in soil, and seedlings were transplanted to a new pot at 2 weeks. Plants were maintained at 22°C and 50% humidity under a 16 h light (100 μmol m−2 sec−1)/8 h dark cycle.

Insect cell expression of Arabidopsis HPT

The mature Arabidopsis HPT was expressed in Spodoptera frugiperda (Sf-9) insect cells using the Bac-2-Bac system (Invitrogen, www.invitrogen.com), and microsomes from recombinant cells were prepared and stored as previously described (Yang et al., 2011).

HPT assay

HPT assays were performed using synthesized [U-14C]-homogentisate (specific activity 450 mCi mmol−1) and cold GGDP and PDP (American Radiolabeled Chemicals, www.arc-inc.com), and microsomes from Sf-9 insect cells expressing mature Arabidopsis HPT with products from competition assays separated by reversed-phase TLC, as described previously (Yang et al., 2011).

Constitutive expression vector construction

p35S-TyrA

The open reading frame of TyrA was amplified from E  coli genomic DNA using the sense and antisense oligonucleotides 5′-TATTCCATGGTTGCTGAATTGACCGC-3′ and 5′-ATTTCTAGATTACTGGCGATTGTCATTC-3′, respectively (added restriction sites are underlined). All PCR reactions were performed using Phusion polymerase (New England Biolabs, www.neb.com). The PCR product was linked as a 5′ NcoI/3′ XbaI fragment to the coding sequence for the plastid transit peptide for coriander ∆4 palmitoyl-acyl carrier protein desaturase as described previously (Cahoon and Shanklin, 2000). The resulting coding sequence of the plastid transit peptide and TyrA was amplified using the antisense oligonucleotide whose sequence is given above and the sense oligonucleotide 5′-TATAGAATTCAAAATGGCCATGAAACTGAATGCC-3′. The resulting product was linked as a 5′ EcoRI/3′ XbaI fragment to the CaMV 35S promoter at its 5′ end and to the 3′ UTR of octapine synthase at its 3′ end. The promoter and 3′ UTR cassette was derived from pART7 (Gleave, 1992) with flanking AscI sites cloned into pBluescript SK(+) (Agilent Genomics, www.genomics.agilent.com). The entire transgene cassette was introduced as an AscI fragment into the previously described binary vector pEC92 containing a plant kanamycin resistance marker (Cahoon et al., 2006b). The resulting plasmid was designated p35S-TyrA.

pSUbi-HPPD

The open reading frame of Arabidopsis HPPD was amplified from a cDNA library prepared from flowers using the oligonucleotides 5′-TAATGGATCCAAAATGGGCCACCAAAACGCC-3′ and 5′-TATACCATGGTGTTCATCCCACTAACTGTTTGGC-3′. The product was cloned as a 5′ BamHI/3′ NcoI fragment behind the super-ubiquitin promoter in the pAKK1517 vector (Yang et al., 2011). The resulting plasmid was digested with AscI, and the released fragment containing the complete HPPD expression cassette was cloned into the corresponding site of the previously described binary vector DsRed-GlyRed2 (Yang et al., 2011), which contains a DsRed marker for transgenic plant selection. The resulting plasmid was designated pSUbi-HPPD.

pSUbiHPPD-35STyrA

An AscI fragment from p35S-TyrA containing the complete expression cassette was cloned into the MluI site of pSUbi-HPPD to form pSUbiHPPD-35STyrA.

The previously described pSH24 vector (Cahoon et al., 2003) was used for expression of barley HGGT under the control of the CaMV 35S promoter. Expression of HPPD and TyrA in transgenic lines was determined by Western blot analyses as described in Methods S1.

Seed-specific expression vector construction

pBConHPPD-GlyTyrA

The 5′ EcoRI/3′ XbaI fragment containing the TyrA openreading frame with the plastid transit peptide coding sequence described above was ligated into the corresponding sites of the binary vector pBinGlyRed3, which contains regulatory sequences for seed-specific transgene expression and a DsRed marker gene under the control of the cassava mosaic virus 35S promoter. This resulted in placement of the TyrA sequence under the control of the strong seed-specific soybean glycinin-1 promoter at its 5′ end and the 3′ UTR of the glycinin-1 gene on its 3′ end to generate pGlyTyrA. The Arabidopsis HPPD sequence was amplified as described above using the oligonucleotides 5′-TAATGCGGCCGCAAAATGGGCCACCAAAACGCC-3′ and 5′-TATAGCGGCCGCTGTTCATCCCACTAACTGTTTGGC-3′. The product was digested with NotI and ligated into the corresponding site of pBCon123-Hyg in the 5′→3′orientation, with the PCR product linked at its 5′ end to the strong soybean seed-specific promoter for the α′ subunit of the β-conglycinin gene and at its 3′ end to the 3′ UTR of the phaseolin gene. The resulting plasmid was digested with AscI, whose sites flank the entire expression cassette, and cloned into the corresponding site of pGlyTyrA to form pBConHPPD-GlyTyrA.

