The Wuschel related homeobox (WOX) family proteins are key regulators implicated in the determination of cell fate in plants by preventing cell differentiation. A recent WOX phylogeny, based on WOX homeodomains, showed that all of the Physcomitrella patens and Selaginella moellendorffii WOX proteins clustered into a single orthologous group. We hypothesized that members of this group might preferentially share a significant part of their function in phylogenetically distant organisms. Hence, we first validated the limits of the WOX13 orthologous group (WOX13 OG) using the occurrence of other clade specific signatures and conserved intron insertion sites. Secondly, a functional analysis using expression data and mutants was undertaken.
The WOX13 OG contained the most conserved plant WOX proteins including the only WOX detected in the highly proliferating basal unicellular and photosynthetic organism Ostreococcus tauri. A large expansion of the WOX family was observed after the separation of mosses from other land plants and before monocots and dicots have arisen. In Arabidopsis thaliana, AtWOX13 was dynamically expressed during primary and lateral root initiation and development, in gynoecium and during embryo development. AtWOX13 appeared to affect the floral transition. An intriguing clade, represented by the functional AtWOX14 gene inside the WOX13 OG, was only found in the Brassicaceae. Compared to AtWOX13, the gene expression profile of AtWOX14 was restricted to the early stages of lateral root formation and specific to developing anthers. A mutational insertion upstream of the AtWOX14 homeodomain sequence led to abnormal root development, a delay in the floral transition and premature anther differentiation.
Our data provide evidence in favor of the WOX13 OG as the clade containing the most conserved WOX genes and established a functional link to organ initiation and development in Arabidopsis, most likely by preventing premature differentiation. The future use of Ostreococcus tauri and Physcomitrella patens as biological models should allow us to obtain a better insight into the functional importance of WOX13 OG genes.
Homeodomain (HD) containing transcription factors are key regulators implicated in the determination of cell fate and cell differentiation in both plants and animals. In Angiosperms, a gene called WUSCHEL (WUS) has been isolated from many different species. WUS was the first identified member of the
The specific expression of the WOX genes in different plant organs and cell types [2-7] suggested an important role for them during organogenesis. The WUS gene is expressed in a restricted region of the meristem called the organizing centre located below the stem cells of the shoot apical meristem (SAM) and functions non-cell autonomously to control the stem cell fate . Interestingly Gallois et al  showed recently that ectopic expression of WUS in the root establishes shoot stem cells and leaf development. Other members of the WOX family might also prevent premature cell differentiation in developing organs or tissues. Recently, WOX5 was shown to be involved in stem cell maintenance signaling in the root . Wu et al  have shown that WOX9 (Stimpy/STIP) prevents premature cell differentiation during organ growth, in addition to contributing positively to WUS expression. This work also revealed that stip mutants can be rescued by sucrose, a modulator of cell proliferation trough cyclin D induction [11,12]. Two other WOX genes, named Narrow Sheath 1 and 2, that are maize orthologues of Arabidopsis thaliana AtWOX6 were shown to be involved in the recruitment of the lateral founder cells within the SAM prior to the primordia development . Last, Haecker et al  analyzed the expression dynamics of the WOX genes during A. thaliana embryo development and showed that members of the WOX family mark cell fate decisions during embryonic patterning. Taken together, these results suggest that not only WUS but several other WOX genes play a role in regulating cell division and preventing cell differentiation.
Characterized genomes of Angiosperms contain more than 10 different WOXs and phylogenetic studies suggest that the WOX family existed in the last common ancestor of monocots and core eudicots [14,15]. A recent WOX phylogeny based on WOX HDs from different Angiosperms, the moss Physcomitrella patens and the lycophyte Selaginella moellendorffii showed that all of the P. patens and eight of the S. moellendorffii WOX genes clustered in the same clade . Indeed, the phylogenetic analyses of a transcription factor family like WOX suffer from some limitations due to the relatively short length of the HD, which is the only conserved sequence between the WOX protein family members, and the relatively few land plant species with a completely characterized proteome. As a consequence, the patterns of phylogenic trees of large families of plant transcription factors are often weakly supported by statistics and may be affected by long branch attraction . Nevertheless, clade specific gene features and protein domains were observed in transcription factors [18,19]. Thus, in this study, we used clade specific signatures beyond the HD and clade specific conservation of intron insertion sites, to provide additional data in support of the orthology links predicted by the WOX phylogeny. In order to carry out the latter analyses, complete genome sequences with accurate structural annotations are essential. Our results strongly support the existence of a WOX13 orthologous group (WOX13 OG) that, contrary to the two other WOX OGs, contains genes from many different members of the plant kingdom, including basal species, and a branch apparently specific to Brassicaceae. To obtain information on the function of the genes belonging to the WOX13 clade, the cell specificity of At WOX13 and At WOX14 expression was established experimentally and phenotypes resulting from mutations within these two genes were characterized in the plant model A. thaliana.
