Members of the Sox gene family of transcription factors are defined by the presence of an 80 amino acid homology domain, the High Mobility Group (HMG) box.Here we report the cloning and initial analysis of murine Sox-13. The 984 amino acids Sox-13 protein contains a single HMG box, a leucine zipper motif and a glutamine-rich stretch. These characteristics are shared with another member of the Sox gene family, Sox-6. High level embryonic expression of Sox-13 occurs uniquely in the arterial walls of 13.5 days post coitum (dpc) mice and later. Low level expression was observed in the inner ear of 13.5 dpc mice and in a limited number of cells in the thymus of 16.5 dpc mice, from which Sox-13 was originally cloned. At 18.5 dpc, Sox-13 is expressed in the tracheal epithelium below the vocal cord and in the hair follicles. The Sox-13 protein binds to the consensus HMG box motif, AACAAAG, but does not transactivate transcription through a concatamer of this motif. Sox-13, like other members of the Sox family likely plays an important role in development.
The cloning of the mammalian sex-determining gene, SRY, has led to the identification of a large number of related genes (1–6). These so called Sox (Sry-related HMG box) genes are characterized by the presence of a DNA-binding domain, the HMG box (High Mobility Group). The HMG box is a 80 amino acids motif, which mediates sequence specific DNA binding (7). A key role for Sox genes in development has been identified via the analysis of individuals with naturally occurring mutations as well as via the analysis of mice created by targeted gene-disruption.
Mutations in the sex-determining gene SRY lead to sex reversal in XY individuals and have been extensively documented (1,2). Development of the indifferent gonads into testis depends on the presence of an intact HMG box in the SRY protein. SRY from XY individuals suffering from gonadal dysgenesis exhibits an impaired DNA binding or bending activity caused by mutations in its HMG domain (8,9).
The human autosomal XY sex reversal locus, SRA1, has been associated with the skeletal malformation syndrome campomelic dysplasia (CMPD1). This locus has been placed at distal 17q (10). Analysis of this region reveals that patients suffering from these linked syndromes display inactivating mutations within or near the SOX-9 gene (11). Expression of Sox-9 in embryonic cartilage and genital ridges of XY embryos strengthen this observation and imply an essential role for Sox-9 in development of cartilage, and sex determination. It has been suggested that Sox-9 functions downstream of Sry (12).
The transcriptional activator Sox-4 is expressed at several sites in the embryo, but in adult mice its expression is restricted to immature B and T lymphocytes (5). Sox-4−/− mice display a severe defect in cardiac outflow tract formation and die due to a circulatory failure at 14 days post coitum (dpc). Detailed analysis of Sox-4−/− hearts revealed a lack of fusion of the endocardial ridges, impaired development of the semilunar valves and common arterial trunk (13). The arrangement of the ventriculoarterial connection corresponds to transposition of the great arteries (Yin Ya and W.Lamers, submitted). The study of haematopoiesis in lethally irradiated mice reconstituted with Sox-4−/− fetal liver cells revealed a specific block in B cell development at the pro-B cell stage (13). In addition to an impaired B cell development, The Sox-4 mutation also affects the T lineage. Explanted fetal thymic organ cultures (FTOC) of Sox-4−/− thymi yielded 10–50-fold fewer maturing thymocytes than FTOC of littermates (14).
In recent years other Sox genes with possible roles in differentiation and development have been molecularly characterized (15). Sox-1, -2 and -3 are mainly expressed in the developing central nervous system (16). Sox-3 is located on the X chromosome and appears to be the closest relative of Sry, together arguing for an evolutionary relationship (17). In the adult mouse, spermatogenesis is associated with expression of Sox-5 and Sox-6, suggesting overlapping functions of these genes in the testis (3,18,19). Sox-6 may also play a role in the developing nervous system based on its neuronal expression in 9.5–12.5 dpc embryos.
In this study we have searched for Sox genes expressed in 15.5 dpc thymus using degenerate PCR. Besides Sox-4, we identified the expression of Sox-13, a fragment of which had previously been described by Wright et al. (6). We have cloned full length Sox-13 cDNA. The developmental expression of murine Sox-13 was studied by in situ hybridization from 10.5 dpc onwards. Low level expression was observed in thymus, inner ear, vocal cord and hair follicles. High level expression was restricted to arteries during late embryogenesis. Sox-13 binds to the consensus HMG box DNA motif, but does not transactivate transcription from multimeric sites.
