PAK-interacting guanine nucleotide exchange factor (βPix; also known as Arhgef7) has been implicated in many actin-based cellular processes, including spine morphogenesis in neurons. However, the molecular mechanisms by which βPix controls spine morphology remain elusive. Previously, we have reported the expression of several alternative spliced βPix isoforms in the brain. Here, we report a novel finding that the b isoform of βPix (βPix-b) mediates the regulation of spine and synapse formation. We found that βPix-b, which is mainly expressed in neurons, enhances spine and synapse formation through preferential localization at spines. In neurons, glutamate treatment efficiently stimulates Rac1 GEF activity of βPix-b. The glutamate stimulation also promotes Src-mediated phosphorylation of βPix-b in both an AMPA receptor- and NMDA receptor-dependent manner. Tyrosine 598 (Y598) of βPix-b is identified as the major Src-mediated phosphorylation site. Finally, Y598 phosphorylation of βPix-b enhances its Rac1 GEF activity that is critical for spine and synapse formation. In conclusion, we provide a novel mechanism by which βPix-b regulates activity-dependent spinogenesis and synaptogenesis via Src-mediated phosphorylation.

Dendritic spines are small actin-rich protrusions on the shaft of neuronal dendrites, which establish postsynaptic sites of most excitatory synapses for receiving synaptic input and transmitting information induced by synaptic activity (Koch and Zador, 1993; Korobova and Svitkina, 2010; Shepherd, 1996; Yuste and Majewska, 2001). Because spines play a vital role in building synaptic connectivity in neurons, spine development and maintenance are critical to normal cognitive function and underlie neural processes such as learning and memory. Abnormal spine formation and morphology are associated with numerous neurological disorders (Calabrese et al., 2006; Lai and Ip, 2013; Penzes et al., 2011; van Spronsen and Hoogenraad, 2010). Spines are highly dynamic structures: immature filopodia-like protrusions differentiate into mature spines with various morphologies such as stubby, thin, mushroom-shaped and cup-shaped protrusions (Harris, 1999; Hering and Sheng, 2001; Peters and Kaiserman-Abramof, 1970). The morphology of dendritic spines is largely sustained by activity-dependent rearrangement of actin cytoskeleton (Calabrese et al., 2006; Cingolani and Goda, 2008; Fortin et al., 2012; Lai and Ip, 2013; Matsuzaki et al., 2004; Matus, 2000; Saneyoshi and Hayashi, 2012). Hence, regulatory proteins of the dynamic actin cytoskeleton play essential roles in spine morphogenesis (Calabrese et al., 2006; Fortin et al., 2012; Lin and Webb, 2009; Tolias et al., 2011). The key regulators for the organization of actin cytoskeleton in spines are Rho family GTPases including RhoA, Rac1 and Cdc42 (Newey et al., 2005). Among Rho GTPases, Rac1 is well established for its role in spine morphogenesis in both animals and cultured neurons (Govek et al., 2005). The activity of Rho GTPases is tightly modulated through GDP/GTP cycling between inactive GDP-bound and active GTP-bound forms. Rho GTPases are activated by guanine nucleotide exchange factors (GEFs), which catalyze the exchange of GDP for GTP on Rho GTPases (Schmidt and Hall, 2002), and inactivated by GTPase-activating proteins (GAPs), which facilitate hydrolysis of GTP to GDP (Bernards and Settleman, 2004). For example, several Rac1 GEFs including Tiam1, kalirin and GEFT (also known as Arhgef25) have been reported to be localized in dendritic spines and to promote their development (Bryan et al., 2004; Penzes et al., 2001; Tolias et al., 2005; Xie et al., 2007; Zhang and Macara, 2006).

βPix (also known as Arhgef7) is a Rac1 GEF involved in synaptogenesis (Saneyoshi et al., 2008; Zhang et al., 2003, 2005). βPix is a multidomain protein consisting of an Src homology 3 (SH3) domain, a Dbl homology (DH)-Pleckstrin homology (PH) Rac/Cdc42 GEF domain, a GIT1-binding domain (GBD) and a proline-rich region (Rosenberger and Kutsche, 2006). βPix interacts with Rac1 or PAK proteins, downstream effector kinases of Rac1, through its SH3 domain and has GEF activity through its DH domain (Bagrodia et al., 1998; ten Klooster et al., 2006). In neurons, βPix has been shown to be localized to spines and involved in the promotion of spine formation in a CaMKI-dependent mechanism (Saneyoshi et al., 2008). Previously, we reported several spliced variants of βPix including βPix-a, -b and -d, showing different expression patterns (Kim et al., 2000; Kim and Park, 2001; Oh et al., 1997). Of these, βPix-a is expressed in all tissues, whereas βPix-b and -d, the βPix isoforms containing the unique insert region, are mainly expressed in the brain, suggesting that βPix-b or -d may play an important role in the regulation of neuron-specific processes. Interestingly, overexpressing the ubiquitous βPix-a isoform has been reported to increase immature filopodia-like protrusions while inhibiting the formation of mature dendritic spines in cultured hippocampal neurons (Zhang et al., 2003). These findings suggest the specific role of neuronal βPix isoforms in spine maturation, yet their potential function remains largely unexplored.

Here, we investigated the roles of βPix-b in spine and synapse formation and uncover the molecular mechanisms underlying the regulation of its activity. We demonstrate that βPix-b is highly expressed in neurons, and overexpression of βPix-b enhances spine development and synapse formation. In addition, glutamate stimulation specifically increases the GEF activity of βPix-b among βPix isoforms. We also found that only βPix-b displays high levels of tyrosine phosphorylation upon glutamate stimulation and its GEF activity correlates with its phosphorylation level. Tyrosine phosphorylation of βPix-b upon glutamate stimulation is mediated by Src family kinases. We identified that tyrosine 598 (Y598) of βPix-b is the major Src-mediated phosphorylation site, and demonstrate that Y598 phosphorylation of βPix-b plays an essential role in regulating its GEF activity, which is critical for spine development and synapse formation in neurons. These findings suggest a novel function of βPix-b in neurons and provide new understanding of the molecular mechanisms underlying activity-dependent spine morphogenesis and synaptic plasticity.

Expression patterns of βPix isoforms in tissues

Previous reports have demonstrated the presence of βPix splice variants in the brain (Bagrodia et al., 1998; Kim et al., 2000; Kim and Park, 2001; Koh et al., 2001). The modular structure of βPix isoforms varies; βPix-b and -d isoforms, but not the ubiquitous βPix-a, contain a common insert region (Fig. 1A). In addition, the βPix-d isoform also lacks the C-terminal coiled-coil domain that is known to mediate dimerization of βPix (Kim and Park, 2001). However, the differential molecular function of these isoforms in the brain remains unclear. To investigate the role of the βPix isoforms, we first examined their expression levels in rat tissues. Western blot analysis of rat tissue samples with anti-βPix antibody recognizing its SH3 domain or the insert domain showed various expression patterns of the βPix isoforms. βPix-b and -d were mainly expressed in the brain. In contrast, βPix-a was expressed in all tissues investigated (Fig. 1B). To further characterize whether expression of βPix-b and -d is specific to neurons, immunoblotting for the βPix isoforms in cultured rat hippocampal neurons, cortical neurons and glial cells was performed. βPix-b and -d were detected in hippocampal and cortical neurons, but not in glial cells, whereas βPix-a was expressed in both neurons and glial cells (Fig. 1C). Immunohistochemistry of rat brain sections with anti-βPix insert antibody also showed that βPix-b and -d, the insert region-containing βPix isoforms, were specifically expressed in neurons of the cerebral cortex and hippocampus (Fig. 1D, arrows). To examine the expression levels of βPix isoforms during brain development, embryonic (E18) and adult rat brain extracts were analyzed by immunoblotting with anti-βPix SH3 antibody. Interestingly, βPix-a was abundant in E18 embryonic brains, but its expression was drastically reduced in adult brains. In contrast, βPix-b and -d expression were consistently high at both stages (Fig. 1E). These data suggest that βPix-b and -d may play important roles in neuronal development and neuronal maintenance in adult brains.

Fig. 1.

Expression patterns of βPix isoforms. (A) Schematic representation of βPix isoforms. CC, coiled-coil domain; DH, Dbl homology domain; GBD, GIT1-binding domain; INS, insert region; PH, Pleckstrin homology domain; PXXP, proline-rich region; SH3, Src homology 3 domain. (B) Expression of βPix isoforms is shown in adult rat tissues. Lysates were analyzed by immunoblotting with anti-SH3 or anti-insert antibodies. Br, brain; He, heart; Ki, kidney; Li, liver; Lu, lung; Mu, muscle; Sp, spleen; Th, thymus. (C) βPix-b is specifically expressed in neurons. 18 DIV rat hippocampal neurons (Hp), cortical neurons (Ct) and 21 DIV glial cells (Gl) were analyzed by immunoblotting with anti-SH3 or anti-insert antibodies. (D) Immunohistochemistry of adult rat brain with anti-insert antibody in the cerebral cortex and hippocampal CA1 region is shown. Positive staining was observed as dark-brown precipitates of diaminobenzidine (arrows). Scale bar: 100 µm. (E) βPix expression is developmentally regulated. Brain extracts of E18 embryos and adult rats were analyzed by immunoblotting with anti-SH3 antibody.

Fig. 1.

Expression patterns of βPix isoforms. (A) Schematic representation of βPix isoforms. CC, coiled-coil domain; DH, Dbl homology domain; GBD, GIT1-binding domain; INS, insert region; PH, Pleckstrin homology domain; PXXP, proline-rich region; SH3, Src homology 3 domain. (B) Expression of βPix isoforms is shown in adult rat tissues. Lysates were analyzed by immunoblotting with anti-SH3 or anti-insert antibodies. Br, brain; He, heart; Ki, kidney; Li, liver; Lu, lung; Mu, muscle; Sp, spleen; Th, thymus. (C) βPix-b is specifically expressed in neurons. 18 DIV rat hippocampal neurons (Hp), cortical neurons (Ct) and 21 DIV glial cells (Gl) were analyzed by immunoblotting with anti-SH3 or anti-insert antibodies. (D) Immunohistochemistry of adult rat brain with anti-insert antibody in the cerebral cortex and hippocampal CA1 region is shown. Positive staining was observed as dark-brown precipitates of diaminobenzidine (arrows). Scale bar: 100 µm. (E) βPix expression is developmentally regulated. Brain extracts of E18 embryos and adult rats were analyzed by immunoblotting with anti-SH3 antibody.

βPix-b enhances spine development and synapse formation through its localization at spines

In a previous study, it was shown that expression of βPix-a enhanced the formation of dendritic filopodia in cultured neurons. However, these filopodia failed to develop into mature spines and synapses (Zhang et al., 2003). Here, we examined whether βPix-b is involved in the formation of mature spines and synapses. Here, we focused on the b isoform because its expression was enriched in neurons and its domain structures were identical to those of βPix-a, except for the presence of the neuronal isoform-specific insert region (Fig. 1). To examine the roles of βPix-b in spine development, 7 days in vitro (DIV) rat hippocampal neurons were transfected with GFP, GFP-βPix-a or GFP-βPix-b constructs. Their effects on spines and synapses were then analyzed at 19 DIV. Synapses were identified by immunostaining for synaptophysin (presynaptic marker) (Fig. 2A). First, we examined the fluorescence intensity profiles of βPix-a and -b to compare their localization in spines. The preferential localization to spines was assessed using the spine localization index described in the Materials and Methods. The index indicates the ratio of fluorescence intensity in the spine to that in the shaft; when the value is greater than 1, it indicates that the fluorescence signal localizes more in the spine than in the shaft. In the neurons overexpressing βPix-a, the spine localization index was increased by 57.5% when compared with control (GFP only), supporting localization of βPix-a in spines. Notably, the localization index in the neurons overexpressing GFP-βPix-b was significantly higher (160% increase compared with control) than that of GFP-βPix-a, indicating that βPix-b more preferentially localizes in spine structures than βPix-a (Fig. 2B). This finding is consistent with the notion that βPix-b is the βPix isoform that plays a primary role in the regulation of spine development. Indeed, neurons overexpressing GFP-βPix-b strongly increased spine density without altering filopodia density compared with that in GFP control (Fig. 2C). In contrast, GFP-βPix-a overexpression resulted in decreased dendritic spine density but increased filopodia density, consistent with Zhang et al. (2003). In addition, GFP-βPix-a overexpression reduced synapse formation, whereas GFP-βPix-b overexpression enhanced synapse formation (Fig. 2D). Taken together, these results indicate that, unlike βPix-a, βPix-b overexpression enhances the formation of spines and synapses, suggesting differential roles of βPix-a and βPix-b in spine morphogenesis and synapse formation.