A WOX distance tree from model genomes
We carefully screened several completely sequenced eukaryote genomes for WOX genes, i.e., two Angiosperms, A. thaliana and O. sativa, three algae, O. tauri, O. lucimarinus and Chlamydomonas reinhardtii and two yeasts, Saccharomyces cerevisiae and Schizosaccharomyces pombe. No WOX genes were identified outside of the plant kingdom and Ostreococcus revealed only one WOX gene per genome. Moreover, using the HMMsearch tool , we found only one HD gene in C. reinhardtii, 2 in S. pombe, 4 in O. tauri and 8 in S. cerevisiae compared to the 92 HD genes in A. thaliana.
Distance and phylogenic trees were computed from the HD sequences of the WOX proteins from the five plant model genomes, O. tauri, P. patens, S. moellendorffii, A. thaliana and O. sativa (see method section). The clustering of the WOX proteins in 3 different OGs were consistent using either the Neighbor-Joining (Figure 1) or the Maximum likelihood (see additional file 1: Figure 1) methods and different outgroups. The topology of the trees clearly showed three different OGs namely the WOX1, WOX8 and WOX13 MOGs (for Model plant OG). Each OG contained rice and Arabidopsis proteins, but the single WOX protein of O. tauri, the three P. patens and the three S. Moellendorffii WOX proteins were all found in the same group, the WOX13 MOG. The WOX1 MOG contained AtWOX1-7 and AtWUS, the WOX8 OG associated AtWOX8, 9, 11 and 12 and the WOX13 OG included AtWOX10, 13 and 14. Thus, the previous observation of the presence of the P. patens proteins only in the WOX13 OG group was extended to O. tauri and S. moellendorffii. However, the introduction of OtWOX and the SmWOXs did not change the main branching of the trees nor significantly modified the bootstrap values at the internal node of the tree. The 14 available WOX sequences from different Gymnosperms were distributed between the three OGs (data not shown). Furthermore, the trees we obtained were not consistent with the phylogenetic tree of plant species as it is now known . Thus the WOX trees we produced suggest probably several rounds of gene deletions in O. tauri and P. patens in order to explain the branching of the WOX13 OG. The finding that all the WOX sequences of O. tauri and P. patens, as well as those of S. moellendorffii were clustered in the WOX13 OG suggested to us that AtWOX10, 13, 14 might represent the most interesting genes to study the functional diversification of the WOX genes although they are the WOX genes for which we have the least functional information (reviewed in ). This prompted us to look at the specificity of the WOX13 OG with regards to gene structure and conserved amino acid signatures other than the HD. Conservation in gene structure is an indication of a common ancestor while conserved amino acid signatures are linked to functions such as protein-protein interactions or regulatory sites.
Additional file 1. Phylogenetic trees. Figure 1 : Maximum-likelihood tree of the WOX proteins from the 5 completely sequenced genomes of the green lineage: Arabidopsis thaliana (AT prefix), Oryza sativa (OS prefix), Physcomitrella patens (Ppa prefix), Ostreococcus tauri (OT prefix) and the three full-length proteins of Selaginella moellendorffii (Sm prefix). The significance of each node was tested using 1000 bootstrap replicates. Only bootstrap values above 50% are shown. The scale, at the right of the tree, indicates 0.2 substitution per site. Figure 2 : Parsimony tree of the WOX proteins from the 5 completely sequenced genomes of the green lineage: Arabidopsis thaliana (AT prefix), Oryza sativa (OS prefix), Physcomitrella patens (Ppa prefix), Ostreococcus tauri (OT prefix) and the three full-length proteins of Selaginella moellendorffii (Sm prefix). The significance of each node was tested using 1000 bootstrap replicates. Only bootstrap values above 50% are shown. Figure 3 : Maximum-likelihood tree of the WOX proteins from the WOX13 Orthology Group. The significance of each node was tested using 1000 bootstrap replicates. Only bootstrap values above 50% are shown. The scale, at the right of the tree, indicates 0.1 substitution per site. Species names, in alphabetic order, are: Afo: Aquilegia formosa; AT: Arabidopsis thaliana; Bna: Brassica napus; Bra: Brassica rapa; Ccl: Citrus clementina; Csi: Citrus sinensis; Cte: Citrus temple; Ees: Euphorbia esula; Ghi: Gossypium hirsutum; Gma: Glycine max; Gra: Gossypium raimondii; Han: Helianthus annuus; Hvu: Hordeum vulgare; Les: Lycopersicon esculentum; Lja: Lotus japonicus; Lsa: Lactuca sativa; Lvi: Lactuca virosa; Mdo: Malus domestica; Mtr: Medicago truncatula; OS: Oryza sativa; OT: Ostreococcus tauri; Ppa: Physcomitrella patens; Psi: Picea sitchensis; Pta: Pinus taeda; Ptr: Populus tremula × Populus tremuloides; Pvu: Phaseolus vulgaris; Stu: Solanum tuberosum; Vvi: Vitis vinifera; Zma: Zea mays. Figure 4 : Parsimony tree of the WOX proteins from the WOX13 Orthology Group. The significance of each node was tested using 1000 bootstrap replicates. Only bootstrap values above 50% are shown. Species names, in alphabetic order, are: Afo: Aquilegia formosa; AT: Arabidopsis thaliana; Bna: Brassica napus; Bra: Brassica rapa; Ccl: Citrus clementina; Csi: Citrus sinensis; Cte: Citrus temple; Ees: Euphorbia esula; Ghi: Gossypium hirsutum; Gma: Glycine max; Gra: Gossypium raimondii; Han: Helianthus annuus; Hvu: Hordeum vulgare; Les: Lycopersicon esculentum; Lja: Lotus japonicus; Lsa: Lactuca sativa; Lvi: Lactuca virosa; Mdo: Malus domestica; Mtr: Medicago truncatula; OS: Oryza sativa; OT: Ostreococcus tauri; Ppa: Physcomitrella patens; Psi: Picea sitchensis; Pta: Pinus taeda; Ptr: Populus tremula × Populus tremuloides; Pvu: Phaseolus vulgaris; Stu: Solanum tuberosum; Vvi: Vitis vinifera; Zma: Zea mays.
