Genomic and Post-Translational Modification Analysis of Leucine-Rich-Repeat Receptor-Like Kinases in Brassica rapa

Among several receptor-like kinases (RLKs), leucine-rich-repeat receptor-like kinases (LRR-RLKs) are a major group of genes that play crucial roles in growth, development and stress responses in plant systems. Given that they have several functional roles, it is important to investigate their roles in Brassica rapa. In the present study, 303 LRR-RLKs were identified in the genome of B. rapa and comparative phylogenetic analysis of 1213 combined LRR-RLKs of B. rapa, Arabidopsis thaliana, Oryza sativa and Populus trichocarpa helped us to categorize the gene family into 15 subfamilies based on their sequence and structural similarities. The chromosome localizations of 293 genes allowed the prediction of duplicates, and motif conservation and intron/exon patterns showed differences among the B. rapa LRR-RLK (BrLRR-RLK) genes. Additionally, computational function annotation and expression analysis was used to predict their possible functional roles in the plant system. Biochemical results for 11 selected genes showed variations in phosphorylation activity. Interestingly, BrBAK1 showed strong auto-phosphorylation and trans-phosphorylation on its tyrosine and threonine residues compared with AtBAK1 in previous studies. The AtBAK1 receptor kinase is involved in plant growth and development, plant innate immunity, and programmed cell death, and our results suggest that BrBAK1 might also be involved in the same functions. Another interesting result was that BrBAK1, BrBRI1, BrPEPR1 and BrPEPR2 showed activity with both anti-phosphotyrosine and anti-phosphothreonine antibodies, indicating that they might have dual-specificity kinase activity. This study provides comprehensive results for the BrLRR-RLKs, revealing expansion of the gene family through gene duplications, structural similarities and variations among the genes, and potential functional roles according to gene ontology, transcriptome profiling and biochemical analysis.


Introduction
The Brassicaceae family is one of the major groups in the plant kingdom, composed of 340-360 genera and over 3,700 species distributed worldwide [1]. Among the six most cultivated species, the mesopolyploid Brassica rapa (AA genome, 2n = 2x = 20) is an economically, agronomically and scientifically important crop. Plants, animals and other multi-cellular organisms carry out inter-and intracellular communications with cell-surface receptor-like kinases (RLKs) through chemical signals. The first RLK was identified in maize [2] and subsequently RLKs were found in other plant species, including Arabidopsis and rice, which contain 610 and 1132 genes, respectively [3]. Furthermore, Arabidopsis RLKs were subdivided into 44 subfamilies according to phylogenetic analysis, and roughly 21 different extracellular domains were predicted [4].
Leucine-rich-repeat receptor-like kinases (LRR-RLK) form one of the major super-families in plant species and consist of an N-terminal LRR as an extracellular domain (ECD) and a Cterminal kinase domain. The LRR domain is 20-30 amino acids in length, is the major class of ECD [5] found among all of the RLKs in plant species, and is predicted to be involved in protein-protein interactions [6]. Additionally, LRR motifs comprise at least eight different classes with varying numbers and arrangements of LRRs in the sequence [4,5]. The protein kinase domain forms a lobe structure that consists of approximately 250-300 amino acids and has two sub-divisions, protein-serine/threonine kinases and protein tyrosine kinases. It is also divided into twelve sub-domains that are highly conserved and involved in catalytic activity, specifically in functions related to cell development [7]. Previously, these genes were predicted to be involved in different functions and placed in three different groups: (1) plant organ growth and development, (2) functions related to biotic and abiotic stresses and (3) multiple functions. For example, In A. thaliana, SOMATIC EMBRYOGENESIS RECEPTOR-LIKE KINASE (SERK) is involved in anther development and disease resistance [8], ERECTA (ER) functions in both ovule development and bacterial wilt [9,10], STRUBBELIG-RECEPTOR FAMILY (SRF) that functions in development of organs like leaves, stem and flower [11], FLAGELLIN-SENSITIVE 2 (FLS2) and PEP1 RECEPTOR (PEPR) is involved in both pathogen recognition and defense response and similarly, EF-TU RECEPTOR (EFR) is involved in enhancing immunity against few bacterial disease in rice, [12,13,14]. In addition, PHYTOSULFOKIN RECEPTOR 1 (PSKR1) is involved in cell proliferation and also in defense response against bacterial pathogens in Arabidopsis [15], BAK1-INTERACTING RECEPTOR-LIKE KINASE 1 (BIR1) is involved in both cell death and plant growth in temperature dependent manner [16], and BRASSINOSTEROID INSENSITIVE (BRI1) along with BRI1 ASSOCIATED RECEPTOR KINASE1 (BAK1) is involved in Brassinosteroid signaling [17]. Generally, Kinases that are involved in phosphorylation are conserved with positively charged arginine (R) and negatively charged aspartate (D) amino acid residues at the activation site, based on the conservation of the arginine (R) residue these kinases are divided into RD and non-RD type [18,19]. Likely, RD kinases are capable of auto-phosphorylation of the stimulation loop that is necessary for kinase activity. Conversely, non-RD kinases are mainly involved in plant immunity against microbes [18,19].
In this work, the whole genome sequence (WGS) of B. rapa [20] was used to identify B. rapa LRR-RLK (BrLRR-RLK) genes genome-wide. As these genes are important for functions related to plant morphological characters and plant stress resistance [5], characterizing this gene family will be useful for the genetic improvement of B. rapa. Because of the breeding and scientific importance of Chiifu (Chinese cabbage) among all the Chinese cabbage varieties, it was used for the identification of 303 BrLRR-RLK genes in the B. rapa genome. Furthermore, using these genes along with LRR-RLK genes from rice, Arabidopsis and Populus, a comparative phylogeny was constructed and we determined the motif patterns and intron/ exon structures for all of the genes. Additionally, the computational functional annotations, the expression patterns among six tissues [21] and the biochemical activity of selected genes were also investigated.

