SAP97 Directs the Localization of Kv4.2 to Spines in Hippocampal Neurons

The pore-forming α-subunit Kv4.2 is a key constituent of the A-type channel and critically involved in the regulation of dendritic excitability and plasticity. Here we show that Kv4.2 is enriched in the postsynaptic density (PSD) fraction and specifically interacts with synapse-associated protein 97 (SAP97). This interaction requires an intact C terminus of Kv4.2 and occurs via the PDZ domains of SAP97. Pharmacologically induced translocation of SAP97 to spines also drives Kv4.2 to the PSD, whereas SAP97 lentivirally based RNA interference reduces Kv4.2 in the PSD. In addition, calcium/calmodulin-dependent protein kinase II (CaMKII)-dependent SAP97 phosphorylation regulates the subcellular localization of Kv4.2. These results show that SAP97-CaMKII pathway plays an important role for the trafficking of Kv4.2 to dendrites and spines.

The efficient transmission and processing of information in neurons depend on the precise subcellular localization and distribution of ion channels in different compartments and at sites of synaptic communication (1,2). Dendrites and spines are the primary sites of synaptic input and express several ligand-and voltage-gated ion channels. Among those are the transient, fast-inactivating A-type channels. The density of A-type channels increases at least 5-fold from the soma to about 350 m in dendrites (3) and appears to be high in the oblique dendrites as well. The A-type current plays an important role in regulating dendritic excitability by reducing the amplitude of excitatory postsynaptic potentials and back propagating action potential and by regulating the induction and expression of long term potentiation. These A-type channels are also modulated by a variety of neurotransmitters such as norepinehrine, acetylcholine, and dopamine (4).
Pharmacological and molecular biological studies have implicated Kv4.2 as a key subunit of the A-type channel (5). Biochemical studies have demonstrated the phosphorylation of Kv4.2 by various kinases in agreement with functional studies; for example, phosphorylation of Kv4.2 by CaMKII 2 was shown to increase the surface expression of Kv4.2 without affecting its biophysical properties (6). Despite this observation, very little is known about how K ϩ channels, and in particular A-type channels, are targeted and expressed in dendrites and spines. In heterologous expression systems, K ϩ channels have been shown to interact with members of the membrane-associated guanylate kinase (MAGUK) family. In particular, interaction with PSD-95 has been shown to facilitate surface expression and clustering of the K ϩ channels (7)(8)(9). SAP97 and PSD-95 have been shown to exhibit distinct mechanisms for regulating the surface expression and clustering of Kv1-type channels. SAP97 is also of importance for the surface expression of Kv1.5 (10) and the inwardly rectifying channel Kir2.1 (11). In addition, SAP97 regulates biosynthesis and surface expression of AMPA receptors (12,13). SAP97 localization to spines appears to be regulated by CaMKII (14). In particular, we have previously shown that CaMKII can in vivo phosphorylate SAP97 into two different residues, Ser-39 and Ser-232. These two residues are located in crucial domains responsible for regulating SAP97 protein trafficking as well as binding of SAP97 to interacting proteins (15).
Here, we have investigated the molecular details of Kv4.2 interaction with MAGUK proteins in organotypic slice cultures of the hippocampus and the role of this interaction for the subcellular localization of Kv4.2 in hippocampal neurons. In addition, we have found evidence for the SAP97-CaMKII pathway in regulating the subcellular localization of Kv4.2.
were expressed at similar expression levels (data not shown). The Kv4.2 IRES constitutively active calmodulin-dependent kinase II ␣ (IRES tCaMKII) was generated by transferring the IRES tCaMKII construct kindly provided by Y. Hayashi (Massachusetts Institute of Technology) into EGFP Kv4.2 in pSinRep 5. The integrity of these constructs was verified by automatic sequencing.
