J Mol Evol (2014) 78:118–129 DOI 10.1007/s00239-013-9609-5
ORIGINAL ARTICLE
On the Origin and Evolution of Plant Brassinosteroid Receptor Kinases Hao Wang • Hongliang Mao
Received: 6 September 2013 / Accepted: 18 December 2013 / Published online: 27 December 2013 Ó Springer Science+Business Media New York 2013
Abstract Brassinosteroid (BR) signaling pathway is so far the best-understood receptor-kinase signaling pathway in plants. In Arabidopsis, the activation of this pathway requires binding of BRs to the receptor kinase BRASSINOSTEROID-INSENSITIVE I (AtBRI1). Although the function of AtBRI1 has been extensively studied, it is not known when the binding function emerged and how this important component of BR signaling pathway and related genes (the BRI1–BRL gene family) have evolved in plants. We define BRI1–BRL genes in sequenced plant genomes, construct profiles for critical protein domains, scan them against all accessible plant gene/EST resources, and reveal the evolution of domain configuration of this family. We also investigate its evolutionary pattern through phylogenetic analysis. The complete BR receptor domain configuration originates through two domain gain events in the ancestral receptor-like kinase: first juxtamembrane domain gained during the early diversification of land plants, and then island domain (ID) acquired in the common ancestor of angiosperms and gymnosperms after its divergence from spike moss. The 70 amino acid ID has characteristic sequences of BRI1–BRL family and this family keeps relative stable copy numbers during the history of angiosperms and the majority of duplications and losses have Electronic supplementary material The online version of this article (doi:10.1007/s00239-013-9609-5) contains supplementary material, which is available to authorized users. H. Wang H. Mao T-Life Research Center, Department of Physics, Fudan University, Shanghai 200433, People’s Republic of China H. Wang (&) Department of Genetics, University of Georgia, 120 Green Street, Athens, GA 30602, USA e-mail: [email protected]
123
occurred in terminal taxa in current taxon sampling. This study reveals important events shaping structural and functional characteristics of plant BR receptors. It answers the question of how and when BR receptors originates, which provide insights into the origin and evolution of the BR signaling pathway. Keywords Brassinosteroid BRI1 Protein domain Gene family Evolution
Introduction Brassinosteroids (BRs) are a group of plant steroid hormones that play critical roles in a wide range of developmental and physiological processes such as cell elongation, vascular differentiation, root growth, light response, stresses resistance, and senescence (Clouse and Sasse 1998; Kim and Wang 2010; Vert et al. 2005). The last two decades have observed great advances in assembling the BR signal transduction pathway (or network Wang et al. 2012) in Arabidopsis. A number of component genes involved in the BR transduction pathway have been defined and important signal transduction steps, from BR perception by the receptor kinase to the activation of the most upstream transcription factors of the BR-dependent transcriptional network, have been revealed (e.g. see reviews in Clouse 2011; Kim and Wang 2010). In early events that activate the Arabidopsis BR signaling network, the binding of BR requires the function of BR receptor kinase BRASSINOSTEROID-INSENSITIVE I (BRI1 or AtBRI1). Genetic and biochemical studies have established AtBRI1 as the major receptor of BRs (Kinoshita et al. 2005; Li and Chory 1997) in Arabidopsis, and great efforts have been made to elucidate the functional
J Mol Evol (2014) 78:118–129
regions or protein domains in this gene (see review in Kim and Wang 2010). The currently accepted domain configuration of AtBRI1 is LRR{20}–ID–LRR21–LRR{3}–TM– JM–KD–CT (Kim and Wang 2010; Vert et al. 2005), where LRR, ID, LRR21, TM, JM, KD, and CT denote leucinerich repeats, island domain, the 21st LRR domain, transmembrane region, juxtamembrane region, kinase domain, and C-terminal region, respectively. LRR21 and its upstream adjacent ID play important roles together so this LRR is marked specially. The numbers in braces are copy numbers of tandem domain units. For instance, LRR{20} means 20 tandem LRR domains. AtBRI1 in fact has 25 LRRs (Kinoshita et al. 2005), but the one located at N-terminal is irregular (She et al. 2011). If taking the irregular LRR as the first LRR, ID is located between LRR-21 and -22. In this study, we count the first regular LRR as the first LRR, following the numbering system used by several previous works (Kim and Wang 2010; Li and Chory 1997; Vert et al. 2005). Several homologs of AtBRI1 have been identified: (1) AtBRI1 orthologs have been defined in a number of crops such as tomato (Curl3/tBRI1/SR160; Koka et al. 2000; Montoya et al. 2002), rice (OsBRI1; Yamamuro et al. 2000), barley (HvBRI1; Chono et al. 2003), cotton (GhBRI1; Sun et al. 2004), grape (VvBRI1; Symons et al. 2006), and pea (LKA/PsBRI1; Nomura et al. 2003), and their function in BR perception and plant growth have been confirmed by mutational researches (see Clouse 2011; Morillo and Tax 2006 for brief reviews). (2) AtBRI1 is the major, but not the only BR receptor in Arabidopsis. Three paralogs of AtBRI1 have been cloned and two of them, AtBRL1 and AtBRL3, can also bind to BR and mediate cell-type-specific BR response and rescue the bri1 mutant when expressed under the control of the BRI1 promoter (Cano-Delgado et al. 2004). The investigation of the counterparts of AtBRI1 and AtBRL genes in rice seems to support the model that BRI1 performing as major BR receptor and some of its homologs as partial functional backup also work in rice (Nakamura et al. 2006). The structure and function of AtBRI1 have been extensively studied. In contrast, the evolution of AtBRI1 and related genes (called the BRI1–BRL gene family hereafter) is largely unrevealed. AtBRI1 belongs to LRR receptor-like kinases (LRR RLKs) which are characterized by extracellular LRR arrays, a single-pass TM and a cytoplasmic KD (Shiu and Bleecker 2001a). Though the relationship and diversity of LRR and KD have attracted many attentions and these studies have established that BRI1–BRL genes form a clade in the KD phylogenetic tree (Dolan et al. 2007; Kobe and Kajava 2001; Lehti-Shiu et al. 2009; Matsushima et al. 2009, 2010; Shiu and Bleecker 2001a, b; Shiu et al. 2004), none of these researches have focused on the BRI1–BRL family per se. To date, the
119
relationship between BRI1 and BRL genes only has been discussed in quite limited organisms (Cano-Delgado et al. 2004; Gish and Clark 2011; Nakamura et al. 2006). The pioneering work based on the phylogeny of the 10 available BRI1–BRL genes (Cano-Delgado et al. 2004) discovered that the early duplications splitting BRI2 from BRI1 and BRL1–3 occurred prior to the divergence of monocots and dicots. However, due to the lack of data, this work failed to correctly resolve the time of the split of BRL1 and BRL3 and could not discuss many other questions such as the birth and death patterns and lineage distribution of family members. In short, systematic investigations of this family with an extensive taxon sampling have not been reported yet and how and when this family originated still remains an open question. Recent genomics revolution has uncovered genomes from many major lineages of plant tree of life and deposited a large number of EST/transcriptome data for lineages without fully sequenced genomes. With the aim of illuminating the origin and evolution of structure and function of plant BRI1–BRL genes, we have performed comparative analyses of the BRI1–BRL gene family in major clades of the plant kingdom. Focusing on the 44 sequenced plant genomes [32 land plants and 12 green algae (Supporting Information Table S1)], we have identified the BRI1–BRL gene repertoire in 29 genomes using bioinformatic method and tried to answer the following questions: (1) how and when the family originated, (2) relationship between family members, and (3) pattern of gene duplication and loss. We also have discussed a possible picture of the origin of the single-exon structure of this family.
Materials and Methods Sequence Data Fully sequenced plant genomes and gene annotations were downloaded from phytozome version 8 (http://phytozome. net/). Data of chlorophytes that were deposited in JGI but not included in phytozome were downloaded from JGI genome portal (http://genome.jgi.doe.gov/). The genome and annotations of Amborella trichopoda were downloaded from Amborella Genome Database (http://www.amborella. org/). ESTs, EST assemblies, transcriptome assemblies, and cDNAs (mRNAs) of conifers as well as other nonangiosperm plants were collected from TreeGene (http://den drome.ucdavis.edu/treegenes/) and plantGDB (http://www. plantgdb.org/). A translated EST dataset was constructed by predicting coding regions of the ESTs by ORFPredictor (http://proteomics.ysu.edu/tools/OrfPredictor.html). Proteins of plants that deposited in Uniprot (http://www.uni prot.org/) were also downloaded. The combined protein set
123
120
was made of nonredundant sequences from genome annotation, translated EST dataset, and Uniprot. BRI1–BRL Gene Identification The combined protein set was scanned against PFAM v26.0 (Punta et al. 2012) to obtain all genes containing KD using the Pfam_scan.pl script (ftp://ftp.sanger.ac.uk/pub/ databases/Pfam/Tools). These genes were then assigned to subfamilies according to the similarity of their KDs with known KDs of plant RLK/Pelle genes (Lehti-Shiu et al. 2009). The previous study (Shiu and Bleecker 2003) suggested that AtBRI1 and AtBRL genes belonged to the LRR-Xb-1 subfamily (called RLK/Pelle-LRR-Xb-1 genes hereafter). All genes belonged to the RLK/Pelle-LRR-Xb-1 subfamily were extracted and the amino acids neighborjoining (NJ) tree was built using their KD sequences (see below). The genes fell within the same well-supported clade with known BRI1–BRL genes were extracted as BRI1–BRL candidates (Supporting Information Fig. S1). We manually cured the BRI1–BRL candidates from fully sequenced organisms to improve gene models. Every gene locus was re-annotated through AUGUSTUS (Stanke et al. 2008) with EST evidence. Each predicted gene model was compared to that released by genome sequencing centers and the better model was chosen based on EST support and sequence alignment quality. This manual inspection excluded nine models for further study because they were far shorter than other BRI1–BRL genes or lacked other essential domains or had no EST support. Protein Domain Profile Construction Multiple sequence alignment (MSA) of protein sequences of BRI1–BRL genes was constructed using MUSCLE (Edgar 2004) and the alignment was manually inspected. In the alignment, blocks representing ID, LRR21, JM, and KD domains were extracted according to their locations in AtBRI1 (Vert et al. 2005). Profile hidden Markov models (HMMs) of domains were then constructed by using HMMER v3.0 package (Finn et al. 2011). Domain Configuration Identification The identification of ID, JM, and KD was performed by hmmsearch (E value = 1e-5) using profile constructed above. Besides hmmsearch, PFAM profiles (v26.0) were also scanned against genome sequences of the 12 chlorophytes to detect the LRR–KD configuration in these species. The presence of TM in genes was predicted using TMHMM webserver (http://www.cbs.dtu.dk/services/ TMHMM/).
123
J Mol Evol (2014) 78:118–129
Phylogenetic Analysis KD regions were aligned using MUSCLE, and the NJ tree (Saitou and Nei 1987) of all RLK/Pelle-LRR-Xb-1 genes was built using MEGA 5 (Tamura et al. 2011) with the following parameters: Poisson model; uniform rate; and pairwise deletion of gaps/missing data. Here, we excluded sequences shorter than 140 aa (50 % of average size of experimentally verified BRI1–BRL genes) and so the size of KD sequences used in this NJ tree construction was from 140 to 355 aa. Amino acid maximum likelihood (ML) phylogenetic tree of BRI1–BRL genes from sequenced angiosperms was constructed by RAxML (Stamatakis 2006) using conserved region stretching from ID to KD (ID–LRR{4}–TM–JM– KD; sequence lengths 536–648 aa). Sequence alignments were generated by MUSCLE and low quality regions were excluded from alignment using trimAL (Capella-Gutierrez et al. 2009) with the option ‘‘automated1’’. Protein model selection was performed by Prottest 3.0 (Abascal et al. 2005). Parameters of ML tree construction are as follows: JTT ? I ? G model; duplications of rapid bootstrap = 100. In 227 Arabidopsis genes of which KDs belong to the LRR RLK domain group, 21 highly diverged ones (could not calculate valid evolutionary distance with others) were excluded. The DNA MSA of the coding sequences of KD domain (sequence lengths 417–900 nt) was constructed with MUSCLE and trimmed with trimAL. ML tree was built by RAxML with the following parameter settings: model = GTR ? I ? G; duplications of rapid bootstrap = 100. Model selection was performed by jModeltest (Posada 2008). Gene Duplication–Loss History The species tree used in this study was constructed by modifying Phytozome 8 plant tree of life: A. trichopoda was added as sister taxon of monocots and dicots and Panicum virgatum as sister species of Setaria italica. Gene tree and species tree reconciliation was performed by Notung 2.6 (Chen et al. 2000). When reconstructing the gene duplication–loss history, weakly supported edges (bootstrap value \90 %) of the gene tree were allowed to be rearranged to minimize duplication and loss events. Selection Analysis Coding DNA sequence alignments of paralogous pairs were obtained under the guidance of protein alignments using PAL2NAL (Suyama et al. 2006). Dn, Ds, and Dn/Ds values were calculated by the yn00 program of the PAML
J Mol Evol (2014) 78:118–129
package (Yang 2007) according to the Nei–Gojobori method (Nei and Gojobori 1986).
