Phylogeny of the plant receptor-like kinase (RLK) gene family and expression analysis of wheat RLK genes in response to biotic and abiotic stresses

Background The receptor-like kinase (RLK) gene families in plants contains a large number of members. They are membrane proteins with an extracellular receptor domain and participate in biotic and abiotic stress responses. Results In this study, we identified RLKs in 15 representative plant genomes, including wheat, and classified them into 64 subfamilies by using four types of phylogenetic trees and HMM models. Conserved exon‒intron structures with conserved exon phases in the kinase domain were found in many RLK subfamilies from Physcomitrella patens to Triticum aestivum. Domain distributions of RLKs were also diagrammed. Collinearity events and tandem gene clusters suggested that polyploidization and tandem duplication events contributed to the member expansions of T. aestivum RLKs. Global expression pattern analysis was performed by using public transcriptome data. These analyses were involved in T. aestivum, Aegilops tauschii and Brachypodium distachyon RLKs under biotic and abiotic stresses. We also selected 9 RLKs to validate the transcriptome prediction by using qRT‒PCR under drought treatment and with Fusarium graminearum infection. The expression trends of these 9 wheat RLKs from public transcriptome data were consistent with the results of qRT‒PCR, indicating that they might be stress response genes under drought or F. graminearum treatments. Conclusion In this study, we identified, classified, evolved, and expressed RLKs in wheat and related plants. Thus, our results will provide insights into the evolutionary history and molecular mechanisms of wheat RLKs. Supplementary Information The online version contains supplementary material available at 10.1186/s12864-023-09303-7.


Introduction
Receptor-like kinases (RLKs) are membrane proteins with an extracellular receptor domain, such as leucine rich repeats (LRRs), lectin (Lec), lysine motif (LysM) or wall associated kinases (WAK) [1]. RLKs play important roles in resistance signalling pathways of various abiotic stresses, including drought, high temperature and low temperature. They can also defend against pathogen infection by activating immune signalling in plants. RLK gene families from various plants have been identified in a large number of articles. We summarized these findings in the following Introduction section.
(1) Whole RLKs: In 2003 and 2004, Shiu et al. identified more than 600 RLKs in Arabidopsis thaliana and more than 1200 RLKs in Oryza sativa [2,3]. They play important roles in plant growth, development, and defence responses to stresses. More than 440 RLKs from O. sativa might have originated from domain fusion events after the split of rice and Arabidopsis in evolution. In 2009, Lehti-Shiu et al. found that the expansion of RLK members coincided with the establishment of land plants [4]. In 2018, Lin et al. identified 563 RLK genes in Jilin ginseng (Panax ginseng C.A. Meyer) and analysed their evolution, functional diversity and coexpression networks [5].
(2) LRR-RLKs (subfamily of RLK): In 2013, Zan et al. identified 379 LRR-type RLK genes in Populus trichocarpa. A total of 312 PtLRR-RLK genes out of 379 are located in segmental duplication blocks. Genomewide analysis of microarray data showed that some PtLRR-RLKs responded to shoot organogenesis, low ammonium feeding, wounding, hypoxia and seasonal dormancy [6]. In 2017, Liu et al. studied the origin and diversity of LRR-RLKs in plants and found that most LRR-RLKs were established in early land plants [7]. In 2018, Sun et al. identified 1641 LRR-RLK genes in four Gossypium species (Gossypium arboreum, Gossypium barbadense, Gossypium hirsutum, and Gossypium raimondii). Tandem duplication played an important role in the expansion of the Gossypium LRR-RLK gene family. Expression pattern analysis showed that Gossypium LRR-RLKs were widely involved in various stress defences and diverse developmental processes [8]. In 2020, Meng et al. identified 329 LRR-RLK genes in Medicago truncatula. Analysis of classification, duplication events, exon/intron organization, and expression profiling were performed in M. truncatula LRR-RLKs [9]. In 2021, 437 LRR-RLK genes were identified in Saccharum spontaneum and categorized into 14 groups. Analysis of promoter sequences and expression profiles showed that SsLRR-RLKs were strongly regulated by various environmental stimuli, transcription factors and phytohormonal factors, suggesting that they respond to various diverse biotic and abiotic stresses [10]. In 2022, Song et al. identified 444 BnLRR-RLKs in the Brassica napus cultivar "Zhongshuang 11" and classified them into 22 subfamilies. Based on CRISPR/Cas9 technology, they obtained six partial knockouts of BnBRI1 homologues to generate semidwarf lines without decreased yield compared with controls [11]. In 2022, 15 TaRPK1 (receptor-like protein kinase 1, a calcium-independent serine-threonine kinase that belongs to the subfamily LRR-RLK) genes were identified in Triticum aestivum. Eighteen putative miRNA targeting and cis-regulatory elements (light-related, hormone responsiveness, and stress elements) were identified in TaRPK1 genes. In silico expression analysis and qRT-PCR validated that TaRPK1 genes exhibited higher expression in the roots of drought-tolerant varieties than in drought-susceptible varieties [12].
