Epidermal Cell Differentiation in Arabidopsis Determined by a Myb Homolog, CPC

See allHide authors and affiliations

Science  22 Aug 1997:
Vol. 277, Issue 5329, pp. 1113-1116
DOI: 10.1126/science.277.5329.1113


The roots of plants normally carry small hairs arranged in a regular pattern. Transfer DNA–tagged lines of Arabidopsis thaliana included a mutant with few, randomly distributed root hairs. The mutated gene CAPRICE (CPC) encoded a protein with a Myb-like DNA binding domain typical of transcription factors involved in animal and plant development. Analysis in combination with other root hair mutations showed that CPCmay work together with the TTG gene and upstream of the GL2 gene. Transgenic plants overexpressingCPC had more root hairs and fewer trichomes than normal. Thus, the CPC gene determines the fate of epidermal cell differentiation in Arabidopsis.

The cellular organization of the primary root of Arabidopsis thaliana is relatively simple and invariant (1). During the maturation of the root epidermis in A. thaliana, each cell ultimately becomes either a root hair (trichoblast; which we shall hereafter term a root hair cell) or a hairless cell (atrichoblast) (2, 3). This choice may be determined by the position of the cell relative to the underlying cortical cell layer. Epidermal cell files that make contact with two cortical cell files by lying over the junction between the two cortical cell files are root hair cells. Epidermal cells that contact only one cortical cell file are hairless cells. Primary roots in wild-type Arabidopsis normally have eight files of cortical cells (Fig. 1F). Root hairs are tip-growing, tubular-shaped outgrowths that help to anchor roots, interact with soil microorganisms, and assist in the uptake of water and nutrients. TRANSPARENT TESTA GLABRA (TTG) andGLABRA2 (GL2) are genes that determine whether epidermal cells differentiate into root hair cells or hairless cells (3, 4). In ttg and gl2 mutants, all of the epidermal cell files differentiate into root hair cells independent of their position relative to the underlying cortical cells. TheGL2 gene encodes a homeodomain protein that is expressed preferentially in the differentiating hairless epidermal cells (4, 5). Although the TTG gene has not been cloned yet, it is believed to encode a protein with a Myc-like domain or a protein positively regulating a Myc-like gene, because the phenotype of the ttg mutant can be complemented by introducing a maize Myc gene, R, into the mutant. When the R gene is overexpressed in a wild-type plant, all of the root epidermal cells differentiate into hairless cells (3, 6). Thus,TTG and GL2 may inhibit the differentiation of root epidermal cells into root hair cells.

Figure 1

Patterns of root hair formation in the major root of 5-day-old seedlings. Root surfaces were photographed with a binocular microscope (Olympus SZH). (A) Wild type, (B) cpc mutant, (C) cpcmutant complemented by the C fragment (see Fig. 2A), (D) transgenic plant harboring 35S::CPCconstruct, and (E) enlarged picture of (D) taken with a camera attached to an Olympus PROVIS AX microscope. All epidermal cell files have root hairs. Scale bars: 200 μm in (A) to (D), and 30 μm in (E). (F) Drawing of a transverse section of the root. Eight epidermal cells that form root hair cells are indicated by arrowheads.

From a T3 population of transfer DNA (T-DNA)–tagged lines (7), we isolated a mutant with fewer than normal root hairs, which we named caprice (cpc) for the irregular distribution of root hairs (Fig. 1B). cpc is a nuclear mutation, not allelic to other known mutations. Heterozygous plants show the wild-type phenotype. From a cross between heterozygotes, about one-fourth (67/324) of the offspring had few root hairs, which indicated that cpc is a single, recessive mutation. The number of root hairs in the primary root of the cpc mutant was about one-fourth of that of the wild type (Table1). The morphology and size of the root hairs produced by the cpc mutant were indistinguishable from those of wild-type hairs. The addition of 1-aminocyclopropane-1-carboxylic acid (ACC), an ethylene precursor, at 5 × 10-6 M induced root hair production incpc seedlings; however, the number of hairs was about 30% of that of the ACC-treated wild type, indicating that ethylene cannot rescue the phenotypic deficiency of the cpc mutant (8).

Table 1

Root hair number and epidermal cell length of various plant lines. Values represent the mean ± SD. Hair number indicates the number of root hairs formed on a segment of a root with an average length of epidermal cells (=ab/1000). The control of the complementation experiment is of plants transformed by the vector pARK5 without genomic fragments. N.T., not tested.

View this table:

To examine how the CPC gene works in combination with theGL2 and TTG genes, we analyzed the phenotype of double mutants (Table 1). The cpc gl2 double mutant had about the same number of root hairs as the gl2 mutant, showing the gl2 mutation to be epistatic to thecpc mutation. The cpc ttg double mutant had more root hairs than the cpc mutant, but less than thettg mutant, indicating either that the CPC and the TTG gene products work together or they work in two independent pathways that control the number of root hairs.

We obtained a 2-kb genome sequence adjacent to the left border of T-DNA by inverse polymerase chain reaction (PCR) and confirmed the close linkage between the cpc locus and the T-DNA insertion (9). The chromosomal location of CPC was mapped by restriction fragment length polymorphism (RFLP) to the lower arm of chromosome II, 3.1 centimorgans from the m336 marker (10). From a genomic library and a cDNA library made from the roots of 5-day-old seedlings (11), we obtained four genomic and five cDNA clones carrying CPC. The five cDNA clones were isolated by screening 3 × 106 clones. No CPC mRNA was detected by Northern hybridization. These two results indicated that transcription of CPC is very rare in the root tissues. The CPC gene has an open reading frame of 1170 base pairs (bp) composed of three exons of 233, 88, and 263 bp and two introns of 73 and 513 bp (Fig. 2A). The longest cDNA (Fig. 2B) is 584 bp long, and the predicted CPC protein contains 94 amino acid residues. The CPC gene has a potential TATA box about 30 bp upstream of the putative transcriptional initiation site. An additional open reading frame in the inverse orientation was found 3 kb upstream of the 5′ terminal of CPC. The corresponding cDNA clone, number 51, was isolated, but the function of the gene is unknown.