Results BRI1–BRL Genes in Sequenced Genomes Phylogenetically, the BRI1–BRL gene family, in this study, was defined as the sub-clade of RLK/Pelle-LRR-Xb1 genes (Lehti-Shiu et al. 2009) that included all of the previously cloned BRI1–BRL genes. We built a NJ tree for all detected KD in the combined plant protein set made of whole gene repertoire of the 44 sequenced genomes, all plant proteins in the Uniprot database and all translated ESTs of nonangiosperm plants (see ‘‘Materials and Methods’’ section for details). In the tree (Fig. S1), the BRI1–BRL clade (bootstrap value = 95 %) included a total of 220 sequences, with 136 being full-length BRI1–BRL gene models (Table S2). Here, full-length means the open reading frame containing both a start and stop codon. 117 out of the 136 full-length genes were from 29 sequenced species (Table S3). Our manual inspection predicted 4 novel models in the apple genome (Malus domestica) and suggested that 35 gene models should be modified based on expression and/or conservation evidence. The correspondence between corrected models and automatic gene annotation was shown in Table S3. The other 19 full-length BRI1–BRL genes were from Uniprot. They belonged to organisms without whole-genome sequences when this study was done (Table S2). It is worth noting that all of the BRI1–BRL genes are from angiosperms and gymnosperms. Domain Configuration of BRI1–BRL Family The MSA of the BRI1–BRL family showed that most domains were quite conservative (Fig. S2). Low conservation was found at the first several LRRs, CT, and internal region of TM. Overall, LRR10–24 (covering ID) and KD were the two most conserved parts in the entire alignment. We compared the domain configuration of AtBRI1 with other previously identified BRI1–BRL genes and found all of them had the identical configuration with AtBRI1. Therefore, we constructed profiles of the LRR, ID, JM, and KD domains (see ‘‘Materials and Methods’’ section) and used them to scan the above 220 BRI1–BRL genes. 152 sequences exhibited a domain configuration like ‘‘LRR array’’–ID–‘‘LRR array’’–TM–JM–KD (Table S2), including 136 full-length genes (117 in sequenced species ? 19 in other species from Uniprot), 11 partial proteins from conifer EST/transcriptome assemblies, and 5 partial proteins from Uniprot. The other 68 sequences had KD but not ID: 23 from Uniprot, 6 from plantGDB, and 39
121
from TreeGene EST/transcriptome assemblies. Unlike protein sequences, several EST/transcriptome assemblies might come from the same gene and thus the number of EST/transcriptome assemblies might not correctly reflect number of genes. However, the occurrence of a certain domain in these assemblies was sufficient to confirm their occurrence in corresponding species. In this scan, we identified TM domains through the TMHMM webserver, but not through domain profile analysis because of low conservation at sequence level. Sequences of the CT region were also highly divergent (Fig. S2) and the instability of CT was reported previously (Xu et al. 2009). Therefore, although it might plays an inhibitory role in BRI1 function (Wang et al. 2005), CT was not suitable for profile analysis and excluded from this study. In short, all full-length genes had a domain configuration like ‘‘LRR array’’–ID–‘‘LRR array’’–TM–JM–KD, so this was the valid definition of BRI1–BRL gene family in term of domain configuration. Presence and Absence of Domains in Major Lineages of the Plant Kingdom We refined the profile for the four domains (LRR, ID, JM, and KD) based on the MSA of the 136 full-length BRI1– BRL genes and scanned them against the combined plant protein set. The results showed that (1) the LRR–KD combination was detected in 22 genes from 6 chlorophytes (Table S4). In contrast, we failed to detect the LRR–KD configuration in the sequenced red algae (Cyanidioschyzon merolae). In 13 cases, LRR occurred as long tandem arrays (C5 LRR units). (2) TM domain was located in between some (nine cases) but not all of the LRR–KD structure. (3) JM domain was not detected in any chlorophyte, but was found in all available major groups of land plants including liverworts, mosses, lycophytes, and seed plants such as cycads, gnetophytes, conifers (Table S5), and angiosperms. We found that whenever full-length proteins are available, their JMs were found co-occurring with LRR array, TM, and KD, but not always with ID. (4) IDs were only detected in gymnosperms (conifers and gnetophytes) and angiosperms. We found that a total of 29 gymnosperm EST/transcriptome assemblies contained IDs, with 11 of them showing complete BRI1–BRL gene domain configuration (Table S6). In contrast, this domain configuration analysis failed to detect BRI1–BRL members in plant genomes that diverged before seed plants and this result was consistent with the KD phylogenetic analysis described above. In summary, according to current available data, the first occurrence of LRR–TM–KD, LRR–TM–JM– KD, and LRR–ID–LRR–TM–JM–KD were observed in chloroplasts, liverworts, and the common ancestor of seed plants, respectively (Fig. 1).
123
122
Fig. 1 Domain configuration evolution of BRI1–BRL family. The relationship within seed plants remains unresolved because of controversial results from different datasets and methods (see e.g. Mathews 2009). The three stages of BRI1–BRL domain configuration evolution are mapped to the tree. Arrows point to the internal nodes which are lower bounds of the first presence of configurations. Right table shows the presence/absence of domains. ‘‘?’’: presence and ‘‘-’’: absence
ID Was the Diagnostic Domain of BRI1–BRL Genes In all full-length sequences within the combined plant protein set, we found that whenever ID was present, the entire BRI1–BRL domain configuration, i.e. ‘‘LRR array’’ –ID–‘‘LRR array’’–TM–JM–KD, occurred automatically. Furthermore, we found that if a sequence contains both ID and KD, the KD fell in the BRI1–BRL clade. These observations indicate that domain configuration and phylogenetic relationship of KD give equivalent definitions of BRI1–BRL family and ID is the domain only present in the BRI1–BRL family. This family now can be described as plant RLKs containing ID.
Characteristics of ID Domain The investigation of the alignment profile of ID domain of BRI1–BRL genes (Figs. 2a; S3) identified six residues conserved in all genes: C23–G27–L29–E31–C49–Y57, where letters were codes of residues and numbers were the locations of residues in the alignment. Besides the CGLECY hexad, another six highly conserved residues or motifs (i.e., occurring in [95 % of the sequences) were also revealed within ID domain (Fig. 2a). Moreover, the ID domain sequences could be further categorized as three groups according to their conservation pattern (Fig. 2b). The three ID groups were congruent to the three major clades of the BRI1–BRL family (Fig. 2b and see below).
Phylogenetics of Plant BRI1–BRL Genes The ML phylogeny of BRI1–BRL genes from the 29 sequenced angiosperms (Fig. 3) supported with 100 %
123
J Mol Evol (2014) 78:118–129
bootstrap value that the BRI1–BRL family was composed of three major clades: the basal split was between BRL2 and all others, and the second split was between BRI1 and BRL1–3 genes. Here, based on the names of the Arabidopsis and rice members, we called the three clades as BRI1, BRL1–3 and BRL2, respectively. Reconciliation the BRI1–BRL gene tree with sequenced angiosperm species tree (Fig. S4) suggested that the three major clades were derived from two gene duplication events before the divergence of angiosperms (Fig. S5). These results were consistent with the previous results from far less sequences (Cano-Delgado et al. 2004; Gish and Clark 2011). We investigated gene duplication and loss pattern of the family. (1) the majority of duplications (85 %, 28 out the 33) and losses (77 %, 10 out of 13) happened after the divergence of leaf nodes and their closest relative species in current taxon sampling (Fig. S5). This indicated that no drastic expansion or contraction of gene number occurred in the ancestral nodes. (2) Copy number variation of BRI1– BRL genes was found in all the three clades (Table 1) with the highest number of genes in M. domestica (nine; three in each clade) and lowest in Medicago truncatula (one from the BRI1 clade). However, 62 % (18 out of 29) species had three or four genes and, except for M. domestica, all variations of gene copy number within species were within mean ± 2 9 SD. In this sense, gene numbers showed no strong organism bias. (3) The three major clades had similar numbers of genes (41 in BRI1, 41 in BRL1–3, and 35 in BRL2). These observations support that gene numbers were quite stable during the evolution of this gene family. According to the phylogeny of BRL1–3 (Fig. 3b), the Arabidopsis and rice BRL1 and BRL3 genes had independent origins. AtBRL1 and AtBRL3 were derived from a duplication in the common ancestor of Brassicaceae, which left two descendants in each of the five sequenced Brassicaceae organisms. Rice BRL1 and BRL3 originated through a duplication in the common ancestor of grass after divergence from dicots and both descendants were also preserved in the most investigated grass species, except for Brachypodium. In switch grass (P. virgatum), a second round of duplications occurred independently in each of the two copies and resulted four BRL1–3 genes. Selection on Paralogs The 28 terminal species duplications generated 22 groups of in-paralogs (Koonin 2005): 17 pairs related by 1 duplication, 4 triplets by 2, and 1 quartets by 3 duplications (Fig. S5). Investigating selection pressure using Nei– Gojobori method exhibited Dn/Ds values \1 in all withingroup gene pairs (Table S7) and detected no positive selection.
J Mol Evol (2014) 78:118–129
Discussion Origin of the BRI1–BRL Family: A Three-Stage Process of Domain Gain The evolution of the domain configuration provides a scenario of gene family evolution at the level of organization of structural and/or functional units. Our results exhibit a perfect consistency between phylogeny and domain configuration when defining the BRI1–BRL gene family. If the origin of the family is marked by formation of the ‘‘LRR array’’–ID–‘‘LRR array’’–TM–JM–KD domain configuration, we have revealed a stepwise domain acquirement process in its ‘‘pre-histoy’’. LRRs present in numerous proteins from all major branches of the tree of life and with diverse functions (Kobe and Kajava 2001). To date, at least seven groups of LRRs have been identified and at least most LRRs have been found following the so-called the mutual exclusive rule, i.e., LRRs from different groups are found to never occur simultaneously in the same protein. Although phylogenetic relationships of the seven groups have not been well resolved yet (Andrade et al. 2000; Kajava 1998), it is clear that LRRs presented in BRI1–BRL genes belong to the plant-specific group that are found in plants and protists (Kobe and Kajava 2001). The ancient origin of KDs of BRI1–BRL genes has been uncovered for more than 10 years: this domain belongs to plant RLKs domain family, which together with animal Pelle and related cytoplasmic KDs form the RLK/Pelle domain family. The RLK/Pelle domain family, animal receptor serine/threonine kinases, animal receptor tyrosine kinases, and Raf kinases form a clade within the serine/ threonine/tyrosine kinases (Shiu and Bleecker 2001b). The domain configuration of RLKs is defined as ‘‘extracelluar domain’’–KD, where extracellular domain represents a wide range of domains including LRR (Shiu and Bleecker 2001a). It has been suggested that such domain combination is the structural basis of new signaling pathway evolution (Lehti-Shiu et al. 2009). Lehti-Shiu et al. (2009) suggested that the LRR–KD configuration first occurred in streptophytes after its split from chlorophytes and before the divergence of land plants (embryophytes). However, their result of absence of LRR– KD configuration in chlorophytes was based on limited data: only two species: Chlamydomonas reinhardtii and Ostreococcus tauri at that time. In this study, we have revisited this question by identifying domain configuration of the 12 sequenced chlorophytes genomes deposited in DOE-JGI genome portal with two methods (see ‘‘Materials and Methods’’ section and Table S4). We have added the red algae C. merolae data (Matsuzaki et al. 2004) in this investigation since red algae have been placed as the sister
123 Fig. 2 ID domain of BRI1–BRL family. a Alignment after redun- c dancy elimination at the cutoff of 90 % of sequence identity. The complete alignment is shown in Fig. S3. In each subgroup, conserved residues of ID are highlighted with dots. Blue stripes mark the column of conserved residues. Numbers on the top of alignments shows the locations of residues in the alignment. b Schematic representation of the ID domain for each group. ID domains are represented by horizontal blue stripes. Vertical bars with amino acid symbols above domains exhibit locations of conserved residues. Vertical bars at two terminals of domains exhibit locations of the first and last residues. The relationship between the three groups is shown on the left. This topology is based on the NJ tree of ID sequences but only the basal branching pattern of the three groups is shown. Bootstrap values (1,000 replicates) of the three groups are shown in branches (Color figure online)
group of green plants in the tree of life by many molecular phylogenetic studies (Adl et al. 2005). Our results support that the plant LRR–KD configuration originate in green plants after the split between them and red algae but before the split of streptophytes from chlorophytes. JM domains are found in all land plant major lineages but not in any of the 12 chlorophytes. In contrast, ID only occurs in gymnosperms and angiosperms. This indicates that in the evolutionary trajectory of BRI1–BRL genes, the acquirement of JM is prior to the acquirement of ID. If the absence of JM domain in the 12 chlorophytes correctly reflects the absence of this domain in chlorophytes, the origin of JM probably is estimated to happen after the streptophytes diverged from chlorophytes but before the divergence of liverworts from other land plants. It has been suggested that Charales and Coleochaetales have a closer relationship to land plants than other green algae (Karol et al. 2001; Qiu et al. 2006). However, whole-genome data on these lineages are not available, so we cannot decide if the origin of JM is before or after the plant colonization of land. Both KD similarity and domain configuration analysis suggest that the upper and lower bound of the origin of ID are in euphyllophytes after its split from lycophytes but before the divergence of angiosperms and extant gymnosperms. Ferns and horsetails form a major lineage diverged after lycophytes but before seed plants. Unfortunately, current accessible data is not enough to resolve the occurrence of ID in this lineage. The diversification of this three clades, and maybe the differentiation of functions (see below), can also be dated as before the differentiation of angiosperms and gymnosperms: in the RLK/Pelle-LRR-Xb-1 NJ tree (Fig. S1), all of the three major clades of BRI1–BRL family have members from conifers (bootstrap support at some nodes are low). If gymnosperms form the sister clade of angiosperms, as argued by the recent molecular studies (see review in Mathews 2009), the origin of the three clades probably can be traced to the ancestor of seed plants. If not,
123
124
the lower bound can be placed at the common ancestor of angiosperms and conifers. In short, the domain configuration evolution of BRI1– BRL family can be resolved as a three stage process (Fig. 1): first, the LRR–TM–KD configuration was ancient
123
J Mol Evol (2014) 78:118–129
to all green plants; then in streptophytes, either before or after the divergence of Coleochaetales and Charales, JM domain added in; and at last, ID appeared in the common ancestor of angiosperms and gymnosperms after differentiated from lycophytes.
