STRUBBELIG defines a receptor kinase-mediated signaling pathway regulating organ development in Arabidopsis

Abstract

An open question remains as to what coordinates cell behavior during organogenesis, permitting organs to reach their appropriate size and shape. The Arabidopsis gene STRUBBELIG (SUB) defines a receptor-mediated signaling pathway in plants. SUB encodes a putative leucine-rich repeat transmembrane receptor-like kinase. The mutant sub phenotype suggests that SUB affects the formation and shape of several organs by influencing cell morphogenesis, the orientation of the division plane, and cell proliferation. Mutational analysis suggests that the kinase domain is important for SUB function. Biochemical assays using bacterially expressed fusion proteins indicate that the SUB kinase domain lacks enzymatic phosphotransfer activity. Furthermore, transgenes encoding WT and different mutant variants of SUB were tested for their ability to rescue the mutant sub phenotype. These genetic data also indicate that SUB carries a catalytically inactive kinase domain. The SUB receptor-like kinase may therefore signal in an atypical fashion.

It remains a salient challenge in biology to understand the coordination of cell behavior that underlies organogenesis and allows organs to develop to their correct size and shape. The task should be easier in plants as cell-division patterns are readily traced and plant cells do not move relative to each other (1). Plant organogenesis is a postembryonic event, and the aboveground organs typically originate at the periphery of the shoot apical meristem, located at the apex of the main shoot (2).

Signaling involving receptor-like kinases (RLKs) constitutes an essential aspect of plant cell communication and contributes to plant–pathogen interactions, hormone signaling, and development (3–5). In Arabidopsis 417 genes are predicted to encode such proteins (6). A function is known for only a handful of these loci. A major player in meristem development is the RLK CLAVATA1 (CLV1). CLV1 participates in a feedback loop maintaining the size of the stem cell population (for reviews see refs. 7 and 8). The CLV1 extracellular domain is characterized by 21 tandem copies of a leucine-rich repeat (LRR), a motif involved in protein–protein recognition (9, 10). The LRR-RLK ERECTA (ER) plays a more direct role in oganogenesis (11–13). ER is expressed in the shoot apical meristem and young lateral organs. Plants lacking WT ER function display a more compact stature, shorter inflorescence internodes, shorter pedicels, and shorter fruits with blunted tips. The main cellular basis of the er phenotype appears to be a reduction of cortex cell numbers (14). Members of the ER family of RLKs collectively promote cell proliferation and organ development (15). In corn, CRINKLY4 (CR4) is involved in cell differentiation in the leaf epidermis and specification of the aleurone layer of the endosperm (16, 17). Its extracellular domain is characterized by seven “crinkly” repeats and a domain containing three repeats also present in mammalian TNF receptor (16). ACR4, one of five Arabidopsis homologs of CR4, represents the CR4 ortholog (18). ACR4 does not seem to affect the endosperm but regulates cell proliferation patterns, cell size, and cell shape in the epidermis of several tissues. It appears to be important for the development of the integuments of ovules, the seed coat, sepal margins, and to a lesser extent for cuticle formation in the leaf (19–21).

Here, we show that STRUBBELIG (SUB) encodes a putative LRR-RLK. Our sub mutant phenotype analysis indicates that SUB affects the formation of the outer integument and the shape of organs such as carpels and petals and is necessary for the shape and height of the stem. Evidence based on a combination of in vitro kinase assays and genetics suggests that SUB carries an enzymatically inactive kinase domain and may signal in an atypical fashion.