(3) LecRLKs (subfamily of RLK): In 2020, 46 putative lectin receptor-like kinases were identified in cucumber (Cucumis sativus L.) genome and were classified into three groups, including 23 G-type, 22 L-type, and one C-type CsLecRLK genes. Analysis of promoter regulatory elements and expression patterns revealed that some CsLecRLKs were associated with phytohormones and stress responses [13]. In 2020, Singh et al. identified 73 putative VrLecRLKs in mung bean (Vigna radiata L. Wilczek) and classified them into three families: G-type, L-type, and C-type VrLecRLKs [14]. In 2021, 1311 AhR-LKs, including AhLRR-RLKs and AhLec-RLKs, were identified from the peanut (Arachis hypogaea) genome. The results of mining transcriptome data showed that 14 of 90 Al-responsive AhRLKs were expressed specifically in root tissue [15]. These CRKs were tightly clustered into four phylogenetic groups and were variably conserved in exons/introns, domains and motifs. Tissue-specific expression analysis suggested that some CRK genes are involved in plant development [19].
(6) PERKs (proline-rich extensin-like receptor kinases, subfamily of RLK): In 2004, 15 AtPERKs were predicted in A. thaliana, and some AtPERK members were identified as tissue-specific genes [20]. In 2022, 37 TaPERKs were identified in wheat (T. aestivum L.) and were classified into eight well-defined groups. Analysis of cis-acting regulatory elements and expression profiles revealed that TaPERKs may respond to phytohormones and various biotic and abiotic stresses [21].
Some articles have reported that RLKs play important roles in the response to biotic and abiotic stresses, such as drought, heat, salinity and cold. Overexpression of PdERECTA (an LRR-type RLK from Populus deltoides) improves water use efficiency and enhances drought resistance in transgenic Arabidopsis plants [22]. Rice (O. sativa) OsSIK1 (LRR RLK) improves tolerance to drought and salt stress. Transgenic rice plants overexpressing OsSIK1 exhibit enhanced tolerance to salt and drought stresses, while knock-out and RNA interference plants exhibit sensitivity to drought and salt stresses [23]. OsLecRLK overexpression and downregulation (through artificial miRNA) transgenic lines showed that rice OsLecRLK enhances salinity tolerance through ion homeostasis [24]. The interaction network of 255 Arabidopsis LRR-RLKs (567 pairs of interaction relationships) was established by using a sensitized high-throughput interaction assay. Plants have evolved LRR-RLK networks to process extracellular signals into cells, functioning in plant growth and immunity [25]. MtDMI2 (a Leu-rich repeat-type receptor kinase) and MtPUB2 (a novel plant U-box (PUB)-type E3 ligase) interact to form a negative feedback loop, playing an important role in nodulation homeostasis [26]. The systemin receptor SYR1 (an LRR-RLK) of tomato is not decisive for local and systemic wound responses but is important for defence against insect herbivory [27]. The A. thaliana lectin RLK AtLecRK-IX.2 can modify the pathogen effector AvrPtoB to dampen its virulence in Arabidopsis [28]. By using a map-based cloning strategy, Duriez et al. identified the sunflower protein HaOr7 (LRR-RLK) as an effector that confers resistance to Orobanche cumana [29]. Rao et al. constructed nine higher-order mutants of A. thaliana receptor-like cytoplasmic kinase (RLCK) subfamily VII, revealing that numerous RLCK VII members are involved in plant development and pattern-triggered immune signalling [30]. A. thaliana AtRIPK (RPM1-INDUCED PROTEIN KINASE), an RLCK VII subfamily member, contributes to ROS (reactive oxygen species) production in the plant immune system [31].