Materials and Methods
Identification, mapping, and duplication analysis of LRR-RLK genes in B. rapa Previously published Arabidopsis thaliana LRR-RLK protein sequences were retrieved and these genes were submitted to a Pfam batch search (http://pfam.sanger.ac.uk/search#tabview= tab1) to find the Pfam IDs. A protein alignment was retrieved for the Pfam IDs (PF00560, PF08263, PF12799, PF13516, PF13855, PF00069, PF07714) in the Stockholm format from the Pfam database (http://pfam.xfam.org/family/pf00069#tabview=tab3) and using this as a reference, we searched for LRR-RLKs among the protein sequences of B. rapa downloaded from the BRAD database (http://brassicadb.org/brad/). Redundant sequences were removed using our own Perl script and further confirmed by manual and Pfam batch searches. The BLAST-like Alignment Tool (BLAT) and TBlastN software were employed to map the positions of the LRR-RLK genes on the B. rapa chromosomes. Multiple sequence alignment and comparative analysis were performed to identify the duplicated BrLRR-RLK genes in the whole genome of B. rapa. The candidate genes were named BrLRR-RLK1 to BrLRR-RLK303 based on their positions counted from top to bottom on chromosomes A01 to A10. An A. thaliana ortholog search was carried out with the mega BLAST tool, which compared the full coding sequences of the BrLRR-RLK genes and all available A. thaliana genes. Highly similar (e-value > e- 45) homologous pairs were filtered out with in-home developed Perl scripts and further the resulted orthologs were cross checked with BRAD database and considered as orthologous genes.

Intron/exon analysis
The evolution and functions of a particular gene can be predicted through intron and exon patterns. Thus, to determine the intron and exon distribution and splicing phases of the 303 BrLRR-RLKs, we compared the BrLRR-RLK coding sequences (CDSs) and corresponding genomic sequences using the online Gene Structure Display Server (GSDS) (http://gsds.cbi. pku.edu.cn).
The molecular function, biological process and cellular component annotations were mapped from plant proteins that had high-scoring segment pairs (HSPs) with the BrLRR-RLK proteins in a BlastP alignment using the Blast2Go software and the data was retrieved in the text file format.

Phylogenetic tree construction
Using the ClustalX-2.0 software package with default parameters [25], multiple sequence alignments were performed with the LRR-RLK protein sequences of A. thaliana, B. rapa, Oryza sativa, and Populus trichocarpa. The neighbor-joining (NJ) method with 1000 bootstrap iterations was used for phylogenetic tree construction with the aligned LRR-RLK protein sequences. A circular phylogenetic tree was generated using the Molecular Evolutionary Genetics Analysis (MEGA) V5.1 software package (http://www.megasoftware.net/mega.php) [26].
Heat map expression pattern of LRR-RLK genes in B. rapa using RNAseq data For BrLRR-RLK gene expression analysis, RNA-seq data under accession no. GSE43245 previously produced and analyzed by Tong et al. [21] were downloaded from the NCBI Gene Expression Omnibus database (http://www.ncbi.nlm.nih.gov/geo/). These data included six tissues of B. rapa accession Chiifu-401-42: callus, root, stem, leaf, flower, and silique. Transcript abundance was recorded as fragments per kilobase of exon model per million mapped (FPKM), and a heat map of the BrLRR-RLK genes was generated using the CIMminer online software (http://discover.nci.nih.gov/cimminer/oneMatrix.do).

