Introduction

The highly conserved eukaryotic Ca2+-binding proteins calreticulin (CRT) and calnexin (CNX) are the central players in the so-called CRT/CNX cycle of glycoprotein folding quality control. CRT possesses a C-terminal domain with a (K/H)DEL endoplasmic reticulum (ER) retrieval signal (Michalak et al. 2009), while CNX is a ER membrane-bound protein (Jin et al. 2009). CRT/CNX cycle is part of the N-glycan-dependent quality control mechanism that takes place in the ER lumen (Hammond et al. 1994). Glycan processing starts with its transfer to Asn residues in nascent proteins within ER. Several subsequent glycosyl hydrolysis exposes the Glc1Man9GlcNAc2 epitope that is then recognized by CRT and CNX that specifically bind monoglucosylated polymannose glycans (Ware et al. 1995; Caramelo and Parodi 2008).

CNX and CRT were first described in plants in 1993 and 1998, respectively (Huang et al. 1993; Crofts and Denecke 1998). Plant’s CRTs were further classified into two groups of homologs, CRT1/2 and CRT3, which were initially thought to be resulting from a gene duplication event occurring before the divergence between monocots and eudicots (Persson et al. 2003). A recent work (Jin et al. 2009) suggested that CRT3 group is present in basal land plants and Arabidopsis CRT3-specific function on the retention of defective brassinosteroid receptor EFR in the ER, which is a specific function of plant’s CRT3 without functional overlapping with CRT1 and 2 (Christensen et al. 2010). EFR accumulation and signaling are impaired in Arabidopsis crt3 mutant, affecting the immune response to the bacterial epitope elf18 (Saijo et al. 2009) suggesting a role for CRT3 in bacterial pathogen-associated molecular pattern (PAMP), while CRT1 and 2 are possibly involved in more general chaperone functions (Li et al. 2009; Christensen et al. 2010). Plant’s CRTs were also implicated in several physiological processes such as virus defense (Chen et al. 2005), ER calcium buffering (Persson et al. 2001; Christensen et al. 2010), plasmodesma cell–cell transport (Baluska et al. 1999; Laporte et al. 2003), and stress response and tolerance (Jia et al. 2008). In the following sections, I will show detailed phylogenetic analyses of CRT and CNX genes in Viridiplantae kingdom using 18 completed genomes and ESTs from diverse lineages such as green algae, basal non-vascular and vascular land plants, gymnosperms and angiosperms.

Methods

Comparative sequence analyses

Predicted proteomes for Arabidopsis thaliana, version 8.0–http://www.arabidopsis.org; Arabidopsis lyrata, version 1.0–http://genomeportal.jgi-psf.org/Araly1/Araly1.home.html; Populus trichocarpa, version 1.1–http://genome.jgi-psf.org/Poptr1_1/Poptr1_1.home.html; Glycine max, version 1.0–http://www.phytozome.net/soybean.php; Ricinus communis, version 0.1–http://castorbean.jcvi.org/downloads.php; Oryza sativa, version 5.0–http://rice.plantbiology.msu.edu; Sorghum bicolor, version 1.4–http://genome.jgi-psf.org/Sorbi1/Sorbi1.home.html; Selaginella moellendorffii, version 1.0–http://genome.jgi-psf.org/Selmo1/Selmo1.home.html; Physcomitrella patens patens, version 1.1–http://genome.jgi-psf.org/Phypa1_1/Phypa1_1.home.html; Volvox carteri, version 1.0–http://genome.jgi-psf.org/Volca1/Volca1.home.html; Chlamydomonas reinhardtii, version 4.0–http://genome.jgi-psf.org/Chlre4/Chlre4.home.html; Ostreococcus lucimarinus, version 2.0–http://genome.jgi-psf.org/Ost9901_3/Ost9901_3.home.html; Ostreococcus tauri, version 2.0–http://genome.jgi-psf.org/Ostta4/Ostta4.home.html; Ostreococcus sp. RCC809, version 2.0–http://genome.jgi-psf.org/OstRCC809_2/OstRCC809_2.home.html; Micromonas pusilla CCMP1545, version 2.0–http://genome.jgi-psf.org/MicpuC2/MicpuC2.home.html; Micromonas sp. RCC299, version 2.0–http://genome.jgi-psf.org/MicpuN2/MicpuN2.home.html; Chlorella vulgaris, version 1.0–http://genome.jgi-psf.org/Chlvu1/Chlvu1.home.html and Chlorella sp. NC64A–http://genome.jgi-psf.org/ChlNC64A_1/ChlNC64A_1.home.html were downloaded and pooled together (Viridiplantae 3.0–530.234 sequences). I performed Hidden Markov Model (HMM) searches using HMMER3 software (http://hmmer.janelia.org/) against Viridiplantae 3.0 in order to identify possible CRT and CNX homologs with an e-value threshold of e−5. I used two different alignments as queries in two independent HMM searches, one of them containing Arabidopsis CRT1, 2 and 3 protein sequences and the other containing Arabidopsis CNX1 and CNX2 protein sequences.

