Constructing a Plant Cell. The Genetic Control of Root Hair Development

TRANSPARENT TESTA GLABRA (TTG)

TTG was the first locus to be identified as essential for root epidermal cell patterning in Arabidopsis (Galway et al., 1994). The homozygous recessive ttg mutants possess root hairs on almost every root epidermal cell, implying that the normal role of the TTG product is either to promote specification of the non-hair cell type or to repress specification of the root hair cell type in immature cells located in the N position. Becausettg mutations alter all aspects of non-hair cell differentiation (including early developmental characteristics such as cell division rate, cytoplasmic density, and vacuolation rate), TTG is likely to be an early-acting component in the cell specification process (Galway et al., 1994; Berger et al., 1998b).

The TTG gene encodes a small protein with WD40 repeats (Walker et al., 1999). Although the protein sequence does not provide an immediate mechanistic understanding of TTG's role, other WD-repeat proteins are involved in protein-protein interactions, suggesting that TTG might be a component of a signal transduction pathway or might interact with transcription factors that specify epidermal cell fate. At present, the pattern of TTG gene expression in the root is unknown.

Basic Helix-Loop-Helix Protein (bHLH)

The analysis of the ttg mutant led to the notion that a bHLH transcription factor in Arabidopsis contributes to root epidermal cell specification. Galway et al. (1994) showed that ttg can be functionally complemented by expressing the cDNA of the maize (Zea mays) R gene (encoding a bHLH transcriptional activator; Ludwig et al., 1989) under the control of the cauliflower mosaic virus 35S promoter in Arabidopsis. In addition to restoring the normal specification of non-hair cells in the N position, the35S::R transgene causes cells in the H position to develop as non-hair cells (i.e. ectopic non-hair cells; Galway et al., 1994). Together, these findings suggest that an R-like bHLH protein exists in Arabidopsis to promote non-hair cell specification and acts downstream from TTG.

GLABRA2 (GL2)

Homozygous recessive gl2 mutations cause nearly all root epidermal cells to produce root hairs (Masucci et al., 1996). However, in contrast to the ttg mutations, gl2mutations do not affect any of the other non-hair cell developmental phenotypes; i.e., the usual position-dependent differences in cell division rate, cytoplasmic density, and cell vacuolation rate are observed in the gl2 root epidermis (Masucci et al., 1996;Berger et al., 1998b). This implies that the GL2 product only influences a portion (or one branch) of the non-hair cell specification pathway.

The GL2 gene encodes a homeodomain transcription factor protein (Rerie et al., 1994; DiCristina et al., 1996), and it is preferentially expressed in the differentiating non-hair epidermal cells within the meristematic and elongation regions of the root (Masucci et al., 1996). GL2 gene expression is influenced by TTG and by the maize R bHLH protein, with ttg mutations causing a reduction in GL2 promoter activity and35S::R expression causing ectopic GL2promoter activity (Hung et al., 1998). Together, these findings suggest that TTG acts at an early stage in epidermis development to activate an R-like bHLH transcription factor, which in turn positively regulates the expression of GL2 (and perhaps other as yet unidentified genes) in a cell position-dependent manner to specify the non-hair cell type.

CAPRICE (CPC)

The cpc mutant produces a reduced number of root hair cells, implying that CPC is a positive regulator of root hair cell specification (Wada et al., 1997). Double mutant analysis shows that the gl2 mutation is epistatic to cpc, which suggests that CPC promotes root hair cell specification by acting as a negative regulator of GL2.

A possible explanation for CPC's negative action is provided by the nature of its gene product: CPC encodes a small protein with an MYB-like DNA binding domain but without a typical transcriptional activation domain (Wada et al., 1997). Thus, one possibility is that CPC acts as a transcriptional repressor ofGL2 by binding to its promoter and blocking its activation in epidermal cells located in the H position. Consistent with this possibility, expression of the CPC gene under control of the cauliflower mosaic virus 35S promoter (35S::CPC) induces ectopic root hair cells (Wada et al., 1997).