Cytoplasmic domain construction for recombinant proteins
Total RNA was extracted from leaf tissues of B. rapa Chiifu (Chinese cabbage) using TRI reagent according to the manufacturer's protocol (Applied Biosystems, Carlsbad, CA, USA) [27] and 2 μg of DNase-treated RNA was reverse transcribed into single-strand cDNA with Superscript III reverse transcriptase (Applied Biosystems). Eleven predicted BrLRR-RLK gene sequences from the Brassica database were compared by BlastN with the Arabidopsis genome information (AGI) CDS dataset to confirm their sequence identity. Then, the 11 selected genes were synthesized using gene-specific primers (S1 Table) and Accuprime pfu DNA polymerase (Applied Biosystems) with PCR conditions of 95°C for 2 min, 35 cycles of 95°C for 20 sec, 60°C for 40 sec, and 72°C for 2 min, with a final 72°C for 5 min. The amplified fragments were purified and cloned into the pENTRY-TOPO vector (Applied Biosystems) and confirmed as BrLRR-RLK genes through sequence analysis. Recombinant Flag-BrLRR-RLKs were amplified using pFLAG-MAC (Sigma-Aldrich, St Louis, MO, USA) and the PCR products were digested with specific enzymes and purified by agarose gel electrophoresis. The resulting fragments were ligated into pFLAG-MAC, which was digested with enzymes, followed by heat shock transformation into E. coli (DH5α) and plasmid extraction. The mutations were confirmed by sequence analysis. quantification was done using Coomassie Brilliant Blue staining and immunoblots were scanned using an Odyssey Infrared Imaging System (LI-COR Bioscience, Lincoln, NE, USA) for visualization.

