OPEN ACCESS ARTICLE

This Article Free via Open Access: OA OA Abstract Full Text (PDF) Supplemental Data OA All Versions of this Article: 20/7/1879    most recent tpc.108.061150v1 Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via CrossRef Google Scholar Articles by Zhang, H. Articles by Gong, Z. PubMed PubMed Citation Articles by Zhang, H. Articles by Gong, Z. Agricola Articles by Zhang, H. Articles by Gong, Z.

Dolichol Biosynthesis and Its Effects on the Unfolded Protein Response and Abiotic Stress Resistance in Arabidopsis^[W],[OA]

^a State Key Laboratory of Plant Physiology and Biochemistry, College of Biological Sciences, China Agricultural University, Beijing 100193, China
^b RIKEN Plant Science Center, Tsurumi-ku, Yokohama, Kanagawa 230-0045, Japan
^c Biochimie et Physiologie Moléculaire des Plantes, Institut de Biologie Intégrative des Plantes, Unité Mixte de Recherche 5004, Centre National de la Recherche Scientifique/Unité Mixte de Recherche 0386, Institut National de la Recherche Agronomique/Montpellier SupAgro/Université Montpellier 2, F-34060 Montpellier Cedex 1, France
^d Kihara Institute for Biological Research, Yokohama City University, Yokohama, Kanagawa 244-0813, Japan
^e Department of Botany and Plant Sciences, Institute for Integrative Genome Biology, University of California, Riverside, California 92521
^f China Agricultural University–University of California-Riverside Center for Biological Sciences and Biotechnology, Beijing 100193, China
^g National Center for Plant Gene Research, Beijing 100193, China

² Address correspondence to gongzz{at}cau.edu.cn

	ABSTRACT

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION METHODS REFERENCES

 
Dolichols are long-chain unsaturated polyisoprenoids with multiple cellular functions, such as serving as lipid carriers of sugars used for protein glycosylation, which affects protein trafficking in the endoplasmic reticulum. The biological functions of dolichols in plants are largely unknown. We isolated an Arabidopsis thaliana mutant, lew1 (for leaf wilting1), that showed a leaf-wilting phenotype under normal growth conditions. LEW1 encoded a cis-prenyltransferase, which when expressed in Escherichia coli catalyzed the formation of dolichol with a chain length around C₈₀ in an in vitro assay. The lew1 mutation reduced the total plant content of main dolichols by 85% and caused protein glycosylation defects. The mutation also impaired plasma membrane integrity, causing electrolyte leakage, lower turgor, reduced stomatal conductance, and increased drought resistance. Interestingly, drought stress in the lew1 mutant induced higher expression of the unfolded protein response pathway genes BINDING PROTEIN and BASIC DOMAIN/LEUCINE ZIPPER60 as well as earlier expression of the stress-responsive genes RD29A and COR47. The lew1 mutant was more sensitive to dark treatment, but this dark sensitivity was suppressed by drought treatment. Our data suggest that LEW1 catalyzes dolichol biosynthesis and that dolichol is important for plant responses to endoplasmic reticulum stress, drought, and dark-induced senescence in Arabidopsis.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION METHODS REFERENCES

 
Dolichols are long-chain unsaturated polyisoprenoids present in all eukaryotic cells; they consist of 15 to 23 isoprenic units depending on the species (Chojnacki and Dallner, 1988; Swiezewskaa and Danikiewiczb, 2005). In animals, dolichols are broadly distributed in different cells. They are localized within the two halves of the phospholipid bilayer, with high contents in the plasma membranes, peroxisomes, Golgi vesicles, and lysosomes. Dolichols exist as free alcohols or products derived by phosphorylation, esterification, or glycosylation (Chojnacki and Dallner, 1988; Swiezewskaa and Danikiewiczb, 2005). The biosynthesis of dolichols takes place mainly in the endoplasmic reticulum (ER) and to a lesser extent in peroxisomes (Chojnacki and Dallner, 1988). This biosynthesis is initiated by the formation of farnesyl diphosphate (FPP), primarily through the mevalonate pathway (Chojnacki and Dallner, 1988; Swiezewskaa and Danikiewiczb, 2005). The formation of longer chain polyprenyl diphosphate is catalyzed by enzymes called cis-prenyltransferases, which sequentially add isopentenyl diphosphates (IPPs) onto FPP (Grabinska and Palamarczyk, 2002). The number of IPPs added is species-dependent. Polyprenyl diphosphate is further converted to dolichol and dolichyl phosphate by a series of reactions (Grabinska and Palamarczyk, 2002). Dolichyl phosphate serves as a glycosyl carrier lipid in the biosyntheses of protein C- and O-mannosylation, in glycosylphosphatidylinositol anchors, and is one of the rate-limiting factors in N-linked protein glycosylation in yeast and mammalian cells (Burda and Aebi, 1999; Jones et al., 2005).

The first eukaryotic cis-prenyltransferase–encoding gene, RER2, which is responsible for dolichol synthesis, was cloned from the yeast mutant rer2 (Sato et al., 1999). The multicopy suppressor of rer2, STR1, encodes a protein similar to RER2, but with different physiological roles during cell growth (Sato et al., 2001). rer2 shows many phenotypes, including temperature and hygromycin sensitivity, slow growth, defects in N- and O-glycosylation, and abnormal accumulation of ER and Golgi membranes (Sato et al., 1999). By contrast, the str1 mutant does not have noticeable phenotypes (Sato et al., 2001). Through comparison of cis-prenyltransferase sequences with plant genomic sequences, a homologous gene (AT2G23410) was cloned from Arabidopsis thaliana (Cunillera et al., 2000; Oh et al., 2000). Expression of this gene in yeast rer2 complements its thermosensitive phenotype (Cunillera et al., 2000). A recombinant protein produced from AT2G23410 shows dehydrodolichyl diphosphate synthase activity, and dolichols could be synthesized in an in vitro assay (Cunillera et al., 2000; Oh et al., 2000). It should be noted that Escherichia coli cis-prenyltransferase synthesizes undecaprenyl diphosphate (C₅₅), which functions as a carrier lipid in cell wall polysaccharide synthesis (Kato et al., 1999). However, no dolichol-deficient mutants have been isolated from any multicellular organisms, including plants, and the physiological roles of dolichols in plants are largely unknown.

N-Glycosylation of secreted or membrane proteins occurs, shortly after synthesis, in the lumen of the ER. Inhibition of N-glycosylation often reduces the folding efficiency, enhances incorrect folding, and accelerates the degradation of the hypoglycoprotein. Chaperone proteins such as BiP (for BINDING PROTEIN) play crucial roles in assisting protein folding during ER stress, which can result from N-glycosylation defects (Wilson, 2002). Increases in secretory activity and in the accumulation of unfolded proteins induce the transcripts of chaperones through the unfolded protein response (UPR) pathway (Wilson, 2002). Recently, two BASIC DOMAIN/LEUCINE ZIPPER (bZIP) transcription factor genes, bZIP60 and bZIP28, were isolated, both of which have been shown to activate BiP expression, probably through ER stress response element–like sequences (Iwata and Koizumi, 2005; Liu et al., 2007a). bZIP60 and bZIP28 are membrane proteins that are anchored to the ER membrane under normal conditions. Upon ER stress, bZIP28, and probably also bZIP60, is cleaved and the bZIP domain translocates to the nucleus to activate the expression of BiP (Iwata and Koizumi, 2005; Liu et al., 2007a).