J Mol Evol (2014) 78:118–129
125
Fig. 3 Amino acids ML phylogeny of BRI1–BRL genes in sequenced plant genomes. a Relationship of deep nodes and the BRI1 clade. KD of AT5G07280 is used as outgroup. b, c Phylogeny of the BRL1–3 and BRL2 clades, respectively. Bootstrap values of 100 replicates are shown in branches. Only values of 50 % or more
are shown. The format of names of leaf nodes is ‘‘three-letter species code’’ ‘‘Gene ID’’/‘‘location of the region used in tree construction’’. The correspondence between species code and the standard species names can be found in Table S1. The detailed information about Gene ID can be found in Table S3
Origin of BR Recognition via BRI1
possibility that this new gene is recruited as the receptor to initiate the BR signaling cascade. Subsequently, an interesting question is when and how many times this occurred. Although exceptions have been reported (Lynch and Wagner 2008; Nehrt et al. 2011), the orthology conjecture, i.e. orthologues carry out equivalent functions, whereas paralogues undergo functional diversification, seems applicable in general (Gabaldon and Koonin 2013). In our case, the AtBRI1 orthologs in tomato (Holton et al. 2007) have also been established as active BR receptors. In rice, evidence also suggests that OsBRI1 is a BR receptor (Yamamuro et al. 2000; Zhao et al. 2002). If the orthology conjecture is valid in our case and the orthology of BRI1 genes in Arabidopsis, tomato and rice reflects conservation of ancient function, our results suggest that recruiting of BRI1 gene as a component of BR signaling pathway has a single origin in plant evolution. We failed to detect BRI1–BRL homologs with ID in nonseed plants in this study. However, BRs have been detected throughout the plant kingdom in every species that has been examined, including nonseed plants such as fern (Equisetum arvense), liverwort (Marchantia polymorpha), and green algae (Chlorella vulgaris, Hydrodictyon
Although island regions which interrupt LRR in proteins containing LRR arrays have been widely observed, their origin, evolution, and function are little known (Matsushima et al. 2009). To date, only in very few genes, e.g. AtBRI1 in Arabidopsis, DcPSKR in Zinnia elegans and Toll in Drosophila, island regions have been proposed to be functional (Gibbard et al. 2006; Kinoshita et al. 2005; Shinohara et al. 2007). In all the three cases, the functional island regions (or ID) have been inferred playing a role in ligand interaction. In AtBRI1, the essentiality of ID (along with its downstream LRR) in BR binding has been confirmed (Kinoshita et al. 2005). Recently resolved three-dimensional structure of AtBRI1 has revealed the details of how ID functions in BR perception (Hothorn et al. 2011; She et al. 2011): ID folds back into the interior of the LRR superhelix to form a surface pocket which the brassinolide, the first isolated BR, can bind to. Since ID acquirement is the last step during the domain configuration evolution of the BRI1–BRL family, this domain gain event provides a LRR RLK gene the ability to recognize BRs and the
123
126
J Mol Evol (2014) 78:118–129
Table 1 BRI1–BRL genes in sequenced angiosperm genomes Organism
BRI1
BRL1–3
BRL2
Sum
Aquilegia coerulea
1
0
1
2
Arabidopsis lyrata
1
2
1
4
Arabidopsis thaliana
1
2
1
4
Amborella trichopoda
1
0
2
3
Brachypodium distachyon
1
1
1
3
Brassica rapa
3
2
1
6
Citrus clementina
1
1
1
3
Carica papaya
1
1
1
3
Capsella rubella
1
2
1
4
Cucumis sativus
1
0
1
2
Citrus sinensis Eucalyptus grandis
1 1
1 2
1 2
3 5
Glycine max
2
2
2
6
Linum usitatissimum
4
0
2
6
Malus domestica
3
3
3
9
Manihot esculenta
2
2
2
6
Mimulus guttatus
2
1
1
4
Medicago truncatula
1
0
0
1
Oryza sativa
1
2
1
4
Panicum virgatum
2
4
0
6
Prunus persica
1
1
1
3
Populus trichocarpa
2
2
2
6
Phaseolus vulgaris
1
1
1
3
Ricinus communis
1
1
1
3
Sorghum bicolor
1
1
1
3
Setaria italica
1
2
1
4
Thellungiella halophila Vitis vinifera
1 1
2 1
1 1
4 3
Zea mays
1
2
1
4
41
41
35
117
Total
reticulatum) (Bajguz and Tretyn 2003; Clouse 2011). Therefore, if BR signaling exists in these ‘‘lower’’ plant lineages, they should use a different BR receptor. Collectively, we estimate that the origin of the Arabidopsis BR signaling paradigm is no earlier than the divergence of lycophytes. The Consistency Between Phylogenetic and Functional Differentiation Our results show strong correlation between evolutionary pattern and gene function: it is known (Cano-Delgado et al. 2004; Clay and Nelson 2002; Li and Chory 1997; Nakamura et al. 2006) that Arabidopsis BRI1–BRL genes can be functionally categorized as two groups: (1) AtBRI1, AtBRL1 and AtBRL3 perform as receptors of BR. (2) AtBRL2 does not induce BR response but plays a role in
123
transduction of other extracellular spatial and temporal signals into downstream cell differentiation responses in provascular/procambial cells (Ceserani et al. 2009; Clay and Nelson 2002). In the BR receptor group, AtBRI1 is the major BR receptor which ubiquitously expresses in growing cells, while AtBRL1 and AtBRL3 perform as functional redundancy partners of AtBRI1 and induce cell-typespecific BR response in vascular tissues (Cano-Delgado et al. 2004). Consistent with this functional diversity, the ML phylogeny (Fig. 3) exhibits that the basal split of BRI1–BRL genes is between BRL2 and all others, and the second split is between BRI1 and BRL1–3 genes and the two splits are caused by duplications that induce neo- and sub-functionalization. At nine sites in the ID, amino acids are conserved in BR binding genes AtBRI1, AtBRL1–3, and OsBRI1, but not in the confirmed non-BR-binding gene AtBRL2. Moreover, seven out of the nine sites (i.e. 9K–11Y/F–61T–64T–68N– 69G–71S) are highly conserved (identity [90 %) in all BRI1 and BRL1–3 gene models (Fig. 2). Therefore, it is possible that one or several of these substitutions may have played a role in the functional diversification of AtBRL2. It is worth noting that substitution in ID domain may not be the (or the only) source of the functional alternation of AtBRL2—changes in other domain may also contribute to or be fully responsible for that. Further analyses are needed to reveal why AtBRL2 does not function as a BR receptor. A Possible Picture of the Evolution of BRI1–BRL Exon–Intron Structure 95 % (111 out of the 117) BRI1–BRL genes were singleexon genes. Mapping exon–intron structures into the phylogeny of BRI1–BRL family and related genes could resolve the evolution of the exon–intron structure. However, if constructing the phylogeny using all BRI1–BRL related genes (i.e. all LRR RLK genes) from the 29 sequenced species, the relatively short size of KD domain (\300 aa) and the huge number of genes (Lehti-Shiu et al. 2009) would make the tree highly unreliable. Here, we used a sampling strategy by investigating LRR RLK genes in the Arabidopsis genome. The reasons for choosing Arabidopsis thaliana were as follows: (1) its LRR genes were distributed in most of the subclades of the LRR RLK domain family (Shiu et al. 2004) and (2) this genome had so far the best gene annotation quality in plants. In this study, we only investigated exon–intron structure within the protein-coding gene ORFs but not untranslated regions because the prediction of untranslated regions is less reliable. The relationship between DNA sequences of KDs of Arabidopsis LRR RLK genes is shown in Fig. S6. Consistent with the previous results (Shiu and Bleecker 2003;
J Mol Evol (2014) 78:118–129
Shiu et al. 2004), this phylogeny suggested that BRI1–BRL family and its sister gene EMS1 (EXCESS MICROSPOROCYTES1, Ath:AT5G07280, a putative LRR RLK gene that controls somatic and reproductive cell fates in Arabidopsis Zhao et al. 2002) formed a monophyletic group (bootstrap value = 55 %) and the most intron-containing genes were diverged earlier. If the grouping of BRI1–BRL and EMS1 as a clade was valid and old LRR RLK genes were intron containing, the origin of singleexon structure of the BRI1–BRL family could be parsimoniously explained as intron loss event(s) in their common ancestor before the divergence of EMS1. However, we note that both observations have only rather poor statistical support and further studies are needed to fully resolve this question. Mapping the domain configuration of Arabidopsis LRR RLK genes to the phylogeny (Fig. S6) identified only five genes containing JM domain: the four BRI1–BRL genes and EMS1. With low statistical support, the pattern that all of the JM containing genes formed a clade indicated that the acquirement of JM might be posterior to the occurrence of the single-exon structure of BRI1–BRL family. Intron-Gain Events The other six BRI1–BRL genes with two introns were probably derived from intron gain. Flanking coding sequences of these introns were highly conservative (two examples are shown in Fig. S7) and all of these introns were located at different sites in the genes, and therefore the six introns were likely to be gained through independent events. Nature of Ancient Intron Loss Ancient intron loss event(s) that shaping current exon– intron structure of BRI1–BRL genes may have removed single or multiple introns at a time. Both reverse transcriptase-mediated (RT-mediated) intron loss (Derr 1998; Roy and Gilbert 2006) and retroposition (Brosius 1991; McCarrey and Thomas 1987) can lead to simultaneous loss of multiple introns. Retroposition usually generates an intronless copy that is located at a different locus from the original copy through the activity of retroelements like LINEs or LTR retrotransposons. RT-mediated intron loss, or gene conversion by a cDNA, however, removes introns in the original gene, but does not change its physical position in the genome. The two mechanisms can thus be tested when compared species are evolutionarily close enough so that synteny is detectable at target loci. In the BRI1–BRL case, however, it is difficult to distinguish the two mechanisms because of the rapid erosion of gene synteny in plant genome evolution.
127 Acknowledgments This work was supported by the National Basic Research Program of China (973 Project No. 2007CB814800 and 2013CB834100) and the Shanghai Leading Academic Discipline Project (No. B111). This study was also supported in part by resources and technical expertise from the Georgia Advanced Computing Resource Center, a partnership between the University of Georgia’s Office of the Vice President for Research and Office of the Vice President for Information Technology. Conflict of interest of interest.
