Review

Plant Receptor-Like Kinase Gene Family: Diversity, Function, and Signaling

See allHide authors and affiliations

Science's STKE  18 Dec 2001:
Vol. 2001, Issue 113, pp. re22
DOI: 10.1126/stke.2001.113.re22

Abstract

Plant receptor-like kinases (RLKs) are transmembrane proteins with putative amino-terminal extracellular domains and carboxyl-terminal intracellular kinase domains, with striking resemblance in domain organization to the animal receptor tyrosine kinases such as epidermal growth factor receptor. The recently sequenced Arabidopsis genome contains more than 600 RLK homologs, representing nearly 2.5% of the annotated protein-coding genes in Arabidopsis. Although only a handful of these genes have known functions and fewer still have identified ligands or downstream targets, the studies of several RLKs such as CLAVATA1, Brassinosteroid Insensitive 1, Flagellin Insensitive 2, and S-locus receptor kinase provide much-needed information on the functions mediated by members of this large gene family. RLKs control a wide range of processes, including development, disease resistance, hormone perception, and self-incompatibility. Combined with the expression studies and biochemical analysis of other RLKs, more details of RLK function and signaling are emerging.

Introduction

Signal perception through cell-surface receptors is a common feature among living organisms. In plants, several different types of cell-surface receptors perceive diverse signals and stimuli from the environment. Some of these receptors, such as the receptor histidine kinases in plants that mediate responses to the hormones ethylene (1) and cytokinin (2), have not yet been found in metazoan genomes. These represent an evolutionary novelty that distinguishes plants from animals (3).

At the same time, the presence of a large number of transmembrane kinases resembling animal growth factor receptor tyrosine kinases represents one of the similarities between plants and animals in signal transduction (Fig. 1) (4). These transmembrane kinases are collectively named the receptor-like kinases (RLKs). They are defined by the presence of a signal sequence, an amino-terminal domain with a transmembrane region, and a carboxyl-terminal kinase domain (Fig. 2) (5, 6). Since the discovery of the first RLK a decade ago, substantial efforts have been devoted to a few specific receptor-kinase genes, including brassinosteroid insensitive 1 (BRI1) (7), CLAVATA1 (8), Pto (9), S-locus receptor kinase (SRK) (10), and Xa21 (11). RLKs participate in a diverse range of processes, including self-versus-nonself recognition, disease resistance, regulation of development, and hormone perception. In addition, several interacting components upstream or downstream to the receptor protein kinases have also been identified.

Fig. 1.

Common themes in plant and animal receptor protein kinases.

Fig. 2.

Representative members of the receptor-like kinase (RLK) family. On the basis of the presence or absence of extracellular domains, members of this gene family are categorized as RLKs or RLCKs (receptor-like cytoplasmic kinases). The gray line represents the plasma membrane except for TAK, which is localized to the thylakoid membrane. The domains above the gray line are the putative extracellular domains with signal sequences. The area below the gray line represents the cytoplasmic side. PERK is the only representative with an extracellular domain but no signal sequence. The signal peptides are presumably absent in mature proteins and are displayed to demonstrate their presence in the RLKs. These representative RLKs are numbered as follows: 1, A5g12000; 2, NAK; 3, TAK; 4, PERK; 5, RKF3; 6, CrRLK1 7, LRK10; 8, At5g56890; 9, Xa21; 10, CLAVATA1; 11, BRI1; 12, TMKL1; 13, At1g53340; 14, TMK1; 15, LRRPK; 16, SERK; 17, At1g52310; 18, At3g26700; 19, WAK1; 20, PR5K; 21, LecRK1; 22, RKF2; 23, SRK; 24, CRINKLY4 and 25, CHRK1. TM, transmembrane region; DUF, domain of unknown function; EGF, epidermal growth factor; B-lectin, agglutinin; C-lectin, C-type lectin; L-lectin, legume lectin; LRR, leucine-rich repeat; LysM, lysin motif; PAN, plasminogen/apple/nematode protein domain; TNFR, tumor necrosis factor receptor.

During the time that these important discoveries were made in understanding the functions of RLKs, the genome of Arabidopsis thaliana was sequenced. RLKs represent one of the largest gene families in the Arabidopsis genome, with at least 610 members (12). Just as animal cells have receptor tyrosine kinases and nonreceptor tyrosine kinases (13), plant RLKs include receptor and nonreceptor protein kinases and are collectively named "the RLK family" (Fig. 2). The possibility that plants use hundreds of receptors as entry points into signaling pathways has broad implications for development and environmental responses. The challenge ahead is daunting. Fewer than 2% of the total RLKs identified have known functions, and much less is known about their signaling components or ligands.

In this review, we summarize the recent advances in studies of the RLK multigene family. Based on the completed Arabidopsis genome and what is known about RLKs from other plant species, we have outlined the diversity of the plant RLK family and the sequence motifs found in their extracellular domains. This information is compiled to facilitate discussion on the potential functions of RLKs. Finally, the results from studies on several specific RLKs are summarized to provide an overview of the known signaling pathways and signaling mechanisms.

Diversity of the RLK Gene Family

Definition and Size of the RLK Family

Although RLKs were originally defined according to their domain organization as transmembrane receptors, it should be noted that, in our definition, the RLK family is delineated on the basis of the phylogenetic relationships of the kinase domains among members, regardless of the presence of an extracellular domain (12). In this review, RLK refers to the receptor protein kinase members with an extracellular domain, and RLCK (receptor-like cytoplasmic kinase) refers to the cytoplasmic kinase members of this family.

RLKs and RLCKs have been cloned from more than 20 flowering plant species. To date, full-length RLK genes have not been cloned from lower vascular plants or conifers, but expressed sequence tags (ESTs) of RLK homologs from these plant species are present in the GenBank nucleotide database. Preliminary analysis of the Arabidopsis genome indicated the presence of more than 300 RLKs (3, 4). On the basis of a molecular phylogenetic analysis of the RLK kinase domain sequences (12), more than 610 members were found to belong to this gene family. They can be classified into 44 subfamilies (Table 1). The RLK gene family is among the largest in Arabidopsis and represents 2.5% of the organism's protein-coding genes.