Materials and Methods
Production of cDNA from 15.5 dpc thymocytes
Embryos (15.5 dpc) were collected from timed-pregnant mice to dissect embryonic thymi. Thymi were pressed through nylon to recover only thymocytes. After NP-40 lysis mRNA was isolated usingoligo-d(T)beads according to manufacturer's instructions (Dynal). The bound mRNA was directly reverse transcribed using the heat stable reverse transcriptase RtTh (Perklin Elmer) and the oligo-d(T)bead as primer at 72°C for 60 min. After removal of the mRNA a second strand was generated using random 9mers and Klenow at 42°C for 45 min. The two strands were melted, the first strand was extracted using a magnet and stored for re-use at 4°C, the second strand was used for PCR.
Cloning of Sox homologues by guess-mer PCR
PCR was performed using oligonucleotides 5′-gggaattcatgga(tc)gc(gatc)tt(tc)at(gatc)gt(gatc)tgg-3′ and 5′-gggaagctt(gatc)- gg(gatc)cg(ag)ta(ct)tt(ga)ta(ga)t(tc)(gatc)gg-3′ in 50 µl reaction volumes containing 2 U SuperTaq according to the manufacturer's instructions (HT Biotechnology Ltd). Cycle conditions were 40 s at 94°C, 40 s at 50°C, 60 s at 72°C with a final extension of 10 min at 72°C after 45 cycles. The PCR reaction was checked by gel electrophoresis, size selected and purified. The amplified material was subsequently blunted and cloned into EcoRV-digested pBluescript. After transformation into DH10β, colonies were screened on insert-size and analysed for Sox-related products by sequencing.
Isolation of cDNA clones and RACE
An incomplete murine Sox-13 cDNA clone was isolated from a 15 dpc whole mouse embryo cDNA library in lambda-gt11, by standard screening using the Sox-13 PCR fragment as a probe. For the isolation of the missing 5′-end of the cDNA the RACE technique was applied as described by Frohman et al. (20) with minor adaptations. In short, after first strand synthesis, the product was spun over an S-400 column (Pharmacia). After TdT-tailing, the second strand was synthesized using an oligo-d(T)-XbaI primer at 42°C annealing temperature, subsequently the temperature was raised to 56°C and nested primers were added. Cycle conditions: 5 min 94°C, 3 min 42°C, 3 min 72°C, 5 min 94°C and 40 cycles (30 s 94°C, 30 s 56°C, 2 min 72°C) with a final extension of 10 min at 72°C). Primers used for the RACE experiment; oligo-d(T)-XbaI: 5′-ctcgaacctgtctagatc(t)16-3′, Lo-XbaI: 5′-ctcgaacctgtctagatc-3′, Sox-13-primers: 5′-atccatgtcacagctcag-3′, 5′-ctggcttacagggaatgggctg-3′ and 5′-ctcctgctgctgtctggcc-3′, 5′-tctgctcagagtgggccg-3′.
Production of Sox-13 protein
Full length Sox-13 was cloned into pGLOMYC by ligation of NcoI blunted-KpnI RACE fragment and a KpnI-XhoI cDNA fragment into EcoRV-XhoI digested pGLOMYC. pGLOMYC was constructed by inserting a HindIII-BamHI fragment containing a MYC-tag and the 5' untranslated region of β-globin into HindIII-BamHI digested pCDNA. Subsequently, a SpeI-XbaI fragment comprising the 3′ untranslated region of β-globin was inserted into this product digested with XbaI. In vitro transcribed and translated Sox-13 was made according to the manufacturer's instructions (Promega) using T7 RNA polymerase and pGLOMYC-Sox-13 (S) or pGLOMYC (L).