The authors declare that they have no conflict
References Abascal F, Zardoya R, Posada D (2005) ProtTest: selection of best-fit models of protein evolution. Bioinformatics 21:2104 Adl SM, Simpson AG, Farmer MA, Andersen RA, Anderson OR, Barta JR, Bowser SS, Brugerolle G, Fensome RA, Fredericq S, James TY, Karpov S, Kugrens P, Krug J, Lane CE, Lewis LA, Lodge J, Lynn DH, Mann DG, McCourt RM, Mendoza L, Moestrup O, Mozley-Standridge SE, Nerad TA, Shearer CA, Smirnov AV, Spiegel FW, Taylor MF (2005) The new higher level classification of eukaryotes with emphasis on the taxonomy of protists. J Eukaryot Microbiol 52:399 Andrade MA, Ponting CP, Gibson TJ, Bork P (2000) Homologybased method for identification of protein repeats using statistical significance estimates. J Mol Biol 298:521 Bajguz A, Tretyn A (2003) The chemical characteristic and distribution of brassinosteroids in plants. Phytochemistry 62:1027 Brosius J (1991) Retroposons—seeds of evolution. Science 251:753 Cano-Delgado A, Yin Y, Yu C, Vafeados D, Mora-Garcia S, Cheng JC, Nam KH, Li J, Chory J (2004) BRL1 and BRL3 are novel brassinosteroid receptors that function in vascular differentiation in Arabidopsis. Development 131:5341 Capella-Gutierrez S, Silla-Martinez JM, Gabaldon T (2009) trimAL: a tool for automated alignment trimming in large-scale phylogenetic analyses. Bioinformatics 25:1972 Ceserani T, Trofka A, Gandotra N, Nelson T (2009) VH1/BRL2 receptor-like kinase interacts with vascular-specific adaptor proteins VIT and VIK to influence leaf venation. Plant J 57:1000 Chen K, Durand D, Farach-Colton M (2000) NOTUNG: a program for dating gene duplications and optimizing gene family trees. J Comput Biol 7:429 Chono M, Honda I, Zeniya H, Yoneyama K, Saisho D, Takeda K, Takatsuto S, Hoshino T, Watanabe Y (2003) A semidwarf phenotype of barley uzu results from a nucleotide substitution in the gene encoding a putative brassinosteroid receptor. Plant Physiol 133:1209 Clay NK, Nelson T (2002) VH1, a provascular cell-specific receptor kinase that influences leaf cell patterns in Arabidopsis. Plant Cell 14:2707 Clouse SD (2011) Brassinosteroid signal transduction: from receptor kinase activation to transcriptional networks regulating plant development. Plant Cell 23:1219 Clouse SD, Sasse JM (1998) BRASSINOSTEROIDS: essential regulators of plant growth and development. Annu Rev Plant Physiol Plant Mol Biol 49:427 Derr LK (1998) The involvement of cellular recombination and repair genes in RNA-mediated recombination in Saccharomyces cerevisiae. Genetics 148:937 Dolan J, Walshe K, Alsbury S, Hokamp K, O’Keeffe S, Okafuji T, Miller SF, Tear G, Mitchell KJ (2007) The extracellular leucinerich repeat superfamily; a comparative survey and analysis of
123
128 evolutionary relationships and expression patterns. BMC Genomics 8:320 Edgar RC (2004) MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Res 32:1792 Finn RD, Clements J, Eddy SR (2011) HMMER web server: interactive sequence similarity searching. Nucleic Acids Res 39:W29 Gabaldon T, Koonin EV (2013) Functional and evolutionary implications of gene orthology. Nat Rev Genet 14:360 Gibbard RJ, Morley PJ, Gay NJ (2006) Conserved features in the extracellular domain of human toll-like receptor 8 are essential for pH-dependent signaling. J Biol Chem 281:27503 Gish LA, Clark SE (2011) The RLK/Pelle family of kinases. Plant J 66:117 Holton N, Cano-Delgado A, Harrison K, Montoya T, Chory J, Bishop GJ (2007) Tomato BRASSINOSTEROID INSENSITIVE1 is required for systemin-induced root elongation in Solanum pimpinellifolium but is not essential for wound signaling. Plant Cell 19:1709 Hothorn M, Belkhadir Y, Dreux M, Dabi T, Noel JP, Wilson IA, Chory J (2011) Structural basis of steroid hormone perception by the receptor kinase BRI1. Nature 474:467 Kajava AV (1998) Structural diversity of leucine-rich repeat proteins. J Mol Biol 277:519 Karol KG, McCourt RM, Cimino MT, Delwiche CF (2001) The closest living relatives of land plants. Science 294:2351 Kim TW, Wang ZY (2010) Brassinosteroid signal transduction from receptor kinases to transcription factors. Annu Rev Plant Biol 61:681 Kinoshita T, Cano-Delgado A, Seto H, Hiranuma S, Fujioka S, Yoshida S, Chory J (2005) Binding of brassinosteroids to the extracellular domain of plant receptor kinase BRI1. Nature 433:167 Kobe B, Kajava AV (2001) The leucine-rich repeat as a protein recognition motif. Curr Opin Struct Biol 11:725 Koka CV, Cerny RE, Gardner RG, Noguchi T, Fujioka S, Takatsuto S, Yoshida S, Clouse SD (2000) A putative role for the tomato genes DUMPY and CURL-3 in brassinosteroid biosynthesis and response. Plant Physiol 122:85 Koonin EV (2005) Orthologs, paralogs, and evolutionary genomics. Annu Rev Genet 39:309 Lehti-Shiu MD, Zou C, Hanada K, Shiu SH (2009) Evolutionary history and stress regulation of plant receptor-like kinase/Pelle genes. Plant Physiol 150:12 Li J, Chory J (1997) A putative leucine-rich repeat receptor kinase involved in brassinosteroid signal transduction. Cell 90:929 Lynch VJ, Wagner GP (2008) Resurrecting the role of transcription factor change in developmental evolution. Evolution 62:2131 Mathews S (2009) Phylogenetic relationships among seed plants: persistent questions and the limits of molecular data. Am J Bot 96:228 Matsushima N, Mikami T, Tanaka T, Miyashita H, Yamada K, Kuroki Y (2009) Analyses of non-leucine-rich repeat (non-LRR) regions intervening between LRRs in proteins. Biochim Biophys Acta 1790:1217 Matsushima N, Miyashita H, Mikami T, Kuroki Y (2010) A nested leucine rich repeat (LRR) domain: the precursor of LRRs is a ten or eleven residue motif. BMC Microbiol 10:235 Matsuzaki M, Misumi O, Shin IT, Maruyama S, Takahara M, Miyagishima SY, Mori T, Nishida K, Yagisawa F, Yoshida Y, Nishimura Y, Nakao S, Kobayashi T, Momoyama Y, Higashiyama T, Minoda A, Sano M, Nomoto H, Oishi K, Hayashi H, Ohta F, Nishizaka S, Haga S, Miura S, Morishita T, Kabeya Y, Terasawa K, Suzuki Y, Ishii Y, Asakawa S, Takano H, Ohta N, Kuroiwa H, Tanaka K, Shimizu N, Sugano S, Sato N, Nozaki H, Ogasawara N, Kohara Y, Kuroiwa T (2004) Genome sequence
123
J Mol Evol (2014) 78:118–129 of the ultrasmall unicellular red alga Cyanidioschyzon merolae 10D. Nature 428:653 McCarrey JR, Thomas K (1987) Human testis-specific PGK gene lacks introns and possesses characteristics of a processed gene. Nature 326:501 Montoya T, Nomura T, Farrar K, Kaneta T, Yokota T, Bishop GJ (2002) Cloning the tomato curl3 gene highlights the putative dual role of the leucine-rich repeat receptor kinase tBRI1/SR160 in plant steroid hormone and peptide hormone signaling. Plant Cell 14:3163 Morillo SA, Tax FE (2006) Functional analysis of receptor-like kinases in monocots and dicots. Curr Opin Plant Biol 9:460 Nakamura A, Fujioka S, Sunohara H, Kamiya N, Hong Z, Inukai Y, Miura K, Takatsuto S, Yoshida S, Ueguchi-Tanaka M, Hasegawa Y, Kitano H, Matsuoka M (2006) The role of OsBRI1 and its homologous genes, OsBRL1 and OsBRL3, in rice. Plant Physiol 140:580 Nehrt NL, Clark WT, Radivojac P, Hahn MW (2011) Testing the ortholog conjecture with comparative functional genomic data from mammals. PLoS Comput Biol 7:e1002073 Nei M, Gojobori T (1986) Simple methods for estimating the numbers of synonymous and nonsynonymous nucleotide substitutions. Mol Biol Evol 3:418 Nomura T, Bishop GJ, Kaneta T, Reid JB, Chory J, Yokota T (2003) The LKA gene is a BRASSINOSTEROID INSENSITIVE 1 homolog of pea. Plant J 36:291 Posada D (2008) jModelTest: phylogenetic model averaging. Mol Biol Evol 25:1253 Punta M, Coggill PC, Eberhardt RY, Mistry J, Tate J, Boursnell C, Pang N, Forslund K, Ceric G, Clements J, Heger A, Holm L, Sonnhammer EL, Eddy SR, Bateman A, Finn RD (2012) The PFAM protein families database. Nucleic Acids Res 40:D290 Qiu YL, Li L, Wang B, Chen Z, Knoop V, Groth-Malonek M, Dombrovska O, Lee J, Kent L, Rest J, Estabrook GF, Hendry TA, Taylor DW, Testa CM, Ambros M, Crandall-Stotler B, Duff RJ, Stech M, Frey W, Quandt D, Davis CC (2006) The deepest divergences in land plants inferred from phylogenomic evidence. Proc Natl Acad Sci USA 103:15511 Roy SW, Gilbert W (2006) The evolution of spliceosomal introns: patterns, puzzles and progress. Nat Rev Genet 7:211 Saitou N, Nei M (1987) The neighbor-joining method: a new method for reconstructing phylogenetic trees. Mol Biol Evol 4:406 She J, Han Z, Kim TW, Wang J, Cheng W, Chang J, Shi S, Yang M, Wang ZY, Chai J (2011) Structural insight into brassinosteroid perception by BRI1. Nature 474:472 Shinohara H, Ogawa M, Sakagami Y, Matsubayashi Y (2007) Identification of ligand binding site of phytosulfokine receptor by on-column photoaffinity labeling. J Biol Chem 282:124 Shiu SH, Bleecker AB (2001a) Plant receptor-like kinase gene family: diversity, function, and signaling. Sci STKE 2001:re22 Shiu SH, Bleecker AB (2001b) Receptor-like kinases from Arabidopsis form a monophyletic gene family related to animal receptor kinases. Proc Natl Acad Sci USA 98:10763 Shiu SH, Bleecker AB (2003) Expansion of the receptor-like kinase/ Pelle gene family and receptor-like proteins in Arabidopsis. Plant Physiol 132:530 Shiu SH, Karlowski WM, Pan R, Tzeng YH, Mayer KF, Li WH (2004) Comparative analysis of the receptor-like kinase family in Arabidopsis and rice. Plant Cell 16:1220 Stamatakis A (2006) RAxML-VI-HPC: maximum likelihood-based phylogenetic analyses with thousands of taxa and mixed models. Bioinformatics 22:2688 Stanke M, Diekhans M, Baertsch R, Haussler D (2008) Using native and syntenically mapped cDNA alignments to improve de novo gene finding. Bioinformatics 24:637
J Mol Evol (2014) 78:118–129 Sun Y, Fokar M, Asami T, Yoshida S, Allen RD (2004) Characterization of the brassinosteroid insensitive 1 genes of cotton. Plant Mol Biol 54:221–232 Suyama M, Torrents D, Bork P (2006) PAL2NAL: robust conversion of protein sequence alignments into the corresponding codon alignments. Nucleic Acids Res 34:W609 Symons GM, Davies C, Shavrukov Y, Dry IB, Reid JB, Thomas MR (2006) Grapes on steroids. Brassinosteroids are involved in grape berry ripening. Plant Physiol 140:150 Tamura K, Peterson D, Peterson N, Stecher G, Nei M, Kumar S (2011) MEGA5: molecular evolutionary genetics analysis using maximum likelihood, evolutionary distance, and maximum parsimony methods. Mol Biol Evol 28:2731 Vert G, Nemhauser JL, Geldner N, Hong F, Chory J (2005) Molecular mechanisms of steroid hormone signaling in plants. Annu Rev Cell Dev Biol 21:177 Wang X, Li X, Meisenhelder J, Hunter T, Yoshida S, Asami T, Chory J (2005) Autoregulation and homodimerization are involved in the activation of the plant steroid receptor BRI1. Dev Cell 8:855
129 Wang ZY, Bai MY, Oh E, Zhu JY (2012) Brassinosteroid signaling network and regulation of photomorphogenesis. Annu Rev Genet 46:701 Xu G, Ma H, Nei M, Kong H (2009) Evolution of F-box genes in plants: different modes of sequence divergence and their relationships with functional diversification. Proc Natl Acad Sci USA 106:835 Yamamuro C, Ihara Y, Wu X, Noguchi T, Fujioka S, Takatsuto S, Ashikari M, Kitano H, Matsuoka M (2000) Loss of function of a rice brassinosteroid insensitive1 homolog prevents internode elongation and bending of the lamina joint. Plant Cell 12:1591 Yang Z (2007) PAML 4: phylogenetic analysis by maximum likelihood. Mol Biol Evol 24:1586 Zhao DZ, Wang GF, Speal B, Ma H (2002) The excess microsporocytes1 gene encodes a putative leucine-rich repeat receptor protein kinase that controls somatic and reproductive cell fates in the Arabidopsis anther. Genes Dev 16:2021
123
ORIGINAL ARTICLE
On the Origin and Evolution of Plant Brassinosteroid Receptor Kinases Hao Wang • Hongliang Mao
Received: 6 September 2013 / Accepted: 18 December 2013 / Published online: 27 December 2013 Ó Springer Science+Business Media New York 2013
Abstract Brassinosteroid (BR) signaling pathway is so far the best-understood receptor-kinase signaling pathway in plants. In Arabidopsis, the activation of this pathway requires binding of BRs to the receptor kinase BRASSINOSTEROID-INSENSITIVE I (AtBRI1). Although the function of AtBRI1 has been extensively studied, it is not known when the binding function emerged and how this important component of BR signaling pathway and related genes (the BRI1–BRL gene family) have evolved in plants. We define BRI1–BRL genes in sequenced plant genomes, construct profiles for critical protein domains, scan them against all accessible plant gene/EST resources, and reveal the evolution of domain configuration of this family. We also investigate its evolutionary pattern through phylogenetic analysis. The complete BR receptor domain configuration originates through two domain gain events in the ancestral receptor-like kinase: first juxtamembrane domain gained during the early diversification of land plants, and then island domain (ID) acquired in the common ancestor of angiosperms and gymnosperms after its divergence from spike moss. The 70 amino acid ID has characteristic sequences of BRI1–BRL family and this family keeps relative stable copy numbers during the history of angiosperms and the majority of duplications and losses have Electronic supplementary material The online version of this article (doi:10.1007/s00239-013-9609-5) contains supplementary material, which is available to authorized users. H. Wang H. Mao T-Life Research Center, Department of Physics, Fudan University, Shanghai 200433, People’s Republic of China H. Wang (&) Department of Genetics, University of Georgia, 120 Green Street, Athens, GA 30602, USA e-mail: [email protected]