Drought stress is one of the most damaging environmental conditions that limit plant growth and distribution. Drought stress induces the accumulation of the phytohormone abscisic acid, which, as an endogenous signal, plays crucial roles in mediating plant responses in order to adjust the water deficit (Xiong et al., 2002; Christmann et al., 2006; Shinozaki and Yamaguchi-Shinozaki, 2007). During drought stress, many genes are upregulated. Expression of some of these genes is mediated by abscisic acid, while others are abscisic acid–independent (Xiong et al., 2002; Christmann et al., 2006; Shinozaki and Yamaguchi-Shinozaki, 2007). The expression of BiP was found to be induced by water stress in soybean (Glycine max) (Cascardo et al., 2000), and overexpression of BiP in transgenic tobacco (Nicotiana tabacum) improves plant drought tolerance (Alvim et al., 2001). However, the upregulation of BiP is not necessarily related to increased drought tolerance (Koiwa et al., 2003). Gene expression profiling of soybean leaves shows that some genes are induced by both osmotic and UPR stresses, suggesting that the two pathways may converge in plants (Irsigler et al., 2007). These results suggest complex crosstalk between drought stress responses and the UPR pathway (Alvim et al., 2001; Koiwa et al., 2003; Irsigler et al., 2007).

Through a genetic screen, we isolated the Arabidopsis mutant lew1 (for leaf wilting1). Here, we show that lew1 is altered in a gene encoding a cis-prenyltransferase for dolichol synthesis. lew1 mutant plants showed a leaf-wilting phenotype under normal growth conditions due to impaired membrane integrity. lew1 mutant plants were hypersensitive to tunicamycin and had reduced levels of protein glycosylation. Drought stress induced higher expression of the UPR genes BiP and bZIP60 as well as earlier expression of the abiotic stress–responsive genes COR47 and RD29A in lew1 than in the wild type. Furthermore, drought treatment suppressed dark-induced senescence in lew1. These data implicate LEW1 as having crucial roles in dolichol synthesis, the UPR pathway, and the abiotic stress response in Arabidopsis.

	RESULTS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION METHODS REFERENCES

 
lew1 Mutant Plants Exhibit a Leaf-Wilting Phenotype under Normal Growth Conditions Despite Increased Hydraulic Conductivity in Roots
The lew1 mutant was isolated during a genetic screen for leaf-wilting phenotypes (Figure 1A ) in an ethyl methanesulfonate–mutagenized Arabidopsis M2 population (Chen et al., 2005). We found that lew1 plants showed a clear leaf-wilting phenotype even under well-watered normal growth conditions. The wilting in lew1 was observed mainly in the margins of old leaves but not in young leaves. This differs from leaf wilting normally caused by water deficit, in that it was not seen in all leaves and was not associated with changed water status of the seedlings. The overall size of lew1 mature plants was smaller than that of wild-type plants (Figure 1A). There was no apparent difference in flowering time between lew1 and wild-type plants (Figure 2A ).

View larger version (69K): [in this window] [in a new window] [as a PowerPoint slide]   Figure 1. The Phenotypes of the lew1 Mutant. (A) Wild-type and lew1 mutant plants grown in soil. Left, lew1 (arrows point to wilted parts of leaves); right, the wild type. (B) Comparison of root:shoot ratio. Data are means ± SE (n = 41). (C) Lpr of wild-type and lew1 roots. The numbers in parentheses indicate the numbers of plants analyzed. Data are means ± SE. (D) and (E) Effects of aquaporin inhibitors on Lpr of wild-type and lew1 plants: azide at 1 mM, 30 min (D) and mercury at 50 µM, 60 min (E). Data are means ± SE. The numbers in parentheses indicate the numbers of plants analyzed. (F) Structure of xylem in leaves (top row) and stems (bottom row) of wild-type and lew1 plants. Arrows point to xylem. Bars = 5 µm. (G) Grafting experiments. Left, lew1 shoot with wild-type root; right, wild-type shoot with lew1 root.

View larger version (64K): [in this window] [in a new window] [as a PowerPoint slide]   Figure 2. Analysis of Drought Resistance and Related Parameters in lew1. (A) Comparison of wild-type and lew1 plant growth under normal (left) and drought stress (right) conditions. For drought treatment, 3-week-old seedlings were subjected to water withholding for 2 weeks. (B) Stomatal conductance in the wild type and lew1. Data are means ± SE. The numbers in parentheses indicate the numbers of analyzed plants. Two to four leaves were used per plant for the wild type, and one to two leaves were used per plant for lew1. (C) Water loss from detached leaves. Three independent experiments were performed, and 15 leaves per genotype were used in each experiment. Data are means ± SE (Student's t test; ** statistically significant difference [P < 0.01]). (D) Stomata closure rates of the wild type and lew1. For lew1 plants, stomata from wilting parts were measured. About 40 stomata were measured in each of three experiments for both the wild type and lew1. Data are means ± SE (Student's t test; ** statistically significant difference [P < 0.05]). (E) Stomata comparison between the wild type and lew1. The stomata on the wilting part of a leaf are shown. Bars = 20 µm. (F) Comparison of stomata in the wild type and lew1 under conditions that promote full stomatal opening. The stomata from the wilting part of a lew1 leaf were not open under conditions when all stomata in the wild-type leaf were open. Bars = 20 µm. (G) Comparison of stomata in the wild type and lew1. Stomata on nonwilting parts in the leaf base of lew1 are shown. Bars = 20 µm. (H) Osmotic potential in the wild type and lew1 under normal growth conditions. Values are means ± SE, n = 3 (Student's t test; * statistically significant difference [P < 0.05]). (I) Comparison of leaf electrolyte leakage between lew1 and the wild type. Values are means ± SE, n = 3 (Student's t test; * statistically significant difference [P < 0.05]). (J) Typical recording trace showing measurements of turgor in epidermal cells of a wild-type leaf. Following successful impalement of a cell with the pressure probe, the intracellular recording was continued until a stationary turgor was reached. The trace shows successive recordings in three individual cells. (K) Mean (±SE) epidermal cell turgor in areas of leaves of wild-type and lew1 plants showing no obvious sign of wilting (wild type, n = 22 cells; lew1, n = 21 cells). There is a statistically significant difference (P = 0.0005) between the two genotypes.

