Subtle functional defects in the Arf-specific guanine nucleotide exchange factor IQSEC2 cause non-syndromic X-linked intellectual disability

Abstract

Mutations in IQSEC2, a guanine nucleotide exchange factor for the ADP-ribosylation factor (Arf) family of small GTPases have recently been shown to cause non-syndromic X-linked intellectual disability (ID), characterised by substantial limitations in intellectual functioning and adaptive behaviour. This discovery was revealed by a combination of large-scale resequencing of the X chromosome, and key functional assays that revealed a reduction, but not elimination, of IQSEC2 GEF activity for mutations affecting conserved amino acids in the IQ-like and Sec7 domains. Compromised GTP binding activity of IQSEC2 leading to reduced activation of selected Arf substrates (Arf1, Arf6) is expected to impact on cytoskeletal organization, dendritic spine morphology and synaptic organisation. This study highlights the need for further investigation of the IQSEC gene family and Arf GTPases in neuronal morphology and synaptic function, and suggests that the genes encoding the ArfGEFs IQSEC1 and IQSEC3 should be considered as candidates for screening in autosomal ID.

This article refers to:

Introduction

Intelligence describes the ability of the brain to comprehend, infer, make decisions, reason, use language, learn and plan. These cognitive abilities require the correct assembly of complex neuronal structures during development and a continued ability to adapt in response to internal and external cues. Disruption of any of these elements by environmental and/or genetic factors can lead to deficits in cognitive function. For example, intellectual disability (ID) is defined1 as significantly impaired cognitive functioning coupled with a deficit in adaptive behaviour, with onset before the age of 18. ID is characterised by significantly below-average scores in test of mental ability or intelligence and limitations in the ability to function in areas of daily life, such as communication, self-care, getting along in social situations or school activities. ID is frequent in the population, with as many as 1 in every 50 individuals2 worldwide directly affected. Whilst the causes of ID are highly heterogeneous, genetic studies of X chromosome-linked intellectual disability (XLID) are relatively advanced. Subdividing XLID into ‘non-syndromic’ (where ID is the only consistent clinical abnormality) and ‘syndromic’ forms (where other, consistent defects may manifest) and the collection of large cohorts of families by international consortia (e.g., the IGOLD project; http://goldstudy.cimr.cam.ac.uk/) has facilitated the discovery3 of associated genes. However, although more than 90 XLID genes have been reported, the underlying defect remains unresolved in over 50% of families with non-syndromic XLID. Until recently, identification of the causative mutations in XLID has been achieved primarily by positional and functional candidate gene approaches and targeted re-sequencing of specific regions of the X chromosome. The application of the systematic re-sequencing of the entire X chromosome exome4 has removed bias associated with candidate gene selection and revealed previously unidentified XLID-associated genes such as SYP, ZNF711 or CASK. Despite the success of this approach, several challenges remain to be overcome, including incomplete coverage due to PCR failure, lack of analysis of gene elements outside coding regions (e.g., promoters/enhancers, non-coding RNAs, etc.,) or copy number variations and reliance on automated sequence analysis. Even when changes were detected, distinguishing pathogenic mutations from rare synonymous or non-synonymous variants is not easy in such large data sets, particularly when non-pathogenic truncating variants are present4,5 in 1% of X-chromosome genes. In a small-scale follow-up study6 on one gene, IQSEC2—a GEF for Arf family GTPases (Fig. 1)—we have highlighted how comprehensive data analysis, genetic and functional assays can be used to link DNA sequence variation to ID pathology.