Antibodies-The monoclonal antibody to Kv4.2 (K57/1) was purchased from Neuromab and has been described previously (18). The monoclonal antibody was developed by and/or obtained from the University of California, Davis, NINDS/ NIMH NeuroMab Facility. Monoclonal antibody to synaptophysin was purchased from Roche Applied Science; mouse ␣CaMKII antibody was purchased from Chemicon International, Inc. (Temecula, CA); polyclonal SAP97 antibody, polyclonal SAP102 antibody, and monoclonal PSD-95 antibody were purchased from Affinity BioReagents Inc. (Golden, CO); monoclonal SAP97 antibody was purchased from StressGen (Victoria, British Columbia, Canada); polyclonal anti-GFP and AlexaFluor 488, 555, 568, and 633 secondary antibodies were purchased from Molecular Probes (Eugene, OR).
Neuronal Culture and Transfection-Hippocampal neuronal cultures were prepared from embryonic day 18 to 19 rat hippocampi as described previously, with minor modifications (19). Neurons were transfected using the calcium phosphate precipitation method at 7 days in vitro (DIV).
Organotypic Slice Cultures and Sindbis Virus Infections-Organotypic slice cultures were prepared from postnatal day 7 (P7) animals and cultured on Millicell inserts (Millipore) as described previously in detail (17). Typically, slice cultures were infected after 5-7 DIV. Sindbis pseudovirions were prepared using a modified less cytotoxic variant. The cDNAs were subcloned into a modified form of pSinRep5 and packaged using the DHBB helper. Psesudovirions were harvested typically after 72 h and used to focally infect the slice cultures as described previously (17).
Lentiviral RNA Interference of SAP97-For the siRNA expressing lentivirus vector, an RNAi stem-loop (12) has been cloned in the lentivirus-based vector pLL3.7 (20; Massachusetts Institute of Technology Center for Cancer Research, Cambridge), and an empty pLL3.7 vector has been used to generate the control lentivirus vectors. The lentiviral infecting particles were prepared as described previously (21). Neurons were infected with SAP97i-lentivirus or control virus at DIV6, and TIF preparation was performed at DIV10. Subcellular Fractionation-TIF were isolated from neurons harvested at 10 -14 DIV or hippocampal slices as described previously (15). PSDs from rat hippocampus were purified as described previously (22). COS-7 Cell Culture and Transfection-COS-7 cells at 20 -50% confluence (24 h after plating on glass coverslips in 12-well plates) were transfected by Superfect transfection reagent (Qiagen, Valencia, CA) with cDNA expression constructs (1-1.5 g of DNA/well) for 3 h at 5% CO 2 , 37°C. Cells were washed twice with phosphate-buffered saline, fed with Dulbecco's modified Eagle's medium, 10% fetal bovine serum, and 1% penicillin/streptomycin, and grown for 24 -48 h before fixation for immunocytochemistry or before metabolic labeling experi-ments. Mutated products were obtained by using the QuikChange TM site-directed mutagenesis kit (Stratagene, La Jolla, CA).
Immunofluorescence Labeling, Image Acquisition, Quantification, and Statistical Analysis-Hippocampal neurons were fixed in 100% methanol at Ϫ20°C for 15 min. Primary and secondary antibodies were applied in GDB buffer (30 mM phosphate buffer (pH 7.4) containing 0.2% gelatin, 0.5% Triton X-100, and 0.8 M NaCl). Confocal images were obtained using a Nikon ϫ60 objective with sequential acquisition setting at 1024 ϫ 1024 pixel resolution. Each image was a z series projection of ϳ8 -12 images taken at 0.5-1-m depth intervals. Transfected COS-7 cells, transfected neurons, and pharmacologically treated neurons were chosen randomly for quantification from two to five coverslips from three to five independent experiments. Quantification of confocal experiments was performed using Laserpix software (Bio-Rad). Image acquisition, quantification of the fluorescence signal, and co-localization analysis were performed by investigators who were 'blind' to the experimental condition. Quantification of Western blot analysis was performed by means of computer-assisted imaging (Quantity-One System; Bio-Rad), and statistical evaluations were performed according to one-way analysis of variance followed by Bonferroni as post hoc comparison test; if the experiment includes only two experimental conditions, paired Student's t test was used. All data are presented as mean Ϯ S.E. and, if not indicated otherwise, as percentage of control derived from three to six independent experiments.
