The role of receptor-like proteins in Arabidopsis development

An intriguing and long-standing question in developmental biology is how plant cells communicate with each other and sense signals from their surrounding environment. Through research over past decades, it became clear that plant cells use membrane-localized receptors to perceive signals from their environment, which subsequently results in the initiation of downstream signaling (Kobe and Kajava, 2001; Torii, 2005). The membrane-associated receptors often form in multimeric complexes that contain receptor-like proteins (RLP) (Song et al., 1997; Jeong et al., 1999; Fritz-Laylin et al., 2005) as well as receptor-like kinase (RLK) proteins (Shiu and Bleecker, 2001a; 2001b). This is the case for the development of the shoot apical meristem (SAM) involving the CLAVATA1 (CLV1) and CLAVATA2 (CLV2) receptor molecules, as well as a small secreted polypeptide CLAVATA3 (CLV3) (Clark et al., 1993; Clark et al., 1995; Clark et al., 1997; Kayes and Clark, 1998; Fletcher et al., 1999; Jeong et al., 1999; Brand et al., 2000; Schoof et al., 2000). The Arabidopsis gene CLV2 encodes an LRR RLP acting in a functional CLV receptor complex that is involved in restricting the size of shoot meristem (Jeong et al., 1999). A total of 57 AtRLPs, which share sequence similarity and domain composition, have been identified in the Arabidopsis genome (Wang et al., 2008). However, the function of most AtRLPs remains elusive despite a genome-wide functional study into the roles of AtRLPs has been carried out recently (Wang et al., 2008). Given that fact that many AtRLPs originated from duplication events (Fritz-Laylin et al., 2005), it is very likely that the lack of identification of biological functions for AtRLP genes may be explained by functional redundancy, a phenomenon that typically obscures studies employing a reverse genetics strategy, as has been described for many RLK gene family members (Cano-Delgado et al., 2004; Shpak et al., 2004; Albrecht et al., 2005; DeYoung et al., 2006; Hord et al., 2006). In the future, RNA interference (RNAi) or artificial microRNA (amiRNA) approaches to target the expression of multiple AtRLP genes simultaneously could be followed to circumvent the functional redundancy (Chuang and Meyerowitz, 2000; Miki and Shimamoto, 2005; Ellendorff et al., 2008; Ossowksi et al., 2008). Alternatively, these multiple mutants of closely related AtRLP genes or potential co-expressed AtRLP genes should be combined to reveal the function of these genes. The clv2 mutant displays weaker, although similar phenotypes as clv1 and clv3 mutants (Kayes and Clark, 1998; Diévart et al., 2003; Chapters 2; 3.1), while loss-of-function mutants of clv3 are phenotypically stronger than clv1 or clv2 null mutants (Fletcher et al., 1999; Kayes and Clark, 1998; Dievart et al., 2003). In addition, CLV2 has a broader expression pattern, which may suggest a wider role for CLV2 in more developmental processes than only meristem development (Kayes and Clarks, 1998; Chapters 2; 3.1). Besides the broader role of CLV2, it is interesting to note that clv2 mutations, similar to clv1 mutations, are significantly affected by natural variation (Diévart et al., 2003; Chapters 3.1; 5), as has also been shown for the strubbelig (sub) and brassinosteroid insensitive1 (bri1) mutants (Chevalier et al., 2005; Cano-Delgado et al., 2004). The phenotype of clv2 in Col-0 (atrlp10) is also significantly enhanced by introduction into Ler background. Although our genetic interaction experiments indicated that the effect does not depend on the ER locus (Chapter 3.1), our observations imply that (a) CLV1/CLV2-modifying factor(s) exist(s) to co-regulate meristem development (Diévart et al., 2003; Chapters 3.1; 5). It would be interesting to identify these co-factor(s) in the future. Several lines of evidence suggest a role for CLV2 in root development. Over-expression of CLV3, CLE19 and CLE40 leads to an arrest of root growth (Casamitjana-Martinez et al., 2003; Hobe et al., 2003; Fiers et al., 2004), while the clv2 mutant can suppress the short-root phenotype caused by ectopic expression of CLE19 (Fiers et al., 2005). In addition, clv2 failed to respond to exogenously supplied synthetic CLE peptide, which corresponds to the conserved CLE motif of the CLV3/ESR gene family (Fiers et al., 2005; Ito et al., 