Figure 2

Chromosome structure ofCPC and the deduced amino acid sequence of its gene product. (A) Genomic structure of the CPCregion. Genomic fragments used for complementation experiments are shown above the map: A, 7.3-kb Xba I–Sau 3AI fragment; B, 3-kb Hinc II fragment; C, 4.4-kb Eco RI fragment. Open boxes indicate the exons. (B) The deduced amino acid sequence of CPC. The region homologous to the Myb DNA binding domain is highlighted. The nucleotide sequence data has been submitted to the DNA Data Bank of Japan, European Molecular Biology Laboratory, and GenBank databases with accession number AB004871. (C) Sequences of the Myb domain of CPC compared with that of the C1, Pl, and P proteins from maize (14); the GL1 protein from Arabidopsis(15); the Mixta protein from Antirrhinum majus(16); and the human c-Myb protein (13). Highlighted residues indicate amino acids that are identical to those of CPC. The W residues marked by dots indicate the position the tryptophan residues conserved in the Myb domain. Single-letter abbreviations for the amino acid residues are as follows: A, Ala; C, Cys; D, Asp; E, Glu; F, Phe; G, Gly; H, His; I, Ile; K, Lys; L, Leu; M, Met; N, Asn; P, Pro; Q, Gln; R, Arg; S, Ser; T, Thr; V, Val; W, Trp; and Y, Tyr.

We used genomic fragments, designated A, B, and C (Fig. 2A), to complement the cpc mutation by transformation (12). All three fragments were able to complement thecpc mutation (Fig. 1C); root hair numbers were equivalent to the wild type (Table 1). Because fragment C contained 525 bp of sequence upstream of the 5′ terminal of the cDNA, and fragment B contained 309 bp of sequence downstream of the cDNA, these untranscribed regions may be sufficient for the correct expression ofCPC.

The deduced amino acid sequence of the CPC protein showed homology to the DNA binding domain of the proto-oncogene Myb (Fig. 2C). Although the other known Myb proteins have two or three Myb domains, theCPC gene encodes only one Myb-like domain, which most resembles the second Myb-like repeat in other plant Myb proteins and the third repeat in the Myb proteins of vertebrates. The Myb DNA binding domains are sequences of about 50 amino acids that form helix-turn-helix folds (13). The two tryptophan residues are conserved in the same position as in other Myb proteins. CPC has no proline-rich or acidic domain typical of a transcriptional activation domain. Insertion of a T-DNA into the CPC Myb homologous region converted the codon of the 75th amino acid residue, tyrosine (TAT), to a stop codon (TAA). Therefore, the CPC protein of the mutant lacks the COOH-terminal 19 amino acids that include part of the Myb domain. Several plant genes carrying Myb domains have been isolated. Mutation of maize genes Pl, C1, or Presults in a defect of pigment synthesis (14). AnArabidopsis mutant carrying a mutation in the GL1gene lacks trichomes (15). Mutation of the Mixtagene in A. majus modifies the cell shape of the petal epidermis (16).

To examine the function of the CPC protein, we analyzed the phenotype of transgenic Arabidopsis plants in which the CPCgene was overexpressed under the control of the 35S promoter of the cauliflower mosaic virus (17). Northern blot experiments showed that the expression of CPC was increased in both the roots and leaves of the transgenic plants (18). The 35S::CPC transgenic plants had ectopic root hairs in the roots (Fig. 1, D and E). Almost all of the root epidermal cells formed a root hair in two independent transgenic lines. The phenotype was the same as that of the gl2 andttg mutants (Table 1 and Fig. 1E). In addition to the ectopic root hair formation, the35S::CPC transgenic plants lacked trichomes on the leaves, stems, and sepals (Fig.3). Trichome formation is controlled by another Myb gene, GL1 (15). It is supposed that in the wild type, GL1 is a Myb gene expressed in shoots, whereas CPC is one expressed in roots. The phenotype of35S::CPC seedlings could be interpreted to indicate that the ectopic expression of CPC might interfere with the action of GL1 on TTG orGL2 (both of which block root hair formation) but induce trichome formation. GL2 is the more likely relevant target, because the phenotype of cpc gl2 and cpc ttgdouble mutants indicated that CPC may work together withTTG and upstream of GL2 in the developmental pathway of root hair formation. Expression of GL2 may be regulated negatively by CPC but positively byGL1, because the GL1 protein carries a transcriptional activation domain, whereas the CPC protein does not. If this is true, it is reasonable that transgenic plants overexpressing theCPC gene would show the same phenotype of the gl2mutant.

Figure 3

Trichome pattern on the surface of the rosette. (A) Wild type. (B) Transgenic plant harboring35S::CPC construct. Scale bar, 500 μm. Photos were taken with a camera attached to a binocular microscope (Leica wild M420).

Both trichomes and root hairs are outgrowths of epidermal cells. Our results may indicate that the two cell differentiation systems are controlled by certain common pathways regulated by the Myb-like proteins.

  • * Present address: Biomolecular Engineering Research Institute, 6-2-3 Furuedai, Suita, Osaka, 565, Japan.

  • To whom correspondence should be addressed at the Department of Botany, Graduate School of Science, Kyoto University, Kyoto, 606-01, Japan. E-mail: kiyo{at}


Stay Connected to Science

Navigate This Article