The body of vertebrates is supported by the vertebral column, a series of segmental skeletal elements that provide both stability and mobility. Axial skeleton development is a multistep process which starts with the formation of somites from the unsegmented paraxial mesoderm that develop at both sides of the neural tube (Tam and Trainor, 1994; Christ and Ordahl, 1995). Shortly after formation, somites become compartmentalized into myotome, dermatome, and sclerotome. In the anuran Xenopus laevis, most somites are myotomal, consisting of cells that will form muscle while the dermatome, a small dorsal lateral part of the somite, will give rise to the dermis of the back, and the sclerotome will form the skeletal elements of the vertebral column and ribs (Keller, 2000).
The sclerotome, the smallest compartment of the somite, made up of polymorphic cells located at the somite ventromedial edge, is adjacent to the notochord, and only becomes distinct from the myotome after the tail bud stage (st. 30) (Youn and Malacinski, 1981a, 1981b; Christ et al., 2004). Thereafter, these cells migrate from the somite to the perinotochordal and perineural space, where they proliferate and differentiate into cartilage and bone of the vertebrae and intervertebral discs (Mookerjee, 1931). In anurans, the vertebral column diverges widely from that of other vertebrates and is characterized by a reduction in the number of vertebrae, absence of discrete caudal vertebrae, and a truncated axial skeleton (Handrigan and Wassersug, 2007). This divergent morphology emerges as a product of the evolutionary modulation of the generalized vertebrate developmental machinery. The acquisition of these morphological features could be associated with changes in the expression patterns and/or functional diversification of the genes involved in the regulation of the proliferation and differentiation of the sclerotome (Handrigan and Wassersug, 2007). Nevertheless, little is known about the way in which the sclerotomal cells change as development proceeds and how these events are regulated at the molecular level in Xenopus laevis (Keller, 2000).
Pax1 and Pax9 play an important role in the formation of the vertebrate axial skeleton, regulating cell proliferation and chondrogenic specification in the sclerotome (Wallin et al., 1994; Peters et al., 1999; Rodrigo et al., 2003). They belong to the Pax family of transcription factors, characterized by the presence of the paired-box domain and of the octapeptide sequence that interacts with the Groucho corepressor and by the absence of a homeobox domain (Strachan and Read, 1994; Eberhard et al., 2000). Both genes are derived from an ancestral preduplication Pax1/9 gene present in ascidians and amphioxus (Holland et al., 1995). While Pax1/9 is expressed only in the endoderm of the developing pharynx (Holland et al., 1995; Ogasawara et al., 1999; Hetzer-Egger et al., 2000), Pax1 and Pax9 have an additional expression in the sclerotomal cells of vertebrates (Wallin et al., 1994; Neubuser et al., 1995; Muller et al., 1996; Nornes et al., 1996; Mise et al., 2008). The functional importance of Pax genes for vertebrate development is demonstrated by the defects of several classical mouse and human mutations (reviewed by Tremblay and Gruss, 1994). After Pax1/9 was duplicated into Pax1 and Pax9, their roles were subfunctionalized as evolution progressed. In medaka fish, Pax1 and Pax9 showed similar expression patterns, except for a strong expression of Pax9 in mesodermal cells of the tail bud that was not observed for Pax1 (Mise et al., 2008). In chicken and mouse, Pax1 and Pax9 exhibited more significant differences in their expression patterns: while Pax9 is expressed mainly in the dorsolateral regions of the sclerotome, Pax1 is expressed in the ventromedial region of the mouse sclerotome and in almost all cells of the chicken sclerotome (Muller et al., 1996; Neubuser et al., 1995). Subfunctionalization of Pax1 and Pax9 roles was supported by the functional studies performed in fish and mouse, where regionalization of their expressions in the sclerotome was found to regulate the differentiation of the distinct components of the vertebrae (Peters et al., 1998, 1999; Wilm et al., 1998; Mise et al., 2008).
