Slitrk2 controls excitatory synapse development via PDZ-mediated protein interactions

Members of the Slitrk (Slit- and Trk-like protein) family of synaptic cell-adhesion molecules control excitatory and inhibitory synapse development through isoform-dependent extracellular interactions with leukocyte common antigen-related receptor protein tyrosine phosphatases (LAR-RPTPs). However, how Slitrks participate in activation of intracellular signaling pathways in postsynaptic neurons remains largely unknown. Here we report that, among the six members of the Slitrk family, only Slitrk2 directly interacts with the PDZ domain-containing excitatory scaffolds, PSD-95 and Shank3. The interaction of Slitrk2 with PDZ proteins is mediated by the cytoplasmic COOH-terminal PDZ domain-binding motif (Ile-Ser-Glu-Leu), which is not found in other Slitrks. Mapping analyses further revealed that a single PDZ domain of Shank3 is responsible for binding to Slitrk2. Slitrk2 forms in vivo complexes with membrane-associated guanylate kinase (MAGUK) family proteins in addition to PSD-95 and Shank3. Intriguingly, in addition to its role in synaptic targeting in cultured hippocampal neurons, the PDZ domain-binding motif of Slitrk2 is required for Slitrk2 promotion of excitatory synapse formation, transmission, and spine development in the CA1 hippocampal region. Collectively, our data suggest a new molecular mechanism for conferring isoform-specific regulatory actions of the Slitrk family in orchestrating intracellular signal transduction pathways in postsynaptic neurons.

Synapses are asymmetric intercellular junctions that permit transmission of signals from presynaptic to postsynaptic neurons. The transmission of synaptic information is coordinated by distinct multiprotein complexes that are compartmentalized at the presynaptic active zone, synaptic cleft, and postsynaptic density 1 . In particular, synaptic cell-adhesion molecules have been recognized as key components that bidirectionally organize the transfer and processing of synaptic information 2 . These proteins are thought to not only mediate the physical alignment of pre-and postsynaptic neurons, but also to orchestrate multiple trans-cellular signaling cascades in both pre-and postsynaptic neurons, lending specific properties to synapse types. Although a variety of synaptic cell-adhesion molecules have been recently demonstrated to organize 'extracellular' synaptic adhesion pathways, how diverse extracellular signals are directionally transduced, and whether different synaptic adhesion pathways can activate distinct 'intracellular' signaling pathways, remains largely unknown. Protein phosphorylation is one of the crucial mechanisms in the regulation of neural function, and a subset of synaptic cell-adhesion molecules directly regulate aspects of this process, particularly tyrosine phosphorylation and dephosphorylation, mediated by receptor tyrosine kinases and receptor tyrosine phosphatases, respectively (reviewed in 3 ).
Slit-and Trk-like (Slitrk) family proteins, composed of six members (Slitrk1-6), are highly expressed in the central nervous system of vertebrates and are postsynaptically localized, controlling both excitatory and inhibitory synapse development [4][5][6][7] . All six members of the Slitrk family are structurally similar, particularly in extracellular sequences, which mediate interactions with three members of the leukocyte common antigen-related receptor tyrosine phosphatase (LAR-RPTP) family-PTPσ, PTPδ, and LAR-through the first cluster of leucine-rich repeat (LRR) domains 5,8,9 . Intriguingly, Slitrk family proteins control formation of distinct synapse-types in an isoform-dependent manner 10 . Although distinct extracellular interactions of Slitrks with specific LAR-RPTP family members likely shapes the establishment of specific adhesion pathways at distinct synapse types, it is plausible that unique intracellular signaling pathways underlying individual Slitrk members are also involved, in keeping with sequence divergence in intracellular sequences of Slitrk family members 11 .
Here, we report that Slitrk2, an excitatory synapse-specific Slitrk, directly binds to the PDZ (PSD-95/Dlg/ ZO-1)-containing excitatory scaffolds, PSD-95 and Shank3. To explore the functions of Slitrk2-PDZ interactions, we performed various experiments in both cultured hippocampal neurons and hippocampal CA1 pyramidal neurons in vivo, revealing that PDZ proteins positively regulate excitatory synaptic targeting of Slitrk2, as well as Slitrk2-mediated excitatory (but not inhibitory) synapse formation and transmission.

Materials and Methods
Animals. All animal experiments were approved by the Institutional Animal Care and Use Committee (IACUC) of the Daegu Gyeongbuk Institute of Science and Technology (DGIST), and have been performed accordingly. All methods were performed in accordance with the relevant guidelines and protocols (DGIST-IACUC-17122104-01). Mice (C57BL/6N) for viral injection and pregnant rats for primary cultures were purchased from Daehan Biolink.

