Vitamin E biosynthesis: functional characterization of the monocot homogentisate geranylgeranyl transferase


  • Wenyu Yang,

    1. Donald Danforth Plant Science Center, 975 North Warson Road, Saint Louis, MO 63132, USA
    Search for more papers by this author
  • Rebecca E. Cahoon,

    1. Donald Danforth Plant Science Center, 975 North Warson Road, Saint Louis, MO 63132, USA
    2. Department of Biochemistry and Center for Plant Science Innovation, 1901 Vine Street, University of Nebraska, Lincoln, NE 68588, USA
    Search for more papers by this author
  • Sarah C. Hunter,

    1. Plant Genetics Research Unit, United States Department of Agriculture-Agricultural Research Service, Donald Danforth Plant Science Center, 975 North Warson Road, Saint Louis, MO 63132, USA
    Search for more papers by this author
    • Present address: Think Tank, Robert L. Nugent Building, 1212 E. University Boulevard, University of Arizona, Tucson AZ 85721, USA.

  • Chunyu Zhang,

    1. Donald Danforth Plant Science Center, 975 North Warson Road, Saint Louis, MO 63132, USA
    2. Department of Biochemistry and Center for Plant Science Innovation, 1901 Vine Street, University of Nebraska, Lincoln, NE 68588, USA
    3. National Key Laboratory of Crop Genetic Improvement, Huazhong Agricultural University, Wuhan 430070, China
    Search for more papers by this author
  • Jixiang Han,

    1. Donald Danforth Plant Science Center, 975 North Warson Road, Saint Louis, MO 63132, USA
    Search for more papers by this author
  • Trissa Borgschulte,

    1. Plant Genetics Research Unit, United States Department of Agriculture-Agricultural Research Service, Donald Danforth Plant Science Center, 975 North Warson Road, Saint Louis, MO 63132, USA
    Search for more papers by this author
    • Present address: SAFC Biosciences, 2909 Laclede Avenue, Saint Louis, MO 63103, USA.

  • Edgar B. Cahoon

    Corresponding author
    1. Donald Danforth Plant Science Center, 975 North Warson Road, Saint Louis, MO 63132, USA
    2. Department of Biochemistry and Center for Plant Science Innovation, 1901 Vine Street, University of Nebraska, Lincoln, NE 68588, USA
    3. Plant Genetics Research Unit, United States Department of Agriculture-Agricultural Research Service, Donald Danforth Plant Science Center, 975 North Warson Road, Saint Louis, MO 63132, USA
    Search for more papers by this author

(fax +01 402 472 3139; e-mail


The biosynthesis of the tocotrienol and tocopherol forms of vitamin E is initiated by prenylation of homogentisate. Geranylgeranyl diphosphate (GGDP) is the prenyl donor for tocotrienol synthesis, whereas phytyl diphosphate (PDP) is the prenyl donor for tocopherol synthesis. We have previously shown that tocotrienol synthesis is initiated in monocot seeds by homogentisate geranylgeranyl transferase (HGGT). This enzyme is related to homogentisate phytyltransferase (HPT), which catalyzes the prenylation step in tocopherol synthesis. Here we show that monocot HGGT is localized in the plastid and expressed primarily in seed endosperm. Despite the close structural relationship of monocot HGGT and HPT, these enzymes were found to have distinct substrate specificities. Barley (Hordeum vulgare cv. Morex) HGGT expressed in insect cells was six times more active with GGDP than with PDP, whereas the Arabidopsis HPT was nine times more active with PDP than with GGDP. However, only small differences were detected in the apparent Km values of barley HGGT for GGDP and PDP. Consistent with its in vitro substrate properties, barley HGGT generated a mixture of tocotrienols and tocopherols when expressed in the vitamin E-null vte2-1 mutant lacking a functional HPT. Relative levels of tocotrienols and tocopherols produced in vte2-1 differed between organs and growth stages, reflective of the composition of plastidic pools of GGDP and PDP. In addition, HGGT was able to functionally substitute for HPT to rescue vte2-1-associated phenotypes, including reduced seed viability and increased fatty acid oxidation of seed lipids. Overall, we show that monocot HGGT is biochemically distinct from HPT, but can replace HPT in important vitamin E-related physiological processes.


Tocotrienols and tocopherols comprise the vitamin E class of lipid soluble antioxidants in plants. These molecules are composed of a polar chromanol head group derived from the shikimate pathway bound to a C20 isoprenoid-derived hydrocarbon tail. Tocotrienols and tocopherols differ only in their degree of unsaturation: the tocotrienol hydrocarbon chain contains three trans double bonds, whereas the tocopherol hydrocarbon chain is fully saturated. Within each class of vitamin E, four forms occur in plants α, β, γ and δ that differ in the numbers and positions of methyl residues on the chromanol head group. The α form of tocotrienols and tocopherols contains three methyl groups on the chromanol ring, the β and γ forms contain two methyl groups on the chromanol ring, but in different positions, and the δ form contains only one methyl group. Collectively, the eight forms of tocotrienols and tocopherols are referred to as tocochromanols.

Tocotrienols and tocopherols are potent lipid soluble antioxidants. This property is exemplified in planta by the demonstrated role of tocopherols in promoting seed longevity by reducing the accumulation of lipid oxidation products during storage (Sattler et al., 2004). Tocotrienols have also been shown to reduce lipid peroxidation in tobacco leaves engineered to produce this form of vitamin E, under growth conditions of high light and low temperature (Matringe et al., 2008). Although tocotrienols and tocopherols both function as antioxidants, these two classes of tocochromanols and the individual forms of each have distinct biological activities and physical properties. α-Tocopherol, for example, is generally considered to be the most nutritionally beneficial form of vitamin E because it is the most readily absorbed and retained by the body (Kamal-Eldin and Appelqvist, 1996). Yet tocotrienols have been shown to inhibit the growth of breast cancer cells and act as serum cholesterol-lowering agents (Qureshi et al., 1986; Nesaretnam et al., 1995, 1998). In addition, the γ and δ forms of tocopherols and especially tocotrienols confer the greatest degree of oxidative stability to vegetable oils exposed to prolonged higher temperatures (Wagner and Elmadfa, 2000; Wagner et al., 2001; Warner et al., 2003). This property is particularly important for the performance of vegetable oils in food processing and bio-based lubricants (Durrett et al., 2008).