Maize cDNA sequences were obtained from MAGI (http://magi.plantgenomics.iastate.edu/), and ESTs used in this study were downloaded from TIGR Plant Transcript Assemblies (http://plantta.jcvi.org/) and included 202.387 assembled ESTs (unisequences) from Ceratopteris richardii (4.492), Cycas rumphii (4.335), Ginkgo biloba (4.178), Marchantia polymorpha (10.721), Picea abies (5.204), Picea glauca (49.412), Picea sitchensis (25.425), Pinus pinaster (13.067), Pinus taeda (78.873), and Welwitschia mirabilis (6.680). ESTs and cDNAs presenting less than 30% of protein query coverage were discarded.

Phylogenetic analyses

All significantly similar sequences found by HMM were automatically recovered using an in-house algorithm (Del Bem and Vincentz 2010) and manually checked. Sequences were aligned using MAFFT 6.717b (Katoh and Toh 2008) under L–INS–i parameters, and all gaps were removed. Three phylogenetic methods were used to infer the presented trees. Neighbor joining (NJ; Saitou and Nei 1987) using PAM 001 matrix to calculate the genetic distances (Dayhoff et al. 1978) and maximum parsimony (MP; Eck and Dayhoff 1966) under default parameters, both conducted in MEGA 4.0 software (Tamura et al. 2007). Bayesian analyses using the Markov chain Monte Carlo technique were performed using MrBayes3 software (Ronquist and Huelsenbeck 2003) under WAG model of protein substitution (Whelan and Goldman 2001). The Bayesian log-likelihood scores were found to stabilize after 10.000 generations. Therefore, I discarded the initial 10.000 generation trees and sampled one out of every 100 generations from the remaining 0.99 million generations (9,900 trees) to calculate posterior probabilities of each branch. The alignments used contained 329 and 338 amino acids of CNXs and CRTs homologs, respectively. The consensus trees presented in Fig. 1a, b and Supplemental Figs. 1 and 2 were constructed with the three different analyses by consense software from PHYLIP suite (Felsenstein 1989) using the majority-rule consensus tree method. The resulting consensus trees were drawn with MEGA4 tree display tool (Tamura et al. 2007). PoGOs were defined as described in Del Bem and Vincentz (2010).

Fig. 1
figure 1

Phylogenetic trees and evolutionary profile of CNX and CRT genes in green plants. a Phylogenetic tree showing the evolutionary relationship between plants CNXs. Tree topology is a consensus from NJ, MP, and Bayesian analyses. Bootstrap values and posterior probabilities from the original trees higher than 50% are shown (NJ/MP/Bayesian). Triangles represent compacted groups of orthologs that appear in detail in Supplemental Fig. 1. b Phylogenetic tree showing the evolutionary relationship between plants CRTs. Tree topology is a consensus from NJ, MP, and Bayesian analyses. Bootstrap values and posterior probabilities from the original trees higher than 50% are shown (NJ/MP/Bayesian). Triangles represent compacted groups of orthologs that appear in detail in Supplemental Fig. 2. c Evolutionary profile of CNX and CRT genes in green plants. The arrows mark duplication events shared along the descendent lineages, and the squares mark lineage-specific duplication events

Full size image

Results and discussion

CRT and CNX possess two conserved domains: the calcium-binding P-domain ‘InterPro IPR009033’ (http://www.ebi.ac.uk/interpro/IEntry?ac=IPR009033) and the lectin-like N-Domain ‘InterPro IPR018124’ (http://www.ebi.ac.uk/interpro/IEntry?ac=IPR018124), which are indicative of a possible common origin. The fact that CRTs and CNXs genes are present along animals and plants along with my results showing the presence of those genes in genomes of green algae such as Micromonas, Volvox, Chlorella and Ostreococcus (Fig. 1a, b; Supplemental Figs. 1 and 2; Supplemental Table 1) strongly indicate that they originated by an ancestral gene duplication prior to the divergence between Chlorophyta and Embryophyta. This duplication event could even take place in early eukaryotes.

