Introduction

Asporin was first identified as a new member of the class I small leucine-rich repeat proteoglycan (SLRP) family (Lorenzo et al. 2001; Henry et al. 2001) that is up-regulated in the early stages of osteoarthritis (OA) (Lorenzo et al. 2001). The cDNA of the human asporin gene (ASPN); GenBank accession number, NM_017680.2) is 2,466 bp long; it encodes a 382 amino-acid protein with polymorphic aspartic acid (D) repeat in the N-terminus. The predicted amino-acid sequence of ASPN is 91% identical to its mouse counterpart. The human gene was mapped to chromosome 9q22, as part of an SLRP gene cluster composed of the genes encoding extracellular matrix protein 2, osteomodulin, and osteoglycin (Henry et al. 2001).

Osteoarthritis is a common disease that is characterized by breakdown of cartilage in the joint. In a recent case-control association study, we identified ASPN as one of the susceptibility genes for OA (Kizawa et al. 2005). In two independent Japanese populations, a (GAT)14 allele of ASPN encoding 14 aspartic acid repeats (D14) is associated with knee OA; the frequency of this allele increased with radiographic severity of OA. In addition, the D14 allele is also over-represented in hip OA. Subsequent functional analyses revealed that asporin acts a negative regulator of chondrogenesis—it inhibits function of TGF-β through direct physical interaction with TGF-β.

In this report, we provide a high-resolution single nucleotide polymorphism (SNP) map within the 33.4-kb genomic region containing ASPN, in which we detected a total of 19 SNPs, nine of which are novel.

Materials and methods

We screened for SNPs in the 33.4-kb region containing ASPN by polymerase chain reaction (PCR) followed by direct sequence analysis as described previously (Iida et al. 2005). We used genomic DNAs from 48 Japanese individuals with knee OA. We obtained written informed consent to participation in this study, which was approved by the ethical committee of the RIKEN SNP Research Center.

We defined the exon–intron structures of ASPN by comparing its cDNA and genomic sequences (accession numbers: AL157827.17 and AL137848.5, respectively). By direct sequencing of 14 PCR amplicons using a total of 48 sequence primers, we explored SNPs in the 5′-flanking region, all exons, intronic sequences flanking each exon, and the 3′-flanking region. We excluded repetitive sequences from the analysis using RepeatMasker (http://www.repeatmasker.org/). PCR conditions and primer sequences are freely available from the JSNP website (http://www.snp.ims.u-tokyo.ac.jp). For SNP discovery, we performed each PCR using a mixture of genomic DNAs from three individuals; 16 PCR products (corresponding to 48 individuals) amplified from the PCR experiments served as templates for direct sequencing. All SNPs detected by the Polyphred computer program were confirmed by sequencing both strands of each PCR product. Since the genotype of each individual cannot be obtained by this method, we subsequently performed genotyping of individual samples using another 96 Japanese samples. Calculation of Hardy–Weinberg equilibrium linkage disequilibrium (LD) coefficient (D′) was as described (Yamada et al. 2001). Haplotype frequencies were estimated using SNPAlyze software (DYNACOM, Chiba, Japan).

Results and discussion

We sequenced a total of 12-kb of genomic DNA from the ASPN region. We identified 19 SNPs, which are distributed an average of 632 nucleotide apart (Fig. 1, Table 1). We found seven SNPs in the 5′-flanking region, eight in introns, and four in the 3′-untranslated region. After comparing our data with the NCBI dbSNP database (late August 2005), we found 9 of 19 SNPs to be novel. The overall frequencies of nucleotide substitutions were 11% for A/C, 21% for A/G, 11% for C/G, and 58% for C/T. Transitions occurred 3.8 times more frequently than transversions. The substitution rate deviated from 2:1 transition-to-transversion theory and previous results in a gene-based SNP study (Haga et al. 2002). In addition, we identified two insertion–deletion polymorphisms in the 5′-flanking region and intron 1.

Fig. 1
figure 1

Genomic organization of the asporin gene (ASPN) and locations of 21 sequence variations within the gene. The eight exons of ASPN are represented by rectangles. The PCR amplicons are drawn as horizontal bars below the map. Single nucleotide polymorphisms (SNPs—corresponding to the left-most column of Table 1) are indicated above the “chromosome” line. The locations of two sequence variations other than SNPs are indicated below the map

Full size image
Table 1 Characterization of 21 genetic variations in the asporin gene (ASPN) locus. Both 5′ and 3′ flanking sequences to each variation are denoted by lower case letters. Variant nucleotides are indicated by capital letters. SNP Single nucleotide polymorphism
Full size table

To examine the frequencies of the SNPs discovered in ASPN, we genotyped another 96 Japanese samples. The distributions of genotypes and minor-allele frequencies for the 19 SNPs are summarized in Table 2. All genotype distributions were in Hardy–Weinberg equilibrium (P>0.05). The SNP i-ASPN-12 (4071T>C) was not identified in the cohort, probably because of the rarity of this SNP. In earlier report, we used seven genetic variations plus a D-repeat polymorphism to perform an association study of knee OA, and found that all seven variations were in strong LD with each other (Kizawa et al. 2005). Using the same seven genetic variations, we constructed haplotypes around ASPN and found that five haplotypes covered more than 90% of those in the population (Table 3). These haplotypes can be represented by three SNPs: i-ASPN-5, i-ASPN-21, and i-ASPN-17. This information on haplotype structure and “tag” SNPs will be a useful resource with which to examine ASPN associations.

Table 2 Genotype and minor-allele frequencies of 19 SNPs in the ASPN locus (96 chromosomes were genotyped). MAF Minor allele frequency, HWE Hardy–Weinberg equilibrium
Full size table
Table 3 Haplotype structures of ASPN in Japanese samples
Full size table

Asporin is an extracellular matrix (ECM) protein that works as a principal negative regulator of TGF-β signaling in chondrocytes (Kizawa et al. 2005). TGF-β is a cytokine playing critical roles in the regulation of proliferation and differentiation of various types of cells and tissues. Disruption of the functional relationship between TGF-β activity and ECM proteins is known to contribute to several disease phenotypes (Border et al. 1992; Isaka et al. 1996; Neptune et al. 2003). Here, we report a high-resolution SNP map of the 33.4-kb region containing ASPN, showing locations of 21 sequence variations found among a Japanese cohort. The sequence variations identified here should provide a useful resource, not only for further examination of the relationships between genotypes and susceptibility to OA, but also for screening of genes associated with other complex diseases involved in TGF-β signaling cascades, in particular bone and joint diseases.