Fig. 2.

βPix-b, a major neuronal βPix isoform, enhances dendritic spine development and synapse formation. (A,B) βPix-b preferentially localizes to dendritic spines. Rat hippocampal neurons were transfected with GFP (control), GFP-βPix-a or GFP-βPix-b at 7 DIV, and Dsred (volume marker) was co-expressed to show the overall morphology of the neurons. Transfected neurons were fixed and immunostained for synapses with anti-synaptophysin antibodies. Arrowheads indicate spines. Scale bars: 5 µm. The bar graph illustrates the spine:shaft fluorescence intensity ratio of primary, secondary and tertiary dendrites randomly selected (246–405 dendrite segments per groups from 59–88 neurons). (C,D) βPix-b enhances spine and synapse formation. Quantification of spine density and synapse formation in neurons transfected with the indicated constructs. C illustrates the number of dendritic spines and filopodia per 10 µm. D illustrates the synapse:spine percentage in neurons. Dendritic spines were randomly selected from 246–405 dendrite segments (662–1317 spines per groups). *P<0.05, ***P<0.00001 for spines and synapses; #P<0.05, ###P<0.00001 for filopodia; ns, not significant; ANOVA followed by Bonferroni-corrected t-test. Error bars indicate s.e.m.

Fig. 2.

βPix-b, a major neuronal βPix isoform, enhances dendritic spine development and synapse formation. (A,B) βPix-b preferentially localizes to dendritic spines. Rat hippocampal neurons were transfected with GFP (control), GFP-βPix-a or GFP-βPix-b at 7 DIV, and Dsred (volume marker) was co-expressed to show the overall morphology of the neurons. Transfected neurons were fixed and immunostained for synapses with anti-synaptophysin antibodies. Arrowheads indicate spines. Scale bars: 5 µm. The bar graph illustrates the spine:shaft fluorescence intensity ratio of primary, secondary and tertiary dendrites randomly selected (246–405 dendrite segments per groups from 59–88 neurons). (C,D) βPix-b enhances spine and synapse formation. Quantification of spine density and synapse formation in neurons transfected with the indicated constructs. C illustrates the number of dendritic spines and filopodia per 10 µm. D illustrates the synapse:spine percentage in neurons. Dendritic spines were randomly selected from 246–405 dendrite segments (662–1317 spines per groups). *P<0.05, ***P<0.00001 for spines and synapses; #P<0.05, ###P<0.00001 for filopodia; ns, not significant; ANOVA followed by Bonferroni-corrected t-test. Error bars indicate s.e.m.

Next, we asked how spine localization of βPix-b is regulated. Previous studies demonstrated that βPix-a can homo- or heterodimerize via its C-terminal region including coiled-coil domain and the PDZ binding motif that is important for the intracellular localization of GEF function in both non-neuronal cells and neurons (Kim et al., 2001; Koh et al., 2001; Zhang et al., 2003). Because βPix-b also has the same C-terminal region, we examined whether the C-terminal region of βPix-b affects its localization in neurons using a mutant, βPix-bΔCC (a βPix-b construct devoid of coiled-coil domain and PDZ binding motif) (Fig. S1). Neurons expressing GFP-βPix-bΔCC showed impaired spine localization of the mutant protein and disturbed spine development and synapse formation. These data show that βPix-b localizes to spine through its C-terminal region that is critical for the development of spine and synapse.

βPix-b has higher GEF activity than βPix-a

In addition to the differences in spine localization between βPix-a and -b, their differential GEF activity for Rac1 may contribute to the distinct phenotypes of βPix-a and -b in spine formation. Llano et al. (2015) showed that βPix-b-induced Rac1 activation is drastically inhibited in neurons expressing βPix-b construct devoid of DH domain (βPix-bDHm), indicating that the DH domain of βPix-b has GEF activity for Rac1. Here, we tested whether there is a difference in GEF activity between βPix-a and -b. To compare Rac1-GTP levels activated by βPix-a and -b, Rac1 activation assay using glutathione S-transferase (GST)-PBD (p21-binding domain of PAK that binds to GTP-bound Rac1) was performed in Cos7 cells (Fig. 3A). Expression of GFP-βPix-b caused a significant increase in active Rac1 (Rac1-GTP) levels (2.68-fold increase compared with control), whereas expression of GFP-βPix-a did not. Similar results were also observed in Neuro2a cells (Fig. 3B), indicating that βPix-b has higher GEF activity for Rac1 than βPix-a.

Fig. 3.

βPix-b has higher Rac1 GEF activity than other βPix isoforms. (A,B) βPix-b has higher GEF activity for Rac1 than βPix-a. (A) Cos7 cells were transfected with construct expressing GFP (control), GFP-βPix-a or GFP-βPix-b. After 17–20 h, cell lysates were subjected to GST-PBD (p21-binding domain) pulldown assay (PD; Rac1 activation assay) and immunoblotted (IB) with anti-Rac1 antibodies. Ratio of active Rac1 to input Rac1 levels was normalized to GFP expression level. *P<0.05; ANOVA followed by Bonferroni-corrected t-test. Error bars indicate s.e.m. (B) Lysates from Neuro2a cells expressing Flag (control), Flag-βPix-a or Flag-βPix-b were subjected to GST-PBD pulldown assay and immunoblotted with anti-Rac1 antibodies. Ratio of active Rac1 to input Rac1 level was normalized to Flag expression level. (C) Glutamate stimulation increases the GEF activity of βPix-b, but not that of βPix-a and -d. Left: 19 DIV rat cortical neurons were treated with 50 µM glutamate/5 µM glycine for 5 min. Cell lysates were subjected to GST-Rac1G15A pulldown assay (active GEF assay) and immunoblotted with anti-insert (βPix) and anti-SH3 (βPix) antibodies. Right: quantification of active βPix-b levels normalized to input βPix-b. *P<0.05; Student's t-test. Error bars indicate s.e.m. CB, Coomassie Blue staining.

Fig. 3.

βPix-b has higher Rac1 GEF activity than other βPix isoforms. (A,B) βPix-b has higher GEF activity for Rac1 than βPix-a. (A) Cos7 cells were transfected with construct expressing GFP (control), GFP-βPix-a or GFP-βPix-b. After 17–20 h, cell lysates were subjected to GST-PBD (p21-binding domain) pulldown assay (PD; Rac1 activation assay) and immunoblotted (IB) with anti-Rac1 antibodies. Ratio of active Rac1 to input Rac1 levels was normalized to GFP expression level. *P<0.05; ANOVA followed by Bonferroni-corrected t-test. Error bars indicate s.e.m. (B) Lysates from Neuro2a cells expressing Flag (control), Flag-βPix-a or Flag-βPix-b were subjected to GST-PBD pulldown assay and immunoblotted with anti-Rac1 antibodies. Ratio of active Rac1 to input Rac1 level was normalized to Flag expression level. (C) Glutamate stimulation increases the GEF activity of βPix-b, but not that of βPix-a and -d. Left: 19 DIV rat cortical neurons were treated with 50 µM glutamate/5 µM glycine for 5 min. Cell lysates were subjected to GST-Rac1G15A pulldown assay (active GEF assay) and immunoblotted with anti-insert (βPix) and anti-SH3 (βPix) antibodies. Right: quantification of active βPix-b levels normalized to input βPix-b. *P<0.05; Student's t-test. Error bars indicate s.e.m. CB, Coomassie Blue staining.

Next, we further examined the GEF activity of βPix-b in glutamate-stimulated neurons. Glutamate is known to activate glutamate receptors and alter the status of phosphorylation in numerous proteins that is essential for normal brain function. Above all, neuronal activity induced by glutamate has been reported to affect the morphology and plasticity of dendritic spines through regulators of actin cytoskeleton (Lai and Ip, 2013; Matsuzaki et al., 2004; Matus, 2000). Tiam1 and kalirin-7, Rac1 GEFs, induce actin remodeling in dendritic spines in response to glutamate receptor stimulation (Tolias et al., 2005; Xie et al., 2007). In active GEF assay using GST-Rac1G15A (a nucleotide-free form of Rac1 that binds to activated Rac1 GEF proteins) (Fig. 3C), βPix-a or -d isoforms with active GEF function were not detected even after glutamate treatment. In contrast, βPix-b with active GEF function was present even in the basal state (without stimulation), and the GEF activity level was further increased (by 127%) by glutamate stimulation. Together, these results show that βPix-b is a more efficient GEF protein than ubiquitous βPix-a and the GEF activity of βPix-b can be regulated effectively in spines by glutamate stimulation.

Glutamate stimulation triggers Src-mediated tyrosine phosphorylation of βPix-b in an AMPA receptor- and NMDA receptor-mediated manner

Next, we set out to investigate the mechanism by which the GEF activity of βPix-a and -b is differentially regulated upon glutamate stimulation. In previous studies, several phosphorylated residues of βPix-a were uncovered (Mayhew et al., 2007), and the phosphorylation of βPix-a is known to be essential for regulation of its GEF activity in neurons and non-neuronal cells (Feng et al., 2006, 2010; Saneyoshi et al., 2008; Shin et al., 2002, 2004). Moreover, neuronal activity induces phosphorylation of βPix-a at serine 516 by a Ca2+-dependent mechanism through N-methyl-D-aspartate (NMDA) receptor stimulation, which is critical for regulating spine formation (Saneyoshi et al., 2008). We hypothesized that glutamate stimulation triggers specific phosphorylation of βPix-b to activate its GEF activity. To test this possibility, we examined phosphorylation levels of βPix-b in glutamate-stimulated neurons. Immunoprecipitation of βPix isoforms in 19 DIV glutamate-stimulated cortical neurons and immunoblotting with anti-4G10 (phosphotyrosine) antibody showed that tyrosine phosphorylation of βPix-b was elevated by glutamate stimulation, by 108% compared with control (Fig. 4A). In contrast, there was no significant increase in phosphotyrosine levels of βPix-a and -d, correlating with the GEF activity assay results in Fig. 3C. These data suggest that the glutamate-dependent increase in both tyrosine phosphorylation and GEF activity is probably specific for βPix-b.

Fig. 4.

Glutamate stimulation increases Src-mediated tyrosine phosphorylation of βPix-b in an AMPA receptor- and NMDA receptor-dependent manner. (A) Among the βPix isoforms, βPix-b is preferentially phosphorylated upon glutamate stimulation. Left: 19 DIV rat cortical neurons were stimulated with 50 µM glutamate/5 µM glycine for 5 min. Cell lysates were incubated with anti-SH3 antibody for immunoprecipitation (IP) and immunoblotted (IB) with anti-4G10 (phosphotyrosine) and anti-SH3 (βPix) antibodies. Right: quantification of βPix-b phosphotyrosine level normalized to immunoprecipitated βPix-b. (B) Src family kinases mediate tyrosine phosphorylation of βPix-b. Left: 50 µM glutamate/5 µM glycine treatment was administered to 19 DIV rat cortical neurons for 5 min. 10 µM PP2 or 10 µM PP3 pretreatment was administered for 15 min before adding glutamate. Cell lysates were incubated with anti-SH3 antibody for immunoprecipitation and immunoblotted with anti-4G10 and anti-SH3 antibodies. Right: quantification of phosphotyrosine levels of βPix-b normalized to immunoprecipitated βPix-b. (C) AMPA receptor and NMDA receptor mediate phosphorylation of βPix-b. Left: 100 µM D-AP5 or 30 µM CNQX pretreatment was administered to 19 DIV rat cortical neurons for 15 min before adding glutamate. Right: quantification of phosphotyrosine levels of βPix-b normalized to immunoprecipitated βPix-b. *P<0.05; ns, not significant; Student's t-test. Error bars indicate s.e.m.

Fig. 4.