Another gene involved in the development of the axial skeleton is Uncx4.1. This gene encodes a paired-type homeobox transcription factor expressed in the developing somite and sclerotome (Saito et al., 1996; Mansouri et al., 1997). In mouse, it is first expressed in the entire caudal half of the newly formed somite and later only in the sclerotome (Mansouri et al., 1997). Additionally, functional studies have shown that Uncx4.1 is required for the condensation of mesenchymal cells of the lateral sclerotome, which is necessary for the specification of pedicles, transverse processes, and proximal ribs (Leitges et al., 2000; Mansouri et al., 2000).
In this study, we isolated the full-length cDNAs of pax1, pax9, and uncx (orthologue of Uncx4.1) of Xenopus laevis and characterized their expression profiles in the developing Xenopus embryo by whole-mount in situ hybridization and reverse transcriptase-polymerase chain reaction (RT-PCR) assay. We found that these three genes have similar but not identical expression patterns located in the sclerotome and in the endoderm of the pharyngeal pouch during Xenopus laevis development.
RESULTS AND DISCUSSION
Isolation and Characterization of Xenopus pax1, pax9, and uncx cDNA
The clones of pax1, pax9, and uncx transcription factors were identified in a screening of an X. laevis expressed sequence tag (EST) library. Sequence analysis of pax1, pax9 and uncx cDNA revealed that they contain an open reading frame of 1077, 1083, and 492 nucleotides that encode proteins predicted to be of 358, 361 and 164 amino acids respectively (GenBank accession no. JQ929179, JQ929180, and JQ929181).
Comparison of the deduced amino acid sequence of Xenopus Pax1 with chick, human, mouse, and zebrafish Pax1 proteins revealed 83, 76, 74, and 71% sequence identity respectively, while the comparison of the deduced amino acid sequence for Xenopus Pax9 with its orthologue in these organisms showed a value of 91, 83, 83, and 70% sequence identity, respectively.
Previous studies have shown that Pax1 and Pax9 genes derive from an ancestral Pax1/9 gene (Holland et al., 1995). This phylogenetic relationship is reflected in the homology between Pax1 and Pax9 Xenopus proteins, which show 71% sequence identity.
An unrooted phylogenetic tree including Pax1 and Pax9 as well as the ancestral Pax1/9 protein sequences clearly indicates the close relationship between these proteins, where Pax1 and Pax9 were grouped with their orthologues (Fig. 1C).
The comparison of the paired-box domain and octapeptide sequences of vertebrates Pax1 and Pax9, and the ancestral Pax1/9 protein of Hemichordata (Ptychodera flava), Urochordata (Ciona intestinalis), and Cephalochordata (Branchiostoma lanceolatum) revealed a striking conservation of the these domains (Fig. 1A,B). It is interesting to note that this gene has remained almost invariable for more than 500 million years, since the divergence between hemichordata and vertebrate lineages (Blair and Hedges, 2005). However, the comparison of the C-terminal region where the transactivation domain is located presented low homology between Pax1 and Pax9 proteins. This could be related to the fact that the C-terminal region undergoes alternative splicing, leading to functionally distinct Pax1 and Pax9 genes (Short and Holland, 2008).
The comparison of the deduced amino acid sequence of Xenopus Uncx with Xenopus tropicalis, human, mouse, chick, and zebrafish Uncx4.1 protein reveals 95, 85, 85, 85, 83% sequence identity, respectively. Alignment of Uncx4.1 protein sequences from different species shows that the homeodomain region is highly conserved (Fig. 2A). A phylogenetic tree based on amino acid sequences indicated that the Uncx4.1 orthologues of zebrafish and X. tropicalis form a group distinct from mammalian, while X. laevis is widely distant from both groups (Fig. 2B). The sequence of the X. laevis Uncx protein is much smaller than that in other organisms, including X. tropicalis. Although we performed an exhaustive search in different gene databases of X. laevis, we were unable to find isoforms of this gene. Consequently, we cannot conclude if this difference observed in the size of the protein corresponds to an isoform or is characteristic of the uncx gene in Xenopus laevis.
