Positional Signaling Mediated by a Receptor-like Kinase in Arabidopsis

See allHide authors and affiliations

Science  18 Feb 2005:
Vol. 307, Issue 5712, pp. 1111-1113
DOI: 10.1126/science.1105373


The position-dependent specification of root epidermal cells in Arabidopsis provides an elegant paradigm for cell patterning during development. Here, we describe a new gene, SCRAMBLED (SCM), required for cells to appropriately interpret their location within the developing root epidermis. SCM encodes a receptor-like kinase protein with a predicted extracellular domain of six leucine-rich repeats and an intracellular serine-threonine kinase domain. SCM regulates the expression of the GLABRA2, CAPRICE, WEREWOLF, and ENHANCER OF GLABRA3 transcription factor genes that define the cell fates. Further, the SCM gene is expressed throughout the developing root. Therefore, SCM likely enables developing epidermal cells to detect positional cues and establish an appropriate cell-type pattern.

A fundamental feature of development in multicellular organisms is the specification of distinct cell types in appropriate patterns. In Arabidopsis, the patterning of the root-hair cells and non–hair cells in the root epidermis is determined by a position-dependent mechanism (1, 2). Developing epidermal cells overlying two cortical cells (the “H” cell position) preferentially differentiate as root-hair cells, whereas cells located outside a single cortical cell (the “N” position) adopt the non–hair cell fate. Both symplastic (e.g., direct cell-cell movement) and apoplastic (e.g., receptor-mediated cell-cell interaction) explanations have been proposed for the putative signaling mechanism (3, 4).

A regulatory network including at least eight putative transcription factors is required to specify the root epidermal cell types through the general mechanism of lateral inhibition with feedback [reviewed in (35)]. WEREWOLF (WER, a MYB transcription factor), TRANSPARENT TESTA GLABRA (TTG, a WD40-repeat protein), and GLABRA3 and ENHANCER OF GLABRA3 (GL3 and EGL3, related bHLH transcription factors) appear to act in a central transcriptional complex in the N cells to promote the non–hair cell fate and mediate lateral inhibition. One of the likely targets of this complex is GLABRA2 (GL2), which encodes a homeodomain transcription factor required for non–hair cell differentiation (6). TTG/GL3/EGL3/WER also positively regulate the CAPRICE (CPC) gene, which encodes a small one-repeat MYB protein that may mediate lateral inhibition by moving to the adjacent H cells and inhibiting the action of the WER MYB (7, 8).

Although much of the transcriptional machinery used to define the two cell types is known, no information is currently available concerning the molecular basis of the positional signaling mechanism that regulates this transcriptional network. To identify mutants defective in position-dependent patterning of the root epidermal cell types, we conducted a new genetic screen based on visual examination of cell-type–specific reporter expression in mutagenized seedling roots. We reasoned that a defect in the production or perception of positional cues would not necessarily lead to a change in hair density, and therefore might require direct examination of cell pattern to detect. The GL2-promoter β-glucuronidase (GUS) fusion (GL2::GUS) was chosen as the reporter because it is expressed in the developing non–hair cells, which generates roots with easily visible “stripes” (or files) of GUS activity (6) (Fig. 1, A and B). An ethylmethane sulfonate–mutagenized population of the GL2::GUS line was generated (in the WS eco-type), and seedlings were screened for changes in GUS activity pattern within M2 families, as a result of the destructive nature of the GUS assay. One family segregated a recessive mutation that caused a patchy or “scrambled” distribution of GL2::GUS-expressing cells in the root epidermis (Fig. 1, A and B). Because the mutation causes a disorganized GL2::GUS pattern, the novel gene affected in this line was named SCRAMBLED (SCM).

Fig. 1.

Identification and cloning of the SCM gene. (A) Expression of the GL2::GUS reporter gene during root development. Four-day-old seedlings were histochemically assayed for GUS activity. The scm-1 mutation disturbs the normal file-specific GL2 expression pattern. Root hairs are not visible in this developmental region; they form on cells above the field of view in these photos. Scale bar, 50 μm. (B) Spatial expression of the GL2::GUS reporter gene in transverse root sections. Plastic sections were taken from the meristematic region of wild-type and scm-1 seedling roots harboring the GL2::GUS transgene. E, epidermis; C, cortex; N, endodermis; S, stele. Scale bar, 25 μm. (C) Seedling-root phenotype of wild-type and scm-1 mutant plants. Scale bar, 200 μm. (D) The intron and exon organization of the SCM locus. The positions of the scm-1 and scm-2 mutations and the 8.4-kb complementing ApaI-SalI fragment are shown. (E) Complementation of the scm mutant phenotype. Expression of the GL2::GUS reporter gene during root development in the scm-1 and scm-2 mutants and in transgenic lines harboring the 8.4-kb SCM genomic fragment (gSCM). Scale bar, 50 μm. (F) Northern blot with total RNA from shoots (S) and roots (R) of 4-day-old wild type, scm-1, and scm-2 mutants hybridized with a SCM probe. The rRNA levels were used as loading controls. (G) Domain organization of the predicted SCM LRR-RLK protein.