In recent years, the biological functions of wheat RLKs mediating the response to biotic stress have been reported. In 2016, Rajaraman et al. found that the barley LRR-type RLK gene HvLEMK1 was a factor mediating nonhost resistance in barley and quantitative host resistance in wheat to the wheat powdery mildew fungus [32]. In 2018, Saintenac et al. discovered that the RLK gene Stb6 (a conserved wall-associated receptor kinase (WAK)-like protein, subfamily of RLK) in wheat is a natural resistance gene to the fungal pathogen Zymoseptoria tritici [33]. In 2019, Wang et al. found that the LRRtype RLK gene TaXa21 in wheat is highly homologous to the rice bacterial blight resistance gene Xa21. They also found that TaXa21 is a positive regulator of wheat high-temperature seedling plant (HTSP) resistance to Puccinia striiformis f. sp. tritici. This process is mediated by H 2 O 2 and ethylene (ET) signalling pathways and is associated with the transcription factors TaWRKY76 and TaWRKY62 [34]. In 2020, Gu et al. discovered a novel cysteine-rich RLK gene, TaCRK2, which positively regulates leaf rust resistance in wheat [35]. Using comparative genomics, mutagenesis and complementation, Saintenac et al. identified a wheat cysteine-rich RLK gene, Stb16q, which exhibited resistance against Septoria tritici blotch (Stb, pathogen Zymoseptoria tritici) and localized at the plasma membrane in the infection cycle [36]. In 2021, Guo et al. identified a novel CRK RLK gene, TaCRK3, that could defend against Rhizoctonia cerealis in wheat. The TaCRK3 protein contains two DUF26 (DOMAIN OF UNKNOWN FUNCTION 26) domains that can inhibit the growth of R. cerealis mycelia [37]. In 2021, the wheat wall-associated receptor-like kinases (WAKs, subfamily of RLKs), TaWAK-6D and TaWAK7D, were identified to mediate broad resistance to fungal pathogens (Fusarium pseudograminearum and R. cerealis) [38,39]. In 2022, TaPsIPK1 (a wheat receptor-like cytoplasmic kinase gene) was identified as a susceptibility gene for stripe rust (caused by P. striiformis f. sp. tritici) effectors [40].
In this study, we performed genome-wide identification, classification, and evolutionary analysis of the RLK gene family in 15 representative plants, including four wheat and Aegilops tauschii species. Global expression analyses of stresses were performed in individual T. aestivum RLKs. qRT-PCR procedures of 9 selected RLKs were performed to validate the prediction of the transcriptome under drought conditions and Fusarium graminearum infection. Our results provide insights into the evolutionary history and molecular mechanisms of wheat RLKs.

Genome-wide identification and classification of the RLK gene family in wheat, Ae. tauschii and other plants
We identified RLK genes with typical kinase domains and corresponding HMM models (see Methods) in 15 representative plants, including four wheat and Ae. tauschii (Table 1, Table S1). Other PKs (protein kinases) with typical kinase domains were also identified and classified in 15 representative plants (Table S1). The results showed that most proportions of RLKs in PKs were 75-78% in 15 representative plants (Table 1).