Mutations in IQSEC2 Cause XLID

Three non-synonymous SNPs in IQSEC2 were identified by Tarpey et al. in separate, multigenerational families with non-syndromic ID (Fig. 2A). In MRX1, the first family where non-syndromic ID was mapped7 to the X chromosome, a c.2587C>T change was found in exon 8 leading to a p.R863W substitution. By contrast, in the MRX18 family, a c.2402A>C change in exon 6 leading to a p.Q801P substitution was noticed. A third family ascertained in the USA (US166) had a c.2273G>A change in exon 5 leading to a p.R758Q substitution. Although these variants were not detected in the rest of the resequenced cohort4 or numerous control DNAs, IQSEC2 was not immediately assigned as a new XLID gene.4 Part of the reason for this was that these variants were amongst over 550 unique non-recurrent variants identified in 208 XLID families. Due to the size of this data set, tracking of all variants in all relevant families could not be performed in a timely manner. In addition, these IQSEC2 variants were not amongst the most highly ranked in an assessment4 of likely pathogenicity based on conservation score. However, our follow-up study6 demonstrated that these changes and an additional c.1075C>T change in exon 4 leading to a p.R359C substitution (found in an Australian family, AU128) clearly segregated with the ID phenotype in the respective families. IQSEC2 is also known as BRAG1 or IQ-ArfGEF (Table 1) and occurs in two isoforms6,8 generated by alternative splicing. IQSEC2 is enriched in the PSD at excitatory synapses8,9 and associates with PSD95, a PDZ domain-containing synaptic protein involved in NMDAR clustering (Table 1). Curiously, IQSEC2 transcripts are localised in both hippocampal neuronal cell bodies and dendritic processes,9 suggesting that local translation of IQSEC2 mRNA in an activity-dependent manner could contribute to synaptic plasticity. Other binding partners10,11 include a 53 kDa insulin receptor tyrosine kinase substrate IRSp53 and RIBEYE, a unique scaffolding protein found at retinal ribbon synapses. The IQSEC2 protein contains an IQ-like domain, a Sec7 domain, a PH domain and a C-terminal PDZ domain binding motif (Fig. 2A). The p.R359C substitution is found in the IQ-like domain, which in other GEFs has been shown12 to bind calmodulin and accelerate the exchange activity. Interestingly, the IQSEC2 IQ-like motif lacks the G and second basic residue of the complete motif (Fig. 2B) suggesting that the binding of calmodulin to IQSEC2 is calcium dependent. Since the p.R359C substitution disrupts the conserved basic residue in the IQ-like domain consensus motif (FILV)Qxxx(RK)xxxxxx, this change is predicted to result in disrupted calcium-dependent calmodulin binding, leading to an altered exchange of GDP for GTP on Arf6. Consistent with this theory, the other three sequence variants result in the substitutions p.R758Q, p.Q801P, p.R863W in the Sec7 domain (Fig. 2C), which catalyses GDP-GTP exchange on Arf GTPases8,9 such as Arf1 and Arf6.

Functional Analysis of IQSEC2 Mutations

Since all four amino acid substitutions were located within key functional domains of IQSEC2 and conserved in the related IQSEC1 and IQSEC3 proteins (Fig. 2B and C) we conducted further bioinformatic analysis using online analysis tools such as SIFT (http://sift.jcvi.org/) and Polyphen (http://genetics.bwh.harvard.edu/pph/) which indicated that these changes were likely to be damaging. This suggested that a loss of GEF activity was the most likely underlying molecular mechanism for IQSEC2 XLID. To obtain definitive proof of IQSEC2 dysfunction, we assessed the effect of the three different Sec7 domain mutations on the ability of IQSEC2 to catalyze GDP-GTP exchange on Arf6 in vitro, using recombinant GST-tagged Sec7 domain fusion proteins. GTP binding mediated by the Sec7 domain mutants p.R758Q, p.Q801P and p.R863W was similar to that of a control dominant-negative mutation (p.E849A), demonstrating significantly diminished GTP binding to Arf6 compared to wild-type IQSEC2 (Fig. 3A). Subsequently, we tested the GEF activity of full-length IQSEC2 in a cellular model, using a pull-down assay with the adaptor protein Golgi-localized, γ ear-containing Arf-binding protein 3 (GGA3). GGAs specifically interact with active, GTP-bound Arf, but not inactive Arfs. HA-tagged Arf6 and FLAG-tagged wild-type or mutant versions of IQSEC2 or empty vector were co-expressed in HEK293 cells and cell lysates were subjected to a pull-down assay with GST-GGA3 immobilized on glutathione-agarose. Co-transfection of Arf6 with wild-type IQSEC2 resulted in an approximately 12-fold increase in GTP-bound Arf6, but only a 6- to 8-fold increase when co-expressed with IQSEC2 harbouring mutations in the Sec7 domain (Fig. 3B). These subtle changes in function may provide a plausible reason why IQSEC2 variants showed only moderate conservation scores4 and highlight the importance of functional testing of missense variants. Taken together, these results strongly suggested all four missense changes in IQSEC2 were both functionally relevant and likely to be the disease-causing mutations in these families. We would also note that some carrier females from the four families present with a mild learning disability (although this could not be unequivocally assigned to their IQSEC2 mutation carrier status) and that a complete loss of IQSEC2 function might be expected to yield a more severe neuro-developmental phenotype.