WEREWOLF (WER)

Homozygous recessive mutations affecting the wer locus cause nearly all root epidermal cells to differentiate as root hair cells, beginning at an early stage of epidermis development. TheWER gene encodes a typical MYB-type transcription factor of the R2R3 class and is preferentially expressed in cells located in the N position of the developing epidermis (Lee and Schiefelbein, 1999). The wer mutations also disrupt the position-dependent pattern of GL2 gene expression, leading to a reduced and largely random pattern of expression (Lee and Schiefelbein, 1999). Inspection of the GL2 promoter region known to be critical for position-dependent gene expression (Hung et al., 1998) reveals the presence of putative MYB-binding site elements. Together, these results support the possibility that WER directly regulatesGL2 transcription in cells located in the N position.

WER action is closely associated with a bHLH protein. The ability of the maize R bHLH protein (encoded by the 35S::Rtransgene) to influence root epidermis cell specification is dependent on WER activity, and the WER protein physically interacts with the R bHLH protein in the yeast two-hybrid assay (Lee and Schiefelbein, 1999). Thus, a WER-bHLH interaction may be responsible for the transcriptional regulation of GL2 and other genes to specify the non-hair cell fate in the Arabidopsis root epidermis.

The MYB encoded by WER and the truncated MYB encoded byCPC appear to exert opposing effects on root epidermis patterning. For example, the phenotype of the wer cpc double mutant is intermediate, as compared with the wer orcpc single mutants (Lee and Schiefelbein, 1999). The similar nature of the WER and CPC gene products and their opposing effects on epidermal cell patterning suggest that they act competitively to regulate epidermal cell fate.

A Model for Specification

The findings from these molecular genetic studies suggest a simple view for the control of root epidermal cell specification in Arabidopsis (Fig. 1). In brief, the fate of an epidermal cell may be determined by the relative abundance of the two MYB-type transcription factors, WER and CPC. In this model, each of these MYBs interacts with a common bHLH protein and the TTG protein. The WER-bHLH-TTG interaction generates an active transcription complex (due to the complete R2R3 MYB encoded by WER), whereas the CPC-bHLH-TTG interaction generates a complex that is unable to activate downstream gene transcription (due to the truncated MYB encoded by CPC; Fig. 1). According to this model, the normal epidermal cell pattern is caused by a relatively high concentration of WER in the N-position cells (which leads to activation of the non-hair cell differentiation pathway), and a relatively high concentration of CPC in the H-cell position (which leads to inactivation of the non-hair cell differentiation pathway and, by default, hair cell differentiation). Although this simple model accounts for the results to date, future work will be necessary to test and extend it. In particular, it will be necessary to identify the putative bHLH protein and test the hypothesis that it forms a transcriptional complex with the other components. It will also be important to define the mechanism that guides the position-dependent accumulation of the WER and CPC transcription factors.

Additional Specification Genes

Although their relationship to the WER/TTG/CPC/GL2 pathway remains to be established, additional cell specification components have recently been identified by genetic analyses. These include theroot hairless (rhl) mutants, rhl1,rhl2, and rhl3, which possess nearly hairless roots and show an early disruption in position-dependent cell differentiation (Schneider et al., 1997). Based on their mutant phenotype, this group of genes is likely to define root hair cell identity (Fig. 2). One of these,RHL1, has been cloned and encodes a novel nuclear-localized protein that does not regulate GL2 gene expression (Schneider et al., 1998). A simple possibility is that GL2acts as a negative regulator of RHL1 in a linear pathway for hair cell specification, although this is not consistent with the observed early developmental defects in the rhl1 mutants and instead suggests independent pathways of control by these components.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2.

Genetic pathway for root hair development in Arabidopsis. Each gene is positioned along the pathway at the point where it is required, according to its mutant phenotype. In most cases, the precise relationship between genes located at the same step in the pathway is unclear. The basal end of these cells (the end that would be nearest the root apex) is at the bottom of each image.

Additional mutations are also known to influence non-hair cell specification, including the recessive ectopic root hair1(ehr1), pom-pom1 (pom1), andectopic root hair3 (erh3) (Hauser et al., 1995;Schneider et al., 1997). Each of these mutants produces an excessive number of root hair cells, suggesting a role in the proper specification and/or differentiation of non-hair cells (Fig. 2). It is interesting that each of these also exhibits a significant increase in radial cell expansion, suggesting a linkage between the control of expansion and cell type differentiation. The molecular analysis of these additional genes will likely lead to refinements in the simple model outlined in Figure 1.