Cloning, Expression, and Purification of GST Fusion Protein-SAP97 fragments were subcloned downstream of glutathione S-transferase (GST) in the BamHI and HindIII sites of the expression plasmid pGEX-KG by PCR using plaque-forming unit polymerase (Promega). The inserts were fully sequenced with the ABI Prism 310 genetic analyzer (ABI Prisma). SAP97-GST fusion proteins were expressed in Escherichia coli, purified on glutathione-agarose beads (Sigma), and eluted as described previously (23).
Pulldown Assay-Aliquots of PSD containing 10 g of proteins were diluted with Tris-buffered saline, 0.1% SDS to a final volume of 200 l and incubated (1 h, 37°C) with glutathioneagarose beads saturated with GST fusion proteins or GST alone. The beads were extensively washed with Tris-buffered saline, 0.1% Triton X-100. Bound proteins were resolved by SDS-PAGE and subjected to immunoblot analysis with a monoclonal Kv4.2 antibody.

SAP97 Co-localization and Interaction with Kv4.2 from Rat
Hippocampus-We first examined the relative abundance of voltage-gated potassium channel Kv4.2 in rat hippocampal subcellular compartments by means of a biochemical fractionation method, as described previously (15). Postsynaptic densities (PSD) were purified from rat hippocampus, and the expression of Kv4.2 as well as of pre-and postsynaptic markers was investigated in various subcellular compartments through Western blotting analysis. As shown in Fig. 1a, Kv4.2 channel was enriched in the PSD fraction and in the Triton-insoluble "PSD-enriched" fraction (TIF). Kv4.2 was present with a similar

SAP97-Kv4.2 Interaction in Hippocampal Neurons
distribution in the total homogenate and in the crude membrane fraction (P2) and at a low level in the synaptosomal membrane and in the low speed supernatant (S1) fractions. In the same samples, we examined the subcellular distribution of synaptophysin (presynaptic marker), ␣CaMKII, PSD-95, SAP102, and SAP97. As expected, synaptophysin was present in all subcellular compartments analyzed but not in the PSD and in the TIF-purified fractions, whereas PSD-95 and ␣CaMKII have a similar distribution pattern being enriched in synaptosomes and PSD fractions (Fig. 1a). The partition pattern of Kv4.2 channel was similar to that of SAP97 and SAP102.
Previous works described the presence of a specific binding between Kv4.2 and PSD-95 (7,8). Because we found a comparable subcellular localization of Kv4.2 with both SAP97 and SAP102, we investigated whether Kv4.2 forms complexes with different members of the MAGUK protein family. First of all, co-immunoprecipitation experiments were performed from homogenate of rat hippocampus (Fig. 1b). As shown in Fig. 1b (top panel), homogenates were incubated with antibodies directed against PSD-95, SAP97, and SAP102, and the precipitates were probed with the Kv4.2 antibody. Anti-SAP97 and anti-PSD-95, but not anti-SAP102, co-immunoprecipitated Kv4.2 ( Fig. 1b, top panel); the absence of any Kv4.2 signal in the No Ab lane tends to exclude that in our experimental conditions, and the co-precipitation of Kv4.2 reflects an unspecific immunoprecipitation of insoluble proteins. Accordingly, anti-Kv4.2 was able to co-precipitate SAP97 and PSD-95 but not SAP102 (Fig. 1b, bottom panel). Furthermore, we performed Kv4.2-MAGUK co-immunoprecipitation experiments from different subcellular fractions, i.e. synaptosomes and PSD to evaluate the capability of SAP97 and PSD-95 to bind Kv4.2 in specific subcellular compartments. Fig. 1c (top panel) shows that, in all tested compartments, Kv4.2 co-precipitates with SAP97 suggesting the subsistence of the binding between the proteins also at synaptic sites. On the other hand, the PSD-95-Kv4.2 interaction was more pronounced in the PSD fraction (Fig. 1c, bottom panel).
Interaction between SAP97 and Kv4.2 in Virally Infected Organotypic Slices-To further confirm an interaction of Kv4.2/SAP97 in neurons and to confirm in neurons the identity of Kv4.2 and SAP97 domains needed for the interaction, we used organotypic slices virally infected with SAP97 and Kv4.2 constructs.
