The epidermis-specific extracellular BODYGUARD controls cuticle development and morphogenesis in Arabidopsis

Abstract

The outermost epidermal cell wall is specialized to withstand pathogens and natural stresses, and lipid-based cuticular polymers are the major barrier against incursions. The Arabidopsis thaliana mutant bodyguard (bdg), which exhibits defects characteristic of the loss of cuticle structure not attributable to a lack of typical cutin monomers, unexpectedly accumulates significantly more cell wall-bound lipids and epicuticular waxes than wild-type plants. Pleiotropic effects of the bdg mutation on growth, viability, and cell differentiation are also observed. BDG encodes a member of the alpha/beta-hydrolase fold protein superfamily and is expressed exclusively in epidermal cells. Using Strep-tag epitope-tagged BDG for mutant complementation and immunolocalization, we show that BDG is a polarly localized protein that accumulates in the outermost cell wall in the epidermis. With regard to the appearance and structure of the cuticle, the phenotype conferred by bdg is reminiscent of that of transgenic Arabidopsis plants that express an extracellular fungal cutinase, suggesting that bdg may be incapable of completing the polymerization of carboxylic esters in the cuticular layer of the cell wall or the cuticle proper. We propose that BDG codes for an extracellular synthase responsible for the formation of cuticle. The alternative hypothesis proposes that BDG controls the proliferation/differentiation status of the epidermis via an unknown mechanism.

Figure 1.

Phenotype of the bdg Mutant of Arabidopsis. The panels show photographs ([A] to [E]) and scanning electron micrographs ([F] to [I]). Bars = 10 mm for (A), 5 mm for (B) and (C), 2 mm for (D), 200 μm for (F), 100 μm for (G), and 10 μm for (H) and (I). (A) When grown naturally in short daylengths (8 h of light), 6-week-old mutants (right) have smaller, often deformed leaves and may stunt their growth. Compare with the wild-type plant at left. (B) and (C) Wild-type (B) and bdg (C) plants can be easily distinguished at the age of 3 weeks. Note the small size of bdg leaf blades and the reduced numbers of trichomes on them. (D) Typical necrotic lesions affecting both the epidermis and the mesophyll at the tips of leaves in bdg (arrows). (E) Deformation of a bdg plant caused by graft-like fusions of rosette leaves. Note the rupture of two petioles (lines 1 and 2 show upper and lower rips) caused by extremely strong fusions between growing organs. (F) General view of the leaf surface showing defects in the epidermis in bdg. Some trichomes are fully developed, but many are misshapen and arrested at earlier stages. In many cases, trichomes become flattened, bent, and eventually adhere to the leaf surface (arrow). (G) Collapse of mature trichomes. (H) A suture zone (arrow) showing no signs of cuticle proper. (I) Germination of wild-type pollen induced by contact with bdg leaf epidermis.

Figure 2.

Defects in the Epidermal Cell Surface of bdg Leaves Revealed by Leaching of Chlorophyll and Staining with a Cationic Dye. (A) Intact rosette leaves of bdg release chlorophyll faster than do wild-type leaves when immersed in 80% ethanol, probably because both the solvent and solute more easily penetrate the leaf surface. Each value is an average of six replicates. Bars indicate se. Differences between Columbia (Col) and bdg were significant at P < 0.01. (B) Representative rosette leaves of bdg (top row) and Col (bottom row) are shown from adaxial and abaxial sides after 2 min of staining with an aqueous solution (0.05%) of toluidine blue, which is able to bind selectively to free anionic groups such as carboxylate, phosphate, and sulfate radicals (Tanaka et al., 2004). When stained, leaves of bdg show a patchy coloration pattern, in contrast with wild-type leaves, which remain virtually unstained because the intact cuticle excludes the dye. Notice that the dye penetrates more rapidly through the cuticle on the abaxial side. The two fused leaves of bdg shown at left (arrows) are reminiscent of staining to unfused leaves. This figure also illustrates changes in leaf morphology. In bdg, the leaves are elongated and narrow with a smooth margin, compared with the broad serrated leaves of the wild type.

Figure 3.

Ultrastructure of the Cuticle of the Leaf Epidermis. The images were obtained with a transmission electron microscope. Bars = 500 nm. (A) Regular electron-dense cuticle proper (membrane) formed by wild-type plants. Compare with the deformed cuticles in bdg ([B] to [F]). (B) The mutant cell wall exhibiting a paste-like appearance with irregular, misshapen boundaries. The cuticular membrane is much thinner than in the wild type and is puckered, probably as a result of pressure from the underlying material associated with the cell wall. (C) In places (arrows), the electron-opaque cuticle proper shows large-scale disruption and may be practically absent altogether. (D) Pockets or caverns in the cuticular zone (see also [C]). (E) A series of irregular layers formed by bulging of the less opaque cell wall material. The cuticular membrane is hardly detectable along the surface of the protuberance (arrows). (F) A fusion zone (arrow) devoid of intervening cuticular membrane between two fused leaves.

