A new method for rapid visualization of defects in leaf cuticle reveals five intrinsic patterns of surface defects in Arabidopsis

Authors

  • Toshihiro Tanaka,

    1. Division of Biological Science, Graduate School of Science, Nagoya University, Furo-cho, Chikusa-ku, Nagoya 464-8602, Japan,
    Search for more papers by this author
  • Hirokazu Tanaka,

    1. College of Bioscience and Biotechnology, Chubu University, 1200 Matsumoto-cho, Kasugai, Aichi 487-8501, Japan, and
    2. CREST, Japan Science and Technology Corporation, Japan
    Search for more papers by this author
  • Chiyoko Machida,

    1. College of Bioscience and Biotechnology, Chubu University, 1200 Matsumoto-cho, Kasugai, Aichi 487-8501, Japan, and
    2. CREST, Japan Science and Technology Corporation, Japan
    Search for more papers by this author
  • Masaru Watanabe,

    1. Division of Biological Science, Graduate School of Science, Nagoya University, Furo-cho, Chikusa-ku, Nagoya 464-8602, Japan,
    Search for more papers by this author
  • Yasunori Machida

    Corresponding author
    1. Division of Biological Science, Graduate School of Science, Nagoya University, Furo-cho, Chikusa-ku, Nagoya 464-8602, Japan,
      For correspondence (fax +81 52 789 2966; e-mail yas@bio.nagoya-u.ac.jp).
    Search for more papers by this author

For correspondence (fax +81 52 789 2966; e-mail yas@bio.nagoya-u.ac.jp).

Summary

The epidermis of higher plants generates the cuticle layer that covers the outer surface of each plant. The cuticle plays a crucial role in plant development, and some mutants with defective cuticle exhibit morphological abnormalities, such as the fusion of organs. The way in which the cuticle forms and its contribution to morphogenesis are poorly understood. Conventional detection of the cuticle by transmission electron microscopy (TEM) requires laborious procedures, which include fixation, staining with osmium, and preparation of ultra-thin sections. It is also difficult to survey entire surfaces of expanded leaves because of the limited size of specimens that can be examined. Thus, TEM is unsuitable for large-scale screening for mutants with defective cuticle. We describe here a rapid and inexpensive method, designated the toluidine-blue (TB) test, for detection of cuticular defects in whole leaves. We demonstrated the validity of the TB test using mutants of Arabidopsis thaliana, including abnormal leaf shape1 (ale1), fiddlehead (fdh), and five eceriferum (cer) mutants, in which the structure and/or function of the cuticle is abnormal. Genetic screening for mutants using the TB test allowed us to identify seven loci. The cuticle-defective regions of leaves of the mutants revealed five intrinsic patterns of surface defects (classes I through V), suggesting that formation of functional cuticle on leaves involves various spatially regulated factors.

Introduction

Terrestrial animals and plants often find themselves exposed to hostile environmental conditions. Extreme dehydration and harmful radiation, for example, can cause critical damage. Thus, during their evolution, terrestrials acquired the ability to survive in changing and unfavorable environments. Land plants, in particular, require sophisticated protective mechanisms to withstand dehydration, radiation, heat, cold, and fatal attacks by herbivores.

The epidermis that covers the outermost surface of land plants and animals plays a central protective role and has highly specialized features in both plants and animals. In flowering plants, the protoderm and the epidermis generate an exocytoskeletal cuticle on the outer surface of the embryo proper (Tanaka et al., 2001) and of the aerial portions after germination (reviewed by Martin and Juniper, 1970). The cuticle of flowering plants is a complex matrix that consists mainly of lipids, namely cutins and waxes (Kolattukudy, 1980), whereas the main components of animal cuticles are chitins and proteins. The cuticular materials are derived from the epidermis and are deposited specifically on the outer surface, and thus, cuticle formation obviously depends on the appropriate functions of the protoderm and the epidermis.