Slices were infected with a modified Sindbis virus expressing EGFP-SAP97 or EGFP-PSD-95 constructs (17). Co-immunoprecipitation studies were performed from infected slice homogenates by means of an EGFP antibody. Fig. 2a (left panel) shows that both infected SAP97 and PSD-95 are able to bind endogenous Kv4.2. No immunostaining for Kv4.2 was detected when the precipitating EGFP antibody was omitted. In addition, experiments performed by transfecting slices with an EGFP-SAP97-⌬PDZ construct lacking the three PDZ domains of SAP97 (Fig. 2a, right panel) confirmed the specificity of the co-precipitation assay as well as the existence of a PDZ-medi-ated interaction between SAP97 and Kv4.2 as already suggested by the pulldown assay (see Fig. 1d). To further strengthen this point, viral transfection of organotypic slices and subsequent co-precipitation experiments were repeated by using EGFP-Kv4.2 constructs. Under these experimental conditions, the EGFP antibody was able to co-precipitate endogenous SAP97 (Fig. 2b) from EGFP-Kv4.2 transfected slice homogenates; no bands corresponding to SAP97 were found in the EGFP lane indicating the absence of any interaction between the exogenously added EGFP tag and endogenous SAP97 present in organotypic slice culture. Viral infection of the EGFP-Kv4.2construct lacking the Kv4.2 C terminus, including VSAL PDZbinding domain, did not produce any immunostaining for SAP97 in the co-immunoprecipitated material; these data confirm the requirement of the PDZ-mediated interaction to detect Kv4.2 binding to SAP97.
Recent studies have shown that Kv4.2 as well as SAP97 are substrates for CaMKII in hippocampal neurons (6,15); to check whether CaMKII phosphorylation could affect Kv4.2-SAP97 protein-protein interaction, active ␣CaMKII (1-290 truncation) was co-transfected with SAP97 and Kv4.2 (Fig. 2c). The co-precipitation assay showed the presence of Kv4.2-SAP97 coprecipitation also in presence of the active form of the kinase (Fig. 2c), suggesting that the constitutively active CaMKII does not qualitatively alter the interaction between Kv4.2 and SAP97.
SAP97-mediated Trafficking of Kv4.2 in Cultured Hippocampal Neurons-To address the role of SAP97 in modulating Kv4.2 localization to the postsynaptic compartment, we infected primary hippocampal cultures with SAP97 RNAi expressing lentivirus. Nearly 100% of neurons in culture can be infected with lentivirus with minimal cytotoxicity (data not shown); the effect of SAP97 RNAi could therefore be quantified through a biochemical approach across the entire population of neurons in culture. TIF was obtained from control and SAP97 RNAi neurons (15,24), and protein levels were measured in the homogenate and TIF. The same amount of proteins from homogenate and TIF was loaded on the SDS-PAGE for Western blot analysis. Compared with empty virus infection at the same viral titer, lentivirus expressing SAP97 RNAi caused profound and specific knockdown of SAP97 as shown by immunoblotting of total homogenate of hippocampal cultures (Fig. 3a). As shown in Fig. 3a, siRNA knockdown of SAP97 leads a significant reduction in Kv4.2 localization in the postsynaptic compartment (p Ͻ 0.05, Ϫ30.0 Ϯ 7.4% SAP97 RNAi versus control) suggesting a specific role for SAP97 in the correct Kv4.2 localization at synaptic sites. No alterations of SAP102 or ␣CaMKII levels were observed in total lysates as well as in TIF fractions, confirming the specificity of SAP97 RNAi knockdown (Fig. 3a).
