Rho-Linked Mental Retardation Genes 215 in syndromic and nonsyndromic forms of X-linked MR led to the hypothesis that a disproportionally high density of genes influencing cognitive abilities reside on the X chromosome, recent estimates suggest a downturn in X-linked MR prevalence (Mandel and Chelly, 2004; Poirier et al., 2006). This is further supported by the emerging high frequency of pathogenic autosomal copy number variations. 1.2 Mental Retardation Is Associated with Abberations in Spine Structure and Synaptic Function A major challenge has been to connect the genetic causes of MR to the relevant cel- lular processes and pathways underlying the pathophysiology of human cognitive disorders. One consistent feature of neurons in patients with MR is alterations in the size, shape, stability, and/or number of dendritic spines, which are highly special- ized and dynamic structures on the dendrites on which most excitatory synapses in the brain are located (Kaufmann and Moser, 2000; Fiala et al., 2002; Newey et al., 2005; Tada and Sheng, 2006). More than three decades ago, studies using Golgi impregnations of postmortem material from normal and mentally retarded children demonstrated that changes in dendritic spine density and shape were associated with MR (Huttenlocher, 1970; Marin-Padilla, 1972; Huttenlocher, 1974; Purpura, 1974; Kaufmann and Moser, 2000). For instance, Purpura reported a loss of spines with an absence of short thick spines and a predominance of abnormally long thin spines (resembling immature filopodia) in patients with MR of unknown etiology (Purpura, 1974). Similar alterations with the presence of long tortuous spines have been observed in other defined classes of MR, such as fragile-X syndrome (Rudelli et al., 1985; Hinton et al., 1991; Wisniewski et al., 1991; Irwin et al., 2001). The latter is caused by a trinucleotide CGG repeat expansion and hypermethylation in the 5 untranslated region of the FMR1 gene, which is the most frequent single-gene cause of MR (see further Section 2.2.4). In the case of Down’s syndrome, which results from trisomy of chromosome 21, a reduction in spine density in the neocortex and hippocampus is a common feature; however, short thin spines and spines with large heads and thin necks have also been described (Marin-Padilla, 1972, 1976; Suetsugu and Mehraein, 1980; Takashima et al., 1981; Ferrer and Gullotta, 1990; Takashima et al., 1994). Subsequent studies using mouse models with genetically generated MR (e.g., FMR1 knock-out mouse and partially trisomic mouse, Ts65Dn) consistently reported defects in dendritic spine morphology (Comery et al., 1997; Nimchinsky et al., 2001; Belichenko et al., 2002; Bakker and Oostra, 2003; Dierssen et al., 2003; Galdzicki and Siarey, 2003; Grossman et al., 2006). Thus, alterations in the shape, stability, and/or number of dendritic spines are likely to be a contributive factor to MR. What are the functional implications of alterations in spine morphology and how are they linked to MR? There are numerous observations indicating that spine size, which can range over two orders of magnitude, is of physiological importance. For instance, larger spines can greatly outlast small spines (months compared with hours) (Holtmaat et al., 2005). It is important to note that large 216 N.N. Kasri and L. Van Aelst spines contain large synapses (Harris et al., 1992) with more glutamate-sensitive AMPA receptors (AMPARs), the principal receptors for fast excitatory neurotrans- mission in the mammalian central nervous system (Baude et al., 1995; Nusser et al., 1998; Kharazia and Weinberg, 1999; Takumi et al., 1999), and thus are functionally stronger than small spines. This strong positive correlation between spine size and synaptic strength is maintained in the face of plasticity (Matsuzaki et al., 2004; Kopec et al., 2006). Indeed, increasing evidence indicates that synap- tic plasticity is associated with changes in spine morphology. These morphological changes of spines depend on NMDAR activation and are thought to contribute to activity-dependent formation and elimination of synaptic connections. Two forms of synaptic plasticity, which are considered to be major cellular mechanisms underlying learning and memory, are long-term potentiation (LTP) and long-term depression (LTD) (reviewed in Citri and Malenka, 2008). LTP-inducing stimuli, associated with the addition of AMPARs at the postsynaptic site, cause the formation of new spines and/or the enlargement of existing spines, whereas LTD-inducing