Format: PDF Size: 614KB Download file
This file can be viewed with: Adobe Acrobat Reader
Figure 1. WOX genes from model plant genomes. The neighbor-joining tree of the WOX proteins from the 5 completely sequenced genomes of the green lineage: Arabidopsis thaliana (AT prefix), Oryza sativa (OS prefix), Physcomitrella patens (Ppa prefix), Ostreococcus tauri (OT prefix) and Selaginella moellendorffii (Sm prefix). The tree topology was inferred from Jones-Taylor-Thornton matrix distances. The significance of each node was tested using 1000 bootstrap replicates. Only bootstrap values above 50% are shown. The scale, at the top left of the tree, indicates 0.1 substitutions per site. Different structural features are represented by different signs and colors as described at the bottom of the figure.
WOX gene structures
The intron-exon organization of WOX genes from the model genomes showed different gene structures (see additional file 2: Figure 1). Alignments supporting the conservation of intron insertion sites are given in additional file 2: Figure 2. The organization of the gene structures along the phylogeny tree did not reveal any contradiction between gene structures and tree pattern. Intron site conservations were exclusively observed within each of the three OGs and never between them. In this context, it was interesting to observe that the insertion site of the first intron in AtWOX13 is conserved in almost all the genes (from moss to Angiosperms) belonging to the WOX13 OG. The only exception was PpaWOX03, a possible pseudogene (see below). This observation cannot be extended to O. tauri as its genes do not have introns. S. moellendorffii genes have introns and the three full-length SmWOXs have one conserved intron insertion site downstream from the WOX13 motif. The latter intron of SmWOXs is in a similar position (relative to the coding region) to the first highly conserved intron of WOX13 OG genes in moss and Angiosperms. Nevertheless, there is no trace of amino acid sequence conservation in the corresponding protein region between S. moellendorffii and the other genes in the WOX13 OG (see additional file 3: Table 3). When taken together, our analyses of intron insertion sites support the boundaries of the WOX13 OG as defined by phylogeny. In addition to the highly conserved intron upstream of the HD coding sequence found in the WOX13 OG, AtWOX10 and AtWOX14 share another conserved intron insertion site downstream of the HD while OS01G60270 and AtWOX13 share a different intron insertion site in the 3' region of the gene. The tree pattern implies first a loss of the orthologues of AtWOX10 and 14 in O. sativa and in P. patens and secondly that AtWOX13 and OS01G60270 are orthologous. Interestingly, the two latter genes share a specific intron insertion site.
Additional file 2. WOX gene structures and protein motifs. Figure 1 : The structure of WOX genes from model genomes. The gene structures from 4 model genomes (Arabidopsis thaliana, AT; Oryza sativa, OS; Physcomitrella patens, Ppa; Ostreococcus tauri, OT) and three Selaginella moellendorffii (Sm) genes are displayed following the order of the Neighbour-joining tree in Figure 1 and aligned on their translation start. Rectangles stand for exons while lines denote introns. Blue is for exonic regions in UTRs and light grey is for translated exons. Five different gene structures are given for PpaWOX02 due to the 5' UTR variants supported by transcripts. Figure 2 : WOX proteins: Sequence alignments around the virtual intron insertion sites. For each WOX OG, 15 amino acid sequences both upstream and downstream introns are given. The intron insertion sites are indicated by stars. Numbers on the left of sequences give the intron length. Digits between stars stand for the phase of the intron insertion site. The intron phase is defined by the position of its insertion into a codon. An intron can be located between two codons (phase 0) or within a codon, lying either after the first or after the second base pair (phase 1 and phase 2 respectively). Boxes indicate conserved introns with the same color code than in Figure 1. Figure 3 : Sequence logos of protein motifs in the WOX family. Each logo http://weblogo.berkeley.edu webcite consists on stack of letters, one stack for each position in the sequence. The overall height of the stack indicates the sequence conservation at that position, while the letter's height within the stack indicates the relative information content of each amino acid at that position. The amino acid color code is dark blue for large polar, light blue for basic, black for small polar, green for nonpolar, orange for cysteine and purple for ambivalent amino acids. Figure 4 : WOX motifs: alignments for motifs without Logo. Motifs defined by less than 6 sequences cannot be shown as logo since the sequence information at each position is too weak. Thus multiple alignments performed with Clustalw with a majority consensus are shown here for theses motifs. Amino acid color code is the same than in Figure 2.