Glutamate stimulation increases Src-mediated tyrosine phosphorylation of βPix-b in an AMPA receptor- and NMDA receptor-dependent manner. (A) Among the βPix isoforms, βPix-b is preferentially phosphorylated upon glutamate stimulation. Left: 19 DIV rat cortical neurons were stimulated with 50 µM glutamate/5 µM glycine for 5 min. Cell lysates were incubated with anti-SH3 antibody for immunoprecipitation (IP) and immunoblotted (IB) with anti-4G10 (phosphotyrosine) and anti-SH3 (βPix) antibodies. Right: quantification of βPix-b phosphotyrosine level normalized to immunoprecipitated βPix-b. (B) Src family kinases mediate tyrosine phosphorylation of βPix-b. Left: 50 µM glutamate/5 µM glycine treatment was administered to 19 DIV rat cortical neurons for 5 min. 10 µM PP2 or 10 µM PP3 pretreatment was administered for 15 min before adding glutamate. Cell lysates were incubated with anti-SH3 antibody for immunoprecipitation and immunoblotted with anti-4G10 and anti-SH3 antibodies. Right: quantification of phosphotyrosine levels of βPix-b normalized to immunoprecipitated βPix-b. (C) AMPA receptor and NMDA receptor mediate phosphorylation of βPix-b. Left: 100 µM D-AP5 or 30 µM CNQX pretreatment was administered to 19 DIV rat cortical neurons for 15 min before adding glutamate. Right: quantification of phosphotyrosine levels of βPix-b normalized to immunoprecipitated βPix-b. *P<0.05; ns, not significant; Student's t-test. Error bars indicate s.e.m.

We next investigated which tyrosine kinase is responsible for the phosphorylation of βPix-b in neurons. Among several candidates, Src family kinases are known as key kinases for regulation of various signaling cascades in neurons and have recently been revealed as important regulators of dendritic spine formation (Morita et al., 2006; Repetto et al., 2014; Webb et al., 2007). Moreover, Src was previously reported as an upstream kinase for phosphorylation of βPix-a (Feng et al., 2006, 2010). In non-neuronal cells, βPix-a was phosphorylated at tyrosine 442 in an Src-dependent manner. Hence, we examined whether Src family kinases phosphorylate βPix-b in glutamate-stimulated neurons. Glutamate with PP2, a specific Src-family kinase inhibitor, or PP3, the inactive analog of PP2, was applied to 19 DIV cortical neurons and phosphorylation levels of βPix-b were examined with anti-4G10 antibody (Fig. 4B). Immunoprecipitation in glutamate-stimulated rat cortical neurons at 19 DIV showed that phosphorylated tyrosine levels of βPix-b were increased by glutamate treatment either with or without PP3 pretreatment. However, PP2 pretreatment blocked tyrosine phosphorylation, indicating that glutamate-induced tyrosine phosphorylation of βPix-b is mediated by an Src-dependent mechanism.

Next, we searched for glutamate receptors involved in the activation of Src family kinases. Previous studies demonstrated the activation of Src family kinases through α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptor activation (Hayashi et al., 1999; Socodato et al., 2012; Wu et al., 2004) and AMPA receptor-mediated synaptic stimulation involved in the maintenance of dendritic spines (Mateos et al., 2007; McKinney et al., 1999; Takahashi et al., 2009). As mentioned, NMDA receptor-dependent neuronal activity also plays an important role in the enhancement of spine and synapse formation through phosphorylation of βPix-a for Rac1 activation in dendritic spines (Saneyoshi et al., 2008). However, the major βPix isoform(s) actually responsible for activity-dependent Rac1 activation in neurons has yet to be clarified in the signaling pathway from AMPA or NMDA receptors to actin reorganization. Our findings indicated that βPix-b appeared to be more effective at activating Rac1 compared with βPix-a (Fig. 3). To investigate which glutamate receptor is involved in the regulation of βPix-b, we employed the specific antagonists for each glutamate receptor and first examined their effects on βPix-b phosphorylation. Rat cortical neurons at 19 DIV were pretreated with CNQX, an AMPA receptor antagonist, or D-AP5, an NMDA receptor antagonist, before glutamate application and immunoprecipitation experiments were performed. Following pretreatment with either CNQX or D-AP5, the glutamate-induced phosphorylation of βPix-b tyrosine residues was reduced to the control level (Fig. 4C), showing that both AMPA receptor- and NMDA receptor-dependent neuronal activity induces Src-mediated tyrosine phosphorylation of βPix-b. Since both inhibitors completely blocked the tyrosine phosphorylation of βPix-b, we additionally investigated whether the inhibitors may affect baseline phosphorylation without glutamate stimulation. However, the baseline phosphorylation of CNQX or D-AP5 treatment in unstimulated neurons was not significantly different from that of the control (Fig. S2). This result suggests that AMPA- and NMDA-receptor pathways may interact to synergistically induce the tyrosine phosphorylation of βPix-b upon glutamate stimulation. Together, these data suggest that βPix-b is specifically phosphorylated by Src family kinases in an AMPA receptor- and NMDA receptor-dependent manner, and the phosphorylation levels correlate with the GEF activity shown in Fig. 3.

Tyrosine 598 (Y598) residue in the insert region of βPix-b is the major phosphorylation site by Src

Among the βPix isoforms, glutamate stimulation exclusively elevates phosphorylation levels of the βPix-b isoform. Hence, we hypothesized that there is a βPix-b-specific residue(s) for which phosphorylation is relevant to the regulation of its GEF function. Because the insert region is unique structural differences between βPix-b and βPix-a (Fig. 5A), we investigated the putative tyrosine phosphorylation sites in the insert region. The amino acid sequence of βPix-b was analyzed by the NetPhos program (http://www.cbs.dtu.dk/services/NetPhos/) to assess the probability of phosphorylation. The analysis revealed that tyrosine 598 (Y598) residue in the insert region has the highest probability (probability=0.971, Fig. 5B) among tyrosine residues in βPix-b. To examine whether Y598 is the authentic phosphorylation site, we generated a βPix-b mutant, Flag-βPix-bY598F, by introducing a phenylalanine substitute for Y598 to block phosphorylation and compared tyrosine phosphorylation between wild-type Flag-βPix-b and Flag-βPix-bY598F. Immunoblotting with anti-4G10 antibody in Neuro2a cells revealed that wild-type βPix-b was strongly phosphorylated at tyrosine, whereas the βPix-bY598F mutant was not noticeably phosphorylated (Fig. 5C). Interestingly, Y598 phosphorylation seems to be spontaneous in Neuro2a cells grown in serum-containing media. It is known that Src is basally active in Neuro2a cells and that neurons have high Src activity during development (Kuo et al., 1997). To investigate whether Y598 is phosphorylated by Src, constitutively active Src was co-expressed with Flag-βPix-b or Flag-βPix-bY598F in Cos7 cells and their tyrosine phosphorylation was examined (Fig. 5D). Unlike Neuro2a cells, Cos7 cells did not exhibit spontaneous Y598 phosphorylation. Expression of active Src strongly induced tyrosine phosphorylation of βPix-b. However, this phosphorylation was drastically reduced in βPix-bY598F, indicating that Y598 of βPix-b is the major Src-mediated phosphorylation site.

Fig. 5.

Tyrosine 598 (Y598) of βPix-b is the major site for Src-mediated phosphorylation. (A,B) Predicted tyrosine phosphorylation sites in the insert region of βPix-b. (A) Schematic diagram of βPix-a and βPix-b domain structures. (B) Potential tyrosine phosphorylation sites in the βPix-b sequence were predicted by NetPhos program. Residues showing score >0.5 were defined as predicted phosphorylation sites. Scores closer to 1 indicate higher probability of phosphorylation. The sequence containing the Y598 in the insert region of βPix-b exhibits the highest probability. (C,D) Y598 of βPix-b is the major phosphorylation site by Src. (C) Lysates from Neuro2a cells expressing Flag-βPix-b or Flag-βPix-bY598F were immunoprecipitated (IP) with anti-Flag (M2) antibody. The immunoprecipitates were subjected to immunoblotting (IB) with anti-4G10 and anti-Flag (M2) antibodies. (D) Cos7 cells expressing Flag (control), Flag-βPix-b or Flag-βPix-bY598F were co-expressed with or without active Src. Cell lysates were immunoprecipitated and subjected to immunoblotting as in C. (E) Anti-pY598 antibody specifically recognizes Y598 phosphorylation of βPix-b. Neuro2a cells expressing Flag (control), Flag-βPix-b or Flag-βPix-bY598F were co-expressed with or without active Src and incubated in the presence or absence of serum. Cell lysates were immunoprecipitated and subjected to immunoblotting with anti-pY598 (Y598-phosphorylated βPix-b) and anti-Flag (M2) antibodies.

Fig. 5.

Tyrosine 598 (Y598) of βPix-b is the major site for Src-mediated phosphorylation. (A,B) Predicted tyrosine phosphorylation sites in the insert region of βPix-b. (A) Schematic diagram of βPix-a and βPix-b domain structures. (B) Potential tyrosine phosphorylation sites in the βPix-b sequence were predicted by NetPhos program. Residues showing score >0.5 were defined as predicted phosphorylation sites. Scores closer to 1 indicate higher probability of phosphorylation. The sequence containing the Y598 in the insert region of βPix-b exhibits the highest probability. (C,D) Y598 of βPix-b is the major phosphorylation site by Src. (C) Lysates from Neuro2a cells expressing Flag-βPix-b or Flag-βPix-bY598F were immunoprecipitated (IP) with anti-Flag (M2) antibody. The immunoprecipitates were subjected to immunoblotting (IB) with anti-4G10 and anti-Flag (M2) antibodies. (D) Cos7 cells expressing Flag (control), Flag-βPix-b or Flag-βPix-bY598F were co-expressed with or without active Src. Cell lysates were immunoprecipitated and subjected to immunoblotting as in C. (E) Anti-pY598 antibody specifically recognizes Y598 phosphorylation of βPix-b. Neuro2a cells expressing Flag (control), Flag-βPix-b or Flag-βPix-bY598F were co-expressed with or without active Src and incubated in the presence or absence of serum. Cell lysates were immunoprecipitated and subjected to immunoblotting with anti-pY598 (Y598-phosphorylated βPix-b) and anti-Flag (M2) antibodies.

To further examine the phosphorylation of βPix-b, we generated an anti-pY598 antibody specific for Y598 phosphorylation. This antibody specifically recognized serum- and Src-induced phosphorylation of βPix-b in Neuro2a cells and did not recognize Flag-βPix-bY589F (Fig. 5E), showing its specificity for pY598 of βPix-b. Using anti-pY598 antibody, we investigated the localization of the endogenous βPix-b phosphorylated at Y598 (Fig. S3). Hippocampal neurons at 25 DIV were immunostained with anti-pY598 antibody, and the pattern of immunostaining was compared with those of total βPix isoforms and various synaptic markers. pY598 immunostaining in dendrites overlapped with the total βPix immunostaining by anti-βPix SH3 antibody only at the spine-like structures (Fig. S3, top row). In addition, the anti-pY598 immunostaining colocalized well with the postsynaptic marker, PSD95, and the AMPA receptor subunit, GluR2, and with a somewhat lesser degree with the presynaptic marker, synaptophysin (Syn; also known as Syp) (Fig. S3, bottom three rows), showing that Y598 phosphorylation of βPix-b occurs spontaneously in neurons and specifically localizes to spines.

Glutamate induces Y598 phosphorylation of βPix-b in an Src-dependent manner

Our data demonstrate that βPix-b is phosphorylated by Src family kinases upon glutamate stimulation (Fig. 4B) and Y598 of βPix-b is the major phosphorylated residue (Fig. 5). Therefore, we examined whether Y598 phosphorylation can be modulated by glutamate-mediated synaptic activity. Glutamate treatment was administered to 19 DIV rat cortical neurons, and Y598 phosphorylation of βPix-b was examined with anti-pY598 and anti-4G10 antibody. Glutamate treatment increased Y598 phosphorylation of βPix-b, whereas PP2 pretreatment blocked the phosphorylation (Fig. 6A), suggesting that glutamate induces Y598 phosphorylation of βPix-b in an Src-dependent manner. This was also confirmed by immunostaining in 25 DIV hippocampal neurons using anti-pY598 antibody (Fig. 6B). Glutamate-induced pY598 βPix-b that colocalized with GluR2 (Fig. 6B, top row) disappeared with PP2 pretreatment (Fig. 6B, middle row), and reappeared with glutamate treatment after washing out PP2 (Fig. 6B, bottom row). Taken together, these results suggest that Src-mediated Y598 phosphorylation of βPix-b can be modulated by glutamate-induced neuronal activity.

Fig. 6.