pNapin-HGGT

The open-reading frame of barley HGGT was amplified using the oligonucleotides 5′-TTTTTTGGCCGCCGAGGATGCAAGCCGTCACGGCC-3′ and 5′- TTTTTGGCGGCCGCCTTGTACAAATTTCACTGCAC-3′. Prior to PCR amplification, the NotI site in the native barley HGGT sequence was removed by introduction of a silent point mutation. The PCR product was digested with NotI and cloned 5′→3′ in the corresponding site of vector pMS4 which contains a strong seed-specific promoter for the Brassica napus napin gene separated by a NotI site from the 3′ UTR of the soybean glycinin-1 gene. pMS4 containing barley HGGT was digested with AscI, whose sites flank the entire expression cassette, and the resulting fragment was cloned into the corresponding site of the binary vector pKAN2, which is a variant of pEC92 (see above) that contains an additional MluI restriction site to generate pNapin-HGGT.

Arabidopsis transformation and selection

Binary vectors were introduced into Agrobacterium tumefaciens C58 by electroporation, and transgenic plants were prepared by the floral-dip method (Clough and Bent, 1998). Transgenic seeds were selected either on LS medium (Phytotechnology Labs, www.phytotechlab.com) containing kanamycin (50 mg/l) or by fluorescence of the DsRed marker (Jach et al., 2001).

SNP marker development

Col-0 ecotype and vte2-1 were distinguished with an SNP marker introduced by the vte2-1 nonsense point mutation (from TGG to TGA) at amino acid 208 of 393 of the Arabidopsis HPT gene, resulting in an extra SacI restriction site in vte2-1 (Sattler et al., 2004). Using genomic DNA of wild-type and vte2-1 as templates, a partial Arabidopsis HPT fragment was amplified using oligonucleotides 5′-GTTAACAAGCCCTATCTTCC-3′ and 5′-GTTGCGAGTATAACATGACC-3′ flanking the mutation in vte2-1. After amplification, PCR products were digested with SacI at 37°C for 3 h and then separated by electrophoresis.

HPLC analysis of tocochromanol content and composition

Vitamin E tocochromanols were analyzed from freshly harvested leaves (50–100 mg) or dried seeds (10 mg) by HPLC with fluorescence detection, and quantified relative to the internal standard 5,7-dimethyltocol (Matreya, www.matreya.com). Tocochromanol extraction was performed by homogenization of plant material in 1 ml methanol/dichloromethane (9:1 v/v) to which the internal standard was added prior to homogenization. HPLC analyses were performed as previously described (Yang et al., 2011). Tocopherols and tocotrienols were resolved on an Eclipse XDB-C18 reversed-phase column (4.6 × 150 mm length; 5 μm particle size, Agilent, www.agilent.com) using a solvent system comprising methanol/water (95:5 v/v) at flow rate of 1.5 ml/min. Analytes were detected and quantified by fluorescence (excitation at 292 nm; emission at 330 nm) using response factors determined for each tocopherol and tocotrienol form.

Homogentisate measurement

The homogentisate content of plant tissues was measured using a Prominence UPLC LC-MS system (Shimadzu, www.shimadzu.com) linked to a QTRAP4000 (ABSciex, www.absciex.com) mass spectrometer, as described by Karunanandaa et al. (2005), with modifications that are described in detail in Methods S1.

Acknowledgements

We thank Dean DellaPenna (Department of Biochemistry, Michigan State University, East Lansing, Michigan) for Arabidopsis vte2-1 seeds. This research was supported by a grant from the US Department of Agriculture National Research Initiative 2004-35318-14887. We also thank the National Natural Science Foundation of China (grant number 31071453) and the 111 Project (grant number B07041) for funding. We thank Tara Nazarenus for expert technical support.

Ancillary