We classified the RLKs into 64 subfamilies by HMM models (Table S2). We only selected 1-3 members from every subfamily as the representative sequences to construct phylogenetic trees. To confirm the classification from HMM models, four types of phylogenetic trees (including Bayesian tree, maximum likelihood (ML) tree, neighbour-joining (NJ) tree with JTT model, and NJ tree with p-distance model) were constructed based on the truncated kinase domain sequences ( Fig. 1, Fig. S1A-D and Table S3). The results showed that almost all the classifications of RLK subfamilies from HMM models and four phylogenetic trees were the same (Table S3). Among 64 RLK subfamilies, we noticed that all RLK-Pelle_RLCK-IXb clades from Bayesian and NJ phylogenetic trees contained a green algae sequence (PNW75571), suggesting that the RLK-Pelle_RLCK-IXb clade might be the ancestral subfamily of 64 RLK subfamilies. The HMM scan results showed that both green algae sequences (PNW75571 and PNW75214) belonged to the RLK-Pelle_RLCK-IXb clade.

Evolution and conserved exon-intron structures of RLK gene subfamilies
To obtain further insight into RLK evolution, we diagrammed the exon-intron structures within the kinase domain in the 15 investigated plants (Fig. S2). The results showed that some conserved exon-intron structures were present in the same RLK subfamilies across the investigated plants, especially in the kinase domain. We summarized these conserved exon-intron structures in six representative plants, including T. aestivum, B. distachyon, Vitis vinifera, Amborella trichopoda, Selaginella moellendorffii, and Physcomitrella patens (Fig. S3). For instance, a conserved exon-intron structure with exon phase "0112-0" existed in the RLK-Pelle_LRR-I-1 subfamily from P. patens to T. aestivum (Fig. 2). A similar example within the "0112-0" exon-intron structure was also found in the RLK-Pelle_ RLCK-IXa subfamily (Fig. 2).

Chromosome location and duplication events of wheat RLKs
The chromosome locations of T. aestivum RLKs were mapped on 21 chromosomes ( Fig. S5 and Table S4). The distribution of T. aestivum RLKs was among the A, B and D subgenomes.
To study the whole genome duplication (WGD) events of wheat RLKs, we identified 2114 RLK gene pairs related to collinearity events by MCscanX (Table S5). Most Ks values of these wheat RLK collinearity events ranged from 0 to 0.35 and formed a peak of Ks at 0-0.15 (Fig.  S6). These collinearity events of wheat RLKs were visualized on 21 chromosomes ( Fig. 3A-C). The results showed that collinearity events within the Ks values of 0-0.35 were mainly located among the corresponding subgenomes, such as between 1A, 1B and 1D, suggesting that these collinearity events occurred along the polyploidization of T. aestivum. Moreover, collinearity events within Ks > 0.35 were mainly located across the corresponding subgenomes.
A total of 232 clusters of tandem duplications (TDs) were identified in T. aestivum RLKs ( Fig. S7 and Table 2). Chromosomes 2D, 3B and 5D contained more than 15 clusters, which were 19, 17 and 18, respectively (Table  S6). Some of these clusters belonged to the RLK-Pelle_ DLSV subfamily. For example, a cluster on chromosome 5D within 4 genes (TraesCS5D02G357700.1, TraesC-S5D02G357800.1, TraesCS5D02G358600.1 and TraesC-S5D02G358700.1) belonged to the RLK-Pelle_DLSV subfamily. The largest cluster within 14 members was an RLK-Pelle_DLSV cluster, which was at the end of chromosome 2A. The second largest cluster within 13 members was also an RLK-Pelle_DLSV cluster, which was at the end of chromosome 7B.