Of the two classes of tocochromanols, tocopherols occur more widely in plants. Tocopherols, typically in the α form, are abundant in leaves of all plants, and are also enriched in seeds of most dicots and seed embryos of monocots (Cahoon et al., 2003; Horvath et al., 2006a,b; Falk and Munne-Bosch, 2010). By contrast, the occurrence of tocotrienols is limited primarily to the seed endosperm of monocots and some dicots, including tobacco, grape and members of the Apiaceae family, where they are the major class of tocochromanols (Horvath et al., 2006b; Falk and Munne-Bosch, 2010). Tocotrienols have also been reported in a variety of non-seed portions of diverse plant species, including the latex of Hevea brasiliensis (Whittle et al., 1966; Chow and Draper, 1970).

The biosynthesis of tocopherols occurs in plastids of plant cells. The initial step in tocopherol biosynthesis is the condensation of homogentisate and phytyl diphosphate (PDP) to form 2-methyl-6-phytylbenzoquinol (Soll et al., 1980; Soll and Schultz, 1980; Soll, 1987) (Figure 1). This reaction is catalyzed by homogentisate phytyltransferase (HPT), which is encoded by VTE2 in Arabidopsis (Collakova and DellaPenna, 2001; Savidge et al., 2002). For the synthesis of α-tocopherol, the initial HPT-catalyzed condensation reaction is followed by methylation, cyclization to form the chromanol head group and a second methylation (Hunter and Cahoon, 2007; Figure 1). HPT is a member of the membrane-associated UbiA prenyltransferase family that includes enzymes such as chlorophyll synthase (Savidge et al., 2002; Venkatesh et al., 2006). Arabidopsis HPT is most active with PDP, but also has low activity with the C20 isoprenoid geranylgeranyl diphosphate (Sadre et al., 2006). HPT, however, has no detectable activity with C45 isoprenoid solanesyl diphosphate (Sadre et al., 2006).

Figure 1.

 Tocotrienol and tocopherol biosynthesis.
The initial prenylation reactions in tocotrienol and tocopherol biosynthesis are catalyzed by homogentisate geranylgeranyl transferase (HGGT) and homogentisate phytyltransferase (HPT), respectively. Tocotrienol biosynthesis is initiated by the prenylation of homogentisate with geranylgeranyl diphosphate (GGDP) to form 2-methyl-6-geranylgeranyl benzoquinol, and tocopherol biosynthesis is initiated by the prenylation of homogentisate with phytyl diphosphate (PDP) to form 2-methyl-6-phytyl benzoquinol. PDP is synthesized by the reduction of geranylgeranyl, either as a diphosphate or as the side-chain of chlorophyll, by the activity of geranylgeranyl reductase.

Tocotrienol biosynthesis is believed to involve reactions analogous to those associated with tocopherol biosynthesis (Soll et al., 1980) (Figure 1). The only difference is that the initial condensation reaction is presumed to use geranylgeranyl diphosphate (GGDP) instead of PDP, given the similarity in unsaturation between GGDP and the hydrocarbon chain of tocotrienols (Horvath et al., 2006b). Consistent with this, we have previously reported the isolation of cDNAs for a structural variant of HPT from the monocots barley (Hordeum vulgare), wheat (Triticum aestivum) and rice (Oryza sativa), designated ‘homogentisate geranylgeranyl transferase’ (or ‘HGGT’) (Cahoon et al., 2003). We also demonstrated that the transgenic expression of HGGT alone is sufficient to confer tocotrienol biosynthesis to plant organs and cells, such as Arabidopsis leaves and tobacco callus, which do not normally accumulate this form of vitamin E (Cahoon et al., 2003). Monocot HGGTs identified to date are related to HPTs, including those from monocot species, but share <50% amino acid sequence identity (Cahoon et al., 2003; Venkatesh et al., 2006). Consistent with the restricted accumulation of tocotrienols in seed endosperm of monocots, expression of HGGT in barley was detected in seeds but was absent from leaves and roots (Cahoon et al., 2003). Overall, our previous results provide evidence for the evolution of a novel enzyme HGGT in monocots that is responsible for the production of the tocotrienols in the seeds of these plants (Cahoon et al., 2003). Apart from our previous identification of monocot HGGTs, no detailed characterization of the biochemical and in planta properties of this enzyme has been described. In this report, we provide a detailed characterization of the monocot HGGT, including information on its subcellular localization and expression patterns in monocots. We also show monocot HGGT is most active with GGDP, but has activity with PDP, and can yield mixtures of tocotrienols and tocopherols in planta, the relative levels of which are likely to be dictated by the available pools of these substrates. We further show, using a vitamin E-null mutant of Arabidopsis, that the monocot HGGT can functionally replace HPT in physiological processes such as seed longevity and the associated oxidative stability of seed oils.


Monocot HGGT-mediated tocotrienol synthesis is localized to plastids

The biosynthesis of tocotrienols is hypothesized to occur in plastids because of the close relationship between the tocopherol and tocotrienol biosynthetic pathways. To gain insight into the localization of tocotrienol biosynthesis in monocot seeds, the coding sequence for the N-terminal 49 amino acids of the barley HGGT was linked to the 5′ end of the GFP coding sequence. This length of sequence was chosen in part because it is predicted to encode a plastid transit peptide based on in silico analysis using the ChloroP v1.1 server (data not shown). The amino acid sequence also immediately precedes the start of homology between HGGTs and HPTs. As a control, the coding sequence for the N-terminal 36 amino acids of the Arabidopsis HPT was linked to the 5′ end of the GFP coding sequence. The resulting transgenes were assembled under the control of the CaMV 35S promoter, and were transiently expressed in Nicotiana tabacum (tobacco) leaves. Expression of both genes resulted in co-localization of GFP with chlorophyll fluorescence in chloroplasts (Figure 2). This result is consistent with the ability of N-terminal sequences of the Arabidopsis HPT and barley HGGT to confer plastid localization. By contrast, expression of GFP without added N-terminal amino acids resulted in no detectable chloroplast localization of GFP (Figure 2). These results indicate that barley HGGT, like Arabidopsis HPT, contains a plastid transit peptide. It is notable that the N-terminal sequences of wheat and rice HGGTs share significant homology with the barley HGGT transit peptide (data not shown). These results are consistent with the localization of tocotrienol biosynthesis in plastids of monocots.

