[1] 159w Autism-related Shank1, Shank2, and Shank3 are major postsynaptic scaffold proteins of excitatory glutamatergic synapses. A few studies, however, have already indicated that within a neuron, the presence of Shank family members is not limited to the postsynaptic density. By separating axons from dendrites of developing hippocampal neurons in microfluidic chambers, we show that RNA of all three Shank family members is present within axons. Immunostaining confirms these findings as all three Shanks are indeed found within separated axons and further co-localize with well-known proteins of the presynaptic specialization in axon terminals.Therefore, Shank proteins might not only serve as postsynaptic scaffold proteins, but also play a crucial role during axonal outgrowth and presynaptic development and function. This is supported by our findings that shRNA-mediated knockdown of Shank3 results in up-regulation of the NMDA receptor subunit GluN1 in axon terminals. Taken together, our findings will have major implications for the future analysis of neuronal Shank biology in both health and disease.

[1] 106w We first analyzed Shank expression in primary hippocampal cultures at distinct time points of neural development. RNA analysis revealed that each family member was expressed from the first day in vitro (DIV1) onwards and exhibited an individual RNA expression profile (Fig. 1a, upper panel). We next over-expressed GFP-Shank1, GFP-Shank2, and GFP-Shank3 fusion proteins in HEK293T cells and confirmed by western blot analysis that the anti-Shank antibodies used in this study were not cross-reactive (Fig. 1b). Corresponding to our RNA data (Fig. 1a, upper panel), each family member was also present on the protein level and exhibited an individual expression profile over time (Fig. 1a, lower panel).

[2] 120w We next analyzed subcellular localization of each Shank family member at distinct stages of development of primary hippocampal neurons in culture (Dotti et al. 1988). In line with our expression analysis, all three Shanks were already detected in stage 1 neurons at DIV1 (Fig. 2a). In unpolarized stage 2 neurons, each Shank appeared in both soma and neurites and accumulated within GAP43-positive growth cones (Fig. 2b). In polarized stage 3 neurons, besides its presence in soma, neurites, and growth cones (data not shown), each Shank also appeared within the Tau-positive neurite that had developed to become the axon. Interestingly, Shank immunoreactivity was particularly strong within the structure that should later become the axon terminal or presynaptic specialization, respectively (Fig. 2c).

[3] 166w Shank1, Shank2 and Shank3 are expressed in axons and predominantly localized in axon terminals and presynaptic specializations To spatially separate axons from the somatodendritic compartment of primary hippocampal neurons, we established a microfluidic chamber system according to a previously described protocol (Taylor et al. 2005) (Fig. 3a). We first isolated RNA from the axonal compartment and demonstrated that RNA of each Shank was present within axons at DIV7 (Fig. 3b), a time point of neural development in culture when the axon has already matured and when its terminal could in principle serve as presynaptic specialization (Fletcher et al. 1994). Next we immunostained the microfluidic chambers with our anti-Shank antibodies and showed that all three Shank proteins were present in Tau-positive axons at DIV7 and DIV14, respectively (Fig. 3c). Further immunostaining revealed that each Shank co-localized with both the presynaptic scaffold protein Bassoon and the presynaptic vesicular glutamate transporter VGluT1 within axon terminals at DIV7 and DIV14 in the axonal compartment of the chamber system (Fig. 3d).

[1] 96w Due to the increasing relevance of SHANK mutations for the pathogenesis of several neurodevelopmental disorders, we crucially need a deeper and more comprehensive understanding of basic Shank neurobiology to efficiently study and understand the underlying pathomechanisms in appropriate model systems (Schmeisser 2015). Although unanimously classified as 'master scaffolds of the PSD' (for most recent review on the Shanks see Sala et al. 2015), some studies have already implicated that the subcellular localization of Shanks in neurons is not limited to the postsynaptic compartment (Du et al. 1998;Durand et al. 2012;Grabrucker et al. 2014;Wang et al. 2014).