Glutamate induces Src-mediated Y598 phosphorylation of βPix-b at dendritic spines. (A) Glutamate stimulation induces Src-dependent Y598 phosphorylation of βPix-b. 50 µM glutamate treatment was administered to 19 DIV cortical neurons for 10 min. 10 µM PP2 pretreatment was administered for 15 min before adding glutamate, where indicated. Cell lysates were immunoprecipitated with anti-insert antibodies and subjected to immunoblotting (IB) with anti-4G10 and anti-pY598 antibodies. (B) Glutamate induces Y598 phosphorylation of βPix-b at dendritic spines. 25 DIV hippocampal neurons were treated with 50 µM glutamate for 10 min with or without PP2 pretreatment, fixed and immunostained with anti-pY598 and anti-GluR2 antibodies. The top row shows control neurons without PP2 pretreatment; the middle row shows neurons pretreated with 10 µM PP2 for 15 min before glutamate addition; the bottom row shows neurons treated with 50 µM glutamate (Glu) for 10 min after washing out the pretreated 10 µM PP2. Scale bar: 10 µm.

Fig. 6.

Glutamate induces Src-mediated Y598 phosphorylation of βPix-b at dendritic spines. (A) Glutamate stimulation induces Src-dependent Y598 phosphorylation of βPix-b. 50 µM glutamate treatment was administered to 19 DIV cortical neurons for 10 min. 10 µM PP2 pretreatment was administered for 15 min before adding glutamate, where indicated. Cell lysates were immunoprecipitated with anti-insert antibodies and subjected to immunoblotting (IB) with anti-4G10 and anti-pY598 antibodies. (B) Glutamate induces Y598 phosphorylation of βPix-b at dendritic spines. 25 DIV hippocampal neurons were treated with 50 µM glutamate for 10 min with or without PP2 pretreatment, fixed and immunostained with anti-pY598 and anti-GluR2 antibodies. The top row shows control neurons without PP2 pretreatment; the middle row shows neurons pretreated with 10 µM PP2 for 15 min before glutamate addition; the bottom row shows neurons treated with 50 µM glutamate (Glu) for 10 min after washing out the pretreated 10 µM PP2. Scale bar: 10 µm.

From these data, we also found that although βPix-d has Y598 residue in the insert region that is identical to βPix-b, its Y598 phosphorylation was not enhanced upon glutamate stimulation (Fig. 6A), consistent with the results shown in Fig. 4. Our data suggest that Y598 phosphorylation is a specific modification on βPix-b among the βPix isoforms in neurons.

Y598 phosphorylation of βPix-b enhances its GEF activity for Rac1 that is critical for spine development and synapse formation

Activated Rac1 GTPase regulates actin cytoskeleton that is essential for development, maintenance and plasticity of dendritic spines in neurons (Tashiro and Yuste, 2004). Hence, we investigated whether Y598 phosphorylation of βPix-b affects the regulation of its GEF function for Rac1 activation. First, active GEF assay with GST-Rac1G15A in Cos7 cells showed that GEF activity of wild-type GFP-βPix-b was markedly elevated when constitutively active Src was co-expressed (2.7±0.36 vs 7.3±2.63; mean±s.e.m.) (Fig. 7A). However, GEF activity of GFP-βPix-bY598F was not significantly changed by co-expression of active Src (not significant; 1.37±0.63). Rac1 activation assay with GST-PBD also showed similar results. Rac1-GTP levels in cells expressing GFP-βPix-bY598F with active Src were modestly increased compared with GFP control (3.42±1.48), whereas expression of GFP-βPix-b with active Src drastically increased Rac1 activation (15.1±3.13) (Fig. 7B). Interestingly, βPix-bY598F exhibited some interesting differences in two different assays: active GEF assay and Rac1 activation assay. As mentioned above, the active GEF level of βPix-bY598F was similar to that of control, while its Rac1-GTP level remained higher than that of control (3.42±1.48). We suppose that the discrepancy is due to active Src co-expression affecting broad Src-dependent pathways inducing Rac1 signaling in cells.

Fig. 7.

Y598 phosphorylation enhances Rac1 GEF activity of βPix-b. (A,B) Src-mediated Y598 phosphorylation enhances Rac1 GEF activity of βPix-b. (A) Left: GFP (control), GFP-βPix-b or GFP-βPix-bY598F were co-expressed with or without active Src in Cos7 cells. After 17–20 h, cell lysates were subjected to active GEF assay (PD) and immunoblotted (IB) with anti-SH3 antibodies. Right: ratio of active βPix-b or βPix-bY598F to input actin (loading control) levels were normalized to GFP expression level. (B) Left: cell lysates were prepared as in A. After 17–20 h, cell lysates were subjected to Rac1 activation assay (PD) and immunoblotted with anti-Rac1 antibodies. Right: ratio of active Rac1 to input Rac1 levels were normalized to GFP expression level. (C) Y598 phosphorylation of βPix-b induces Rac1 activation in neurons. Dissociated hippocampal neurons from P0 neuron-specific βPix isoform KO mice were transfected with Raichu-Rac1 (Rac1 activation biosensor), together with Flag (control), Flag-βPix-b or Flag-βPix-bY598F at 7 DIV and imaged at 19 DIV. Left: representative FRET images of Rac1 activation are shown in the intensity-modulated display mode. In the intensity-modulated mode, eight colors from red to blue were used to represent the FRET:CFP ratio. Scale bar: 10 µm. Right: quantification of Raichu-Rac1 FRET signal in dendrites (13–17 neurons per condition). The values were normalized to that of cells expressing Raichu-Rac1 only. *P<0.05, **P<0.001; ns, not significant; ANOVA followed by Bonferroni-corrected t-test. Error bars indicate s.e.m. CB, Coomassie Blue staining.

Fig. 7.

Y598 phosphorylation enhances Rac1 GEF activity of βPix-b. (A,B) Src-mediated Y598 phosphorylation enhances Rac1 GEF activity of βPix-b. (A) Left: GFP (control), GFP-βPix-b or GFP-βPix-bY598F were co-expressed with or without active Src in Cos7 cells. After 17–20 h, cell lysates were subjected to active GEF assay (PD) and immunoblotted (IB) with anti-SH3 antibodies. Right: ratio of active βPix-b or βPix-bY598F to input actin (loading control) levels were normalized to GFP expression level. (B) Left: cell lysates were prepared as in A. After 17–20 h, cell lysates were subjected to Rac1 activation assay (PD) and immunoblotted with anti-Rac1 antibodies. Right: ratio of active Rac1 to input Rac1 levels were normalized to GFP expression level. (C) Y598 phosphorylation of βPix-b induces Rac1 activation in neurons. Dissociated hippocampal neurons from P0 neuron-specific βPix isoform KO mice were transfected with Raichu-Rac1 (Rac1 activation biosensor), together with Flag (control), Flag-βPix-b or Flag-βPix-bY598F at 7 DIV and imaged at 19 DIV. Left: representative FRET images of Rac1 activation are shown in the intensity-modulated display mode. In the intensity-modulated mode, eight colors from red to blue were used to represent the FRET:CFP ratio. Scale bar: 10 µm. Right: quantification of Raichu-Rac1 FRET signal in dendrites (13–17 neurons per condition). The values were normalized to that of cells expressing Raichu-Rac1 only. *P<0.05, **P<0.001; ns, not significant; ANOVA followed by Bonferroni-corrected t-test. Error bars indicate s.e.m. CB, Coomassie Blue staining.

To confirm the effect of Y598 phosphorylation on βPix-b-mediated Rac1 activation in neurons, we performed a fluorescence resonance energy transfer (FRET)-based assay in cultured hippocampal neurons using a Rac1 activation biosensor, Raichu-Rac1 construct (Itoh et al., 2002). We transfected Flag-βPix-b or Flag-βPix-bY598F mutant with Raichu-Rac1 probe to 7 DIV hippocampal neurons from neuron-specific βPix isoform knockout (KO) mice and analyzed the FRET signals at 19 DIV. These KO mice lack expression of neuron-specific βPix isoforms, βPix-b and -d. We found that βPix-b overexpression increased Rac1 activation in dendrites and protrusions by 49% (Fig. 7C), consistent with the effect of βPix-b expression in a previous FRET study (Llano et al., 2015). Importantly, this effect was fully abolished by the Y598F mutation. These results suggest that Y598 phosphorylation regulates activity of βPix-b as a Rac1 GEF in both Cos7 cells and neurons.

Finally, we tested whether Y598 phosphorylation of βPix-b is critical for regulation of spine morphogenesis and synapse plasticity. Hippocampal neurons from neuron-specific βPix isoform KO mice were cultured and transfected at 7 DIV with GFP, GFP-βPix-b, GFP-βPix-bY598E or GFP-βPix-bY598F constructs, as indicated in Fig. 8A. βPix-bY598E, a phospho-mimicking mutant, was characterized in Neuro2a cells in which the Y598 phosphorylation occurs spontaneously due to the Src being basally active. We found that βPix-bY598E was recognized by anti-pY598 antibody and exhibited higher Rac1 GEF activity than βPix-bY598F. Notably, no further enhancement of the GEF activity was observed by βPix-bY598E expression compared with the expression of wild-type βPix-b (Fig. S4). Synapses were stained using anti-synaptophysin antibody and analyzed at 19 DIV. Expression of wild-type GFP-βPix-b and GFP-βPix-bY598E strongly increased the density of dendritic spine, whereas neurons expressing GFP-βPix-bY598F, non-phosphorylatable mutant, did not significantly change the spine density (Fig. 8A,C). Filopodia density was not significantly changed by any of the βPix-b constructs. In addition, GFP-βPix-b and GFP-βPix-bY598E increased the formation of synapses, but GFP-βPix-bY598F did not (Fig. 8D), suggesting that phosphorylation of βPix-b at Y598 regulates spine development and synapse formation. Interestingly, wild-type GFP-βPix-b, GFP-βPix-bY598E and GFP-βPix-bY598F all localized preferentially at dendritic spines without noticeable differences (Fig. 8B), indicating that Y598 phosphorylation of βPix-b does not affect the localization of βPix-b. In addition, we observed that GFP-βPix-bY598E did not significantly enhance spine density and synapse formation when compared with wild-type βPix-b (not significant). This is consistent with the results shown in Fig. S4, suggesting that wild-type βPix-b may already be fully activated in neuronal cells and neurons in an active Src-dependent manner. Collectively, these results demonstrate that Src-mediated Y598 phosphorylation of βPix-b regulates its GEF function for Rac1 activation that is critical for spine development and synapse formation.

Fig. 8.

Y598 phosphorylation of βPix-b is critical for spine development and synapse formation. (A,B) Y598 phosphorylation does not affect the localization of βPix-b to spines. Hippocampal neurons cultured from P0 neuron-specific βPix isoform KO mice were transfected at 7 DIV with Dsred and one of the following constructs: GFP (control), GFP-βPix-b, GFP-βPix-bY598E or GFP-βPix-bY598F. Synapses were labeled by anti-synaptophysin immunostaining. Arrowheads indicate spines. Scale bars: 5 µm. The bar graphs illustrate the spine:shaft fluorescence intensity ratio in randomly selected dendrites (112–140 dendrite segments per groups from 31–36 neurons). (C,D) Y598 phosphorylation of βPix-b regulates spine and synapse formation. Quantification of spine density and synapse formation in neurons transfected with the indicated constructs. C illustrates the number of dendritic spines per 10 µm. D illustrates the synapse:spine percentage in neurons. Dendritic spines were randomly selected from 112–140 dendrite segments (317–452 spines per groups). **P<0.001, ***P<0.00001; ns, not significant; ANOVA followed by Bonferroni-corrected t-test. Error bars indicate s.e.m. (E) A prospective model for a novel role of βPix-b in activity-dependent spinogenesis and synaptogenesis. βPix-b preferentially localizes at the spine and promotes spine and synapse formation. Once glutamate-dependent neuronal activity activates Src family kinases in an AMPA receptor- and NMDA receptor-mediated manner, Y598 of βPix-b is specifically phosphorylated. Y598 phosphorylation of βPix-b increases its GEF activity for Rac1 and enhances spine development and synapse formation.

Fig. 8.