Figure 2.

 The N-terminal amino acid sequences of HGGT and HPT contain plastid transit peptides.
Subcellular localization studies were conducted with the translational fusions of the N-terminal 49 amino acids of barley HGGT and the N-terminal 36 amino acids of Arabidopsis HPT with GFP, as well as with GFP lacking added N-terminal sequences. The coding sequences for these fusion proteins were transiently expressed in leaves of Nicotiana tabacum. Proteins were localized with confocal fluorescence microscopy. Shown are chlorophyll fluorescence (left panel), subcellular localization of GFP (middle panel), and overlays of chlorophyll fluorescence and GFP fluorescence micrographs (right panels). Scale bars: 10 μm.

HGGT and HPT genes are differentially expressed in barley and wheat

We have previously shown by Northern blot analysis that barley HGGT expression is detectable in the caryopsis, but is not detectable in leaves and roots (Cahoon et al., 2003). Rice HGGT is also highly expressed in caryopsis, as indicated by microarray analysis (Chaudhary and Khurana, 2009). The availability of the Barley1 22k GeneChip probe array allows for a more detailed analysis of HGGT expression, particularly during seed development and within different portions of the caryopsis, as well as comparative analysis of HGGT and HPT expression (Druka et al., 2006). Data shown in Figure 3 were obtained from the minimum information about a microarray experiment (MIAME)-compliant BarleyBase with probe sets corresponding to the HPT and HGGT. Based on these data, HPT is more highly expressed in all vegetative and floral organs, except anthers. During development of the caryopsis, HGGT expression is considerably higher than that of HPT, especially at 10 and 16 days after pollination (DAP). However, dissection of the caryopsis at 22 DAP revealed that HGGT expression is largely limited to the endosperm, whereas HPT expression predominates in the embryo. A similar dichotomy of HPT and HGGT expression between embryo and endosperm was also found for wheat using data from the publicly available Wheat 61k GeneChip (Schreiber et al., 2009) (Figure S1). The expression differences of HPT and HGGT in the monocot caryopsis are consistent with a previous report showing that tocotrienols are nearly the exclusive tocochromanol in the endosperm of two barley cultivars, and that conversely, tocopherols are nearly the exclusive tocochromanol in the embryos of these cultivars (Falk et al., 2004).

Figure 3.

 Microarray expression analysis of barley HPT and HGGT genes.
Comparison of expression profiles of barley HPT and HGGT genes derived from the Barley1 22k microarray. Data were obtained from Experiment BB3 from the BarleyBase database with probe sets derived from Sequence ID hv11h02u_at (HPT) and Sequence ID contig9138_at (HGGT). The results shown are from three independent hybridizations (±SD) with mas v5.0 normalization.

For comparison, the expression patterns of additional genes for enzymes in the vitamin E tocochromanol biosynthetic pathways in barley and wheat were examined using public microarray data (Figure S1). Unlike HGGT, genes for other vitamin E biosynthesis pathways, including tocopherol/tocotrienol cyclase (VTE1), 2-methyl-6-prenylbenzoquinol and γ-tocopherol/tocotrienol methyltransferases (VTE3 and VTE4), geranylgeranyl reductase and hydroxyphenylpyruvate dioxygenase, are expressed in the endosperm and embryo of seeds, as well as in vegetative and floral organs of barley and wheat. This finding is consistent with the specialized involvement of HGGT in the biosynthesis of tocotrienols, and the general activity of other vitamin E biosynthetic enzymes in the synthesis of both tocotrienols and tocopherols in monocots.

Barley HGGT and Arabidopsis HPT have distinct substrate properties

In vitro assays were conducted to establish the substrate preference of barley HGGT for GGDP, PDP and solanesyl diphosphate (SDP), a C45 isoprenoid and precursor of plastoquinone. Assays were complicated by difficulties in expressing HGGT in Escherichia coli. Numerous attempts to express the barley HGGT in E. coli failed to yield detectable levels of recombinant protein or active enzyme, probably because of toxicity from the membrane-associated nature of HGGT. Given the reported success of insect cells as hosts for expression of the Synechocystis HPT (Savidge et al., 2002), a similar strategy was used for barley HGGT. For these studies, barley HGGT was expressed under the control of the polyhedron promoter in Sf-9 cells using baculovirus-mediated transfection. The Arabidopsis HPT was also expressed in these cells for the comparison of its substrate properties. Homogentisate prenyltransferase activity was detectable in microsomes of Sf-9 cells expressing the barley HGGT and Arabidopsis HPT. This activity was absent from non-transfected Sf-9 (data not shown). Each enzyme was assayed in Sf-9 microsomes with GGDP, PDP or SDP and [14C]homogentisate. The barley HGGT was approximately six times more active with GGDP than with PDP (Figure 4a). Conversely, the Arabidopsis HPT was approximately nine times more active with PDP than with GGDP (Figure 4b). Neither enzyme displayed detectable activity with SDP (results not shown). These findings clearly show that the barley HGGT and Arabidopsis HPT are biochemically distinct enzymes.

Figure 4.

 Characterization of the activities of recombinant barley HGGT and Arabidopsis HPT.
Enzyme activities were measured using microsomal extracts of insect Sf-9 cells expressing barley HGGT or Arabidopsis HPT. Activities were determined by measurement of the prenylation of [U-14C]homogentisate with GGDP or PDP.
(a) Activity of barley HGGT with GGDP or PDP.
(b) Activity of Arabidopsis HPT with GGDP or PDP.
(c) Kinetics of barley HGGT activity with differing concentrations of GGDP or PDP. Measurements shown in (a) and (b) were obtained from three independent assays (±SD) with 3 μm of GGDP or PDP, and with reaction conditions yielding linear activities.

