Work

Genome-Wide Annotation of Remorins, a Plant-Specific Protein Family: Evolutionary and Functional Perspectives

Year: 2007

Type: article

Abstract: Remorins were discovered in a screen for plasma membrane (PM) proteins differentially phosphorylated in the presence of oligogalacturonides (Farmer et al., 1989). The first remorin was initially designated as pp34, as it corresponded to a phosphorylated protein with a molecular mass of 34 kD in purified PM preparations from potato (Solanum tuberosum; Jacinto et al., 1993). After a section in J.L. Borge's “Book of Imaginary Beings,” in which the remora fish is described to attach itself to larger fish and ships, this protein was renamed remorin (Reymond et al., 1996) due to its hydrophilic profile and its ability to attach to PMs (E. Farmer, personal communication). Several proteomic approaches on PM preparations from tobacco (Nicotiana tabacum) leaves and Arabidopsis (Arabidopsis thaliana) seedlings now suggest localization of remorins in association with PMs (Watson et al., 2003; Marmagne et al., 2004; Mongrand et al., 2004; Sazuka et al., 2004; Nelson et al., 2006; Valot et al., 2006). Interestingly, remorins have also been found in detergent-resistant membrane fractions, called lipid rafts (Mongrand et al., 2004; Shahollari et al., 2004; Bhat and Panstruga, 2005; Laloi et al., 2007; Lefebvre et al., 2007). Lipid rafts are characterized by an enrichment of sterols, sphingolipids, and steryl-glycosides, and are believed to be functional microdomains in PMs (Peskan et al., 2000; Mongrand et al., 2004; Borner et al., 2005; Laloi et al., 2007). Most recently, Lefebvre and coworkers demonstrated the presence of remorins in lipid raft fractions from root PMs of the model legume Medicago truncatula (Lefebvre et al., 2007). While little is known about the function of these microdomains in plants, key signaling processes during HIV-1 infection (Manes et al., 2000; Ono and Freed, 2001), spreading of malarial parasites and the Newcastle disease virus, as well as prion conversion occur at lipid rafts in mammalian systems (Samuel et al., 2001; Nishina et al., 2004; Laliberte et al., 2006). Some remorins share physical properties with viral movement proteins due to their hydrophobic N-terminal region. The C-terminal domain is generally predicted to consist of coiled-coil domains, a feature that appears to be common to all remorin proteins. This region was found to be responsible for the ability of remorins to form oligomeric filamentous structures in vitro, suggesting that these proteins might be structural components of plant cytoskeletons and/or the membrane skeletons (Bariola et al., 2004). In addition, expression of a remorin from tomato (Solanum lycopersicum) was observed in embryonic, apical, and vascular tissue (Bariola et al., 2004). However, astonishingly, there is no published genetic evidence in any plant species that can help us to decipher putative functions of remorins. The few reported attempts to generate plant mutants in which one or more of the ubiquitously expressed remorin genes were knocked out either failed or did not lead to an obvious phenotype (Reymond et al., 1996; Bariola et al., 2004). Searches in databases of fully sequenced plant species revealed the existence of eight, 16, and 19 remorin genes in Populus trichocarpa, Arabidopsis, and rice (Oryza sativa), respectively (Supplemental Table S1). Patterns of differentially regulated remorin genes and proteins have been reported in an increasing number of transcriptomic and proteomic studies. Since many of these are annotated as “unknown function,” they need to be properly classified and given suitable nomenclature. Here, we report on a genome-wide survey of the remorin family throughout fully sequenced plant genomes. We combined a phylogenetic approach with various in silico sequence-scanning tools (alignments, and motif and domain predictions) and analysis of available genetic and genomic data to identify a specific remorin signature and to define several groups within the remorin family. We used this classification to propose a general nomenclature for remorins that should help to differentiate between the different members in future studies. Keyword and BLAST searches among available data allowed us to retrieve approximately 1,000 remorin sequences corresponding to more than 130 remorin proteins from 55 different plant species (Supplemental Table S1). Importantly, no significant hits were identified