Y598 phosphorylation of βPix-b is critical for spine development and synapse formation. (A,B) Y598 phosphorylation does not affect the localization of βPix-b to spines. Hippocampal neurons cultured from P0 neuron-specific βPix isoform KO mice were transfected at 7 DIV with Dsred and one of the following constructs: GFP (control), GFP-βPix-b, GFP-βPix-bY598E or GFP-βPix-bY598F. Synapses were labeled by anti-synaptophysin immunostaining. Arrowheads indicate spines. Scale bars: 5 µm. The bar graphs illustrate the spine:shaft fluorescence intensity ratio in randomly selected dendrites (112–140 dendrite segments per groups from 31–36 neurons). (C,D) Y598 phosphorylation of βPix-b regulates spine and synapse formation. Quantification of spine density and synapse formation in neurons transfected with the indicated constructs. C illustrates the number of dendritic spines per 10 µm. D illustrates the synapse:spine percentage in neurons. Dendritic spines were randomly selected from 112–140 dendrite segments (317–452 spines per groups). **P<0.001, ***P<0.00001; ns, not significant; ANOVA followed by Bonferroni-corrected t-test. Error bars indicate s.e.m. (E) A prospective model for a novel role of βPix-b in activity-dependent spinogenesis and synaptogenesis. βPix-b preferentially localizes at the spine and promotes spine and synapse formation. Once glutamate-dependent neuronal activity activates Src family kinases in an AMPA receptor- and NMDA receptor-mediated manner, Y598 of βPix-b is specifically phosphorylated. Y598 phosphorylation of βPix-b increases its GEF activity for Rac1 and enhances spine development and synapse formation.

Here, we demonstrate that Src-mediated Y598 phosphorylation of βPix-b regulates spine and synapse development. This conclusion is based on our findings that (1) βPix-b is mainly expressed in neurons; (2) βPix-b overexpression enhances spine development and synapse formation with its preferential localization at spines; (3) βPix-b exhibits higher basal GEF activity than ubiquitous βPix-a and its GEF activity is further elevated upon glutamate-stimulation in neurons; (4) glutamate induces Src-mediated tyrosine phosphorylation of βPix-b; (5) tyrosine 598 (Y598) in the insert region is the major phosphorylation site in βPix-b upon glutamate stimulation; and (6) inhibiting Y598 phosphorylation (βPix-bY598F) drastically decreases the GEF activity of βPix-b for Rac1 signaling and further disrupts spine development and synapse formation. Our data provide evidence that βPix-b acts as a positive regulator to enhance spinogenesis and synaptogenesis through its GEF activity regulated by Y598 phosphorylation upon neuronal activation. We previously reported that several βPix isoforms are present in the brain. Among those, βPix-a is ubiquitously expressed, whereas βPix-b and -d are highly enriched in brain tissues, suggesting that βPix-b and -d are involved in neuronal processes (Kim et al., 2000; Kim and Park, 2001; Oh et al., 1997). In our present data, we confirmed that βPix-b and -d are mainly expressed in neurons (Fig. 1). βPix functions have been implicated in spine and synapse formation (Saneyoshi et al., 2008). However, a previous study showed that βPix-a overexpression in cultured neurons inhibited the development of mature spine and synapse, but enhanced the formation of dendritic filopodia (Zhang et al., 2003), suggesting that βPix-a overexpression induces an antimorphic phenotype of βPix, possibly due to mislocalization of βPix-a in neurons. Because the molecular structure of βPix-b is identical to that of βPix-a, except for the presence of the insert region in βPix-b, we primarily compared the roles of these two isoforms in spine development in this study. Interestingly, overexpression of βPix-b, but not βPix-a, enhances spine development and synapse formation in cultured neurons (Fig. 2). Our data support that the functional difference between βPix-a and -b is due to their localization (Fig. 2) and GEF activity (Fig. 3). First, we found that overexpressed βPix-b preferentially localizes in the spines and, at the same time, spine density and synapse formation are enhanced. In the case of βPix-a overexpression, however, spine density and synapse formation are inhibited, consistent with Zhang et al. (2003), although its localization still slightly prefers spines to shafts. Our data show that overexpression of βPix-b elevates its localization at spine heads and results in promotion of spine and synapse formation, suggesting that βPix-b is important for the regulation of molecular effectors in spine development.

Next, our results show that βPix-b exhibits a higher GEF activity than βPix-a toward Rac1, which is known as a downstream target of βPix (Klooster et al., 2006) (Fig. 3), suggesting that βPix-b is a more efficient GEF protein than βPix-a. In addition, our GEF activity assay of βPix in glutamate-stimulated neurons shows that only βPix-b responds to the stimulation with enhanced GEF activity. Taken together, these results indicate that differences in localization and GEF activity between βPix-a and -b lead to their distinct roles in the regulation of spine morphology and synapse plasticity in neurons.

What could be a possible mechanism regulating the GEF activity of βPix-b? Previous studies reported that βPix-a has numerous phosphorylated residues that also exist in βPix-b, and its phosphorylation is known as a critical modification to activate GEF function (Feng et al., 2006, 2010; Mayhew et al., 2007; Saneyoshi et al., 2008; Shin et al., 2002; Shin et al., 2004). We predicted that the phosphorylation of βPix-b is correlated with its GEF activity. To test this hypothesis, we examined the phosphorylation level of βPix-b in glutamate-stimulated neurons (Fig. 4). As predicted, our data suggest that tyrosine phosphorylation of βPix-b regulates its GEF function.

To understand the upstream mechanisms for the phosphorylation of βPix-b, we investigated which kinases cause phosphorylation of βPix-b (Fig. 4). A glutamate stimulation-dependent increase in the tyrosine phosphorylation levels of βPix-b was inhibited when PP2 was pretreated, showing that Src family kinases mediate the phosphorylation of βPix-b. Further, we investigated which glutamate receptor is involved in the activation of Src (Fig. 4). Our results show that the increase in the tyrosine phosphorylation of βPix-b was abrogated by both CNQX and D-AP5, which are antagonists for the AMPA receptor and NMDA receptor, respectively. These data indicate that Src activation induces the phosphorylation of βPix-b by both AMPA receptor- and NMDA receptor-mediated pathways.

Because βPix-b has the insert region, which is deficient in βPix-a, we hypothesized that an unknown tyrosine phosphorylation site(s) in the insert region of βPix-b is critical for regulation of its GEF activity. We found that Y598 in the insert region of βPix-b is the major phosphorylation site by Src (Figs 5 and 6; Fig. S3). Among several prospective tyrosine phosphorylation sites, Y598 in the insert region showed the highest probability for phosphorylation. To confirm the site of phosphorylation, βPix-bY598F mutant for blocking Y598 phosphorylation and anti-pY598 antibody for targeting Y598 phosphorylation were generated. We observed that the Y598F mutation blocks tyrosine phosphorylation of βPix-b even when Src is constitutively activated, indicating that Y598 of βPix-b in the insert region is the major phosphorylation site (Fig. 5). In neurons, we also confirmed that Y598 phosphorylation can be modulated by glutamate-induced Src activation (Fig. 6). The anti-pY598 antibody reveals that Y598-phosphorylated βPix-b is highly enriched in spines and that Y598 phosphorylation of βPix-b occurs spontaneously in neurons (Fig. S3). Our other results also support that Y598-phosphorylated βPix-b localizes in spines. First, βPix-b is preferentially localized in spines (Fig. 2). Second, βPix-b is phosphorylated by Src family kinases in a stimulation-dependent manner (Figs 46). Third, wild-type βPix-b, the phospho-mimicking mutant βPix-bY598E and the phospho-blocking mutant βPix-bY598F do not display different spine localization indices (Fig. 8B), indicating that Y598 phosphorylation of βPix-b takes place in spines, but the phosphorylation itself does not affect the localization of βPix-b. Furthermore, we found that βPix-b is the only Y598-phosphorylated βPix isoform detected in neurons (Fig. 6A). The βPix-d isoform, containing the identical insert region, does not show a detectable Y598 phosphorylation upon glutamate stimulation. This result is consistent with the finding that only βPix-b is tyrosine phosphorylated upon glutamate stimulation (Fig. 4), suggesting that the Y598 phosphorylation is a specific modification for βPix-b. Why is Y598 of βPix-d not phosphorylated by Src in glutamate-stimulated neurons? We suppose that lack of C-terminal domain in βPix-d interrupts its localization at spines, resulting in inefficient Y598 phosphorylation. βPix-d has a unique 11-amino-acid stretch instead of the C-terminal domain found in other βPix isoforms that is necessary for proper intracellular localization, as shown in Fig. S1. Indeed, βPix-d expression shows diffused localization in both spines and dendrites (data not shown), similar to the localization of βPix-bΔCC, which is structurally identical to βPix-d except for the 11 amino acids unique to βPix-d (Fig. S1). Unlike βPix-b, βPix-d probably cannot properly localize to spines due to the lack of C-terminal domain, leading to spatial sequestration of βPix-d from Src that localized at the membrane of spines.

Finally, we asked what the role of Y598 phosphorylation of βPix-b in neuronal development was (Figs 7 and 8). We observed that introducing a non-phosphorylatable mutation to βPix-b Y598 residue (Y598F) abrogates the GEF activity of βPix-b for Rac1 in Rac1 activation assays in Cos 7 cells and in neurons. Furthermore, βPix-bY598F is functionally inactive and its overexpression fails to recover spine and synapse formation in hippocampal neurons from neuron-specific βPix isoform KO mice, unlike the expression of wild-type βPix-b. These observations demonstrate that Y598 phosphorylation of βPix-b regulates GEF function for Rac1 activation, and this regulation is critical for spine development and synapse formation. Interestingly, we also show that there are no differences in localization between wild-type βPix-b and its mutants, the phospho-mimicking βPix-bY598E and the phospho-blocking βPix-bY598F, indicating that Y598 phosphorylation of βPix-b regulates its GEF activity, but not its localization in spines.

Based on these data, we suggest that βPix-b expression and its Y598 phosphorylation in dendritic spines regulate the proper morphology and development of spines in both basal state and stimulated condition. A prospective model of this regulation is schematically described in Fig. 8E. Although our study reveals the important role of βPix-b in spinogenesis, a few questions remain unanswered.

We have not yet identified the reason why βPix-b is more localized at spine heads than ubiquitous βPix-a (Fig. 2). We suppose that there are probably binding partners specifically interacting with the βPix-b isoform for spine localization. Recent studies reported that KCC2, a neuron-specific KCl co-transporter that localizes at the dendritic spines, plays a critical role in the regulation of spine actin dynamics via interaction with βPix-b (Llano et al., 2015). It will be interesting to examine whether KCC2 specifically interacts with βPix-b among βPix isoforms and whether KCC2 recruits βPix-b to proper localization at spines.

In neurons, Rac1 and Cdc42 are important for promoting the formation, development and maintenance of spines and synapses (Newey et al., 2005; Tolias et al., 2011). Because βPix is known as the GEF protein for Cdc42 as well as Rac1 (Bagrodia et al., 1998; Manser et al., 1998), we also examined whether βPix-b has the potential to promote Cdc42 activity. Cos7 cells expressing GFP-βPix-b showed higher Cdc42 activity than those expressing GFP-βPix-a (2.76-fold increase compared with control) (Fig. S5). This result is similar to the results of the Rac1 activation assay shown in Fig. 3A, suggesting that Cdc42 activity can also be regulated by βPix-b. Whether Y598 phosphorylation of βPix-b modulates spine development through the Cdc42-activating ability is to be investigated in future studies.

Next, because PP2 is a broad-spectrum inhibitor of Src family kinases, we cannot exclude the possible involvement of other Src family kinases in βPix-b phosphorylation. Among the Src family kinases, Src is expressed ubiquitously, but shows a high expression level especially in neurons (Thomas and Brugge, 1997). Fyn and Yes (also known as Yes1) are also ubiquitously expressed, and Lyn and Lck are expressed in brain (Thomas and Brugge, 1997). Thus, these Src family kinases can potentially be involved in Y598 phosphorylation of βPix-b. In addition, it is also possible that the Src family kinases might indirectly regulate the phosphorylation of βPix-b by activating other kinases.

The functional roles and the regulatory mechanisms of βPix-d are also to be discovered in future studies. Like βPix-b, βPix-d is highly expressed in neurons and contains the insert domain (Fig. 1). However, βPix-d has a unique stretch of 11 amino acids replacing the C-terminal region of other βPix isoforms and is not phosphorylated by Src, suggesting that βPix-d is regulated via different mechanisms from those regulating βPix-b.

In conclusion, we identify βPix-b as a major neuronal βPix isoform primarily regulating the formation of dendritic spines and synapses through its GEF activity for Rac1. The regulatory pathway requires Src-mediated Y598 phosphorylation of βPix-b that is necessary for its GEF activity. We show that modulation of the GEF activity by Y598 phosphorylation is critical for regulating spine development and synapse formation. These findings provide novel understanding of the molecular mechanisms underlying activity-dependent spine morphogenesis and synapse formation.