Recent observations indicate that activation of ryanodine receptors in the hippocampus, through caffeine treatment, can play a role in synaptic plasticity events by means of elevation of CaMKII activity (14), addressing CaMKII as a potential enzymatic target of the calcium-induced calcium release from ER ryanodine stores. Interestingly, calcium-induced calcium release from ryanodine receptors, as induced by caffeine treatment, has been shown necessary to trigger CaMKII-dependent

SAP97-Kv4.2 Interaction in Hippocampal Neurons
SAP97 trafficking (14,24). Based on these observations, we tested whether caffeine treatment (10 mM, 15 min) was sufficient to induce not only SAP97 but also Kv4.2 trafficking from the ER. These experiments were performed in the presence of D-2-amino-5-phosphopentanoic acid (APV) to block NMDA receptors as a source of extracellular calcium. As shown in Fig.  3b, caffeine treatment significantly increased Kv4.2 immunostaining in TIF without affecting the total Kv4.2 protein level in the homogenate (*, p Ͻ 0.01; ϩ60.2 Ϯ 13.1%, caffeine versus control expressed as Kv4.2 ratio TIF/homogenate). As expected, treatment of hippocampal cultures with caffeine also leads to a higher staining of SAP97 in TIF (*, p Ͻ 0.01; ϩ51.1 Ϯ 8.3%, caffeine versus control expressed as SAP97 ratio TIF/ homogenate) confirming previous observations (14). On the other hand, no significant effect of caffeine treatment on ␣CaMKII, SAP102 and PSD-95 localization in the TIF was found (Fig. 3b).
We have very recently described that treatment of hippocampal neurons with ryanodine (100 M), to block calcium-induced calcium release from intracellular ER stores, and with a brief pulse of NMDA (50 M), to promote a raise in calcium levels into the postsynaptic compartment, does not influence SAP97-Ser-39 phosphorylation and consequently does not modify SAP97 subcellular distribution (24). Based on these data, we treated primary hippocampal neurons with NMDA ϩ ryanodine to check whether not only SAP97 but also Kv4.2 localization was unaffected by this experimental treatment. As shown in Fig. 3c, no significant effect of both SAP97 (p Ͻ 0.05; Ϫ3.7 Ϯ 9.3%, NMDA ϩ ryanodine versus control expressed as SAP97 ratio TIF/homogenate) and Kv4.2 (p Ͼ 0.05; Ϫ9,1 Ϯ 13.2%, NMDA ϩ ryanodine versus control expressed as Kv4.2 ratio TIF/homogenate) localization in the TIF was induced by NMDA ϩ ryanodine treatment.
To further study the role of SAP97-CaMKII pathway on subcellular localization of native Kv4.2 in neurons, we analyzed by confocal microscopy the effect of caffeine treatment on the endogenous distribution of Kv4.2 in primary hippocampal cultures. It has been shown that Kv4.2 channels in neurons are primarily localized in cell soma and dendrites (4,(25)(26)(27). As shown in Fig. 3d (left panel), endogenous Kv4.2 displayed an intense labeling in the cell soma; immunoreactivity was also detectable in dendrites where a diffuse staining was present. Interestingly, caffeine (Fig. 3d,  right panel) was able to affect Kv4.2 immunostaining leading to a higher signal in the dendritic shafts paralleled to a decreased immunofluorescence in the cell soma. Co-immunoprecipitation experiments were performed to check the possible modulation of the MAGUK-Kv4.2 interaction by caffeine. A significant increase of SAP97-Kv4.2 co-precipitation was found (Fig.  3e) both in homogenate (*, p Ͻ 0.05; ϩ30.2 Ϯ 6.9%, caffeine versus control) and TIF (*, p Ͻ 0.01; ϩ76.1 Ϯ 10.5%, caffeine  SEPTEMBER 28, 2007 • VOLUME 282 • NUMBER 39 versus control) confirming the hypothesis of a Kv4.2 SAP97mediated trafficking induced by caffeine (Fig. 3e). PSD-95-Kv4.2 complex was detected in the TIF only after caffeine treat-ment, confirming the increased localization of Kv4.2 in the postsynaptic compartment. No PSD-95-Kv4.2 interaction was found in the homogenate under these experimental conditions (Fig. 3e). No interaction of SAP102 to Kv4.2 was found confirming the specificity of the binding of SAP97 and PSD-95 to the channel (Fig. 3e).