In plants, several genes have been shown to be involved in surface functions or formation of leaf epidermis. These genes encode proteins that appear to participate in the biosynthesis of cuticle (Pruitt et al., 2000; Wellesen et al., 2001; Yephremov et al., 1999), in signal transduction (Becraft et al., 1996; Tanaka et al., 2001), and in the expression of epidermis-specific genes (Abe et al., 2003). For example, mutations in the FIDDLEHEAD (FDH) gene, which encodes an enzyme that might catalyze the elongation of fatty acids, result in the production of leaves and floral organs with altered surface function such that organs are fused to one another (Lolle et al., 1992, 1998; Pruitt et al., 2000; Yephremov et al., 1999). Furthermore, loss-of-function mutations in the ABNORMAL LEAF SHAPE1 (ALE1) gene, which encodes a subtilisin-like serine protease, result in fused cotyledons and leaves. Ultrastructural analysis by transmission electron microscopy (TEM) of surfaces of ale1 cotyledons revealed a discontinuous cuticle, with partial exposure of cell walls (Tanaka et al., 2001). The crinkly4 (cr4) gene of maize is also involved in the formation of cuticle and encodes a kinase that resembles the tumor necrosis factor (TNF) receptor kinase (Becraft et al., 1996). These observations suggest the existence of a signaling pathway that controls the formation of the cuticle. However, the genetic and biochemical relationships among these genes and gene products are poorly understood, and the mechanisms controlling the differentiation and/or function of the plant epidermis remain to be characterized.

As it is the epidermal cells that specifically generate a plant cuticle, cuticle formation can be regarded as an indicator of appropriate epidermal differentiation. The cuticle of Arabidopsis thaliana can be stained with osmic acid and visualized by TEM, but such observations require the laborious preparation of samples and are unsuitable for large-scale screening for cuticle-defective mutants. It is also difficult to observe cuticle defects on the entire surface of an expanded leaf by TEM because of limits to the size of each specimen that can be examined. In an attempt to simplify visual analysis of plant cuticle to facilitate investigations of epidermal differentiation and cuticle formation, we developed the toluidine-blue (TB) test. Application of the TB test to genetic screening yielded several novel mutants with five patterns of cuticular defects. Our findings suggest that the TB test will be very useful for monitoring cuticular and/or epidermal defects.

Results

Cuticular defects in the ale1 and fdh mutants were easily visualized after staining with TB

The cuticle layer, a hydrophobic extracellular matrix derived from the epidermis, restricts the movement of water across the epidermal surface of land plants. We postulated that leaf surfaces without a functional cuticle might be permeable to water-soluble molecules, and we anticipated that the cuticular defects might be revealed by staining with an aqueous solution of a dye, such as TB, which is a hydrophilic dye in general use for histological staining (Figure 1a). We subjected plants that were homozygous for the ale1 mutation to staining with TB. The ale1 seedlings, with cuticular defects, were partially stained with TB (Figure 1c) whereas wild-type plants were not (Figure 1b). It has been suggested that the fdh mutant (Lolle et al., 1992, 1997; for detail of the allele, see Experimental procedures), in which organs are fused, might have an altered cuticle, and this mutant was also sensitive to TB staining or the ‘TB test’ (Figure 1e).

Figure 1.

Schematic representations and application of the TB test to the detection of permeable epidermal surfaces.

(a) A conceptual diagram of the TB test. Plant with a normal cuticle repels TB (top). A deficient cuticle allows TB to permeate the epidermal surface. An incomplete cuticle results in patchy staining (middle) and complete loss of cuticle results in overall staining (bottom).

(b, c) Seven-day-old plants were stained with TB: (b) Ler (wild type); and (c) ale1-1 mutant, in which patchy staining was observed.

(d, e) 21-day-old fdh-12 mutant before (d) and after (e) staining. Scale bars: (b,c), 1 mm; (d,e), 5 mm.

The TB test allowed detection of epidermal defects

Judging from the results of the TB test in the ale1 and the fdh mutants, we postulated that mutations that cause cuticular defects might result in plants with TB sensitivity. Indeed, we found that subsets of ethyl methanesulfonate (EMS)-mutagenized populations were stained with TB to various extents and were easily distinguishable from the wild type. Moreover, several mutant lines yielded segregation ratios typical of monogenic, recessive mutations (see below). Thus, we reasoned that the TB test might be a useful method for visualizing inherited defects in the cuticle layer.

While refining the TB test, we found that wild-type plants were also stained with TB after growth at high humidity, when seedlings became vitreous (data not shown). We were able to circumvent such problems by modifying the growth conditions (see Experimental procedures).