123
occurred in terminal taxa in current taxon sampling. This study reveals important events shaping structural and functional characteristics of plant BR receptors. It answers the question of how and when BR receptors originates, which provide insights into the origin and evolution of the BR signaling pathway. Keywords Brassinosteroid BRI1 Protein domain Gene family Evolution
Introduction Brassinosteroids (BRs) are a group of plant steroid hormones that play critical roles in a wide range of developmental and physiological processes such as cell elongation, vascular differentiation, root growth, light response, stresses resistance, and senescence (Clouse and Sasse 1998; Kim and Wang 2010; Vert et al. 2005). The last two decades have observed great advances in assembling the BR signal transduction pathway (or network Wang et al. 2012) in Arabidopsis. A number of component genes involved in the BR transduction pathway have been defined and important signal transduction steps, from BR perception by the receptor kinase to the activation of the most upstream transcription factors of the BR-dependent transcriptional network, have been revealed (e.g. see reviews in Clouse 2011; Kim and Wang 2010). In early events that activate the Arabidopsis BR signaling network, the binding of BR requires the function of BR receptor kinase BRASSINOSTEROID-INSENSITIVE I (BRI1 or AtBRI1). Genetic and biochemical studies have established AtBRI1 as the major receptor of BRs (Kinoshita et al. 2005; Li and Chory 1997) in Arabidopsis, and great efforts have been made to elucidate the functional
J Mol Evol (2014) 78:118–129
regions or protein domains in this gene (see review in Kim and Wang 2010). The currently accepted domain configuration of AtBRI1 is LRR{20}–ID–LRR21–LRR{3}–TM– JM–KD–CT (Kim and Wang 2010; Vert et al. 2005), where LRR, ID, LRR21, TM, JM, KD, and CT denote leucinerich repeats, island domain, the 21st LRR domain, transmembrane region, juxtamembrane region, kinase domain, and C-terminal region, respectively. LRR21 and its upstream adjacent ID play important roles together so this LRR is marked specially. The numbers in braces are copy numbers of tandem domain units. For instance, LRR{20} means 20 tandem LRR domains. AtBRI1 in fact has 25 LRRs (Kinoshita et al. 2005), but the one located at N-terminal is irregular (She et al. 2011). If taking the irregular LRR as the first LRR, ID is located between LRR-21 and -22. In this study, we count the first regular LRR as the first LRR, following the numbering system used by several previous works (Kim and Wang 2010; Li and Chory 1997; Vert et al. 2005). Several homologs of AtBRI1 have been identified: (1) AtBRI1 orthologs have been defined in a number of crops such as tomato (Curl3/tBRI1/SR160; Koka et al. 2000; Montoya et al. 2002), rice (OsBRI1; Yamamuro et al. 2000), barley (HvBRI1; Chono et al. 2003), cotton (GhBRI1; Sun et al. 2004), grape (VvBRI1; Symons et al. 2006), and pea (LKA/PsBRI1; Nomura et al. 2003), and their function in BR perception and plant growth have been confirmed by mutational researches (see Clouse 2011; Morillo and Tax 2006 for brief reviews). (2) AtBRI1 is the major, but not the only BR receptor in Arabidopsis. Three paralogs of AtBRI1 have been cloned and two of them, AtBRL1 and AtBRL3, can also bind to BR and mediate cell-type-specific BR response and rescue the bri1 mutant when expressed under the control of the BRI1 promoter (Cano-Delgado et al. 2004). The investigation of the counterparts of AtBRI1 and AtBRL genes in rice seems to support the model that BRI1 performing as major BR receptor and some of its homologs as partial functional backup also work in rice (Nakamura et al. 2006). The structure and function of AtBRI1 have been extensively studied. In contrast, the evolution of AtBRI1 and related genes (called the BRI1–BRL gene family hereafter) is largely unrevealed. AtBRI1 belongs to LRR receptor-like kinases (LRR RLKs) which are characterized by extracellular LRR arrays, a single-pass TM and a cytoplasmic KD (Shiu and Bleecker 2001a). Though the relationship and diversity of LRR and KD have attracted many attentions and these studies have established that BRI1–BRL genes form a clade in the KD phylogenetic tree (Dolan et al. 2007; Kobe and Kajava 2001; Lehti-Shiu et al. 2009; Matsushima et al. 2009, 2010; Shiu and Bleecker 2001a, b; Shiu et al. 2004), none of these researches have focused on the BRI1–BRL family per se. To date, the
119
relationship between BRI1 and BRL genes only has been discussed in quite limited organisms (Cano-Delgado et al. 2004; Gish and Clark 2011; Nakamura et al. 2006). The pioneering work based on the phylogeny of the 10 available BRI1–BRL genes (Cano-Delgado et al. 2004) discovered that the early duplications splitting BRI2 from BRI1 and BRL1–3 occurred prior to the divergence of monocots and dicots. However, due to the lack of data, this work failed to correctly resolve the time of the split of BRL1 and BRL3 and could not discuss many other questions such as the birth and death patterns and lineage distribution of family members. In short, systematic investigations of this family with an extensive taxon sampling have not been reported yet and how and when this family originated still remains an open question. Recent genomics revolution has uncovered genomes from many major lineages of plant tree of life and deposited a large number of EST/transcriptome data for lineages without fully sequenced genomes. With the aim of illuminating the origin and evolution of structure and function of plant BRI1–BRL genes, we have performed comparative analyses of the BRI1–BRL gene family in major clades of the plant kingdom. Focusing on the 44 sequenced plant genomes [32 land plants and 12 green algae (Supporting Information Table S1)], we have identified the BRI1–BRL gene repertoire in 29 genomes using bioinformatic method and tried to answer the following questions: (1) how and when the family originated, (2) relationship between family members, and (3) pattern of gene duplication and loss. We also have discussed a possible picture of the origin of the single-exon structure of this family.
Materials and Methods Sequence Data Fully sequenced plant genomes and gene annotations were downloaded from phytozome version 8 (http://phytozome. net/). Data of chlorophytes that were deposited in JGI but not included in phytozome were downloaded from JGI genome portal (http://genome.jgi.doe.gov/). The genome and annotations of Amborella trichopoda were downloaded from Amborella Genome Database (http://www.amborella. org/). ESTs, EST assemblies, transcriptome assemblies, and cDNAs (mRNAs) of conifers as well as other nonangiosperm plants were collected from TreeGene (http://den drome.ucdavis.edu/treegenes/) and plantGDB (http://www. plantgdb.org/). A translated EST dataset was constructed by predicting coding regions of the ESTs by ORFPredictor (http://proteomics.ysu.edu/tools/OrfPredictor.html). Proteins of plants that deposited in Uniprot (http://www.uni prot.org/) were also downloaded. The combined protein set
123
120
was made of nonredundant sequences from genome annotation, translated EST dataset, and Uniprot. BRI1–BRL Gene Identification The combined protein set was scanned against PFAM v26.0 (Punta et al. 2012) to obtain all genes containing KD using the Pfam_scan.pl script (ftp://ftp.sanger.ac.uk/pub/ databases/Pfam/Tools). These genes were then assigned to subfamilies according to the similarity of their KDs with known KDs of plant RLK/Pelle genes (Lehti-Shiu et al. 2009). The previous study (Shiu and Bleecker 2003) suggested that AtBRI1 and AtBRL genes belonged to the LRR-Xb-1 subfamily (called RLK/Pelle-LRR-Xb-1 genes hereafter). All genes belonged to the RLK/Pelle-LRR-Xb-1 subfamily were extracted and the amino acids neighborjoining (NJ) tree was built using their KD sequences (see below). The genes fell within the same well-supported clade with known BRI1–BRL genes were extracted as BRI1–BRL candidates (Supporting Information Fig. S1). We manually cured the BRI1–BRL candidates from fully sequenced organisms to improve gene models. Every gene locus was re-annotated through AUGUSTUS (Stanke et al. 2008) with EST evidence. Each predicted gene model was compared to that released by genome sequencing centers and the better model was chosen based on EST support and sequence alignment quality. This manual inspection excluded nine models for further study because they were far shorter than other BRI1–BRL genes or lacked other essential domains or had no EST support. Protein Domain Profile Construction Multiple sequence alignment (MSA) of protein sequences of BRI1–BRL genes was constructed using MUSCLE (Edgar 2004) and the alignment was manually inspected. In the alignment, blocks representing ID, LRR21, JM, and KD domains were extracted according to their locations in AtBRI1 (Vert et al. 2005). Profile hidden Markov models (HMMs) of domains were then constructed by using HMMER v3.0 package (Finn et al. 2011). Domain Configuration Identification The identification of ID, JM, and KD was performed by hmmsearch (E value = 1e-5) using profile constructed above. Besides hmmsearch, PFAM profiles (v26.0) were also scanned against genome sequences of the 12 chlorophytes to detect the LRR–KD configuration in these species. The presence of TM in genes was predicted using TMHMM webserver (http://www.cbs.dtu.dk/services/ TMHMM/).
123
J Mol Evol (2014) 78:118–129
Phylogenetic Analysis KD regions were aligned using MUSCLE, and the NJ tree (Saitou and Nei 1987) of all RLK/Pelle-LRR-Xb-1 genes was built using MEGA 5 (Tamura et al. 2011) with the following parameters: Poisson model; uniform rate; and pairwise deletion of gaps/missing data. Here, we excluded sequences shorter than 140 aa (50 % of average size of experimentally verified BRI1–BRL genes) and so the size of KD sequences used in this NJ tree construction was from 140 to 355 aa. Amino acid maximum likelihood (ML) phylogenetic tree of BRI1–BRL genes from sequenced angiosperms was constructed by RAxML (Stamatakis 2006) using conserved region stretching from ID to KD (ID–LRR{4}–TM–JM– KD; sequence lengths 536–648 aa). Sequence alignments were generated by MUSCLE and low quality regions were excluded from alignment using trimAL (Capella-Gutierrez et al. 2009) with the option ‘‘automated1’’. Protein model selection was performed by Prottest 3.0 (Abascal et al. 2005). Parameters of ML tree construction are as follows: JTT ? I ? G model; duplications of rapid bootstrap = 100. In 227 Arabidopsis genes of which KDs belong to the LRR RLK domain group, 21 highly diverged ones (could not calculate valid evolutionary distance with others) were excluded. The DNA MSA of the coding sequences of KD domain (sequence lengths 417–900 nt) was constructed with MUSCLE and trimmed with trimAL. ML tree was built by RAxML with the following parameter settings: model = GTR ? I ? G; duplications of rapid bootstrap = 100. Model selection was performed by jModeltest (Posada 2008). Gene Duplication–Loss History The species tree used in this study was constructed by modifying Phytozome 8 plant tree of life: A. trichopoda was added as sister taxon of monocots and dicots and Panicum virgatum as sister species of Setaria italica. Gene tree and species tree reconciliation was performed by Notung 2.6 (Chen et al. 2000). When reconstructing the gene duplication–loss history, weakly supported edges (bootstrap value \90 %) of the gene tree were allowed to be rearranged to minimize duplication and loss events. Selection Analysis Coding DNA sequence alignments of paralogous pairs were obtained under the guidance of protein alignments using PAL2NAL (Suyama et al. 2006). Dn, Ds, and Dn/Ds values were calculated by the yn00 program of the PAML
J Mol Evol (2014) 78:118–129
package (Yang 2007) according to the Nei–Gojobori method (Nei and Gojobori 1986).