Figure 4.

In Vitro Culture of the bdg Mutant. The panels show photographs ([B] to [D] and [J] to [L]) and scanning electron micrographs ([E] to [H]). Bars = 3 mm for (B) to (D), 100 μm for (E) and (F), 300 μm for (G) and (H), 10 mm for (J), and 250 μm for (K) and (L). (A) Graph showing the mean dry weight of root and leaf tissue in bdg compared with the wild type. The plants from germinated seeds were grown for 7 d on vertical agar plates. Dry weight refers to the weight of tissue after it has been dried at 60°C. Results are presented as means ± se for 16 bdg and 32 wild-type plants. (B) Twenty-four-day-old bdg and wild-type plants. Rosette leaves in bdg are severely misshapen (see also [D]) and have short, thick petioles, deformed laminae, and, occasionally, leaf fusions. (C) At the age of 4 weeks, wild-type plants produced inflorescences, whereas mutant plants became vitrified and developed a calli-like appearance. (D) A typical bdg leaf at the age of 24 d that did not exhibit any fusions with other leaves. It can be distinguished by misshapen lamina compared with the wild-type leaf. (E) Branched trichomes on the abaxial surface of a mutant rosette leaf such as the one shown in (D). Note that bdg trichomes do not form bulbous structures at their bases as wild-type trichomes, shown in the inset for comparison, do (arrow). Subsidiary cells (sc) forming a ring around the trichome cell appear unaffected. (F) Typical collapsed trichome in bdg. It appears similar to the trichomes shown in Figure 1G on the rosette leaves of bdg plants grown in a greenhouse. (G) Epidermal ruptures (arrows) in the bdg rosette leaf such as the one shown in (D). (H) Degenerative changes in the bdg epidermis resulting from the progression of tears. (I) Graph showing the mean number of secondary roots in bdg and wild-type plants grown on vertical agar plates. Results are presented as means ± se for 16 bdg and 32 wild-type plants. (J) Four bdg and four wild-type plants grown on a representative vertical plate for 1 week. (K) and (L) Higher magnification of bdg roots (K) compared with wild-type roots (L). Mutant roots produce more and longer root hairs.

Figure 5.

GC-MS Analysis of Residual Bound Lipids in bdg and Wild-Type Leaves. Graphs (A) and (B) show the amount of each compound (expressed in micrograms per square centimeter and as a percentage of the total, respectively). Each value is the mean ± se of six replicates. The numbers correspond to the order of the peaks in the GC-MS scan. Monomers found in Arabidopsis cutin in an independent experiment (Franke et al., 2005) are indicated by black circles; see Franke et al. (2005) and Kurdyukov et al. (2006) for details of the method.

Figure 6.

Wax Biosynthesis in bdg. (A) Analysis of the composition of epicuticular wax in bdg and wild-type leaves. Compound classes and carbon chain lengths are indicated on the x axis. Each bar shows the relative amount of a specific constituent as the mean ± se of five replicates. (B) Increased expression levels of genes in the decarbonylation pathway. Quantification of mRNA was performed by RT-PCR. Transcript abundance is expressed relative to the amount of product amplified using primers for the control transcript, actin. The SHN1/WIN1 gene codes for an ethylene response factor–type transcription factor, which can activate several genes involved in wax biosynthesis in Arabidopsis, including CER1. CER1 encodes a putative fatty aldehyde decarbonylase, or a transmembrane protein that regulates it, or a protein involved in the secretion of alkanes (Aarts et al., 1995; Broun et al., 2004). WAX2/YRE has high sequence homology with CER1, but its function in the decarbonylation pathway is even less clear; it may encode an aldehyde-generating enzyme in the metabolic step preceding that catalyzed by CER1 (Chen et al., 2003; Kurata et al., 2003). The number of repetitions was 3 for WAX2/YRE and 12 for SHN1/WIN1 and CER1. Bars indicate means ± se. P values calculated by a nonparametric Wilcoxon-Mann-Whitney test (Statistica 6.0 package from StatSoft) are depicted above the graph bars if they are ≤0.05.

Figure 7.

Mutant Alleles of the BDG Gene. The location and orientation of En/Spm transposons in various mutant alleles of BDG (At1g64670, GenBank accession number AJ781319) are shown by the arrows. The bdg-2 allele is a derivative of bdg-1 and contains a transposon footprint (7-bp deletion) (see Supplemental Figure 1 online). Exons, introns, and 5′ and 3′ untranslated regions are boxed; exons are shown as thicker boxes.

