Wheat coding sequence (CDS), and functional annotations of wheat genes were downloaded from IWGSC archive v.1.1 (https://urgi.versailles.inra.fr/download/iwgsc/IWGSC_RefSeq_Annotations/v1.1/). Functional annotations were filtered for Protein family database (Pfam) identifiers of the LysM domain (PF01476). The proteins with at least one LysM domain were selected as family member. A total of 65 sequences were selected. Splice variants were excluded and only the longest variant was kept for further analysis. Transmembrane helices were predicted by TMHMM Server v 2.0 [30]. The subcellular localization of these proteins was predicted by TargetP-2.0 Server (http://www.cbs.dtu.dk/services/TargetP/) [31]. The pI and MW were calculated by Expasy’s ProtParam (https://web.expasy.org/protparam/). The conserved domain was predicted by PFAM (http://pfam.xfam.org/) [32].

Protein sequences of A. thaliana, O. sativa, Brachypodium distachyon (L.) P. Beauv. and T. aestivum were aligned by MUSCLE. The phylogenetic tree was constructed by using neighbor-joining and maximum likelihood method with default parameters in Mega 7.0 [33].

We suggest a consistent naming pattern for all LysM genes, considering of their subgroup, phylogenetic relationships, motif contained as well as their subgenome location (A, B or D). Each gene name starts with an abbreviation of T. aestivum (Ta).

The genome annotation information of wheat was used to analyzed the chromosomal localization. The local blast searches of wheat and itself was conducted for considerable pairs of homologous genes. Then TBtools was employed to perform synteny analysis. Ka, Ks, and Ka/Ks values of wheat LYP gene pairs were calculated [34].

Expression level was downloaded from Wheat Expression Browser (http://www.wheat-expression.com/). We considered a gene expressed when its average expression per treatment was >0.5 tpm in at least one treatment. A heatmap was generated by R studio. The expression data of development was calculated as log2 [(transcript per million)+1] [35]. Genes were clustered according to their expression using K-means. The expression data of biotic and abiotic treatment was normalized by control and calculated as log2 (relative amount of expression). The expression data of biotic treatment contained powdery mildew, stripe rust and wheat head scab.

The relative expression levels of the A, B and D subgenome were analyzed only when the gene triads across the three subgenomes comply with 1:1:1. To standardize the relative expression of each homolog across the triad, we normalized the absolute tpm to the expression ratio of three subgenome. The tenary diagram was plotted by R package ggtern [36].

For qRT-PCR, wheat leaves inoculated with Puccinia triticina (race PHJ), the causal agent of wheat leaf rust, were harvested at 0, 6,18,48,168 h post-inoculation (hpi) for detection of the transcript levels of selected wheat LYP genes. The infected leaves were ground in liquid nitrogen, and RNA was isolated using MiniBEST Plant RNA Extraction Kit (TaKaRa) following the manufacturer’s instructions. qRT-PCR was carried out using Bio-Rad CFX 96, and the 2(^−ΔΔCt) analysis method was used to determine the relative expression levels of selected genes. Wheat GADPH (GenBank No. AF251217) was used as an internal reference gene [37]. Three independent biological replicates were performed per treatment. Primers were listed in Appendix A.

A total of 65 coding sequences were identified by HMM search using HMM profile (PF01476) in IWGSC wheat genome (Appendix B). The dataset was simplified by retaining 62 coding genes with only one splice variant from each genomic locus for further analysis (Table 1).

In order to classify the LYP genes, phylogenetic analyses were performed and the domains contained in these LYPs were considered simultaneously (Appendix B). Wheat LYP proteins were separated into six major groups: LYP, LYK, LysMn, EMSA, LysMe and LysMFbox. All the proteins have at least one LysM domain. In addition to LysM domain, LYP family contain an another LysM domain. Protein kinase domain (Pfam ID: PF00069) or protein tyrosine and serine/threonine kinase domain (Pfam ID: PF07714) were the characteristic for LYK family. LysMn family was predicted to be intracellular protein. EMSA family members were homologs of OsEMSA1. LysMe family contained proteins with only one LysM domain and have signal peptide. F-box domain (Pfam ID: PF00646) was the characteristic of LysMFbox family.

The characteristics of the wheat LYPs are shown in Table 1. The lengths of LYP protein sequences ranged from 100 to 749 amino acids and the molecular weights were 10.31 to 80.56 kD. Protein isoelectric points (PI) ranged from 4.57 to 9.29. The majority of the LYPs were predicted as secreted protein. The grand average of hydropathy (GRAVY) for LysMe and most of LYP subgroup was positive indicating hydrophobic character, while that of LysMFbox, EMSA, LysMn and most of LYK subgroup as negative indicating hydrophilic character.