among sequence data from other kingdoms (fungi, animals, archaeabacteria, and eubacteria). Unexpectedly, no remorin genes were found in algae (neither by BLAST against publicly available data nor in the Ectocarpus siliculosus full-genome sequence; M. Cock, personal communication), while remorins were identified in mosses and ferns. Canonical remorins, such as StREM (Reymond et al., 1996; now StREM1.3), are characterized by the presence of an N-terminal and a C-terminal PFAM domain (Remorin_N: PF03766; Remorin_C: PF03763). Higher sequence variability in N-terminal regions of remorins suggests structural and functional divergence and indicates that a suitable classification and nomenclature are needed. In contrast, the Remorin_C domain, identified by InterProScan using default parameters, appeared to be well conserved among all remorin proteins. It was thus considered as the remorin signature and was used to validate potential family members. The remorin signature. Highly conserved residues within the Remorin_C domain of the 85 remorin proteins that were considered for phylogenetic analysis are shown. Prediction of the position along the heptad repeat of the coiled-coil structure is given in the top panel. pos., Position in Remorin_C domain alignment; ρ, aromatic residues; +, positively charged residues; ϕ, aliphatic residues; % refers to the number of sequences showing the indicated residue at the indicated position, among the list of 85 remorins given in Supplemental Table S1. Some particularly conserved residues can be outlined: position 11 of Remorin_C is occupied by an aromatic residue (mainly Trp); positions 18, 22, and 25 by positively charged residues; and position 30 by an aliphatic residue for all but one sequence (Fig. 1). Position 33 harbors a Trp in all sequences, except in Physcomitrella patens and Ceratopteris richardii, and positions 75 and 80 are Alas in most of the proteins. Because of their very high conservation, these amino acids can be considered as identifying features of the Remorin_C domain. In addition, a 23-amino-acid-long coiled-coil structure between Remorin_C residues 40 to 63 is predicted with a very high probability (Marcoils probability >90%). Prediction of the different positions in heptad repeats (Fig. 1, top diagram) reveals aliphatic residues for sites a and d that form the core of the coiled-coil region. In comparison with coiled-coil domains from other proteins, in remorins these motifs consist of a higher number of charged residues (56% for StREM1.3) in positions b, c, d, and e. This characteristic could be responsible for the high stability of protein interactions mediated by this domain (Burkhard et al., 2001). With the aim of considering only full-length proteins and to avoid a bias of the study toward only highly expressed family members, we focused mainly on genes/proteins from plants with annotated genomes. As recent proteomic studies report the identification of remorins in legumes (Watson et al., 2003; Wienkoop and Saalbach, 2003; Valot et al., 2006; Lefebvre et al., 2007), all 10 remorin sequences of the partially sequenced model legume M. truncatula have been included in the analysis as representative remorins for the Fabaceae family. Considering that plant species that diverged early in evolution are likely to contain ancestral remorin sequences, all available remorins from ferns, mosses, and gymnosperms were also incorporated. This led to a list of 85 remorins (Supplemental Table S1) on which phylogenetic analysis was performed, as described in detail in Supplemental Materials and Methods S1. Phylogenetic and domain organization of the remorin family based on their protein sequences. The right box shows the color code used to indicate evolutionary positions within the plant kingdom (according to the Angiosperm Phylogeny Group, 2003) and domain elements used for classification. In the diagrams domain lengths are proportional to the average protein sequence length (except for groups 5 and 6, for which the representation of the N-terminal region is intercepted by “//,” indicating a variable length of this module within these two groups). Underlined sequences contain Remorin_N domains identified by InterProScan. Bootstrap values (%) are given along the most important branch points. True branch lengths were inferred by maximum-likelihood analysis; the upper-left scale represents a probability of 0.1 amino acid changes per site. The tree was reconstructed using parsimony