Results from kinetic analyses revealed only small differences in the apparent Km values of recombinant barley HGGT with GGDP (apparent Km = 0.14 ± 0.02 μm, n = 3 ± SD) and PDP (apparent Km = 0.40 ± 0.03 μm, n = 3 ± SD) (Figure 4c). Instead, the Vmax of barley HGGT with GGDP was six- to eightfold higher than with PDP. For example, in the experiment shown in Figure 4c, the Vmax of barley HGGT was 0.43 pmole min−1 mg protein−1 with GGDP and 0.06 pmole min−1 mg protein−1 with PDP. The small difference in Km with GGDP and PDP, but large difference in Vmax with these substrates indicates that HGGT has a similar affinity for GGDP and PDP, but has a higher turnover rate with GGDP.

These results along with the findings above indicate that HGGT can function in the synthesis of tocopherols, in addition to its primary role in tocotrienol synthesis, depending on the relative pool sizes of GGDP and PDP. To test this, competition studies were conducted by measuring barley HGGT activity with different ratios of GGDP and PDP mixtures. The 2-methyl-6-geranylgeranylbenzoquinol and 2-methyl-6-phytylbenzoquinol products of HGGT activity with GGDP and PDP, respectively, were resolved by reverse-phase thin layer chromatography, and the incorporation of 14C into each product was measured. With a 10:1 relative concentration of GGDP : PDP, 2-methyl-6-geranylgeranylbenzoquinol was nearly the exclusive product of HGGT (Figure 5). At a relative concentration of 1:1 of the two substrates, seven times more 2-methyl-6-geranylgeranylbenzoquinol than 2-methyl-6-phytylbenzoquinol was formed. This was consistent with the relative difference in activity when GGDP and PDP were assayed in separate reactions (Figure 4a). Using a 1:10 relative concentration of GGDP : PDP, HGGT still produced nearly 1.5-fold more 2-methyl-6-geranylgeranylbenzoquinol than 2-methyl-6-phytylbenzoquinol. It is notable that the total activity of HGGT declined in the presence of higher relative concentrations of PDP, which probably reflects the lower activity of HGGT with PDP (Figure 4a). Overall, the results from substrate competition assays show that HGGT primarily uses GGDP as a substrate, but is capable of generating the 2-methyl-6-phytylbenzoquinol precursor of tocopherols, particularly in the presence of elevated ratios of PDP : GGDP.

Figure 5.

 Substrate competition assays with barley HGGT.
Enzyme assays were conducted with differing ratios of GGDP and PDP added as substrates. Reactions were conducted with 10 μm GGDP and 1 μm PDP (10:1), 1 μm GGDP and 1 μm PDP (1:1) or 1 μm GGDP and 10 μm PDP (1:10). Assays were conducted with reaction conditions yielding linear activity, and products from each substrate were separated by reverse-phase TLC. Activity was measured from three independent assays for each ratio of substrates (±SD).

In planta characterization of monocot HGGT activity

The activity of a monocot HGGT in planta was examined by expression of the barley enzyme in the vte2-1 Arabidopsis mutant (Figure 6a). This mutant lacks detectable levels of tocochromanols because of the disruption of HPT (Sattler et al., 2004). Constitutive expression of the barley HGGT in the vte2-1 background produced primarily tocotrienols in leaves of 4-week-old plants (Figure 6b), whereas expression of a wild-type HPT resulted in the accumulation of tocopherols (Figure 6c). Consistent with the mixed activity of HGGT with GGDP and PDP in vitro, tocopherols were also detected in leaves of 4-week-old vte2-1 plants expressing the barley enzyme, although at relative levels of ≤10% of the total tocochromanol content.

Figure 6.

 Tocotrienol and tocopherol composition of leaves from 4-week-old Arabidopsis vte2-1 expressing barley HGGT or Arabidopsis HPT.
Shown in each panel are chromatograms from HPLC analyses of tocotrienols (T3) and tocopherols (T), as detected by fluorescence emission.
(a) Leaves from non-transformed vte2-1 plants.
(b) Leaves from vte2-1 plants constitutively expressing barley HGGT.
(c) Leaves from vte2-1 plants constitutively expressing Arabidopsis HPT.
(d) Tocotrienol and tocopherol standards α, β, γ and δ forms of each tocochromanol class. (Note: β and γ forms of tocotrienols and tocopherols are not resolved by this HPLC method.)

It was expected that relative ratios of tocotrienols and tocopherols should vary with leaf age, and in other organs of the vte2-1 plants expressing the barley HGGT in a manner that reflects the pool sizes of GGDP and PDP, based on the enzyme assays described above. In this regard, tocotrienols accounted for 85–90% of the tocochromanols in leaves from 2-, 4- and 6-week-old plants (Figure 7). However, at 8 and 11 weeks, the relative level of tocotrienols decreased to ∼60% of the total tocochromanol content (Figure 7). The increase in tocopherol content coincided with the onset of senescence in the vte2-1 plants. It is known that tocopherol content increases during senescence presumably because of the increased availability of PDP from chlorophyll degradation (Collakova and DellaPenna, 2003; Ischebeck et al., 2006). Therefore, an increase in the content of PDP relative to GGDP as substrates for HGGT in vivo probably explains the increased tocopherol content in the older leaves (Collakova and DellaPenna, 2003; Ischebeck et al., 2006). To examine the converse metabolic scenario, barley HGGT was expressed in a geranylgeranyl reductase mutant that contains GGDP but lacks the ability to produce the PDP substrate for tocopherol synthesis (Figure 1). In this background, expression of HGGT was accompanied by the production of just the tocotrienol form of vitamin E (Figures 8 and S2).

Figure 7.

 Expression of barley HGGT in the vitamin E null vte2-1 mutant results in the production of tocotrienol and tocopherol mixtures, the composition of which are age-dependent and differ among organs and seeds.
Measurements of total tocotrienol and tocopherol content were conducted on leaves of differing ages, roots, flowers and mature seeds from vte2-1 plants constitutively expressing barley HGGT. The values shown are from between three and five measurements of independent samples (±SD) for leaves of each age, and for organs and seeds.