Production of recombinant adeno-associated virus (AAV).
HEK293T cells were co-transfected with the indicated AAV vectors and pHelper and AAV1.0 (serotype2/9) capsids vectors. Seventy-two hours later, transfected HEK293T cells were collected, lysed, and mixed with 40% polyethylene glycol and 2.5 M NaCl, and centrifuged at 2000 × g for 30 min. The cell pellets were resuspended in HEPES buffer (20 mM HEPES; 115 mM NaCl, 1.2 mM CaCl 2 , 1.2 mM MgCl 2 , 2.4 mM KH 2 PO 4 ) and an equal volume of chloroform was added. The mixture was centrifuged at 400 × g for 5 min, and concentrated three times with a Centriprep centrifugal filter (Millipore) at 1,220 × g for 5 min each and with an Amicon Ultra centrifugal filter (Millipore) at 16,000 × g for 10 min. Before titering AAVs, contaminating plasmid DNA was eliminated by treating 1 μl of concentrated, sterile-filtered AAVs with 1 μl of DNase I (Sigma-Aldrich) for 30 min at 37 °C. After treatment with 1 μl of stop solution (50 mM ethylenediaminetetraacetic acid) for 10 min at 65 °C, 10 μg of protease K (Sigma-Aldrich) was added and AAVs were incubated for 1 h at 50 °C. Reactions were inactivated by incubating samples for 20 min at 95 °C. The final virus titer was measured by quantitative reverse transcription-PCR (qRT-PCR) detection of EGFP sequences and subsequent reference to a standard curve generated using the pAAV-U6-EGFP plasmid. All plasmids were purified using a Plasmid Maxi Kit (Qiagen GmbH).

Stereotaxic injection of rAAVs into mice.
Five-week-old C57BL/6 mice were anesthetized by intraperitoneal injection of Avertin (400 mg/kg body weight). Viral solutions (titers ≥ 1 × 10 11 viral genomes/ml) were injected with a NanoFil syringe (World Precision Instruments) at a flow rate of 0.1 μl/min. The coordinates used for the CA1 region of the dorsal hippocampus were AP −2.5 mm, ML ± 1.5 mm, DV + 1.5 mm (from the dura). The site at DV + 1.5 mm received a 1-μl injection. Injected mice were allowed to recover for at least 14 d following surgery prior to use in experiments.
Immunohistochemistry. Mice were transcardially perfused first with PBS and then with 4% paraformaldehyde. After post-fixation overnight, mouse brains were slowly sectioned at 40 μm using a vibratome (VT1200S; Leica) and washed with PBS. Brain sections were incubated in blocking solution containing 10% horse serum, 0.2% bovine serum albumin, and 2% Triton X-100 for 1 h at room temperature (RT), and then incubated overnight at 4 °C with primary antibodies against VGLUT1 (JK111; 1:200) or GAD65 (JK158; 1:200). After washing three times, sections were incubated with Cy3-or FITC-conjugated secondary antibodies (Jackson ImmunoResearch, West Grove, PA, USA) for 2 h at RT. Sections were then washed extensively and mounted on glass slides with Vectashield Mounting Medium (Vector Laboratories). Images were acquired using a confocal laser-scanning microscope (LSM700; Carl Zeiss).
Electrophysiology. Whole-cell voltage-clamp recordings were performed in acute mouse brain slices, as previously described 15 . Mouse brain slices were transferred to a recording chamber and perfused with a bath solution of aerated (O 2 95%/CO 2 5% mixed gas) artificial cerebrospinal fluid (aCSF) consisting of 124 mM NaCl,  www.nature.com/scientificreports www.nature.com/scientificreports/ intracellular solution was 290-300 mOsm and the pH was 7.3 (adjusted with CsOH). For mEPSC recordings, 1 μM TTX (Tocris) and 50 μM picrotoxin (Tocris) added to bath solution to block Na + currents and GABA A receptors. For mIPSC recordings, 1 μM TTX, 10 μM CNQX (Sigma Aldrich) and 50 μM D-AP-5 (Tocris) added to block Na + currents, AMPAR and NMDAR. Statistical analysis. All data are presented as means ± SEM. All experiments were repeated using at least three independent cultures, and data were statistically evaluated using a Mann-Whitney U test or Kruskal-Wallis test followed by Dunn's pairwise post hoc test, as appropriate. Prism8.0 (GraphPad Software) was used for analysis of data and preparation of bar graphs. P-values < 0.05 were considered statistically significant (individual p-values are presented in respective figure legends).