Antibodies and reagents

Rabbit polyclonal antibodies for the insert and the SH3 domains of βPix were prepared as described previously (Kim et al., 2000; Oh et al., 1997). Antibodies specifically recognizing phosphorylated Y598 of βPix-b were generated by Labfrontier (Republic of Korea) by immunizing rabbits with phosphorylated peptides EDSE(pY598)DSIW (where pY is phosphorylated Tyr) purchased from BioSynthesis (USA). Specific phosphotyrosine antibodies were purified from the rabbit serum with immobilized peptides, first with non-phosphorylated peptides for negative selection and then with phosphorylated peptides for positive selection. The following antibodies were purchased: monoclonal anti-phosphotyrosine 4G10 (#05-321) from Upstate; monoclonal anti-Rac1 (#610650) from BD Transduction Laboratories; monoclonal anti-Cdc42 (#C70020) from Transduction Laboratories; polyclonal anti-phospho-Src(Tyr416) (#2101) from Cell Signaling Technology; monoclonal anti-Flag M2 (#F7425), monoclonal anti-β-tubulin (#T4026) and monoclonal anti-actin (#A4700) from Sigma-Aldrich; polyclonal anti-synaptophysin (#MAB5258), polyclonal anti-PSD-95 (#MAB1596) and polyclonal anti-GluR2 (#AB1506) from Chemicon. Reagents purchased were PP2 (#529573), a specific Src family kinase inhibitor, and PP3 (#529574), a negative control for PP2, from Calbiochem; and CNQX (#1045), AMPA receptor antagonist, and D-AP5 (#0106), NMDA receptor antagonist, from Tocris.

DNA constructs

The coding region of βPix-b was subcloned into pFlag-CMV2 (Eastman Kodak Co.) and pEGFP-N1 (Clontech) by PCR. To generate βPix-b mutant constructs, site-directed mutagenesis was performed using a QuikChange site-directed mutagenesis kit (Stratagene), following the manufacturer's instruction. The following mutagenic primers were used: pEGFPN1-βPix-b(Y598E), 5′-TCGGAAGACTCTGAGGAAGACAGTATATGGACA-3′ and 5′-TGTCCATATACTGTCTTCCTCAGAGTCTTCCGA-3′; pEGFPN1-βPix-b(Y598F), 5′-TCGGAAGACTCTGAGTTTGACAGTATATGGACA-3′ and 5′-TGTCCATATACTGTCAAACTCAGAGTCTTCCGA-3′; pEGFPN1-βPix-b(ΔCC), 5′-GACAAGCTTCGATGACTGATAACAAC-3′ and 5′-CGCGTCGACGTACTCTATCACTGTCTG-3′. These primers were also used for mutagenesis of pFlagCMV2-βPix-b(Y598F). The mutants were verified by automated DNA sequencing. We used a GST-PBD construct described previously (Park et al., 2012). GST-Rac1G15A construct was a kind gift from J. G. Hanley (University of Bristol, Bristol, UK). Raichu-Rac1 probe for FRET was kindly provided by M. Matsuda (Kyoto University, Kyoto, Japan).

Mice

Neuron-specific βPix isoform KO mice were generated by replacing exon 19 with a neomycin-resistance cassette flanked by LoxP sites to specifically delete the insert domain-coding region. In order to allow expression of the ubiquitous βPix-a isoform, the mouse line described above was crossed to Sox2-Cre mice to excise the neomycin-resistance cassette in the germ line (Hayashi et al., 2002; Kos, 2004). In these mice, expression of βPix-b and -d containing the insert region was removed, but βPix-a remained expressed in brain (data not shown). Their genotypes were confirmed using PCR techniques. All mice were maintained and used according to the guidelines of the Seoul National University Institutional Animal Care and Use Committees.

Cell line culture and transfection

Cos7 cells were cultured in Dulbecco's modified Eagle medium supplemented with 10% fetal bovine serum (c-FBS, Welgene) and 1% penicillin-streptomycin (Gibco) at 37°C with 5% CO2. Neuro2a cells were cultured in minimum essential medium (MEM) supplemented with 10% c-FBS at 37°C with 5% CO2. Cos7 cells and Neuro2a cells were transfected using Lipofectamine Plus (Invitrogen) according to the manufacturer's instructions.

Primary neuronal culture and transfection

Dissociated hippocampal (for staining experiments) and cortical (for biochemical experiments) neurons were prepared from embryonic day 18 (E18) Sprague-Dawley rat embryos and postnatal day 0 (P0) pups of neuron-specific βPix isoform KO mice of either sex as described previously (Beaudoin et al., 2012; Park et al., 2012). We used protocols described in Park et al. (2012) and Beaudoin et al. (2012) for neuron cultures, with minor modifications. In brief, dissociated hippocampal and cortex tissues were treated with papain (20 µg/ml, Worthington) and DNase (10 U/µl, Sigma-Aldrich) for 25 min at 37°C. The tissues were dissociated by trituration with a glass Pasteur pipette and then plated in 60-mm or 100-mm dishes coated with poly-D-lysine (10 ng/ml, Sigma-Aldrich). Cultures were grown in Neurobasal medium (Invitrogen) supplemented with B27 (Invitrogen) and 0.5 mM L-glutamine (Welgene). The cultures were maintained by discarding and replacing half of the original medium with fresh medium every 4–7 days. Hippocampal neurons were transiently transfected at 7 DIV with the calcium phosphate method using a CalPhos Transfection Kit (Calbiochem). Glial cells were cultured from 18-day embryos as described previously (Goslin et al., 1998). Dissociated cells were plated in 60-mm or 100-mm diameter dishes coated with poly-D-lysine (1 mg/ml). Cultures were grown in MEM (Invitrogen) supplemented with 5% FBS (Invitrogen), penicillin-streptomycin, 0.4% glucose and 0.5 mM L-glutamine.

Active GEF assay, Rac1 activation assay and Cdc42 activation assay

Active GEF assay and Rac1 activation assay were performed with GST fusion proteins expressed in Escherichia coli (BL21) cells and purified by glutathione sepharose beads (GE Healthcare). The fusion proteins bound to the beads were then eluted by glutathione elution buffer (20 mM glutathione, 100 mM Tris-HCl, pH 8.0, 120 mM NaCl) and quantified using Bradford quantification method (Bio-Rad). Activity of GEF proteins was assessed by measuring the amount of GEF protein bound to GST-Rac1G15A (nucleotide-free Rac1 GTPase) proteins, which form a high-affinity complex with active Rac1 GEFs, as described previously (Blanco-Suárez et al., 2014). Active Rac1 levels were measured using GST-PBD (p21-binding domain of PAK) proteins, which bind to GTP-bound Rac1 (Shin et al., 2004). Briefly, glutathione sepharose beads and 5 μg GST fusion proteins were incubated in assay buffer (50 mM HEPES, pH7.4, 1% NP-40, 150 mM NaCl, 15 mM NaF, 1 mM sodium vanadate, 1 µg/ml leupeptin, 1 µg/ml aprotinin and 1 µg/ml pepstatin) for 1 h at 4°C. Cells transfected with the indicated constructs for 17–20 h were washed with PBS and lysed with assay buffer for 20 min at 4°C. Cell lysates were centrifuged at 22,500 g for 30 min at 4°C, and then the supernatant was added to the pre-incubated mixture of GST fusion proteins and beads and incubated for 1 h at 4°C. The beads were washed three times with assay buffer. These assays were also performed with rat neuronal cultures. Rat cortical neurons at 19 DIV were incubated in Tyrode solution (119 mM NaCl, 2.5 mM KCl, 2 mM CaCl2, 2 mM MgCl2, 25 mM HEPES, pH 7.4, and 30 mM glucose) for 30 min before treatment with 50 μM glutamate and 5 μM glycine for 5 min. The bead-bound proteins were denatured in Laemmli sample buffer, separated by SDS-PAGE and immunoblotted using anti-SH3 (βPix), anti-Rac1 and anti-Cdc42 antibodies. The GST fusion proteins were visualized by Coomassie Blue staining.

Immunoprecipitation

Protein G sepharose beads (GE Healthcare) and appropriate antibodies were incubated for 1 h at 4°C. Transiently transfected cells were washed with PBS and lysed in immunoprecipitation buffer (50 mM HEPES, pH7.4, 1% NP-40, 300 mM NaCl, 15 mM NaF, 1 mM sodium vanadate, 1 µg/ml leupeptin, 1 µg/ml aprotinin and 1 µg/ml pepstatin). Lysates were clarified by centrifugation at 22,500 g for 30 min. Protein concentrations were determined using Bradford reagent. Clarified lysates were added to samples and incubated for 1 h at 4°C. The immunoprecipitates were washed five times with immunoprecipitation buffer to remove unbound proteins. The same protocol was performed with cultured rat neurons. Rat cortical neurons at 19 DIV were incubated in Tyrode solution (119 mM NaCl, 2.5 mM KCl, 2 mM CaCl2, 2 mM MgCl2 and 30 mM glucose in 25 mM HEPES, pH 7.4) for 30 min before stimulation with 50 µM glutamate and 5 µM glycine for 5 or 10 min. The following reagents were administered as pretreatment for 15 min before adding glutamate: 10 µM PP2, 10 µM PP3, 30 µM CNQX or 100 µM D-AP5. Proteins bound to antibodies were loaded in Laemmli sample buffer, separated by SDS-PAGE and immunoblotted.

Immunoblotting

For examination of βPix isoform expression levels, adult rat brain and organ tissues were homogenized in 8 ml homogenization buffer (50 mM Tris-HCl, pH 8.5, 1 mM EDTA, 1 mM EGTA, 150 mM NaCl, 2 mM sodium vanadate, 15 mM NaF, 1 µg/ml leupeptin and 1 µg/ml pepstatin) with a Polytron homogenizer and incubated for 1 h at 4°C after addition of 1% Triton X-100. The homogenates were centrifuged at 22,500 g for 30 min at 4°C and the supernatants used for immunoblotting. Lysates from cultured cell lines or primary neurons were prepared by washing with PBS and lysing in SDS sample buffer (62.5 mM Tris-HCl, pH 6.7, 10% glycerol and 2% SDS) or in assay buffer. Protein concentrations were determined with BCA reagent (Pierce) or Bradford reagent (Bio-Rad) by using bovine serum albumin solutions as standards. Samples were loaded in 1× Laemmli sample buffer (5× sample buffer: 60 mM Tris-HCl, pH 6.8, 30% glycerol, 2% SDS, 14.4 mM β-mercaptoethanol and 0.2% Bromophenol Blue) and equal amounts of protein were resolved by SDS-PAGE and transferred to a PVDF or nitrocellulose membrane. Blots were blocked with 5% bovine serum albumin (BSA) in 1× TBS-T (0.1% Tween 20 in 1× TBS-T) for 30 min. The blots were incubated with primary antibodies for 1 h and washed three times for 10 min with TBS-T. Then, the blots were incubated with horseradish peroxidase-conjugated secondary antibodies (Jackson ImmunoResearch Laboratories) for 45 min, washed three times with TBS-T and analyzed by enhanced chemiluminescence.

Immunohistochemistry

We performed fixation and sectioning according to the procedures of Schaeren-Wiemers and Gerfin-Moser (1993). Frozen sections of adult rat brains on slides were sectioned into 40-μm slices, fixed with 4% paraformaldehyde for 15 min and washed with PBS. All procedures were performed at room temperature (RT). Specimens were incubated with 0.3% H2O2/methanol solution for quenching of endogenous peroxidase activity for 20 min and washed three times with PBS. Non-specific signals were blocked by incubation with 1.5% horse serum/0.3% Triton X-100 in PBS. Then, specimens were incubated with primary antibodies, biotinylated secondary antibodies and Ultra-Sensitive ABC Peroxidase reagent (Pierce). Signals were visualized by diaminobenzidine staining (Pierce).

Immunocytochemistry and microscope image acquisition

Hippocampal neurons seeded on 18-mm coverslips were fixed at 19 or 25 DIV in PBS containing 3.7% paraformaldehyde for 10 min at RT. Where indicated, cells were treated with 10 µM PP2 for 15 min and then 50 µM glutamate for 10 min in Tyrode solution for 30 min before fixation. Fixed neurons were permeabilized with 0.5% Triton X-100 in PBS for 10 min and incubated in blocking solution (3% BSA, 0.5% gelatin, 0.1% Triton X-100 in PBS) for 30 min. Blocked coverslips were then incubated with primary antibodies diluted in blocking solution for 1 h at RT. After washing with PBS containing 0.1% Triton X-100, the neurons were incubated with secondary antibodies for 45 min at RT. Coverslips were mounted with Vectashield mounting medium (Vector Laboratories). Fluorescent images were acquired with a Zeiss Axiovert fluorescence microscope equipped with a 100×, 1.4 NA plan-apochromat objective lens and a Zeiss Axiocam HRm CCD camera. Quantification of various parameters of spine morphology was performed with ImageJ (NIH) software.