We previously showed that leaf wilting in lew2 is caused by impeded water transport due to a collapsed xylem (Chen et al., 2005). To investigate a possible xylem defect, we examined cross sections of vascular bundles in the xylem of stems and leaves and found no difference between lew1 and wild-type plants (Figure 1F). We also performed grafting experiments to test whether leaf wilting in lew1 might have been due to an overall defect of water supply to the shoot. When lew1 shoots were grafted onto wild-type roots, leaf wilting still occurred. By contrast, when grafted onto lew1 roots, wild-type shoots did not show any wilting (Figure 1G).

In conclusion, the leaf-wilting phenotype of lew1 was not caused by defects in root water uptake or xylem transport.

The lew1 Mutation Impairs Cell Turgor in Leaves by Causing Increased Electrolyte Leakage and Reduces Leaf Transpiration
In contrast to the increased Lp_r, increased transpiration could cause leaf wilting. Surprisingly, we found that stomatal conductance, as measured using a porometer, was reduced by 75% in lew1 (Figure 2B). Consistent with this, spontaneous water loss from detached leaves was slower in lew1 than in the wild type (Figure 2C). Importantly, lew1 plants grown in soil were more resistant to drought stress than were wild-type plants (Figure 2A).

A closer inspection of leaf epidermis showed that, under normal growth conditions, 60% of stomata were closed in the wilted parts of lew1 leaves but only 15% of stomata were closed in wild-type leaves (Figures 2D and 2E). Even under conditions (e.g., high light) that promote stomatal opening, when all stomata of wild-type leaves were fully open, most stomata in the wilted parts of lew1 leaves remained partly closed (Figure 2F). By contrast, no apparent difference in stomatal opening was found in the unwilted parts of lew1 and wild-type leaves (Figure 2G). These results suggested that, in the wilted parts of lew1 leaves, guard cells and possibly surrounding cells were not able to establish a normal turgor. To assess further the basis of the apparent leaf-wilting phenotype of lew1, leaf cell turgor was investigated using the cell pressure probe technique. Because lew1 leaves could show signs of early senescence, the measurements were performed in young, growing leaves that did not show any of these symptoms. Yet, local zones of wilting, mostly in the most distal part of the blade, could be observed in lew1. Among all leaf cell types investigated, the large fusiform epidermal cells located at the edge of the blade yielded the most stable micropipette impalements and turgor recordings. In wild-type plants, cell turgor measurements could be performed all along the leaf edge, yielding a mean stationary turgor value of P₀ = 4.70 ± 0.21 bars (n = 22 cells) (Figure 2J). By contrast, impalement of the wilted area in lew1 leaves did not allow any stable turgor recording. Measurements performed in other areas of the leaf provided a mean turgor value of P₀ = 3.34 ± 0.29 bars (n = 21 cells), which was significantly lower (P = 0.0005) than the value in wild-type leaves. Thus, leaf cells of lew1 showed a significantly reduced turgor, even in areas showing no clear sign of wilting.

Under normal growth conditions, the osmotic potential of lew1 leaves was slightly higher than that of wild-type leaves (Figure 2H). This was probably due to a slight osmotic compensation in the wilted lew1 cells. We then examined cell membrane integrity by measuring electrolyte leakage. As shown in Figure 2I, lew1 leaves had a higher electrolyte ion leakage than did wild-type leaves, suggesting that the lew1 mutation impaired membrane integrity.

In conclusion, these results suggest that the leaf-wilting phenotype of lew1 is due to increased solute leakage caused by cell membrane lesions. The reduced cell turgor would lead to partial stomatal closure and less transpiration in lew1.

Mapping-Based Cloning of the LEW1 Gene
To clone the LEW1 gene, lew1 plants in the Columbia gl1 accession were crossed with wild-type Landsberg plants, and F2 progeny was grown in soil under normal conditions. lew1 plants with a leaf-wilting phenotype were selected for genetic mapping. LEW1 was initially mapped to the upper arm of chromosome 1. Fine-mapping narrowed the lew1 mutation to a region covered by four BAC clones, T23J18, F25C20, F12F1, and T28K15 (Figure 3A ). Further fine-mapping delimited lew1 on BAC clones F25C20 and F12F1. We sequenced all of the open reading frames from the lew1 mutant in this region and found a single G-to-A mutation in the open reading frame of AT1G11755, which contains seven exons (Figure 3A). The mutation occurred in the second exon and changed GGT (encoding Gly-21) to GAT (encoding Asp-21) in lew1 (Figure 3A).

View larger version (66K): [in this window] [in a new window] [as a PowerPoint slide]   Figure 3. Position Cloning of LEW1 and Sequence Analysis of Long-Chain cis-Prenyltransferases. (A) Positional cloning of LEW1. The LEW1 locus was mapped to between two polymorphic markers on BAC clones T23J18 (recombinants, 11 of 1638) and T28K15 (recombinants, 17 of 1638). Further mapping delimited the LEW1 locus to a region within BACs F12F1 (recombinants, 3 of 1638) and T28K15. All candidate open reading frames were sequenced in this region, and only one mutation (G159A) was found in AT1G11755. (B) (a) Complementation of the lew1 mutant. Shown are the wild type (gl1), the lew1 mutant, the lew1 mutant transformed with 35S-LEW1 (cDNA), the wild type (Columbia [Col]), and a heterozygous plant of lew1 crossed with a line in which the LEW1 gene was disrupted by a T-DNA insertion. The T-DNA was inserted at position 205 counting from the first putative ATG of AT1G11755. (b) RT-PCR analysis of LEW1 transcripts in the wild type, lew1, and a heterozygous plant of lew1 crossed with the T-DNA line. rRNA is shown as a control for loading. (C) Alignment of long-chain cis-prenyltransferases from different organisms: rice, Os02g0197700 and Os03g0197000; human, NP_612468; mouse, BAE39779; Xenopus, NP_001016090; fruit fly, AAF51407; E. coli, P60472; Micrococcus luteus, BAA31993; Saccharomyces cerevisiae, BAA36577 (RER2); Arabidopsis, AT2G23410 and AT1G11755 (LEW1). (D) Phylogenetic tree of LEW1 and related proteins from different organisms: Arabidopsis, At2g23400, At5g58780, At5g58782, At5g58784, At5g58770, At5g60510, At5g60500, and At2g17570; rice, Os01g0857200, Os06g0167400, and Os07g0607700; monkey, XP_001110733; dog, XP_541219; Aedes aegypti, EAT38464; rat, XP_001058252; S. cerevisiae, NP_013819/SRT1 and NP_010088; the others are the same as in (C). The phylogenetic tree was constructed on the alignment using MEGA (version 4.0) (see Supplemental Data Set 1 online). Bootstrap values were calculated from 1000 trials and are shown at each node. The extent of divergence according to the scale (relative units) is shown at bottom.