Fig. 3.

Signal transduction pathways of selected RLKs. (A). CLAVATA1 (CLV1)-mediated signaling pathway. CLV1 is an RLK with 20 LRRs (green). It is activated by the ligand CLAVATA3 (CLV3), with the involvement of CLAVATA2 (CLV2), for which the exact role remains unknown. The intracellular kinase domain (blue) may interact with KAPP, a negative regulator, and Rho GTPase (ROP). The activation of CLV1 inhibits the function of the transcriptional regulator WUSCHEL (WUS) and POLTERGEIST (POL) and negatively controls meristem proliferation. (B). SRK-mediated signaling pathway. SRK is activated by its ligand SCR/SP11. SLG, a soluble protein resembling the SRK extracellular domain may also participate in ligand binding. SRK activation may lead to the activation of ARC1, subsequently inhibiting the action of a potential component, aquaporin, and finally leading to the self-incompatibility response. The kinase activity of SRK is negatively regulated by THL1, THL2, and KAPP. (C) Pto-mediated signaling pathway. Bacteria-derived avrPto binds directly to the kinase domain of Pto and activates the kinase activity of Pto. Prf is also involved, but its exact location in the pathway is unknown. The activated Pto activates two separate pathways, which lead to different aspects of disease resistance. Pti1, also an RLK member, is phosphorylated by Pto. Pti4-5-6 are EREBP-type transcription factors able to activate pathogenesis-related protein genes. (D) FLS2 signaling pathway. Flagellin is a candidate for the elusive FLS2 ligand. The identity of FLS1 remains unknown, but may be one of the RLPs in close proximity to FLS2 gene on the chromosome. AtMPK6 is found to modulate flagellin sensitivity and may also be in the pathway.

Table 1.

The classification of RLKs. The number of subfamilies is the number of phylogenetically distinct groups in a given RLK type. The sequence motifs are based on the annotation of SMART (123) and Pfam (124) databases. The InterPro database (column 3) can be accessed from the URL at (125). The number of genes is based on the Arabidopsis genes only. NA, not applicable; ND, not determined. *The assignment of this subfamily is based on the similarity of kinase domains only.

RLK family members vary greatly in their domain organization and the sequence identity of their extracellular domains (Fig. 2). Among them, 75% have a receptor configuration where signal sequences and transmembrane regions are both present; a hallmark for type I membrane proteins. In contrast, proline extensin receptor kinases (PERKs) from Brassica and Arabidopsis have only a transmembrane region (14), and the thylakoid-associated kinase 1 (TAK1) has a single hydrophobic region near the beginning of the protein (15). The remaining members of the RLK family are cytoplasmic kinases. As mentioned earlier, the relationships between plant RLKs and RLCKs are similar to those between animal receptor tyrosine kinases and nonreceptor tyrosine kinases (13). This relationship in domain organization may reflect the independent origin of receptor protein kinases in plants and animals.

Motifs in the extracellular domains

Diverse sequence motifs are present in the putative extracellular domains of RLKs (Table 1), and several motifs are found in both plant and animal proteins, whereas others seem to be unique to plant RLKs. Some motifs are implicated in protein-protein interactions, whereas other motifs are implicated in binding to various carbohydrate substrates, including plant and microbial cell-wall components or glycoproteins of various origins. One interesting exception is BRI1, which may bind directly to a steroid hormone.

In Arabidopsis, 235 of the RLKs contain 1 to 32 leucine-rich repeats (LRRs) in the extracellular domain. LRRs often participate in protein-protein interactions (16). The crystal structure of the porcine ribonuclease inhibitor indicates that each LRR repeat forms a β-α structure, and a number of LRRs form a horseshoe structure, providing a docking site for the binding of ribonuclease (17). Therefore, it seems that the LRR domain of RLKs may also function in binding to proteinaceous ligands, although there may be exceptions to this scheme. So far, the CLAVATA1 RLK represents the only example of this class for which a putative polypeptide ligand, CLAVATA3, has been identified (18, 19).

The LRR-RLKs can be subdivided into 12 subfamilies on the basis of the amino acid relationships between their kinase domains. Members of each subfamily have characteristic numbers and locations of LRRs in their extracellular domains (Fig. 2). Several of these subfamilies have unrelated sequences or "islands" between their LRRs (for example, BRI1 in Fig. 2). It is interesting that the island in BRI1 might mediate binding to its steroid hormone ligand, brassinosteroid (20). LRRs make up less than 50% of the extracellular domains in some RLKs. The nature of the ligands for other LRR RLK subfamilies remains to be established.

In animals, LRRs are found in various membrane proteins such as Drosophila Toll, human Toll-like receptors (21), and Slit, the transmembrane ligand for the Drosophila Roundabout receptor (22). However, none of the animal LRR-containing proteins identified to date is a receptor protein kinase.

The second largest class of extracellular motifs found in RLKs are various sugar-binding motifs or lectins. In the case of lectin receptor protein kinases (LecRKs) (23), their extracellular domains contain legume lectin motifs originally found in the seeds of leguminous plants (24). Legume lectins are very well characterized in terms of their three-dimensional structure, sugar-ligand interaction, and specificity. Ironically, very little is known about their biological roles in plants. Legume lectins can bind various disaccharides such as glucose-mannose, galactose-N-acetylgalactosamine, fucose, chitobiose, and other complex sugars (25). A BLAST database search of amino acid sequences showed that the legume lectin motifs in RLKs are similar to those of Ulex lectin II, a chitobiose-binding lectin (26). This suggests that these RLKs may also bind to chitobiose, a cell-wall component of fungi, insects, and nematodes.