Features of the Cell Specification Mechanism

In addition to defining the molecular components of the pathway, the cell specification genes have been used to uncover basic features of the specification mechanism. One of the important questions is: At what point during epidermis development do cells begin to adopt a particular position-dependent fate? As mentioned above, it has long been known that epidermal cells at the H and N positions exhibit cytological differences within the meristematic region of the root, indicating that cell specification must begin at a relatively early stage of root development (Cutter, 1978; Dolan et al., 1994; Galway et al., 1994). This view has been extended by the recent use of two reporter gene fusions, a GL2::GUS fusion construct (Masucci et al., 1996) and an enhancer-trap GFP construct (line J2301; Berger et al., 1998c), each of which express their reporter gene in epidermal cells located in the N position within the meristematic region of the root. Careful examination using these sensitive reporters reveals position-dependent gene expression within, or just one cell beyond, the epidermal/lateral root cap initials, which implies that positional information may be provided (and cell fates begin to be defined) within these initial cells (Masucci et al., 1996;Berger et al., 1998a). Furthermore, the analysis of these reporters shows that the specification pattern becomes established during embryonic root development in Arabidopsis. Reporter expression is first detected at the heart stage, when the epidermal tissue becomes distinct from the ground tissue, and expression is later detected in a position-dependent pattern that mirrors the postembryonic pattern (Berger et al., 1998a). Thus, it appears that specification signals are provided during embryonic root development to establish the proper pattern of gene activities that ultimately leads to appropriate postembryonic cell type differentiation.

The establishment of position-dependent cell specification activity during embryogenesis opened the possibility that positional information is only provided to the developing epidermis during embryonic root development and not postembryonically. To examine this possibility, two sorts of experiments were conducted. In one, a detailed analysis of peculiar epidermal cell clones was performed (Berger et al., 1998a). The clones examined were generated by rare postembryonic longitudinal divisions of epidermal cells, which cause the two resulting daughter cells to occupy different positions relative to the underlying cortical cells. In these clones, the cells were observed to express marker genes and to exhibit characteristics that are appropriate for cells in their new position (Berger et al., 1998a). In a second set of experiments, specific differentiating epidermal cells were subjected to laser ablation such that neighboring epidermal cells were able to invade the available space. Regardless of the original state of the ablated cell or invading cell (differentiating hair cell or non-hair cell), the final characteristics of the invading cell were nearly always determined by its new location rather than its old (Berger et al., 1998a). Therefore, in each of these sets of experiments, cells had effectively undergone a postembryonic change in their position and, in response, exhibited a change in their developmental fate. This supports the view that positional information is provided postembryonically, not only embryonically, to ensure appropriate cell specification in the Arabidopsis root epidermis.

Root Hair Initiation

The first visible sign of root hair morphogenesis is the formation of a protrusion at the site of hair outgrowth. Genetic analyses have shown that this step is composed of two parts: The cell must first be capable of initiating an outgrowth (root hair initiation) and then the cell must generate a protrusion of the appropriate size at the proper place (bulge formation).

Although technically difficult to distinguish from the class of mutants affected in hair cell specification (discussed above), mutations that alter root hair initiation are defined by a cytologically normal pattern of immature epidermal cells (i.e. cells in the H and N positions exhibit the typical differences in cytoplasmic and vacuolar features) and an abnormal number of root hairs. All of the root hair initiation mutants identified to date are related in some manner to the action of the plant hormones auxin and ethylene. For example, theauxin resistant2 (axr2; auxin, ethylene, and abscisic acid resistant; Wilson et al., 1990) and the axr3 (altered auxin response; Leyser et al., 1996) mutants produce very few root hairs, although early cell specification appears normal. Mutations in another root hair initiation gene, CONSTITUTIVE TRIPLE RESPONSE1(CTR1), which encodes a Raf-like protein kinase proposed to negatively regulate the ethylene signal transduction pathway (Kieber et al., 1993), causes root hairs to form on epidermal cells that are normally hairless (Dolan et al., 1994). The hair initiation defect of the rhd6 mutant, which produces nearly hairless roots, can be suppressed by the inclusion of 1-aminocyclopropane-1-carboxylic acid (ACC; an ethylene precursor) or indole-3-acetic acid (an auxin) in the growth media (Masucci and Schiefelbein, 1994). Further evidence for the critical role of ethylene and/or auxin is illustrated by the ability of aminoethoxyvinyl-Gly (an ethylene biosynthesis inhibitor) or Ag⁺ (an inhibitor of ethylene perception) to block root hair formation (Masucci and Schiefelbein, 1994; Tanimoto et al., 1995) and ACC to induce some ectopic root hairs in Arabidopsis (Tanimoto et al., 1995; Masucci and Schiefelbein, 1996).