Results BRI1–BRL Genes in Sequenced Genomes Phylogenetically, the BRI1–BRL gene family, in this study, was defined as the sub-clade of RLK/Pelle-LRR-Xb1 genes (Lehti-Shiu et al. 2009) that included all of the previously cloned BRI1–BRL genes. We built a NJ tree for all detected KD in the combined plant protein set made of whole gene repertoire of the 44 sequenced genomes, all plant proteins in the Uniprot database and all translated ESTs of nonangiosperm plants (see ‘‘Materials and Methods’’ section for details). In the tree (Fig. S1), the BRI1–BRL clade (bootstrap value = 95 %) included a total of 220 sequences, with 136 being full-length BRI1–BRL gene models (Table S2). Here, full-length means the open reading frame containing both a start and stop codon. 117 out of the 136 full-length genes were from 29 sequenced species (Table S3). Our manual inspection predicted 4 novel models in the apple genome (Malus domestica) and suggested that 35 gene models should be modified based on expression and/or conservation evidence. The correspondence between corrected models and automatic gene annotation was shown in Table S3. The other 19 full-length BRI1–BRL genes were from Uniprot. They belonged to organisms without whole-genome sequences when this study was done (Table S2). It is worth noting that all of the BRI1–BRL genes are from angiosperms and gymnosperms. Domain Configuration of BRI1–BRL Family The MSA of the BRI1–BRL family showed that most domains were quite conservative (Fig. S2). Low conservation was found at the first several LRRs, CT, and internal region of TM. Overall, LRR10–24 (covering ID) and KD were the two most conserved parts in the entire alignment. We compared the domain configuration of AtBRI1 with other previously identified BRI1–BRL genes and found all of them had the identical configuration with AtBRI1. Therefore, we constructed profiles of the LRR, ID, JM, and KD domains (see ‘‘Materials and Methods’’ section) and used them to scan the above 220 BRI1–BRL genes. 152 sequences exhibited a domain configuration like ‘‘LRR array’’–ID–‘‘LRR array’’–TM–JM–KD (Table S2), including 136 full-length genes (117 in sequenced species ? 19 in other species from Uniprot), 11 partial proteins from conifer EST/transcriptome assemblies, and 5 partial proteins from Uniprot. The other 68 sequences had KD but not ID: 23 from Uniprot, 6 from plantGDB, and 39
121
from TreeGene EST/transcriptome assemblies. Unlike protein sequences, several EST/transcriptome assemblies might come from the same gene and thus the number of EST/transcriptome assemblies might not correctly reflect number of genes. However, the occurrence of a certain domain in these assemblies was sufficient to confirm their occurrence in corresponding species. In this scan, we identified TM domains through the TMHMM webserver, but not through domain profile analysis because of low conservation at sequence level. Sequences of the CT region were also highly divergent (Fig. S2) and the instability of CT was reported previously (Xu et al. 2009). Therefore, although it might plays an inhibitory role in BRI1 function (Wang et al. 2005), CT was not suitable for profile analysis and excluded from this study. In short, all full-length genes had a domain configuration like ‘‘LRR array’’–ID–‘‘LRR array’’–TM–JM–KD, so this was the valid definition of BRI1–BRL gene family in term of domain configuration. Presence and Absence of Domains in Major Lineages of the Plant Kingdom We refined the profile for the four domains (LRR, ID, JM, and KD) based on the MSA of the 136 full-length BRI1– BRL genes and scanned them against the combined plant protein set. The results showed that (1) the LRR–KD combination was detected in 22 genes from 6 chlorophytes (Table S4). In contrast, we failed to detect the LRR–KD configuration in the sequenced red algae (Cyanidioschyzon merolae). In 13 cases, LRR occurred as long tandem arrays (C5 LRR units). (2) TM domain was located in between some (nine cases) but not all of the LRR–KD structure. (3) JM domain was not detected in any chlorophyte, but was found in all available major groups of land plants including liverworts, mosses, lycophytes, and seed plants such as cycads, gnetophytes, conifers (Table S5), and angiosperms. We found that whenever full-length proteins are available, their JMs were found co-occurring with LRR array, TM, and KD, but not always with ID. (4) IDs were only detected in gymnosperms (conifers and gnetophytes) and angiosperms. We found that a total of 29 gymnosperm EST/transcriptome assemblies contained IDs, with 11 of them showing complete BRI1–BRL gene domain configuration (Table S6). In contrast, this domain configuration analysis failed to detect BRI1–BRL members in plant genomes that diverged before seed plants and this result was consistent with the KD phylogenetic analysis described above. In summary, according to current available data, the first occurrence of LRR–TM–KD, LRR–TM–JM– KD, and LRR–ID–LRR–TM–JM–KD were observed in chloroplasts, liverworts, and the common ancestor of seed plants, respectively (Fig. 1).
123
122
Fig. 1 Domain configuration evolution of BRI1–BRL family. The relationship within seed plants remains unresolved because of controversial results from different datasets and methods (see e.g. Mathews 2009). The three stages of BRI1–BRL domain configuration evolution are mapped to the tree. Arrows point to the internal nodes which are lower bounds of the first presence of configurations. Right table shows the presence/absence of domains. ‘‘?’’: presence and ‘‘-’’: absence
ID Was the Diagnostic Domain of BRI1–BRL Genes In all full-length sequences within the combined plant protein set, we found that whenever ID was present, the entire BRI1–BRL domain configuration, i.e. ‘‘LRR array’’ –ID–‘‘LRR array’’–TM–JM–KD, occurred automatically. Furthermore, we found that if a sequence contains both ID and KD, the KD fell in the BRI1–BRL clade. These observations indicate that domain configuration and phylogenetic relationship of KD give equivalent definitions of BRI1–BRL family and ID is the domain only present in the BRI1–BRL family. This family now can be described as plant RLKs containing ID.
Characteristics of ID Domain The investigation of the alignment profile of ID domain of BRI1–BRL genes (Figs. 2a; S3) identified six residues conserved in all genes: C23–G27–L29–E31–C49–Y57, where letters were codes of residues and numbers were the locations of residues in the alignment. Besides the CGLECY hexad, another six highly conserved residues or motifs (i.e., occurring in [95 % of the sequences) were also revealed within ID domain (Fig. 2a). Moreover, the ID domain sequences could be further categorized as three groups according to their conservation pattern (Fig. 2b). The three ID groups were congruent to the three major clades of the BRI1–BRL family (Fig. 2b and see below).
Phylogenetics of Plant BRI1–BRL Genes The ML phylogeny of BRI1–BRL genes from the 29 sequenced angiosperms (Fig. 3) supported with 100 %
123
J Mol Evol (2014) 78:118–129
bootstrap value that the BRI1–BRL family was composed of three major clades: the basal split was between BRL2 and all others, and the second split was between BRI1 and BRL1–3 genes. Here, based on the names of the Arabidopsis and rice members, we called the three clades as BRI1, BRL1–3 and BRL2, respectively. Reconciliation the BRI1–BRL gene tree with sequenced angiosperm species tree (Fig. S4) suggested that the three major clades were derived from two gene duplication events before the divergence of angiosperms (Fig. S5). These results were consistent with the previous results from far less sequences (Cano-Delgado et al. 2004; Gish and Clark 2011). We investigated gene duplication and loss pattern of the family. (1) the majority of duplications (85 %, 28 out the 33) and losses (77 %, 10 out of 13) happened after the divergence of leaf nodes and their closest relative species in current taxon sampling (Fig. S5). This indicated that no drastic expansion or contraction of gene number occurred in the ancestral nodes. (2) Copy number variation of BRI1– BRL genes was found in all the three clades (Table 1) with the highest number of genes in M. domestica (nine; three in each clade) and lowest in Medicago truncatula (one from the BRI1 clade). However, 62 % (18 out of 29) species had three or four genes and, except for M. domestica, all variations of gene copy number within species were within mean ± 2 9 SD. In this sense, gene numbers showed no strong organism bias. (3) The three major clades had similar numbers of genes (41 in BRI1, 41 in BRL1–3, and 35 in BRL2). These observations support that gene numbers were quite stable during the evolution of this gene family. According to the phylogeny of BRL1–3 (Fig. 3b), the Arabidopsis and rice BRL1 and BRL3 genes had independent origins. AtBRL1 and AtBRL3 were derived from a duplication in the common ancestor of Brassicaceae, which left two descendants in each of the five sequenced Brassicaceae organisms. Rice BRL1 and BRL3 originated through a duplication in the common ancestor of grass after divergence from dicots and both descendants were also preserved in the most investigated grass species, except for Brachypodium. In switch grass (P. virgatum), a second round of duplications occurred independently in each of the two copies and resulted four BRL1–3 genes. Selection on Paralogs The 28 terminal species duplications generated 22 groups of in-paralogs (Koonin 2005): 17 pairs related by 1 duplication, 4 triplets by 2, and 1 quartets by 3 duplications (Fig. S5). Investigating selection pressure using Nei– Gojobori method exhibited Dn/Ds values \1 in all withingroup gene pairs (Table S7) and detected no positive selection.
J Mol Evol (2014) 78:118–129
Discussion Origin of the BRI1–BRL Family: A Three-Stage Process of Domain Gain The evolution of the domain configuration provides a scenario of gene family evolution at the level of organization of structural and/or functional units. Our results exhibit a perfect consistency between phylogeny and domain configuration when defining the BRI1–BRL gene family. If the origin of the family is marked by formation of the ‘‘LRR array’’–ID–‘‘LRR array’’–TM–JM–KD domain configuration, we have revealed a stepwise domain acquirement process in its ‘‘pre-histoy’’. LRRs present in numerous proteins from all major branches of the tree of life and with diverse functions (Kobe and Kajava 2001). To date, at least seven groups of LRRs have been identified and at least most LRRs have been found following the so-called the mutual exclusive rule, i.e., LRRs from different groups are found to never occur simultaneously in the same protein. Although phylogenetic relationships of the seven groups have not been well resolved yet (Andrade et al. 2000; Kajava 1998), it is clear that LRRs presented in BRI1–BRL genes belong to the plant-specific group that are found in plants and protists (Kobe and Kajava 2001). The ancient origin of KDs of BRI1–BRL genes has been uncovered for more than 10 years: this domain belongs to plant RLKs domain family, which together with animal Pelle and related cytoplasmic KDs form the RLK/Pelle domain family. The RLK/Pelle domain family, animal receptor serine/threonine kinases, animal receptor tyrosine kinases, and Raf kinases form a clade within the serine/ threonine/tyrosine kinases (Shiu and Bleecker 2001b). The domain configuration of RLKs is defined as ‘‘extracelluar domain’’–KD, where extracellular domain represents a wide range of domains including LRR (Shiu and Bleecker 2001a). It has been suggested that such domain combination is the structural basis of new signaling pathway evolution (Lehti-Shiu et al. 2009). Lehti-Shiu et al. (2009) suggested that the LRR–KD configuration first occurred in streptophytes after its split from chlorophytes and before the divergence of land plants (embryophytes). However, their result of absence of LRR– KD configuration in chlorophytes was based on limited data: only two species: Chlamydomonas reinhardtii and Ostreococcus tauri at that time. In this study, we have revisited this question by identifying domain configuration of the 12 sequenced chlorophytes genomes deposited in DOE-JGI genome portal with two methods (see ‘‘Materials and Methods’’ section and Table S4). We have added the red algae C. merolae data (Matsuzaki et al. 2004) in this investigation since red algae have been placed as the sister
123 Fig. 2 ID domain of BRI1–BRL family. a Alignment after redun- c dancy elimination at the cutoff of 90 % of sequence identity. The complete alignment is shown in Fig. S3. In each subgroup, conserved residues of ID are highlighted with dots. Blue stripes mark the column of conserved residues. Numbers on the top of alignments shows the locations of residues in the alignment. b Schematic representation of the ID domain for each group. ID domains are represented by horizontal blue stripes. Vertical bars with amino acid symbols above domains exhibit locations of conserved residues. Vertical bars at two terminals of domains exhibit locations of the first and last residues. The relationship between the three groups is shown on the left. This topology is based on the NJ tree of ID sequences but only the basal branching pattern of the three groups is shown. Bootstrap values (1,000 replicates) of the three groups are shown in branches (Color figure online)
group of green plants in the tree of life by many molecular phylogenetic studies (Adl et al. 2005). Our results support that the plant LRR–KD configuration originate in green plants after the split between them and red algae but before the split of streptophytes from chlorophytes. JM domains are found in all land plant major lineages but not in any of the 12 chlorophytes. In contrast, ID only occurs in gymnosperms and angiosperms. This indicates that in the evolutionary trajectory of BRI1–BRL genes, the acquirement of JM is prior to the acquirement of ID. If the absence of JM domain in the 12 chlorophytes correctly reflects the absence of this domain in chlorophytes, the origin of JM probably is estimated to happen after the streptophytes diverged from chlorophytes but before the divergence of liverworts from other land plants. It has been suggested that Charales and Coleochaetales have a closer relationship to land plants than other green algae (Karol et al. 2001; Qiu et al. 2006). However, whole-genome data on these lineages are not available, so we cannot decide if the origin of JM is before or after the plant colonization of land. Both KD similarity and domain configuration analysis suggest that the upper and lower bound of the origin of ID are in euphyllophytes after its split from lycophytes but before the divergence of angiosperms and extant gymnosperms. Ferns and horsetails form a major lineage diverged after lycophytes but before seed plants. Unfortunately, current accessible data is not enough to resolve the occurrence of ID in this lineage. The diversification of this three clades, and maybe the differentiation of functions (see below), can also be dated as before the differentiation of angiosperms and gymnosperms: in the RLK/Pelle-LRR-Xb-1 NJ tree (Fig. S1), all of the three major clades of BRI1–BRL family have members from conifers (bootstrap support at some nodes are low). If gymnosperms form the sister clade of angiosperms, as argued by the recent molecular studies (see review in Mathews 2009), the origin of the three clades probably can be traced to the ancestor of seed plants. If not,
123
124
the lower bound can be placed at the common ancestor of angiosperms and conifers. In short, the domain configuration evolution of BRI1– BRL family can be resolved as a three stage process (Fig. 1): first, the LRR–TM–KD configuration was ancient
123
J Mol Evol (2014) 78:118–129
to all green plants; then in streptophytes, either before or after the divergence of Coleochaetales and Charales, JM domain added in; and at last, ID appeared in the common ancestor of angiosperms and gymnosperms after differentiated from lycophytes.