The second type of lectin motif found in RLKs is the agglutinin motif that specifically binds α-D-mannose (27); the motif is located in the NH2-terminus of the extracellular domain of S-domain-containing RLKs, such as SRK (28). The putative polypeptide ligand for SRK, S-locus protein 11 (also known as S-locus Cys-rich) (SCR or SP11), has been identified (29); however, it is not known whether the agglutinin motif participates in ligand binding.

The third type of lectin motif found in plant RLKs is the C-type lectin (30). C-type lectin motifs are found in proteins that mediate innate immune responses in mammals, and they bind to a diverse range of sugar moieties on the surface of nonself biological entities and cells (31, 32). The presence of a C-type lectin motif in a subfamily of RLKs raises the question of whether these proteins function in pathogen detection and plant response. Although a large number of extracellular or membrane proteins in animals contain C-lectins, none of these proteins is a receptor protein kinase.

In addition to the motifs named after lectins, other types of sugar-binding motifs are also present in RLKs. The first is the lysin motif found in several bacterial enzymes that function in bacterial wall biosynthesis and in several eukaryotic proteins of unknown function. The lysin motif may represent a general peptidoglycan-binding motif (33). Its presence in a plant receptor protein kinase raises the intriguing possibility that it is responsible for sensing cell-wall fragments of bacteria and that it serves as a molecular trigger eliciting plant defense systems.

Other types of molecular trigger RLKs might detect other biotic factors. The clearest example of an RLK that appears to function in the detection of fungi (or nematodes, or even insects) can be found in tobacco chitinase receptor kinase 1 (CHRK1), an RLK with a chitinase extracellular domain (34). CHRK1 lacks an essential glutamic acid for chitinase function and shows no catalytic activity toward chitin, suggesting that it simply serves as a chitin-binding motif. Another motif implicated in binding to fungal cell walls is the thaumatin domain, which is found in the extracellular region of Arabidopsis PR5K (35). Thaumatin domains have antifungal activity and in vitro chitinase activity (36). Whether thaumatin-containing RLKs have such a function remains to be determined.

In addition to motifs implicated in interacting with microbial cell walls, some RLKs bind to plant cell-wall components. The extracellular domains of cell wall-associated kinase (WAK)-type RLKs are associated with pectin, an ingredient in the middle lamella and primary cell wall (37); however, the region responsible for pectin binding has yet to be identified. Some of the RLKs contain proline-rich sequences resembling those found in the plant protein extensin, a structural component of the cell wall (14, 36). Thus, the proline-rich sequences in RLKs that are similar to extensin may interact with cell-wall components directly. Perhaps these RLKs with wall-bound extracellular domains function in sensing the integrity or dynamic status of the cell wall.

The EGF repeat is the only motif found in both plant RLKs and animal receptor tyrosine kinases so far. Examples of animal receptors with EGF repeats are the receptors Tie and Tek, which are essential for angiogenesis. The crystal structure of the Ca2+-binding EGF repeats indicates that these repeats are directly involved in protein-protein interactions (38). This study provides a structural basis for the function of multiple tandem-linked Ca2+-binding EGF-like motifs, but it is not clear whether the one-to-two non-Ca2+-binding type EGF repeats found in plant RLKs behave in a similar manner. To date, all EGF repeats in plants are found in receptor-like kinases of the WAK-type (39) and SRK-type or secreted proteins closely resembling the extracellular domains of SRKs.

The tumor necrosis factor receptor (TNFR) motif is found in a number of animal receptors (without COOH-terminal kinases) regulating processes ranging from cell proliferation to programmed cell death and is directly involved in binding to the cognate ligands of TNFR (40). In all plant proteins examined so far, only the CRINKLY4-type RLKs contain a single TNFR motif in their extracellular domains, and it is hypothesized that it may be involved in binding to its putative protein ligand (41); however, the ligand for CRINKLY4 remains unknown. In addition to the TNFR motif, the extracellular domain of CRINKLY4 also contains seven tandem repeats with unknown functions.

A large number of RLKs do not have any established sequence motif (Fig. 2 and Table 1). BLAST searches using these extracellular domains against the GenBank nonredundant database yielded the same RLK sequences or receptor-like proteins resembling the extracellular domains of RLKs. No nonplant sequences were obtained during these searches. These extracellular domain sequences may represent RLK-specific motifs unique to plants.

Expression Patterns and Functions of RLKS

Functional categories of RLKs

The large number of RLKs present in the Arabidopsis genome and their diversity suggest that RLKs may function in perception of a wide range of signals or stimuli. Although RLKs with known functions represent fewer than 2% of the total RLKs identified, the functions for several RLKs have been studied in detail. The identity of the RLK extracellular domains also reveals tantalizing clues to the potential functions of orphan RLKs.

The functions of plant RLKs can be divided into two broad categories (Table 2). The first category includes RLKs involved in the control of plant growth and development under normal growth conditions. The second category includes RLKs involved in plant-microbe interactions and stress responses. The functions of RLK homologs in animals, the kinases Drosophila Pelle and human interleukin 1-receptor-associated kinase (IRAK), also fall into these two functional categories. On the basis of kinase phylogeny, plant RLKs and animal Pelle kinases appear to belong to a monophyletic group and are each other's closest relatives in the eukaryotic kinase superfamily (12). Pelle is involved in determining both dorsal and ventral patterning and innate immunity in fly (42). Although it is not known whether IRAK functions in development, IRAK does participate in innate immune responses (43). The evolutionary relationships between these proteins and what appear to be overlapping functions suggest that the kinases ancestral to these two gene families may have been involved in these two "functions" before the divergence between plants and animals (12, 19, 44).

Table 2.

The functions and expression patterns of RLKs. *These two categories are not mutually exclusive, because most expression studies do not test potential conditions exhaustively.

Aside from these RLKs, SRK from Brassica is a single member of a unique category. For some of the RLKs with no known function, the expression pattern or conditions required for their induction have been determined, and they are discussed in the following sections.