To further study the mechanism of RLK duplication events, comparative syntenic maps of T. aestivum, B. distachyon and O. sativa RLKs were constructed ( Fig. 4A-D). A total of 1081 RLK syntenic gene pairs were detected between T. aestivum and B. distachyon. Similarly, 961 RLK syntenic gene pairs were detected between T. aestivum and O. sativa (Table S7). Some RLK syntenic gene pairs shared the same T. aestivum RLK member associated with B. distachyon and O. sativa, suggesting that they might have descended from a single common ancestral sequence before the Graminaceae split in evolution. For instance, the Tae-Bid gene pair (TraesC-S7D02G241500.1 and KQJ98873) and Tae-Osa gene pair (TraesCS7D02G241500.1 and Os08t0501200-00) shared the T. aestivum RLK gene (TraesCS7D02G241500.1), and all three of these genes belonged to the RLK-Pelle_WAK subfamily. We also calculated the Ks values between the Tae-Bid and Tae-Osa syntenic gene pairs (Table S7 and

Expression patterns of T. aestivum, Ae. tauschii and B. distachyon RLKs under drought stress
We studied the expression patterns of T. aestivum, Ae. tauschii and B. distachyon RLKs under biotic and abiotic stresses by using public transcriptome data at NCBI. According to the quality control performed by FastQC software, two transcriptome samples were excluded in the following analysis (Table S8). To search the expression pattern of wheat RLKs under drought stress, three public transcriptome datasets of wheat were selected for study ( 111" and "TAM 112" (grown in southern America) have excellent drought tolerance. This RNA-seq analysis was conducted to compare gene expression differences in flag leaves of "TAM 111" and "TAM 112" under wet and dry conditions. We extracted the expression patterns of RLKs from this RNA-seq analysis. These results showed that some RLKs exhibited different expression patterns between "TAM 111" and "TAM 112", suggesting that different drought-tolerance mechanisms existed in "TAM 111" and "TAM 112". For example, TraesCS5B02G059000 (RLK-Pelle_DLSV) from the TAM112_Heading and TAM112_GrainFilling samples exhibited downregulation (log 2 fold change, -1.71 and -2.07), while TraesCS5B02G059000 (RLK-Pelle_DLSV) from the TAM111_Heading and TAM111_GrainFilling samples exhibited upregulation (1.35 and 1.52). (2) "Svevo" and "IL20-2" (Bioproject: 686121): Two wheat genotypes, "Svevo" and "IL20-2", were treated under well-watered and water-limited conditions. Then, the wheat seedling root tissues of "Svevo" and "IL20-2" underwent RNA extraction to perform transcriptome analysis. We studied the expression patterns of RLKs from this RNA-seq analysis. We noticed that some RLKs exhibited the same expression trends between "Svevo" and "IL20-2". For instance, the expression trends of TraesCS2B02G008400 (RLK-Pelle_L-LEC) were downregulated with log 2 FC values of -1.57 and -1.99 in "IL20-2" and "Svevo", respectively. (3) "L-82" and "Marvdasht" (Bioproject: 450487): Two wheat genotypes, "L-82" (drought-tolerant) and "Marvdasht" (drought-sensitive), were treated under well-watered and drought conditions. Then, RNA was extracted from the wheat root tissues of "L-82" and "Marvdasht" to perform transcriptome analysis. These results showed that some RLKs exhibited different expression patterns between "L-82" and "Marvdasht", suggesting that different molecular mechanisms Drought is an abiotic stress that seriously affects wheat yield and quality. In our previous study, we found that some RLKs are involved in the drought response by bioinformatics analysis (Fig. 5 of article [41]). To further explore the expression patterns of RLK genes in the early stage of drought stress, we selected six wheat RLKs to examine their expression patterns under PEG (drought) treatment for 0, 3, 6, 12 and 24 h (hours) by using qRT-PCR (Fig. 5A-B). We only compared the expression patterns between the public transcriptome ("TAM 111" and "TAM 112", Bioproject: 659916) and qRT-PCR because all of their sample tissues were from wheat leaves (the other two datasets were from root tissues). qRT-PCR experiments also determined the expression patterns of RLKs from public transcriptome data under drought stress. The results showed that the expression trends were almost the same between the public transcriptome and qRT-PCR. For example, the log 2 values of TraesC-S7D02G355800 (RLK-Pelle_LRR-XII-1) were 2.93, 1.50, 1.50 and 3.97 in the TAM112_Heading, TAM112_Grain-Filling, TAM111_Heading and TAM111_GrainFilling samples, respectively. Our qRT-PCR results validated that the expression trend of TraesCS7D02G355800 was the same as the transcriptome (Bioproject: 659916), exhibiting a peak within more than 1000-fold upregulation at 24 h.