Format: PDF Size: 173KB Download file
This file can be viewed with: Adobe Acrobat Reader
Additional file 3. WOX sequences and HMM motifs. Table 1: Gene and protein sequences of the WOX family in the model genomes A. thaliana, O. sativa, P. patens and O. tauri. Table 2: Specificity and sensitivity of HMM for WOX motifs. Scores can give indications to distinguish potential family -or subfamily-members from false positive. A clear drop in the score could be detected in most cases, indicating that sequence below this threshold did not fulfill the family model as well as above . HMMER bit scores were obtained for each motif by querying the HMM against three protein sequence sets: WOX13 OG (33 sequences), model genome WOX (32 sequences) and GenBank NR (Release 157). The score drops above 50 and indicating the motifs exhibiting the highest specificities with a given database are in bold. a: no hit was found in the set. b: no specific hit was found in the GenBank NR database. Note that Physcomitrella sequences from genome model WOX set are not present in GenBank NR. c: best hits are not belonging to the WOX family and no score drop was found. Table 3: Vitis vinifera WOX protein prediction with HMM WOX motifs. HMM search scores above a drop give indication to assign novel WOX proteins to the different WOX OGs. AM(...) IDs are contigs from Pinot noir V. vinifera while contig(...) IDs are from PN40024 V. vinifera. Dash indicates that no score was found above the drop. a: no hit was found with this HMM profile. (*): 17.4 score is due to an undefined C or T nucleotide at one position leading to a stop in the CDS. After fixing the nucleotide to C, the score is 27.7.
Format: PDF Size: 87KB Download file
This file can be viewed with: Adobe Acrobat Reader
Specific motifs in the WOX proteins
A search for specific motifs other than the HD used to build the distance and phylogeny trees led us to identify HMMs that were specific to each clade and located either within the N or the C terminal regions of the different proteins. These HMMs are illustrated by sequence logos (see additional file 2: Figure 3) and by alignments (see additional file 2: Figure 4). Only proteins of the WOX1 clade have a small motif of 10 amino acids (i.e., the TLxLFP motif) in the C-terminal region, but at relatively variable distances from the HD. TLxLFP has previously been described as a characteristic motif of WUS proteins [1,23], but we show here that it is common to the entire WOX1 clade, i.e., it is observed not only in WUS, but also in the AtWOX1-7 proteins. However, the WUS proteins of A. thaliana (AT2G17950-AtWUS) and O. sativa (OS04G56780) share specific features not observed in any of the other proteins of the WOX1 MOG. Indeed, they display 1) the LELxL WUS motif  known as EAR-like or ERF-like motifs that are located in the outmost C-terminal region, similar to a motif found in potent transcriptional repressors in plants i.e., Superman  and Aux/IAA ; 2) the same intron exon structure in the coding regions of each gene; and 3) an additional amino acid at a conserved position in the homeodomain.
Nearly all proteins of the WOX8 MOG contain two contiguous motifs of 30 amino acids situated near to their C-terminus (i.e., VFIN-WOX8 and LQxG-WOX8 MOG motifs). The latter motifs are surrounded by two conserved introns. Remarkably, these introns are also found in OS05G48990 in which the VFIN-WOX8 MOG motif is replaced by a non-conserved sequence of a similar length.
All the proteins of the WOX13 MOG, except for PpaWOX03, share a 39 amino acid motif that is found upstream from the HD and named the WOX13 MOG motif. This observation provides further support for the validity of the WOX13 OG as defined by phylogeny and to the hypothesis of a specific function of the genes clustered within. For all of these reasons, it was decided to investigate further the WOX13 clade.
WOX13 OG distance tree
Thirty three full length WOX proteins were used to reconstruct the distance tree of the WOX13 clade (Figure 2), using OtWOX as the outgroup. The neighbor-joining tree showed two main branches. The first branch contained all of the Gymnosperm and the moss WOXs. The second branch contained AtWOX13 together with many Angiosperm sequences. This branch was further separated into two branches, one contained AtWOX13 and many proteins from different Angiosperms while the second contained AtWOX10, AtWOX14 and a Brassica rapa sequence.