J Mol Evol (2014) 78:118–129
125
Fig. 3 Amino acids ML phylogeny of BRI1–BRL genes in sequenced plant genomes. a Relationship of deep nodes and the BRI1 clade. KD of AT5G07280 is used as outgroup. b, c Phylogeny of the BRL1–3 and BRL2 clades, respectively. Bootstrap values of 100 replicates are shown in branches. Only values of 50 % or more
are shown. The format of names of leaf nodes is ‘‘three-letter species code’’ ‘‘Gene ID’’/‘‘location of the region used in tree construction’’. The correspondence between species code and the standard species names can be found in Table S1. The detailed information about Gene ID can be found in Table S3
Origin of BR Recognition via BRI1
possibility that this new gene is recruited as the receptor to initiate the BR signaling cascade. Subsequently, an interesting question is when and how many times this occurred. Although exceptions have been reported (Lynch and Wagner 2008; Nehrt et al. 2011), the orthology conjecture, i.e. orthologues carry out equivalent functions, whereas paralogues undergo functional diversification, seems applicable in general (Gabaldon and Koonin 2013). In our case, the AtBRI1 orthologs in tomato (Holton et al. 2007) have also been established as active BR receptors. In rice, evidence also suggests that OsBRI1 is a BR receptor (Yamamuro et al. 2000; Zhao et al. 2002). If the orthology conjecture is valid in our case and the orthology of BRI1 genes in Arabidopsis, tomato and rice reflects conservation of ancient function, our results suggest that recruiting of BRI1 gene as a component of BR signaling pathway has a single origin in plant evolution. We failed to detect BRI1–BRL homologs with ID in nonseed plants in this study. However, BRs have been detected throughout the plant kingdom in every species that has been examined, including nonseed plants such as fern (Equisetum arvense), liverwort (Marchantia polymorpha), and green algae (Chlorella vulgaris, Hydrodictyon
Although island regions which interrupt LRR in proteins containing LRR arrays have been widely observed, their origin, evolution, and function are little known (Matsushima et al. 2009). To date, only in very few genes, e.g. AtBRI1 in Arabidopsis, DcPSKR in Zinnia elegans and Toll in Drosophila, island regions have been proposed to be functional (Gibbard et al. 2006; Kinoshita et al. 2005; Shinohara et al. 2007). In all the three cases, the functional island regions (or ID) have been inferred playing a role in ligand interaction. In AtBRI1, the essentiality of ID (along with its downstream LRR) in BR binding has been confirmed (Kinoshita et al. 2005). Recently resolved three-dimensional structure of AtBRI1 has revealed the details of how ID functions in BR perception (Hothorn et al. 2011; She et al. 2011): ID folds back into the interior of the LRR superhelix to form a surface pocket which the brassinolide, the first isolated BR, can bind to. Since ID acquirement is the last step during the domain configuration evolution of the BRI1–BRL family, this domain gain event provides a LRR RLK gene the ability to recognize BRs and the
123
126
J Mol Evol (2014) 78:118–129
Table 1 BRI1–BRL genes in sequenced angiosperm genomes Organism
BRI1
BRL1–3
BRL2
Sum
Aquilegia coerulea
1
0
1
2
Arabidopsis lyrata
1
2
1
4
Arabidopsis thaliana
1
2
1
4
Amborella trichopoda
1
0
2
3
Brachypodium distachyon
1
1
1
3
Brassica rapa
3
2
1
6
Citrus clementina
1
1
1
3
Carica papaya
1
1
1
3
Capsella rubella
1
2
1
4
Cucumis sativus
1
0
1
2
Citrus sinensis Eucalyptus grandis
1 1
1 2
1 2
3 5
Glycine max
2
2
2
6
Linum usitatissimum
4
0
2
6
Malus domestica
3
3
3
9
Manihot esculenta
2
2
2
6
Mimulus guttatus
2
1
1
4
Medicago truncatula
1
0
0
1
Oryza sativa
1
2
1
4
Panicum virgatum
2
4
0
6
Prunus persica
1
1
1
3
Populus trichocarpa
2
2
2
6
Phaseolus vulgaris
1
1
1
3
Ricinus communis
1
1
1
3
Sorghum bicolor
1
1
1
3
Setaria italica
1
2
1
4
Thellungiella halophila Vitis vinifera
1 1
2 1
1 1
4 3
Zea mays
1
2
1
4
41
41
35
117
Total
reticulatum) (Bajguz and Tretyn 2003; Clouse 2011). Therefore, if BR signaling exists in these ‘‘lower’’ plant lineages, they should use a different BR receptor. Collectively, we estimate that the origin of the Arabidopsis BR signaling paradigm is no earlier than the divergence of lycophytes. The Consistency Between Phylogenetic and Functional Differentiation Our results show strong correlation between evolutionary pattern and gene function: it is known (Cano-Delgado et al. 2004; Clay and Nelson 2002; Li and Chory 1997; Nakamura et al. 2006) that Arabidopsis BRI1–BRL genes can be functionally categorized as two groups: (1) AtBRI1, AtBRL1 and AtBRL3 perform as receptors of BR. (2) AtBRL2 does not induce BR response but plays a role in
123
transduction of other extracellular spatial and temporal signals into downstream cell differentiation responses in provascular/procambial cells (Ceserani et al. 2009; Clay and Nelson 2002). In the BR receptor group, AtBRI1 is the major BR receptor which ubiquitously expresses in growing cells, while AtBRL1 and AtBRL3 perform as functional redundancy partners of AtBRI1 and induce cell-typespecific BR response in vascular tissues (Cano-Delgado et al. 2004). Consistent with this functional diversity, the ML phylogeny (Fig. 3) exhibits that the basal split of BRI1–BRL genes is between BRL2 and all others, and the second split is between BRI1 and BRL1–3 genes and the two splits are caused by duplications that induce neo- and sub-functionalization. At nine sites in the ID, amino acids are conserved in BR binding genes AtBRI1, AtBRL1–3, and OsBRI1, but not in the confirmed non-BR-binding gene AtBRL2. Moreover, seven out of the nine sites (i.e. 9K–11Y/F–61T–64T–68N– 69G–71S) are highly conserved (identity [90 %) in all BRI1 and BRL1–3 gene models (Fig. 2). Therefore, it is possible that one or several of these substitutions may have played a role in the functional diversification of AtBRL2. It is worth noting that substitution in ID domain may not be the (or the only) source of the functional alternation of AtBRL2—changes in other domain may also contribute to or be fully responsible for that. Further analyses are needed to reveal why AtBRL2 does not function as a BR receptor. A Possible Picture of the Evolution of BRI1–BRL Exon–Intron Structure 95 % (111 out of the 117) BRI1–BRL genes were singleexon genes. Mapping exon–intron structures into the phylogeny of BRI1–BRL family and related genes could resolve the evolution of the exon–intron structure. However, if constructing the phylogeny using all BRI1–BRL related genes (i.e. all LRR RLK genes) from the 29 sequenced species, the relatively short size of KD domain (\300 aa) and the huge number of genes (Lehti-Shiu et al. 2009) would make the tree highly unreliable. Here, we used a sampling strategy by investigating LRR RLK genes in the Arabidopsis genome. The reasons for choosing Arabidopsis thaliana were as follows: (1) its LRR genes were distributed in most of the subclades of the LRR RLK domain family (Shiu et al. 2004) and (2) this genome had so far the best gene annotation quality in plants. In this study, we only investigated exon–intron structure within the protein-coding gene ORFs but not untranslated regions because the prediction of untranslated regions is less reliable. The relationship between DNA sequences of KDs of Arabidopsis LRR RLK genes is shown in Fig. S6. Consistent with the previous results (Shiu and Bleecker 2003;
J Mol Evol (2014) 78:118–129
Shiu et al. 2004), this phylogeny suggested that BRI1–BRL family and its sister gene EMS1 (EXCESS MICROSPOROCYTES1, Ath:AT5G07280, a putative LRR RLK gene that controls somatic and reproductive cell fates in Arabidopsis Zhao et al. 2002) formed a monophyletic group (bootstrap value = 55 %) and the most intron-containing genes were diverged earlier. If the grouping of BRI1–BRL and EMS1 as a clade was valid and old LRR RLK genes were intron containing, the origin of singleexon structure of the BRI1–BRL family could be parsimoniously explained as intron loss event(s) in their common ancestor before the divergence of EMS1. However, we note that both observations have only rather poor statistical support and further studies are needed to fully resolve this question. Mapping the domain configuration of Arabidopsis LRR RLK genes to the phylogeny (Fig. S6) identified only five genes containing JM domain: the four BRI1–BRL genes and EMS1. With low statistical support, the pattern that all of the JM containing genes formed a clade indicated that the acquirement of JM might be posterior to the occurrence of the single-exon structure of BRI1–BRL family. Intron-Gain Events The other six BRI1–BRL genes with two introns were probably derived from intron gain. Flanking coding sequences of these introns were highly conservative (two examples are shown in Fig. S7) and all of these introns were located at different sites in the genes, and therefore the six introns were likely to be gained through independent events. Nature of Ancient Intron Loss Ancient intron loss event(s) that shaping current exon– intron structure of BRI1–BRL genes may have removed single or multiple introns at a time. Both reverse transcriptase-mediated (RT-mediated) intron loss (Derr 1998; Roy and Gilbert 2006) and retroposition (Brosius 1991; McCarrey and Thomas 1987) can lead to simultaneous loss of multiple introns. Retroposition usually generates an intronless copy that is located at a different locus from the original copy through the activity of retroelements like LINEs or LTR retrotransposons. RT-mediated intron loss, or gene conversion by a cDNA, however, removes introns in the original gene, but does not change its physical position in the genome. The two mechanisms can thus be tested when compared species are evolutionarily close enough so that synteny is detectable at target loci. In the BRI1–BRL case, however, it is difficult to distinguish the two mechanisms because of the rapid erosion of gene synteny in plant genome evolution.
127 Acknowledgments This work was supported by the National Basic Research Program of China (973 Project No. 2007CB814800 and 2013CB834100) and the Shanghai Leading Academic Discipline Project (No. B111). This study was also supported in part by resources and technical expertise from the Georgia Advanced Computing Resource Center, a partnership between the University of Georgia’s Office of the Vice President for Research and Office of the Vice President for Information Technology. Conflict of interest of interest.
