During growth and development, plants are affected by many different kinds of signal stimulation from cell-cell and cell-environment interactions, which trigger a series of signal transductions. The regulatory mechanism of reversible phosphorylation that involves protein kinases (PKs) plays an important role in cellular signal transduction processes and all living cells can perceive and process external or internal signals via cell surface PKs, which mostly belong to the receptor-like kinase (RLK) family [1, 2]. RLKs are ubiquitous in the plant kingdom. Since the first leucine-rich repeat receptor-like kinase (LRR-RLK) gene was isolated in maize (Zea mays) [3], more and more LRR-RLK genes have been confirmed [2, 4]. However, the function of most RLKs has not been confirmed and the natural ligands of most receptor proteins have not been identified.

LRR-RLK is the largest and most highly conserved group of the plant RLK gene family [5]. Previous studies have shown that the approximately 225 LRR-RLK gene members in Arabidopsis can be grouped into 23 subfamilies based on the sequence similarity in their kinase domains [2, 6, 7]. A typical LRR-RLK contains an extracellular tandem arrayed leucine-rich repeat domain, a single-pass transmembrane domain, and an intracellular functional protein kinase domain. The major class of extracellular domain is 20–30 amino acids long. With regards to tertiary structure, each LRR domain forms an α/ß helix and is located on the surface of the protein, enabling it to participate in protein-protein interactions [2, 8]. Because of their structural characteristics, LRR-RLKs play essential roles in mediating cell signal transduction pathways and critically control plant growth, development, responses to stresses, and immunity. In the Brassinosteroid (BR) and abscisic acid (ABA) signal pathways, LRR-RLKs play a pivotal role. BRs are important regulatory substances in plant development. The receptor kinase Brassinosteroid-insensitive 1 (BRI1), which has 25 LRR tandem structures in its extracellular portion, can perceive and deliver BR signaling [9]. BRI1-associated kinase 1 (BAK1) is another LRR-RLK that has five LRR tandem structures in its extracellular portion. The cross-talk between BRI1 and BAK1 makes up a dimer, the key step in the BR signal transduction [10]. In Arabidopsis, receptor-like protein kinase 1 (RPK1) is an LRR-RLK on the surface of the plasma membrane, the LRR domain of which is key to regulating ABA sensitivity [11]. Previous research has shown that the rpk1 mutant exhibits decreased sensitivity to ABA-induced senescence in Arabidopsis leaves, but does not exhibit a change in jasmonic acid- and ethylene-induced senescence [12]. BAK1 not only plays a role in the process of BR signal transmission, but also forms a dimer with the resistance gene flagellin-sensitive 2 (FLS2), and is involved in plant immune responses. BAK1 may also combine with other ligands and participate in other LRR-RLK signaling pathways [13].

There are a large number of LRR-RLKs in plants, but intensive studies on their functions have only been performed for a few genes. Because of the complexity of plant signal transduction pathways and complementary functions between different proteins, very few LRR-RLK ligands and complex LRR-RLK signaling pathways have been found. Thus, there is a great need for system analyses of the evolution and expression differentiation of the LRR-RLK gene family using bioinformatics tools, as well as to identify genes that have potential functions in plant responses to environmental factors; this will also help us to improve our understanding of LRR-RLK functions in regulating networks during plant growth. Recently, LRR-RLK gene family have been analyzed in soybean [14], two Citrus species [15], Amborella trichopoda [16], and evolutionary studies in various species [17, 18]. Specifically, Fischer et al. investigated the involvement of selection during the expansion of the LRR-RLK gene family among angiosperms by analyzing the phylogeny of 7554 LRR-RLK genes from 31 fully sequenced flowering plant genomes, including two genomes of Rosaceae, apple (Malus domestica) and peach (Prunus persica) [18]. However, a more comprehensive investigation of LRR-RLK genes in a wider range of Rosaceae genomes is needed. In this study, we aimed to elucidate the structure, expression, and evolution of Rosaceae LRR-RLK genes. To this end, we analyzed a total of 1622 LRR-RLK genes in 23 LRR-RLK subfamilies in five Rosaceae species [strawberry (Fragaria vesca), peach, mei (Prunus mume), apple and Chinese white pear (Pyrus bretschneideri)] and Arabidopsis, including their chromosomal location, gene structure, duplication events, expansion, expression, and co-expression relationships, which should provide genome-level insights on LRR-RLK genes.