2006; Kondo et al., 2006), indicating that CLV2 is able to perceive the CLE ligands in the root. However, the clv2 mutant exhibits no visible root phenotype under normal growth conditions, suggesting that a redundant protein, most likely another AtRLP gene, compensates for the loss of CLV2 function in the root. We found that only a few AtRLP genes are expressed in the root, although their expression is quite low (Chapters 2; 3.1). Possibly, a root defect only will become apparent in combination of multiple mutants for these genes. This observation argues that a CLV-like pathway also operates in roots (Fiers et al., 2007), but no RLK involved in this process has been found, although some of the RLKs that are expressed in the root (Birnbaum et al., 2003; Nawy et al., 2005). As such, CORYNE/Suppressor of overexpression of LLP-2 (CRN/SOL2) and Barely Any Meristem1-3 (BAM1-3) might be the logical RLK candidates for the redundant role in root development, because of their pronounced expression in roots (DeYoung et al., 2006; Müller et al., 2008; Miwa et al., 2008). Specifically, like clv2 mutants, crn/sol2 mutants did not respond to CLE peptide treatments (Müller et al., 2008; Miwa et al., 2008), suggesting that CRN, like CLV2, is involved in transmitting CLE signals. Therefore, studies with different combinations of mutants of these genes will help to clarify their biological function in root development (Chapter 2; 3.1; 4). Interestingly, WOX5, a homologue of WUS, marks the root Quiescent Centre (QC) identity and is expressed very early in the hypophysial cell (Sarkar et al., 2007), which is strikingly similar to the role of WUS in the shoot meristem (Haecker et al., 2004; Sarkar et al., 2007). Furthermore, it has been shown recently that POL and PLL, in addition to their role in Arabidopsis SAM maintenance (Song and Clark, 2005; Song et al., 2006), also act in the root meristem development through regulating the expression of the WUS homolog WOX5 (Song et al., 2008). These findings strengthen the hypothesis that a CLV-like pathway exists in the root meristem, which might include CLV2, CRN, WOX5, POL and PLL1. Similarly, our results as well as previous reports (Song and Clark, 2005; DeYoung et al., 2006; Müller et al., 2008) suggest that a CLV-related signaling pathway is involved in the regulation of leaf shape/size (Chapter 3.1) and pedicel length (Chapter 5). Interestingly, all these developmental pathways share some conserved factors such as POL and PLL, indicating common regulatory mechanisms exist in the developmental regulation of SAM, root meristem, leaf shape/size and pedicel growth. We provide evidence that two CLV2-related AtRLPs, AtRLP2 and AtRLP12, were capable to rescue clv2 mutants when expressed under the control of the CLV2 promoter (Chapter 4), suggesting that functional specification of these two AtRLPs may reside, at least in part, in their cis-regulatory elements. The importance of variation in expression pattern for the specificity in function of closely-related genes from multi-gene families while the proteins are interchangeable, has been documented for many genes such as the CLV1-like genes (BAM1-3, DeYoung et al., 2006; Hord et al., 2006), ERECTA-family (ER and ERL1-2, Shapk et al., 2004) and BRL-family members (BRI1, BRL1and BRL3; Cano-Delgado et al., 2004). However, the double mutant combinations of atrlp2 and atrlp12 mutants with atrlp10/clv2 did not show additive effects on growth of meristems and other organs when compared to that of the atrlp10/clv2 mutant, suggesting that other AtRLPs, such as AtRLP3 and AtRLP11 that are duplication counterparts of AtRLP2 and AtRLP12 respectively, can also replace the function of CLV2 in the regulation of the meristem development (Chapter 4). Therefore, additional phenotypes and the role of other AtRLP family members will become apparent only in the absence of the entire CLV2 close-related members as has been shown for ER, ERL1 and ERL2 (Shpak et al., 2004). Our studies in Chapter 4 revealed that several members of the AtRLP family can replace each other and are functionally equivalent. However, there are also family members with a similar protein domain organization, but that are clearly distinct from CLV2. This raised the question what determines the specificity in these proteins. We determined the function of the different domains by deletion analysis