To analyze the phylogenetic relationships of LYPs from different species, lysin motif contained protein sequences from A. thaliana, O. sativa, B. distachyon and T. aestivum were used to construct a neighbor-joining tree. As shown in Fig. 1A, wheat LYPs shared high homology with that from other species. Regardless of species, LYK was the largest group.

The amount of group numbers was calculated for each species. Rice, Arabidopsis and Brachypodium, despite their phylogenetic distance, have a similar number of LYP genes (15, 11 and 13, respectively) (Figs. 1C–1E). However, the number of LYPs in wheat is as high as 62 (Fig. 1B). The number of LYP genes was 4-fold than those in A. thaliana, O. sativa and B. distachyon (Fig. 1F). One of the main reasons is that wheat is hexaploidy which the number of genes is theoretically three times that of other diploid species. When corrected for ploid level, gene expanding in LysMe, LYK and EMSA subgroup was the main reason (Fig. 1F).

We found that 62 TaLYPs were localized on chromosomes (1A, 1B, 1D, 3A, 3B, 3D, 4A, 4B, 4D, 5A, 5B, 5D, 6A, 6B, 6D, 7A, 7B, 7D) and the unassembled scaffolds (Un). No gene was located on chromosome 2A, 2B and 2C (Fig. 2A).

In genetics, Ka/Ks represent the ratio between the nonsynonymous substitution rate (Ka) and the synonymous substitution rate (Ks) of two protein-coding genes. The value of Ka/Ks can be used as an indicator of selective pressure on a protein-coding gene. The Ka/Ks ratio was less than one almost in all gene pairs. The EMSA group showed relatively high Ka/Ks ratio, suggesting that these genes evolved at a faster evolutionary rate, which is a feature of new genes (Fig. 2B).

8.39% of LYP genes triads across the three subgenomes were comply with 1:1:1. The percentage of LYP genes with homolog-specific duplication is much higher than in all wheat genes (32.26% vs. 5.76%) [38]. Loss of one homolog is less pronounced in LYP genes (3.23% vs. 13.22%) (Table 2).

36 genes (58%) of the 62 genes were expressed in at least one developmental stage (based on the >0.5 TPM criteria). The remaining 26 genes which tpm <0.5 were considered not expressed. The LysMe clade has been expanded during wheat evolution. Many of the genes from this clade were not expressed or expressed on a very low level. However, the expression of LysMn and LysMFbox genes showed no significant changes during development. The expression peak of the EMSA subclade genes appeared in root at the reproductive stage. TaLYP5 and TaLYP1 clusters were expressed at seedling and vegetative period, but not at reproductive period. TaLYK1 genes and TaLYK4 genes were not expressed at any developmental period. TaLYK3 were highly expressed on leaves at reproductive period and spike (Fig. 3A).

We also analyzed the expression pattern of each triad. TaEMSA2 was expressed only from B subgenome. TaLYP4 was B suppressed. TaLysMFbox was D-suppressed in root at reproductive period. TaLYP1 was A-suppressed in root at reproductive period. TaLYP3 was B-suppressed in root at vegetative period and in spike at reproductive period. TaLYP3 was D-dominant in leaves/shoots at reproductive period (Fig. 3B).

We also analyzed the expression pattern of wheat LYPs to various biotic and abiotic stress (Fig. 4). Like the expression pattern during development, the LysMe subgroup were nearly not expressed upon various treatment. The expression of TaLysMFbox family was not changed significantly in response to various stress. For the LysMn family, genes were upregulated slightly when treated by chitin, flg22 or abiotic stress.

Except for the genes not expressed, most of the wheat LYK family members were upregulated in response to biotic and abiotic stress. Interestingly, the expression of TaLYK5s and TaLYK6s (except TaLYK6-6A) increased dramatically when treated by chitin and flg22, which may infer that they were important for PAMPs recognition. They were also slightly increased upon drought stress. TaLYK2s were induced by fungal infection and PAMPs treatment. TaLYK1-3B was dramatically increased when treated by heat or heat and drought after 6 h (Fig. 4).

The expression levels of TaLYP2 clusters were upregulated when inoculated with fungal disease or PAMPs. In this study, TaLYP3 and TaLYP4 clusters were not significantly changed under biotic stress. However, TaLYP4 gene cluster were induced by cold stress, drought, and heat treatment. Meanwhile, TaLYP3 cluster were depressed when treated by drought and heat stress.

We further used qRT-PCR analysis to investigate the expression of selected genes in response to wheat leaf rust. The selected four genes were upregulated during wheat leaf rust infection with similar expression patterns. The expression of TaLysMFbox (Fig. 5A) and TaLysMn (Fig. 5B) increased steadily from 6 to 18 h, fell markedly at 48 h and increased slightly at 168 h. TaLYK5 showed remarkably upregulated during the infection with a peak at 18 h (Fig. 5C). In particular, TaLYK6 was upregulated 300 to 400-fold at 6 and 18 h upon wheat leaf rust infection (Fig. 5D).