analysis. For clarity of the figure, a shortened nomenclature was used (e.g. At1.1 for AtRem1.1). The first number in the proposed name indicates the corresponding group (e.g. At1.3 is the third member of group 1). Abbreviations: Ac, Adiantum capillus-veneris; Ak, Amborella trichopoda; Ap, Allium cepa; At, Arabidopsis; Cr, C. richardii; Mc, Mesembryanthemum crystallinum; Mt, M. truncatula; Na, Nuphar advena; Nt, tobacco; Os, rice; Pa, Persea americana; Pd, Pinus taeda; Pi, Pinus pinaster; Pp, P. patens; Ps, Picea sitchensis; Pt, P. trichocarpa; Sl, tomato; St, Solanum tuberosum; Wm, Welwitschia mirabilis; Zm, Zea mays. For the final grouping leading to the proposed nomenclature, results from the phylogenetic analysis were combined with domain features and similarities between different remorin proteins. Due to significant differences, mainly in the N-terminal regions, the remorin family was subdivided into six separate groups. Such subdivision based on domain features enabled us to discriminate between groups 1, 2, and 3 on those parts of the tree where phylogenetic separation was less clear (Fig. 2; Table I Structural characteristics of remorins Structural features were identified for all remorins and were combined with phylogenetic clustering of the different remorin isoforms to propose a group classification. Numbers of exons were obtained from public databases for genes of all sequenced plant species or newly assembled for M. truncatula sequences (T. Ott, unpublished data). Details of MEME motifs can be found in Supplemental Table S2. *, Number of amino acids. **, Average e value for domain recognition: +, >10−10; ++, >10−15; +++, >10−30; ++++, <10−30 (3). ***, Two exceptions with four exons: OsREM1.2 for group 1a and AtREM1.3 for group 1b. NA, Not applicable; Rem_C, Remorin_C domain; Rem_N, Remorin_N domain. Structural characteristics of remorins Structural features were identified for all remorins and were combined with phylogenetic clustering of the different remorin isoforms to propose a group classification. Numbers of exons were obtained from public databases for genes of all sequenced plant species or newly assembled for M. truncatula sequences (T. Ott, unpublished data). Details of MEME motifs can be found in Supplemental Table S2. *, Number of amino acids. **, Average e value for domain recognition: +, >10−10; ++, >10−15; +++, >10−30; ++++, <10−30 (3). ***, Two exceptions with four exons: OsREM1.2 for group 1a and AtREM1.3 for group 1b. NA, Not applicable; Rem_C, Remorin_C domain; Rem_N, Remorin_N domain. This group gathers all remorins that possess both Remorin_N and Remorin_C domains identified by InterProScan. Genomic data support coherence of this group as most genes consist of five exons, except for AtREM1.3 and OsREM1.2, in which the last two exons were fused by deletion of an intron. Generally, group 1 remorins are characterized by a high Pro content in the N-terminal region (only group 4 members have similar numbers of Pros in this region). Group 1 remorins were subdivided into group 1a (8.9% Pros) and group 1b (14.4% Pros) based on their N-terminal amino acid composition (Table I; Supplemental Fig. S2). Group 1a remorins are about 200 amino acids long with a very homologous N-terminal region (about 82 amino acids). Four well-conserved motifs were identified by MEME/MAST (Multiple Em for Motif Elicitation/Motif Alignment and Search Tool; Bailey and Elkan, 1994) in these N-terminal regions (Supplemental Table S2). They bear two ELMs (Eukaryotic Linear motifs; Puntervoll et al., 2003): a casein kinase Ser/Thr phosphorylation site and a motif found in cargo adaptor proteins that bind clathrin heavy chains. These elements, although not functionally characterized in remorins, underline the potential importance of phosphorylation events and membrane association for these proteins. Phylogenetic clustering of this group is remarkable as it assembles branches of ancestral plants: One is comprised of sequences from ferns and gymnosperms, the second consists of sequences from early angiosperms, and two other branches contain exclusively monocot sequences. Only two dicot remorin sequences from Arabidopsis and potato belong to this group. Coexpression analysis of remorins throughout the Arabidopsis transcriptome. Expression data from hybridizations using At22k Affymetrix GeneChips were used to retrieve a list containing 100 genes showing highest Pearson correlation coefficients for different remorin