Methods

Plant Work and Genetics. Plants were either grown as described (22) or in a greenhouse under Philips SON-T Plus 400-W fluorescent bulbs on a long day cycle (16 h light). Arabidopsis thaliana (L.) Heynh. var. Landsberg (erecta mutant) (Ler) was used as WT strain. The three sub-1–3 mutants were isolated in an ethyl methane sulfonate (EMS) mutagenesis in a Ler background (22). Previous notations were 132A3 (sub-1), 180F4 (sub-2), and 88C12 (sub-3). The sub-4 and sub-5 alleles were isolated in a screen for sub-like mutants by using EMS-mutagenized Ler seeds purchased from Lehle Seeds (Round Rock, TX). All five sub alleles were tested for complementation and backcrossed at least twice before being investigated. Initial analysis indicated that sub-1 and sub-3 mutants also displayed variably enlarged apical meristems. Further investigation revealed, however, that an additional mutation, resulting in an environmentally sensitive enlarged meristem phenotype, was present at a low frequency in the original Ler population used for mutagenesis. This mutation had apparently been coselected in the progeny of the sub-1 and sub-3 outcrosses. The mutation was genetically separated from sub-1 and sub-3 mutations, and only the clean sub alleles were used for further analysis.

Molecular Work, DNA Sequencing, and cDNA Isolation. The map-based cloning of SUB is described in Supporting Text, which is published as supporting information on the PNAS web site. For DNA and RNA work standard molecular biology techniques were used (23). Sequences were obtained by standard cycle sequencing with an Applied Biosystems 373 sequencer. All PCR products and genomic and cDNA clones were sequenced on both strands. Three SUB cDNA clones were available as ESTs isolated from the Columbia ecotype. SUBc1 and SUBc2 (GenBank accession nos. W43625 and W43624, respectively) were isolated from 3-day-old seedling hypocotyls (24), and SUBc3 (AV521707) was isolated from 2- to 6-week-old aboveground organs (25). Sequencing the ends of SUBc1 and SUBc2 indicated they were of exactly the same size. Both were slightly larger than SUBc3. Thus, only the Subc1 clone was fully sequenced. Additional 5′ and 3′ RACE experiments revealed Subc1 to be full length. The SUB sequence has been deposited under GenBank accession no. AF399923.

Computer-Based Sequence Analysis. Homology searches were done with the blast tool (26). The signal peptide sequence and the transmembrane domains were determined by using the PSORT web site (http://psort.nibb.ac.jp). Sequence alignments were done with clustalw by using default parameters (27).

In Situ Hybridization. Digoxigenin-based nonradioactive in situ hybridization experiments were carried out as described (28). To generate antisense and sense probes SUBc1 was linearized with XbaI and XhoI, respectively, in vitro-transcribed by using T7 (NEB) or T3 RNA polymerase (Roche Diagnostics) as described (29). Ler tissue was used in the experiments.

Ectopic Expression of WT and Mutant Versions of SUB. A detailed description of the procedures is given in Supporting Text. In summary, the full-length SUB ORF was amplified from SUBc1 by PCR and cloned in sense orientation behind the cauliflower mosaic virus 35S promoter of pART7 (30). Then, various mutations in the SUB cDNA were generated with the help of the QuikChange XL site-directed mutagenesis kit from Stratagene, according to the manufacturer's recommendations. This cassette was subsequently transferred to the plant transformation vector pMLBART (31). Plant transformation of sub-1 plants was performed by using the floral dip method (32).

Microscopy and Artwork. Scanning electron microscopy and artwork have been described (22, 33). Preparation and analysis of propidium iodide-stained samples for confocal laser scanning microscopy was done essentially as described (33, 34). Analysis was performed with the help of a LSM 510 microscope from Zeiss. Pictures of mature Arabidopsis flowers were taken with a Stemi SV11 dissecting scope from Zeiss, coupled to a Kodak Professional DCS 760 digital camera.