Results and Discussion
LRR-RLK gene identification and chromosomal locations in the B. rapa genome A total of 303 non-redundant BrLRR-RLK genes were identified in the B. rapa genome and were confirmed through manual and Pfam searches for the presence of both ECDs and kinase domains (KDs) in the sequences. The homologous LRR-RLK genes identified in B. rapa in our study are reported in A. thaliana, O. sativa and P. trichocarpa to have important functional roles related to plant growth and development [29][30][31][32]. Most of the BrLRR-RLK genes contained leucine rich repeats (LRRs) in their ECDs, which play a major role in protein-protein interactions [6], whereas transmembrane (TM) domains and signal peptides (SP) are signatures of membrane RLK proteins [5]. In total, 293 genes, BrLRR-RLK1 to BrLRR-RLK293, were physically located on the 10 chromosomes and whereas the remaining 10 genes, BrLRR-RLK294 to BrLRR-RLK303 were mapped on scaffolds using the BLAT and TblastN software. The BrLRR-RLK genes were randomly distributed among all 10 chromosomes, with A01 and A06 having the most (39 genes) and A10 having the least (10 genes) (Fig 1 and Table 1).
Among non-RD types, the highest number (23 genes) was found on chromosome A01 and the remaining 109 genes were distributed randomly among the other chromosomes including scaffolds. In the case of RD type genes, 25 genes were mapped on A06, followed by 21 on A03, and the least number (4 genes) were found to be on scaffolds (S2 Table). Orthologous genes are homologous genes found in two different species but which originated from a common ancestor, and if a gene is duplicated and occupies two different positions in the same genome, then the two copies are considered paralogous [33]. In our analysis, around 143 genes were duplicated and formed 143 paralogous copies (Fig 1 and S3 Table). We also identified 11 BrLRR-RLK genes that were not orthologous to A. thaliana LRR-RLK genes (S2 Table). Surprisingly, 31 BrLRR-RLK genes were missing from crucifier genome building blocks that were confirmed by computational and manual curation, whereas the remaining 272 genes were categorized under 23 genomic blocks (GB) with a maximum of 30 in GB "U" and a minimum of 1 in "G"; moreover, 35 and 12 BrLRR-RLK genes have undergone tandem and segmental duplications (Fig 1, S2 Table and S3 Table). These results indicate that the duplication and uneven distribution of the genes on the chromosomes and GBs occurred because of duplication events during the process of evolution [34][35][36].
Comparative phylogenetic analysis of LRR-RLK genes in A. thaliana, B. rapa, O. sativa, and P. trichocarpa To identify the evolutionary relationships between the BrLRR-RLKs and other species, multiple alignments of 222, 303, 309 and 379 LRR-RLK proteins of A. thaliana, B. rapa, O. sativa and P. trichocarpa, respectively, were used for comparative phylogenetic analysis. The 1213 LRR-RLK proteins were divided into 15 subfamilies (I to XIV) according to the node support between the LRR-RLK genes [4,29,37] and to differentiate the subfamilies, each was indicated with a random color (Fig 2 and S1 Fig). Subfamily XII contained the most genes (186), followed by subfamilies III (181), XI (162), I (117), X (92), VII (88), VIII-2 (86), II (68), VI (52), V (43), VIII-1 (41), XIII (35) and IX (29), with the least number of genes in subfamilies XIV and IV (16 each) (S1 Fig). The origin of the RLK genes was predicted to be before the speciation of plants, The RD (red color) and non-RD (aqua color) type genes are distributed on chromosomes A01 to A10 in inner circle and paralogous copies are marked with different colored lines in center. animals, fungi and protists but expansion of this super-family might have occurred throughout land plant evolution [4,37]. The LRR-RLK genes of B. rapa and A. thaliana were evenly distributed among all the subfamilies but the rice and Populus LRR-RLK genes formed monophyletic clusters consisting of species-specific genes [30,38]. Similarly, monocot and dicot LRR-RLK genes were observed in all the subfamilies, indicating that expansion of the family might have occurred before their divergence [39]. Additionally, in our comparative phylogenetic analysis we observed that all the subfamilies comprised 30-80% monocot genes, but most contained 50-60% monocot genes. Surprisingly, subfamilies XII and VII contained the least number of dicot genes (20% and 25%, respectively). Comparison between monocot and dicot LRR-RLK gene numbers among the subfamilies revealed that gene duplication might have occurred largely within subfamilies XI in monocots and dicots, III and I in dicots and XII in monocots (S1 Fig). Since sister pairs denote the closest relatives of a group in a phylogenetic tree, we analyzed and predicted 342 pairs, this covered 56.3% of the total number of genes used in this study. These sister pairs included 288 pairs in the same species and 154 cross-species pairs,   Additionally, subfamilies II, VIII-1, IX, XIII and XIV consisted of RD type BrLRR-RLK genes and subfamilies III, IV, V and XII were composed of non-RD type genes, while subfamilies I, VI, VII, VIII-2, X and XI had both RD and non-RD type BrLRR-RLK genes (S2 Table). Non-RD kinases are known to function mainly as pathogen recognition receptors (PRRs) while RD kinases are involved in activation of the non-RD kinases [19,40]. Moreover, in our phylogeny we determined some of the important RD and non-RD LRR-RLK genes in different subfamilies that are involved in several functions related to plant growth and development, defense and plant immunity (S2 Table) [29,30,41].
Among the important RD and non-RD BrLRR-RLK genes, SRF2-9, FEI2, ERL1, RHS16, MRH1 and MEE39 had 10 or more introns, SERK1/3/4, TOAD2 and NIK1/2 had 5-10 introns, and TMK1, IMK2, TMKL1, MRLK, PRK2, FLS2, EFR, GSO2, SOBIR1, BAM1-3, IKU2, PEPR1/ 2, FLO5, CLV1, HSL1/2, MEE62, BRI1, PSKR1, BIR1, BRL1/3, VH1 and EXS had fewer than five introns and several genes had only one exon (Fig 3A and 3B, S3 Table). Similarly, previous studies have been shown that the introns influence gene expression through Intron Mediated Expression (IME) and Intron-Dependent Spatial Expression (IDSE) and a study on ERECTA gene given that multiple introns in specific location enhances its expression [43][44][45]. In general, position of a intron indicates its conservation and site of gene expression in a plant [42,43], so in this study we performed manual calibration and found that almost all of the BrLRR-RLK genes had a greater number of splicing phase 2 introns, indicating a highly prone site for mutation in the intron and much less conservation. Intron phase 1 is intermediate and was the second most common phase found among the genes, but phase 0, which indicates highly conserved introns, was rare among these genes [46]. The results of our study, number of introns and splicing phases can be used further as a resource for identifying the affect of introns on BrLRR-RLK genes in different tissues.