[2] 207w The data presented in this study clearly demonstrate that all three Shanks, which exhibit individual expression profiles in primary hippocampal neurons throughout development, are found in growth cones of young unpolarized neurons. Intriguingly, in polarized neurons at later stages, they are expressed and localized in axons and accumulate within axon terminals filled with classical presynaptic proteins ready to form glutamatergic synaptic contacts in culture (Grabrucker et al. 2009). The anatomical and functional importance of these findings is supported by the following three facts: (i) Similar data have been obtained for both the NMDA and a-amino-3-hydroxy-5-methylisoxazole-4-propionate (AMPA) receptor (NMDAR, AMPAR) in developing hippocampal neurons (Schenk et al. 2003;Wang et al. 2011;Gill et al. 2015), (ii) Both the NMDAR and AMPAR are essentially clustered and regulated by the Shank platform in the PSD thus making the existence of presynaptic NMDAR/ AMPAR-Shank complexes highly probable and (iii) The major and converging molecular finding in various mouse models and human iPSC-derived neurons modeling Shankopathies is a dysregulation of both NMDAR and AMPAR levels and physiology at glutamatergic synapses (Jiang and Ehlers 2013;Shcheglovitov et al. 2013). Most interestingly in this context, we provide the first evidence that Shank3 is indeed involved in the modulation of NMDA receptor levels also at axon terminals.

[3] 160w Future studies on basic Shank cell biology and model systems related to human SHANK mutations alike should therefore include the analysis of axonal development and presynaptic function to reveal hitherto undiscovered neurites and growth cones. Growth cones were visualized by anti-GAP43 immunostaining. (c) Immunostaining of polarized stage 3 neurons revealed localization of each Shank family member within the Tau-positive axon as indicated. Note Shank accumulation within the axon terminal (a-c). Neurons were identified by anti-b3-Tubulin immunostaining. Scale bars: 10 lm. alterations and to better understand disease-relevant pathomechanisms. At this moment, we can only speculate about the precise axonal and/or presynaptic role of the Shank family. For example, it will be intriguing to study if the axonal presence of Shanks is an isoform-specific phenomenon. Furthermore, it will be of significant interest to analyze in detail the ultrastructural localization of Shanks in axons and to compare the axonal/presynaptic Shank interactome with the dendritic/postsynaptic one, so as to possibly determine functional molecular differences.

[4] 36w Taken together, our findings on the presence of Shank proteins in axons and presynaptic specializations will surely open up novel avenues for a better and more comprehensive understanding on Shank biology in both health and disease.

[1] 76w HEK293T cells transfected with green fluorescent protein (GFP)-Shank constructs or primary hippocampal neurons were lysed in Triton X-100 lysis buffer (150 mM NaCl, 50 mM TrisHCl, 1% Triton X-100, pH 8.0, protease inhibitor mix, Roche Applied Science, Germany). Protein concentrations were determined by the Bradford method, 15 lg of protein was loaded per lane for sodium dodecyl sulfate-polyacrylamide gel electrophoresis and western blot analysis was conducted as described previously (Proepper et al. 2011;Cochoy et al. 2015).

[2] 25w Preparation and transfection of primary hippocampal neurons from rat was performed as described previously with minor modifications (Schmeisser et al. 2009(Schmeisser et al. , 2013)).

[3] 26w Immunocytochemistry was performed as described previously (Grabrucker et al. 2011b;Schmeisser et al. 2012a). The actin cytoskeleton was visualized in some experiments using Alexa Fluor 568 phalloidin.

[4] 106w The microfluidic chamber system has been established according to a previously described protocol (Taylor et al. 2005). A quantity of 5 lL of primary hippocampal cell suspension was directly placed into the somal compartment of a Dulbecco's modified Eagle's medium-coated microfluidic chamber system (250 lm) followed by incubation for 15 min at 37°C to allow cell attachment. Neurobasal medium complemented with B-27 supplement, 2 mM L-glutamine and penicillin/streptomycin at 100 U/ml was added to both somal and axonal compartments. Primary hippocampal neurons were cultured in the microfluidic chamber system until day in vitro (DIV) 7 or DIV14 and further processed for immunocytochemistry or RNA analysis, respectively.

[5] 92w To isolate pure axonal RNA from microfluidic chambers, the medium of the somal compartment was aspirated continuously to avoid contamination of the axonal compartment with somal RNA. Incubation of the axonal compartment or of conventional primary hippocampal cultures with lysis buffer for 30 sec was followed by total RNA isolation with RNeasy â Kits (Qiagen, Germany). Quantitative real-time PCR (qRT-PCR) was performed using the Rotor-Gene TM SYBR â -Green RT-PCR FAST Kit (Qiagen) as previously described (Schmeisser et al. 2012b). For all genes analyzed, we used commercially available QuantiTect primers from Qiagen.