Figure 8.

 Expression of barley HGGT in an Arabidopsis geranylgeranyl reductase mutant results in the accumulation of only the tocotrienol form of vitamin E.
Shown are HPLC chromatograms of tocotrienol (T3) and tocopherol (T) analyses from 3-week-old Arabidopsis seedlings maintained on plates. Transgenic lines were obtained by transformation of a heterozygous geranylgerarnyl reductase ggr mutant with constitutively expressed barley HGGT. Transformants were selected with a DsRed marker and genotyped to determine the homozygosity of the ggr mutant allele (Figure S2).
(a) Non-transformed seedlings homozygous for the ggr mutant allele.
(b) Transgenic seedlings expressing barley HGGT in a homozygous ggr background.
(c) Tocotrienol and tocopherol standards: α, β, γ and δ forms for each tocochromanol class. Samples and standards also contain the internal standard (IS) 5′,7′-dimethyltocol.

Examination of other organs in vte2-1 plants expressing barley HGGT revealed marked differences in the levels of tocotrienols and tocopherols. For example, tocotrienols accounted for ∼60% of tocochromanols in mature seeds, but composed ∼90% of tocochromanols in roots and ∼75% of tocochromanols in flowers (Figure 7). In the case of roots, PDP pools that support HGGT-mediated tocopherol synthesis are likely to be at a low level because of the near absence of chlorophyll, whereas PDP is likely to be more abundant in seeds, as chlorophyll is degraded during maturation (Ischebeck et al., 2006).

Monocot HGGT can functionally replace HPT

Although vitamin E tocochromanols are not required for growth in Arabidopsis, vte2-1 plants have distinct phenotypic defects, including low seed viability following extended storage at room temperature that is manifested by the elevated accumulation of fatty acid oxidation products (Sattler et al., 2004). In addition, the growth of vte2-1 plants is impaired upon prolonged exposure to low temperatures, which has been attributed in part to defects in phloem loading (Maeda et al., 2006).

To test the ability of HGGT to rescue these phenotypic abnormalities, seeds from vte2-1 plants expressing the barley HGGT gene under the control of the CaMV 35S promoter were exposed to accelerated aging by incubation at 38°C for 48 h at 100% humidity (tocochromanol compositions of seeds are shown in Table S1). Following this treatment, the germination rate of seeds from non-transformed vte2-1 plants was reduced to 10% (Figure 9a). By comparison, seeds from vte2-1 lines expressing the barley HGGT had germination rates ranging from 80 to 98%. This was similar to that of seeds from wild-type Arabidopsis Col-0 plants, which had germination rates of 89% after accelerated aging. Consistent with their large decline in viability, artificially aged vte2-1 seeds contained nearly four-fold higher levels of lipid hydroperoxides compared with non-aged vte2-1 seeds (Figure 9b). In contrast, only small increases in lipid hydroperoxides were detected in wild-type seeds and in vte2-1 seeds expressing the barley HGGT. These findings indicate that HGGT can rescue the loss of viability and the associated increase in lipid hydroperoxides observed in aged seeds of the vte2-1 mutant.

Figure 9.

 Expression of barley HGGT rescues reduced viability and accumulation of fatty acid oxidation products in artificially aged vte2-1 seeds.
(a) Germination rates of seeds with or without exposure to artificial aging.
(b) Lipid hydroperoxide concentrations in seeds with or without exposure to artificial aging. Shown are data from seeds of wild-type Col-0, vte2-1 and independent lines of vte2-1 engineered to constitutively express barley HGGT.

We also observed that exposure of the vte2-1 mutant to low temperatures for extended periods resulted in a reduction in rosette size, similar to previous reports (Maeda et al., 2006), and chlorosis of mature leaves relative to those of wild-type Col-0 (Figure 10). As with the seed viability phenotype, expression of HGGT was able to restore rosette size to that observed in the wild type, and to prevent the extensive chlorosis of mature leaves (Figure 10).

Figure 10.

 Expression of barley HGGT rescues low-temperature growth defects in vte2-1.
Col-0, vte2-1 and vte2-1 expressing barley HGGT were grown at 22°C for 3 weeks and then shifted to 8.5°C for 8 weeks, as shown.


This report builds on our previous identification of the monocot HGGT, and our demonstration that transgenic expression of HGGT can confer tocotrienol biosynthesis to plants and organs that do not normally produce this form of vitamin E (Cahoon et al., 2003). Here we show that HGGT, like HPT, contains a transit peptide for plastid localization. Unlike HPT, which is expressed in the embryo, HGGT expression is detected primarily in the endosperm of monocot seeds. In addition, HGGT was found to have the highest activity with the tocotrienol side-chain precursor GGDP, but also retains an approximately sixfold lower activity with the tocopherol side-chain precursor PDP. Despite the differences in activities with GGDP and PDP, HGGT has similar Km values for both substrates. Consistent with these biochemical properties, it was further demonstrated that HGGT expression in an HPT-null mutant of Arabidopsis can generate tocotrienols and tocopherols, but can only generate tocotrienols in a PDP-deficient geranylgeranyl reductase ggr mutant. These results indicate that the functional outcome of HGGT in planta is also dependent on the sizes and availability of GGDP and PDP substrate pools. Finally, we show that HGGT can functionally replace HPT to rescue vitamin E-deficiency phenotypes, including reduced seed longevity and the associated accumulation of lipid hydroperoxides from polyunsaturated fatty acid oxidation.