The LEW1 Gene Encodes a Novel cis-Prenyltransferase Protein
The LEW1 gene encodes a polypeptide of 254 amino acids with a predicted molecular mass of 28.52 kD. Because the genes encoding cis-prenyltransferases form a large family with an overall low level of sequence similarity, we selected for comparison only those candidates that were the most closely related to LEW1 (Figure 3C). The previously identified AT2G23410 and yeast RER2 were included in this analysis. Figure 3D shows that in Arabidopsis, LEW1 (AT1G11755) has only one copy, whereas in rice (Oryza sativa), there are two homologs (Os 03G0197000 and Os 02G0197700). The product of AT2G23410, RER2, and a RER2 homolog in E. coli (P60472) belong to another homology subgroup (Sato et al., 1999; Cunillera et al., 2000; Oh et al., 2000). The analysis indicates that LEW1 is a new member of the cis-prenyltransferase family.

In order to confirm that LEW1 encodes a functional cis-prenyltransferase, we expressed LEW1 in E. coli (Figure 4A ) and analyzed its enzyme activity in vitro. The proteins extracted from E. coli carrying LEW1 or an empty vector were incubated with ¹⁴C-labeled IPP and FPP, and the reaction products were analyzed by normal-phase thin layer chromatography (TLC) (Figure 4B). The authentic solanesols with chain lengths of C₄₅ and C₉₀ were used as standard compounds (Figure 4B). Because E. coli is able to catalyze the synthesis of dolichol with a chain length of C₅₅, the labeling of C₅₅ products was equally detected in both reactions, as expected. However, larger reaction products were detected in the extract from E. coli expressing LEW1. These products migrated slightly faster than the C₉₀ solanesol standards and therefore must be somewhat smaller (Figure 4B). These results indicate that LEW1 possesses cis-prenyltransferase activity for dolichol synthesis in vitro.

View larger version (50K): [in this window] [in a new window] [as a PowerPoint slide]   Figure 4. Biochemical Characterization of LEW1 and Complementation of the Yeast rer2 Mutant. (A) Ectopic expression of LEW1 in E. coli. Proteins extracted from E. coli cells were analyzed on 12% SDS-PAGE gels and stained with Coomassie blue. The arrow points to the expressed LEW1 protein in E. coli. M, marker; Con, empty vector control; LEW1, isopropylthio-β-galactoside–induced LEW1. (B) TLC analysis of LEW1 reaction products. [14C]IPP and [14C]FPP were incubated with total proteins isolated from E. coli strains transformed with either the empty vector (Con) or LEW1. At the completion of the reaction, lipid products were extracted and subjected to TLC analysis. Solanesols C45 and C90 were used as standards. A C55 band (representing a product that can be synthesized by the native cis-prenyltransferase in E. coli) was equally detected in E. coli transformed with either empty vector or LEW1, indicating that the reactions were successful. Ori, original spot. (C) LEW1 partially complemented the temperature-sensitive phenotype of the yeast rer2 mutant. rer2 cells transformed with an empty vector, LEW1, or DPS (AT2G23410) were grown overnight. Serial decimal dilutions were spotted onto plates of synthetic dropout medium with all amino acids (URA+) or excluding URA (URA–). URA is a selection marker of the transformed clone. Plates were incubated at the indicated temperatures and photographed after 4 d.

Dolichol Contents Are Reduced in lew1 Mutant Plants but Are Increased in Transgenic Plants Overexpressing LEW1
Dolichols with chain lengths of C₇₅ and C₈₀ have been reported to be the most abundant forms in Arabidopsis (Gutkowska et al., 2004). We performed lipid analysis to compare the dolichol contents among wild-type, lew1 mutant, and transgenic plants overexpressing LEW1 (lines OE14 and OE34). Figure 5F shows that LEW1 transcripts increased in the two overexpression lines. The dolichols in Arabidopsis exist as free lipids, lipids covalently bound to proteins, or lipid carriers linked to oligosaccharides for protein glycosylation. The total dolichols were extracted and analyzed by liquid chromatography–mass spectrometry (LC-MS) following previously established methods (Skorupinska-Tudek et al., 2003; Gutkowska et al., 2004). Dol-C_65-105 (a mixture of dolichols C₆₅ to C₁₀₅) was purchased and used as a standard compound. Because we did not have pure standard samples for each dolichol to quantify the exact amounts, here we only show the relative dolichol contents in each sample. As indicated in Figure 5A, we detected Dol-C₇₅ and Dol-C₈₀ at different retention times (Dol-C₇₅, 25 min, mass analysis of m/z 1063 to 1065; Dol-C₈₀, 26 min, m/z 1131 to 1133). Based on this information, we compared the peak areas of Dol-C₇₅ and Dol-C₈₀ of wild-type (Figure 5B), lew1 (Figure 5C), OE14 (Figure 5D), and OE34 (Figure 5E) plants. Compared with wild-type levels, the relative contents of Dol-C₇₅ and Dol-C₈₀ were 13 and 15% in lew1, 109 and 103% in line OE14, and 111 and 109% in line OE34, respectively. These results strongly suggest that LEW1 catalyzes the biosynthesis of Dol-C₇₅ and Dol-C₈₀ in Arabidopsis.

View larger version (44K): [in this window] [in a new window] [as a PowerPoint slide]   Figure 5. Analysis of the Relative Dolichol Contents in Wild-Type, lew1, and LEW1 Overexpression Plants by LC-MS. The peak area in the wild type was taken as 100%, and the relative contents in other plants were compared with that in the wild type. (A) The dolichol standards (left, Dol-C75; right, Dol-C80). Arrows indicate the peaks. (B) The wild type. (C) lew1 mutant. (D) Overexpression line 14 (OE14). (E) Overexpression line 34 (OE34). (F) RNA gel blot analysis of LEW1 expression in OE14 and OE34. Due to its low expression, LEW1 transcript could not be detected in the wild type and the lew1 mutant by RNA gel blot under the conditions used. OE15 and OE21, two other transgenic lines not used in this study, are also included. The bottom panel shows an RNA gel as a loading control.

View larger version (112K): [in this window] [in a new window] [as a PowerPoint slide]   Figure 6. lew1 Impairs Protein N-Glycosylation, and lew1 Plants Are Hypersensitive to Tunicamycin. (A) Comparison of protein glycosylation patterns between the wild type and lew1 by PAS staining. Thirty micrograms of total protein were resolved by SDS-PAGE and detected by PAS staining. The arrow points to the bands missing from lew1. (B) Concanavalin A–Sepharose binding assay for N-glycosylated proteins. Total proteins extracted from the wild type and lew1 were incubated with concanavalin A–Sepharose, and the bound proteins were eluted, treated with (+) or without (–) endoglycosidase H (Endo H), resolved by SDS-PAGE, and stained with Coomassie blue. The arrows at left point to the two protein bands found in the wild type but not in lew1. These two bands were recovered and analyzed by MALDI-TOF MS. The top one corresponds to SKU5, and the bottom one corresponds to TTG1. Endoglycosidase H treatment resolved the proteins found in both the wild type and lew1. M, marker. (C) lew1 seedlings are sensitive to tunicamycin (TM). Seedlings were germinated on MS medium or MS medium containing 0.1 µg/mL tunicamycin and grown for 10 d in a growth chamber. WT, wild type (gl1); lew1, lew1 mutant; 35S-LEW1, a wild-type line overexpressing LEW1 under the control of the 35S promoter (line 14); lew+35S-LEW1, a lew1 mutant line overexpressing LEW1 under the control of the 35S promoter (line 5).