Conserved motif identification in B. rapa LRR-RLK genes
The 303 BrLRR-RLK genes were divided into different groups (I-XIV) (S3 Fig) based on the phylogeny [4,5,37] and 25 motifs were identified through the MEME online tool and used to annotate the domain types. In total, 10 LRR motifs (9-13, 15, 17-19 and 25) were found in the BrLRR-RLK genes, with a common motif pattern among most of the genes in a subfamily and variation among the subfamilies. All 11 protein kinase motifs (1-8, 14, 16 and 20) were found in most of the genes in subfamilies II, IV, V, VII, VIII-1, VIII-2, IX, X and XIII but the remaining families did not have all 11 kinase motifs, which may have been because of the kinase domain size (Fig 4). Some additional motifs were also identified. The DUF1191 motif (22) was Genomic Characterization of BrLRR-RLK Genes restricted to subfamilies VIII-2, VIII-1 and I. Similarly, the melactin motif (24) was identified in the genes of VIII-2 and one gene each in subfamilies VIII-1 and VI. Furthermore, unknown motifs (21,23) were observed in subfamilies III, IV, XI, VII, XIII and X; motif 21 was found in II, I, VI and XII, and motif 23 was in V, XIII and XIV (Fig 4 and S4 Fig). These results support our BrLRR-RLK subfamily classification based on the phylogeny and also suggest that genes with different sets of protein domains in their sequences may have subfamily-specific functions in plant growth and development [29,37,39,47]. In addition, the motifs had varying lengths (11-36 aa) and numbers of sites , which are shown in Table 2. In the case of LRR motif types, the 303 BrLRR-RLK genes mostly contained LRR_8, followed by the leucine-richrepeat N-terminal domain (LRRNT), LRR_6 and LRR_4, whereas the protein kinase motifs were mixed with protein kinase and protein kinase_Tyr (tyrosine kinase) [3,7] (Table 2).
In general, kinases are divided into two types, arginine-aspartate (RD) and non-arginineaspartate (non-RD), based on the conservation of the amino acid Arg (R) in their sub-domains [48]. In this study, we identified 171 RD type and 132 non-RD BrLRR-RLK genes based on the conservation of the 'RD' amino acids (S2 Table and S4 Table). Similarly, previous results suggest that the percentage of non-RD type kinases is low in Arabidopsis and rice, and, moreover, this type of kinase plays a crucial role in pathogen recognition activity [48]. The average number of motifs was greater in subfamilies VII, X, XI and XII, and a few genes in XIII, but there was no specific difference between RD and non-RD type kinases. Additionally, among the 303 BrLRR-RLK genes, only 234 genes contained a transmembrane domain, and regarding SPs, 276 genes contained one SP, whereas eight genes had two SPs, which were mostly found in subfamilies III and XI, and only one gene had three SPs (in subfamily III); the remaining eight genes lacked a SP (S2 Table).

Classification of B. rapa LRR-RLRR-RLK genes according to their functions
Gene ontology annotations for the 303 BrLRR-RLK genes were retrieved using the Blast2Go software under the biological process, molecular function and cellular component categories. In the biological process category, the BrLRR-RLK genes were involved in 918 different processes with the highest number (249 genes) involved in cellular processes, followed by cellular metabolic processes (239), and phosphorylation and phosphate-containing compound metabolic processes (232) (S3 Table). Similarly, previous reports published on Arabidopsis suggest that LRR-RLK genes play roles in different cellular processes in relation to organ development and phosphorylation [10,[49][50][51]. Furthermore, the BrLRR-RLK genes were associated with 74 different molecular functions, with the highest number of genes involved in transferase activity, transferring phosphorus-containing groups (268), kinase activity (267), binding (233), phosphotransferase activity, and 'alcohol group as acceptor and protein kinase activity' (241). Consistent with our observations, previous studies have reported the important roles of LRR-RLK genes in relation to metabolic activities [3,5,40,51]. In general, receptor kinases are important for cell-to-cell interactions, so many of these proteins are associated with the cell surface [5,37]. Here, the BrLRR-RLK genes were associated with 66 different cellular components including the membrane (251), 'cell and cell part' (221), and 'cell periphery and plasma membrane' (207) (S5 Table).