It is likely that HGGT arose from the more widely occurring HPT given the close structural relationship between these enzymes. This raises the question of why monocots evolved a specialized enzyme for the synthesis of an alternative form of vitamin E. Based on our findings and previous reports on tocopherol synthesis, the monocot HGGT probably evolved to fit a biochemical and functional niche. Tocopherol synthesis requires PDP, whereas tocotrienol synthesis requires GGDP. One possibility is that PDP pools in monocot endosperm are too low to support synthesis of sufficient levels of tocopherols for extended viability of monocot seeds. This isoprenoid diphosphate is also a precursor of chlorophyll, and the flux of PDP appears to be primarily directed towards chlorophyll synthesis based on studies of the phytol kinase vte5-1 mutant of Arabidopsis (Figure 1) (Ischebeck et al., 2006; Valentin et al., 2006). Phytol kinase is associated with the conversion of phytol released from chlorophyll to PDP, which can be used for tocopherol synthesis (Ischebeck et al., 2006). Consistent with this, the vte5-1 mutant retains only 20% of the wild-type levels of tocopherols in seeds (Valentin et al., 2006). It is known that chlorophyll synthase can readily use GGDP for incorporation as the hydrophobic side chain of chlorophyll (Soll et al., 1983; Tanaka et al., 1999), and that geranylgeranyl reductase can use geranylgeranyl chlorophyll as a substrate for the synthesis of phytol (Keller et al., 1998). This information collectively indicates that PDP is synthesized predominantly, but not exclusively, for tocopherol production via chlorophyll metabolism. Given the lack of chlorophyll in monocot endosperm, it is possible that these cells lack PDP. It also cannot be excluded that monocot endosperm has low geranylgeranyl reductase activity for phytol synthesis. Either metabolic scenario would result in a deficiency in PDP to maintain the synthesis of tocopherols. Instead, monocot endosperm may be more enriched in GGDP, which is also a precursor of other molecules, including carotenoids and gibberellins. The evolution of HGGT would therefore allow the synthesis of tocotrienols in monocot endosperm in the absence or deficiency of PDP. It is notable that the lack of reliable methods to measure GGDP and PDP, and the inability to distinguish cytosolic and plastidic GGDP, preclude the direct quantitation of the relative composition and total concentrations of pools of these isoprenoid diphosphates in plant extracts.

Although tocopherols are not essential in Arabidopsis, their absence in the HPT vte2-1 mutant results in the severe reduction of seed viability following prolonged storage, as a result of the increased oxidation of polyunsaturated fatty acids (Sattler et al., 2004). This finding indicates that tocopherols provide a selective advantage to Arabidopsis. Tocotrienols are also potent antioxidants, and have been shown in a variety of in vitro studies to be superior to tocopherols for quenching reactive oxygen species (Serbinova et al., 1991; Suarna et al., 1993; Suzuki et al., 1993; Serbinova and Packer, 1994; Kamal-Eldin and Appelqvist, 1996; Wagner et al., 2001). Tocotrienols have also been shown to have antioxidant capacity in the leaves of transgenic tobacco (Matringe et al., 2008). As such, the presence of HGGT in the monocot endosperm probably facilitates the synthesis of sufficient levels of antioxidants for extended seed viability. Our finding that HGGT can rescue the reduced viability of artificially aged vte2-1 seeds suggests that HGGT is important for promoting the longevity of monocot seeds.

It is notable that tocotrienols are absent or detected at low levels in most dicots and in most organs of monocots. Instead, HPT activity in these species and organs predominantly gives rise to tocopherols. This result is surprising in light of our observation that transgenic expression of HGGT can generate both tocotrienols and tocopherols in different organs of vte2-1 mutants as a result of the activity of HGGT with GGDP and PDP. Not unlike HGGT, HPT displays in vitro activity with PDP and GGDP, although this enzyme is approximately nine times more active with PDP. HPT-mediated tocotrienol synthesis is also not detected in antisense geranylgeranyl reductase-suppression lines of tobacco (Tanaka et al., 1999), or in the geranylgeranyl reductase mutant of Arabidopsis that is deficient in PDP but retains the ability to synthesize GGDP (Figure 8). These observations suggest that PDP pools are typically more enriched in most plant cells, or that GGDP pools are typically too low to support HPT-mediated tocotrienol synthesis, given that HPT competes for GGDP with other biosynthetic pathways (e.g. carotenoid and gibberellin biosynthetic pathways). However, high levels of tocotrienol production have been reported in tobacco, Arabidopsis and soybean (Glycine max) seeds engineered to synthesize elevated levels of the homogentisate substrate of tocopherols and tocotrienols, in the absence of transgenic expression of HGGT (Herbers, 2003; Rippert et al., 2004; Karunanandaa et al., 2005). Assuming that tocotrienols are produced under these conditions by HPT-mediated prenylation, high levels of HGA synthesis may enhance the total pool size of GGDP, perhaps by the activation of the plastid isoprenoid pathway, to enable the synthesis of tocotrienols.

The HPTs and HGGTs are members of the UbiA prenyltransferase family, and share 40–50% amino acid sequence identity (Cahoon et al., 2003). Their active sites are characterized by the aspartic acid-containing motifs required for binding Mg2+ and the diphosphate residue of the isoprenoid substrate (Melzer and Heide, 1994; Brauer et al., 2004, 2008). The sequence conservation between HGGTs and HPTs makes it possible to explore the structural basis for their different substrate properties. A particularly useful tool is cpdl (conserved property difference locator), which identifies amino acid residues that may confer functional differences between closely related proteins (Mayer et al., 2005). Whereas most alignment programs account for the strict conservation of amino acid residues, cpdl flags residues that share similar functional properties (e.g. size, hydrophobicity, charge, polarity and aromaticity) when comparing two groups of related sequences. Using cpdl, seven conservative and 10 non-conservative amino acid differences were identified between HGGT and HPT enzyme classes (Appendix S1; Figure S3), with high-stringency parameters for residue conservation in the HGGT group. Membrane topology programs such as sosui (Hirokawa et al., 1998) predict that HPT and HGGT enzymes contain seven transmembrane helices, with three active site regions located on the same side of the membrane (Figure S4). Seven positions identified by cpdl fall near active site residues, and show differences in size and aromaticity between HGGTs and HPTs, and at one position, a difference in charge is also present. Five positions are predicted to fall on the membrane face opposite the active site residues (Figures S3 and S4). The remaining positions occur within predicted transmembrane helices and tend to be hydrophobic residues that vary in size and aromaticity, with HPTs containing larger or more aromatic residues compared with the HGGTs. These analyses suggest that a relatively small number of amino acid differences account for the distinct isoprenoid diphosphate substrate preferences of these enzymes. Identification of more diverse HGGT genes may aid in precisely mapping the structural basis for evolution of HGGT activity in plants. Possible sources of these genes include not only additional monocot species, but also dicot species from families such as the Apiaceae, which predominantly accumulate tocotrienols in their seeds (Horvath et al., 2006b).