Temporal Expression of pax1, pax9, and uncx
The temporal expression of pax1, pax9, and uncx was analyzed by reverse transcriptase-polymerase chain reaction (RT-PCR) using total RNAs isolated from different developmental stages of Xenopus laevis embryos (Fig. 3). The expression of these genes begins during early somitogenesis (st. 17). The transcripts were most abundant during the tail bud stage (st. 25) and expression continued throughout the tadpole stage (st. 30–35-45), suggesting that in X. laevis pax1, pax9, and uncx may be required not only for the differentiation of sclerotomal cells but also for the initiation and maintenance of sclerotome development.
Spatial Expression of pax1, pax9, and uncx
We examined the spatial distribution of Xenopus pax1, pax9, and uncx transcripts by whole-mount in situ hybridization and subsequent histological analyses. The three genes showed similar but not identical expression patterns. The expression of pax1, pax9, and uncx was detected from early tail bud stage in the branchial arches and somites around stage 21–22 (Figs. 4A, 5A, 6A). However, the expression of uncx in somitic region was observed later (Fig. 6B). As development continues, the expression of these genes in the somites progressed from rostral to caudal direction (Figs. 4B–D, 5B–D, 6B–D). These transcripts are expressed in a similar way in the newly formed branchial arches. Histological examination of late tail bud stage embryos (st. 35) enabled a more detailed analysis of the location of the transcripts. Transverse sections of somite and pharyngeal pouch showed that the expression of pax1, pax9, and uncx is located in the sclerotome and endodermal pharyngeal pouch (Figs. 4G,H, 5G,H, 6G,H), as reported for other vertebrates (Neubuser et al., 1995; Peters et al., 1995; Saito et al., 1996; Mansouri et al., 1997; Mise et al., 2008). In this study we detected differences between the spatial expression pattern of Xenopus pax1 and pax9, which in turn differs from the pattern of their orthologues observed in other vertebrates. In fish, the expression of pax1 and pax9 mRNA in the sclerotome is restricted to the ventromedial region of the somite and no differences were detected in their spatial patterns (Mise et al., 2008). However, in amniotes, Pax1 and Pax9 subfunctionalized their roles in the development of the sclerotome (Mise et al., 2008). In mice, while Pax9 is mainly expressed in the lateral region of the sclerotome, Pax1 is expressed medially and ventromedially (Deutsch et al., 1988; Neubuser et al., 1995). Similarly, in chicken, PAX1 is expressed in almost all sclerotome cells, whereas PAX9 expression is found mainly in dorsolaterally located sclerotomal cells (Muller et al., 1996). In the anuran Xenopus laevis, we found that the expression of pax1 is located in the sclerotomal cells of the center of the somites surrounding the notochord and neural tube (Fig. 4E,G,H), whereas pax9 is expressed mainly in the anterior half of the sclerotome around the neural tube (Fig. 5C,E). These sublocations of pax1 and pax9 in the sclerotome could be involved in the regionalization of the sclerotome and in the differentiation of the vertebral column components.
For Xenopus uncx factor, we observed that it is expressed only in the sclerotomal cells surrounding the notochord (Fig. 6E,G,H). This sublocalization in the sclerotome differs from the one observed in mouse, in which Uncx4.1 is expressed in the entire caudal half of the sclerotome (Mansouri et al., 1997). To the best of our knowledge, in other orders of vertebrates the expression pattern of Uncx4.1 has not been described yet.
Pax1, Pax9, and Uncx4.1 genes play a central role in the formation of the axial skeleton of vertebrates. As in other vertebrates, the expression of pax1, pax9, and uncx genes in Xenopus was subregionalized within the sclerotome, so that they would possibly regulate the formation of different components of the vertebral column. The spatial expression pattern of these genes in Xenopus differs from those observed in other vertebrate models, which would make them at least partly responsible for the divergent morphogenesis of the vertebral column in anurans.