Results

The sub Mutant Phenotype. Segregation and outcrossing experiments showed all five alleles to be recessive. Despite the different predicted alterations in the putative SUB protein (see below) they all resulted in similar phenotypes. Originally, sub mutants were isolated on the basis of their defects in ovule development (22) (Fig. 1). Arabidopsis ovules undergo a highly regular and stereotypic development (33, 35, 36). We analyzed at least 300 ovules of different stages per sub allele. The ovule phenotype of sub mutants is variable, ranging from severely affected to normal and fertile ovules. As an example we describe the sub-1 ovule phenotype in more detail (Fig. 1). Seventy percent of ovules from sub-1 mutants display irregularities (n = 141 sub-1 ovules for percentage calculation). Irregularities become first visible in ovules of about late stage-2-III/early stage-2-IV (stages according to ref. 33). Upon initiation at an abaxial position, the outer integument sometimes fails to spread around the circumference of the ovule, resulting in a scoop-like incomplete outer integument (Fig. 1C). Often, however, sections of the outer integument develop. This irregular growth can lead to ovules with exposed inner integuments that are surrounded by outer integument tissue with “gaps” and resembling a “multifingered clamp” (Fig. 1D). Ovules often fail to form an embryo sac. The two-cell-layered organization of the outer integument can also be disrupted with certain areas featuring an aberrant cell layer structure, most likely the result of misoriented planes of cell divisions (Fig. 1F). Furthermore, cells at the micropylar end of outer integuments can appear more spheric as compared with the normal brick-like appearance. Ovules of sub mutants have been reported to display protrusions from the distal end of otherwise normal outer integuments (22). This phenotype, however, is caused by the interaction of sub with another, unlinked, mutation (see Methods).

Fig. 1.

Ovule development in WT and sub plants. (A-D) Scanning electron micrographs. (E and F) Specimens stained with propidium iodide. Pictures were taken by confocal laser scanning microscopy. (A, B, and E) WT. (C and D) sub-1. (F) sub-4. (A) Young WT ovule at stage-2-III. (B) WT ovule at ≈stage 4-V. (C) Stage-2-III ovule of sub-1. The ovules show no deviation from WT. (Inset) A stage-2-IV ovule. Note the partial outer integument. (D) About stage-4-V ovules of sub-1. Note the aberrant formation of the outer integument resulting in “finger-like clamps.” (E) Midoptical section through an early stage-4 WT ovule. (F) Midoptical section through a stage-4-V ovule from a sub-4 mutant. Endothelium differentiation occurred. An outer integument is mostly absent at the adaxial side. An embryo sac cannot be discerned. The arrow marks a spot in the distal abaxial outer integument where irregular cell divisions occurred and the normally two-layered organization is not maintained. Note the aberrant cell shape and the aberrant orientation of the planes of cell division. The group of cells indicated by * are part of an outer integument that is mostly located outside this plane of focus. es, embryo sac; et, endothelium; fu, funiculus; ii, inner integument; mp, micropyle; nu, nucellus; oi, outer integument. (Scale bars: 20 μm.)

We observed additional alterations in above-ground organs of sub mutants. For example, 30-day-old sub plants exhibit reduced plant height compared with WT (Fig. 6, which is published as supporting information on the PNAS web site) and show twisted stems (Fig. 2 F–I and see Fig. 5F). Twisting was not regular, in the sense of a continued helical growth, rather we observed regions exhibiting clockwise or anticlockwise twists on the same stem. These zones were separated by regions with unclear orientation of the twisting. We could also observe twisting of leaf petioles in sub individuals (50% of sub plants showed at least one leaf with an obviously twisted petiole, n = 100). The hypocotyl was apparently unaffected. We assayed apical meristem size at different stages by cell counts and/or distance measurements along the meristem diameter (Table 1, which is published as supporting information on the PNAS web site). These experiments revealed minimal differences, if any, between WT and sub mutants. In apical meristems of 10- and 30-day-old sub plants (and also in floral meristems, see below), however, we observed a less regular L2 layer and occasional periclinal divisions in this layer. To investigate mature main stem tissue we took horizontal sections cut just above the first secondary inflorescence branch point of 30-day-old sub-1 plants (Fig. 2 H–I and Table 2, which is published as supporting information on the PNAS web site). We observed a reduced number of epidermal (20%), cortex (30%), and pith cells (20%). The pith cells in particular appeared smaller. Furthermore, the outlines of epidermis, cortex, and pith cells were irregular. We also observed periclinal divisions in the L2 layer of young floral meristems of sub mutants (Fig. 7 and Table 3, which are published as supporting information on the PNAS web site), and the L2 layer and the shape of its cells seemed more irregular. At the macroscopic level we noted that ≈70% of petals from stage-13 to -15 flowers of 30-day-old sub mutants showed twisting (Fig. 2 B and C) (131 twisted petals of 185 total, 50 sub-1 flowers counted). In addition, we detected twisted carpels in all examined stage-13 to -15 flowers of sub mutants (Fig. 2E, 100 flowers per sub allele scored). Additional abnormalities included fusions of two neighboring sepals (≤10 occurrences per 100 flowers scored per sub allele), partially or completely unfused carpels (≤7/100), petals with sepal or anther characteristics (≤4/100), smaller but still obvious petals (≤7/100), and a minor apparent reduction in the number of petals and stamens (Table 4, which is published as supporting information on the PNAS web site). We did not notice any of those alterations in WT Ler flowers (n = 50).