Differential expression analysis of B. rapa LRR-RLK genes
The expression abundance of a gene in different tissues can be used as a source of information for identifying its function in particular tissues; thus, expression data for the 303 BrLRR-RLK genes were extracted separately from a whole transcriptome dataset [21]. The genes were grouped according to subfamily, and a heat map was generated to analyze the expression variation among six different tissues of Chiifu-callus, root, stem, leaf, flower and silique (Fig 5). Furthermore, a few BrLRR-RLK genes were named after their A. thaliana homologs (S2 Table). Numerous expression variations were observed in the heat map of BrLRR-RLK genes so these genes were clustered into groups according to their expression levels. In addition to silique, the genes BrLRR-RLK34, 207 and 224 of subfamily VIII-2 had detectable differential expression in all the tissues whereas the remaining genes of this subfamily have failed to express in any of the tissues so the expression results can be depicted as may be they have functional roles during different developmental stages (Fig 5). The genes BrLRR-RLK245, 248 and 261 of subfamily VI and BrLRR-RLK26, 64, 212 and 258 of subfamily XI had a high expression level in all tissues except that BrLRR-RLK245 and 212 had moderate expression in the silique. Jones et al. [52] reported that MRH1 is mainly involved in root hair elongation during tip growth in Arabidopsis; similarly in our results, a homolog of MRH1, BrLRR-RLK12 of subfamily VI, had high expression in the root compared with other tissues. The genes of subfamily VIII-1 showed Table 2. The amino acid patterns and characteristics of the 25 motifs identified in the 303 BrLRR-RLK genes.

Motif
Sequence Length(aa) Sites Type  detectable to high expression abundance in all the tissues except for BrLRR-RLK249, which was not expressed in the flower or silique. Generally, SERK genes function in anther development, male gametophyte maturation and disease resistance, whereas NIK genes function in signaling during virus infection [8,53,54]. Here, the genes BrLRR-RLK95 (SERK4) and 168 of subfamily II were only expressed in siliques, BrLRR-RLK110 was not expressed in any tissue and most of the remaining genes of this subfamily had detectable expression variation among all the tissues. In subfamily I, the genes BrLRR-RLK84, 126 and 215 had moderate to high expression in all tissues but the remaining genes of this subfamily had undetectable or no expression in a few tissues; moreover, the MEE and RHS genes of this subfamily are involved in embryo and root hair development [55,56] (Fig 5). A low to moderate expression level was observed for all three genes of subfamily XIV; conversely, in subfamily IV, four genes had moderate to high expression levels in most of the tissues. Among the 13 BrLRR-RLK genes of subfamily V, most showed detectable transcript abundance in all six different tissues; additionally, their homologous genes SRF1-8 and SUB/ SRF9 of A. thaliana play major roles in leaf, stem, flower and inflorescence morphology and are also involved in cell wall biology, basic cell processes and abiotic stress [11,57]. Among the 52 genes of subfamily III, most had high expression levels in vegetative tissues like roots, stems and leaves compared with callus, flowers and siliques (Fig 5). Few homologs of the subfamily III genes have been published. Valon et al. [58] showed that TMKL1 is transcriptionally active in different organs and also regulates the development of silique maturation. Similarly, AtPRK2 is involved in pollen tube growth [59] and IMK functions in cell proliferation and cell fate determination [30].

Motif1 PI[V/I]HR[D/N][I/V/L]K[S/A/P]SN[I/V]LLDED[F/L]E[A/P]K[V/L/I][S/A]DFGL[A/S][K/R
In general, the subfamily XII genes EFR and FLS2 are up-regulated during biotic stress [41]. Accordingly, all 25 genes of subfamily XII had detectable expression or a complete absence of transcripts in most of the tissues. In subfamilies XI, VII, XIII and X, most of the genes showed high expression levels in roots, stems and leaves, and in addition, a few genes also showed moderate to high expression levels in callus, flowers and siliques (Fig 5). Their homologous genes are involved in different functions: CLV1, BAM1, BAM2 and BAM3 in cell differentiation and control of size in the shoot and floral meristem, GSO in normal development of the epidermal surface of embryos, HSL2 in flower abscission, TOAD2 and RPK1 in embryonic pattern formation, FEI2 in synthesis of cell wall components, ERL1 and ERL2 in promoting organ growth and flower development and BRL1, BRL3 and BRI1 in normal vascular differentiation in Arabidopsis shoots [10,49,[60][61][62][63][64][65]. Likewise, SOBIR1, PEPR1, PEPR2 and PSKR1 are mainly involved in defense responses, BRI1 and BAK1 in brassinosteroid signaling, and BIR1 in plant resistance [14][15][16][17]66]. These expression data may be a valuable resource for further investigation of the functional significance of the BrLRR-RLK genes in B. rapa.