Third, we recovered from SDS-PAGE two bands that were present in the total protein extracts from the wild type but not from lew1 (Figure 6B, arrows). Tryptic peptides were resolved by matrix-assisted laser-desorption ionization time of flight (MALDI-TOF) MS (Autoflex II TOF/TOF; Bruker) and analyzed by Mascot at www.Matrixscience.com. The smaller band corresponded to the glycoprotein β-thioglucoside glucohydrolase encoded by TGG1 in Arabidopsis (Xue et al., 1995; Husebye et al., 2002). TGG1 is localized in the vacuole, and underglycosylated TGG1 was also found in the stt3a mutant of Arabidopsis in a previous study (Koiwa et al., 2003). The partial sequence of the larger product was identical to that of a glycoprotein named SKU5 (Sedbrook et al., 2002). SKU5 is localized to both the plasma membrane and the cell wall and is involved in directional root growth in Arabidopsis (Sedbrook et al., 2002). These results further demonstrate that LEW1 plays critical roles in protein glycosylation.

lew1 Mutants Are Hypersensitive to the Glycosylation Inhibitor Tunicamycin
Tunicamycin is a specific inhibitor of UDP-N-acetylglucosamine-dolichyl-phosphate N-acetylglucosaminephosphotransferase (EC 2.7.8.15), the first enzyme in the biosynthesis of dolichol-linked oligosaccharides for protein glycosylation modification (Lehrman, 1991; Koizumi et al., 1999). Blocking glycosylation by tunicamycin induces UPR, causing enhanced expression of some UPR pathway genes (Koizumi et al., 1999; Lukowitz et al., 2001; Martinez and Chrispeels, 2003). Because the lew1 mutation impairs protein glycosylation, we tested the sensitivity of lew1 to tunicamycin. As shown in Figure 6C, lew1 plants were more sensitive than wild-type plants to exposure to 0.1 µg/mL tunicamycin. Overexpression of LEW1 in the wild type did not obviously increase the resistance of transgenic plants to tunicamycin, suggesting that LEW1 is not a limiting factor for protein glycosylation.

The lew1 Mutant Is Hypersensitive to Dark Stress
Defects in protein glycosylation greatly affect the biogenesis of membrane and secreted proteins by interfering with their folding, trafficking, degradation, and targeting to their final sites in the cell; therefore, defects in protein glycosylation generally lead to ER stress. We observed that mutant plants with one copy of the T-DNA insertion allele of lew1 and the other copy as the lew1 point mutation allele showed an early-senescence phenotype (Figure 3B). We further tested whether lew1 exhibits a more severe membrane-damage phenotype in the dark, because prolonged dark treatment induces membrane damage and senescence (Fujiki et al., 2000). lew1 and wild-type plants grown under well-watered conditions for 2 weeks were transferred to darkness at 22°C for 10 d. All of the lew1 plants became yellow, but wild-type plants did not. lew1 plants transformed with the LEW1 overexpression construct (35S-LEW1) exhibited similar phenotypes as wild-type plants (Figure 7A ). Electrolyte leakage from leaves of lew1 plants was greatly increased and reached 35% after 10 d in the dark, whereas wild-type leaves showed only 10% leakage under the same conditions (Figure 7B). These results show that more severe membrane damage occurred in lew1 than in the wild type under prolonged dark treatment.

View larger version (79K): [in this window] [in a new window] [as a PowerPoint slide]   Figure 7. lew1 Plants Are Hypersensitive to Dark. (A) Wild-type and lew1 seedlings (3 weeks old) were grown in light (top row) or dark (bottom row) at 22°C for 10 d. Also shown is a lew1 transgenic line overexpressing LEW1 (line 5) and a wild-type transgenic line overexpressing LEW1 (line 14). (B) Electrolyte leakage assay of wild-type and lew1 plants in the dark. Two-week-old seedlings were grown in soil in the dark at 22°C for the indicated times, and the leaves were taken for ion leakage assays. Results were from three independent experiments. Data are means ± SE (* P < 0.05, ** P < 0.01). (C) Tunicamycin treatment accelerated dark-induced senescence in lew1. Three-week-old seedlings grown in soil were sprayed with water (Dark) or with 0.1 µg/mL tunicamycin (Dark+TM) and then incubated in the dark for 10 d before photographs were taken. (D) Expression of dark-inducible genes in lew1 and the wild type. Two-week-old seedlings were grown in the dark for 24 or 48 h, and total RNA was analyzed by RNA gel blot. Also shown are two wild-type transgenic lines (OE14 and OE34) overexpressing LEW1 (35S-LEW1). The two bottom panels show rRNA and tubulin as loading controls.

Darkness Stress–Responsive Genes Are Hyperinduced in lew1 by Dark Treatment
We selected two late dark-inducible marker genes, DIN2 (for dark-inducible2) and DIN9, to compare their expression patterns in lew1 and wild-type plants. DIN2 encodes a β-glucosidase, and DIN9 encodes a mannose-6-phosphate isomerase (Fujiki et al., 2000). We also examined a leaf senescence marker gene, YLS4 (for YELLOW-LEAF-SPECIFIC GENE4), which is induced in old leaves by darkness (Fujiki et al., 2005). Two-week-old seedlings were kept under darkness for 24 and 48 h. RNA gel blot analysis revealed DIN2, DIN9, and YLS4 transcripts to be substantially induced in lew1 at 24 and 48 h in the dark, whereas the transcripts were not yet or less detected in the wild type and two transgenic lines overexpressing LEW1 at these time points in the dark (Figure 7D).

Drought or Osmotic Stress Suppresses the Darkness Sensitivity of lew1 Plants, and lew1 Seedlings Are More Tolerant to Osmotic Stress
Because lew1 plants are more tolerant to drought stress, we next analyzed the response of lew1 to drought stress in the dark. Ten-day-old seedlings grown in the light were pretreated by withholding water for 7 d prior to the dark treatment, which produced only a modest drought stress. Control plants were watered every third day, and both control and drought-treated plants were kept in the dark or light for various times. After an additional 15 d of drought treatment, wild-type plants grown in the light were seriously wilted (Figure 8A , 15 d in light) and did not recover growth after rewatering. By contrast, lew1 plants, which developed clear drought-tolerant phenotypes in light, were able to regrow after watering (see Supplemental Figure 1 online). These results are consistent with previous data (Figure 2A) and indicate that lew1 is tolerant to drought stress at different developmental stages.