Fig. 2.

Flowers, carpels, and stems of 30-day-old WT and sub-1 plants. (A, D, F, and H) WTLer. (B, C, E, G, and I) sub-1. (A-C) Flowers at stage 13. (D and E) Scanning electron micrographs of mature carpels. (F and G) Scanning electron micrographs of mature stems. (H and I) Horizontal stem sections (2 μm) stained with toluidine blue. (A) WT flower. (B) Flower of sub-1 mutant. Arrows indicate twisted petals. (Inset) An example of a “fused” petal with distal notch. (C) sub-1 flower. Note the twisted petal (arrowheads). (D) Mature WT carpel of a stage-14 flower. (E) Twisted carpel of a sub-1 mutant. (F) WT stem. (G) Twisted stem of a sub-1 mutant. (H) Horizontal section through a WT stem. (I) Horizontal section of a sub-1 stem. Note the altered size and shape of epidermis, cortex, and pith cells. The stem circumference is less regular, has a ragged appearance, and shows invaginations. c, cortex; e, epidermis; p, pith; v, vascular tissue. (Scale bars: 0.5 mm, A–C; 100 μm, F and G; and 200 μm, D, E, H, and I.)

Fig. 5.

Phenotype of sub-1 plants overexpressing several SUB variants. (A–C) Flowers at stage 13. (D–F) Stems of 30-day-old plants. (A and D) 35S::SUB sub-1. Stem and flower appear as in WT. (B and E) 35S::SUBK525E sub-1. Stem and flower appear normal. Plants with the genotypes 35S::SUBG506A sub-1 and 35S::SUBE539A sub-1 looked identical. (C and F) 35S::sub-4 sub-1. The plants display a sub-1 phenotype. Note the twisted stem and the missing petal. The 35S::sub-3 sub-1 plants looked identical. (Scale bars: 0.5 mm, A–C; 1 mm, D–F.)

The sub phenotype is sensitive to ecotype background (data not shown). Crossing different sub alleles isolated in the Ler background to Columbia WT plants resulted in sub plants displaying a greatly reduced mutant phenotype, with carpel twisting, weak stem twisting, and a weak ovule phenotype being the prominent aspects. The effect does not depend on the ER locus (D.C. and K.S., unpublished work). A similarly weak phenotype is also observed in plants carrying apparent sub null mutations caused by T-DNA insertions in the Col background (M.B. and K.S., data not shown). Such an effect makes it presently difficult to infer additional information about SUB function from T-DNA-mediated gene knockouts.