Figure 8.

Structure and Sequence Alignment of BDG with α/β-Hydrolases. The amino acids that form the catalytic triad in α/β-hydrolases are indicated by triangles. (A) Alignment of BDG with the representative, relatively high-scoring BioH protein from GenTHREADER searches (McGuffin and Jones, 2003). The known secondary structure sequences of BioH and predicted secondary structure sequences of BDG, represented by sequences of helixes (h), extended strands (E), and random coils (c), are correlated with the corresponding amino acid sequences. (B) Proposed domain structure of the predicted BDG protein shown relative to a Kyte–Doolittle hydropathy plot generated by the MacVector sequence analysis package (Accelrys). In the plot, hydrophilic portions are above 0, and hydrophobic areas are below 0. Peaks with scores of >1.8 using a window size of 19 correspond to a possible N-terminal signal peptide (1) and a hydrophobic region in the α/β-hydrolase domain (2). NTHD, N-terminal hydrophilic domain; SP, signal peptide. (C) Active-site Ser regions of various α/β-hydrolases. Shading highlights amino acid identities found in >35% of cases. Shown at left are functionally and structurally related gene families: (1) cutinases and cutinase-like proteins; (2) BioH family; (3) BDG family; (4) monoglyceridelipase and lysophospholipases; (5) putative lysophospholipases of Arabidopsis; and (6) various lipases and esterases.

Figure 9.

Tissue-Specific Expression of BDG. Series of images showing GFP fluorescence in transgenic Arabidopsis plants expressing GFP under the control of the BDG promoter ([A] to [E]) and in situ mRNA expression patterns of BDG ([F] to [I]) in wild-type plants. The GFP images were acquired using a confocal laser scanning microscope. GFP fluorescence is green (channel 520 to 540 nm), and autofluorescence is red (channel >590 nm). For mRNA expression analysis, tissue sections were hybridized with a digoxigenin-labeled riboprobe antisense to BDG cDNA. After immunodetection, using an alkaline phosphatase–conjugated anti-digoxigenin antibody, the sections were inspected with a light microscope. Bars = 100 μm for (A) to (C), 200 μm for (D) to (F), and 50 μm for (G) to (I). (A) Longitudinal section through a vegetative apex. (B) Lateral (secondary) root primordium. (C) Optical section through the lateral root tip. (D) Longitudinal section through a young floral bud showing the pistil (pi) and the stamen (st). (E) Cross section through a floral bud at a later stage showing that strong GFP fluorescence is retained at the abaxial (external) side of the pistil. A weaker GFP signal is also visible in ovules (ov), petals (pe), and sepals (se). (F) Cross section through a vegetative shoot apex of a short-day-grown wild-type plant. Note the particularly strong expression of BDG mRNA in the epidermis on the flanks of leaf primordia. (G) The part of (F) inside the rectangle is magnified and displayed. (H) Longitudinal section through a vegetative shoot apex of a long-day-grown wild-type plant. BDG expression is detected in the epidermis. (I) Cross section through a floral bud showing BDG expression in the adaxial epidermis in prefusion carpels. At this stage, BDG is hardly detectable in other flower organs.

Figure 10.

Subcellular Immunolocalization of the Strep-Tag Epitope-Tagged BDG Protein. Cross sections were cut from the paraffin-embedded vegetative shoot apexes of short-day-grown plants. After incubation of the sections with an epitope-specific monoclonal antibody, the fluorescent tyramide signal amplification method was used to detect BDG–Strep-tag. Images were taken with a fluorescence microscope ([A] to [C]) or a confocal laser scanning microscope ([D] to [G]). Bars = 500 μm for (A) and (B), 250 μm for (C), 40 μm for (D) and (E), and 15 μm for (F) and (G). (A) A vegetative shoot apex in bdg used as a negative control together with that of a wild-type plant (data not shown). (B) to (G) A vegetative shoot apex in bdg complemented with a construct carrying the BDG protein C-terminally tagged with the Strep-tag epitope sequence under the control of the native BDG promoter. Note that deformation in the shape of the leaf primordia occurs in bdg (A) but not in the complemented mutant (B). Specific fluorescent signals in the epidermis (arrows) were detected only in bdg BDG–Strep-tag plants ([B] to [G]) but not in negative controls. A magnified view of the leaf primordia boxed in (B) is shown in (C). (D) and (F) Transverse sections corresponding to the light microscope images in (E) and (G), respectively. The Strep-tag–specific fluorescent signal is found localized in the outermost epidermal cell walls (arrows).