Results
The C-terminal PDZ domain-binding sequence in Slitrk2 mediates interactions with PSD-95 and Shank3. Among the six Slitrks, only Slitrk2 contains a canonical type-I PDZ domain-binding motif at its C-terminus that recognizes the PDZ domain 19 (Fig. 1A), raising the possibility that Slitrk2 interacts with intracellular PDZ domain-containing scaffolds. To test this possibility, we performed co-immunoprecipitation analyses in HEK293T (human embryonic kidney 293T) cells expressing the indicated Slitrk isoforms (Slitrk1-5) alone or coexpressing Slitrks and PSD-95 or Shank3. We found that Slitrk2 robustly and specifically interacted with PSD-95 and Shank3 (Figs. 1C,F, S1A,C). These interactions were completely disrupted by deleting the last four amino acids residues of Slitrk2 (Ile-Ser-Glu-Leu; ISQL) (Figs. 1B,D,E,G,H, S1B,D), suggesting a canonical PDZ domain-mediated interaction between Slitrk2 and the tested PDZ domain-containing proteins.

The C-terminal PDZ domain-binding sequence of Slitrk2 is required for Slitrk2 promotion of excitatory synapse development in cultured hippocampal neurons.
To investigate the functional significance of PDZ domain-mediated interactions for Slitrk2 targeting to excitatory synapses, we first transfected cultured hippocampal neurons at DIV10 with expression vectors encoding HA-tagged Slitrk2 wild-type (WT) and a Slitrk2 variant lacking the C-terminal four amino acids (ΔPBM), and immunostained the transfected neurons for the excitatory presynaptic marker VGLUT1 (vesicular glutamate transporter 1) at DIV14. Investigation of the subcellular localization of recombinant Slitrk2 variants, visualized by monitoring expression of HA immunofluorescence, revealed that Slitrk2 WT was mainly distributed to dendritic spines, whereas recombinant Slitrk2 ΔPBM showed less distribution to dendritic spines compared with Slitrk2 WT (Fig. 4A,B). In addition, the number of VGLUT1-positive dendritic spines was reduced following overexpression of recombinant Slitrk2 ΔPBM, although both Slitrk2 WT and ΔPBM exhibited comparable levels of dendritic targeting when expressed in mature cultured hippocampal neurons (Fig. 4A-D).
Next, to determine whether Slitrk2 ΔPBM also compromises the ability of Slitrk2 WT to promote excitatory synapse development in cultured hippocampal neurons, we introduced Slitrk2-specific knockdown (KD) vectors into cultured neurons at DIV8, and immunostained the transfected neurons for EGFP and various excitatory synaptic markers (VGLUT1, PSD-95, and pan-Shank) at DIV14 (Fig. 4E). As previously reported 10,12 , Slitrk2 KD significantly decreased the linear density of excitatory synaptic clusters (Fig. 4E,F). Coexpression of short hairpin RNA (shRNA)-resistant forms of Slitrk2 WT completely rescued these deficits in the numbers of excitatory synaptic clusters (Fig. 4E,F). However, coexpression of shRNA-resistant Slitrk2 ΔPBM failed to reverse the reduction in the density of excitatory synaptic clusters induced by Slitrk2 KD (Fig. 4E,F). Moreover, and in keeping with our previous observations 10,12 , Slitrk2 KD significantly decreased the number of dendritic spines, which was completely rescued by coexpression of Slitrk2 WT, but not by coexpression of Slitrk2 ΔPBM (Fig. S4). In addition, overexpression of Slitrk2 WT led to a significant increase in the number of dendritic spines in transfected neurons (Fig. 4G,H), whereas overexpression of Slitrk2 ΔPBM did not (Fig. 4G,H). These data suggest that Slitrk2 promotes the development of excitatory synapses in cultured hippocampal neurons via PDZ domain-mediated interactions.