The authors declare that they have no conflict
References Abascal F, Zardoya R, Posada D (2005) ProtTest: selection of best-fit models of protein evolution. Bioinformatics 21:2104 Adl SM, Simpson AG, Farmer MA, Andersen RA, Anderson OR, Barta JR, Bowser SS, Brugerolle G, Fensome RA, Fredericq S, James TY, Karpov S, Kugrens P, Krug J, Lane CE, Lewis LA, Lodge J, Lynn DH, Mann DG, McCourt RM, Mendoza L, Moestrup O, Mozley-Standridge SE, Nerad TA, Shearer CA, Smirnov AV, Spiegel FW, Taylor MF (2005) The new higher level classification of eukaryotes with emphasis on the taxonomy of protists. J Eukaryot Microbiol 52:399 Andrade MA, Ponting CP, Gibson TJ, Bork P (2000) Homologybased method for identification of protein repeats using statistical significance estimates. J Mol Biol 298:521 Bajguz A, Tretyn A (2003) The chemical characteristic and distribution of brassinosteroids in plants. Phytochemistry 62:1027 Brosius J (1991) Retroposons—seeds of evolution. Science 251:753 Cano-Delgado A, Yin Y, Yu C, Vafeados D, Mora-Garcia S, Cheng JC, Nam KH, Li J, Chory J (2004) BRL1 and BRL3 are novel brassinosteroid receptors that function in vascular differentiation in Arabidopsis. Development 131:5341 Capella-Gutierrez S, Silla-Martinez JM, Gabaldon T (2009) trimAL: a tool for automated alignment trimming in large-scale phylogenetic analyses. Bioinformatics 25:1972 Ceserani T, Trofka A, Gandotra N, Nelson T (2009) VH1/BRL2 receptor-like kinase interacts with vascular-specific adaptor proteins VIT and VIK to influence leaf venation. Plant J 57:1000 Chen K, Durand D, Farach-Colton M (2000) NOTUNG: a program for dating gene duplications and optimizing gene family trees. J Comput Biol 7:429 Chono M, Honda I, Zeniya H, Yoneyama K, Saisho D, Takeda K, Takatsuto S, Hoshino T, Watanabe Y (2003) A semidwarf phenotype of barley uzu results from a nucleotide substitution in the gene encoding a putative brassinosteroid receptor. Plant Physiol 133:1209 Clay NK, Nelson T (2002) VH1, a provascular cell-specific receptor kinase that influences leaf cell patterns in Arabidopsis. Plant Cell 14:2707 Clouse SD (2011) Brassinosteroid signal transduction: from receptor kinase activation to transcriptional networks regulating plant development. Plant Cell 23:1219 Clouse SD, Sasse JM (1998) BRASSINOSTEROIDS: essential regulators of plant growth and development. Annu Rev Plant Physiol Plant Mol Biol 49:427 Derr LK (1998) The involvement of cellular recombination and repair genes in RNA-mediated recombination in Saccharomyces cerevisiae. Genetics 148:937 Dolan J, Walshe K, Alsbury S, Hokamp K, O’Keeffe S, Okafuji T, Miller SF, Tear G, Mitchell KJ (2007) The extracellular leucinerich repeat superfamily; a comparative survey and analysis of
123
128 evolutionary relationships and expression patterns. BMC Genomics 8:320 Edgar RC (2004) MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Res 32:1792 Finn RD, Clements J, Eddy SR (2011) HMMER web server: interactive sequence similarity searching. Nucleic Acids Res 39:W29 Gabaldon T, Koonin EV (2013) Functional and evolutionary implications of gene orthology. Nat Rev Genet 14:360 Gibbard RJ, Morley PJ, Gay NJ (2006) Conserved features in the extracellular domain of human toll-like receptor 8 are essential for pH-dependent signaling. J Biol Chem 281:27503 Gish LA, Clark SE (2011) The RLK/Pelle family of kinases. Plant J 66:117 Holton N, Cano-Delgado A, Harrison K, Montoya T, Chory J, Bishop GJ (2007) Tomato BRASSINOSTEROID INSENSITIVE1 is required for systemin-induced root elongation in Solanum pimpinellifolium but is not essential for wound signaling. Plant Cell 19:1709 Hothorn M, Belkhadir Y, Dreux M, Dabi T, Noel JP, Wilson IA, Chory J (2011) Structural basis of steroid hormone perception by the receptor kinase BRI1. Nature 474:467 Kajava AV (1998) Structural diversity of leucine-rich repeat proteins. J Mol Biol 277:519 Karol KG, McCourt RM, Cimino MT, Delwiche CF (2001) The closest living relatives of land plants. Science 294:2351 Kim TW, Wang ZY (2010) Brassinosteroid signal transduction from receptor kinases to transcription factors. Annu Rev Plant Biol 61:681 Kinoshita T, Cano-Delgado A, Seto H, Hiranuma S, Fujioka S, Yoshida S, Chory J (2005) Binding of brassinosteroids to the extracellular domain of plant receptor kinase BRI1. Nature 433:167 Kobe B, Kajava AV (2001) The leucine-rich repeat as a protein recognition motif. Curr Opin Struct Biol 11:725 Koka CV, Cerny RE, Gardner RG, Noguchi T, Fujioka S, Takatsuto S, Yoshida S, Clouse SD (2000) A putative role for the tomato genes DUMPY and CURL-3 in brassinosteroid biosynthesis and response. Plant Physiol 122:85 Koonin EV (2005) Orthologs, paralogs, and evolutionary genomics. Annu Rev Genet 39:309 Lehti-Shiu MD, Zou C, Hanada K, Shiu SH (2009) Evolutionary history and stress regulation of plant receptor-like kinase/Pelle genes. Plant Physiol 150:12 Li J, Chory J (1997) A putative leucine-rich repeat receptor kinase involved in brassinosteroid signal transduction. Cell 90:929 Lynch VJ, Wagner GP (2008) Resurrecting the role of transcription factor change in developmental evolution. Evolution 62:2131 Mathews S (2009) Phylogenetic relationships among seed plants: persistent questions and the limits of molecular data. Am J Bot 96:228 Matsushima N, Mikami T, Tanaka T, Miyashita H, Yamada K, Kuroki Y (2009) Analyses of non-leucine-rich repeat (non-LRR) regions intervening between LRRs in proteins. Biochim Biophys Acta 1790:1217 Matsushima N, Miyashita H, Mikami T, Kuroki Y (2010) A nested leucine rich repeat (LRR) domain: the precursor of LRRs is a ten or eleven residue motif. BMC Microbiol 10:235 Matsuzaki M, Misumi O, Shin IT, Maruyama S, Takahara M, Miyagishima SY, Mori T, Nishida K, Yagisawa F, Yoshida Y, Nishimura Y, Nakao S, Kobayashi T, Momoyama Y, Higashiyama T, Minoda A, Sano M, Nomoto H, Oishi K, Hayashi H, Ohta F, Nishizaka S, Haga S, Miura S, Morishita T, Kabeya Y, Terasawa K, Suzuki Y, Ishii Y, Asakawa S, Takano H, Ohta N, Kuroiwa H, Tanaka K, Shimizu N, Sugano S, Sato N, Nozaki H, Ogasawara N, Kohara Y, Kuroiwa T (2004) Genome sequence
123
J Mol Evol (2014) 78:118–129 of the ultrasmall unicellular red alga Cyanidioschyzon merolae 10D. Nature 428:653 McCarrey JR, Thomas K (1987) Human testis-specific PGK gene lacks introns and possesses characteristics of a processed gene. Nature 326:501 Montoya T, Nomura T, Farrar K, Kaneta T, Yokota T, Bishop GJ (2002) Cloning the tomato curl3 gene highlights the putative dual role of the leucine-rich repeat receptor kinase tBRI1/SR160 in plant steroid hormone and peptide hormone signaling. Plant Cell 14:3163 Morillo SA, Tax FE (2006) Functional analysis of receptor-like kinases in monocots and dicots. Curr Opin Plant Biol 9:460 Nakamura A, Fujioka S, Sunohara H, Kamiya N, Hong Z, Inukai Y, Miura K, Takatsuto S, Yoshida S, Ueguchi-Tanaka M, Hasegawa Y, Kitano H, Matsuoka M (2006) The role of OsBRI1 and its homologous genes, OsBRL1 and OsBRL3, in rice. Plant Physiol 140:580 Nehrt NL, Clark WT, Radivojac P, Hahn MW (2011) Testing the ortholog conjecture with comparative functional genomic data from mammals. PLoS Comput Biol 7:e1002073 Nei M, Gojobori T (1986) Simple methods for estimating the numbers of synonymous and nonsynonymous nucleotide substitutions. Mol Biol Evol 3:418 Nomura T, Bishop GJ, Kaneta T, Reid JB, Chory J, Yokota T (2003) The LKA gene is a BRASSINOSTEROID INSENSITIVE 1 homolog of pea. Plant J 36:291 Posada D (2008) jModelTest: phylogenetic model averaging. Mol Biol Evol 25:1253 Punta M, Coggill PC, Eberhardt RY, Mistry J, Tate J, Boursnell C, Pang N, Forslund K, Ceric G, Clements J, Heger A, Holm L, Sonnhammer EL, Eddy SR, Bateman A, Finn RD (2012) The PFAM protein families database. Nucleic Acids Res 40:D290 Qiu YL, Li L, Wang B, Chen Z, Knoop V, Groth-Malonek M, Dombrovska O, Lee J, Kent L, Rest J, Estabrook GF, Hendry TA, Taylor DW, Testa CM, Ambros M, Crandall-Stotler B, Duff RJ, Stech M, Frey W, Quandt D, Davis CC (2006) The deepest divergences in land plants inferred from phylogenomic evidence. Proc Natl Acad Sci USA 103:15511 Roy SW, Gilbert W (2006) The evolution of spliceosomal introns: patterns, puzzles and progress. Nat Rev Genet 7:211 Saitou N, Nei M (1987) The neighbor-joining method: a new method for reconstructing phylogenetic trees. Mol Biol Evol 4:406 She J, Han Z, Kim TW, Wang J, Cheng W, Chang J, Shi S, Yang M, Wang ZY, Chai J (2011) Structural insight into brassinosteroid perception by BRI1. Nature 474:472 Shinohara H, Ogawa M, Sakagami Y, Matsubayashi Y (2007) Identification of ligand binding site of phytosulfokine receptor by on-column photoaffinity labeling. J Biol Chem 282:124 Shiu SH, Bleecker AB (2001a) Plant receptor-like kinase gene family: diversity, function, and signaling. Sci STKE 2001:re22 Shiu SH, Bleecker AB (2001b) Receptor-like kinases from Arabidopsis form a monophyletic gene family related to animal receptor kinases. Proc Natl Acad Sci USA 98:10763 Shiu SH, Bleecker AB (2003) Expansion of the receptor-like kinase/ Pelle gene family and receptor-like proteins in Arabidopsis. Plant Physiol 132:530 Shiu SH, Karlowski WM, Pan R, Tzeng YH, Mayer KF, Li WH (2004) Comparative analysis of the receptor-like kinase family in Arabidopsis and rice. Plant Cell 16:1220 Stamatakis A (2006) RAxML-VI-HPC: maximum likelihood-based phylogenetic analyses with thousands of taxa and mixed models. Bioinformatics 22:2688 Stanke M, Diekhans M, Baertsch R, Haussler D (2008) Using native and syntenically mapped cDNA alignments to improve de novo gene finding. Bioinformatics 24:637
J Mol Evol (2014) 78:118–129 Sun Y, Fokar M, Asami T, Yoshida S, Allen RD (2004) Characterization of the brassinosteroid insensitive 1 genes of cotton. Plant Mol Biol 54:221–232 Suyama M, Torrents D, Bork P (2006) PAL2NAL: robust conversion of protein sequence alignments into the corresponding codon alignments. Nucleic Acids Res 34:W609 Symons GM, Davies C, Shavrukov Y, Dry IB, Reid JB, Thomas MR (2006) Grapes on steroids. Brassinosteroids are involved in grape berry ripening. Plant Physiol 140:150 Tamura K, Peterson D, Peterson N, Stecher G, Nei M, Kumar S (2011) MEGA5: molecular evolutionary genetics analysis using maximum likelihood, evolutionary distance, and maximum parsimony methods. Mol Biol Evol 28:2731 Vert G, Nemhauser JL, Geldner N, Hong F, Chory J (2005) Molecular mechanisms of steroid hormone signaling in plants. Annu Rev Cell Dev Biol 21:177 Wang X, Li X, Meisenhelder J, Hunter T, Yoshida S, Asami T, Chory J (2005) Autoregulation and homodimerization are involved in the activation of the plant steroid receptor BRI1. Dev Cell 8:855
129 Wang ZY, Bai MY, Oh E, Zhu JY (2012) Brassinosteroid signaling network and regulation of photomorphogenesis. Annu Rev Genet 46:701 Xu G, Ma H, Nei M, Kong H (2009) Evolution of F-box genes in plants: different modes of sequence divergence and their relationships with functional diversification. Proc Natl Acad Sci USA 106:835 Yamamuro C, Ihara Y, Wu X, Noguchi T, Fujioka S, Takatsuto S, Ashikari M, Kitano H, Matsuoka M (2000) Loss of function of a rice brassinosteroid insensitive1 homolog prevents internode elongation and bending of the lamina joint. Plant Cell 12:1591 Yang Z (2007) PAML 4: phylogenetic analysis by maximum likelihood. Mol Biol Evol 24:1586 Zhao DZ, Wang GF, Speal B, Ma H (2002) The excess microsporocytes1 gene encodes a putative leucine-rich repeat receptor protein kinase that controls somatic and reproductive cell fates in the Arabidopsis anther. Genes Dev 16:2021
123