and generation of hybrid molecules (Chapter 4). CLV2 is still fully functional when the island domain is removed, while the C3-F region can be replaced by a close homologue (Chapter 4). Taken together, this study provided valuable information on the function of CLV2 domains that contribute to functional specificity and conservation. Despite these findings, little is known about the roles and specificity of other CLV2 domains, such as the transmembrane domain which is proposed to be the site for dimerization between CLV2 and CRN/SOL2 (Müller et al., 2008; Miwa et al., 2008). Of particular interest for future investigations are conserved residues flanking the LRR domain, which could be mutated to determine whether they are essential for CLV2 activity (Fritz-Laylin et al., 2005; van der Hoorn et al., 2005; Chapters 1; 3.1; 4). Previous studies proposed that CLV2 dimerizes with CLV1 to form an active receptor complex that binds the CLV3 ligand and initiates the downstream signaling pathway required for the maintenance of the stem cell population in the shoot apical meristem (Jeong et al. 1999; Trotochaud et al., 1999; Rojo et al., 2002; Dievart and Clark, 2004). The CLV2 protein was regarded as a stabilizer for the CLV1 protein based on the observation that CLV1 protein levels were reported to be reduced by over 90% in clv2 loss-of-function mutants (Jeong et al., 1999). In addition to CLV1, the receptor kinase CRN/SOL2 acts closely together with CLV2 to transmit the CLV3 signal independently but in parallel with CLV1 (Müller et al., 2008; Miwa et al., 2008). CRN/SOL2 has a kinase domain which might create a fully functional transmembrane receptor kinase together with CLV2 through dimerization in the transmembrane domains (Müller et al., 2008). This raises the hypothesis that CLV3 could bind CLV2 directly. However, whether CLV3 peptide can directly bind to the extracelluar domain of CLV2 remains to be validated. Nevertheless, the CLV3 signal is probably transduced through two separate receptor complexes, comprising CLV1/CLV2 (or CLV1 alone) and CRN/CLV2 (Müller et al., 2008). Despite these observations, the role of CLV2 in these signaling pathways remains largely unresolved. The CLV pathway is largely built on genetic data whereas direct biochemical evidence for the mode of action of the proteins involved is largely missing (Jeong et al. 1999; Trotochaud et al., 1999; Müller et al., 2008). Only very recently, it has been shown that the CLV3 peptide directly binds to the CLV1 ectodomain (Ogawa et al., 2008). In an effort to understand the physical interactions of the CLV proteins and their localization, we created fluorescently tagged versions of CLV1 and CLV2 (Chapter 5). These fusion proteins appeared to be targeted to the plasma membrane, displaying a common subcellular localization (Chapter 5). The functionality of the fusion proteins was confirmed by complementation of the respective mutants (Chapter 5). It has been postulated that the CLV1/CLV2 receptor complex resides in the plasma membrane to perceive the CLV3 ligand (Jeong et al. 1999; Trotochaud et al., 1999; Diévart et al., 2003). Our localization studies support this scenario. Unfortunately, it was not possible to determine a direct interaction between CLV1 and CLV2 in the framework of this Ph.D study. However, it is obvious that the interaction study can be the immediate next step using the available fluorescently-tagged CLV1 and CLV2 proteins and the stable transgenic lines generated in this study (Chapter 5). Furthermore, the study can be extended to visualize the components of the CLV signaling complex and follow their dynamics during plant growth. Furthermore, the labelled CLV-receptors can also be used for the isolation and identification of unknown components of the CLV receptor complex. For instance, it would be interesting to investigate whether CLV2 and CRN/SOL2 interact directly as proposed (Müller et al., 2008; Miwa et al., 2008). Indeed, a preliminary study using a similar approach supports the interaction of CLV2 and CRN/SOL2 (Y-F. Zhu and C-M. Liu, personal communication). It also intrigues to determine whether the CLV3 or other possible CLE(s) are directly interacting with the CLV2/CRN receptor complex. Undoubtedly, the tools generated in this study will be of great help for future experiments aiming to unravel the CLV signaling pathway.