Isolation of mutants using the TB test

In order to identify mutations that might affect permeability of the leaf surface, we screened an EMS-mutagenized population for mutations that allowed staining of plantlets with TB. We initially examined the cotyledons of approximately 50 seedlings descended from each of 3305 independent M2 lines using the TB test, and we obtained 130 TB-positive lines. However, at the time of this first screening, we had not yet optimized the TB test. Therefore, the initial candidates compromised a population that included both true mutants and potential false-positives. We performed a second screening, based on the TB sensitivity of rosette leaves of 3-week-old plantlets, and identified 49 secondary positive lines. Among these lines, we have confirmed, to date, that 19 lines show inheritability of defects with appropriate Mendelian segregation. In this report, we describe 9 of the 19 mutant lines for which loci were mapped. The results of genetic mapping, an examination of allelism, and phenotypic classification indicated that these mutations occurred at seven loci (see below). We refer collectively to such loci as permeable leaves (pel), with the sequential serial numbers shown in Table 1, except in the case of the fdh-13 mutant, which had a mutation in the FDH gene (see below; Pruitt et al., 2000; Yephremov et al., 1999).

Table 1.  Details of the alleles identified in the present study
AlleleOriginal
line number
Map
positiona
Staining
patternb
FusionscMorphological change
in epidermal pavement cellsd
Fertilitye
  • a

    Details of positional information are shown in Figure 4.

  • b

    Staining patterns were divided into five classes. For the detail of each class, see Figure 2.

  • c

    Fusion phenotype of vegetative and reproductive organs: +, fusion with no variations among individuals; +/−, fusion with variations among individuals; −, no fusion observed.

  • d

    Examined by SEM: ++, severe phenotype; +, mild phenotype; NT, not tested. For the details of the phenotypes, see Supplementary Material (Figure S2).

  • e

    Spontaneous fertility: +, fertile; +/−, significantly reduced fertility; −, loss of fertility.

  • f

    The pel6 allele was allelic to the wax2/yre allele, as discussed in the text.

pel1-11077Chromosome 1 topII+++
pel1-23254Chromosome 1 topII+++
pel2-10003Chromosome 1 middleI++
pel2-20619Chromosome 1 middleI++
fdh-133033Chromosome 2 middleI, II+(NT)
pel30269Chromosome 5 topII, III++++/−
pel42230Chromosome 2 middleIII(NT)+
pel51812Chromosome 1 bottomIII(NT)+
pel6 (wax2/yre)f1806Chromosome 5 bottomV+/−(NT)

The TB test allowed discrimination between pel and cer mutations

Table 1 shows the annotations of the loci identified by the TB test. All of the mutations were recessive. The staining patterns of the mutants could be divided into five classes (Table 1). We also examined the sensitivity to staining with TB of leaves of 20 eceriferum (cer) mutants, which were originally identified by a deficiency in epicuticular wax (Koornneef et al., 1989). Details of the results for these mutants are shown in Figure S1, which can be found as Supplementary Material. Five cer mutants exhibited characteristic staining patterns.

Figure 2 shows our classification of staining patterns into five classes. These staining patterns were reproducible. The staining pattern designated class I was characterized by uniform staining and was observed mainly in the pel2 mutant exclusively (Figure 3h). Many more mutants fell into classes II and III. Class II was characterized by patchy and random staining. The pel1, fdh-13, pel3, and cer14 mutations belonged to this class (Figures 3g,i,j and S1). The staining pattern designated class III, in which proximal regions of the lamina and the petiole itself were stained predominantly, was often observed in the pel4, pel5, cer5, and cer12 mutants (Figure 3p,q and S1). The sensitivity to TB staining of the mutants in classes II and III was stronger in the abaxial and basal regions than in the adaxial and distal regions of leaves (data not shown). The cer19 mutant was unique in terms of the distal staining of leaves (Figure S1). This pattern, designated class IV, could be regarded as complementary to the class III pattern. Class V was defined as a distinct pattern of staining that was focussed on the trichomes. This class included the pel6 and cer10 mutants (Figure 3r–t and S1). In addition to the trichomes, the basal parts of leaves were also slightly stained in these mutants.

Figure 2.

Classification and distribution of TB-staining patterns.

Staining patterns were divided into five classes on the basis of the staining of adaxial surfaces of rosette leaves, as indicated. The intensity of shading beside each allele represents the frequency with which particular classes of staining pattern were observed. Refer also to photographs of pel mutants in Figure 3 and to Supplementary Material (Figure S1) for cer mutants.

