Identification and characterization of the human long form of Sox5 (L-SOX5) gene

Abstract

The Sox (Sry-type HMG box) group of transcription factors, which is defined by a high-mobility group (HMG) DNA-binding domain, is categorized into six subfamilies. Sox5 and Sox6 belong to the group D subfamily, which is characterized by conserved N-terminal domains including a leucine-zipper, a coiled-coil domain and a Q-box. Group D Sox genes are expressed as long and short transcripts that exhibit differential expression patterns. In mouse, the long form of Sox5, L-Sox5, is co-expressed and interacts with Sox6; together, these two proteins appear to play a key role in chondrogenesis and myogenesis. In humans, however, only the short form of Sox5 has previously been identified. To gain insight into Sox5 function, we have identified and characterized human L-SOX5. The human L-SOX5 cDNA encodes a 763-amino-acid protein that is 416 residues longer than the short form and contains all of the characteristic motifs of group D Sox proteins. The predicted L-SOX5 protein shares 97% amino acid identity with its mouse counterpart and 59% identity with human SOX6. The L-SOX5 gene contains 18 exons and shows similar genomic structure to SOX6. We have identified two transcription start sites in L-SOX5 and multiple alternatively spliced mRNA variants that are distinct from the short form. Unlike the short form, which shows testis-specific expression, L-SOX5 is expressed in multiple tissues. Like SOX6, L-SOX5 shows strong expression in chondrocytes and striated muscles, indicating a likely role in human cartilage and muscle development.

Introduction

The Sox (Sry-type HMG box) family of transcription factors is related to the testis-determining gene Sry and is defined by the presence of a high-mobility group (HMG) DNA-binding domain. Sox proteins interact with DNA through this domain in a sequence-specific manner, binding in the minor groove of DNA and inducing a significant bend (Ferrari et al., 1992, Connor et al., 1994). Therefore, Sox proteins are thought to function as architectural proteins that organize local chromatin structure, assemble other DNA-binding transcription factors and induce correct gene expression (Werner and Burley, 1997, Wolffe, 1994). Sox function is critical to a number of developmental processes, including sex determination (SOX9) (Foster et al., 1994, Wagner et al., 1994), lens development (Sox1, 2 and 3) (Kamachi et al., 1998), T-cell differentiation (Sox4) (van de Wetering et al., 1993), endocardial ridge development (Sox4) (Schilham et al., 1996), developing cardiac and skeletal muscle systems (Sox6) (Hagiwara et al., 2000) and neural crest cell differentiation (Sox10) (Southard-Smith et al., 1998).

Sox proteins are categorized into six subfamilies based on sequence homology within the HMG box and other domains (Pevny and Lovell-Badge, 1997). The group D subfamily consists of Sox5, Sox6 and Sox13. Although many Sox proteins are encoded by a single exon, group D Sox genes contain multiple exons with conserved genomic structure (Wunderle et al., 1996, Argentaro et al., 2000). They also share conserved N-terminal domains, including a leucine zipper, coiled-coil domains and a glutamine-rich region (Q-box), as well as the highly conserved HMG domain (Kido et al., 1998), which are considered critical to their function.

Group D Sox genes are known to express two types of alternatively spliced transcripts, short and long forms (Hiraoka et al., 1998, Argentaro et al., 2000). The short form is a component of the long form, and it lacks the characteristic N-terminal domains. The two forms exhibit different expression patterns: in mouse, the short forms of Sox5 and Sox6 are predominantly expressed in testis (Connor et al., 1995, Denny et al., 1992). In contrast, the long form of Sox6 is expressed in multiple tissues, especially in skeletal muscle, and the long form of Sox5 (L-Sox5) is primarily expressed in cartilage (Lefebvre et al., 1998). Co-expressed with Sox9 during chondrogenesis, L-Sox5 and Sox6 heterodimerize via their coiled-coil domains and activate the type II collagen gene (Col2A1), which encodes a major matrix component protein in cartilage (Lefebvre et al., 1998).

In humans, only the short form of the SOX5 gene has previously been identified (Wunderle et al., 1996). This short form shares high homology with mouse Sox5. Human SOX6, which is predicted to interact with the long form of SOX5, has recently been identified (Cohen-Barak et al., 2001). These facts strongly support the existence of a long form of human SOX5 (L-SOX5). Identification of the human L-SOX5 gene will be an important first step toward determining the precise role of the group D SOX genes in human chondrogenesis and other developmental pathways.

Here we report the isolation and characterization of the human L-SOX5 gene. L-SOX5 contains all cardinal motifs common to group D SOX proteins. Like its mouse counterpart, L-SOX5 has multiple transcriptional start sites and multiple alternative splicing variants, but it shows a unique expression pattern in human tissues.