stimuli, associated with internalization of AMPARs, lead to shrink- age and/or retraction of spines (reviewed in Malinow and Malenka, 2002; Bredt and Nicoll, 2003; Shepherd and Huganir, 2007; Citri and Malenka, 2008). Given that learning deficit is a constant feature of patients with MR, alterations in synap- tic structure and function are thought to be attributed to some of the MR conditions. Consistent with this notion, increasing evidence suggests that impairments in synap- togenesis and synaptic plasticity contribute to mental and neurological disorders, including MR (Fiala et al., 2002; Bagni and Greenough, 2005; Halpain et al., 2005; Chahrour and Zoghbi, 2007; Dolen and Bear, 2008) (see also Section 2.2). Hence, it is not surprising that large efforts have been devoted towards unraveling the molec- ular and cellular mechanisms underlying synaptic structure, function, and plasticity (reviewed in Shepherd and Huganir, 2007; Citri and Malenka, 2008). Ample evidence points to an active contribution of actin to the modulation of spine morphology and the efficacy of pre- and postsynaptic terminals (reviewed in Cingolani and Goda, 2008). Actin filaments form the main cytoskeleton of den- dritic spines, which are remarkably dynamic. It is widely believed that the regulated polymerization and/or depolymerization of actin underlie spine motility, growth, and shape (Tada and Sheng, 2006; Cingolani and Goda, 2008). Moreover, several observations support the view that dynamic actin filaments are a prerequisite for synapse formation. For instance, activity-dependent synaptogenesis is blocked by actin depolymerizing agents; disruption of signaling pathways implicated in synap- tic actin reorganization results in synaptogenesis defects; and finally, actin plays a part in the development of dendritic spines, thus linking the synapse with actin (reviewed in Cingolani and Goda, 2008). The actin network is also directly involved in synaptic regulation at mature synapses, such as LTP and LTD. Actin-GFP FRET experiments demonstrated that changes in actin polymerization/depolymerization occur in response to different patterns of synaptic stimulation. In particular, these studies showed that tetanic stimulation causes a shift of actin equilibrium towards filamentous actin (F-actin), whereas prolonged low-frequency stimulation causes a shift in actin equilibrium towards G-actin, resulting in a loss of postsynaptic actin Rho-Linked Mental Retardation Genes 217 (Okamoto et al., 2004). Furthermore, F-actin has been shown to be required for stable LTP, suggesting that nascent actin filaments stabilize synaptically delivered AMPARs (Kim and Lisman, 1999; Krucker et al., 2000; Fukazawa et al., 2003; Chen et al., 2004; Okamoto et al., 2004; Lin et al., 2005; Matus, 2005; Honkura et al., 2008). Interestingly, a more recent study showed that synaptic insertion of the AMPAR subunit GluR1, independent of its role of increasing synaptic strength, is required for stable spine enlargement after plasticity-inducing stimuli (Kopec et al., 2007). These findings suggest that AMPARs and nascent actin filaments are interde- pendent and mutually stabilizing. Together, these studies point to an important role for actin in functional and structural plasticity. The central role of actin in the regulation of synaptic structure and func- tion pointed to Rho GTPase family members as central contributors, inasmuch as they are key regulators of actin dynamics and organization (Van Aelst and D’Souza-Schorey, 1997). Indeed, Rho proteins emerged as key regulators of spine morphogenesis, and more recently have been implicated in synapse formation and synaptic plasticity. Furthermore, and significantly, mutations in regulators and effec- tors of the Rho GTPases have been found to underlie various forms of MR. In the remainder of this review, we first briefly discuss the role of Rho GTPases in spine and synapse formation, and subsequently describe in more detail some of the Rho GTPase signaling pathways involved in different forms of MR. 2 Rho GTPases 2.1 Rho GTPases Control Synaptic Structure and Function The Rho family of small GTPases are low-molecular-weight guanine nucleotide- binding proteins, which act as molecular switches cycling between an active GTP-bound form and an inactive GDP-bound form (see Fig. 1). Their activity is tightly controlled by dedicated guanine nucleotide exchange factors (GEFs), which promote GTP-loading, GTPase activating proteins (GAPs), which enhance hydroly- sis of the bound