Figure 3.

Gross morphology and the TB-staining patterns of mutants.

(a–e,k–m) Gross morphology of the 18-day-old plants grown on plates.

(f–j,p–r) Permeability of leaf surfaces was examined by the TB test. White frames in panels (m,n,r,s) show the regions depicted in panels (n,o,s,t), respectively.

(a,f) Ler (wild type); (b,g) pel1-1; (c,h) pel2-1; (d,i) fdh-13; (e,j) pel3-1; (k,p) pel4-1; (l,q) pel5-1; (m–o,r–t) pel6-1. Scale bars: (a–j,k–m,p–r), 5 mm; (n,s), 1 mm; (o,t), 0.5 mm.

Characterization of the fusion phenotype of pel mutants

In some pel mutants, organs appeared to have fused. Conspicuous fusion was evident in the pel1, pel2, and fdh-13 (Figure 3b–d) mutants. In the pel3 mutant, the extent of fusion was relatively limited (Figure 3e). Such organ fusion also occurred in the reproductive organs (data not shown). All these mutants exhibited complete loss of fertility, with the exception of the pel3 mutant (Table 1). Some cuticle-defective mutants, including ale1 and lacerata (lcr), have been reported to show simplistic morphology of epidermal pavement cells as well as the fusion phenotype (Tanaka et al., 2001; Wellesen et al., 2001). The pel1 and pel3 mutants also had similar morphological changes (Table 1; Figure S2).

The mutations at the pel4 and pel5 loci generated TB-sensitive plantlets but the leaves did not fuse (Figure 3k,l). Both the vegetative organs and the reproductive organs of these mutants were indistinguishable from those of wild-type plants. These mutants also did not exhibit any cer-like phenotype.

Fusions between trichomes were observed in the pel6 mutant (Figure 3n,o). The resultant bridge-like trichomes generated a force that opposed the expansion of the lamina, which resulted in distorted lamina (Figure 3n,o). The reproductive stage of the pel6 mutant was characterized by a glossy inflorescence stem (data not shown), which is typical of the phenotype of cer mutants (Koornneef et al., 1989; data not shown), as well as loss of fertility (Table 1).

Genetic mapping of the pel loci

Genetic mapping revealed that the pel loci were distributed at different positions and, in addition, that pel4 mapped close to the fdh locus (Figure 4). However, the pel4 mutant never exhibited the fusion phenotype of the fdh mutant, and moreover, the pel4 mutant genome had no mutations in the FDH coding region (data not shown). Thus, it seems unlikely that the pel4 and the fdh loci are allelic. Direct sequencing of the FDH coding region, using fdh-13 genomic DNA as template for amplification, revealed a single base change from G to A in exon 1 of the FDH gene. This mutation generates a termination codon in place of a tryptophan residue at position 92, resulting in a truncated protein. The pel5 locus was mapped close to ale1. However, it was not allelic to ale1 because recombinants were obtained with respect to the polymorphism within the ALE1 coding region (bacterial artificial chromosome clone F24O1; Figure 4).

Figure 4.

Map positions of genes or alleles associated with defects in epidermal function.

The alleles identified here are shown in blue letters. Solid triangles indicate genes whose molecular characterization has been reported. Roughly mapped alleles are shown at their approximate positions. Semi-fine mapping was performed for pel1, pel2, pel3, and pel5. In each close-up view of these loci, numbers of recombination events at each DNA marker (shown as a vertical line), which is located on the chromosomal position shown in megabase (Mb) pairs on the scale at top, are indicated. Four clone contigs spanning each chromosomal region are represented as horizontal red bars with their clone name deposited on GenBank. References for known loci are as follows: deadhead (ded), bulkhead (bud), hothead (hth), conehead (cod), cer10, thunderhead (thd), airhead (ahd), and pothead (phd; Lolle et al., 1998); FDH (Lolle et al., 1998; Pruitt et al., 2000; Yephremov et al., 1999); ALE1 (Tanaka et al., 2001); LCR (Wellesen et al., 2001); A. thaliana homolog of cr4 (ACR4; Tanaka et al., 2002); PROTODERMAL FACTOR2 (PDF2; Abe et al., 2003); A. thaliana MERISTEM LAYER1 (ATML1; Lu et al., 1996); and WAX2/YRE; Chen et al., 2003; Kurata et al., 2003).