Gel retardation assays
Oligonucleotides were labelled by T4 kinase with [γ-32P]ATP and annealed to their complementary strand. For a typical binding reaction, 1 µl of in vitro transcribed and translated protein was pre-incubated with 1 ng of poly-dIdC at room temperature for 15 min. Labelled probe (10 000–20 000 c.p.m.) was added in 13 µl buffer containing 10 mM HEPES, 60 mM KCl, 1 mM EDTA, 1 mM DTT and 10% glycerol and the mixture was incubated for an additional 30 min at room temperature. The samples were electrophoresed through a non-denaturing 4.5% polyacrylamide gel run in 0.25× TBE at room temperature. Oligonucleotides used were: 56, 5′-ACTGAGAACAAAGCGCTCT-3′ annealed to 3′-TGACTCTTGTTTCGCGAGA-5′ and 56mut, 5′-ACTGAGGGCAAAGCGCTCT-3′ annealed to 3′-TGACTCCCGTTTCGCGAGA-5′.
Cells were transfected by electroporation. In short, 2.5 × 106 cells were transiently transfected with 1 µg of CAT reporter plasmid and 5 µg of pGLOMYC/pGLOMYC-Sox-13 or 5 µg of pCDM8-Sox-4 expression plasmid in a volume of 250 µl. Pulse conditions were 960 µF and 250 V using a Gene PulserApparatus (Bio-rad). Cells were harvested after 48 h and freeze-thawed in 60 µl of 100 mM NaCl, 10 mM Tris, pH 7.4, 1 mM EDTA, 50 µl of lysate were added to 50 µl CAT cocktail ([14C]chloramphenicol 1 µCi/ml [60 mCi/mmol], 2.5% glycerol, 250 mM Tris, pH 7.5, 3 mM butyryl-CoA) and incubated for 2 h at 37°C. Pristane-xylene extractable c.p.m. representing butyrylated [14C]cloramphenicol were determined by liquid scintillation counting.
pCDM8-Sox-4 and the two different reporter constructs were identical as described previously by van de Wetering et al. (5).
Cells were grown in RPMI-1640 supplemented with 5% fetal calf serum and antibiotics. All cell-lines are of murine origin and described by Oosterwegel et al. (21). The 34.1 early, 34.1 late and 18.2 are all derived from fetal thymus organ culture and transformed with a retroviral myc-raf oncogene.
Northern blot analysis
For the cell-line blots total RNA was isolated according to Chomczynski and Sacchi (22). RNA (15 µg) was electrophoresed, transferred to nitrocellulose and hybridized with Sox-13 cDNA probes, labelled by random oligo priming, all according to standard procedures (23). For the cell-line blot an EcoRV fragment (1593–3181) was used.
In situ hybridization
Embryos of sequential developmental stages were collected from timed-pregnant mice. Embryos of 10.5, 13.5, 16.5 and 18.5 dpc and day 1 neonates were used for in situ hybridization and treated as described by Jonker et al. (24). In short, 35S-labelled transcripts were made by in vitro transcription of linearized pBluescript (SK) containing a 725 bp, 5′ located, EcoRI-NsiI fragment (546–1271) of Sox-13. 20 000 c.p.m./µl hybridization solution was added per section. As a control, sections were incubated with a sense Sox-13 probe.
Results and Discussion
Cloning of Sox-13
The developmental characteristics of several Sox genes prompted us to search for Sox homologues with possible roles in thymocyte development, possibly acting in concert with established lymphoid transcription factors (25). To this end, we isolated RNA from 15.5 dpc thymuses. Low stringency PCR with degenerate primers based on homologies between the HMG boxes of SRY and Sox-1–4 (1) was performed. Products of the predicted size were subcloned into pBluescript (SK) and sequenced. Of 40 clones analyzed, 32 encoded the HMG box of Sox-4 (5), while eight encoded the HMG box of Sox-13, originally identified by Wright et al. (6). Screening of a 15 dpc whole mouse embryo lambda-gt11 library with the Sox-13 HMG box yielded a single incomplete Sox-13 cDNA clone. The 5′ sequence was cloned by RACE (Accession no. AJ000740). The sequence predicted an open reading frame of 984 amino acids (Fig. 1A). The ATG at position bp 3, most likely represented the translation initiation codon, based on comparison with partial human SOX-13 cDNAs which have stop codons preceding this conserved ATG (Roose and Clevers, unpublished). A consensus polyadenylation signal (AATAAA) is positioned at bp 3157. The original cDNA clone starts at nucleotide position 546. Of the RACE clones, 70% start at position bp 1, others at bp 27 and bp 324. The encoded protein contained a central HMG box with high homology to Sox-5 and Sox-6, 92 and 91% respectively (6). A putative leucine zipper motif was identified at amino acid position 120–141, immediately followed by a glutamine-rich region from AA 159 to 195 (Fig. 1A). This leucine zipper-Q region was highly homologous to a similar stretch in the Sox-6/Sox-lz and SOX-LZ proteins (Fig. 1B) (19,26). The leucine zipper was clearly amphipatic and, therefore, likely mediates dimerisation of the protein. Indeed, the homologous zipper region in the, Sox-6 orthologue, rainbow trout protein SOX-LZ mediates homodimerisation, as was shown by Takamatsu et al. (26). The conservation of the leucine zipper-Q region indicates a conserved role for this domain in the three proteins, in addition to providing a dimerisation interface.