Identification and Isolation of cDNA Clones
The NCBI database was screened by using high stringency Tblastx against a pool of Xenopus laevis ESTs from different Xenopus laevis embryonic libraries. The sequences of the Pax1, Pax9, and Uncx4.1 mRNA from mouse (GenBank accession no. NM_008780, NM_011041 and BC051973) were used as probes, and we identified several interesting ESTs. The clones XL085n23 (kindly provided by Dr. Naoto Ueno, NIBB, Okazaki, Japan), BC084222, and BC044278 (acquired through the IMAGE Consortium/LLNL) were fully sequenced, with an open reading frame encoding for X. laevis pax1, pax9, and uncx protein, respectively. Alignments of amino acid sequences were done using the ClustalW algorithm (http://www.ebi.ac.uk/Tools/msa/clustalw2/). The phylogenetic tree was drawn using the Phylodendron application (http://iubio.bio.indiana.edu/treeapp/treeprint-form.html).
Adult male and female Xenopus laevis specimens were stimulated with 400 IU and 800 IU of human chorionic gonadotropin (HCG –Elea Lab.), respectively. Fertilized eggs were obtained after natural single-pair mating and they were dejellied with 2% cysteine hydrochloride (pH 7.8) and cultured in 1× NAM. They were staged according to the Nieuwkoop and Faber tables (1967).
RNA Isolation and RT-PCR Expression Analysis
Total RNA was isolated from whole embryos using TRIREAGENT reagent (MRC) according to manufacturer's instructions. cDNAs were synthesized by M-MLV reverse transcriptase (Promega) with oligo(dT15) priming from 1 μg total RNA extracted from embryos at different stages. PCRs were performed with GoTaq polymerase (Promega) in semi-quantitative amplification conditions and EF-1α was used as an internal standard. The oligonucleotide primers and cycling conditions designed for this study were: EF-1α, 5′-CAG ATT GGT GCT GGA TAT GC-3′ and 5′-ACT GCC TTG ATG ACT CCT AG-3′, 25 cycles; pax1, 5′-CTA CCC TAC CAG CAA CCA ATA TG-3′ and 5′-CAA CTG TCC CAC TAA ATC ACC TC-3′, 29 cycles; pax9, 5′-AGT AGG AAC ACG TTT CAG TCG-3′ and 5′-TTG GAT CCT AGA GAT GAC AGC-3′, 30 cycles; uncx, 5′-TTA AGA CCC TGG AGG TGT GG-3′ and 5′-TGT TTT CGC CCC TGT AAT TC-3′, 30 cycles. The PCR products were analyzed on 1.5% agarose gels. As a negative control, PCR was performed with RNA that had not been reverse-transcribed to check for DNA contamination.
Whole-Mount In Situ Hybridization
The embryos fixation and whole-mount in situ hybridization with digoxigenin-UTP RNA probes were carried out as previously described by Harland (1991), with minor modifications as described by Mancilla and Mayor (1996) and Monsoro-Burq (2007). Antisense RNA probes for pax1, pax9, and uncx were transcribed with T7 RNA polymerase from pax1/pBluescript II, pax9/pCMV-sport6, and uncx/pCMV-sport6 templates linearized with EcoRV, BglII, and HindII, respectively. Digoxigenin-labeled sense probes were used as negative control. Sense probe for pax1 was digested with ApaI and transcribed with T3 RNA polymerase, whereas the pax9 and uncx probes were digested with BamHI and PstI, respectively, and transcribed with SP6 RNA polymerase. Detection of labeled probes was performed using alkaline-phosphatase conjugated anti-digoxigenin Fab fragments and with NBT/BCIP (nitroblue tetrazolium/5-bromo-4-chloro-3-indolyl phosphate) as substrate (Roche). Pigmented embryos were bleached with a solution containing 1% H2O2/PBS.
We thank Naoto Ueno for the EST XL085n23 clone. This research was supported by Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET) and CIUNT grants to S.S.S and Agencia Nacional de Promoción Científica y Tecnológica (FONCYT), Argentina. R.S.S. is a recipient of a CONICET (Argentina) Fellowship. S.S.S. is a Career Investigator of CONICET (Argentina). We also thank Ms. Virginia Mendez for her proofreading.