Download Kwak, 2005 - U of L Class Index

REPORTS
Positional Signaling
Mediated by a Receptor-like
Kinase in Arabidopsis
Su-Hwan Kwak, Ronglai Shen, John Schiefelbein*
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 BH[ cell position) preferentially differentiate as root-hair cells, whereas
cells located outside a single cortical cell (the
BN[ 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 Ereviewed in (3–5)^.
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).
Department of Molecular, Cellular, and Developmental
Biology, University of Michigan, Ann Arbor, MI 48109–
1048, USA.
*To whom correspondence should be addressed.
E-mail: [email protected]
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 b-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 Bstripes[ (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 ecotype), 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 Bscrambled[
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).
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 posi-
www.sciencemag.org
SCIENCE
VOL 307
tion 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 Eall roots examined (N 9 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 receptorlike 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 leucinerich 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.
18 FEBRUARY 2005
1111
REPORTS
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 haircell regulators, which indicates that it acts in
an upstream signaling pathway to influence
the entire cell-fate transcriptional network.
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 È30% 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 T 0.01; scm-2, 1.18 T
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
1112
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
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 mm. (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
mm. (C) Seedling-root phenotype of wild-type and scm-1 mutant plants. Scale bar, 200 mm. (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 mm. (F) Northern blot with total RNA from shoots (S) and roots (R) of 4-dayold 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.
18 FEBRUARY 2005 VOL 307
SCIENCE
www.sciencemag.org
REPORTS
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.
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
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 mm. (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 mm. (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 mm. (D) Expression of the EGL3::GUS reporter
gene in wild-type and scm-2 seedling roots. Scale bar, 50 mm.
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 mm. (B)
Transverse sections of wild-type
and scm-2 mutant roots from
the meristematic region of 4day-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 mm.
(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 mm. (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 mm.
www.sciencemag.org
SCIENCE
VOL 307
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 receptormediated 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.
References and Notes
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
L. Dolan et al., Development 120, 2465 (1994).
M. E. Galway et al., Dev. Biol. 166, 740 (1994).
B. Scheres, Curr. Biol. 12, R804 (2002).
J. C. Larkin, M. L. Brown, J. Schiefelbein, Annu. Rev. Plant
Physiol. Plant Mol. Biol. 54, 403 (2003).
J. Schiefelbein, Curr. Opin. Plant Biol. 6, 74 (2003).
J. D. Masucci et al., Development 122, 1253 (1996).
T. Wada et al., Development 129, 5409 (2002).
T. Wada, T. Tachibana, Y. Shimura, K. Okada, Science
277, 1113 (1997).
S. H. Shiu, A. B. Bleecker, Proc. Natl. Acad. Sci. U.S.A.
98, 10763 (2001).
B. Kobe, A. V. Kajava, Curr. Opin. Struct. Biol. 11, 725
(2001).
S. K. Hanks, A. M. Quinn, Methods Enzymol. 200, 38
(1991).
J. M. Alonso et al., Science 301, 653 (2003).
F. Berger, C. Y. Hung, L. Dolan, J. Schiefelbein, Dev.
Biol. 194, 235 (1998).
M. M. Lee, J. Schiefelbein, Cell 99, 473 (1999).
M. Freeman, J. B. Gurdon, Annu. Rev. Cell Dev. Biol.
18, 515 (2002).
We thank M. M. Lee for technical assistance and
helpful discussions, J. Fuzak and C.-Y. Hung for assistance, and Y. Lin, M. Simon, C. Bernhardt, S. Clark,
and J. Li for helpful discussions. We acknowledge the
Salk Institute Genomic Analysis Laboratory and the
Arabidopsis Biological Resource Center (Columbus, OH)
for providing the scm-2 line. This work was supported by the Post-doctoral Fellowship Program of
Korea Science and Engineering Foundation (KOSEF)
(S.-H.K.) and by a grant from the NSF (IBN-0316312).
Supporting Online Material
www.sciencemag.org/cgi/content/full/1105373/DC1
Materials and Methods
Fig. S1
Tables S1 and S2
References
17 September 2004; accepted 14 December 2004
Published online 23 December 2004;
10.1126/science.1105373
Include this information when citing this paper.
18 FEBRUARY 2005
1113