The preferential presence of amino acid motifs fully supported the separation of the different gene groups within the trees. Obviously, all proteins of the WOX13 clade displayed the WOX HD and the WOX13 motif. Moreover, we also identified specific motifs for the AtWOX10, 14 and for the AtWOX13 branches (see additional file 2: Figure 3). All Angiosperm proteins located inside the WOX13 OG shared two motifs, the YxDpl-WOX13 motif located between the WOX13 MOG motif and the HD, and the ESExE-WOX13 motif just downstream from the HD. The monocots of the AtWOX13 branch also share a specific motif called the monocot-WOX13 motif that was detected also in two other partial proteins of two monocotyledon species.
Proteins of the AtWOX10, 14 branch shared two specific motifs, the YFdPM-WOX10 and QAdDaAVTT-WOX10 motifs. They were located to the same region of the protein as the YxDpl-WOX13 and ESExE-WOX13 motifs, respectively. Hence, the similarity between the YFdPM-WOX10 OG and YxDpl-WOX13 motifs suggested a common origin. Partial protein sequences from Brassica oleracea and Raphanus raphanistrum subsp. raphanistrum (protein sequence derived from the two ESTs, gb|EX762111| and gb|EX766460|) confirmed the specificity of the motifs to Brassicaceae. Thus, the AtWOX14 containing branch appears to be Brassicaceae specific since: 1) all the sequences containing the two AtWOX10, 14 motifs are from Brassicaceae and 2) no species other than those belonging to the Brassicaceae displayed any of these two motifs.
The Gymnosperm and moss branch contained proteins that shared the ESExE-WOX13 OG motif with the Angiosperm WOX13 proteins, located at the same position relative to the HD. They also shared two other motifs specific to this branch: 1) the LxxGQ-gymnosperm/moss-WOX OG motif that was at the same position as the YxDpl-WOX13 OG or YFdPM-WOX10 OG motifs and 2) a supplementary motif in the C-terminal region of the proteins.
Figure 2. The WOX13 Orthology Group. The different elements of the figure are described from the left to the right. 1) The neighbor-joining tree of the WOX proteins from the WOX13 Orthology Group (see material and methods). The tree topology, the significance of the nodes, and the bootstrap values displayed are the same as in Figure 1. The scale at the top left corner indicates 0.05 substitutions per site. 2) The motif composition of each protein is denoted by colored boxes on a bar representing the length of the protein and aligned on the WOX homeodomain (displayed in orange). Species names, in alphabetic order, are: Afo: Aquilegia formosa; AT: Arabidopsis thaliana; Bna: Brassica napus; Bra: Brassica rapa; Ccl: Citrus clementina; Csi: Citrus sinensis; Cte: Citrus temple; Ees: Euphorbia esula; Ghi: Gossypium hirsutum; Gma: Glycine max; Gra: Gossypium raimondii; Han: Helianthus annuus; Hvu: Hordeum vulgare; Les: Lycopersicon esculentum; Lja: Lotus japonicus; Lsa: Lactuca sativa; Lvi: Lactuca virosa; Mdo: Malus domestica; Mtr: Medicago truncatula; OS: Oryza sativa; OT: Ostreococcus tauri; Ppa: Physcomitrella patens; Psi: Picea sitchensis; Pta: Pinus taeda; Ptr: Populus tremula × Populus tremuloides; Pvu: Phaseolus vulgaris; Stu: Solanum tuberosum; Vvi: Vitis vinifera; Zma: Zea mays. The color code of the different motifs is described at the bottom of the figure.
Functional WOX13 OG genes in the three model plants
We investigated the expression profiles of the WOX13 OG genes within 3 plant model species namely O. tauri, P. patens and A. thaliana because of their position within the green lineage. O. tauri is a highly proliferating unicellular alga that has no known differentiated form. In addition to the WOX gene, we also identified a single copy of a KNOX-like gene (Ot04g04460) within the O. tauri genome. In Angiosperms, members of the class-1 KNOX subfamily act to prevent cell differentiation within the SAM . Interestingly, the OtKNOX protein showed the characteristics of both the class-1 and class-2 KNOX conserved domains, suggesting that O. tauri had diverged from a common ancestor before the diversification of the KNOX gene into 2 classes.
In order to monitor the expression of WOX and KNOX genes during the progression of the cell cycle, we partially light-synchronized a cell suspension of O. tauri as described before in . RT-PCR and quantitative PCR analyses of Cyclin B and Histone 4 genes showed a peak of expression at 14 h, confirming the overlapping of the G2/M and S subpopulations (Figure 3A and see additional file 4: Figure 1). The OtRH4 (Ot07g03220), and OtRH21 (Ot10g01950) genes, respectively the orthologues of the Arabidopsis DEAD-box RNA-Helicase 4 and 21  were used as controls of constitutively expressed genes. They did not show any significant variation, and they were not cell cycle regulated (Figure 3A). More importantly, the WOX and the KNOX genes were found to be constitutively expressed and not cell cycle regulated (Figure 3A), with a higher expression level for WOX compared to KNOX (see additional file 4: Figure 1).