Future Prospects

Having established that mutations in IQSEC2 are causative for XLID, a major challenge is to discover the changes in neuronal function triggered by defective IQSEC2. These could include changes in cell morphology, neuronal migration or synapse formation, function and plasticity. Clearly, the regulation of actin dynamics, membrane trafficking and synaptic function play a pivotal role in this process, since mutations in the several GEFs have been implicated in non-syndromic XLID (Table 2). These include ARHGEF6 encoding α-PIX/Cool-2,13 ARHGEF9 encoding collybistin14–16 and the faciogenital dysplasia (Aarskog-Scott syndrome) gene,17–20 FGD1. In addition, mutations in GTPase-activating proteins (GAPs) and effector proteins for other small GTPases (Rho and Rab families), such as oligophrenin-1 and PAK3, also contribute to ID. Deficits in the function of GEFs, GAPs and effectors affecting actin dynamics have provided a mechanistic link to microstructural deficits (e.g., abnormalities of dendritic spines) associated with intellectual disability.21,22 Certainly, pathways affecting synaptic function and the remodelling of neuronal networks have been suggested23 as an underlying mechanism for a broad spectrum of neurocognitive and neuro-psychiatric phenotypes.

IQSEC2 is currently undergoing additional screening for genetic variants in families with XLID that were not part of recent studies (e.g., the EURO MRX Consortium; http://www.euromrx.com/). Given their similar domain structure, neuronal expression and localisation at synapses (Table 1), other IQSEC gene family members—IQSEC1 and IQSEC3—should also be considered as candidates for genetic studies of autosomal ID. In this regard, it is noteworthy that a maternally inherited 1.65 Mb deletion on 12p13.33 encompassing 10 genes (including IQSEC3) was reported24 to cause moderate intellectual disability and behavioural problems in a 15-year-old boy. However, whether dominant de novo mutations or autosomal recessive forms of ID (ARID) result from mutations in IQSEC1 or IQSEC3 remains to be determined. Large scale re-sequencing is clearly impacting on studies of XLID,4 providing a wealth of data for detailed follow-up studies. The discovery of new disease genes in these disorders also opens new avenues for the development of rational pharmacotherapies, which together with novel and most accurate diagnostic tests25 will provide a brighter outlook for affected individuals and their families.

Abbreviations

Arf	=	ADP-ribosylation factor
GEF	=	guanine nucleotide exchange factor
NMDA	=	N-methyl-Daspartate
XLID	=	X-linked intellectual disability

Figures and Tables

Figure 1 Role of Arf6 in neuronal development and synaptic plasticity. Small GTPases of the ADP ribosylation factor (Arf) family play a crucial role in the regulation of vesicular transport, organelle structure, lipid modification, membrane trafficking and actin dynamics.26 In mammals there are six ubiquitously expressed genes/paralogs, divided into three classes based on sequence similarity (Class I: Arf1–3, Class II: Arf4–5, Class III: Arf6). Class I and II Arfs are localised to the Golgi and endosomal compartments. By contrast, Arf6 is localised to cell periphery in association with the plasma membrane and a subset of endosomes.27 The exchange of GDP for GTP on Arfs is catalysed by guanine nucleotide exchange factors (GEFs). There are at least fifteen ArfGEFs in the human genome,28 each characterised by a catalytic Sec7 domain with sequence identity to the yeast ArfGEF sec7p. Active Arf6 (Arf6-GTP) activates several downstream effectors such as phospholipase D (PLD), phosphatidylinositol-4-phosphate 5-kinases (PIP5K) and Rac1, leading to changes in membrane trafficking and actin dynamics. PA, phosphatidic acid; PIP2, phosphatidylinositol bisphosphate. Modified from reference 29.

Figure 2 IQSEC2 mutations in XLID. (A) Schematic of the human IQSEC2 protein with a regulatory IQ-like motif, a catalytic Sec7 domain, a pleckstrin homology (PH) domain and a PDZ binding motif. The relative locations of mutations found in each family are shown. In the MRX 1 family (MIM 309530), a c.2587C>T change was identified in exon 8 leading to a p.R863W substitution. In the MRX 18 family, a c.2402A>C change in exon 6 was detected, leading to a p.Q801P substitution. A third, unpublished family from the USA (US166) had a c.2273G>A change in exon 5 leading to a p.R758Q substitution. In a fourth, Australian family (AU128) a c.1075C>T change in exon 4 was noted, leading to a p.R359C substitution in the IQ-like domain of IQSEC2. (B and C) Sequence alignments of IQSEC1, IQSEC2 and IQSEC3 showing the location of mutations (bold red type above alignments) in the IQ-like and Sec7 domains. Note that IQ-like motifs lack the G and second basic residue found in canonical IQ motifs. Characters within parentheses can substitute for each other. The R359C mutation disrupts the basic (R) residue. Note that IQSEC2 Sec7 domain mutations do not affect predicted GTP binding residues (bold blue type below alignments) in the structure of the related Sec7 domain in the GEF ARNO. Grey shading indicates amino acids that are identical in all four sequences.

