[Note: This project was renewed by NSF on July 1, 2010. To see the abstract of the renewal project, "Arabidopsis 2010: Protein Interacting Networks and Site-Specific Phosphorylation in Leucine-Rich Repeat Receptor-Like Kinase Function", please go to the NSF website and view award abstract 1021363.]

Arabidopsis 2010: Functional Analysis and Phosphorylation Site Mapping of Leucine-Rich Repeat Receptor-Like Kinases in Arabidopsis

National Science Foundation (MCB)0419819

Project Objectives

The purpose of this project was to biochemically characterize kinase function in the entire 200-plus member LRR RLK family and to identify and functionally characterize in planta phosphorylation sites in a subset of 30 of these LRR RLKs. Global, discovery-based proteomics approaches were combined with in depth functional analyses of selected LRR RLKs to generate novel mechanisms on the role of phosphorylation in plant receptor kinase function.

Objective 1 Generation of bacterial and plant expression vectors for all 223 LRR RLKs. As a foundation for this study and as a resource for the Arabidopsis community, we used high-throughput Gateway cloning to generate a complete set of expression vectors yielding epitope tagged proteins for in vitro and in vivo LRR RLK functional analysis. We cloned 201 different LRR RLK cytoplasmic domains (including JM, KD and CT domains) into the Gateway entry vector pDONR/Zeo and verified by sequence analysis. Of these, 175 have been moved to Gateway expression vectors with a Flag epitope Tag and 189 with a His epitope tag, which establishes an important resource for biochemical studies. Most of these have now been examined for expression and kinase activity. A total of 208 full-length LRR RLK cDNAs have been amplified by RT-PCR and cloned into pDONR/Zeo and 194 of these have been fully sequenced (Gou et al., 2010). So far, 142 of these full-length constructs have been moved to a plant transformation vector with a Flag epitope tag and have been expressed in Arabidopsis, while 152 LRR-RLKs have also been expressed with a GFP tag in Arabidopsis. To date, 781 LRR RLK clones have been deposited with the Arabidopsis Biological Resource Center (ABRC) at Ohio State University for distribution, and 139 clones were sent directly to Dr. Wolf Frommer’s 2010 project (0618402) focusing on membrane protein interactions.

Objective 2 Biochemical studies of kinase domain function in the complete LRR RLK family. We transformed 334 His- and/or Flag-tagged bacterial expression constructs for LRR-RLK kinase domains in BL21 E. coli cells and then performed comparative biochemical assays on the recombinant kinase domains. The phosphorylation levels of the expressed proteins were assessed by; i) phosphorylation-specific staining with the dye Pro-Q Diamond; and ii) anti-phosphothreonine immunoblotting. The data set acquired to date suggests that sequence analyses may be sufficient to predict the autophosphorylation status of LRR RLK kinase domains. We continue to find that RD type LRR RLKs show significant protein autophosphorylation, while non-RD types do not.

Objective 3 Selection of 30 LRR RLKs for in depth studies. Proteomic methods involving multidimensional LC/MS/MS analysis were used to profile LRR RLK protein abundance and phosphorylation in whole Arabidopsis seedlings. A total of 30 LRR RLKs were selected for further detailed analysis as described below. We developed an effective membrane protein isolation strategy combining standard two phase-partitioning with methanol-assisted solubilization and extraction, prior to 2D LC/MS/MS analysis which resulted in at least a two- to three-fold increase in the identification of LRR RLKs (Mitra et al., 2007; Mitra et al., 2009). This method was applied to a study of the changes in the abundance of Arabidopsis membrane proteins in response to plant hormones and salt stress, using a label-free approach employing a novel parallel fragmentation mode during LC/MS/MS analysis.

Objective 4 Identification of in vivo and in vitro phosphorylation sites of LRR RLKs. Following our standard approach of combining both a family-wide screen with an intensive focus on a few selected LRR RLKs, we identified at least one in vivo phosphorylation site in 19 LRR-RLKs using the proteomic approach described in Objective 3. However, our most in-depth study was on BRI1 and BAK1 and the results have been published (Wang et al., 2005; Wang et al., 2008). We found that the in vitro BRI1 and BAK1 kinase domain autophosphorylation sites were highly predictive of in vivo phosphorylation and thus identification of in vitro LRR RLK sites can be a valuable preliminary exercise in examining biochemical function (Oh et al., 2000). We developed a high-throughput protocol for analyzing LRR RLK in vitro autophosphorylation sites using our Premier Q-ToF mass spectrometer functioning in both data-dependent acquisition LC/MS/MS and data-independent modes. Analyses with both un-enriched and IMAC-enriched peptide samples for over 100 LRR RLKs, with four injections per LRR RLK, were performed. We are continuing to finish our analysis of the remaining LRR RLKs and are currently compiling and manually assessing the data from these measurements to build upon our phosphorylation database. Hundreds of novel phosphorylation sites have now been identified. One common pattern across the family is that most LRR RLKs have multiple phosphorylation sites, sometimes more than 12 per kinase.

Objective 5 Functional analysis of LRR RLK phosphorylation sites. The functional significance of each of the identified and predicted phosphorylation sites in BRI1 and BAK1 were assessed by site-directed mutagenesis of each specific Ser or Thr to Ala or Asp, and each Tyr to Phe or Asp. Mutated constructs were examined by biochemical analysis in vitro and testing for the ability of the altered construct to rescue the weak bri1-5 BR-insensitive mutant in planta, as described in detail in three publications (Wang et al., 2005; Wang et al., 2008, Oh et al., 2009). Over 50 transgenic lines with mutations in the JM region, activation loop and CT region were generated for BRI1 and another 40 such lines were generated for BAK1, creating a valuable resource for further studies on the role of specific phosphorylation sites in protein-protein interaction. A similar approach is being used to functionally characterize individual phosphorylation sites in several LRR RLKs of known function in collaboration with other research groups.

Objective 6 Studies of LRR/RLK interactions. We are examining in planta interactions between LRR RLKs by co-immunoprecipitation experiments in plants expressing both proteins with different epitope tags. Our prototype for LRR RLK interaction studies is the BRI1/BAK1 model system. Our studies combining LC/MS/MS analysis, functional characterization in mutant backgrounds and in vitro biochemical studies allowed us to develop a novel sequential transphosphorylation model of BRI1/BAK1 interaction that shows plant receptor kinases share some of the properties of both TGF-β and receptor tyrosine kinases in mammals while retaining unique plant-specific features (Wang et al., 2008). It will be of interest to test this model on other LRR RLK interactions uncovered in the project. Using similar methods, we also studied in detail the interaction with BRI1 of the LRR RLKs most closely related to BAK1, including SERK1, SERK2 and BKK1 (SERK4). The results of some of these studies have been published (He et al., 2007) and others are in preparation for submission.