The pel6 locus was mapped in the vicinity of the WAX2/YORE-YORE (YRE) gene (Figure 4; Chen et al., 2003; Kurata et al., 2003), and the phenotypic effects of mutations in this gene were similar to those of pel6. We sequenced the open-reading frame that corresponded to the WAX2/YRE gene in the pel6 mutant and found a single base substitution. The pel6 mutation resulted in generation of a premature termination codon (W534Z) in WAX2 mRNA (GenBank Accession no. AY131334).

Discussion

We demonstrated that defects in the cuticle of A. thaliana can be readily visualized by the TB test. Our results show that this method allows rapid and inexpensive examination of defective cuticle over the entire leaf surface and in large numbers of plants.

The Arabidopsis Genome Initiative (2000) yielded the sequence of the entire genome in 2000. The developing database has enhanced the possibilities for reverse genetics, but standard genetics, with the isolation and analysis of the mutants, remains powerful. The TB test described herein should enhance genetic screening for mutations in genes that control the structure and function of the leaf cuticle. In this test, results can be obtained even at the seedling stage. As some mutants with defective cuticle are likely to exhibit reduced resistance to low humidity (Tanaka et al., 2001), growth conditions for such mutants would have to be carefully controlled. Therefore, the ability to recognize cuticular defects at an early stage of growth saves effort, time, and money. Moreover, mutants with no apparent morphological abnormalities could be identified by the TB test so long as they had a defective cuticle. This advantage is important because so many mutants of A. thaliana with morphologically obvious defects have already been generated. In addition, the TB test paves the way for whole-mount analysis of cuticle formation, which is impossible to achieve by TEM as the maximum size of each specimen that can be examined by TEM is so small.

We have not proved the unequivocal correlation between sensitivity to TB and a cuticular defect. Further phenotypic analysis of the mutants obtained here, as well as molecular cloning of the corresponding genes, will help us to clarify the relationship between the structure of the cuticle layer and its permeability to water-soluble molecules. Recently, we have found the pel3 mutation in an open reading frame that is predicted to encode an enzyme that might be involved in fatty acid biosynthesis (our unpublished data). The amino acid sequence predicted from the PEL3 gene has partial similarity to that from the CER2 gene that has been proposed to function in fatty-acid elongation (Xia et al., 1996), hence in epicuticular wax biosynthesis. This finding demonstrates that the TB test work in a detection of epidermal surface defects as we expected.

Our genetic screening, using the TB test, allowed us to identify at least seven loci, including a new allele of the well-characterized FDH gene (fdh-13; Pruitt et al., 2000; Yephremov et al., 1999) and a new allele of the WAX2/YRE gene (pel6; Chen et al., 2003; Kurata et al., 2003). Whole-mount analysis also allowed us to assess both region-specific and organ-specific defects in the epidermis. As shown in Figure 2, each mutant exhibited a characteristic pattern of staining. The staining pattern might reveal the spatial distribution of the regions in which the corresponding gene function is required. These data, coupled with future analysis of gene expression, will contribute to the functional analysis of the relevant genes.

In this study, we subjected cer mutants to the TB test. The cer phenotype was described as the absence of lipid crystals on the surface of inflorescence stems and/or siliques (e.g. Fiebig et al., 2000; Koornneef et al., 1989; and references therein). The waxy constituents of the stems and leaves of each cer mutant were also reported (e.g. Jenks et al., 1995; Rashotte et al., 2001; and references therein). The sensitivity to TB of the cer mutants that we examined did not necessarily reflect these chemical alterations. It is noteworthy that the cer mutants exhibited a variety of staining patterns as such patterns reveal, for the first time, the regional specificity of cuticular defects in leaves of such mutants. Detailed phenotypic analysis of regions of leaves with altered surface properties in the cer mutants should help us to understand the spatiotemporal contribution of CER genes to cuticle development in leaves.

Some mutations that rendered leaves TB sensitive also caused organ fusion. The fusions observed in the pel1, pel2, pel3, and fdh-13 mutants were typical of known epidermis-defective mutants. In these mutants, there was a good correlation between the extent of TB sensitivity and the severity of the fusion phenotype, suggesting that the results of the TB test reflected epidermal abnormalities accurately. Lolle et al. (1998) described a genetic screening for Arabidopsis mutants with the fusion phenotype. It will be of interest to examine the genetic interactions between the pel mutants and the mutants that they identified.