DNA-binding and transactivating characteristics of Sox-13
To establish whether Sox-13 can bind to the consensus HMG box motif as observed for other Sox proteins, we performed a gel retardation analysis (5). The full length Sox-13 cDNA was cloned in the expression vector pGLOMYC and subsequently transcribed and translated in vitro. As a control, pGLOMYC, containing only the MYC-tag flanked by β-globin untranslated regions was transcribed and translated. The two programmed extracts were incubated with radio labelled gel retardation probes containing either the optimal motif (AACAAAG) or a mutated motif (GGCAAAG). As analysed by gel retardation, Sox-13 (S) bound specifically to the AACAAAG motif (Fig. 2A, lane 2) while no shift was seen using the mutated motif or control extract (L) (lanes 1, 3 and 4). Only one shifted band was observed in the gel retardation of Sox-13, suggesting that Sox-13 binds as a monomer or as a homodimer to the consensus motif in this assay. Homodimerisation of Sox-13 may be mediated by the leucine zipper-Q region in the same fashion as described for SOX-LZ (26).
Having established the binding capacity of Sox-13 to the AACAAAG motif, we subsequently studied its transactivating characteristics. Sox-13 was readily expressed as a nuclear protein from the pGLOMYC vector as analysed in transiently transfected COS cells stained with an anti-myc tag antibody (not shown). pGLOMYC-Sox-13 was co-transfected with pTK(56)7, a reporter-CAT plasmid containing seven copies of the AACAAAG motif inserted upstream of a minimal herpes simplex thymidine kinase (TK) promoter (27). As a control, we used pTK(56sac)7, a TK-CAT vector in which the seven copies of AACAAAG had been replaced by seven CCGCGGT copies (27). CAT transfections were carried out in duplicate and assayed by organic phase separation. Sox-13 did not transactivate transcription through the AACAAAG concatamer in the murine B cell line, AZU II (Fig. 2B), nor in the murine B cell line NS1 (not shown). Sox-4 strongly transactivated Sox-4 strongly transactivated transcription from pTK56(CAT) in the same assay as reported previously (5).
Although some HMG box factors, such as Sox-4 (5), can transactivate transcription from multimerized motifs, it is clearly not a feature of all members of this family. As has been extensively documented for LEF-1, the DNA bending characteristics may endow HMG box factors with potent context dependent transactivation properties (28,29). Such properties can only be studied in the context of natural, complex promoters, that remain to be identified for Sox-13. Alternatively, Sox-13 may need a partner molecule to form a bipartite transcriptional complex, as was shown for Sox-2 (30). Sox-2 is able to form a complex with either the ubiquitous Oct-1 or the embryonic Oct-3, but only in combination with the latter is the ternary complex capable of promoting transcriptional activation (30). As another example, members of the Tcf/LEF family do activate transcription from multimerized motifs when complexed with the Wnt/Wingless effector β-catenin/Armadillo which provides an essential transactivation domain (31).
Lymphoid expression of Sox-13
We next analysed expression of Sox-13 by Northern blot analysis on a selection of well-defined precursor and mature lymphoid cell lines. A single T lineage cell line expressing a Sox-13 species of 4 kb was identified. This cell line, 18.2, was derived from a day 18 fetal thymus (Fig. 3, lane 10). In addition, Northern blot analysis of sorted thymocyte precursors from adult thymuses revealed low levels of Sox-13 expression at all stages (not shown).