CNX genes in green plants were further classified in a single possible group of orthologs (PoGO) that integrate genes from Chlorophyta algae to angiosperms (Fig. 1a; Supplemental Table 1; Supplemental Fig. 1). This PoGO generally remained as a single-copy gene in a very diverse taxonomic ranking of green plants such as in the Chlorophyta Volvox carteri or even in the monocots sorghum and rice. This observation suggests that single-copy green plant’s CNX genes probably retained the ancestral eukaryotic function that is thought to be related to glycoprotein folding quality control (Schrag et al. 2001). In contrast, soybean and the moss Physcomitrella genomes, probably due to recent large-scale genome duplications, contain four CNX paralogs (Supplemental Table 1). I also analysed Arabidopsis thaliana (eudicot), sorghum (monocot), Physcomitrella patens patens (moss), and Volvox carteri (green algae) CNX genes for shared intron positions within their coding sequences (Supplemental Fig. 3). This analysis helped support the suggested phylogenetic relationship between green plants CNX genes.

Differently from CNX, CRT genes in Viridiplantae kingdom have diversified specifically in land plants by an ancient event of gene duplication in the last common ancestor of Tracheophyta (Fig. 1b; Supplemental Table 1; Supplemental Fig. 2). While chlorophytes’ CRTs formed a single PoGO, land plant’s CRTs were further divided into CRT1/2 and CRT3 PoGOs in agreement with the previous literature (Persson et al. 2003). CRT3 PoGO is embryophyte-specific, which means that genes from this group emerged as a single gene in early land plants’ genomes evolving directly from chlorophyte’s ancestral single-copy CRT gene (PoGO Proto-CRT in Fig. 1b). I interpreted CRT1/2 PoGO as been derived from a CRT3 PoGO gene by an ancestral duplication taking place in the last common ancestor of tracheophytes, as evidenced by the presence of a CRT gene from Selaginella moellendorffii in both PoGOs and the absence of Marchantiophyta and Bryophyta genes in CRT1/2 PoGO. The proposed phylogenetic classification is further supported by shared intron positions analysis (Supplemental Fig. 4). However, I cannot fully discard an alternative scenario where CRT1/2 genes were specifically lost in non-vascular land plants.

Genes belonging to CRT3 PoGO appear to have evolved specialized functions in land plants when compared to other eukaryotic CRTs, as illustrated by the specific function in retention of defective EFR brassinosteroid receptor in Arabidopsis thaliana (Jin et al. 2009; Saijo et al. 2009). In fact, the three Arabidopsis CRT genes were shown to be up-regulated by short-term treatments with exogenous glucose and sucrose along with a sugarcane gene belonging to CRT1/2 group (Papini-Terzi et al. 2009). This finding supports the notion that plant’s CRTs expression response to sugars is conserved at least in angiosperms. Grasses also appear to have duplicated CRT1/2 and CRT3 angiosperms’ ancestral genes earlier in their evolution (Fig. 1c; Supplemental Table 1).

My analysis suggests that CRT1/2 founder gene in Spermatophyta has undergone at least three independent duplication events: one occurring in Pinophyta (as seen for Pinus taeda and Picea sitchensis), other in grasses (rice and sorghum), and finally in Brassicaceae Arabidopsis thaliana and A. lyrata (Supplemental Fig. 2; Supplemental Table 1; Fig. 1c). The expression of CRT1 and CRT3 from Arabidopsis thaliana into calreticulin-deficient (crt − / −) mouse embryonic fibroblasts restored the wild-type phenotype of ER Ca2+-holding potential and putative chaperone capabilities, suggesting a strong conserved biochemical function of plants and mammals CRTs (Christensen et al. 2008, 2010.

In conclusion, I presented a catalog of CRT and CNX genes from 18 completed genomes and ESTs from several green plants species (Supplemental Table 1), as well the evolutionary profile of CRT and CNX genes in the Viridiplantae kingdom (Fig. 1).