genes at http://www.bar.utoronto.ca/. Graphs represent the normalized frequency of each function in the list of well-correlated genes compared to randomly selected genes among the Arabidopsis genome (Toufighi et al., 2005). Error bars indicate variation over 100 different sets of randomly selected genes. Group 1b is the second group of proteins comprising both Remorin_N and Remorin_C domains, but their N-terminal regions contain twice the number of Pro residues compared to group 1a remorins. Only dicotyledonous plant species fall into this subgroup, suggesting that the Pro enrichment of the Remorin_N domain occurred after divergence of dicots. Group 1b contains the first discovered remorins, StREM1.3 (Reymond et al., 1996) and LeREM1 (Bariola et al., 2004; now named SlREM1.2), to which most of the current literature refers. The described Arabidopsis remorin AtDBP (AtREM1.3) and most remorins currently identified in proteomic screens of membrane preparations also cluster in this group (Watson et al., 2003; Marmagne et al., 2004; Mongrand et al., 2004; Sazuka et al., 2004; Nelson et al., 2006; Valot et al., 2006). Association with membranes thus seems to be a common feature of group 1b remorins. MEME motifs found in the N-terminal part of group 1b remorins are close to those found for group 1a (Supplemental Table S2). In addition to the Eukaryotic Linear Motif identified in group 1a, group 1b well-conserved motifs also notably bear an endoplasmic reticulum retention and retrieving signal. Abundance of group 1b remorins in proteomic studies may be explained by the high expression level of the corresponding genes. Indeed, AtREM1.2 and AtREM1.3 are among the 10% most highly and ubiquitously expressed genes in Arabidopsis (Supplemental Table S3). AtREM1.2 and AtREM1.4 are likely to represent genes derived from a recent duplication event (Supplemental Fig. S1), and their function may therefore be redundant. Expression of group 1b genes in Arabidopsis is markedly increased in mature and/or senescing tissues (mature leaves, roots, mature pollen, flowers) and in vascular tissue (Bariola et al., 2004; AtGenExpress). Transient gene induction was also found following abiotic stress treatment, such as cold, osmotic, or drought (Bray, 2002; Reddy et al., 2002), and by addition of the phytohormone ABA (AtGenExpress; Lin et al., 2003). Another link with plant hormones is also suggested by remorin AtREM1.3 (Alliotte et al., 1989), which was copurified with the cytokinin-induced response regulator AtARR4 (Yamada et al., 1998). Differential expression of group 1b remorins was also observed during plant-microbe interactions—in Arabidopsis upon infection with Pseudomonas syringae (Journot-Catalino et al., 2006) and in Lycopersicon hirsutum after inoculation with Clavibacter michiganensis (Coaker et al., 2004). Whether a direct correlation between and to be Several genes that were found to be with group 1b remorins in Arabidopsis proteins with the and of have functions to response to and signaling (Fig. Group remorins contain the Remorin_C domain (Fig. while their N-terminal domain shows no significant protein to the Remorin_N domain of group 1 remorins and contains Pro the MEME motifs found in their N-terminal regions from those of group group contains only proteins from legumes by sequences from M. and P. group remorins found in Arabidopsis or Interestingly, a group remorin was found during and approaches to be during interactions between legumes and et al., 2002; Wienkoop and Saalbach, 2003; et al., 2004; et al., 2004). this gene was named Ott, unpublished data). of this gene can only be found in other such as and In addition, a gene a group remorin was found to be during the between the model legume and the et al., 2005). However, the of sequences in rice or other plants that are to a suggests that this remorin is not for such Remorins from group 3 only consist of a Remorin_C domain that is more with positively charged residues by aliphatic amino acids. it the residues of the remorin signature. The of an N-terminal domain suggests that group 3 may and high sequence similarities with the C-terminal regions of and AtREM1.2 These genes are believed to from gene (Supplemental Fig. S1). Differential expression of been found abiotic stress while for no expression data have been in the The Remorin_C domain of group 4 remorins is very similar to that of remorins, and this group also contains genes comprising only two Group 4 remorins are