Overall, these studies extend our knowledge of vitamin E metabolism and function in plants, and highlight the evolution of an enzyme that allows for the production of vitamin E antioxidants in the absence of a key substrate (i.e. PDP) for tocopherol synthesis.

Experimental Procedures

Plant material and growth conditions

Arabidopsis Col-0 vte2-1 seeds were a gift from Dean DellaPenna, Michigan State University. SALK_046604, a ggr heterozygous T-DNA insertion line, was kindly provided by Dr Peter Dörmann, University of Bonn, Germany. Seeds of SALK_046604 were surface sterilized and grown on MS agar plates supplemented with 2% sucrose and 100 mg L−1 ampicillin, and grown under continuous light (100 μmol m−2 s−1) at 22°C. Plants were genotyped by two PCR reactions, using a gene-specific primer pair P1, 5′-ACTTCCAATGGTGTTCACAGC-3′; and P2, 5′-GAGAAGAGTGGAGCCACTGTG-3′ to determine homozygosity, and primer P2 and the T-DNA left border primer PLB, 5′-TGGTTCACGTAGTGGGCCATCG-3′, to test for the presence of the T-DNA. Heterozygous ggr plants were grown under ambient light conditions in the lab. Seeds from vte2-1 lines transformed with the barley HGGT (>T4 generation) were sown directly in soil, and plants were maintained at 22°C and 50% humidity under a 16-h light (100 μmol m−2 s−1)/8-h dark cycle.

Transit peptide subcellular targeting studies

N-terminal sequences corresponding to putative plastid transit peptide sequences (based on ChloroP v1.1 analysis) from barley HGGT and Arabidopsis HPT were linked to soluble GFP for subcellular localization studies. The coding sequences for the putative transit peptides were linked in-frame to the 5′ end of the coding sequence for soluble GFP through a two-step overlap PCR strategy, as described in Appendix S1.

Constructs were introduced into N. tabacum (Xanthi-NN) leaves by biolistic transformation. The subcellular localization of the corresponding GFP, with or without N-terminal fusion, was determined 24–44 h after bombardment by confocal microscopy using a Zeiss LSM 510 META microscope (Zeiss, Excitation of chlorophyll and GFP was conducted at 488 nm, and emission was detected using a 650–710-nm bandpass filter for chlorophyll and a 500–550-nm bandpass filter for GFP.

Microarray data mining

Microarray data for barley (H. vulgare cv. Morex) and wheat (T. aestivum cv. Chinese Spring) were obtained from BarleyBase and WheatPLEX MIAME-compliant databases. The data were derived from gene expression atlases from experiments BB3 for barley using the Affymetrix 22k Barley1 GeneChip and experiment TA3 for wheat using the Affymetrix 61k Wheat GeneChip (Druka et al., 2006; Schreiber et al., 2009). Probe sets for HPT from barley and wheat were obtained from Sequence IDs: hv11h02u_at (barley) and Ta.10966.1.S1_at (wheat). Probe sets for HGGT from barley and wheat were obtained from sequence IDs: contig9138_at (barley) and Ta.28314.1.S1_at (wheat). Three biological replicates were used for the measurement of expression levels in different tissues.

Vector construction

The CaMV 35S barley HGGT expression vector pSH24 was described previously (Cahoon et al., 2003). Barley HGGT was expressed under the control of a ubiquitin promoter in the Arabidopsis ggr mutant with DsRed with the binary vector DsRed-GlyRed2-HGGT. Details of the preparation of this vector are provided in Appendix S1.

Arabidopsis transformation and selection

Binary vectors pSH24 and DsRed-GlyRed2-HGGT were introduced into Agrobacterium tumefaciens C58 by electroporation. Arabidopsis (Col-0) vte2-1 or ggr mutants were transformed with the recombinant agrobacterium by use of the floral-dip method (Clough and Bent, 1998). Transgenic lines transformed with pSH24 were selected by resistance to kanamycin resistance, and transgenic seeds from transformation with DsRed-GlyRed2-HGGT were identified by DsRed fluorescence (Jach et al., 2001; Pidkowich et al., 2007).

Expression of Arabidopsis HPT and barley HGGT in insect Sf-9 cells

The Bac-2-Bac system (Invitrogen, was used for heterologous expression in Spodoptera frugiperda (Sf-9) insect cells. The open reading frame for the mature Arabidopsis HPT (lacking the plastid transit peptide sequence) was amplified by PCR with Phusion polymerase (New England Biolabs, from a corresponding cDNA clone using the forward and reverse oligonucleotides: 5′-TTGGATCCACATGGATTCGAGTAAAG-TTGTCGC-3′ and 5′-TTTTGAGCTCACTTCAAAAAAGGTAACAG-3′ (the added restriction enzyme sites are set in bold). The mature barley HGGT open reading frame was amplified by PCR from a cDNA clone using the forward and reverse oligonucleotides: 5′-TTGGATCCGAGGATGGCTACTACTGATAATG-3′ and 5′-TTTTGAGCTCGTACAAATTTCACTGCACAAATG-3′. PCR products were digested with BamHI and SacI and ligated into the corresponding sites of the pFastBac expression vector (Invitrogen). The resulting plasmids were then used to generate recombinant bacmids as described in the manufacturer’s protocol. The empty pFastBac vector was used as a negative expression control. Sf-9 cells at a density of 1 × 106 cells ml−1 were infected with high-titer virus and cultured for an additional 48 h. Cells were harvested by centrifugation and stored at −80°C.