[6] 46w Statistical analysis was performed by one-way ANOVA with Bonferroni post hoc test. Ctrl vector: n = 12 neurons, Scrambled: n = 12 neurons, shShank3: n = 16 neurons from three independent cultures. **p < 0.01 (Ctrl vector vs. shShank3), ## p < 0.01 (Scrambled vs. shShank3).

[1] 90w The Shank (or ProSAP) family comprises three proteins: Shank1, Shank2, and Shank3. They were originally identified as a novel family of postsynaptic density (PSD) proteins binding to NMDA receptor/PSD-95/GKAP (Guanylate kinase-associated protein) and mGluR-Homer complexes (Boeckers et al. 1999a,b;Naisbitt et al. 1999;Tu et al. 1999;Yao et al. 1999). A special interest in Shank has emerged in recent years due to the identification of mutations in all three SHANK genes in a significant amount of individuals with neurodevelopmental disorders, predominantly autism (Grabrucker et al. 2011a;Leblond et al. 2014;Sala et al. 2015).

[2] 215w Most of the studies addressing Shank biology in neurons strongly support an essential role of these proteins for the postsynaptic protein scaffold (Boeckers et al. 2002;Verpelli et al. 2012). Subcellular localization of all three Shanks within the PSD has been characterized in detail by biochemical analysis of PSD fractions, immunohistochemistry, and electron microscopy (Boeckers et al. 1999a;Lim et al. 1999;Naisbitt et al. 1999;Tu et al. 1999;Bockmann et al. 2002;Tao-Cheng et al. 2010, 2015). Furthermore, direct molecular interactions include other major PSD scaffold proteins, including GKAP or Homer and various cytoskeleton-associated proteins of the dendritic spine (Grabrucker et al. 2011a;Verpelli et al. 2012). However, a few studies have already provided evidence that the localization of Shank within neurons is not limited to the PSD. It has recently been shown, for example, that Shank3 is also present within neuronal nuclei, although it is still unclear if this is an isoform-specific effect or if Shank3 is processed in a stimulus-specific manner to shuttle to the nucleus (Grabrucker et al. 2014;Wang et al. 2014). Interestingly, Shank2 and Shank3 are further present in growth cones of young neurons (Du et al. 1998;Durand et al. 2012). To follow up the latter phenomenon, we characterized the spatial distribution of all three Shanks in primary hippocampal neurons throughout neural development in more detail.

[3] 52w Pregnant rats (Sprague-Dawley) were purchased from Janvier Labs, France. All animal experiments were approved by the review board of the Land Baden Wurttemberg, Permit Number Nr. O.103 and performed in compliance with the guidelines for the welfare of experimental animals issued by the Federal Government of Germany and the Max Planck Society.

[4] 87w Anti-Shank2 and anti-Shank3 antibodies have been described previously (Schmeisser et al. 2012b). The following antibodies were purchased from commercial suppliers: anti-b3-Tubulin (Covance, Princeton, NJ, USA), anti-Bassoon (Enzo Life Sciences, Farmingdale, NY, USA), anti-GAP43 (Millipore Corporation, Bedford, MA, USA), anti-GluN1 (Synaptic Systems, G€ ottingen, Germany), anti-MAP2 (Aves Laboratories, OR, USA), anti-Shank1 (Novus Biologicals, Littleton, CO, USA), anti-Tau (Millipore) and anti-VGluT1 (Synaptic Systems). The Shank cDNA constructs (Gessert et al. 2011) and the targeting sequence of the pSUPER Shank3 shRNA construct (Roussignol et al. 2005) have been described elsewhere.

[5] 99w To obtain first insight into the role of axonal/presynaptic Shank, we focused on Shank3, the Shank family member most relevant to neurodevelopmental disorders (Leblond et al. 2014). Based on the fact that Shank3 is involved in the modulation of GluN1 levels (Duffney et al. 2013(Duffney et al. , 2015;;Shcheglovitov et al. 2013) and that both Shank3 and GluN1 are indeed located within the axon (Fig. 4a), we then analyzed GluN1 immunoreactivity within axon terminals after Shank3 knockdown. Intriguingly, we found a strong upregulation of the GluN1 signal (Fig. 4b) implicating that Shank3 modulates NMDA receptor levels at axon terminals.