characterized by a long and N-terminal region (about amino that not any significant to the group 1 Remorin_N domain. motifs found in group 4 N-terminal regions (Supplemental Table indicate that most of these remorins possess two positions sites from proteins with and and a site predicted to of regions within proteins. similarities suggest that and were derived from each other by a recent duplication event within the Arabidopsis genome (Supplemental Fig. S1). These two genes are expressed in plant and senescing leaves, as well as They are stress showing a after osmotic, and drought and of plants with such as ABA and these genes are upon (Supplemental Table S3). Such are by the expression of group 4 remorins with those of genes with and as well as abiotic or (Fig. Group 5 remorins a more Remorin_C domain, as indicated by values given by protein (Table This group consists of a highly variable N-terminal although all members contain two conserved MEME motifs that bear putative phosphorylation sites (Supplemental Table S2). Due to the of the group 5 remorin from the and the of any significant sequence of their N-terminal regions, it to the of remorins from this group. However, of M. the putative of is one of the remorins expressed in and M. personal communication). The last group contains remorins from to amino acids in The of their N-terminal regions is variable and may the existence of several which are to define based on currently available Interestingly, sequence similarities in the N-terminal regions of members of group were found with a from in and and from P. patens in In addition, four conserved MEME motifs can be in their N-terminal regions, indicating that many remorins of this group contain conserved putative phosphorylation a predicted for a of a predicted responsible for the with an adaptor protein of Arabidopsis remorins and are highest in remarkable in of these genes after and as well as induction after stress was also Supplemental Table S3). Interestingly, study predicted a putative site for the protein a within a group remorin et al., 2005). However, such direct of remorins by and other during plant-microbe interactions to be proteins to be specific to the plant The of remorins in but their presence in mosses, ferns, and higher plants, suggests that the of remorins with the of It is very likely that this of to with particularly in composition is to the of an position and the of vascular in components were found by from of and to vascular For components of such as and components of are in and vascular plants, but are not in acid of revealed that acids were more in ancestral plants than in more recent vascular and that is more in and than in vascular was from all algae and and 2003). Interestingly, the potato StREM1.3 was found to be to bind both and and with or components can be (Reymond et al., It an as to remorins were positively selected with these thus an that allowed of different remorins from the various groups to be but gene expression data suggest that of these proteins might have key functions during to and abiotic and might be in and Transient and induction of gene expression (e.g. and upon such as suggests in early of plant-microbe Whether such can be to and/or functions to be With to the variability of the N-terminal parts among the different groups and the that this region functional of the different functions might be Interestingly, no mutants are available for any remorin gene in Arabidopsis, indicating remorin gene expression is model such as root as described may therefore be to the functions of remorins. The nomenclature suggested should analysis and classification of remorin and proteins during future and lead to a more of this protein family. The following are available in the of this Supplemental S1. positions and duplication events for remorin genes within the Arabidopsis Supplemental S2. of 85 remorin proteins used for phylogenetic analysis. Supplemental Table S1. of remorin sequences from publicly available databases as described in Supplemental Materials and Methods S1. Supplemental Table S2. of characteristic MEME motifs leading to specific remorin Supplemental Table of expression data of remorin genes from Supplemental Materials and Methods S1. of used for this We for and for the BLAST analysis on siliculosus sequences. we to and for of the

Source: PLANT PHYSIOLOGY

DOI: 10.1104/pp.107.108639

Authors: Sylvain Raffaele, Sébastien Mongrand, Pascal Gamas, Andreas Niebel, Thomas Ott