For preparation of microsomes, Sf-9 cell pellets were resuspended in phosphate buffered saline (PBS) containing 1 mm EDTA, 0.4 mm phenylmethylsulfonyl fluoride (PMSF) and 10 mm 2-mercaptoethanol. Cells were lysed by manual homogenization in a 15-ml glass homogenizer on ice with 30 strokes. The lysate was centrifuged for 10 min at 10 000 g to remove cell debris. The resulting supernatant was centrifuged at 500 000 g for 15 min. The microsomal pellet was washed once with the lysis buffer and then resuspended in a buffer consisting of 1′ PBS buffer, 10 mm 2-mercaptoethanol and 20% glycerol, by gentle agitation with a glass rod. Protein concentrations were determined by bicinchoninic acid (BCA) assay (Pierce, with bovine serum albumin (BSA) as a standard. Aliquots of the microsomes were stored at −80°C. Enzyme activity was stable for at least 3 months under these storage conditions.

Prenyltransferase assays

[U-14C]Tyr (specific activity 450 mCi mmol−1), ‘cold’ GGDP and ‘cold’ PDP were purchased from American Radiolabeled Chemicals ( Preparation of the [U-14C]homogentisate (HGA) substrate was synthesized enzymatically in two steps from [U-14C]Tyr, as described in Appendix S1.

Prenyltransferase assays were based on a previously described method (Collakova and DellaPenna, 2003). Each reaction contained PDP and/or GGPP evaporated to dryness under N2 and resuspended in 90 μl of a reaction cocktail consisting of 1–10 μm PDP and/or GGPP, 4.5 μm [U-14C]HGA in 50 mm Tris, pH 7.6, 4 mm MgCl2, 78 mm KPO4 and 10 mm potassium ascorbate. Reactions were initiated by the addition of 100 μg of protein from purified insect microsomes (10 mg ml−1 total protein) containing expressed Arabidopsis HPT or barley HGGT to a final volume of 100 μl. Reactions were incubated at 25°C for between 15 min and 2 h. Lipids were extracted three times with two volumes of acetone : heptane (1:1, v/v). After centrifugation, the ether phases were loaded on a 1-ml Hypersil Si silica solid phase extraction column (ThermoFisher Scientific, equilibrated with 1:5 (v/v) diethyl ether : heptane, and eluted with 0.5 volumes of this solvent. After drying under N2, radioactivity was determined by scintillation counting. Background activities were obtained from control reactions containing 14C-labeled HGA without PDP or GGPP.

For competition assays, the ether layers were collected, concentrated and spotted onto an RP-18 F254S reversed phase thin layer chromatography plate (Merck, The thin layer chromatography plate was developed in acetonitrile : methanol (70:30) and exposed to a phosphorimager screen for 4 days. Radioactive bands were scraped from the TLC plate and quantified by scintillation counting, as described above.

HPLC analysis of tocochromanols

For measurement of tocopherol and tocotrienol concentrations of transgenic Arabidopsis lines, freshly harvested plant material was homogenized in 3 ml of 2:1 (v/v) methanol : chloroform containing butylated hydroxytoluene (0.01% w/v). After 20 min of incubation, 1 ml of chloroform and 1.8 ml of water were added to each sample. Following mixing and centrifugation, the lower organic layer was recovered, dried under nitrogen and resuspended in dichloromethane : methanol (1:5, v/v) for HPLC analysis. 5,7-dimethyltocol (Matreya, was added to each sample as an internal standard prior to extraction. The tocopherol and tocotrienol concentrations of the organic extracts were then determined using an Agilent 1100 HPLC. Resolution of vitamin E species was achieved using an Agilent Eclipse XDB-C18 column (4.6 × 150 mm length; 5 μm particle size) and a solvent system consisting of methanol : water (95:5, v/v) with a flow rate of 1.5 ml min−1. Sample components were detected and quantified by fluorescence with excitation at 292 nm and emission at 330 nm, with detector response factors determined for each tocopherol and tocotrienol species. Sample components were identified by mobility relative to standards and quantified relative to the internal standard.

Controlled deterioration test of seed viability

Seed germination of wild type (Col-0), vte2-1 and vte2-1 lines transformed with the pSH24 CaMV35S-HGGT construct was determined after controlled deterioration (CDT) or accelerated artificial aging. Seeds were placed at 4°C upon harvesting and stored for 1.5 months prior to use in CDT experiments. The CDT was modified based on the procedure described by Bentsink et al. (2000), and performed as follows: bulked seeds were equilibrated at 85% relative humidity and 38°C for 48 h. Immediately after the aging treatment, seeds were surface sterilized with 70% ethanol and 20% bleach, and placed on half-strength MS plus 1% sucrose medium for germination under a temperature regime of 24°C, with a 12-h dark/light cycle and light intensity of 200 μmol m−2 s−1. The germination rate was calculated 4 days after the plating of seeds. Lipids were extracted from ∼6 mg of seeds as described by Welti et al. (2002) and Devaiah et al. (2007). The total lipid extract was dried under N2 and dissolved in 1 ml chloroform and used for ferrous oxidation-xylenol orange (FOX) assay, as previously described, using hydrogen peroxide for generation of a standard curve (DeLong et al., 2002).

Low-temperature growth of vte2-1 with or without expression of HGGT

Arabidopsis (Col-0), vte2-1 and vte2-1 plants expressing HGGT under the control of the CaMV 35S promoter were grown for 3 weeks at 22°C and 50% humidity under a 16-h light (100 μmol m−2 s−1)/8-h dark cycle. Plants were then shifted to growth at 8.5°C and 50% humidity under a 12-h light (100 μmol m−2 s−1)/12-h dark cycle, and phenotypic alterations were noted over 8 weeks of exposure to low temperatures. Tocotrienol and tocopherol production in vte2-1 plants expressing HGGT was confirmed by HPLC analysis.


We thank Dr Dean DellaPenna (Michigan State University) for Arabidopsis vte2-1 seeds, Dr Peter Dörmann (Universität Bonn) for Arabidopsis ggr seeds, and Dr Christopher Taylor (Ohio State University) for the AKK1517 plasmid. We also thank Dr R. Howard Berg (Donald Danforth Plant Science Center) for assistance with microscopy and Ms Jamie Shipp for technical assistance. The research was supported by a grant from the USDA-National Research Initiative 2004-35318-14887.