SUB Encodes a Putative LRR-RLK. SUB was cloned and identified by a map-based approach (Fig. 8, which is published as supporting information on the PNAS web site; gene identifier At1g11130). SUB is predicted to encode a LRR-RLK of 768 aa with a calculated molecular mass of 84.5 kDa (Fig. 3). The putative SUB protein belongs to the nine-member Arabidopsis LRR-V group of LRR-RLKs (6) and is also related to the maize LTK family of LRR-RLKs with unknown functions (37). We termed the genes encoding the other members of the LRRV class STRUBBELIG receptor family 1–8 (SRF1–8) (www.arabidopsis.org/info/genefamily/lrrv.html). Sequence analysis predicts that SUB contains a signal peptide of 24 aa, an amino-terminal domain shared between the LRR-V members, which we call SUB domain, six LRRs, a proline-rich region, a transmembrane domain, a juxta-membrane domain, and a carboxyl-terminal kinase domain. Thus, SUB is most likely involved in transmitting a signal across a membrane. The sub-1, sub-2, and sub-5 alleles are predicted to cause a complete lack of SUB function. All three mutations apparently lead to a shorter putative SUB protein, carrying part of the extracellular domain but lacking the transmembrane and intracellular domains. Thus, in all three mutants it is to be expected that no SUB-based signaling is mediated across a membrane. The sub-3 mutation causes a valine-to-methionine substitution in the SUB domain. Presently, the negative consequences of this alteration cannot be easily inferred from sequence comparisons. Either valine or isoleucine is found at this position in other members of the LRRV family. The sub-4 allele carries an arginine-to-cysteine substitution (R599C) at a conserved position in kinase subdomain VIa (Fig. 9, which is published as supporting information on the PNAS web site). The bri1–8 and bri1–108 mutations affect the corresponding position in the kinase domain of BRASSINOSTEROID INSENSITIVE 1 (BRI1) (38, 39). In addition, domain VIa is altered in er-1, er-101, and er-102 (I750K), leading to a strong erecta phenotype (12). Taken together, the sub-3 and sub-4 mutations indicate that the putative extracellular SUB domain and the cytoplasmic kinase domain, respectively, are important for SUB function.

Fig. 3.

The predicted SUB protein. (A) The N-terminal domain. The putative signal sequence is underlined. (B) The SUB domain. The valine, which is converted to a methionine in the sub-3 mutant, is underlined. The amino acids that are strictly conserved between the LRR-V family members are highlighted in bold. (C) The six LRRs. The conserved amino acids typical for the LRRs are shown in bold. The underlined residue indicates the position of the splicing mutation in sub-1. (D) The proline-rich region. The amino acids replaced by a stop in sub-5 and sub-2, respectively, are underlined. (E) The transmembrane domain. (F) The juxtamembrane domain. (G) The kinase domain. The residues altered in the genetic kinase activity experiments are marked in bold. The unusual amino acids in domain VIa are shown in bold italics. The residue affected by the sub-4 mutation is underlined.

The Temporal and Spatial Expression Pattern of SUB. Northern analysis (Fig. 10, which is published as supporting information on the PNAS web site), RT-PCR analysis (data not shown), and consultation of the AtGenExpress data set (81) indicated that SUB transcripts are present in a broad pattern that includes tissues that show altered morphology in sub mutants, such as leaves, stems, inflorescences, and flowers. Our in situ hybridization experiments indicate a complex SUB expression pattern. A summary of the pattern during floral development is given in Fig. 4. Further description is provided in Figs. 11 and 12, which are published as supporting information on the PNAS web site.

Fig. 4.

SUB mRNA expression pattern in the inflorescence meristem and during floral development. Longitudinal tissue sections (10 μm) are shown. (A) Inflorescence meristem and flowers at stages 2 and 3. (B) Stage-4 flower. (C) Flowers at stages 6 and 7. (D) Flower at stage 13. ca, carpel; se, sepal; st, stamen; ov, ovule. (Scale bars: 20 μm.)