Mutants without an obvious fusion phenotype were also isolated in the screening by the TB test. These mutants, pel4 and pel5, were difficult to distinguish visually from wild-type plants, confirming the advantages of the TB test.

In summary, we propose that the TB test is an efficient method for investigations of the formation and/or function of plant epidermis. Studies of these issues are in progress using mutants that were isolated in the present study.

Experimental procedures

Plant materials

Arabidopsis thaliana ecotype Landsberg erecta (Ler) was used as wild-type strain (Hamada et al., 2000). For genetic screening, an M2 population of Ler seeds, mutagenized by treatment with ethyl methanesulfonate (EMS; cat. M2E04-06; groups 25 through 46) and derived from approximately 32 000 M1 plants, was purchased from Lehle Seeds (Round Rock, TX, USA). The M3 seeds were collected from each M2 plant, and 3305 M3 lines were obtained, each of which was derived from only a single M2 plant. When the plants were grown on soil, they were treated as described elsewhere by Tanaka et al. (2001). Mutants for phenotypic characterization were back-crossed two times for pel1, pel2, and pel3, and once for fdh-13, pel5, and pel6. The pel4 mutant shown in the Figures has yet been back-crossed. Seeds of cer mutants were obtained from the Arabidopsis Biological Resource Center (Columbus, OH, USA). The fdh-12 mutant was obtained by independent screening for the organ-fusion phenotype of an EMS-mutagenized population (ecotype Ler). This allele had a mutation at the 239th codon that resulted in the replacement of a glycine residue by a glutamate residue. The segregation ratio indicated that the fdh-12 allele was semidominant.

TB test

Plants for analysis by the TB test were grown as follows. Seeds were surface-sterilized and sown on plates of Murashige and Skoog's medium (Murashige and Skoog, 1962) solidified with 0.4% (w/v) gellan gum (Wako Pure Chemicals Industries Ltd., Osaka, Japan). The concentration of the gelling agent was critical for elimination of inappropriate staining. When the concentration of gellan gum was too low, we often obtained vitreous plants. Even wild-type plants were sensitive to TB staining under these conditions. After vernalization treatment for 3 days, plants were transferred to a growth chamber (22°C, continuous light). The growth stages of plants are indicated herein as days after vernalization. An aqueous solution of 0.05% (w/v) TB (Sigma, St Louis, MO, USA), which had been filtered through a fiber media filter (pore diameter, 0.2 µm; Spectrum Laboratories Inc., Rancho Dominguez, CA, USA), was poured directly onto the plates on which plants were growing until plants were submerged. After 2 min, the TB solution was removed and plates were washed gently with water to remove excess TB from plants.

Genetic mapping

Each mutant was outcrossed with the Columbia (Col) ecotype. The resultant F1 plants were self-pollinated, yielding a polymorphic F2 population. Genomic DNA was isolated from each individual F2 plant that exhibited a mutant phenotype for PCR-based linkage analyses, using simple sequence length polymorphism and cleaved amplified polymorphic sequence codominant molecular markers (Konieczny and Ausubel, 1993). The information about polymorphisms between Ler and Col was kindly provided by Cereon Genomics (http://www.arabidopsis.org/Cereon/index.html). Details of the primer sets for detection of Ler/Col polymorphisms, in which the most proximal recombinants were identified for each locus, are available on request.

Scanning electron microscopy

Scanning electron microscopy was performed as described by Semiarti et al. (2001).

Acknowledgements

The authors thank the Arabidopsis Biological Resource Center (The Ohio State University, Columbus, OH, USA) for providing seeds of cer mutants. This work was supported in part by a Grant-in-Aid for Scientific Research on Priority Areas (no. 14036216) to Y.M. and by a grant for Core Research in Evolutional Science and Technology (CREST) to C.M. from the Japanese Ministry of Education, Science, Culture, Sports and Technology. H.T. was supported by a grant from CREST.

Supplementary Material

The following material is available from http://www.blackwellpublishing.com/products/journals/suppmat/TPJ/TPJ1946/TPJ1946sm.htm

Figure S1. Discrimination among cer mutants by the TB test.

Figure S2. Morphology of epidermal pavement cells of pel mutants.

Ancillary

Advertisement