Biochemical analysis of BrLRR-RLK genes
For biochemical analysis, we created 11 BrLRR-RLK constructs with the cytoplasmic domain, which consisted of a juxtamembrane domain, kinase domain and C-terminal region. Previous studies suggested that Arabidopsis BAK1 identifies and phosphorylates E. coli proteins in vitro and a wide range of cytoplasmic domains including the juxtamembrane domain, and the C-  terminal region regulates BAK1 kinase activity and specificity through auto-phosphorylation and trans-phosphorylation [67]. We transformed Flag-tagged recombinant protein expression constructs for LRR-RLK cytoplasmic domains into BL21 E. coli cells and then performed comparative biochemical assays with the recombinant proteins. The phosphorylation levels of the expressed recombinant proteins were assessed with phosphorylation-specific antibodies, antiphosphothreonine and anti-phosphotyrosine.
The expression levels of the 11 recombinant LRR-RLK proteins were detected with anti-Flag. As shown in Fig 6, seven LRR-RLKs showed auto-and trans-phosphorylation activity. Even though RD kinases have auto-phosphorylation ability, some of the RD kinases (BrPSKR1 and BrPSY1) in our analysis did not phosphorylate tyrosine and/or threonine residues. This might have been because of a lack of catalytic activity at the target kinase [68]. Coomassie brilliant blue (CBB) staining of the total recombinant E. coli crude extract showed the protein expression. Interestingly, BrBAK1 had strong auto-phosphorylation and trans-phosphorylation of tyrosine and threonine residues compared with AtBAK1 in previous studies [69,70]. These data suggest that BrBAK1 has much stronger kinase activity, which might be a very interesting result because the BrBAK1 signaling pathway is involved in plant growth and development, plant innate immunity, programmed cell death, and other novel mechanisms with a different pattern compared with AtBAK1. Another interesting result was that the highly accumulated BrBAK1, BrBRI1, BrPEPR1 and BrPEPR2 proteins showed activity with both antiphosphotyrosine and anti-phosphothreonine antibodies, indicating that they might have dualspecificity kinase activity.
To further confirm the auto-phosphorylation of the Flag-BrBRI1 receptor kinase, we induced the recombinant protein in E. coli cells and analyzed the protein over a time course of incubation after adding IPTG. As shown in Fig 7, Flag-BrBRI1 auto-phosphorylation showed a gradual increase from 6 to 16 h (approx. size 45 kDa and 150 kDa) and more E. coli proteins were phosphorylated by the Flag-BrBRI1 receptor kinase. Similarly, in anti-phosphothreonine analysis, auto-phosphorylation was observed from 4 h and showed the highest signal intensity at 16 h at around 50 kDa; in addition, several E. coli proteins were trans-phosphorylated. Compared with the phosphorylation of BrBRI-Flag with anti-phosphothreonine at 4 h, tyrosine auto-phosphorylation was observed at 6 h (Fig 7). These data suggest that specific threonine phosphosites were phosphorylated faster than specific tyrosine phosphosites on BrBRI1 in vitro. Similar to our results, previous research also proved that AtBRI1 kinase has the ability to phosphorylate at serine/threonine and tyrosine residues, which helps in different functions related to plant growth and development [71,72]. This work provides a comprehensive analysis of the broad role of Tyr and Thr phosphorylation for the function of B. rapa LRR-RLKs.

Conclusion
Our results for the BrLRR-RLK gene family suggest that duplications led to the expansion of the gene family. Comparative phylogenetic analysis allowed us to predict sequence similarities and identify the genes specific to each subfamily and the variations among monocot and dicot species (A. thaliana, B. rapa, O. sativa and P. trichocarpa). Motif and structural analysis showed motif conservation and intron/exon pattern variations among the subfamilies. In addition, gene ontology and expression analysis revealed various functional roles and tissue-specific expression variations. Finally, biochemical analysis demonstrated the type of receptor kinase activity occurring in cells. These results will be a major resource for studying the adaptive evolution, molecular mechanisms and functions of LRR-RLKs in the near future.