GTP, and guanine-nucleotide-dissociation-inhibitors (GDI), which prevent the exchange of GDP for GTP (Van Aelst and D’Souza-Schorey, 1997). Activated GTP-bound Rho GTPases interact with specific effector molecules to mediate their cellular actions. Of the Rho GTPase family members, RhoA, Rac1, and Cdc42 have been characterized most extensively. These GTPases are best known for their effects on the actin cytoskeleton, and, hence, it is not surprising that they emerged as critical regulators of spine formation and/or maintenance (Govek et al., 2005) (Fig. 1). Several lines of evidence pointed to a role for Rac in spine formation and/or main- tenance and the control of spine morphology in different model organisms ( Govek et al., 2005). These studies largely relied on imaging of individual, fluorescently labeled neurons expressing constitutively active (CA) and dominant-negative (DN) mutant forms of Rac. Expression of CA Rac1 in hippocampal brain slices resulted in the formation of multiple small spines (Nakayama et al., 2000; Tashiro et al., 218 N.N. Kasri and L. Van Aelst G D I e f f e c t o r postsynaptic dendrites dendrite cell body axon synaptic terminals synaptic vessels dendritic spine OPHN1 cytoskeleton organization gene transcription Rho GTP Rho GDP GEF Pi GDP GTP GAP OPHN1 dendritic spine ARHGEF6 PAK3 OCRL1? FMRP MEGAP Fig. 1 Regulatory cycle for the activation and inactivation of Rho GTPases and their involvement in synapse development and maturation. Left panel: Rho GTPases cycle between an inactive GDP and an active GTP bound form. Their activity is tightly controlled by dedicated guanine nucleotide exchange factors (GEFs), which promote GTP-loading; GTPase activating proteins (GAPs), which enhance their intrinsic rate of GTP hydrolysis; and guanine nucleotide dissociation inhibitors (GDIs), which prevent exchange of GDP for GTP and inhibit the intrinsic GTPase activity of GTP-bound GTPases. Only in their active state, Rho GTPases bind to their downstream effectors and exert their effects on various important biological activities. Right panel: Rho GTPases have been implicated in various aspects of neuronal development, including spine/synapse development and maturation. A number of Rho-associated MR gene products are indicated at the appropri- ate positions. Blue label indicates postsynaptic localization and red label presynaptic localization. Abbreviations: ARHGEF6, Rho guanine nucleotide exchange factor 6; FMRP: fragile-X mental retardation-1 protein; MEGAP, Mental disorder-associated GAP protein; OCRL1, the oculocere- brorenal syndrome of Lowe protein 1; OPHN1, oligophrenin-1; PAK3, p21-activated kinase 3 2000; Pilpel and Segal, 2004). This spine phenotype was also observed in trans- genic mice expressing CA Rac1 in Purkinje cells (Luo et al., 1996). Such spines appear to be often engaged in multiple synaptic contacts, which is rarely seen in normal animals (Luo et al., 1996). On the other hand, expression of a DN Rac1 Rho-Linked Mental Retardation Genes 219 mutant in mouse and rat hippocampal slices caused a reduction in spine density and a corresponding reduction in synapse formation (Tashiro et al., 2000; Tashiro and Yuste, 2004). Notably, the s pines of DN Rac transfected neurons were in general sig- nificantly longer than control spines, and detailed analysis revealed that blockade of Rac transforms a subset of existing spines into long, thin filopodia-like protrusions. Furthermore, inhibition of Rac1 reduces spine head growth (particularly in mature neurons), morphological changes, and spine stability (Tashiro and Yuste, 2004). Interestingly, a recent study examining single and double Rac1 and Rac3 (which encodes the closely related, neuron-specific, Rac3 family member) knock-out mice demonstrated that spine formation is strongly hampered only in hippocampal neu- rons lacking both Rac1 and Rac3, implying that Rac1 and Rac3 play complementary roles during late stages of neuronal development. This study additionally showed that the double knock-out mice displayed neurological abnormalities (Corbetta et al., 2009). Recent studies have also coupled Rac1 function to synaptic activity. Wiens et al. found that overexpression of wild-type or CA Rac1 enhances excitatory synaptic transmission and induces clustering of AMPARs in both pre-existing and newly formed dendritic spines, demonstrating that Rac1 can regulate the function of excitatory synapses (Wiens et al., 2005). These findings indicate that Rac1 is not only important for spine morphology and