Slitrk2 promotes excitatory synapse structure and transmission through its C-terminal PDZ-mediated interactions in CA1 hippocampal pyramidal neurons in vivo.
To test the physiological significance of Slitrk2 PDZ domain-mediated interactions in vivo, we performed immunohistochemistry using mice stereotactically injected with adeno-associated viruses (AAVs) expressing Slitrk2-targeting shRNA (sh-Slitrk2) or non-targeting control shRNA (sh-Control) into the hippocampal CA1 areas (Figs. 5A, S5A). Slitrk2 KD in vivo was validated by immunohistochemistry and immunoblotting analyses using the indicated AAV-injected mouse brain tissues (Fig. S5B,C). Quantitative immunofluorescence analyses using the excitatory synaptic marker VGLUT1 revealed a slight (but statistically non-significant) reduction in the intensity of VGLUT1 puncta in both stratum oriens (SO) and stratum radiatum (SR) layers of the hippocampal CA1 (Fig. 5B,C). In contrast, the intensity of GABAergic synaptic marker GAD65 (glutamic acid decarboxylase 65) puncta was not altered in Slitrk2-deficient neurons (Fig. 5D,E). Intriguingly, the VGLUT1 puncta intensity in Slitrk2-KD mice was increased by coexpression of shRNA-resistant Slitrk2 WT, but not by coexpression of shRNA-resistant Slitrk2 ΔPBM, despite the fact that both Slitrk2 WT and Slitrk2 ΔPBM exhibited comparable expression levels (Figs. 5B,C, S5B,C).
To corroborate these anatomical findings, we performed whole-cell electrophysiological recordings of miniature excitatory and inhibitory postsynaptic currents (mEPSCs and mIPSCs, respectively) in acutely prepared mouse brain slices infected with adeno-associated viruses (AAVs) expressing either sh-Slitrk2 or sh-Control. Consistently, we did not detect any noticeable difference in the amplitude or frequency of mEPSCs or mIPSCs in Slitrk2-deficient neurons (Fig. 6A-F) 10 . Again, coexpression of Slitrk2 WT in Slitrk2-deficient neurons led to a significant increase in the frequency (but not amplitude) of mEPSCs, whereas coexpression of Slitrk2 ΔPBM did not (Fig. 6A-C). Interestingly, the frequency of mEPSCs under these conditions exceeded that in controls, a result that may be an artifact of Slitrk2 overexpression reflecting the high level expression of Slitrk2 WT AAV constructs. Taken together, our data suggest that Slitrk2 is not required for basal synaptic transmission at either synapse type in vivo, and that Slitrk2 specifically regulates excitatory synapse organization in hippocampal CA1 neurons through interactions with PDZ domain-containing proteins.