Figure 3. WOX13 OG gene expression in O. tauri, P. patens and A. thaliana. (A) RT-PCR analysis of WOX gene expression in Ostreococcus tauri. 12 h light/dark synchronized or continuous light (cont) cell cultures of 10-day-old were used. Gene name (left side), number of PCR cycles done (right side), time of sampling (top side), light (blank bar) and dark (black bar) periods. The genes OtRH4 and OtRH21 were used as constitutive expression controls. RT were done using oligodT primers except in (*) where random primers were used (B) RT-PCR analysis of WOX gene expression during Physcomitrella patens development: (1) 4-day-old, (2) 10-day-old, (3) 14-day-old, (4) genomic DNA control. (C) RT-PCR analysis of the expression of WOX13 OG genes carried out on 7-day old seedlings of different Arabidopsis thaliana lines: (1) Wild type No ecotype, (2) the pst13645 AtWOX14 mutant line, (3) Wild type WS ecotype, (4) the 193G10 AtWOX13 mutant line, (5) Genomic DNA control.
Additional file 4. Expression analyses. Figure 1 : Real-time PCR analysis of Ostreoccocus tauri gene expression in a 12-hour light/12-hour dark synchronized cell culture. Copy number of WOX, KNOX, Histone4 and CyclinB RNA was quantified using total RNA sampling from the starting of the light period (0 h) or from continuous light grown cell culture. Figure 2 : Alignments of predicted sequences of AtWOX13 (A) and AtWOX14 (B) chimeric proteins with the full length ones. The homeodomain are indicated in bold and the amino acid changes at the end of the chimeric proteins indicated in lower case. Figure 3 : AtWOX13 expression is correlated with the induction of cell proliferation. Data mining analysis of the expression profiles of a set of genes in protoplast culture of Arabidopsis thaliana. (A) AtWOX13 (-◆-), ATHB-7 (■), PCNA1 (▲); (B) Expression profile of a cluster of cell cycle genes including AT4G37490 (CYC1); AT1G44110 (CYCA1;1 (■)); AT5G11300 (CYC3B); AT1G47210 (CYCA3;2); AT3G05330 (CDK); AT2G26760 (CYCB1;4); AT4G35620 (CYCB2;2); AT2G26430 (RCY1); AT2G31400 (ATPDNA Binding). Table 1: Primer sequences designed for expression analysis, genotyping (g primers) and cDNA cloning (2 last row) using primer3 software.
Format: PDF Size: 147KB Download file
This file can be viewed with: Adobe Acrobat Reader
In P. patens, the initial gametophytic phase consists of an array of filamentous tubes that are formed after cell initials are produced from the main filament. Further development proceeds by the formation of buds which form the initial meristem from which originates the leafy adult gametophyte . These processes rely on a tight control of cellular differentiation throughout the gametophyte development. An analysis of WOX expression in P. patens was carried out on 4-, 10- and 14-day-old plants corresponding respectively to the protonema, the bud and the gametophore stages under our growth conditions (data not shown). RT-PCR analyses showed that PpaWOX01 and PpaWOX02 were expressed throughout gametophyte development whereas PpaWOX03 transcripts were not detected (Figure 3B), thus supporting the hypothesis that PpaWOX3 could be a pseudogene.
An exhaustive data mining analysis of A. thaliana expression databases [30-33] indicated that, both AtWOX13 and AtWOX14 transcripts are present at different developmental stages i. e., embryo development, plantlet stage, root formation, bolting stage, in flowers and in response to stress (see discussion). However, no AtWOX10 transcripts were present in any of the databases analyzed. In addition, RT-PCR analyses of seedlings, even with a high PCR cycle number, did not allow us to detect the AtWOX10 mRNA (Figure 3C). These results suggested that the recently duplicated AtWOX10 gene, found only in the A. thaliana genome, might be a pseudogene. Based on these observations, this gene was not further characterized. We conclude that in the three plant model genomes studied at least one copy of the WOX13OG genes is expressed during development.
AtWOX13 shows a dynamic expression profile during A. thaliana development
Compared to other WOX proteins, the function of the AtWOX13 and 14 proteins have received little attention to date (see review ). In order to carry out a more detailed analysis of the WOX13 OG genes in A. thaliana, we cloned the gene coding for the GUS reporter under the control of the AtWOX13 and AtWOX14 upstream regulatory sequences. Based on our data mining analyses (see above), we first analyzed the GUS expression during root development. AtWOX13 promoter activity was detected during lateral root formation (Figure 4A, C, F and 4I). To validate the specificity of these data, in situ hybridization experiments were performed using respectively AtWOX13 (Figure 4B, D, G and 4K), GUS (Figure 4J) and CUC2 as probes (Figure 4E and 4H). The signal obtained using the AtWOX13 and GUS probes matched perfectly with the above GUS staining, whereas the CUC2 probe, as expected, did not give any specific signal. Hence, the GUS staining obtained reflects stable tissue specific mRNA expression and revealed that the AtWOX13 promoter sequence, used to drive the reporter gene GUS expression was sufficient to mimic the tissue specific localization of the endogenous AtWOX13 mRNA.