Expression patterns of T. aestivum RLKs under other abiotic stresses
We also studied the expression patterns of T. aestivum RLKs under other abiotic stresses by using public transcriptome data at NCBI (Fig. S11A-C and Table S11). There were differences in waterlogging tolerance among different wheat varieties. Among the three investigated wheat varieties, the seeds of "Bainong 607" germinated earlier than those of "Bainong 207" and "Zhoumai 22" under waterlogging stress. Some T. aestivum RLKs exhibited upregulation in "Bainong 607" but downregulation in "Bainong 207" and "Zhoumai 22", suggesting that different molecular mechanisms of waterlogging tolerance exist in different wheat varieties. For instance, the

Expression patterns of T. aestivum RLKs under biotic stresses
We studied the expression patterns of T. aestivum RLKs under various abiotic stresses by using public transcriptome data at NCBI (Fig. S12A-D   To further determine the expression patterns of public transcriptome data under Fusarium head blight (FHB) stress, we selected three wheat RLKs to examine their expression patterns with F. graminearum treatments by using qRT-PCR (Fig. 6A-B). The expression trends of three wheat RLKs from public transcriptome data were consistent with the results of qRT-PCR. For instance, the log 2 values of TraesCS1B02G454000 (RLK-Pelle_DLSV) in public transcriptome data were almost all upregulated to form a peak in NIL38 (0.48, -0.38, 0.50, 1.62, 1.00 and
Among 64 RLK subfamilies, the members of some T. aestivum RLK subfamilies, such as RLK-Pelle_DLSV (829), RLK-Pelle_L-LEC (320), RLK-Pelle_LRR-XII-1 (337), RLK-Pelle_SD-2b (331), RLK-Pelle_WAK (385) and RLK-Pelle_LRR-XI-1 (237), were much larger than others, suggesting that certain RLK subfamilies experienced expansion during evolution (Table S2). This was consistent with the result of the 2009 article that RLKs had extensively expanding subfamilies, including DUF26, LRK10L-2, LRR-I, LRR-XII, SD1, SD-2b, and WAK [4]. The allohexaploid bread wheat (T. aestivum) genome contained three closely related subgenomes (A, B, and D). The A and B genomes diverged from a common ancestor approximately 7 million years ago, and the D genome diverged through homoploid hybrid speciation 5-6 million years ago. The bread wheat genome has experienced multiple rounds of hybrid speciation [44]. Our results showed that almost all RLKs were homologous sequences among the subgenome A, B and D chromosomes (Fig. 3). The Ks values of these RLK collinearity events were 0-0.35, suggesting that they were associated with polyploidization events among the wheat A, B and D subgenomes (Table S5 and Fig. S6). It has been reported that polyploidy, tandem duplications, segmental duplications and transposition events are the main mechanisms for the expansion of the wheat expansin gene family [45]. Consistent with their results, we also found some tandem RLK clusters on wheat chromosomes, suggesting that tandem duplication events also contributed to the expansion of RLK members during evolution (Fig. S7).
There are several comprehensive reports about the genome-wide identification and classification of the RLK gene family in many plants [2][3][4][5]. However, our results were more comprehensive. We studied the exon-intron structures of RLKs and found some conserved exonintron structures in evolution. In addition, our work included five different varieties of wheat (T. aestivum, Triticum spelta, Triticum turgidum, Triticum dicoccoides, Triticum urartu), which are diploid, tetraploid and hexaploid wheat, so our results will provide more detailed information on RLK evolution in wheat. Moreover, we studied the expression patterns of wheat RLKs under biotic and abiotic stresses by using public transcriptome data and qRT-PCR. This study will provide candidate RLKs of drought and F. graminearum stresses for researchers.