Growth and developmental functions

A number of RLKs with growth and developmental functions have been identified. These include Petunia pollen-expressed receptor-like kinase 1 (PRK1) (45), maize CRINKLY4 (41), Arabidopsis BRI1 (46), CLAVATA1 (47), ERECTA (48), and HAESA (49). Their compartmental expression patterns, in general, correlate well with their functions, and the extent of their expression in plant tissue ranges from restricted to global. RLKs with restricted expression patterns play specific roles in controlling growth and development in certain cells, tissues, or organs. For example, Petunia PRK is an RLK with five LRRs (45). PRK is expressed in anthers with binucleate pollen, in mature pollen, and in germinated pollen tubes. Decreased amounts of PRK1 transcripts in antisense PRK1-expressing plant lines result in the loss of nuclei and abortion of pollen, suggesting that PRK1 plays an essential role in the postmeiotic development of microspores (50).

The second gene in this category, Arabidopsis CLAVATA1, encodes a protein that contains 20 LRRs in its extracellular domain. It is expressed specifically in the L2 and L3 layers of the shoot apical meristem. Mutations in the CLAVATA1 locus result in the accumulation of undifferentiated cells at the shoot and floral meristems (47, 51). Together with CLAVATA2 and CLAVATA3, CLAVATA1 regulates the balance between cell proliferation and differentiation at the shoot meristem (18, 19, 52, 53).

Another RLK with 20 LRRs, the protein product of HAESA, is also expressed in tissue-specific areas, including floral abscission zones, pedicel bases, and leaf petiole bases. Plants expressing antisense HAESA RNA show a delay in abscission of floral organs, suggesting that HAESA could be involved in the developmental regulation of abscission processes (49). It is noteworthy that the pedicels and leaves of Arabidopsis do not normally undergo abscission, so the significance of HAESA expression in these regions remains unknown. ERECTA, still another gene for an LRR RLK with 20 LRRs from Arabidopsis, affects the initiation and elongation of organs originating from the shoot apical meristem after the floral transition (48).

The expression of ERECTA is less restricted than CLAVATA1 or HAESA because various amounts of ERECTA transcripts are found in all organs except root (48). A detailed localization study showed that strong ERECTA expression can be detected in the shoot apical meristem after the floral transition, in floral organ primordia, and in immature floral organs. The lessrestricted expression pattern is consistent with its relatively global mutant phenotype compared with those of CLAVATA1 and HAESA (54).

In contrast to mutations in genes with relatively specific expression patterns and organ- or tissue-specific phenotypes, mutations in the widely expressed Arabidopsis BRI1 gene result in dwarfism, delayed senescence, reduced fertility, and light-independent development because of insensitivity to the plant steroid hormone brassinosteroid (46). The global phenotype is consistent with its ubiquitous expression in Arabidopsis (55). The BRI1 ortholog in rice has been identified recently, and mutations in rice BRI1 also confer brassinosteroid insensitivity and other related phenotypes. This represents the first example of conservation of RLK function between monocots (rice) and dicots (Arabidopsis) (56). In addition to BRI1, maize CRINKLY4, which encodes a TNFR motif-containing RLK, exhibits a global phenotype when mutated (41). Mutations in CRINKLY4 alter the cell size and morphology of the leaf epidermis and result in fusion between organs; however, its expression pattern remains unknown.

Plant-microbe interactions and stress responses

Several RLK family members have been implicated in controlling disease resistance or in interacting with proteins of microbial origins. These genes include tomato Pto (57), tomato Pti (58), rice Xa21 (59), wheat LRK10 (60), Arabidopsis FLS2 (61), and Arabidopsis PBS1 (62). These genes confer resistance to bacterial and fungal pathogens and include both receptor protein kinases and cytoplasmic kinases.

Three RLCKs--Pto, Pti, and PBS1--mediate bacterial resistance. Pto confers resistance specifically to Pseudomonas syringae pv. tomato expressing the avrPto avirulence gene (57). Pti1, on the other hand, physically interacts with Pto and confers disease resistance when introduced into susceptible plants (58). PBS1 mediates resistance against Pseudomonas syringae pv. phaseolicola (62).

Xa21 and FLS2 are both LRR-containing RLKs with 21 and 28 LRRs, respectively. On the basis of the amino acid sequences of their kinase domains, these two proteins appear to be close relatives belonging to the same RLK subfamily. The Xa21 gene confers resistance to the bacterial blight pathogen, Xanthomonas oryzae pv. oryzae. However, FLS2 confers sensitivity to flagellin, which induces responses reminiscent of pathogen-mediated hypersensitive reactions (63). In addition to resistance to bacterial pathogens, one RLK has been implicated in mediating fungal resistance. Wheat rust resistance is mediated by a single locus, Lr10, and an RLK, LRK10, is found in this locus (60). Wheat rust-resistant lines have a complete copy of LRK10, whereas susceptible lines do not; however, the role of LRK10 has yet to be demonstrated through gain-of-function studies.

Several other genes are implicated in defense responses, although genetic mutants are not yet available. Among them, the best studied are the WAK-like genes (64). The expression of an antisense or a truncated form of WAK1 resulted in increased sensitivity to the salicylic acid analog 2,2-dichloroisonicotinic acid (INA) (65). Because salicylic acid is produced during defense responses and mediates resistance in plants, WAK1 may function in similar processes. However, in another study using plants that expressed antisense WAK2 RNA (to reduce the amount of related WAKs), reduced amounts of WAK protein were correlated with the loss of cell-wall expansion (37). The distinct phenotypes observed in WAK1 and WAK2 studies may indicate functional diversification of related RLK genes. Nevertheless, they may also indicate that the phenotypes observed are context sensitive. It remains to be seen whether the phenotypic effects were due to suppression of multiple WAKs with similar functions and whether WAK1 antisense lines exhibit altered cell-expansion pattern and WAK2 antisense lines have altered INA sensitivity.