Figure 4. AtWOX13 gene expression during plant development. (A to L) Expression profile during the initiation and development of lateral roots obtained with 10-day-old AtWOX13::GUS seedlings. Independent lines were tested. (A, C, F and I) GUS staining. (B, D, and K) In situ hybridization with an AtWOX13 antisense probe. (J) In situ hybridization with a GUS antisense probe used as a positive control. (E, H and L) In situ hybridization with a CUC2 antisense probe used as a negative control. (M to S) Expression profile at different plant developmental stages with independent AtWOX13::GUS lines. (M) 5-day-old seedling with staining in the leaf, the hypocotyl, the root vasculature, the pericycle and the primary root tip. (N) Magnification of a 5-day-old seedling primary root tip showing high staining in the stem cell of the proximal meristem. (O) Expression within the shoot apex of 10-day-old seedlings. Note the staining in the leaf primordia. (P to S) Expression within the flower at different developmental stages where staining is seen in the stigmata and the vasculature (P) in the gynoecium (Q) at the floral stage 12–13 where the ovules (Q) are stained. (R) Young embryo, (S) Mature silique. Scale bars = 100 μm except for A, B, R and S (= 50 μm) and M (= 500 μm).
AtWOX13 expression was detected at the early stage of the lateral root (LR) formation of 10-days-old seedlings and it persisted during LR emergence (Figure 4A–D). The GUS staining was progressively restrained to the distal meristem (Figure 4F, and 4G) and it strongly accumulated in both the columella and the adjacent lateral root cap cells at a later developmental stage (Figure 4I–K). In mature LR or primary roots of 10-day-old seedlings no staining was detected within the root tip.
The AtWOX13 expression pattern was also investigated using the AtWOX13::GUS lines, during plant development. In 5-day-old seedlings, the promoter activity was strong in both cotyledon and root vasculatures, the pericycle and in the stem cells of the primary root meristem (Figure 4M and 4N). No GUS staining was detected in the division zone of the root. In the vegetative apex of 10 day-old seedlings, a strong staining was observed in the leaf primordia (Figure 4O). In the flower, the AtWOX13 promoter was active in the vasculature, the stigma (Figure 4P) and in the gynoecium at the flower stage 13/14 (Figure 4Q, and 4R). At this developmental stage, when fertilization was ongoing and pollen grains were germinating at the tip of the gynoecium, ovules stained blue (Figure 4Q). Later on, the young embryos remained stained (Figure 4R), but no staining was observed in mature siliques indicating that the AtWOX13 promoter was no longer active (Figure 4S), as observed before in all the other mature organs. During germination, soon after seed imbibition, GUS staining was also seen within the vasculature and the root tip, as observed in young seedlings (data not shown). Hence, AtWOX13 expression was restricted to differentiating organs and therefore probably regulated by developmental signals.
AtWOX14 promoter specificity differs from that of AtWOX13
Compared to the AtWOX13 expression patterns, the activity of AtWOX14 was more restrictive during lateral root formation. As for AtWOX13, the AtWOX14 promoter activity, revealed by GUS staining, was detected in the LR primordia, but decreased progressively as the LR emerged becoming restricted in fine to the vasculature up to the LR junction (Figure 5A–C). The AtWOX13::GUS gene continued to be expressed for a longer time (compare Figure 5B with Figure 4C). In 5-day-old seedlings, AtWOX14 driven GUS activity was only found in the vascular system and in the pericycle of the primary root (Figure 5D). In the LR of 14-day-old seedlings (Figure 5E), GUS staining was again observed in the vasculature and the pericycle with AtWOX14 expression being progressively induced in the pericycle and in the vasculature as the LR developed. Compared to the activity of the AtWOX13 promoter, AtWOX14::GUS expression stopped at a longer distance further up from the root tip (Figure 5E, black arrow), exactly at the position where the youngest LR primordia were initiated (see arrow in Figure 5F).
Figure 5. AtWOX14 gene expression during plant development. (A to C) Expression profile during the initiation and development of lateral roots obtained with 14-day-old AtWOX14::GUS seedlings. Independent lines were tested. (D) 5-day-old seedling showing staining in the lower part of the hypocotyl and the upper part of the primary root vasculature. (E) Staining profile in primary and lateral roots of 14-day-old seedlings with expression in the upper root vasculature and pericycle. (F) Magnification at the edge of the GUS coloration showing the presence of the youngest LR initial (black arrowhead). (G to I) Expression within the flower showing strong staining in the anther only at floral stage 11–12 (G and H) before opening. Note the absence of staining in the mature anther (I). Scale bars = 100 μm except for C (= 200 μm) and D, E (= 500 μm).