View larger version (113K): [in this window] [in a new window] [as a PowerPoint slide]   Figure 8. Drought or Osmotic Stress Improves the Survival of lew1 Plants in the Dark or under Weak Light (A) Ten-day-old seedlings in soil were subjected to water withholding for 7 d to give a modest drought treatment (Drought) in the light and then incubated in the dark for 15 or 20 d, or in the light for 15 d, without watering. Wild-type but not lew1 plants died from drought treatment for 15 d in the light. All lew1 plants died from drought at 20 d in the light. Plants without drought treatment (Watered) were kept in the light for 15 d or in the dark for 15 or 20 d. (B) After 20 d, the drought-treated plants in the dark were watered and moved, together with plants without drought treatment in the dark, to light for an additional 3 d. No lew1 plants without drought treatment in the dark survived, but all lew1 plants with drought treatment in the dark survived. No wild-type plants died under the conditions used. (C) Osmotic stress imposed by mannitol, NaCl, or glycerol increased the survival of lew1 seedlings in MS medium without sucrose under weak light. Four-day-old seedlings grown on MS medium with 3% sucrose were moved to MS medium without sucrose (–), with sucrose (3%), or without sucrose but with 100 mM mannitol, 180 mM NaCl, or 1.5% glycerol and cultured for 7 d. All lew1 seedlings on MS medium without sucrose died, while all wild-type seedlings survived. In other treatment conditions, all lew1 and wild-type seedlings survived. (D) lew1 seedlings are more tolerant to high-salt stress than wild-type seedlings. Four-day-old seedlings were moved to MS medium containing 3% sucrose with different concentrations of NaCl (0, 75, 120, or 180 mM) and cultured for 10 d. There was no clear growth difference at 0, 75, and 120 mM NaCl between the wild type and lew1. However, at 180 mM NaCl, lew1 seedlings grew better than wild-type seedlings.

We further tested the growth of lew1 mutant plants on medium supplemented with various osmotic stress agents. As controls, both lew1 and wild-type seedlings grew well on Murashige and Skoog (MS) medium containing 3% sucrose (Figure 8C). When no sugar was added to the MS medium, lew1 seedlings grew poorly and died after 10 d in culture. The growth of wild-type seedlings was also inhibited in the absence of sugar in the medium, but in contrast with lew1, all wild-type seedlings survived and still grew (Figure 8C). When grown on MS medium (without sugar) supplemented with 100 mM mannitol, 180 mM NaCl, or 1.5% glycerol for 10 d, all lew1 seedlings survived, as did the wild type (Figure 8C). These results indicate that hyperosmotic stress is able to suppress the sensitivity of lew1 seedlings to conditions of no sugar in culture medium.

We also compared the growth of lew1 and the wild type on MS medium with 3% sucrose and different concentrations of NaCl. As shown in Figure 8D, lew1 seedlings showed similar sensitivity to NaCl concentrations of <120 mM as the wild type. However, lew1 seedlings grew better than wild-type seedlings on 180 mM NaCl. This result suggests that unlike stt3a and other N-glycan defect mutants, which are sensitive to salt stress (Koiwa et al., 2003; Kang et al., 2008), lew1 mutants were more resistant to high NaCl than the wild type.

Drought Stress Induces Higher Expression of BiP and bZIP60 in lew1 Than in the Wild Type
Cells have evolved elaborate mechanisms to ensure the accuracy of protein folding and assembly. When unfolded proteins are accumulated to a certain level, UPR signaling is activated to induce the expression of specific proteins to protect cells from the ER stress or to induce cell death if homeostasis cannot be reestablished (Lin et al., 2007). A previous study found that overexpression of BiP improved the drought tolerance of transgenic plants, suggesting that UPR signaling might be involved in drought stress tolerance (Alvim et al., 2001). In order to determine whether drought stress could provoke UPR signaling or not, we compared the transcripts of genes in the UPR pathway in the wild type and lew1 under drought stress conditions. We used a BiP probe that detects the transcripts of both BiP1 and BiP2 because of their high sequence identity. BiP transcripts were hardly detected in the wild type after a 5-h drought treatment under our conditions, whereas the BiP transcripts were clearly detected at 2 h and highly expressed at 5 h in lew1 (Figure 9A ). BiP transcripts were also induced by a mannitol treatment at higher levels in lew1 than in the wild type (Figure 9B). bZIP60 is a key bZIP transcription factor responsible for the ER stress response, and its expression is induced by tunicamycin (Iwata and Koizumi, 2005). The expression of bZIP60 was induced to a higher level in lew1 than in the wild type under drought treatment (Figure 9A). The expression of PROTEIN DISULFIDE ISOMERASE (PDI), encoding a disulfide isomerase, is also induced by tunicamycin in Arabidopsis (Martinez and Chrispeels, 2003). However, the level of PDI did not change under our drought stress conditions (Figure 9A). These results indicate that osmotic stress produced from either drought or mannitol treatment can activate the expression of some genes in the UPR pathway, and this response is enhanced by the lew1 mutation.

View larger version (83K): [in this window] [in a new window] [as a PowerPoint slide]   Figure 9. RNA Gel Blot Analysis of Stress-Inducible Genes under Different Stress Conditions. (A) The transcripts of BiP and bZIP60 were induced to higher levels in lew1 than in the wild type by drought stress. Four-week-old seedlings were removed from soil and placed on a laboratory bench for 1 h (drought 1 [Dro1]). Then, seedlings were covered with a transparent film (to slow water loss) for 1 or 4 h (drought 2 h and 5 h [Dro2 and Dro5, respectively]). Total RNA from each treatment was used for RNA gel blot analysis. RNA from shoots without treatment was used as a control at 0 h (Con). rRNAs stained with ethidium bromide were used as loading controls. (B) More BiP transcripts accumulated in lew1 than in the wild type under mannitol treatment. Four-week-old seedlings from soil were dipped into solution containing 300 mM mannitol for 1, 2, or 6 h (Man1, Man2, and Man6, respectively), and total RNA was analyzed by RNA gel blot. rRNAs stained with ethidium bromide served as loading controls. RNA from shoots without treatment was used as a control at 0 h (0). (C) The transcripts of RD29A and COR47 were induced to higher levels in lew1 than in the wild type by drought stress. Drought treatment was done as in (A). rRNA was used as a loading control. (D) Four-week-old seedlings were treated with 150 mM NaCl for 0.5 h (NaCl0.5) or 1 h (NaCl1) or with 300 mM mannitol for 1 h (Man1) or 2 h (Man2). rRNAs stained with ethidium bromide served as loading controls. RNA from shoots without treatment was used as a control at 0 h (Con). (E) Two-week-old seedlings grown on MS medium were treated with 0.1 µg/mL tunicamycin for different times, and total RNA was analyzed by RNA gel blot. rRNAs stained with ethidium bromide served as loading controls.

