Evolutionary Dynamics of the Leucine-Rich Repeat Receptor-Like Kinase (LRR-RLK) Subfamily in Angiosperms

Gene duplications are an important factor in plant evolution, and lineage-specific expanded (LSE) genes are of particular interest. Receptor-like kinases expanded massively in land plants, and leucine-rich repeat receptor-like kinases (LRR-RLK) constitute the largest receptor-like kinases family. Based on the phylogeny of 7,554 LRR-RLK genes from 31 fully sequenced flowering plant genomes, the complex evolutionary dynamics of this family was characterized in depth. We studied the involvement of selection during the expansion of this family among angiosperms. LRR-RLK subgroups harbor extremely contrasting rates of duplication, retention, or loss, and LSE copies are predominantly found in subgroups involved in environmental interactions. Expansion rates also differ significantly depending on the time when rounds of expansion or loss occurred on the angiosperm phylogenetic tree. Finally, using a d_N/d_S-based test in a phylogenetic framework, we searched for selection footprints on LSE and single-copy LRR-RLK genes. Selective constraint appeared to be globally relaxed at LSE genes, and codons under positive selection were detected in 50% of them. Moreover, the leucine-rich repeat domains, and specifically four amino acids in them, were found to be the main targets of positive selection. Here, we provide an extensive overview of the expansion and evolution of this very large gene family.Receptor-like kinases (RLKs) constitute one of the largest gene families in plants and expanded massively in land plants (Embryophyta; Lehti-Shiu et al., 2009, 2012). For plant RLK gene families, the functions of most members are often not known (especially in recently expanded families), but some described functions include innate immunity (Albert et al., 2010), pathogen response (Dodds and Rathjen, 2010), abiotic stress (Yang et al., 2010), development (De Smet et al., 2009), and sometimes multiple functions (Lehti-Shiu et al., 2012). The RLKs usually consist of three domains: an N-terminal extracellular domain, a transmembrane domain, and a C-terminal kinase domain (KD). In plants, the KD usually has a Ser/Thr specificity (Shiu and Bleecker, 2001), but Tyr-specific RLKs were also described (e.g. BRASSINOSTEROID INSENSITIVE1; Oh et al., 2009). Interestingly, it was estimated that approximately 20% of RLKs contain a catalytically inactive KD (e.g. STRUBBELIG and CORYNE; Chevalier et al., 2005; Castells and Casacuberta, 2007; Gish and Clark, 2011). In Arabidopsis (Arabidopsis thaliana), 44 RLK subgroups (SGs) were defined by inferring the phylogenetic relationships between the KDs (Shiu and Bleecker, 2001). Interestingly, different SGs show different duplication/retention rates (Lehti-Shiu et al., 2009). Specifically, RLKs involved in stress responses show a high number of tandemly duplicated genes whereas those involved in development do not (Shiu et al., 2004), which suggests that some RLK genes are important for the responses of land plants to a changing environment (Lehti-Shiu et al., 2012). There seem to be relatively few RLK pseudogenes compared with other large gene families, and copy retention was argued to be driven by both drift and selection (Zou et al., 2009; Lehti-Shiu et al., 2012). As most SGs are relatively old and RLK subfamilies expanded independently in several plant lineages, duplicate retention cannot be explained by drift alone, and natural selection is expected to be an important driving factor in RLK gene family retention (Lehti-Shiu et al., 2009).Leucine-rich repeat-receptor-like kinases (LRR-RLKs), which contain up to 30 leucine-rich repeat (LRRs) in their extracellular domain, constitute the largest RLK family (Shiu and Bleecker, 2001). Based on the KD, 15 LRR-RLK SGs have been established in Arabidopsis (Shiu et al., 2004; Lehti-Shiu et al., 2009). So far, two major functions have been attributed to them: defense against pathogens and development (Tang et al., 2010b). LRR-RLKs involved in defense are predominantly found in lineage-specific expanded (LSE) gene clusters, whereas LRR-RLKs involved in development are mostly found in nonexpanded groups (Tang et al., 2010b). It was also discovered that the LRR domains are significantly less conserved than the remaining domains of the LRR-RLK genes (Tang et al., 2010b). In addition, a study of four plant genomes (Arabidopsis, grape Vitis vinifera], poplar Populus trichocarpa], and rice Oryza sativa]) showed that LRR-RLK genes from LSE gene clusters show significantly more indications of positive selection or relaxed constraint than LRR-RLKs from nonexpanded groups (Tang et al., 2010b).The genomes of flowering plants (angiosperms) have been shown to be highly dynamic compared with most other groups of land plants (Leitch and Leitch, 2012). This dynamic is mostly caused by the frequent multiplication of genetic material, followed by a complex pattern of differential losses (i.e. the fragmentation process) and chromosomal rearrangements (Langham et al., 2004; Leitch and Leitch, 2012). Most angiosperm genomes sequenced so far show evidence for at least one whole-genome multiplication event during their evolution (Jaillon et al., 2007; D’Hont et al., 2012; Tomato Genome Consortium, 2012). At a smaller scale, tandem and segmental duplications are also very common in angiosperms (Arabidopsis Genome Initiative, 2000; International Rice Genome Sequencing Project, 2005; Rizzon et al., 2006). Although the most common fate of duplicated genes is to be progressively lost, in some cases they can be retained in the genome, and adaptive as well as nonadaptive scenarios have been discussed to play a role in this preservation process (for review, see Moore and Purugganan, 2005; Hahn, 2009; Innan, 2009; Innan and Kondrashov, 2010). Whole-genome sequences also revealed that the same gene may undergo several rounds of duplication and retention. These LSE genes were shown to evolve under positive selection more frequently than single-copy genes in angiosperms (Fischer et al., 2014). That study analyzed general trends over whole genomes. Here, we ask if, and to what extent, this trend is observable at LRR-RLK genes. As this gene family is very dynamic and large, and in accordance with the results of Tang et al. (2010b), we expect the effect of positive selection to be even more pronounced than in the whole-genome average.We analyzed 33 Embryophyta genomes to investigate the evolutionary history of the LRR-RLK gene family in a phylogenetic framework. Twenty LRR-RLK SGs were identified, and from this data set, we deciphered the evolutionary dynamics of this family within angiosperms. The expansion/reduction rates were contrasted between SGs and species as well as in ancestral branches of the angiosperm phylogeny. We then focused on genes whose number increased dramatically in an SG- and/or species-specific manner (i.e. LSE genes). Those genes are likely to be involved in species-specific cellular processes or adaptive interactions and were used as a template to infer the potential occurrence of positive selection. This led to the identification of sites at which positive selection likely acted. We discuss our results in the light of angiosperm genome evolution and current knowledge of LRR-RLK functions. Positive selection footprints identified in LSE genes highlight the importance of combining evolutionary analysis and functional knowledge to guide further investigations.