Recent data showed that CaMKIIdependent SAP97 Ser-39 phosphorylation, induced by caffeine treatment, regulates the subcellular localization of SAP97, providing a fine molecular mechanism responsible for the synaptic delivery of SAP97 itself as well as SAP97-interacting proteins (14,15). Based on these observations, we co-transfected in COS-7 cells Kv4.2wt with SAP97(S39D) mutation construct mimicking CaMKIIdependent SAP97 Ser-39 phosphorylation (Fig. 4, c-e). Co-expression of Kv4.2wt with SAP97(S39D) caused a redistribution of Kv4.2 channel staining throughout the cell with a concomitant reduction of the strong fluorescence around the nucleus. Thus, SAP97 mutant mimicking the CaMKII phospho-site appears to facilitate release of the Kv4.2 channel from the internal compartments and recruitment toward the cell surface, without affecting SAP97 clustering with Kv4.2. Conversely, co-transfection of Kv4.2 with SAP97 (S39A) showed a more

SAP97-Kv4.2 Interaction in Hippocampal Neurons
pronounced perinuclear accumulation of the two proteins (Fig. 4,. No effect on Kv4.2 distribution was observed by mutation into aspartate of the two Kv4.2 CaMKII-dependent phospho-sites (6) both in Kv4.2 single transfections (data not shown) and in cotransfection with SAP97wt (Fig. 4, l-n). No significant effect on the degree of co-localization between the two transfected proteins was produced by any mutation constructs used, confirming the idea that CaMKII phosphorylation of Kv4.2 or SAP97 does not affect the co-localization between the two proteins.
To confirm the results obtained in transfected COS-7 cells and to address the role of SAP97 Ser-39 phosphorylation in modulating Kv4.2 localization in neurons, we transfected primary hippocampal cultures with GFP-SAP97wt, GFP-SAP97-S39D, or GFP-SAP97-S39A constructs (Fig. 5, a-d) (14,15). Interestingly, SAP97 staining overlapped Kv4.2 distribution in cultured hippocampal neurons (Fig. 5b); co-localization analysis revealed a high co-localization pattern between the two proteins (76.0 Ϯ 9.2%). Analysis of endogenous Kv4.2 staining in neurons transfected with GFP-SAP97wt revealed no major differences compared with untransfected cells (data not shown; for quantification see Fig. 5d) in agreement with previous observations in COS-7 cells (see Fig.  4, f-h) indicating a high co-localization between the two wild-type proteins without any specific effect on Kv4.2 distribution. On the other hand, hippocampal neurons transfected with GFP-SAP97 S39D showed an increased staining of Kv4.2 in "spine-like" structures where the Kv4.2 channel co-localizes precisely with the S39D mutant form of SAP97 (14) (Fig. 5c). In addition, quantification of the dendritic shaft versus somatic signal of Kv4.2, measured by the relative fluorescence intensity within these structures, revealed that GFP-SAP97 S39D transfected neurons displays an increased Kv4.2 fluorescent signal in dendritic structures (Fig. 5, c and d, p Ͻ 0.005, GFP-SAP97 S39D versus untransfected). Conversely, transfection of GFP-SAP97 S39A (Fig. 5a) resulted in a redistribution of Kv4.2 signal toward the cell soma with a decreased dendritic staining (Fig. 5d, **, p Ͻ 0.001, GFP-SAP97 S39A versus untransfected) suggesting that CaMKII phosphorylation of SAP97-Ser-39 phospho-site can be necessary for synaptic trafficking of SAP97 interacting proteins, such as Kv4.2.

DISCUSSION
The somatodendritic A-type potassium current, I SA , is particularly important for regulating action potential back propagation as well as the local excitability of the dendrite. Modulations in the expression and functional properties of the I SA current have been observed during long term potentiation and epilepsy, and increasing evidence suggests that abnormal regulation of these channels may be acquired during epilepsy (4), leading to an increased excitability of the pyramidal neuron dendrite that contributes to the initiation and prolongation of seizures. It is thereby of importance to understand the regulatory mechanism underlying the trafficking of key constituents of I SA , such as Kv4.2, into dendrites and spines.
Here we show that Kv4.2 is enriched in rodent purified PSDs by means of SAP97-mediated trafficking from the ER. Indeed, we have observed that Kv4.2 interacts with SAP97 and PSD-95, although in different cell compartments suggesting the binding with SAP97 as being instrumental for SAP97 trafficking to and with PSD-95 for localization in the postsynaptic compartment. Based on co-immunoprecipitation experiments, the interaction of Kv4.2 with SAP97 appears to require an intact C terminus of Kv4.2. In GST pulldown assays, we have identified the PDZ domains of SAP97 as being a key constituent for the binding to Kv4.2.