Self-incompatibility

Brassica S-locus receptor kinase (SRK) mediates the self-incompatibility response in Brassicaceae in which self-pollen is rejected on the stigma surface (10). SRK represents a unique intermediate between RLKs mediating developmental and resistance functions, because it functions not only in the regulation of stigma development, but also in self-nonself recognition. SRK was first isolated as a marker cosegregating with the self-incompatible locus (28) and is highly expressed in the stigma (28, 66). The importance of SRK as the stigma determinant of self-incompatibility was later demonstrated through loss-of-function (67) and gain-of-function (68) studies. Although SRK mediates self-incompatibility in Brassica, genes highly similar to SRK can also be found in the self-compatible Arabidopsis thaliana (69). In fact, SRK-like genes in Arabidopsis belong to a large subfamily with 32 members, but the none of their functions is known. It is suggested that the development of the SRK-mediated self-incompatibility response is an evolutionarily relatively recent event in Brassicaceae that may have occurred through the recruitment of preexisting genes that performed other unrelated functions, or through duplicated genes that have undergone functional diversification (70). Consistent with the hypothesis for an ancestral function of SRK-like genes other than self-incompatibility, SRK-like genes in maize may be involved in reproductive development (71).

Genes with only expression data

Aside from the RLKs identified through the genome sequencing project, quite a few were cloned independently from various plant species (Table 2). The expression profiles of some of these genes have been determined. Most of these genes are expressed in specific areas or developmental stages in plants, including tomato PRK1 and PRK2, which are localized to pollen tubes (72); carrot somatic embryogenesis receptor-like kinase (SERK), expressed in the embryonic through the early globular stages (73); RKF1 (receptor-like kinase in flowers 1), expressed in flower primordia and during early stamen development (74); Sorghum bicolor receptor-like kinase 1 (SbRLK1), accumulating preferentially in leaf mesophyll cells (75); WAK4, expressed only in siliques (39); and bean senescence-associated receptor-like kinase (SARK), expressed during leaf senescence (76). These genes may participate in the control of localized development or growth responses.

Relatively few RLKs are expressed globally. Arabidopsis receptor-like transmembrane kinase 1 (TMK1), which encodes an LRR RLK and is expressed in large amounts in many tissues, shows a high degree of global expression (77). Another example, Arabidopsis RPK1, is expressed in all major tissues (78). RPK1 expression is increased upon ABA treatment and under several abiotic stress conditions.

Like that of RPK1, the expression of several RLKs is enhanced or repressed under various conditions. The expression of Arabidopsis LRRPK is repressed in plants exposed to light (79). Rice TMK1 expression is increased by gibberellic acid (GA) (80). The expression of several Arabidopsis RLKs, including that of RKL1, RKS1, RKS2 (68), RKC1 (81), WAKs (39), and RLK3 (82), is increased by salicylic acid treatment. RLK3 expression also increases during oxidative stress and pathogen treatment, suggesting a role for RLK3 during pathogen attack. Brassica S gene family receptor 1 (SFR1) and SFR2 are both SRK-like genes, and their expression is increased by plant wounding and pathogen challenge (83). These genes may function in the defense responses in plants.

The 599 unknowns

Although the above examples provide some clues about the overall function of RLKs, most of the RLKs identified so far do not have defined functions. The relative ease of generating gene knockouts in Arabidopsis is expected to greatly facilitate functional analysis of RLKs. Combined with expression studies, this approach has been useful in identifying the function of HAESA (49); however, the effort to generate a single loss-of-function mutant often does not pay off. For example, we have found that gene silencing of the LRR-type RLK TMK1 results in no discernible phenotype (77).

One of the likely reasons for a lack of observable phenotype conferred by a mutated gene is that other genes with redundant functions are present in the genome. In Arabidopsis genome, 266 out of 610 RLKs have one to six sequences with more than 80% sequence identity in their kinase domains. This high degree of sequence similarity suggests that RLKs have some degree of functional redundancy. Moreover, 33% of the RLK family members are found in tandem duplications, creating yet another challenge in studying their functions. Another potential reason for the absence of phenotype is that the signals or stimuli perceived by the receptors are only present under suitable conditions. For example, the mutation of RLKs that mediate disease resistance or defense would not have any discernible phenotype unless plants were exposed to the proper pathogen or microbe. Just how many of the RLKs may be in this category is an open question.

RLK Signaling Pathways

Many of the components of RLK signaling pathways have been identified by a combination of genetic and biochemical approaches. In particular, our knowledge of the signaling pathways of SRK, CLAVATA1, Pto, and FLS2 is more complete than that of other kinases, and these kinases have been selected as examples for the differences and similarities among RLK signaling pathways. The pathways are still fragmented at best, even for the best-studied RLKs.

CLAVATA1. CLAVATA1, CLAVATA2, and CLAVATA3 were identified originally through genetic screens for mutant plants accumulating undifferentiated cells in the apical meristem (51, 84). CLAVATA1 encodes an LRR RLK, whereas CLAVATA3 encodes a small, secreted protein that acts with CLAVATA1 in controlling meristem proliferation (Fig. 3A) (53). In fact, CLAVATA3 is the ligand of CLAVATA1 (19). CLAVATA2 may take part in CLAVATA1-CLAVATA3-mediated developmental functions and in an independent pathway regulating organ development (85). CLAVATA2 encodes an LRR protein with a signal peptide and a transmembrane domain similar to those in proteins encoded by the Cf genes, which confer disease resistance to pathogenic fungi (86). CLAVATA1 and CLAVATA2 are both required for CLAVATA3 function and CLAVATA3 interaction with CLAVATA1 occurs when CLAVATA2 is present (18, 19). CLAVATA1 and CLAVATA3 are found in the same ~450-kD complex with kinase-associated protein phosphatase (KAPP, a type 2C phosphatase), and a putative Rho GTPase (52). Moreover, this complex is not formed in clv1-10 mutants, which lack the intrinsic kinase activity of CLAVATA1. The role of KAPP in CLAVATA1 signaling has been demonstrated by its overexpression (87) and its reduced expression (88). The data suggest that KAPP serves as a negative regulator of CLAVATA1 function. KAPP also interacts with other RLKs (89); however, the in vivo functions for these interactions remain to be elucidated. Downstream of the multiprotein signaling complex, two genes, WUSCHEL (16, 90) and POLTERGEIST (91), are negatively regulated by CLAVATA1. WUSCHEL is a homeobox transcription factor whose expression is repressed by CLAVATA activity, whereas WUSCHEL expression induces the expression of CLAVATA3, forming a feedback loop in controlling the proliferation of the meristem. The second gene, POLTERGEIST, was isolated as a suppressor of the CLAVATA mutant phenotype; however, the molecular identity of POLTERGEIST is unknown.