Despite the strong effect of scm-1 on the pattern of GL2::GUS expression, scm-1 did not show a detectable effect on root-hair density (Fig. 1C). To determine whether the position-dependent patterning of root epidermal cell types is altered, we analyzed the arrangement of mature root-hair and non–hair cell types in scm-1. In wild-type roots, most cells (92%) in the H position are root-hair cells, and nearly all cells (99%) in the N position are non–hair cells (table S1). In the scm-1 roots, only 66% of cells in the H position are root-hair cells, and 79% of cells in the N position are non–hair cells (table S1). This shows that a functional SCM gene is required for proper position-dependent cell-type patterning.

Because epidermal cell fate is correlated with the cortical cell arrangement, we tested the possibility that the defective cell patterning in the scm mutant is due to an abnormality in root structure. However, examination of wild-type and scm mutant roots revealed no significant differences in the cortical cell number [all roots examined (N > 8 for each line) possessed eight cortical cell files] or the organization of the root tissues (Fig. 1B), which indicates that the scm mutations do not significantly affect root structure.

We cloned the SCM gene by a genetic map–based strategy. Several genes within a 55-kb interval near the top of Arabidopsis chromosome 1 were sequenced, and we identified a single base substitution in scm-1 that generates a premature stop codon at amino acid 324 in the putative coding region of the At1g11130 gene (Fig. 1D). This gene encodes a predicted leucine-rich repeat receptor-like protein kinase (LRR-RLK). No biological function has been reported or assigned to this gene/protein. DNA fragments from this gene region were introduced into scm-1 GL2::GUS mutant plants, and an 8.4-kb fragment containing only At1g11130 sequences was found to restore the normal GL2::GUS expression pattern (Fig. 1, D and E).

The deduced SCM protein is 768 amino acids and possesses all the structural features of a typical LRR-RLK (Fig. 1G), of which there are more than 200 predicted in Arabidopsis (9). The N terminus contains a hydrophobic sequence predicted to act as a signal sequence for secretion, followed by the putative extracellular domain, which includes six tandem copies of a 24-residue leucine-rich repeat (LRR; residues 85 to 231). LRRs are found in diverse eukaryotic proteins and typically participate in protein-protein interactions (e.g., ligand binding and/or receptor dimerization) (10). A single predicted transmembrane domain is near the center of SCM (amino acids 342 to 362), and the C terminus contains a putative intracellular kinase domain that possesses most of the signature sequences for a serine-threonine protein kinase (amino acids 496 to 764) (11) (fig. S1).

A second scm mutant (scm-2) was obtained from the Salk Institute Genomic Analysis Laboratory collection (12). We verified that scm-2 has a transferred DNA (T-DNA) insertion within the third intron of the SCM gene (Fig. 1D). This scm-2 mutant has an altered pattern of epidermal cell types similar to scm-1 (table S1), which provides additional support for the important role of the SCM/At1g11130 gene in epidermal patterning.

Northern blot analysis showed that SCM RNA is present in the roots and shoots of wild-type seedlings, although no obvious shoot phenotype was observed in the scm seedlings (Fig. 1F). The SCM RNA is substantially reduced in scm-1 and undetectable in scm-2 (Fig. 1F). Because scm-2 affects the SCM transcript at an earlier point than scm-1 (Fig. 1D) and lacks detectable SCM mRNA (Fig. 1F), it was used as the scm reference mutant for our subsequent analyses.

To further define the role of the SCM receptor-like kinase, we analyzed the effect of scm-2 on the expression of several transcription-factor genes in the specification pathway. In independent experiments, the non–hair-cell–expressing GL2::GUS, CPC::GUS, and WER::GFP promoter-reporter transgenes were introduced into the scm-2 by crossing. In each line, we found a disorganized pattern of reporter gene expression (Fig. 1E and Fig. 2, A to C). We quantified these effects and discovered that, in each line, scm-2 significantly reduces the frequency of H cells that lack reporter expression and the frequency of N cells that express each reporter (table S2). To determine whether SCM affects the expression pattern of GL2, CPC, and WER independently or coordinately, we generated a scm-2 line that possessed both the GL2::GUS and WER::GFP, and we assayed both reporters in the same roots. We discovered a clear correlation between GUS-expressing cells and GFP-expressing cells in this line (Fig. 2C), implying that the SCM receptor affects the expression of both genes in a similar way in the same cells. We also examined the role of SCM on a hair-cell–expressing transcription-factor reporter, EGL3::GUS, and we observed a similar disorganized pattern of GUS expression in the scm-2 EGL3::GUS root epidermis (Fig. 2D). Together, these results show that the SCM receptor is required for position-dependent transcription of both non–hair-cell and hair-cell regulators, which indicates that it acts in an upstream signaling pathway to influence the entire cell-fate transcriptional network.

Fig. 2.