motility, but is also directly coupled to synaptic function. Positive regulators of Rac examined in the context of dendritic spine morphogenesis and/or synaptic function include Tiam1, Kalirin, and α- and β-PIX (Penzes et al., 2008); see Section 2.2). Evidence has been provided that the Rac-GEF Tiam1 acts as a critical mediator of N-methyl-D-aspartate receptor (NMDAR)-dependent spine development (Tolias et al., 2005). Tolias et al. showed that Tiam1 is necessary for spine and synapse development and that it interacts with the NMDAR. Following glutamate applica- tion, they observed that NMDAR-mediated increases in intracellular calcium causes phosphorylation of Tiam1, with a concomitant increase in Rac1 activity required for spine remodeling (Tolias et al., 2005). In a subsequent study, the same group examined EphB receptors, as they are known to form a complex with NMDARs and positively modulate their function. They found that Tiam1 also mediates EphB receptor-dependent dendritic spine development, and proposed a model in which Tiam1 by functioning downstream of both EphB and NMDAR may act as a con- vergence point to help integrate activity-dependent and -independent signaling pathways during the development and remodeling of synaptic connections (Tolias et al., 2007). Recent work also indicates that the Rac-GEF kalirin-7 is a key com- ponent in coupling NMDAR activation to Rac activation and structural plasticity in mature cortical neurons. Xie et al. found that activation of the calcium/calmodulin-dependent kinase II family member (CaMKII) following NMDAR activation directly phosphory- lates kalirin-7 on its N-terminus, thereby stimulating its GEF activity (Xie et al., 2007). Knock-down of kalirin-7 levels reduces basal spine density and the fre- quency of miniature excitatory postsynaptic currents (mEPSCs) (Xie et al., 2007). Also, kalirin has been linked to the EphB receptor. Activation of the EphB recep- tor by ephrin-B has been shown to translocate kalirin-7 to synapses where it locally 220 N.N. Kasri and L. Van Aelst activates Rac1 and its effector PAK, which presumably regulates the actin cytoskele- ton to contribute thereby to proper dendritic spine development (Penzes et al., 2003). Finally, the Rac-GEF βPIX has been shown to be regulated by NMDAR activation and to be critical for activity-dependent synaptogenesis. In particular, Saneyoshi et al. demonstrated that CaMK kinase (CaMKK)/CaMKI and βPIX form a signaling complex in spines, in which CaMKK/CaMKI phosphorylates and stimulates the GEF activity of βPIX to enhance Rac activity and promote forma- tion/stabilization of mushroom-shaped spines (Saneyoshi et al., 2008). Key effectors mediating the effects of Rac on spine morphogenesis and potentially synaptic func- tion include group 1 PAK kinases and the WAVE proteins, and are further discussed in Section 2.2. As with other cellular functions, RhoA appears to work in an opposite fashion to Rac in the regulation of spine structure. In general, increased RhoA activity has been coupled to reduced spine length, size, and density (Tashiro et al., 2000; Ryan et al., 2005; Elia et al., 2006; Sfakianos et al., 2007; Zhang and Macara, 2008), whereas, conversely, low levels of RhoA have been associated with the maintenance of dendritic and spine structures (Nakayama et al., 2000; Van Aelst and Cline, 2004; Sfakianos et al., 2007). Interestingly, a few groups reported a decrease in endoge- nous RhoA activity upon glutamate receptor activation (Van Aelst and Cline, 2004; Schubert et al., 2006), suggesting a link between synaptic input and regulation of endogenous RhoA activity. Recent studies corroborated this and provided further insight into potential molecular links between synaptic activity and RhoA signaling. Kang et al. found a complex formation between the RhoA-GEF, Lfc/GEF- H1, and AMPARs and showed that these proteins colocalize in spines (Kang et al., 2009). Furthermore, they demonstrated that Lfc/GEF-H1 activity negatively regulates spine density and length through a RhoA signaling cascade, and that AMPAR-dependent changes in spine development were eliminated by downregu- lation of Lfc/GEF-H1. Thus, these data suggest that Lfc/GEF-H1 is a key mediator of AMPAR activity-dependent structural plasticity in hippocampal neurons. Nadif Kasri et al. found that the Rho-GAP, Oligophrenin-1, is regulated by synaptic activ- ity and NMDAR activation, and, significantly, that oligophrenin-1 in turn controls synapse maturation and plasticity at the hippocampal CA3-CA1 