Expression pattern of wheat RLKs under drought and F. graminearum stresses
We studied the RLK expression patterns under drought stress by using public transcriptome data and qRT-PCR. (1) DLSV-RLK: In 2021, Arabidopsis cysteine-rich receptor-like protein kinase AtCRK33 was found to affect drought tolerance and stomatal density. CRKs contain the DUF26 (Domain of Unknown Function 26) domain [47]. The DUF26 domain, also known as the stress-antifungal domain (PF01657, Pfam domain name, Stress-antifungal Family) on the Pfam website [48], could be found in the RLK − Pelle_DLSV subfamily in our results (Fig. S4). We also detected the expression pattern of an RLK − Pelle_ DLSV member (TraesCS6A02G390300) under drought stress by using public transcriptome data and qRT-PCR (Fig. 5). The qRT-PCR results showed that it exhibited a peak of upregulation at 3 h. The log 2 value of TraesC-S6A02G390300 in the TAM111_GrainFilling (c-w) sample was 4.29. (2) LRR-RLK: In 2011, it was reported that overexpression of PdERECTA (Populus deltoides LRR-RLK gene) in Arabidopsis enhances drought resistance [22]. In 2021, Li et al. discovered that overexpression of the PdERECTA gene in poplar improved water use efficiency and enhanced drought tolerance by reducing stomatal density and restricting water consumption [49]. A similar result was also reported in which overexpression of the Sorghum bicolor gene SbERECTA in Arabidopsis and maize enhanced their drought tolerance [50]. In our results, we also checked the expression pattern of two LRR-type RLKs (RLK-Pelle_LRR-VI-1 TraesC-S5D02G437000 and RLK-Pelle_LRR-XII-1 TraesC-S7D02G355800) under drought stress by using public transcriptome data and qRT-PCR (Fig. 5). The qRT-PCR results showed that they all exhibited a peak of upregulation at 24 h. Their log 2 values in four samples were all upregulated. (3) SD-2b: In 2020, an S-domain RLK gene OsESG1 (LOC_Os01g12410) in rice (O. sativa) was identified in early crown root development and drought response by controlling auxin response and distribution [51]. We obtained the protein sequence (LOC_ Os01g12410, Os01t0223800-01) and converted the ID on the website (https:// rapdb. dna. affrc. go. jp/). In our results, it belonged to the RLK subfamily RLK-Pelle_ SD-2b (Table S1).
Some articles have studied the relationships between F. graminearum infection and RLKs in wheat and other plants. After F. graminearum infection, Manes et al. screened 227 RLKs and innate immune response genes in Arabidopsis and identified nine genes (including RLK7) that play roles in F. graminearum resistance [52]. An LRR-RLK gene (GRMZM2G132212) in maize was identified as a defence or recognition gene in the response to fungal pathogens (Cochliobolus heterostrophus and F. graminearum). However, F. graminearum might be able to exploit this LRR-RLK gene (GRMZM2G132212) function to increase its virulence [53]. We converted ID (GRMZM2G132212, Zm00001eb293660) in NCBI and Ensembl plant and found that it belonged to the RLK subfamily RLK-Pelle_LRR-XI-1 in our results (Table S1). Two LRR-RLK genes (HvLRRK-6H and TaLRRK-6D) were found to contribute to Fusarium resistance in cereals (H. vulgare and T. aestivum) [54]. In our results, we also checked the expression pattern of two DLSV-type RLKs (RLK-Pelle_DLSV TraesCS1B02G454000 and RLK-Pelle_ DLSV TraesCS3D02G097000) with F. graminearum infection by using public transcriptome data and qRT-PCR (Fig. 6). qRT-PCR showed that they all exhibited a peak of upregulation at 48 and 96 h. The log 2 values of these genes in some NIL38 and NIL51 samples (24, 36 and 48 h) were all upregulated. They might be new resistance genes to defend against F. graminearum infection. Indeed, it was reported that a novel CRK RLK (DLSV-RLK subfamily in our results) gene, TaCRK3, could defend against another fungal pathogen, R. cerealis, in wheat [37]. org/) in batch mode with an E value of 0.01. Atypical PKs with kinase (PK_Tyr_Ser-Thr or Pkinase) domains covering less than 50% alignment were excluded from the following analysis. Classifications of "typical" sequences of PK subfamilies were performed by HMMER 3.1 with HMM models developed by Legti-Shiu and Shui [55].