S-locus receptor kinase

SRK resides in the S locus, which determines self-incompatibility responses in Brassica. In this system, rejection of pollen occurs when common alleles at the S locus are present in both pollen and female receptive tissues (Fig. 3B). In addition to SRK, two other proteins involved in self-incompatibility are found in the S locus. The first is S-locus glycoprotein (SLG), a soluble protein that resembles the extracellular domains of SRK. Its exact function in self-incompatibility is not clear, although it enhances the response (74). SCR/SP11, the putative ligand for SRK, is also found in the S locus and is the pollen determinant of self-incompatibility in both loss-of-function and gain-of-function studies (29). Components downstream and specific to SRK have also been identified, including the thioredoxins THL1 and THL2 (92) and an armadillo-repeat-containing protein ARC1 (93). KAPP also binds SRK in vitro (89).

The interaction of THL1 and THL2 with SRK does not require the kinase activity of SRK. Thioredoxins inhibit SRK phosphorylation and, therefore, are postulated to act as negative regulators of SRK function (94). The residue crucial for the association of THL1 and THL2 with SRK was mapped to a single cysteine located at the NH2-terminal of the SRK kinase domain (95). These findings are consistent with the hypothesis that the redox capability of thioredoxins is essential for the regulation of SRK and suggest the presence of redox-controlling factors in the SRK signaling pathway. This raises the intriguing question of whether ion channel activity is involved in RLK signaling.

In contrast to THL1 and THL2, ARC1 only interacts with SRK in a phosphorylation-dependent manner. Suppression of ARC1 production in the stigma results in the partial breakdown of self-incompatibility (96). ARC1 is likely to function downstream of SRK. The function of KAPP in SRK signaling pathway is unknown but it is generally assumed that it may regulate SRK signaling by dephosphorylation. In self-compatible Brassica, an aquaporin-like gene was thought to be responsible for the breakdown of self-incompatibility (97); however, its involvement has not been demonstrated unequivocally.

Pto

Pto is a well-studied cytoplasmic kinase member in this gene family (Fig. 3C). The ligand for Pto is avrPto, a small peptide secreted by the pathogen Pseudomonas syringae pv. tomato. AvrPto interacts with Pto directly in a phosphorylation-dependent manner, and their interaction is essential for the onset of disease resistance (98, 99). The Prf gene was identified through a genetic screen for additional genes mediating resistance to Pseudomonas syringae pv. tomato and is required for Pto-mediated resistance. It encodes a protein with a leucine zipper, a nucleotide binding site, and LRRs (100). Prf does not interact with avrPto but is required for the avrPto-Pto-mediated resistance response (101). Other components involved in avrPto-induced resistance response have been identified through yeast two-hybrid screens. Pti1 (58), and the multiprotein complex of Pti4, Pti5, and Pti6 (Pti4-5-6) (102) were found to interact specifically with Pto. Pti1 is a serine-threonine kinase phosphorylated by Pto and is required for the avrPto-Pto-mediated hypersensitive response. Pti4-5-6 are transcription factors similar to tobacco ethylene-responsive element binding proteins (EREBPs) and bind to the promoter regions of genes encoding a large number of pathogenesis-related proteins. Yeast two-hybrid assays also identified proteins that associate with both avrPto and Pto, and proteins interacting with avrPto only (103). So far, these interacting proteins have not been tested for their contribution to avrPto-induced resistance, and it remains to be seen if they are indeed components of the pathway.

Flagellin Insensitive 2

In addition to recognition systems for specific pathogens, plants have evolved perception systems for microbial components. One such component is flagellin, the building block of bacterial flagella (104). Several lines of evidence indicate that FLS2 functions in plant responses to flagellin (Fig. 3D). First of all, mutations in FLS2 resulted in insensitivity to flagellin, and functional FLS2 restored sensitivity. Secondly, transgenic plants overexpressing FLS2 showed increased sensitivity to flagellin. Intuitively, flagellin might serve as the ligand for FLS2 in triggering the sensitivity response. Using iodinated flagellin peptide, flg22, Meindl et al. showed that specific and saturable binding sites to flg22 are present in both intact cells and membrane preparations isolated from Arabidopsis (104). Cross-linking experiments further demonstrated that a 115-kD polypeptide is responsible for this interaction; however, it is not clear whether this 115-kD protein represents FLS2. Wild-type FLS2 restored flagellin binding in two FLS2 mutant alleles (105). Other components in the pathway include FLS1, which was identified through flagellin insensitivity screens (63) AtMPK6, which shows transient increases in phosphorylation in plants treated with flagellin (106) and KAPP, which interacts with the kinase domain of FLS2 in vitro and whose overexpression renders plants flagellin-insensitive (105).

The human flagellin receptor is Toll-like receptor 5 (TLR5) (107). TLR5, like the other Toll-like receptors, is an LRR-containing transmembrane protein and interacts indirectly with interleukin-1-receptor-associated kinase (IRAK) upon ligand binding. It is interesting that the plant RLK FLS2 has an LRR extracellular domain and an intracellular kinase with sequence similarity to IRAK. These correlations suggest that TLR5-IRAK and FLS2 are evolutionarily related and raise the intriguing question of whether, in the ancestral condition, plants had two proteins or a chimera.