We have previously shown that CaMKII phosphorylates SAP97 and causes an enrichment of SAP97 in the postsynaptic compartment (14,15). In addition, it has been demonstrated that CaMKII phosphorylation of Kv4.2 increases the surface expression of Kv4.2 in fibroblasts or cultured hippocampal neurons without altering its biophysical properties (6). Interestingly, here we show that CaMKII-dependent phosphorylation of Kv4.2 in the two previously identified sites (6) does not affect Kv4.2 interaction with SAP97 as well as Kv4.2 distribution in transfected cells.
Interestingly, knockdown of SAP97 significantly reduces the localization of Kv4.2 in the Triton-insoluble postsynaptic compartment addressing a specific role of SAP97 in driving Kv4.2 to synaptic sites. Using site-specific mutants of SAP97 mimicking either its phosphorylation in Ser-39 by CaMKII or abolishing it, we show that the SAP97 Ser-39 phosphorylation increased the presence of endogenous Kv4.2 in dendrites compared with wild type, whereas the mutant with abolished phosphorylation site decreased it.
In addition, we dissected the intracellular biochemical pathways governing trafficking of endogenous Kv4.2 from the ER. In particular, treating neurons with caffeine, a pharmacological tool capable of activating ER-associated CaMKII and consequently driving the exit of SAP97 from the ER (14), we saw an increase in the staining of Kv4.2 in dendrites, as well as its level in a TIF fraction, resembling the composition of the postsynaptic density. On the other hand, a pharmacological strategy blocking Ca 2ϩ efflux from the ER and promoting activation of CaMKII in the spine (24) is not capable of inserting new Kv4.2 subunits in the TIF fraction.
With this view, we here confirm and expand the observation that CaMKII activation results in a direct regulation of SAP97 function; indeed, we show here that CaMKII-dependent phosphorylation of SAP97 at Ser-39 affects trafficking and distribution of SAP97-interacting proteins, i.e. Kv4.2, thus strengthening the hypothesis that phosphorylation of SAP97 entails redistribution of SAP97 complexes from the ER to synaptic compartments. In agreement with these data, Nakagawa et al. (12) recently reported that SAP97 S39A mutant shifts SAP97 balance toward multimerization and, consequently, toward a static localization. In addition, our study confirms previous data showing that SAP97 phosphorylation by CaMKII is a key step in the complex and finely tuned mechanism governing delivery of SAP97-interacting proteins, i.e. GluR1 (14), to the postsynaptic complex. Schluter et al. (29) recently identified a two isoforms of SAP97 and PSD-95. Although the ␣ isoforms promote AMPA receptor insertion in an activity-independent manner, the ␤ isoforms influence AMPA receptor-dependent synaptic strength in a CaMKII-dependent manner. Interestingly, the effect of the ␤ isoform of SAP97 was dependent of the knockdown of endogenous PSD-95. Future studies will have to address a possible differential role of the two isoforms of PSD-95 and SAP97 in combination with acute knockdown and rescue experiments. It will also be interesting to determine how the induction of synaptic plasticity influences the distribution of Kv4.2 to synapses.
Expression of PSD-95 did not change the localization of Kv4.2 in dendrites or spines in cultured neurons (data not shown). A previous report had suggested clustering of Kv4.2 in transfected fibroblasts by co-expressed PSD-95, but we did not notice any apparent differences in the localization of endogenous Kv4.2 upon expression of PSD-95 in transfected neurons. Our co-immunoprecipitation experiments clearly suggest the presence of a PSD-95-Kv4.2 complex only at synaptic sites, suggesting that SAP97 is a key protein responsible for Kv4.2 trafficking, whereas PSD-95 represents an important anchoring element for Kv4.2 once the protein has been delivered at membrane/synaptic compartment.
Finally, our study has clarified the molecular details of Kv4.2 interaction with MAGUK proteins in hippocampal neurons addressing a key role of CaMKII in the regulation of SAP97mediated localization of Kv4.2 at synaptic sites.