In previous research, RLK genes were divided into six groups according to differences in the structures of their RLK extracellular domain, as follows: LRR [30–32], pathogenesis-related protein 5-like receptor kinase (PR5K) [33, 34], epidermal growth factor-like repeats (EGF) [35, 36], lectin-binding domain (LB) [37, 38], tumor necrosis factor receptor-like (TNFR) [39, 40], and S-domain [3, 41]. LRR-RLKs constitute more than half of the total number of RLKs. This group is one of the largest gene families in plants, playing important roles in plant growth and responses to pathogens [42–44]. In this study, we identified 1397 LRR-RLK genes in Rosaceae; of these, 201 genes were in strawberry, 244 in apple, 427 in Chinese white pear, 267 in mei, and 258 in peach (Table 2). These numbers are far greater than the average gene family size, which ranges from 1.71 to 2.17 in Arabidopsis, Chinese white pear, apple, and strawberry [20]. It was also noted that strawberry, apple, Chinese white pear, mei, and peach LRR-RLK genes represented 23.5%, 22.8%, 26.5%, 24.1%, and 25.2% of their total typical PK genes, respectively. These proportions are all higher than that in Arabidopsis (22.3%). In a previous study, it was asserted that LRR-RLK gene family expansion is mainly a consequence of high rates of duplicate retention [45], and that segmental and tandem duplication are two major sources of this expansion. In this study, we found that approximately one-third to two-thirds of LRR-RLK genes in Rosaceae could be explained by segmental duplication or tandem duplication, confirming this hypothesis. Considering the high ratio of LRR-RLK genes compared to PK genes in Rosaceae, the expansion of LRR-RLK genes could partially explain why the PK superfamily is larger in these species than in Arabidopsis. In Rosaceae, the Pyreae tribe was reported to have undergone one recent duplication [46] and it was found that 57% of apple LRR-RLK genes attributed to segmental duplication could be explained by this recent duplication, while the corresponding proportion was 54% in Chinese white pear. This explains why the proportions of total segmental duplicated genes in the Pyreae tribe are higher than in other studied species. We also noted that the genome and proteome sizes in the Pyreae tribe are larger than in other studied species, which could be another reason for the larger gene family size of LRR-RLK in these species. However, there is one exception to this in the apple genome, where the number of LRR-RLK genes identified were much less than in Chinese white pear, because we used the apple genome version 1.0 pseudo-haplotype assembly primary resource in the Genome Database for Rosaceae (GDR) [47]. The total number of genes in this version of the apple genome is 30,294, which is approximately half of that in the apple genome version 1.0. Nevertheless, we still observed a higher proportion of genes attributed to recent duplication.

Novel gene functions are acquired in four processes of genetic evolution: chromosome doubling, chromosome fragment insertion mutation, tandem duplication, and transposition [19, 48, 49]. Duplications can help species resist stress and to adapt to diverse environments by providing abundant genetic materials for evolution and producing numerous genetic variations. As described above, duplications could have driven the lineage-specific expansion of the LRR-RLK gene family, especially in the Pyreae tribe. We also noted that duplications had different effects on specific subfamilies, especially for tandem duplications. For instance, in Arabidopsis, 37 genes assigned to tandem duplication arrays were distributed in only five subfamilies, which represented approximately 16% of the 225 LRR-RLK genes; in addition, in Chinese white pear, 67 genes assigned to tandem duplication arrays were distributed in six subfamilies. Furthermore, the proportions of genes in tandem duplication arrays against the total LRR-RLK genes in each species in Rosaceae were not consistent, from 12% in strawberry and 13% in apple, less than in Arabidopsis, to 16% in Chinese white pear, 19% in mei, and 24% in peach, greater than or equal to that in Arabidopsis (Additional file 1: Table S10). Although different species had different numbers of subfamilies that expanded through tandem duplication, the observation that some subfamilies were consistently expanded through tandem duplication in all studied species (I-1, XI-1, and XII-1) while others were not indicates the preference for tandem duplication. Those subfamilies that expanded markedly through tandem duplication were partly consistent with those subfamilies with the greatest number of members, except subfamily III, which showed tandem duplication events only in apple. In contrast, the expansion of subfamily III was mainly the consequence of segmental duplication.