Figure 3 Missense mutations in IQSEC2 that cause non-syndromic XLID reduce guanine nucleotide exchange activity. (A) Radiometric analysis of the effects of mutations in the Sec7 domain of IQSEC2 on guanine nucleotide exchange activity. Recombinant wild-type (dark blue) or mutated versions of the Sec7 domain of IQSEC2 were incubated with recombinant Arf6 and GTPγ³⁵S to catalyze binding of GTPγ³⁵S to Arf6. Arf6-GTPγ³⁵S was isolated on nitrocellulose membrane and the amount of bound radiolabel was quantified in a scintillation counter. Each XLID mutation in the Sec7 domain (light blue) significantly reduced GTP binding to Arf6. *p < 0.0001 vs. wild-type. Note that mutations to E849 (sky blue) are artificial dominant-negative mutants that strongly reduce GEF activity of the Sec7 domain. (B) Guanine nucleotide exchange activity in vivo analyzed using a pull-down assay. Arf6 and either wild-type (dark blue) or mutated (light blue) full-length IQSEC2 were co-expressed in HEK293 cells. The cells were lysed and Arf6-GTP was isolated on beads coated with the adaptor protein Golgi-localized, γ ear-containing Arf binding protein 3 (GGA), which binds to GTP-bound but not GDP-bound Arfs. Arf6-GTP was stripped from the beads and quantified by western blotting. Each mutation reduced activation of Arf6 in this cellular assay.

Table 1 IQSEC family of proteins

Gene symbol	Other names	Location	OMIM	GTPase specificity*	Interacting proteins/proposed biological function	References
IQSEC1	BRAG2; KIAA0763; ArfGEP100	3p25.2	610166	Arf1, Arf5, Arf6	Alpha-catenin (CTNNA1) binding to IQSEC1 activates Arf6 in E-cadherin recycling and actin re-modelling AMPA receptor (GluA2 subunit) binding to IQSEC1 activates Arf6 in receptor endocytosis in LTD	30–34
IQSEC2	BRAG1; KIAA0522; IQ-ArfGEF	Xp11.22	300522	Arf1, Arf6	Proteins of the PSD95 complex, including PSD95 and NMDA receptors bind IQSEC2 at excitatory synapses Insulin receptor tyrosine kinase substrate p53 (IRSp53) binding to IQSEC2 may play a role downstream of NMDA receptors Proteins of the retinal synaptic ribbon, e.g., RIBEYE interacts with IQSEC2 in adult mouse retina	6, 8–11
IQSEC3	BRAG3; KIAA1110; MGC30156; SynArfGEF(Po)	12p13.33	612118	Arf1	IQSEC3 interacts with PSD95, SAP97 and Homer/Vesl1/PSD-Zip45	35, 36

* GTPase in bold depicts the most specific substrate of each GEF determined experimentally.

Table 2 Guanine nucleotide exchange factors implicated in X-linked intellectual disability

Gene symbol	Other names	Location	OMIM	GTPase specificity*	Disease	References
IQSEC2	BRAG1; KIAA0522; IQ-ArfGEF	Xp11.22	300522	Arf family GTPases, Arf1, Arf6	NS-XLID; MRX1, MRX18 (OMIM 309530)	6
ARHGEF6	α-PIX, Cool-2; PIXA; COOL2; KIAA0006	Xq26	300267	Rho family GTPases - Cdc42, Rac1	NS-XLID; MRX46 (OMIM 300436)	13
ARHGEF9	hPEM2; KIAA0424; Collybistin	Xq22.1	300429	Rho family GTPases - Cdc42	NS-XLID, aggression, anxiety, epilepsy, hyperekplexia (OMIM 300607)	14–16
FGD1	AAS; FGDY; ZFYVE3	Xp11.21	300546	Rho family GTPases - Cdc42	NS-XLMR; Aarskog-Scott syndrome (AAS); Faciogenital Dysplasia (OMIM 305400); genital and skeletal developmental abnormalities.	17–20

* GTPase in bold depicts the most specific substrate of each GEF determined experimentally.

Acknowledgements

The authors acknowledge the support of the Australian NHMRC (J.G. and C.S.), the SMILE Foundation (J.G.), the Women's and Children's Hospital Foundation (C.S.), D. Harwood (J.G.) and the Medical Research Council (R.J.H.).

Extra View to: Shoubridge C, Tarpey PS, Abidi F, Ramsden SL, Rujirabanjerd S, Murphy JA, et al. Mutations in the guanine nucleotide exchange factor gene IQSEC2 cause nonsyndromic intellectual disability. Nat Genet 2010; 42:486 - 488; PMID: 20473311; http://dx.doi.org/10.1038/ng.588

 Google Scholar