SCM is required for position-dependent gene expression. (A) Expression of the CPC::GUS reporter gene in wild-type and scm-2 seedling roots. Scale bar, 50 μm. (B) Expression of the WER::GFP reporter gene in wild-type and scm-2 seedling roots. Propidium iodide (red) was included to visualize cell boundaries. The upper inset shows the relative positions of the underlying cortical cells. Scale bar, 50 μm. (C) Expression of the GL2::GUS and WER::GFP reporter genes in a single scm-2 seedling root. The root was first examined for GFP expression and then was assayed for GUS activity. Asterisks mark ectopic reporter-expressing cells and ectopic reporter-nonexpressing cells. Scale bar, 50 μm. (D) Expression of the EGL3::GUS reporter gene in wild-type and scm-2 seedling roots. Scale bar, 50 μm.

To further examine the importance of the SCM receptor during root epidermis development, we evaluated the relative cell-division rate in the H and N positions of the scm-2 mutant. In wild-type roots, cells in the H position have ∼30% greater cell-division rate than the developing N cells (13). We discovered that the scm-2 mutant has a significant reduction in the relative division rate in the H and N positions (relative H/N cell-division rate: Col wild-type, 1.35 ± 0.01; scm-2, 1.18 ± 0.02). Together with our other findings, this indicates that SCM acts at an early stage of root epidermal development and affects all known cell characteristics.

The SCM receptor may mediate patterning by enabling epidermal cells to detect an asymmetrically distributed positional cue or by being differentially expressed in the H or N cell positions. To address this issue, we analyzed SCM gene expression during root epidermis development, using in situ RNA hybridization and promoter-reporter gene fusion experiments. In the wild-type roots, we observed a strong antisense SCM hybridization signal throughout the developing root, especially within and near the meristem initials, but not in the root cap (Fig. 3A). Transverse sections showed that, in the epidermis, the SCM antisense probe hybridized to cells in both the H and the N cell position (Fig. 3B). For the reporter-gene experiments, we fused the 5′ and 3′ SCM genomic sequences from the complementing SCM fragment to the coding region of the GUS or GFP reporters. Similar to the in situ hybridization results, the SCM::GUS and SCM::GFP reporters were expressed throughout the developing root tissues, including epidermal cells, in the meristematic region (Fig. 3, C and D). Together, these results show that SCM gene expression and mRNA accumulation occurs at the time when position signaling is acting, and it is not confined to a particular type of epidermal cell.

Fig. 3.

Expression of SCM gene in the developing Arabidopsis root. (A) Whole-mount in situ RNA hybridization of SCM mRNA in wild-type and scm-2 mutant (negative control) seedling root tips. Roots were exposed to a digoxigenin-labeled SCM antisense RNA probe. In a separate experiment, a SCM sense RNA probe was used and no signal was detected on wild-type roots (data not shown). S, stele; N, endodermis; C, cortex; E, epidermis. Scale bar, 25 μm. (B) Transverse sections of wild-type and scm-2 mutant roots from the meristematic region of 4-day-old seedlings exposed to a digoxigenin-labeled antisense RNA probe. S, stele; N, endodermis; C, cortex; E, epidermis; L, lateral root cap. Scale bar, 25 μm. (C) Expression of SCM::GUS promoter-reporter gene fusion in the wild-type seedling root tip. Left panel contains longitudinal view; right panels contain transverse sections at two different developmental stages showing GUS activity in epidermal layer. Scale bar, 25 μm. (D) Expression of SCM::GFP promoter-reporter gene fusion in the wild-type seedling root tip. Upper panel contains optical section from median longitudinal view; lower panel contains optical section of the overlying epidermal layer alone, showing GFP accumulation in the epidermal cells. Scale bar, 25 μm.

These results expand our understanding of the establishment of the position-dependent pattern of root epidermal cell types. It is likely that, at an early stage in epidermal development, positional cues from underlying root cells are distributed in a nonuniform manner to differentially activate the SCM receptor in the N and H cells and that, after signal transduction, this leads to a difference in the expression of the cell-fate transcription factors in these cells. Because the WER MYB controls the expression of all other known transcription factors, including GL2 and CPC (4, 14), the WER gene is an attractive candidate for the primary target of this SCM signaling pathway. The SCM-mediated difference in transcription-factor gene expression is likely to be amplified and stabilized by the known lateral-feedback loops (4) and to lead to the establishment of distinct gene expression patterns and cell fates.

Together with previous findings, this work shows that two different cell signaling mechanisms are employed during root epidermis development. One of these (described here) uses a putative plasma membrane–bound receptor kinase, SCM, to enable epidermal cells to perceive extracellular positional cues, and may be analogous to the kind of receptor-mediated mechanisms that help metazoan cells make their developmental decisions (15). The other previously identified signaling mechanism acts later and uses the mobile transcription factor, CPC, which likely moves directly from cell to cell by means of plasmodesmata to mediate lateral inhibition (7). Thus, a distinct combination of apoplastic and symplastic signaling appears to specify the root epidermis cell fates in Arabidopsis.

Supporting Online Material

Materials and Methods

Fig. S1

Tables S1 and S2


References and Notes

View Abstract

Stay Connected to Science

Navigate This Article