Many LRR-RLKs have been reported to play roles in plant disease resistance. For example, dependent on the suppressor of BIR1–1 (SOBIR1), BAK1 negatively regulated Arabidopsis defenses interacted with the RLK BIR1 (BAK1-interacting receptor-like kinase 1) [50]. Moreover, the steroidal hormones BRs were shown to promote pollen tube growth in Arabidopsis [51]. Interaction of BRI1 with BAK1 also play a key role in this promoted signaling network. In addition, BAK1 acts not only as the second receptor of BRs, but is also a co-receptor of LRR-RLK FLS2 and EF-Tu receptor (EFR) [42, 52, 53]. In the processes of plant immune response, BAK1 and FLS2/EFR form a dynamic receptor and work with Botrytis-induced kinase 1 (BIK1). As a member of the Catharanthus roseus receptor-like kinase (CrRLK) family, FERONIA (FER) plays two important roles. First, it works as a pollen tube guidance protein during plant fertilization. Second, it plays a negative role in resistance to powdery mildew fungus [54–56]. The functionally redundant genes ANXUR (ANX1 and ANX2) are close homologs of FER and expressed in pollen. Overexpressed ANXUR could negatively regulate pollen tube growth. However, more research is required to determine whether FER and ANXUR are involved in the process of BR promotion of pollen tube growth and plant immune responses and what roles they play. Although LRR-RLKs and CrRLKs, two important subfamilies of the RLK family, have evolved in different directions, they retain many similar biological functions in the process of plant growth.

According to the analyses of expression data presented in this work, we found that genes in the same subfamily have no consistent expression patterns; some genes in a subfamily might be highly expressed, whereas others are expressed at a low level or not at all. This pattern differs from an analysis of the soybean kinome, where it was found that gene members belonging to the same PK subfamilies showed similar expression patterns [57]. The reason for this difference could be the limitation of available expression data for soybean PKs, just 38%. By examining the expression of some duplicated genes within a subfamily, we found different expression patterns, which could be the result of the sub-functionalization or neo-functionalization of duplicated genes. Considering that some LRR-RLK subfamilies are very large, with over 100 members, the use of averaged expression data for such subfamilies could result in the loss of substantial information. LRR-RLK genes have been proven to have various functions in multiple tissues. Previous studies showed that LRR-RLK genes play important roles in root development; for example, OsRPK1 in rice was shown to affect the root system architecture by negatively regulating polar auxin transport [58]. In this study, tissue expression analyses also showed that LRR-RLK genes were relatively highly expressed in root and stem compared with their levels in other examined tissues.

In our study, subfamily XI-1 was not only the largest subfamily, but was also highly expressed in all five Rosaceae species. It is worth mentioning that a total of 80 tandem duplicated genes were found only in the XI-1 subfamily, representing approximately 33.9% of the 236 tandem duplicated genes in the five Rosaceae species (Additional file 1: Table S10). This indicates that the gene expansion of this subfamily mainly occurred through fragment replication and tandem copies. CLAVATA (CLV), a member of the XI-1 subfamily in Arabidopsis, plays an important role in maintaining stem cell numbers in the apical meristem. CLV1 encodes a typical LRR-RLK domain, the mutation of which leads to increases in the stem cell number of shoot apical meristem, flower number, and stem tip size [59]. Another example is provided by the LRR receptor kinase HAESA (HAE), which belongs to the XI-1 subfamily and is expressed in sepals, petals, the abscission zone of stamens, the bottom of the peduncle, and the base of the petiole. Decreased expression of this gene delays the withering of flower organs [60]. These results indicate that the Rosaceae XI-1 subfamily plays a very important role in growth and development. However, the functions of numerous LRR-RLK genes still remain unclear.

According to the expression analysis results presented here, we found that some genes from the III, V, VI-2, and LRR-XII-1 subfamilies were highly expressed during Chinese white pear pollen growth. Interestingly, a pattern was identified of most of these genes being up-regulated in the SPT stage, while some LRR-RLK genes were shown to promote pollen tube growth. Therefore, these highly expressed genes may function in the process of plant sperm cell release. In addition, there were 40 tandem duplicated genes in the XII-1 subfamily, which represented approximately 60% of the 67 tandem duplicated genes in Chinese white pear. This suggested that the gene tandem copies are one of the main causes of high expression of the XII-1 subfamily in pollen. Notably, the expression data showed that the Xb-1 and II subfamilies were not highly expressed in Chinese white pear pollen, but the RT-PCR experiment revealed that the genes PbrBRL2 and PbrSERK2, which belong to these two subfamilies, were highly expressed in pollen tissues. This features might be associated with the age, location, and biotic and abiotic stresses of the tested samples.

Our analyses of LRR-RLK genes in five Rosaceae species revealed their diversity in terms of member number, gene structure, tissue expression patterns, and expression patterns in response to stress. Expression patterns of these LRR-RLK genes should promote our understanding of the involvement of this gene family in plant growth, development, signal transduction, immunity, and stress responses.