Analysis of neuronal morphology

For quantification of spine:shaft fluorescence intensity ratio, spine density and synapse formation percentage (%), neuron images were processed equally by Adobe Photoshop. All quantifications were repeated, with at least three independent experiments, and 317–1317 spines per group in 116–405 primary, secondary and tertiary dendrites were randomly selected from 31–88 neurons. The fluorescence intensity ratio of spine:shaft, defined as ‘spine localization index’, was measured to assess spine accumulation of indicated proteins as previously described (Okuno et al., 2012):

Spine localization index=(GFP-spine/GFP-shaft)/(Dsred-spine/Dsred-shaft).

Briefly, the images were analyzed by line-scan analyses of ImageJ software to obtain intensity plots. The intensity values were taken at the peak of the green fluorescence intensity plot for GFP value and at the peak of the red fluorescence intensity plot for Dsred value.

Protrusions were classified into spines and filopodia as previously described (Llano et al., 2015; Park et al., 2012; Zhang and Macara, 2006). Filopodia were defined as protrusions without a distinguishable head or with a length of >3.5 µm, whereas spines were defined as protrusions with a head, as determined by head/neck ratio >1.2. Spines were measured manually using ImageJ software. Synapses were defined as spines that colocalized with a presynaptic marker, synaptophysin, or a postsynaptic marker, PSD-95. For the number of synapses, the synaptophysin-positive spines were counted.

FRET

Cultured hippocampal neurons from neuron-specific βPix isoform KO mice were transfected with Raichu-Rac1 probe (Itoh et al., 2002) and Flag-tagged βPix constructs at 7 DIV. At 19 DIV, neurons were fixed by adding 3.7% paraformaldehyde for 10 min at RT and washed three times in PBS. Coverslips were mounted and then imaged using a Zeiss LSM 700 confocal laser-scanning microscope equipped with a 40×, 1.20 NA C-Apochromat objective. Images were obtained with the following settings: FRET, excitation (Ex) 405 nm/emission (Em) 500–600 nm; CFP, Ex 405 nm/Em 420–490 nm; YFP, Ex 488 nm/Em 500–600 nm. After acquisition of a FRET image, intensity of the FRET signal was calculated using the ImageJ plug-in PixFRET (NIH) (Feige et al., 2005).

Statistics

All statistical analyses were performed from at least three independent experiments by Student's t-test or one-way analysis of variance (ANOVA) followed by Bonferroni-corrected t-test using Origin 8.5 software (OriginLab Co., Northampton, MA). Actual P-values are indicated in the figure legends. All histograms representing western blots and morphometric data in neurons are represented as mean±s.e.m.CS224980C66,CS224980C67 

We thank Dr Heiner Westphal and Alex Grinberg, who were former members of Mammalian Genes and Development, Eunice Kennedy Shriver National Institute of Child Health and Human Development, National Institutes of Health, Bethesda, USA, for advice and technical assistance for generating βPix KO mice.

Author contributions

Conceptualization: M.S., S.S., D.P.; Methodology: M.S., S.S., D.P.; Validation: M.S., S.S., J.E.S., S.-H.L., S.-O.H.; Formal analysis: M.S., S.S., J.E.S., D.P.; Investigation: M.S., S.S., J.E.S., S.-H.L., S.-O.H.; Resources: M.S., S.S.; Data curation: M.S., S.S.; Writing - original draft: M.S., S.S., D.P.; Writing - review & editing: J.E.S., S.-H.L.; Visualization: M.S., S.S.; Supervision: D.P.; Project administration: D.P.; Funding acquisition: D.P.

Funding

This work was supported by the Basic Science Research Program through the National Research Foundation of Korea [NRF-2016R1D1A1B03934362 and NRF-2017R1A2B4006259].

