A balanced chromosomal translocation disrupting ARHGEF9 is associated with epilepsy, anxiety, aggression, and mental retardation

Abstract

Clustering of inhibitory gamma-aminobutyric acid(A) (GABA(A)) and glycine receptors at synapses is thought to involve key interactions between the receptors, a "scaffolding" protein known as gephyrin and the RhoGEF collybistin. We report the identification of a balanced chromosomal translocation in a female patient presenting with a disturbed sleep-wake cycle, late-onset epileptic seizures, increased anxiety, aggressive behavior, and mental retardation, but not hyperekplexia. Fine mapping of the breakpoint indicates disruption of the collybistin gene (ARHGEF9) on chromosome Xq11, while the other breakpoint lies in a region of 18q11 that lacks any known or predicted genes. We show that defective collybistin transcripts are synthesized and exons 7-10 are replaced by cryptic exons from chromosomes X and 18. These mRNAs no longer encode the pleckstrin homology (PH) domain of collybistin, which we now show binds phosphatidylinositol-3-phosphate (PI3P/PtdIns-3-P), a phosphoinositide with an emerging role in membrane trafficking and signal transduction, rather than phosphatidylinositol 3,4,5-trisphosphate (PIP3/PtdIns-3,4,5-P) as previously suggested in the "membrane activation model" of gephyrin clustering. Consistent with this finding, expression of truncated collybistin proteins in cultured neurons interferes with synaptic localization of endogenous gephyrin and GABA(A) receptors. These results suggest that collybistin has a key role in membrane trafficking of gephyrin and selected GABA(A) receptor subtypes involved in epilepsy, anxiety, aggression, insomnia, and learning and memory.

Figure 1

Mapping a balanced chromosomal translocation in ARHGEF9. A: Patient photographs showing dysmorphic features (long and narrow face, irregular teeth, and micrognathia). B: Ideograms depicting the translocated chromosomes and their normal homologs. C: FISH analysis with chromosome 18 BAC clones with signals proximal and distal to the breakpoint and chromosome X BAC clone RP11-943J20, which spans the X breakpoint. D: Southern blot analysis of the breakpoint within ARHGEF9 intron 6. SacI, XbaI, and RsaI digests revealed patient-specific bands of approximately 6, 11, 3, and 1.3 kb, which are absent in the control DNA. E: Sequence of the junction fragments and the respective genomic regions of the normal homologs. The four nucleotides identical in the breakpoint region and lost during rearrangement from one of the derivative chromosomes are boxed. F: Schematic diagram depicting ARHGEF9 with the location of the chromosome breakpoint (arrow) as determined by FISH, breakpoint cloning, and RT-PCR experiments. The breakpoint-spanning BAC RP11-943J20 is shown below the gene. G: ARHGEF9 transcript analysis in patient and control cell line RNAs. For the patient cell line, transcripts could be detected from derivative chromosome 18 (e.g., exons 4–6 and 5–6). However, no products could be obtained using primers spanning the breakpoint (exons 5–7) and within exons located proximal to the breakpoint (exons 8–10) on the derivative X chromosome. Specific bands of the expected size are present in all RT-PCRs from the control.

Figure 2

Rapid amplification of cDNA ends reveals defective ARHGEF9 transcripts resulting in a loss of the collybistin PH domain and C-terminus. A: Analysis of collybistin transcripts using 3′ RACE demonstrates the loss of exons 7–10 from collybistin transcripts, which are replaced by cryptic exons from chromosomes X and 18. B: Splicing of these cryptic exons results in the loss of the collybistin PH domain and C-terminus. C: Using immobilized phosphoinositides and purified GST-collybistin we found that the PH domain is functional, binding PI3P/PtdIns-3-P (phosphatidylinositol-3-phosphate) rather than PIP3/PtdIns-3,4,5-P (phosphatidylinositol 3,4,5-trisphosphate) as predicted by the “membrane activation model” of gephyrin clustering [Kneussel and Betz, 2000]. Controls using centaurin α1 verified that these overlay assays were capable of detecting binding to PIP2 and PIP3.

Figure 3

Cellular models of gephyrin clustering demonstrate that hPEM2-18 and hPEM2-X colocalize with gephyrin, but do not direct the formation of submembrane microaggregates. A: Western blots of myc-tagged hPEM2 constructs expressed in HEK293 cells indicate that the aberrant myc-hPEM2-18 and myc-hPEM2-X proteins are stable and expressed at equivalent levels to myc-hPEM2_SH3+ and myc-hPEM2_SH3− when compared to loading control (α-tubulin). B: Schematic diagram showing the domains specified by the different hPEM2 constructs. C–F: Coexpression of myc-tagged hPEM2 proteins with EGFP-gephyrin in HEK293 cells costained with a nuclear marker (ToPro3). As expected, myc-hPEM2_SH3+ normally colocalizes with gephyrin in large intracellular aggregates (C), while myc-hPEM2_SH3− directs EGFP-gephyrin to submembrane clusters (D). Although they lack the regulatory SH3 domain, myc-hPEM2-18 (E) and myc-hPEM2-X (F) proteins colocalize with gephyrin, but do not direct the formation of submembrane microaggregates.

Figure 4

Overexpression of truncated collybistin mutants in neurons results in loss of gephyrin and GABA_AR clusters. Triple staining of neurons transfected with either myc-hPEM2_SH3+ (A,B), myc-hPEM2_SH3− (C,D) or mutant constructs myc-hPEM2-X (E,F) and myc-hPEM2-18 (G,H) with antibodies against the GABA_A receptor γ2 subunit (red), gephyrin (green), and myc-tag (blue). Note that myc-hPEM2_SH3+ or myc-hPEM2_SH3− do not interfere with the colocalization of gephyrin and GABA_A receptors in synaptic clusters (A–D). Quantitative analysis shows that hPEM2_SH3+ or myc-hPEM2_SH3− do not influence gephyrin puncta number (I) although deletion of the regulatory SH3 domain influences puncta size (J). In neurons expressing myc-hPEM2-X (E,F) and myc-hPEM2-18 (G,H), gephyrin and GABA_A receptor immunoreactivity is significantly reduced. Note that gephyrin and GABA_A receptor clustering in neighboring neurons that do not express the myc-hPEM mutants is unaffected (E–H). Quantitative analysis of gephyrin immunofluorescence (I,J) indicates a statistically-significant loss of gephyrin clusters and a reduction in size of remaining clusters. ^***P<0.001, Student’s t-test, n = 17–28 cells per construct.