During flower development, AtWOX14::GUS expression was detected in the stamen at floral stage 11/12 (Figure 5G–I), when mitotic divisions of microspores are known to occur . The staining decreased as the stamen matured, showing that the function of WOX14 is linked to the early stages of organ and tissue development. Hence, as for AtWOX13, AtWOX14 promoter activity also seems to be regulated by developmental signals. The gene expression profile reported in the gene atlas  and the tissues and organs affected by the mutation of AtWOX14 (see below) agree with the expression profile described above.
An AtWOX13 gene truncated downstream of the homeodomain neither alters root nor flower development
The only available wox13 mutant line, identified in GABI-Kat  and numbered 193G10, showed a T-DNA insertion upstream of the sequence coding for the characteristic WFQN motif located at the end of the WOX homeodomain (Figure 6A and see additional file 4: Figure 2A). RT-PCR experiments carried out using polyA+-RNA, prepared from the wox13 193G10 line, and primers matching each side of the insertion site, did not yield any PCR product (see Figure 3C). Because of the position of the insertion at the end of the WOX homeodomain, we also carried out an expression analysis using another couple of primers matching the 5'UTR of the AtWOX13 mRNA and the left border of the T-DNA. The latter experiment led to the amplification of a 500 bp PCR product, the sequence of which coded for a putative chimeric protein of 148 residues composed of the 143 amino acids of the AtWOX13 N-terminal region plus an additional 5 amino acids specified by the T-DNA sequence.
Figure 6. The effects of the WOX13 OG mutations on Arabidopsis thaliana development. (A) Diagram of the AtWOX13 protein of the wox13 line 193G10 and the wild type plants (WT). (B) Diagram of the AtWOX14 protein of the wox14 line pst13645 and the wild type plants (WT). Legends: WOX13MOG (pink), HD (orange)(C and D) Comparative primary root growth of mutant and wild type seedlings during the first 6 days of development. In (C), wild type WS ecotype (grey bar) and wox13 line 193G10 (dot bar). In (D), wild type No ecotype (black bar) and wox14 line pst13645 (white bar). (E) Floral transition of wox13 and WS plants after a long day period of 4 weeks.(F) Floral transition of wox14 and Nossen plants after a long day period of 4 weeks. (G to K) Root development of the wild type plant (G and H) compared to the wox14 line pst13645 (I to K) that overproduces adventitious roots. (M to P) Flower development of wox14 line pst13645 inflorescences that show partial sterility (M). Wild type (upper silique in N and O) and wox14 semi-sterile silique (lower siliques in N and P). (N and P) Pictures obtained after clearing. (Q to S) Pollen viability tests with Alexander staining. (Q) Wild type, (R) wox14 pollen. (S) dmc1/RNAi sterile mutant. (T to V). Mature flower development in wild type plants (T) and the wox14 line (U and V). (V) Picture after clearing. Scale bars: Q, V (= 100 μm), R, S (= 50 μm) and T, U (= 500 μm).
Based on the expression profiles obtained using the GUS reporter gene constructs, an analysis of the primary root growth of the mutant lines was carried out after seed germination. Surprisingly, the above WOX13 mutation neither affected primary root growth (Figure 6C), nor vegetative and flower development. The only phenotype appeared to be a much earlier floral transition (Figure 6E).
A WOX14 knock-out mutant modifies lateral root and stamen development of A. thaliana
Conversely to the wox13 mutant, a wox14 insertion mutant line (i.e., pst13645) with the entire WOX homeodomain deleted (Figure 6B and see additional file 4: Figure 2B) displayed developmental defects in all the organs where GUS staining was observed. The primary root growth was retarded (Figure 6D) and after flowering, the pst13645 insertion line showed partial sterility (Figure 6M), with aborted and shorter siliques compared to the wild type (Figure 6N). The floral transition was also altered, however the phenotype observed was that of a late flowering mutant (Figure 6F).
The wox14 mutant was further characterized during plant development. Whereas in the wild type Nossen a fully developed network of LR was observed (Figure 6G, H), the LR formation in wox14 was strongly inhibited (Figure 6I, J). The pst13645 line displayed extra adventitious roots (see Figure 6K, L) whereas the wild type plants developed only one or two. At the reproductive stage, observations after tissue clearing showed that the shorter siliques of wox14 contained many ovules that did not develop into seeds (Figure 6P), compared to the wild type (Figure 6O). Since the AtWOX14 promoter was only active in the developing stamen, pollen grain viability was checked using Alexander staining. As shown in Figure 6Q and 6R, the anthers of both the wild type and the wox14 mutant contained viable pollen grains when compared to the sterile dmc1 RNAi line . However, stamen development was severely affected in the wox14 mutant. In open flowers of the wild type plants, the stamens were longer than the gynoecium allowing complete fertilization of the ovule (Figure 6T). In contrast, in the mutant line, the stamen did not fully develop and remained shorter than the gynoecium thereby preventing efficient fertilization and causing ovule abortion. In some wox14 flowers, the phenotype was even more severe, with aborted stamens that developed to only half the size of the gynoecium (Figure 6V). These data suggested that AtWOX14 prevents premature organ and tissue differentiation at the floral stage.