Bagrodia
,
S.
,
Taylor
,
S. J.
,
Jordon
,
K. A.
,
Van Aelst
,
L.
and
Cerione
,
R. A.
(
1998
).
A novel regulator of p21-activated kinases
.
J. Biol. Chem.
273
,
23633
-
23636
.
Beaudoin
,
G. M. J.
, III
,
Lee
,
S.-H.
,
Singh
,
D.
,
Yuan
,
Y.
,
Ng
,
Y.-G.
,
Reichardt
,
L. F.
and
Arikkath
,
J.
(
2012
).
Culturing pyramidal neurons from the early postnatal mouse hippocampus and cortex
.
Nat. Protoc.
7
,
1741
-
1754
.
Bernards
,
A.
and
Settleman
,
J.
(
2004
).
GAP control: regulating the regulators of small GTPases
.
Trends Cell Biol.
14
,
377
-
385
.
Blanco-Suárez
,
E.
,
Fiuza
,
M.
,
Liu
,
X.
,
Chakkarapani
,
E.
and
Hanley
,
J. G.
(
2014
).
Differential Tiam1/Rac1 activation in hippocampal and cortical neurons mediates differential spine shrinkage in response to oxygen/glucose deprivation
.
J. Cereb. Blood Flow Metab.
34
,
1898
-
1906
.
Bryan
,
B.
,
Kumar
,
V.
,
Stafford
,
L. J.
,
Cai
,
Y.
,
Wu
,
G.
and
Liu
,
M.
(
2004
).
GEFT, a Rho family guanine nucleotide exchange factor, regulates neurite outgrowth and dendritic spine formation
.
J. Biol. Chem.
279
,
45824
-
45832
.
Calabrese
,
B.
,
Wilson
,
M. S.
and
Halpain
,
S.
(
2006
).
Development and regulation of dendritic spine synapses
.
Physiology (Bethesda)
21
,
38
-
47
.
Cingolani
,
L. A.
and
Goda
,
Y.
(
2008
).
Actin in action: the interplay between the actin cytoskeleton and synaptic efficacy
.
Nat. Rev. Neurosci.
9
,
344
-
356
.
Feige
,
J. N.
,
Sage
,
D.
,
Wahli
,
W.
,
Desvergne
,
B.
and
Gelman
,
L.
(
2005
).
PixFRET, an ImageJ plug-in for FRET calculation that can accommodate variations in spectral bleed-throughs
.
Microsc. Res. Tech.
68
,
51
-
58
.
Feng
,
Q.
,
Baird
,
D.
,
Peng
,
X.
,
Wang
,
J.
,
Ly
,
T.
,
Guan
,
J.-L.
and
Cerione
,
R. A.
(
2006
).
Cool-1 functions as an essential regulatory node for EGF receptor- and Src-mediated cell growth
.
Nat. Cell Biol.
8
,
945
-
956
.
Feng
,
Q.
,
Baird
,
D.
,
Yoo
,
S.
,
Antonyak
,
M.
and
Cerione
,
R. A.
(
2010
).
Phosphorylation of the cool-1/beta-Pix protein serves as a regulatory signal for the migration and invasive activity of Src-transformed cells
.
J. Biol. Chem.
285
,
18806
-
18816
.
Fortin
,
D. A.
,
Srivastava
,
T.
and
Soderling
,
T. R.
(
2012
).
Structural modulation of dendritic spines during synaptic plasticity
.
Neuroscientist
18
,
326
-
341
.
Goslin
,
K.
,
Asmussen
,
H.
and
Banker
,
G.
(
1998
).
Rat hippocampal neurons in low-density culture
. In
Culturing Nerve Cells
(ed.
G.
Banker
and
K.
Goslin
), pp.
339
-
370
.
MA
:
MIT University Press: Cambridge
.
Govek
,
E.-E.
,
Newey
,
S. E.
and
Van Aelst
,
L.
(
2005
).
The role of the Rho GTPases in neuronal development
.
Genes Dev.
19
,
1
-
49
.
Harris
,
K. M.
(
1999
).
Structure, development, and plasticity of dendritic spines
.
Curr. Opin. Neurobiol.
9
,
343
-
348
.
Hayashi
,
T.
,
Umemori
,
H.
,
Mishina
,
M.
and
Yamamoto
,
T.
(
1999
).
The AMPA receptor interacts with and signals through the protein tyrosine kinase Lyn
.
Nature
397
,
72
-
76
.
Hayashi
,
S.
,
Lewis
,
P.
,
Pevny
,
L.
and
McMahon
,
A. P.
(
2002
).
Efficient gene modulation in mouse epiblast using a Sox2Cre transgenic mouse strain
.
Mech. Dev.
119
Suppl. 1
,
S97
-
s101
.
Hering
,
H.
and
Sheng
,
M.
(
2001
).
Dendritic spines: structure, dynamics and regulation
.
Nat. Rev. Neurosci.
2
,
880
-
888
.
Itoh
,
R. E.
,
Kurokawa
,
K.
,
Ohba
,
Y.
,
Yoshizaki
,
H.
,
Mochizuki
,
N.
and
Matsuda
,
M.
(
2002
).
Activation of rac and cdc42 video imaged by fluorescent resonance energy transfer-based single-molecule probes in the membrane of living cells
.
Mol. Cell. Biol.
22
,
6582
-
6591
.
Kim
,
T.
and
Park
,
D.
(
2001
).
Molecular cloning and characterization of a novel mouse betaPix isoform
.
Mol. Cells
11
,
89
-
94
.
Kim
,
S.
,
Kim
,
T.
,
Lee
,
D.
,
Park
,
S. H.
,
Kim
,
H.
and
Park
,
D.
(
2000
).
Molecular cloning of neuronally expressed mouse betaPix isoforms
.
Biochem. Biophys. Res. Commun.
272
,
721
-
725
.
Kim
,
S.
,
Lee
,
S. H.
and
Park
,
D.
(
2001
).
Leucine zipper-mediated homodimerization of the p21-activated kinase-interacting factor, beta Pix. Implication for a role in cytoskeletal reorganization
.
J. Biol. Chem.
276
,
10581
-
10584
.
Koch
,
C.
and
Zador
,
A.
(
1993
).
The function of dendritic spines: devices subserving biochemical rather than electrical compartmentalization
.
J. Neurosci.
13
,
413
-
422
.
Koh
,
C. G.
,
Manser
,
E.
,
Zhao
,
Z. S.
,
Ng
,
C. P.
and
Lim
,
L.
(
2001
).
Beta1PIX, the PAK-interacting exchange factor, requires localization via a coiled-coil region to promote microvillus-like structures and membrane ruffles
.
J. Cell Sci.
114
,
4239
-
4251
.
Korobova
,
F.
and
Svitkina
,
T.
(
2010
).
Molecular architecture of synaptic actin cytoskeleton in hippocampal neurons reveals a mechanism of dendritic spine morphogenesis
.
Mol. Biol. Cell
21
,
165
-
176
.
Kos
,
C. H.
(
2004
).
Cre/loxP system for generating tissue-specific knockout mouse models
.
Nutr. Rev.
62
,
243
-
246
.
Kuo
,
W. L.
,
Chung
,
K. C.
and
Rosner
,
M. R.
(
1997
).
Differentiation of central nervous system neuronal cells by fibroblast-derived growth factor requires at least two signaling pathways: roles for Ras and Src
.
Mol. Cell. Biol.
17
,
4633
-
4643
.
Lai
,
K.-O.
and
Ip
,
N. Y.
(
2013
).
Structural plasticity of dendritic spines: the underlying mechanisms and its dysregulation in brain disorders
.
Biochim. Biophys. Acta
1832
,
2257
-
2263
.
Lin
,
W.-H.
and
Webb
,
D. J.
(
2009
).
Actin and actin-binding proteins: masters of dendritic spine formation, morphology, and function
.
Open Neurosci. J.
3
,
54
-
66
.
Llano
,
O.
,
Smirnov
,
S.
,
Soni
,
S.
,
Golubtsov
,
A.
,
Guillemin
,
I.
,
Hotulainen
,
P.
,
Medina
,
I.
,
Nothwang
,
H. G.
,
Rivera
,
C.
and
Ludwig
,
A.
(
2015
).
KCC2 regulates actin dynamics in dendritic spines via interaction with beta-PIX
.
J. Cell Biol.
209
,
671
-
686
.
Manser
,
E.
,
Loo
,
T.-H.
,
Koh
,
C.-G.
,
Zhao
,
Z.-S.
,
Chen
,
X.-Q.
,
Tan
,
L.
,
Tan
,
I.
,
Leung
,
T.
and
Lim
,
L.
(
1998
).
PAK kinases are directly coupled to the PIX family of nucleotide exchange factors
.
Mol. Cell
1
,
183
-
192
.
Mateos
,
J. M.
,
Lüthi
,
A.
,
Savic
,
N.
,
Stierli
,
B.
,
Streit
,
P.
,
Gähwiler
,
B. H.
and
McKinney
,
R. A.
(
2007
).
Synaptic modifications at the CA3-CA1 synapse after chronic AMPA receptor blockade in rat hippocampal slices
.
J. Physiol.
581
,
129
-
138
.
Matsuzaki
,
M.
,
Honkura
,
N.
,
Ellis-Davies
,
G. C. R.
and
Kasai
,
H.
(
2004
).
Structural basis of long-term potentiation in single dendritic spines
.
Nature
429
,
761
-
766
.
Matus
,
A.
(
2000
).
Actin-based plasticity in dendritic spines
.
Science
290
,
754
-
758
.
Mayhew
,
M. W.
,
Jeffery
,
E. D.
,
Sherman
,
N. E.
,
Nelson
,
K.
,
Polefrone
,
J. M.
,
Pratt
,
S. J.
,
Shabanowitz
,
J.
,
Parsons
,
J. T.
,
Fox
,
J. W.
,
Hunt
,
D. F.
, et al. 
(
2007
).
Identification of phosphorylation sites in betaPIX and PAK1
.
J. Cell Sci.
120
,
3911
-
3918
.
McKinney
,
R. A.
,
Capogna
,
M.
,
Dürr
,
R.
,
Gahwiler
,
B. H.
and
Thompson
,
S. M.
(
1999
).
Miniature synaptic events maintain dendritic spines via AMPA receptor activation
.
Nat. Neurosci.
2
,
44
-
49
.
Morita
,
A.
,
Yamashita
,
N.
,
Sasaki
,
Y.
,
Uchida
,
Y.
,
Nakajima
,
O.
,
Nakamura
,
F.
,
Yagi
,
T.
,
Taniguchi
,
M.
,
Usui
,
H.
,
Katoh-Semba
,
R.
, et al. 
(
2006
).
Regulation of dendritic branching and spine maturation by semaphorin3A-Fyn signaling
.
J. Neurosci.
26
,
2971
-
2980
.
Newey
,
S. E.
,
Velamoor
,
V.
,
Govek
,
E.-E.
and
Van Aelst
,
L.
(
2005
).
Rho GTPases, dendritic structure, and mental retardation
.
J. Neurobiol.
64
,
58
-
74
.
Oh
,
W. K.
,
Yoo
,
J. C.
,
Jo
,
D.
,
Song
,
Y. H.
,
Kim
,
M. G.
and
Park
,
D.
(
1997
).
Cloning of a SH3 domain-containing proline-rich protein, p85SPR, and its localization in focal adhesion
.
Biochem. Biophys. Res. Commun.
235
,
794
-
798
.
Okuno
,
H.
,
Akashi
,
K.
,
Ishii
,
Y.
,
Yagishita-Kyo
,
N.
,
Suzuki
,
K.
,
Nonaka
,
M.
,
Kawashima
,
T.
,
Fujii
,
H.
,
Takemoto-Kimura
,
S.
,
Abe
,
M
, et al. 
. (
2012
).
Inverse synaptic tagging of inactive synapses via dynamic interaction of Arc/Arg3.1 with CaMKIIβ
.
Cell 149, 886-898
.
Park
,
E.
,
Chi
,
S.
and
Park
,
D.
(
2012
).
Activity-dependent modulation of the interaction between CaMKIIalpha and Abi1 and its involvement in spine maturation
.
J. Neurosci.
32
,
13177
-
13188
.
Penzes
,
P.
,
Johnson
,
R. C.
,
Sattler
,
R.
,
Zhang
,
X.
,
Huganir
,
R. L.
,
Kambampati
,
V.
,
Mains
,
R. E.
and
Eipper
,
B. A.
(
2001
).
The neuronal Rho-GEF Kalirin-7 interacts with PDZ domain-containing proteins and regulates dendritic morphogenesis
.
Neuron
29
,
229
-
242
.
Penzes
,
P.
,
Cahill
,
M. E.
,
Jones
,
K. A.
,
VanLeeuwen
,
J.-E.
and
Woolfrey
,
K. M.
(
2011
).
Dendritic spine pathology in neuropsychiatric disorders
.
Nat. Neurosci.
14
,
285
-
293
.
Peters
,
A.
and
Kaiserman-Abramof
,
I. R.
(
1970
).
The small pyramidal neuron of the rat cerebral cortex. The perikaryon, dendrites and spines
.
Am. J. Anat.
127
,
321
-
355
.
Repetto
,
D.
,
Camera
,
P.
,
Melani
,
R.
,
Morello
,
N.
,
Russo
,
I.
,
Calcagno
,
E.
,
Tomasoni
,
R.
,
Bianchi
,
F.
,
Berto
,
G.
,
Giustetto
,
M.
, et al. 
(
2014
).
p140Cap regulates memory and synaptic plasticity through Src-mediated and citron-N-mediated actin reorganization
.
J. Neurosci.
34
,
1542
-
1553
.
Rosenberger
,
G.
and
Kutsche
,
K.
(
2006
).
AlphaPIX and betaPIX and their role in focal adhesion formation
.
Eur. J. Cell Biol.
85
,
265
-
274
.
Saneyoshi
,
T.
and
Hayashi
,
Y.
(
2012
).
The Ca2+ and Rho GTPase signaling pathways underlying activity-dependent actin remodeling at dendritic spines
.
Cytoskeleton (Hoboken)
69
,
545
-
554
.
Saneyoshi
,
T.
,
Wayman
,
G.
,
Fortin
,
D.
,
Davare
,
M.
,
Hoshi
,
N.
,
Nozaki
,
N.
,
Natsume
,
T.
and
Soderling
,
T. R.
(
2008
).
Activity-dependent synaptogenesis: regulation by a CaM-kinase kinase/CaM-kinase I/betaPIX signaling complex
.
Neuron
57
,
94
-
107
.
Schmidt
,
A.
and
Hall
,
A.
(
2002
).
Guanine nucleotide exchange factors for Rho GTPases: turning on the switch
.
Genes Dev.
16
,
1587
-
1609
.
Schaeren-Wiemers
,
N.
and
Gerfin-Moser
,
A.
(
1993
).
A single protocol to detect transcripts of various types and expression levels in neural tissue and cultured cells: in situ hybridization using digoxigenin-labelled cRNA probes
.
Histochemistry 100, 431-440
.
Shepherd
,
G. M.
(
1996
).
The dendritic spine: a multifunctional integrative unit
.
J. Neurophysiol.
75
,
2197
-
2210
.
Shin
,
E. Y.
,
Shin
,
K. S.
,
Lee
,
C. S.
,
Woo
,
K. N.
,
Quan
,
S. H.
,
Soung
,
N. K.
,
Kim
,
Y. G.
,
Cha
,
C. I.
,
Kim
,
S. R.
,
Park
,
D.
, et al. 
(
2002
).
Phosphorylation of p85 beta PIX, a Rac/Cdc42-specific guanine nucleotide exchange factor, via the Ras/ERK/PAK2 pathway is required for basic fibroblast growth factor-induced neurite outgrowth
.
J. Biol. Chem.
277
,
44417
-
44430
.
Shin
,
E.-Y.
,
Woo
,
K.-N.
,
Lee
,
C.-S.
,
Koo
,
S.-H.
,
Kim
,
Y.-G.
,
Kim
,
W.-J.
,
Bae
,
C.-D.
,
Chang
,
S.-I.
and
Kim
,
E.-G.
(
2004
).
Basic fibroblast growth factor stimulates activation of Rac1 through a p85 betaPIX phosphorylation-dependent pathway
.
J. Biol. Chem.
279
,
1994
-
2004
.
Socodato
,
R.
,
Santiago
,
F. N.
,
Portugal
,
C. C.
,
Domingues
,
A. F.
,
Santiago
,
A. R.
,
Relvas
,
J. B.
,
Ambrosio
,
A. F.
and
Paes-de-Carvalho
,
R.
(
2012
).
Calcium-permeable alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptors trigger neuronal nitric-oxide synthase activation to promote nerve cell death in an Src kinase-dependent fashion
.
J. Biol. Chem.
287
,
38680
-
38694
.
Takahashi
,
H.
,
Yamazaki
,
H.
,
Hanamura
,
K.
,
Sekino
,
Y.
and
Shirao
,
T.
(
2009
).
Activity of the AMPA receptor regulates drebrin stabilization in dendritic spine morphogenesis
.
J. Cell Sci.
122
,
1211
-
1219
.
Tashiro
,
A.
and
Yuste
,
R.
(
2004
).
Regulation of dendritic spine motility and stability by Rac1 and Rho kinase: evidence for two forms of spine motility
.
Mol. Cell. Neurosci.
26
,
429
-
440
.
ten Klooster
,
J. P.
,
Jaffer
,
Z. M.
,
Chernoff
,
J.
and
Hordijk
,
P. L.
(
2006
).
Targeting and activation of Rac1 are mediated by the exchange factor beta-Pix
.
J. Cell Biol.
172
,
759
-
769
.
Thomas
,
S. M.
and
Brugge
,
J. S.
(
1997
).
Cellular functions regulated by Src family kinases
.
Annu. Rev. Cell Dev. Biol.
13
,
513
-
609
.
Tolias
,
K. F.
,
Bikoff
,
J. B.
,
Burette
,
A.
,
Paradis
,
S.
,
Harrar
,
D.
,
Tavazoie
,
S.
,
Weinberg
,
R. J.
and
Greenberg
,
M. E.
(
2005
).
The Rac1-GEF Tiam1 couples the NMDA receptor to the activity-dependent development of dendritic arbors and spines
.
Neuron
45
,
525
-
538
.
Tolias
,
K. F.
,
Duman
,
J. G.
and
Um
,
K.
(
2011
).
Control of synapse development and plasticity by Rho GTPase regulatory proteins
.
Prog. Neurobiol.
94
,
133
-
148
.
van Spronsen
,
M.
and
Hoogenraad
,
C. C.
(
2010
).
Synapse pathology in psychiatric and neurologic disease
.
Curr. Neurol. Neurosci. Rep.
10
,
207
-
214
.
Webb
,
D. J.
,
Zhang
,
H.
,
Majumdar
,
D.
and
Horwitz
,
A. F.
(
2007
).
alpha5 integrin signaling regulates the formation of spines and synapses in hippocampal neurons
.
J. Biol. Chem.
282
,
6929
-
6935
.
Wu
,
X.
,
Zhu
,
D.
,
Jiang
,
X.
,
Okagaki
,
P.
,
Mearow
,
K.
,
Zhu
,
G.
,
McCall
,
S.
,
Banaudha
,
K.
,
Lipsky
,
R. H.
and
Marini
,
A. M.
(
2004
).
AMPA protects cultured neurons against glutamate excitotoxicity through a phosphatidylinositol 3-kinase-dependent activation in extracellular signal-regulated kinase to upregulate BDNF gene expression
.
J. Neurochem.
90
,
807
-
818
.
Xie
,
Z.
,
Srivastava
,
D. P.
,
Photowala
,
H.
,
Kai
,
L.
,
Cahill
,
M. E.
,
Woolfrey
,
K. M.
,
Shum
,
C. Y.
,
Surmeier
,
D. J.
and
Penzes
,
P.
(
2007
).
Kalirin-7 controls activity-dependent structural and functional plasticity of dendritic spines
.
Neuron
56
,
640
-
656
.
Yuste
,
R.
and
Majewska
,
A.
(
2001
).
On the function of dendritic spines
.
Neuroscientist
7
,
387
-
395
.
Zhang
,
H.
and
Macara
,
I. G.
(
2006
).
The polarity protein PAR-3 and TIAM1 cooperate in dendritic spine morphogenesis
.
Nat. Cell Biol.
8
,
227
-
237
.
Zhang
,
H.
,
Webb
,
D. J.
,
Asmussen
,
H.
and
Horwitz
,
A. F.
(
2003
).
Synapse formation is regulated by the signaling adaptor GIT1
.
J. Cell Biol.
161
,
131
-
142
.
Zhang
,
H.
,
Webb
,
D. J.
,
Asmussen
,
H.
,
Niu
,
S.
and
Horwitz
,
A. F.
(
2005
).
A GIT1/PIX/Rac/PAK signaling module regulates spine morphogenesis and synapse formation through MLC
.
J. Neurosci.
25
,
3379
-
3388
.

Competing interests

The authors declare no competing or financial interests.

Supplementary information