The SUB Protein May Represent an Atypical RLK. The kinase domain of SUB has the hallmarks of a typical protein kinase (40). Nevertheless, there are two notable alterations within the catalytic loop of the WT SUB kinase domain (Fig. 3). SUB carries an asparagine at a position (N-625) where functional protein kinases usually contain an aspartate. In addition, SUB features a lysine at position 630. In contrast, plant RLKs with experimentally detectable kinase activity feature an asparagine at this position (Fig. 9). As both residues are important for the catalytic mechanism (41, 42), SUB was tested for kinase activity. First, SUB kinase activity was assayed by using bacterially expressed fusion proteins (Fig. 13, which is published as supporting information on the PNAS web site). We were unable to detect autophosphorylation or transphosphorylation of general substrates such as myelin basic protein in these experiments. In a second approach we resorted to genetics. We first generated sub-1 plants, carrying a transgene expressing the WT SUB cDNA under the control of the 35S promoter of cauliflower mosaic virus (43). 35S::SUB sub-1 plants showed a WT phenotype, indicating that the transgene could provide sufficient WT SUB gene function (Fig. 5). Ler plants carrying the 35S::SUB construct also did not display any noticeable alterations from WT. We then repeated the experiment by using three different mutant SUB cDNA versions. The introduced mutations are predicted to affect ATP binding by SUB and should not have gross negative effects on its potential substrate binding capacity (41, 42, 44). In SUB_G506A the conserved glycine at position 506 is converted to alanine. This mutation affects the P loop, which is involved in orienting the ATP through interactions with its phosphates. The equivalent residue in cAMP-dependent protein kinase A has been shown to be the most critical of the three P loop glycines (45). In SUB_K525E the conserved lysine in kinase domain II is altered to glutamic acid. In general, mutating this position leads to loss of kinase activity (46, 47) and has served as a negative control in a number of biochemical kinase assays involving plant RLKs (18, 19, 21, 48–51). In SUB_E539A a conserved glutamic acid is changed to alanine. Transgenic sub-1 plants carrying any of the three constructs showed a WT phenotype (Fig. 5), indicating that catalytic activity is not essential for in vivo SUB function. As a control we also introduced into sub-1 plants cDNA constructs carrying the same mutations as sub-3 and sub-4. The corresponding transgenic sub-1 plants still displayed a sub mutant phenotype (Fig. 5 C and F) although we could observe expression of the transgene (see Supporting Text). In addition, a 35S::SRF5 sub-1 plant exhibited a sub mutant phenotype (At1g78980; Banu Eyueboglu and K.S., unpublished results). The biochemical and genetic data together suggest that SUB carries a catalytically inactive kinase domain.

Discussion

The molecular structure of SUB and the mutant sub phenotype indicate that SUB defines a novel signaling pathway, which is repeatedly required for organogenesis in Arabidopsis. The five sub mutations, predicted to affect different domains of SUB protein and including, for example, the likely null-allele sub-1, result in very similar phenotypes. This result indicates that they perturb SUB function to the same degree. The variability of the sub phenotype and its susceptibility to ecotype effects indicate that processes involving SUB signaling are regulated by partially redundant activities. As SUB is a member of the LRRV gene family (6) redundant activity could be provided by other members of this family. Within the family, however, SUB occupies a solitary position and at the protein level, identity between SUB and other members ranges between 31% and 41%. Redundancy could also be achieved by the action of different genetic pathways as it occurs during polarity establishment in carpel development (52).

At the organ level, our analysis of the sub phenotype indicates that SUB signaling is required for the formation of the outer integument, the correct shape of the gynoecium and the petals, and the shape and height of the stem. What is the cellular function of SUB? The observed SUB expression pattern correlates with regions occupied by actively dividing cells. SUB seems to be required in some tissues for the regulation of cell shape and the orientation of the mitotic division plane. For example, apical and floral meristems of sub plants exhibit cells of altered shape, and at least cells in the L2 layer of apical and floral meristems have a higher than normal chance of undergoing cell division with a misoriented division plane. In addition, ovules of sub mutants feature cells that show aberrant shape and size or had apparently passed through mitosis with disoriented division planes. SUB may have a role in the control of cell proliferation in certain tissues. On the one hand, we could not measure a distinct alteration of the L1 cell number in apical or floral meristems in sub mutants. On the other hand, however, and in addition to effects on their size and shape, the absence of WT SUB activity eventually results in a reduced number of epidermis, cortex, and pith cells in the stem. We think it is possible that the reduced number of cells in stem tissue reflects a positive role for SUB in the regulation of those cell numbers as we could detect SUB expression in stem tissue by RT-PCR. Unfortunately, however, we were unable to visualize SUB expression in regions basal to the apex by in situ hybridization. Incomplete early development of the outer integument and the presence of gaps in this structure could also be interpreted, in a broad sense, as SUB acting as a positive regulator of cell division.