One possible explanation for the involvement of auxin and ethylene pathways in root hair initiation is that it provides a mechanism for environmental factors to influence root hair frequency. In this view, the newly formed epidermal cells in the Arabidopsis root may be tentatively defined to become hair or non-hair cells (due to the action of the early specification genes), but the ultimate proportion of hair-bearing cells may be determined by environmental factors such as humidity, ion concentrations, or shoot nutrient needs (mediated by auxin and/or ethylene pathways).

Bulge Formation

The root hair bulge is believed to result from the localized loosening and yielding of the epidermal cell wall. Mutations in theROOT HAIR DEVELOPMENT1 (RHD1) and TIP GROWTH1 (TIP1) loci lead to larger bulges than normal, suggesting a defect in the spatial control of cell wall loosening (Schiefelbein and Somerville, 1990; Ryan et al., 1998). Inrhd1 mutants, a normal hair ultimately forms from the enlarged bulge, whereas tip1 mutants produce one or more abnormally shaped hairs from each bulge.

One of the fascinating aspects of root hair morphogenesis in Arabidopsis is the predictable location of the root hair outgrowth at the basal end of the cell (the end near the root apex), which implies that the process of bulge formation is influenced by cell polarity (Fig. 2). The location of the bulge site is altered in therhd6, axr2, and ethylene resistant1(etr1) mutants, which exhibit an overall basal shift in the average bulge site, as well as the ethylene overproducer1(eto1) mutant, which has a slight apical shift (Masucci and Schiefelbein, 1994). The effect of these mutations on bulge site selection implies that hormone pathways influence the establishment or perception of cell polarity in developing epidermal cells.

Tip Growth

The major portion of the root hair is formed by the process of tip growth, a specialized type of cell expansion employed to generate tubular-shaped cells. Tip growth is characterized by a highly organized, polar distribution of cellular components geared for efficient localized secretion and cell wall synthesis. Many genes involved in root hair tip growth have been defined by mutations in Arabidopsis, perhaps because tip growth is particularly sensitive to perturbation.

The initiation of tip growth in the root hair bulge is likely to involve a complex reorganization of cytoskeletal and organelle components to generate a new growth axis. In this view, it is perhaps surprising that only one gene, RHD2, appears to be specifically required for the initiation of root hair tip growth. The hair cells of the rhd2 mutant possess bulges of the proper size and location, but there is no subsequent hair elongation (Schiefelbein and Somerville, 1990).

Several loci are required for tip growth to proceed properly in Arabidopsis root hairs, including CAN OF WORMS1(COW1), RHD3, RHD4, TIP1, and WAVY (Schiefelbein and Somerville, 1990; Marks and Esch, 1992; Schiefelbein et al., 1993; Grierson et al., 1997; Ryan et al., 1998; Galway et al., 1999). Mutations in these loci alter the morphology of the hairs, generating swollen, branched, wavy, or otherwise defective shapes, suggesting that these loci generally regulate the spatial organization or activity of tip growth processes. Double mutant analyses suggest a genetic pathway that has COW1, RHD3, RHD4, and TIP1 acting in parallel (Schiefelbein and Somerville, 1990;Grierson et al., 1997). One of the genes in this group,RHD3, has been cloned and encodes a novel GTP-binding protein that appears to influence vacuole biogenesis, illustrating the importance of this process during hair cell enlargement (Galway et al., 1997; Wang et al., 1997).

Duration of Tip Growth

What determines the duration of tip growth, and hence, the length of root hairs? The presence of significantly shorter or longer root hairs in hormone-related mutants indicates that auxin and ethylene play an important role in this process. Mutations in the AXR1,AUX1, ETR1, and EIN2 loci cause hairs to be shorter than normal, but to retain their normal morphology (Pitts et al., 1998). On the other hand, the eto1 andphyB mutants generate longer hairs than the wild type (Reed et al., 1993; Pitts et al., 1998), and treatment of seedlings with the synthetic auxin 2,4-D or ACC also generates longer root hairs (Pitts et al., 1998). Together, these findings suggest that auxin and ethylene are, in part, responsible for determining the extent of root hair tip growth under laboratory growth conditions.