Identification and classification of RLKs in plants
We selected 1-3 members as the representative sequences from every RLK subfamily to construct the phylogenetic trees. Each RLK subfamily was selected by the following criteria: members ≤ 6, 1 RLK; 6 < members ≤ 30, 2 RLKs; members > 30, 3 RLKs. The alignment of truncated RLK sequences in the kinase (PK_Tyr_Ser-Thr or Pkinase) domain was performed by ClustalW v2.0 [56]. A Bayesian phylogenetic tree was constructed using MrBayes v3.2.7 [57] with the mixed amino acid substitution model, and an MCMC chain with 10,000,000 generations was used. Markov chains were sampled every 100 generations, and the first 25% of the trees were discarded as burn-in. The results of MrBayes v3.2.7 were analysed by TreeGraph v2.14 [58] and our Perl scripts. The ML phylogenetic tree was constructed using PhyML v3.1 [59] with 100 bootstrap replicates. The appropriate model of the ML method, including model parameters, was calculated using the Akaike information criterion (AIC) with ProtTest v3.4 [60]. The NJ phylogenetic trees were constructed by Megacc 7.0 [61] with a model (p-distance or JTT) and 1000 bootstrap repetitions. The four types of phylogenetic trees were constructed by the above descriptions in our local server.

Domain and intron-exon structure diagram of RLKs
The domain and intron-exon structures of RLK sequences in these fifteen plants were generated by our Perl and R scripts based on the corresponding GFF file information from Ensembl Plant release-51 (http:// plants. ensem bl. org/). The domain information of pfam-A models was downloaded from pfam 34.0 (http:// pfam. xfam. org/) and then scanned in our local server.

Chromosome locations, duplication events and synthetic analysis of wheat RLKs
Based on the extracted information in the GFF files from Ensembl Plants release-51 (http:// plants. ensem bl. org/), the chromosome locations of T. aestivum RLKs were diagrammed using GenomePixelizer software [62]. BLASTP was performed against RLKs of T. aestivum, B. distachyon and O. sativa with an E value of e-100. Based on the GFF files and BLAST results, tandem duplication and segmental duplication were searched using MCScanX [63]. The Ka and Ks values were calculated by "add_ka_and_ks_to_ collinearity.pl" from MCScanX. Based on the GFF files and MCScanX results, synthetic diagrams among T. aestivum, B. distachyon and O. sativa were generated by using our Perl scripts and Circos software (http:// circos. ca/). The chromosome locations of the tandem duplicate RLKs were mapped on each chromosome by using Mapchart v2.3 (http:// www. wagen ingen ur. nl/ en/ show/ mapch art. htm).

Plant materials and stress treatments
Wheat (T. aestivum L.) cultivar "SuMai 3" was used in this study. The wheat seedlings were planted into pots and grown at 22-25 °C with a photoperiods of 16 h of light and 8 h of darkness. The seedlings of wheat (T. aestivum L.) cultivar "SuMai 3" at the three-leaf and one-heart stage were drought stressed, which including 20% (m/V) PEG-6000 for 0, 3, 6, 12 and 24 h. The wheat leaves were harvested and immediately frozen in liquid nitrogen for expression analysis. At least 15 samples of each experimental replicate were analysed. The spikes of "SuMai 3" were inoculated with 10 μL of an F. graminearum (PH1-1) conidia suspension (5-10 × 10 4 conidia mL −1 ). Then, the inoculated spikes were sealed in plastic bags to retain moisture for 72 h. The inoculated wheat spikes were collected after 0, 12, 24, 48, 72 and 96 h, and used to perform qRT-PCR analysis.

RNA Extraction and qRT-PCR
The total RNA of each sample was extracted using the RNAprep Pure Plant Kit (Tiangen) and reverse transcribed into cDNA using the HiScript III 1st strand cDNA synthesis kit (Vazyme). The cDNA samples were used for qRT-PCR analysis. qRT-PCR analysis was performed using a Roche LightCycler ® 480 (Roche Diagnostics GmbH, Mannheim, Germany). The wheat gene β-Actin was used as an endogenous control. Relative expression levels of genes were calculated using the Formula 2 −ΔΔCT . Each experiment included three biological and technical replicates. All the qRT-PCR primers in this study are provided in Table S13. The melting curves of qRT-PCR samples are provided in Figure S13.