What is the possible common theme underlying the range of cellular defects observed in sub mutants and how does the cellular function of SUB relate to the twisting of organs observed at the macroscopical level? We presently do not know the answer but we speculate that SUB may affect cytoskeleton function. The cytoskeleton is necessary for the regulation of cell shape, cell size, and mitosis (53, 54). Interestingly, particular mutations in α-tubulin genes and defects in a number of genes encoding microtubule-interacting proteins result in alterations in cortical microtubule organization and helical growth of various plant organs (55–60).

SUB Is Likely to Encode an Atypical RLK. The sub-4 mutation indicates that the kinase domain is important for SUB function. Nevertheless, our data suggest that the WT SUB protein kinase domain lacks enzymatic phosphotransfer activity.

A number of so-called dead or atypical receptor kinases that carry a kinase domain but lack enzymatic activity are described in the literature (61). Examples include CCK-4 (62), the Ryk class of receptor tyrosine kinases (63–68), maize atypical receptor kinase (MARK) from corn (69), Arabidopsis ATCRR1 and ATCRR2 (18, 19), and likely also transmembrane kinase-like (TMKL1) (70). In most of these examples, the alterations affect the P loop, the DFG motif at the beginning of the activation loop (kinase domain VII), or both motifs (Fig. 9). ATCRR1 and ATCRR2 display a deletion of the activation loop (18, 19). In a number of cases, and with the exception of human Ryk, mutating the different residues back to consensus-like amino acids did not create active kinases, indicating that other alterations are present in atypical kinase domains (65, 69, 71).

What could be the basis of the likely lacking kinase activity of SUB? Sequence comparisons point to the catalytic loop. As a rule active kinases feature a conserved aspartate and arginine in this domain (40). In SUB an arginine substitutes for this aspartate [N625, the D166 in cAMP-dependent protein kinase A and important for catalysis (41, 72–74)] and a lysine for the arginine (K630). All other members of the LRRV family also carry an arginine at the position equivalent to N625 of SUB except for SRF2 (At5g06820), which features an aspartate at this position (data not shown). The K630 residue is unique to SUB. Recent evidence indicates that a N625-related change may not be an essential alteration in some RLKs (75, 76). The K630 may be responsible as the putative SUB protein seems otherwise fine or critical alterations could be located somewhere else in the protein, not easily detectable by sequence comparisons. Alternatively, SUB does have kinase activity but its mode of function has a level of complexity not accounted for in our experiments. One future strategy to address the issue could involve an x-ray-based structural analysis of the SUB kinase domain by using cocrystals of SUB and a partner naturally interacting with the SUB kinase domain.

What is the mechanistic basis of signaling through atypical RLKs? Little is known in this regard but atypical RLKs may function via regulated protein–protein interactions (61). The SUB kinase domain could still interact with downstream effectors requiring the typical 3D configuration of a kinase domain. This is the case for animal atypical RLKs such as members of the Ryk family (64–66, 77–79). In plants MARK represents a salient example of an atypical RLK although its developmental function remains to be elucidated (69). MARK activity could not be detected; however, in a COS-7 cell expression system MARK was able to interact with the functional MAP4K-class kinase MARK-interacting kinase (MIK). This interaction led to an increase of the intrinsic MIK activity. Thus, it seems that MARK signals through a mitogen-activated protein kinase pathway via its interaction with MIK (69). A different scenario may apply for ATCRR2. The developmental role of ATCRR2 is also not clear but it can be phosphorylated by ACR4 in vitro (18). ATCRR2 may therefore form a heterodimer with the kinase-active RLK ACR4 and thus could be involved in ACR4-mediated signaling.

To our knowledge, SUB appears to be the sole representative of the atypical receptor kinase class in plants with an established developmental role. Because little is known about the function of such receptors, it will be very interesting to unravel the different aspects of the SUB signaling pathway in the future.