Other components associated with RLKs

The rice ortholog of KAPP associates with Xa21 (80). Like its Arabidopsis counterpart, rice KAPP also interacts with several RLKs. This rather nonspecific binding suggests that KAPP may serve as a modulator commonly used in RLK signaling pathways. The extracellular domain of WAK1 binds to a glycine-rich protein, AtGRP-3 (108), and it was therefore suggested that AtGRP-3 may represent the WAK1 ligand. In any case, AtGRP-3 represents the first protein identified with the extracellular domains of RLKs as baits in a two-hybrid screen.

Receptor-like proteins and RLK signaling

One emerging feature of RLK signaling is the involvement of receptor-like proteins (RLPs). They are proteins resembling the extracellular domains of RLKs. Some RLPs are hypothesized to be secreted proteins, whereas others are membrane-bound. Examples of RLPs include SLG (109) and CLAVATA2 (86). In addition, Xa21D, an RLP resembling the extracellular domain of Xa21, can confer partial resistance to plants exposed to pathogens (110); however, the underlying mechanism for the action of Xa21D is still unknown. In the genome of Arabidopsis, a large number of RLPs can be found that resemble the extracellular domains of several types of RLKs (4). The functions of these genes remain elusive. Some of these RLPs from Arabidopsis resemble the Cf resistance gene from tomato, suggesting their potential roles in disease resistance. In addition, the resemblance between RLP and RLK raises the possibility that some RLPs represent a natural form of dominant-negative regulation of RLK activity. Although this is a likely scenario, no RLP examined so far seems to function in this manner.

Biochemical Properties of RLKS

Because RLKs and animal receptor tyrosine kinases have similar domain organizations, it is widely assumed that the mechanism of RLK action may be similar to the mechanism of animal receptor tyrosine kinase activation (13, 111, 112). As mentioned before, animal Pelle kinases are more closely related to plant RLKs than any other kinases. In addition, all RLK and Pelle kinases are serine-threonine kinases--perhaps suggesting that the biochemical properties of RLKs and Pelle kinases are likely to be similar--whereas the majority of animal receptor protein kinases are tyrosine kinases. Recent efforts in RLK research have generated a somewhat clearer understanding of their biochemical properties and are summarized below.

The membrane association and subcellular location of RLKs

The majority of RLK family members are type I membrane proteins. Among them, SRK (113), HAESA (49), BRI1 (55), and TMK1 (77) are localized to the plasma membrane. The PERK-like sequences have single putative transmembrane regions without signal sequence (14). They have no obvious organellar targeting signals and may be localized to the plasma membrane as well. Arabidopsis TAK1 has no extracellular domain but a single, long NH2-terminal hydrophobic region with no cleavage site that may serve as a membrane anchor. It is interesting that TAK1 associates with thylakoid membrane, representing the first RLK member whose localization to an organelle is known (15). At least 10 other TAK-like genes are present in Arabidopsis, and all of them have the same NH2-terminal anchor. RLCKs such as APK1 (114), Pto (57), and PBS1 (62) contain putative myristoylation sites. It remains to be seen if these three gene products are indeed myristoylated in vitro or in vivo.

RLK signaling complexes

Two groups have independently shown that some RLKs may form signaling complexes. CLAVATA1 is a component of a 450-kD complex containing its ligand CLAVATA3, KAPP, a Rho GTPase, and, perhaps, CLAVATA2 (17, 52, 86). KAPP is also found in a complex consisting of AtGRP-3 and WAK1 (108).

An interesting question is whether RLKs form dimers or oligomers in these complexes. Ligand-induced dimerization or oligomerization of receptor tyrosine kinases is the mechanism by which tyrosine phosphorylation is triggered (13, 111). Based on the similarity in the domain organization of RLKs and receptor tyrosine kinases, RLK activation may be similar to that of receptor tyrosine kinases, involving ligand-induced dimerization and autophosphorylation. Giranton et al. found that SRK was present in a complex of large molecular mass after treatment with cross-linking agents, and the authors argue that these larger forms represent oligomers of SRKs (115). The large quantitative differences between the cross-linked products and the SRK monomers suggest that only a small portion of SRK formed oligomers in the sample. It is also possible that the large complex simply represents cross-linking of SRKs to other proteins nonspecifically.

Ligand-activated phosphorylation

Another hallmark of animal receptor protein kinase signaling is the ligand-induced activation of its intrinsic kinase activity (13,111). Among plant RLKs, few have putative ligands, and only BRI1 (20) and SRK (94) have been tested as to whether ligand application activates the intrinsic kinase activities. The steroid hormone brassinosteroid binds to BRI1 in a specific, saturable fashion (20). Brassinosteroid treatment of BRI1-expressing plants results in a shift in the apparent molecular mass of the BRI1 protein. This decreased electrophoretic mobility is abolished in the presence of protein phosphatase. In addition, BRI1 isolated from the kinase inactive bri1-117 background does not show a molecular mass shift in the absence or presence of brassinosteroid. These findings demonstrate, albeit indirectly, that brassinosteroid application results in increased phosphorylation of BRI1 protein. In the case of SRK, Cabrillac et al. showed that pollination with incompatible pollen results in in vivo phosphorylation of SRK, whereas pollination with pollen from compatible plants does not (94). Because the ligand SCR/SP11 is located in the pollen coat, this result suggests that SCR/SP11 may contribute to the change in the phosphorylation status of SRK. This intriguing result is a direct demonstration of specific activation of RLK phosphorylation; however, the active component responsible for this activation remains to be identified.