Stress-Responsive Genes Are Induced to Higher Levels and at Earlier Time Points by Drought Stress in lew1 Than in the Wild Type
We further investigated whether the lew1 mutation also alters the expression of drought-responsive genes. In these experiments, 4-week-old seedlings were removed from soil and subjected to different water stress treatments, and the expression of the drought-responsive genes RD29A and COR47 was analyzed by RNA gel blots. Under drought conditions, the transcripts of RD29A and COR47 were induced to higher levels in lew1 than in wild-type plants (Figure 9C, Dro1 and Dro2). In addition, NaCl and mannitol treatments resulted in earlier induction of RD29A and COR47 in lew1 than in the wild type (Figure 9D).

Because tunicamycin treatment causes ER stress, we also tested the expression of RD29A and COR47 under tunicamycin treatment. We detected higher expression levels of RD29A and COR47 in lew1 than in the wild type at 1 h. The expression levels in wild-type and mutant plants became similar at 2 or 6 h (Figure 9E). The expression of BiP was induced at earlier time points and to higher levels by tunicamycin treatment in lew1 than in the wild type (Figure 9E). These results suggest that lew1 plants are more sensitive to tunicamycin treatment than are wild-type plants in the induction of UPR genes.

	DISCUSSION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION METHODS REFERENCES

 
Here, we report on the genetic and functional analyses of LEW1, a gene encoding a long-chain cis-prenyltransferase in dolichol biosynthesis in Arabidopsis. There are 10 cis-prenyltransferases related to the yeast RER2, and LEW1 belongs to a subgroup that is different from the other 9 cis-prenyltransferases in Arabidopsis. LEW1 is an essential gene in Arabidopsis, since plants with a homozygous T-DNA insertion in LEW1 were nonviable. In yeast, RER2 and STR1 encode two similar cis-prenyltransferases catalyzing the biosynthesis of dolichols with different chain lengths: RER2 for C₇₀ to C₈₅ and STR1 for C₉₅ to C₁₀₅ (Sato et al., 1999, 2001). The rer2 mutant showed various phenotypes that were not observed in the str1 mutant. Although STR1 overexpression could suppress the growth and glycosylation defects of rer2, the biological functions of RER2 and STR1 are considered to be different (Sato et al., 1999, 2001). These results suggest that different cis-prenyltransferases in Arabidopsis might play distinct roles in plant growth and development.

Our in vitro assay indicated that LEW1 is able to catalyze the biosynthesis of dolichol with a chain length of less than C₉₀. Comparison of total dolichol contents between lew1 and the wild type showed that the dolichols with chain lengths of C₇₅ to C₈₀ were reduced to about one-quarter of the wild-type level, suggesting that LEW1 plays a predominant role in catalyzing the biosynthesis of dolichol C₇₅ to C₈₀ in Arabidopsis. In mammals, the chain lengths of dominating dolichyl phosphates used as glycosyl carrier lipids are around C₉₅ to C₁₀₅ (Schenk et al., 2001). Our results indicate that the chain lengths of dolichols in Arabidopsis are similar to those in yeast. Although LEW1 catalyzed the biosynthesis of dolichols with chain lengths similar to those produced by RER2, overexpression of LEW1 only partially complemented the temperature-sensitive phenotype of rer2. By contrast, AT2G23410, which catalyzes the formation of dolichol with a chain length of C₁₂₀, was able to complement the rer2 mutant phenotype (Cunillera et al., 2000; Oh et al., 2000). Similarly, a cDNA that encodes a cis-prenyltransferase for catalyzing the production of dolichols with a chain length of C₈₅ to C₉₅ from human brain complemented the growth and other phenotypes of the rer2 mutant (Shridas et al., 2003). Studies on the rer2 mutant overexpressing SRT1 indicated that yeast cells are able to utilize the longer dolichols in protein glycosylation when shorter ones (C₇₀ to C₈₅) are lacking (Sato et al., 2001). Sequence analysis suggested that LEW1 was more diverged from RER2 than AT2G23410. We speculate that LEW1 expressed in yeast might have low catalytic efficiency for dolichol synthesis, possibly because the LEW1 protein may not be folded correctly in yeast or LEW1 may need some specific cofactors or modifications for full activity.

One of the important functions of dolichols is to serve as the lipid carrier of carbohydrates in the assembly of Dol-P–linked oligosaccharides for protein N-glycosylation (Burda and Aebi, 1999). In plants and animals, some genes involved in the dolichol-related N-linked glycosylation pathway have been identified, and defects in the human genes lead to serious diseases called congenital disorders of glycosylation (Koiwa et al., 2003; Kranz et al., 2004, 2007; Leroy, 2006). In this study, we found that the lew1 mutation reduced protein glycosylation levels and identified two specific proteins with altered glycosylation. There are 10 RER2-related cis-prenyltransferase homologs in Arabidopsis, and at least one of them, AT2G23410, was shown to also be able to catalyze the biosynthesis of dolichols in an in vitro assay (Cunillera et al., 2000; Oh et al., 2000). It is not known yet whether all of these genes play some roles in N-linked protein glycosylation as LEW1 does.

More than 70% of all proteins in eukaryotic cells are predicted to be N-glycosylated (Apweiler et al., 1999). Genes encoding some of the enzymes functioning at various steps of the protein glycosylation pathway have been characterized. These include, for example, GNTI (for N-ACETYL GLUCOSAMINYLTRANSFERASE I, encoded by COMPLEX GLYCAN1 [CGL1]) (von Schaewen et al., 1993; Wenderoth and von Schaewen, 2000; Strasser et al., 2005), GMII (an -mannosidase II, encoded by HYBRID GLYCOSYLATION1 [HGL1]) (Strasser et al., 2006), STT3a (Koiwa et al., 2003), DGL1 (for DEFECTIVE GLYCOSYLATION1), an Arabidopsis homolog of an oligosaccharyltransferase complex subunit (Lerouxel et al., 2005), CYT1 (for CYTOKINESIS-DEFECTIVE1, mannose-1-phosphate guanylyltransferase) (Lukowitz et al., 2001), GCS1/KNF (glucosidase I), and RSW3 (glucosidase II; glucosidases I and II trim the terminal glucose of N-glycans). Mutations in these genes have some overlapping, yet differential, effects on plant growth, development, and stress responses. For instance, a lesion in GNTI caused by the cgl1 mutation impairs the synthesis of Golgi-modified complex N-linked glycans, and cgl1 mutant plants do not show phenotypes under normal conditions but are more sensitive to heat stress or dark treatment (von Schaewen et al., 1993). Similarly, hgl1 mutants do not show any obvious phenotypes under normal conditions (Strasser et al., 2006), whereas dgl1 and cyt1 mutants do (Zablackis et al., 1996; Bonin et al., 1997; Lukowitz et al., 2001; Lerouxel et al., 2005). Interestingly, some mutants, such as dgl1, gcs1/knf, rsw3, and cyt1, show a clear cell wall–defective phenotype, suggesting that N-glycan processing plays crucial roles in plant cell well formation. Mutations in STT3a, a subunit of the oligosaccharyltransferase complex responsible for protein N-glycosylation, may also affect the cell wall, particularly under salt stress (Koiwa et al., 2003). stt3a mutants are hypersensitive to salt and osmotic stress due to cell swelling and cell cycle arrest at the root tip (Koiwa et al., 2003). The stt3a mutations were also reported to cause increased expression of BiP in the root tip (Koiwa et al., 2003). lew1 mutant plants were not hypersensitive to salt or osmotic stress. Rather, the mutant showed increased drought resistance and enhanced survival under high-salt conditions. These results suggest that defects in different steps of the protein N-glycosylation pathway have different consequences on plant stress responses.

