X-linked forms of mental retardation (XLMR) include a variety of different disorders and may account for up to 25% of all inherited cases of mental retardation. So far, seven X-chromosomal genes mutated in nonspecific mental retardation (MRX) have been identified: FMR2, GDI1, RPS6KA3, IL1RAPL, TM4SF2, OPHN1 and PAK3 (refs 2-9). The products of the latter two have been implicated in regulation of neural plasticity by controlling the activity of small GTPases of the Rho family. Here we report the identification of a new MRX gene, ARHGEF6 (also known as alphaPIX or Cool-2), encoding a protein with homology to guanine nucleotide exchange factors for Rho GTPases (Rho GEF). Molecular analysis of a reciprocal X/21 translocation in a male with mental retardation showed that this gene in Xq26 was disrupted by the rearrangement. Mutation screening of 119 patients with nonspecific mental retardation revealed a mutation in the first intron of ARHGEF6 (IVS1-11T-->C) in all affected males in a large Dutch family. The mutation resulted in preferential skipping of exon 2, predicting a protein lacking 28 amino acids. ARHGEF6 is the eighth MRX gene identified so far and the third such gene to encode a protein that interacts with Rho GTPases.

Members of the Rho GTPase family are key regulatory molecules that link surface receptors to the organization of the actin cytoskeleton. It is now well established that these small GTPases are also crucial for neuronal morphogenesis and connectivity. Moreover, mutations in ARHGEF6 (also known as alphaPIX or Cool-2 ), encoding a Rac1/Cdc42-specific guanine nucleotide exchange factor, have been implicated in X-linked mental retardation. In an attempt to get insight into the biological function of ARHGEF6 and the upstream signaling cascades leading to its activation, we used the full-length coding region of ARHGEF6 as bait in yeast-two hybrid screens and identified PARVB (beta-parvin or affixin) as a novel binding partner. The interaction was confirmed by co-immunoprecipitation and GST pull-down. We showed by immunofluorescence that ARHGEF6 and PARVB co-localize at the cell periphery to lamellipodia and ruffles in well-spread and actively spreading cells adhered to fibronectin. In addition, interaction of ARHGEF6 to ARHGEF7 (betaPIX or Cool-1), a close homolog of ARHGEF6, was confirmed. In in vivo assays, two ARHGEF6 mutations identified previously in patients with X-linked non-specific mental retardation, ARHGEF6 deltaaa56-83 and deltaaa396-776, abolished interaction of ARHGEF6 to PARVB. Binding between ARHGEF6 and ARHGEF7 was not affected by ARHGEF6 deltaaa56-83 but did not occur with ARHGEF6 deltaaa396-776. These data suggest that both the N-terminal calponin homology (CH) and C-terminal coiled-coil domains are necessary for the ARHGEF6-PARVB binding. In contrast, it seems that only the coiled-coil domain is required for the interaction and heterodimerization of ARHGEF6 and ARHGEF7. PARVB is known to interact with integrin-linked kinase (ILK) and is involved in the early stage of cell-substrate interaction through integrins. The identification of PARVB as an ARHGEF6 interacting partner together with the co-localization of ARHGEF6 and ILK in spreading cells suggest that ARHGEF6 is involved in integrin-mediated signaling leading to activation of the GTPases Rac1 and/or Cdc42.

A sample of 114 primary breast tumors and corresponding constitutional DNA were tested for loss of heterozygosity (LOH) of the YNZ22 and p53 genes, both located in the 17p13 region. Loss of the p53 allele was found in 28 of 44 primary breast carcinomas (64%). In contrast LOH in only 26 of 61 tumors (43%) was detected with the variable number of tandem repeats (VNTR) probe YNZ22 mapping at 17p13.3 close to the p53 locus at 17p13.1. Among 19 tumors informative for both probes allele loss at 17p13.3 never occurred without p53 involvement. These data suggest, that p53 is the target of 17p13 allelic deletions in human breast cancer. Immunohistochemistry showed overexpression of the p53 protein in 25 of 50 cases (50%) presumably reflecting activating point mutations. Overexpression was not correlated with allele loss but seemed to be closely related to the presence of point mutations in this study. No homozygous deletions or rearrangements of the p53 gene were detected. This would argue for an important role of heterozygous p53 mutations in human breast cancer.