synapse by stabi- lizing AMPARs (see Section 2.2.1). Finally, the p190RhoGAP has been implicated in the regulation of hippocampal synapse stability by regulating Rho activity in the dendritic spine (Sfakianos et al., 2007). The effects of RhoA activity on spine number and morphology are mediated, at least in part, by the RhoA effector, Rho kinase (Nakayama et al., 2000; Tashiro and Yuste, 2004; Yuste and Bonhoeffer, 2004). Different targets of Rho-kinase have been identified, such as LIMK, myosin light chain (MLC), and MLC phosphatase. Rho-kinase phosphorylates and activates LIMK, which in turn phosphorylates and inactivates the actin depolymerization factor (ADF) cofilin (Maekawa et al., 1999; Sumietal.,1999; Ohashi et al., 2000; Amano et al., 2001). Phosphorylation of MLC by Rho-kinase results in the stimulation of myosin–actin interactions (Amano et al., 1996). Rho-kinase can also regulate the amount of phosphorylated MLC by phos- phorylating and inactivating MLC phosphatase (Kimura et al., 1996). Significantly, a recent study has demonstrated that myosinIIB, which binds and contracts actin Rho-Linked Mental Retardation Genes 221 filaments, is essential for spine morphology and dynamics, as well as synaptic function (Ryu et al., 2006). The role for Cdc42 in spine morphogenesis is less well defined. In hippocampal pyramidal neurons in organotypic slices, expression of a CA- or DN-Cdc42 mutant did not have a significant effect on spine density or length (Tashiro et al., 2000). However, Cdc42 has been demonstrated to affect spine formation in other systems. Loss of function of Cdc42 in vertical system (VS) neurons in the Drosophila visual system leads to a reduction in the density of spinelike structures (Scott et al., 2003), and reduced Cdc42 protein expression is associated with reduced cortical pyra- midal neuron spine density and synapses in insulinlike growth factor 1 (Igf1)–/– brains (Cheng et al., 2003). Furthermore, Cdc42 has been shown to mediate the effects of upstream activators such as the EphB receptor and the Cdc42-specific GEF, intersectin-1, on spine morphogenesis in rat hippocampal neurons (Irie and Yamaguchi, 2002). The presence of both the Cdc42 effector N-Wasp and the EphB receptor had a synergizing effect on the GEF activity of intersectin-1, resulting in high levels of Cdc42-GTP; whereas DN mutant forms of intersectin-1, N-Wasp and Cdc42 interfered with spine formation (Irie and Yamaguchi, 2002). These findings have led to a model in which the EphB receptor, in a complex with intersectin-1 and N-Wasp, triggers the activation of Cdc42 to promote actin poly- merization via N-Wasp and the Arp2/3 complex, leading to spherical expansion of dendritic spine heads. A more recent study identified Numb as an intersectin-1 bind- ing protein (Nishimura et al., 2006). They found that Numb enhanced intersectin-1’s GEF activity for filopodia formation, and demonstrated a role for Numb in spine development. Moreover, they found that Numb forms a complex with the EphB2 receptor and NMDA-type glutamate receptors at the postsynapse together with intersectin, which potentially links Numb to EphB and glutamate receptor signal- ing for synaptic development. In addition to N-Wasp, the insulin receptor substrate 53 (IRSp53) and PAK3 have also been shown to mediate the effects of Cdc42 on spine morphogenesis. Of note, whereas IRSp53 seems to bind equally to Cdc42 and Rac1, Pak3 preferentially binds to Cdc42 (Choi et al., 2005; Kreis et al., 2007;see also Section 2.2). Taken together, these studies clearly implicate Rho GTPase signaling in the struc- tural remodeling of dendritic spines. Emerging evidence also points to a critical role for Rho GTPase signaling in the regulation of synaptic function and plastic- ity. Notably, additional regulators and effectors of Rho GTPases implicated in spine morphogenesis (that are not discussed here) have been reported; for a more detailed description of regulators and effectors of Rho GTPases, see reviews: Govek et al. (2005) and van Galen and Ramakers (2005). 