Although we compared the amounts of cellulose and lignin from lew1 and wild-type cell walls and did not find a clear difference (see Supplemental Figure 2 online), we cannot exclude the possibility of a subtle change in cellulose or other cell wall components in lew1 that could not be detected in our experiments. Recent studies suggest that the plant cell wall could be important for stress signaling (Ellis et al., 2002; Zhong et al., 2002; Chen et al., 2005; Hernandez-Blanco et al., 2007; Li et al., 2007). Mutations in genes for cell wall biosynthesis or modifications can activate jasmonate, ethylene, and/or abscisic acid signaling, all of which could be connected to sugar responses (Ellis et al., 2002; Zhong et al., 2002; Chen et al., 2005; Hernandez-Blanco et al., 2007; Li et al., 2007). Our results on lew1 suggest possible complex crosstalk among the different signaling pathways, which may regulate the various stress responses observed in this study.

Dolichols are involved in many different cellular processes. In addition to their roles in protein glycosylation and other modifications, free dolichols also have important biological functions as lipids (Chojnacki and Dallner, 1988; Bizzarri et al., 2003). Because it is an unsaturated lipid, dolichol was postulated to function as a free radical scavenger against reactive oxygen species either produced by chemicals or UV light or accumulated upon aging. However, there is no direct evidence for such a function in vivo (Bergamini et al., 1998; Bizzarri et al., 2003; Sgarbossa et al., 2003). Although we did not observe any phenotypes, such as sensitivity to H₂O₂ treatment (see Supplemental Figure 3 online), we still cannot exclude that the pleiotropic phenotypes observed in lew1 might be a comprehensive outcome of changing lipid peroxidation, plasma membrane characteristics, protein dolichylation, and/or protein glycosylation.

A striking phenotype, which served in the initial isolation of lew1, is leaf wilting even under normal growth conditions. We found that the leaf wilting was not due to a defect in water uptake by roots or in water transport from root to shoot through the xylem. The phenotype was also not caused by excessive transpiration or a low leaf cell osmotic potential. On the contrary, lew1 plants displayed enhanced water permeability in roots and had a reduced stomatal conductance. Our results suggest that the main cause of the leaf-wilting phenotype is probably cell membrane lesions that result in electrolyte leakage and thus impair the ability to maintain cell turgor. The turgor defect may explain, at least partly, the reduced stomatal conductance and enhanced drought resistance of the lew1 mutant. The cell membrane lesions reflect a critical role of dolichol in membrane function and/or its function in protein glycosylation. Because virtually all membrane proteins and secreted proteins are folded and assembled in the ER before being exported and transported to final destinations in the cell, defects in protein glycosylation would greatly affect protein folding, induce ER stress, and eventually reduce the levels and/or activities of certain membrane proteins. Membrane trafficking is known to play crucial roles in abiotic stress responses (Carter et al., 2004). One of the two identified proteins that are affected by lew1, SKU5, is localized in the plasma membrane (Sedbrook et al., 2002). Other plasma membrane proteins, such as H⁺-ATPases and aquaporins, which are critical in turgor maintenance, may also be altered by the lew1 mutation.

The increased dark sensitivity of lew1 and the early-senescence phenotype exhibited by heterozygous plants carrying lew1 and a T-DNA insertion also support membrane defects in the mutant. The dark-inducible genes DIN2 and DIN9 and the leaf senescence marker YLS4 were induced to higher levels in lew1 than in the wild type by the dark treatment. Leaf senescence is considered to be a type of programmed cell death, with plasma membrane breakage as an earlier symptom (Lim et al., 2007). In yeast, defects in protein N-glycosylation induce apoptosis (Hauptmann et al., 2006). Apoptosis, or programmed cell death, is strongly induced by reduced activities of ER and the Golgi apparatus in plant cells (Crosti et al., 2001; Malerba et al., 2004). The fact that tunicamycin-treated wild-type plants mimicked the leaf-senescence phenotype of lew1 in the dark suggests that protein N-glycosylation defects in lew1 contribute to the increased dark-sensitivity and early-senescence phenotypes of the mutant. Nevertheless, dark treatment did not upregulate UPR pathway genes. Interestingly, drought but not tunicamycin treatment greatly improved the survival of lew1 plants during prolonged dark treatment. This suggests that drought treatment suppresses the dark sensitivity of the mutant by a mechanism independent of the UPR pathway.

The transcriptional upregulation of a large number of genes involved in the UPR pathway is the most prominent strategy for the cell to cope with ER stress (Urade, 2007). BiP is a chaperone protein that is induced by chemicals such as the specific N-glycosylation inhibitor tunicamycin (Koizumi et al., 1999). Previous studies have indicated that UPR pathway genes are important for both biotic and abiotic stress responses (Alvim et al., 2001; Wang et al., 2005; Fujita et al., 2007; Liu et al., 2007b). In this study, we observed that drought stress induced the expression of the UPR pathway genes BiP and bZIP60, and the induction was much stronger in lew1 than in the wild type. On the other hand, tunicamycin treatment was able to induce the expression of the drought-responsive genes RD29A and COR47, with stronger induction in the lew1 mutant. These stress marker genes were also more responsive to drought in the mutant. Taken together, these results suggest that drought stress may cause ER stress and activate UPR responses, and part of the drought responses (e.g., induction of RD29A and COR47) may be mediated by the UPR pathway. The enhanced inducibility of UPR genes and drought-responsive genes in lew1 may contribute to the drought resistance and salt/osmotic tolerance phenotypes of the mutant.

	METHODS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION METHODS REFERENCES

 
Plant Growth Conditions
Arabidopsis thaliana ecotype Columbia (gl1) was grown in long-day conditions (16-h-light/8-h-dark cycle) with forest soil and vermiculite (1:1) or on MS medium (M5519; Sigma-Aldrich) containing 3% (w/v) sucrose and 0.8% (w/v) agar at 22°C. All seeds were kept at 4°C for 3 d before sowing. For drought treatment in soil, we grew Arabidopsis seedlings in short-day conditions (12-h-light/12-h-dark