Expression of Sox-13 during embryonic development
We investigated Sox-13 expression during embryonic development by in situ hybridization. At 10.5 dpc, no expression of Sox-13 was observed. Sox-13 was first detected in the wall of the great arteries at 13.5 dpc (Fig. 4A and B). In contrast, the vein walls did not express Sox-13. Sox-13 signal was also found in the saccular component of the inner ear (Fig. 4C). In 16.5 dpc embryos, the intensity of Sox-13 expression in the blood vessels resembled that seen in 13.5 dpc (Fig. 5C and D). In the inner ear, a slight but significant signal was seen in the saccular and utricular components as well as in the ampullae of the semicircular canals (Fig. 5A). Finally, a small number of strongly staining cells were observed in the thymus (Fig. 5D). The hybridization signal seen in 16.5 dpc thymus is most likely caused by early thymocytes situated there, given that we amplified the Sox-13 HMG box originally from 15.5 dpc thymocytes and the fact that it is highly expressed in a day 18 thymocyte cell line. No expression was found in various sections of the 16.5 dpc larynx (Fig. 5B). Just before birth (18.5 dpc, Fig. 6), staining in the wall of the arteries had strongly increased. Staining appeared confined to the media and intima. In addition to the great vessels (Fig. 6B), Sox-13 expression included smaller vessels such as the carotid artery, the basilar artery, the pulmonary arteries, the mesenteric artery and the iliac artery (Fig. 6A–D), but remained absent from the wall of the veins. Staining in the thymus remained intense and confined to relatively few cells (Fig. 6B). Sox-13 expression in the inner ear was no longer above background (not shown). New tissues to acquire Sox-13 expression were the tracheal epithelium below the vocal cord (Fig. 6A) and, to a much lesser extent, the hair follicles in the skin (Fig. 6E). In neonates of 1 day old, expression of Sox-13 in the wall of most arteries further increased in intensity (Fig. 7A and B). However, Sox-13 expression in the coronary arteries was confined to their roots even though small arteries elsewhere strongly expressed Sox-13 (Fig. 7B, arrow). In the lung, the pulmonary arteries showed Sox-13 expression, the bronchi or lung epithelium were negative (Fig. 7C). Furthermore, the wall of the veins remained negative for Sox-13 expression. Sox-13 expression in the thymus was comparable to that seen in earlier stages (Fig. 7B), while it increased in intensity in the hair follicles (Fig. 7D).
The specific high level expression suggests that the prime role of Sox-13 involves the development of arterial walls from generic vessels. Little is known about the formation of arteries from endothelial tubes and smooth muscle cells. Few specific markers discriminating developing venous and arterial vessels at an early stage in embryogenesis have been described to date (32). In aortic smooth muscle cells from fetuses of 8–10 and 20–22 weeks of gestation, expression of -smooth muscle myosin and smooth muscle-myosin heavy chain is observed (33). Expression of two other markers, calpoponin and caldesmon, is absent at this stage, but comes up relatively late in embryogenesis, resulting in expression of all four markers in the adult aortic media (33). During the process of vessel-formation, the initial endothelial tubes become invested by mesenchymal cells of unknown lineage (32). Most perivascular tissue arises from mesodermal cells. However, smooth muscle cells of some arteries originate from neural crest (34). In such arteries, i.e. the common carotid vessels, pulmonary trunk, and part of the definitive aortic arch, investment of mesenchymal cells with endothelial cells coincides with the expression of smooth muscle actin and myosin (32). Sox-13 expression was observed in the media of arteries from both neural crest and mesodermal origin in 13.5 dpc embryos and later stages. Gene knock-out experiments are underway to shed more light on the exact function of Sox-13 in the developing mouse embryo.
We thank Marco Schilham for technical assistance, Frits Meijlink for providing us with the cDNA library and Johan van Es and Nick Barker for reviewing the manuscript. We're grateful for the advice and criticism of all the co-worker of the Clevers lab. This project was supported by a PIONIER grant from NWO-GMW.
- polymerase chain reaction
- in situ hybridization
- amino acids
- cell line
- dna, complementary
- hair follicle
- inner ear
- leucine zippers
- vocal folds
- thymus gland
- transcription factor
- embryologic development
- binding (molecular function)