2.2 Mutations in Regulators and Effectors of Rho GTPases Underlie Various Forms of Mental Retardation As discussed above, MR has been associated with abnormalities in s pine structure and function, and Rho GTPases have been implicated in the regulation of these processes. It is thus not surprising that mutations in several regulators (GEFs and 222 N.N. Kasri and L. Van Aelst GAPs) and effectors of the Rho GTPases have been found to underlie or contribute to various forms of MR. These include syndromic and nonsyndromic X-linked forms of MR, as well as autosomal syndromic MR. Below, we discuss several examples demonstrating the involvement of Rho GTPase signaling in the etiology of different forms of MR. These examples also tackle the emerging view of how mutations in Rho-linked genes could result in MR, that is, by disrupting the normal development, structure, and/or plasticity of neuronal networks via perturbations in the regulation of the actin cytoskeleton and gene expression (see also Fig. 2). 2.2.1 Oligophrenin-1 (OPHN1) OPHN1 was the first identified Rho-linked MR gene (Billuart et al., 1998). It encodes the protein OPHN1 that contains a BAR (Bin, amphiphysin, Rvs) and PH domain at its N-terminus, and a GAP domain shown to negatively regulate Rho family members at its C-terminus (Fauchereau et al., 2003; Govek et al., 2004). OPHN1 was initially identified by the analysis of a balanced translocation t(X;12) observed in a female patient with mild MR (Bienvenu et al., 1997). Subsequent studies have revealed the presence of OPHN1 mutations in families with MR asso- ciated with cerebellar hypoplasia and lateral ventricle enlargement (Tentler et al., 1999; Bergmann et al., 2003; Philip et al., 2003; des Portes et al., 2004; Zanni et al., 2005). Abnormal behavior, impaired language skills, and motor development delays were described for several of the patients (Tentler et al., 1999; Bergmann et al., 2003; Philip et al., 2003). All OPHN1 mutations identified to date have been shown, or predicted, to result in OPHN1 loss of function (Zanni et al., 2005), and, interestingly, inactivation of ophn1 in mice has recently been demonstrated to reca- pitulate some of the human phenotypes, such as behavioral, social, and cognitive impairments (Khelfaoui et al., 2007). The OPHN1 protein is expressed in multiple tissues, although with highest levels in the brain, where it is found in neurons of all major regions, including hippocam- pus and cortex, and is present in axons, dendrites, and spines (Govek et al., 2004). Thus, OPHN1 is present both pre- and postsynaptically in neurons. Recent studies have begun to unveil how mutations in OPHN1 may affect neuronal function. In a first study, it was found that knock-down of OPHN1, by using RNA interference (RNAi), in CA1 pyramidal neurons in hippocampal slices results in a significant decrease in dendritic spine length (Govek et al., 2004). This phenotype was mim- icked using a constitutive active (CA) RhoA mutant and was rescued by inhibiting a key effector of RhoA, termed Rho-kinase (Govek et al., 2004). As discussed above, Rho-kinase can influence the actin cytoskeleton by acting on LIM kinase (LIMK), myosin light chain, and/or MLC phosphatase (see also Fig. 2). These findings support a model in which loss of OPHN1 causes aberrations in spine morphology during development as a result of changes in the actin cytoskele- ton triggered upon elevation of RhoA and Rho-kinase activities. More recently, mice lacking the Ophn1 gene were generated, and analysis of these mice showed a decrease in mature spines (Khelfaoui et al., 2007). Surprisingly, in this mouse knock-out model, no obvious deficits in synaptic transmission or plasticity were Rho-Linked Mental Retardation Genes 223 PAK3 CYFIP2/WAVE complex Active WAVE CYFIP FMRP MLC MLCP MLCK LIMK1 Arp2/3 MLC Drebrin OPHN1 MEGAP OCRL1 ARHGEF6 RhoA Cdc42Rac1 Abeta oligomers Rho-kinase WASP Active FMRP CREB F-actin Local regulation of translation PP mRNP Neuronal activity NMDAR AMPAR Fig. 2 Rho-linked mental retardation proteins and effector pathways connecting Rho GTPases to actin dynamics. Proteins encoded by Rho-linked genes involved in different forms of MR are highlighted in red text. See main text for explanation. Abbreviations: AMPAR, α-amino- 3-hydroxy-5-methyl-4-isoxazolepropionic acid receptor; ARHGEF6, a Rho guanine nucleotide exchange factor 6; Arp2/3, actin-related proteins 2 and 3; Aß, amyloid ß; CREB, cAMP-responsive element-binding protein; CYFIP, cytoplasmic FMR1 interacting protein; FMRP, fragile X syn- drome protein; LIM, Lin-11, Isl-1 and Mec-3 kinase; MEGAP, Mental disorder-associated GAP protein; MLC, myosin light chain; MLCK, myosin light chain kinase; MLCP, myosin light chain phosphatase; NMDAR, N-methyl D-aspartate receptor; OCRL1, oculocerebrorenal syndrome of Lowe; OPHN1, oligophrenin-1; PAK, p21-activated kinases; WASP, Wiskott-Aldrich-syndrome protein; WAVE, WASP family Verprolin-homologous protein observed. The interpretation of these data is, however, complicated by the fact that OPHN1 is absent both pre- and postsynaptically in global ophn1 knock-out mice, and there could also be compensatory adaptations during development in the ophn1 knock-out mice. Indeed, by temporally and spatially manipulating OPHN1 gene expression, Nadif Kasri et al. recently demonstrated that postsynaptic OPHN1 plays a key role in activity-dependent maturation and plasticity of excitatory synapses by regulating their structural and functional stability (Nadif Kasri et al., 2009). 