Although they do not address whether ligand binding results in kinase activation, two findings suggest that the kinase activity of RLKs is important for the stable association of the ligand. Trotochaud et al. found that kinase-inactive CLAVATA1 was not able to associate with its ligand, CLAVATA3 (19). In the case of FLS2, the insensitive mutant fls2-17 has a point mutation in the kinase domain that abolished kinase activity in an in vitro assay (105). It is possible that unphosphorylated RLKs bind to their ligands at low affinity, and ligand-induced phosphorylation transforms RLKs into high-affinity forms. Alternatively, these mutant alleles, despite their lack of kinase activity, may have other defects such as conformational changes preventing the proper receptor complex formation.

Autophosphorylation of RLKs

The loss-of-function alleles of several RLKs have mutations in the kinase domain, suggesting that the kinase activity and the proper conformation of the kinase domain are important for RLK function. Kinase activity is critical for the proper function of SRK (116), Pto (101), CLAVATA1 (19), and BRI1 (20). A large number of kinases in this family have enzymatically active kinases both in vivo and in vitro and phosphorylate serine and threonine. It has been pointed out that different RLKs seem to differ in the relative amount of serine and threonine phosphorylated (117). Petunia PRK1 is the only RLK found to phosphorylate itself on tyrosine (118). Unfortunately, the phosphoamino acid analysis conducted in this paper was separated in only one dimension, which raises a question about the identity of the phosphotyrosine observed.

The kinase domains of RLKs show intramolecular or intermolecular autophosphorylation. These differences in phosphorylation might reflect the differences in the mechanism of receptor activation observed in animal RTKs (13). CrRLK1, Pto, and Xa21 only undergo intramolecular autophosphorylation during their activation (119-121). In addition, the kinase domains of CrRLK1 and Pto are highly similar, and phylogenetic analysis indicates that they belong to the same RLK subfamily; however, Pto does not have an extracellular domain, whereas CrRLK1 does. The similarity in their kinase domains suggests that these two RLKs may share similar signaling mechanisms even though they have distinct domain organizations. Both Xa21 and Pto confer disease resistance, raising the question of whether intramolecular phosphorylation is a common theme among members of this gene family involving in the defense response. All the other RLKs studied so far are capable of trans auto-phosphorylation. It is not clear whether any RLK is activated by both intramolecular and intermolecular autophosphorylation.

Concluding Remarks

The large number of RLKs found in plants suggests the presence of large numbers of potential signal sources that plants can sense and respond to. At this point, fewer than 2% of the total RLKs have known functions. Even fewer have their ligands or other components of their signaling pathways identified. With the completion of Arabidopsis genome sequencing project and the realization that more than 600 genes belong to the RLK gene family in Arabidopsis, we have just started to unravel the complexity of receptor signaling in plants.

The availability of large numbers of T-DNA insertion and transposon-tagged Arabidopsis lines has facilitated the discovery of RLK functions through insertion mutagenesis. However, in many cases, the knockout lines have no discernible phenotypes. Given the size of the RLK gene family and the similarity between subfamily members, it is likely that a substantial number of RLKs have overlapping or redundant functions. Therefore, the generation of double, triple, or quadruple knockouts may be necessary. In addition, half of the RLK family members with known functions mediate defense responses in plants, indicating that a large number of RLKs may have similar functions. This similarity suggests the need for proper conditions for RLKs to function properly and for RLK knockouts to reveal the consequences of loss of RLKs. One possible solution to this dilemma is the use of chimeric receptors, as shown in the BRI1-Xa21 fusion (122), which trigger disease resistance pathways in response to brassinolide application. In theory, a fusion between the extracellular domain of an RLK whose ligand is known and the kinase domain of an orphan RLK may allow the researchers to discern whether the application of ligand results in the elevation of defense response markers or some other observable biological outcome.

The kinase domains of RLKs can undergo inter- or intramolecular phosphorylation. There is some preliminary evidence for receptor complex formation. However, a detailed picture of the mechanisms and participants of any given RLK-mediated response has yet to emerge. The situation is complicated further by the presence of a large number of receptor-like proteins similar to the extracellular domains of RLKs. Do some of these receptor-like proteins participate in the signaling pathways of RLKs by facilitating ligand binding? Or do they serve as natural dominant-negative components in the pathways? Finally, more than 20% of genes in the RLK gene family are cytoplasmic kinases, yet the function of only two of them has been identified in plants. Do they interact with a receptor-like protein or do they respond to stimuli directly as in the case of Pto? These questions represent exciting challenges in understanding this divergent gene family.

References

  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 41.
  42. 42.
  43. 43.
  44. 44.
  45. 45.
  46. 46.
  47. 47.
  48. 48.
  49. 49.
  50. 50.
  51. 51.
  52. 52.
  53. 53.
  54. 54.
  55. 55.
  56. 56.
  57. 57.
  58. 58.
  59. 59.
  60. 60.
  61. 61.
  62. 62.
  63. 63.
  64. 64.
  65. 65.
  66. 66.
  67. 67.
  68. 68.
  69. 69.
  70. 70.
  71. 71.
  72. 72.
  73. 73.
  74. 74.
  75. 75.
  76. 76.
  77. 77.
  78. 78.
  79. 79.
  80. 80.
  81. 81.
  82. 82.
  83. 83.
  84. 84.
  85. 85.
  86. 86.
  87. 87.
  88. 88.
  89. 89.
  90. 90.
  91. 91.
  92. 92.
  93. 93.
  94. 94.
  95. 95.
  96. 96.
  97. 97.
  98. 98.
  99. 99.
  100. 100.
  101. 101.
  102. 102.
  103. 103.
  104. 104.
  105. 105.
  106. 106.
  107. 107.
  108. 108.
  109. 109.
  110. 110.
  111. 111.
  112. 112.
  113. 113.
  114. 114.
  115. 115.
  116. 116.
  117. 117.
  118. 118.
  119. 119.
  120. 120.
  121. 121.
  122. 122.
  123. 123.
  124. 124.
  125. 125.
  126. 126.
  127. 127.
  128. 128.
  129. 129.
  130. 130.
  131. 131.
View Abstract

Stay Connected to Science Signaling

Navigate This Article