Identification and characterization of human ARHGAP23 gene in silico

ARHGAP family genes, such as FNBP2, SRGAP1/ARHGAP13, SRGAP2/ARHGAP14, ARHGAP4 and AHRGAP20/KIAA1391, encode GTPase activating proteins for Rho family proteins (RhoGAPs). Here, we identified and characterized the ARHGAP23 gene by using bioinformatics. KIAA1501 (AB040934.1) was a 5'-truncated partial cDNA derived from the ARHGAP23 gene. Complete coding sequence of human ARHGAP23 cDNA was determined by assembling BM806021 EST, BQ718622 EST, KIAA1501 partial cDNA, and AC115090.8 genome sequence corresponding to exons 7 and 25. ARHGAP23 gene encoded 1491-aa isoform 1 (without exon 23) and 1144-aa isoform 2 (with exon 23) due to alternative splicing. Isoform 2 was C-terminally truncated due to frame-shift within 23-bp exon 23. ARHGAP23 mRNA was expressed in placenta, prostate, hippocampus, brain medulla as well as in brain tumor, salivary gland tumor, head and neck tumor. Mouse 4933428G20 (NM_021493.1) was a 5'-truncated partial cDNA derived from Arhgap23 gene at mouse chromosome 11D. Human ARHGAP23, ARHGAP21 and Xenopus rGAP shared the common domain structure consisting of PDZ, Pleckstrin homology (PH), and RhoGAP domains. ARHGAP23-KIAA1684-MLLT6-RNF110-PIP5K2B-LASP1-PLXDC1-CACNB1 locus at human chromosome 17q12 and CACNB2-PLXDC2-LASP2-MLLT10-BMI1-PIP5K2A-KIAA1217-ARHGAP21 locus at human chromosome 10p12 were paralogous regions (paralogons) with internal inversion. MLLT6, MLLT10 and LASP1 genes are fusion partners of MLL gene in hematological malignancies, while RNF110, PIP5K2B, LASP1 and BMI1 genes are amplified in human tumors. Evolutionary recombination hotspots and oncogenomic recombination hotspots were co-localized around the ARHGAP23-CACNB1 locus and the ARHGAP21-CACNB2 locus.