224 N.N. Kasri and L. Van Aelst Furthermore, they showed that OPHN1’s localization and function in excitatory synapses is dependent on synaptic activity and NMDA receptor activation, and that OPHN1 regulates synaptic structure and function by controlling the stabilization of AMPA receptors. Therefore, defective OPHN1 signaling results in destabilization of synaptic AMPA receptors and spine structure, leading to impairment in plasticity and eventually loss of spines and NMDA receptors. Together, these results indicate that critical levels of OPHN1 are necessary for proper activity-driven glutamater- gic synapse development and suggest a cellular mechanism by which mutations in OPHN1 can contribute to the cognitive deficits observed in OPHN1 patients. Interestingly, Khelfaoui et al. did report a decrease in paired-pulse facilitation (PPF) in the adult ophn1 global knock-out, a measurement of presynaptic release (Khelfaoui et al., 2007). These findings point towards the potential importance of presynaptic function of OPHN1 signaling. In support of this, recent studies demonstrated that reduced/defective OPHN1 signaling impairs synaptic vesicle (SV) retrieval at hippocampal synapses (Nakano-Kobayashi et al., 2009; Khelfaoni et al., 2009). Nakano-Kobayashi et al. further showed that OPHN1 forms a com- plex with endophilin A1, a protein implicated in membrane curvature generation during SV endocytosis. It is important to note that OPHN1 mutants defective in endophilin A1 binding, or with impaired Rho-GAP activity, fail to substitute for wild-type OPHN1, indicating that OPHN1’s interactions with endophilin A1 and Rho GTPases are important for its function in SV retrieval. These data suggest that defects in efficient SV retrieval may also contribute to the pathogenesis of OPHN1-linked cognitive impairment. Taken together, these data suggest that impairments in both long- and short-term plasticity may contribute to the cognitive deficits observed in OPHN1 patients. 2.2.2 p21-Activated Kinase 3 (PAK3) The second Rho-linked MR gene identified is PAK3. Mutations in PAK3 were found to be the cause of nonsyndromic X-linked MR (see below). PAK3 encodes a member of the group I p21-activated serine/threonine kinases (PAK) (Dan et al., 2001a). The group I PAK proteins (including PAK1, PAK2, and PAK3) function as effectors of the Rac1 and Cdc42 GTPases, and have been demonstrated to mediate their effects on the actin cytoskeleton and gene expression (Jaffer and Chernoff, 2002; Bokoch, 2003). One mechanism by which PAKs affect the actin cytoskeleton involves phos- phorylation and activation of LIMK (Stanyon and Bernard, 1999), which in turn phosphorylates and inhibits cofilin, an actin filament depolymerizing/severing fac- tor, thereby stabilizing actin filaments and promoting actin polymerization (Yang et al., 1998; Edwards et al., 1999; Dan et al., 2001). The regulation of myosins is likely to be another component of PAK-mediated cytoskeletal signaling. There is evidence that PAK1 can interfere with myosin light chain function via direct phos- phorylation and inhibition of myosin light chain kinase (MLCK) (Sanders et al., 1999; Bokoch, 2003). This action of PAK may assist in the disassembly of actin stress fibers triggered by PAK (see also Fig. 2). The group I PAK kinases exist in a dormant state in the cytoplasm as a result of an N-terminal autoinhibitory region, . finally, actin plays a part in the development of dendritic spines, thus linking the synapse with actin (reviewed in Cingolani and Goda, 2008). The actin network is also directly involved in synaptic. syn- drome protein; LIM, Lin-11, Isl-1 and Mec-3 kinase; MEGAP, Mental disorder-associated GAP protein; MLC, myosin light chain; MLCK, myosin light chain kinase; MLCP, myosin light chain phosphatase;. inhibiting a key effector of RhoA, termed Rho-kinase (Govek et al., 2004). As discussed above, Rho-kinase can in uence the actin cytoskeleton by acting on LIM kinase (LIMK), myosin light chain,