Discussion
Recent efforts to identify a number of trans-synaptic adhesion molecules and investigate their synaptic roles has significantly contributed to our current understanding of how synapses are formed, refined, and eliminated 2 . The initial conceptualization of these molecules was as building blocks for structural organization, and some candidates in invertebrate synapses have been proposed in this context. However, in vertebrate synapses, this perspective has broadened to encompass all biological processes, including synaptic initiation, assembly, and organization of canonical signaling ensembles 2 . Strikingly, the roles of most trans-synaptic adhesion molecules expressed in mammalian synapses as signaling entities have remained largely unknown. PSD-95, by extension, the PSD-95 family of membrane-associated guanylate kinases (MAGUKs), together with gephyrin are arguably the most extensively studied synaptic proteins that exclusively expressed in excitatory data showing the effects of Slitrk2 overexpression in neurons on dendritic spine density. Data are presented as means ± SEMs from three independent experiments (n = 22-30 neurons; **p < 0.01 vs. control; nonparametric ANOVA with Kruskal-Wallis test followed by post hoc Dunn's multiple comparison test). www.nature.com/scientificreports www.nature.com/scientificreports/ and inhibitory synaptic sites [21][22][23][24][25] . These proteins are structurally modular and form multimeric complexes with neurotransmitter receptors, synaptic adhesion molecules and downstream signaling effectors, serving as scaffolding cornerstones that mediate diverse structural and functional organization. Notably, PSD-95 and gephyrin contribute to the stabilization of interacting trans-membrane proteins (NMDA-and AMPA-type glutamate receptors, and various synaptic adhesion molecules) by suppressing their surface mobility or internalization 21 . Among the domain architectural features of MAGUKs, three PDZ domains are involved in recognizing specific C-terminal motifs present in a variety of trans-membrane proteins 26 . However, to date, only neuroligin-2 (expressed exclusively at GABAergic synaptic sites) has been reported to bind gephyrin, a uniqueness that may largely reflect the lack of an identifiable gephyrin-binding consensus sequence 23 . Thus, efforts to test the putative interactions of MAGUK PDZ domains with various trans-membrane proteins have proved fruitful for describing the molecular organization of excitatory synapses 25 . A number of putative trans-synaptic adhesion molecules have been subsequently shown to interact with MAGUK family proteins, including neuroligins (NLs) 27 , SALMs 28,29 , Netrin-G ligands (NGLs) 30 , ADAM22 31 , LRRTMs 15,32 , Kirrels/Nephs 33 , and IgSF11 34 . In most cases, however, the precise physiological roles of these PDZ domain-mediated interactions are still undefined. NLs are targeted to synaptic sites, independent of PDZ domain-mediated interactions 35 ; for these proteins, targeting instead requires non-PDZ domain-mediated intracellular mechanisms (e.g., activity-dependent posttranslational modifications) 36 . In contrast, NGL-2 and IgSF11 require PSD-95 interactions for excitatory synapse localization, and mediate the physical and functional coupling of PSD-95 with other key components expressed at excitatory synaptic sites, such as AMPA-type glutamate receptors 30,34 . Although these observations suggest the compelling concept that PSD-95 matches synaptic adhesion molecules with various forms of intracellular signaling cascades to regulate excitatory synapse development in postsynaptic neurons, this hypothesis has not been systematically addressed.
Our current study sought to identify intracellular mechanisms governing the actions of Slitrks, an emerging class of postsynaptic adhesion molecules 4,7 . Taking note of the fact that Slitrk2 (among six Slitrk members) possesses a typical type I PDZ domain-binding motif at its C-terminus, we here showed that Slitrk2 directly interacts with two key PDZ domain-containing synapse organizers, PSD-95 and Shank3 (Figs. 1 and 2). We also demonstrated that Slitrk2 promotes excitatory synapse structure and transmission through these PDZ domain-mediated interactions in vitro and in vivo (Figs. 4-6). These observations suggest a tantalizing model in which interactions of Slitrk2 with PTPσ (a member of the LAR-RPTP family with specific functions at excitatory synapses) 5 transduce/propagate extracellular information into specific intracellular complexes involving PSD-95, MAGUKs, and/ or Shank3. However, our observation that other Slitrks do not interact with PSD-95 suggests that the molecular diversity of Slitrk-mediated synapse development depends on specific intracellular protein complexes (Fig. 1). Intriguingly, Slitrk2, -3 and -5 are similar to Trk receptors in that they contain a signature sequence motif (NpXY) at their juxtamembranous regions that usually serves as a binding site for adaptor proteins, such as Shc, following binding to specific neurotrophins and subsequent phosphorylation at a tyrosine residue 37 . Moreover, Slitrks contain various tyrosine residues that are conserved across all six members, hinting at the possibility that tyrosine phosphorylation could be an important regulatory mechanism for Slitrk-mediated synapse development. Prior studies have also demonstrated the significance of tyrosine phosphorylation of NL1 (Y782) in promoting recruitment of functional AMPA-type glutamate receptors and long-term potentiation, and excluding gephyrin binding 38,39 . Thus, tyrosine phosphorylation of synaptic adhesion molecules could be a universally important mechanism for establishing the balance between excitatory and inhibitory synapses during synapse formation, assembly, and stabilization.
Slitrk2 KD in cultured hippocampal neurons clearly reduced the number of excitatory synapses (Fig. 4 10,19 ;), whereas Slitrk2 KD in mouse hippocampal CA1 regions did not significantly affect the density of excitatory www.nature.com/scientificreports www.nature.com/scientificreports/ synapses or excitatory synaptic transmission (Figs. 5 and 6). This apparent phenotypic difference between cultured neurons and in vivo neurons caused by the loss of Slitrk2 has been similarly documented in loss-of-function analyses of other synaptic proteins (e.g., neuroligins) 17,18,40 . Although the present study pinpointed a crucial contribution of the C-terminal PDZ-binding motif, which is unique to Slitrk2, in promoting excitatory synapse development, extracellular mechanisms shared by other excitatory synaptic Slitrks (i.e., Slitrk1, Slitrk4 and Slitrk5) may differentially compensate for the effect of Slitrk2 loss. Moreover, loss-of-function analyses of Slitrk2 were performed on sparsely transfected neurons, whereas AAV-mediated shRNA delivery downregulated Slitrk2 levels in most hippocampal CA1 neurons, suggesting the possibility that intercellular differences in Slitrk2 level are critical for excitatory synapse development, as previously reported for neuroligin-1, ephrin-B3 and TrkC 17,21,41 . These possibilities should be tested using conditional Slitrk2 mice to avoid possible developmental compensation.
Our demonstration that Slitrk2 interacts with Shank3 was unexpected, given the differential laminar organization of PSD scaffolds 42 . PSD-95 is spatially localized beneath PSD membranes, whereas Shank3 lies on the deeper cytoplasmic side of the PSD 42,43 . Because Slitrk2 has a relatively short cytoplasmic region (~203 residues), the direct interaction of Slitrk2 with Shank3 in vivo could be spatially constrained, despite our analyses showing robust interactions of Slitrk2 with Shank3 ( Figs. 1 and 2). Future follow-up studies should address the differential roles of Slitrk2/PSD-95 and Slitrk2/Shank3 complexes in regulating excitatory synapse development in vivo.

Data availability
The datasets generated and/or analyzed during the current study are available from the corresponding author upon request.