Cypin regulates K63-linked polyubiquitination to shape synaptic content

Abstract

An open question in neuroscience is how protein posttranslational modifications regulate synaptic site targeting. Polyubiquitination plays a role in proteasome-mediated protein turnover; however, additional functions for specific types of polyubiquitin linkages have been identified. One type of polyubiquitination, K63-polyubiquitin (K63-polyUb), has been studied for its role in signal transduction within the context of cancer, but little has been done to uncover its role regarding neuronal and synaptic function. Here, we report an emerging function for the cytosolic PSD-95 interactor, cypin, in the regulation of synaptic content by K63-polyUb during neuronal development in vitro and in adult mice in vivo. We identify cypin-promoted K63-polyUb on postsynaptic proteins and also find an important role for cypin in presynaptic function. Our work demonstrates that cypin-promoted changes to K63-polyUb can regulate synaptic content and function on both sides of the synapse, adding important insight into basic mechanisms of neuronal signaling.

K63-polyubiquitin chains enable cypin protein to simultaneously modify pre- and postsynaptic content.

INTRODUCTION

Cognitive tasks, such as learning and memory, require rapid changes to proteins at synapses (1, 2). These changes may include new synthesis of proteins, degradation of proteins by the proteasome, lysozymes, or autophagy, or trafficking of proteins away from and to the synapse, all of which can be regulated by neuronal activity (3). Ubiquitination is a common mechanism underlying these changes, and at first, it was thought only to regulate protein turnover, which in turn modulates synaptic remodeling (4). Most of the synaptic remodeling that takes place at the postsynaptic densities of excitatory synapses relies on the polyubiquitination of proteins (5). However, it has recently become clear that different types of polyubiquitin tags target proteins for different purposes [reviewed in (6)]. Specifically, K48-polyubiquitin (K48-polyUb) linkages are responsible for promoting protein degradation by the proteasome (7), while K63-polyUb linkages not only target proteins for degradation but are also involved in the DNA damage response (8), control protein kinase activation (9–12), mediate protein trafficking (13), and regulate endocytosis and autophagy (14–16). The proteasome has low affinity for K63-polyUb chains, suggesting that this type of polyubiquitination has little effect on proteasome-mediated protein turnover (17–19).

The K63-polyUb linkage has mostly been studied in the context of cancer, and not much is known about the function of K63-polyUb in neurons and, specifically, at synapses. However, a role for K63 polyubiquitination in postsynaptic density 95 (PSD-95) function has recently been reported. K63-polyUb linkages increase the scaffolding efficiency of PSD-95 (20), and cold exposure increases this type of polyubiquitination on PSD-95 in the hippocampi of mice (21). Furthermore, knockout of CYLD lysine 53 deubiquitinase in mice results in aberrations in dendritic morphology and function and the appearance of behavioral phenotypes similar to autism (22). In addition, AMPA receptors are tagged with K63-polyUb in vivo and can be deubiquitinated by ubiquitin-specific peptidase 46 (USP46) (23). There are no reports beyond these that identify signaling molecules that regulate K63-polyUb linkages on synaptic proteins other than ubiquitin ligases, such as ubiquitination factor 4A (UBE4A), which has been implicated in developmental delay and intellectual disability (24).

In the current study, we identified cypin (cytosolic PSD-95 interactor) as a regulator of the levels of K63-polyUb–linked proteins in neurons, specifically at the synapse of rodent hippocampal neurons. We studied both developing neurons in culture and adult male mice and found similar, but distinct, effects of cypin. Cypin overexpression decreases proteasome activity by binding to the β7 proteasome subunit and altering the levels of proteasome subunits in neurons in vitro and proteasomal composition at synapses in vivo. Since proteins that are targeted to the proteasome are tagged with polyubiquitin, we examined changes to two different types of these linkages, K48 and K63, and found that cypin regulates K63 polyubiquitination of proteins in neurons both in vitro and in vivo. In line with these data, cypin regulates synaptogenesis signaling and protein ubiquitination pathways and, specifically, UBE4A, which promotes K63-polyUb linkages. Up-regulation of cypin protein expression in vivo significantly increases the levels of PSD-95, the N-methyl-d-aspartate (NMDA) receptor subunit GluN2A, and the AMPA subunit GluR1, but it had no effect on the levels of GluN2B. Proteomic analysis of K63-polyubiquitinated proteins at synapses in vivo with cypin overexpression and knockout identifies both pre- and postsynaptic proteins, with a greater emphasis on presynaptic processes, including synapse organization and transport. Thus, we report a previously unidentified pathway by which cypin promotes changes to synaptic content and presynaptic content, via K63 polyubiquitination of proteins.

RESULTS

Cypin binds to the β7 subunit of the proteasome and inhibits proteasome activity

To identify proteins that interact with cypin, we performed a yeast two-hybrid screen using cypin lacking the last four amino acids, which bind to PDZ domains in other proteins, as bait. Inclusion of the C-terminal tail of cypin would result in the identification of a number of proteins already known to be interactors. We identified the β7 subunit of the proteasome (PSMB4) and confirmed this interaction with cypin by immunoprecipitation of exogenous proteins in human embryonic kidney (HEK) 293T cells (fig. S1A). To determine whether cypin regulates proteasome activity, we transfected HEK293T cells with plasmids encoding mRFP or cypin-mRFP and proteasome sensors tagged with green fluorescent protein (GFP) reporters (25). We used sensors PzsGreen, which is degraded regardless of ubiquitin (Ub) conjugation (26), and UbG76V, which is degraded only when ubiquitinated (27). Overexpression of cypin resulted in the accumulation of both sensors (fig. S1, B and C), indicating a decrease in proteasome activity regardless of ubiquitination. To further confirm that cypin overexpression regulates proteasome activity, we measured chymotrypsin-, trypsin-, and caspase-like activity in extracts from neuronal cultures using fluorogenic substrates. Cypin overexpression resulted in a decrease of all three catalytic activities of the proteasome (fig. S1D), confirming that cypin can regulate proteasome activity.

Since cypin overexpression results in decreased proteasome activity, we predicted that this would result in the accumulation of ubiquitinated proteins. We cotransfected HEK293T cells with plasmids encoding mRFP or cypin-mRFP and hemagglutinin (HA)–tagged Ub (HA-Ub) and observed that cypin overexpression significantly increases the levels of HA-Ub–tagged proteins (fold change: mRFP: 1.00 ± 0.01; cypin-mRFP: 1.39 ± 0.04; P < 0.0001, Student’s t test, n = 9 replicates) and, unexpectedly, free HA-Ub (fold change: mRFP: 1.00 ± 0.14; cypin-mRFP: 2.07 ± 0.33; P = 0.0146, Student’s t test, n = 6 replicates) (fig. S2A). We also confirmed these increases by probing for endogenous Ub (fold change: mRFP: 1.00 ± 0.02; cypin-mRFP: 1.20 ± 0.04; P < 0.001, Student’s t test, n = 9 replicates) (fig. S2B). As a control, we probed the blots for β-actin and observed no changes (fig. S2, A and B). These results suggest that cypin overexpression increases the levels of ubiquitinated proteins and are consistent with the reduced proteasome peptidase activity observed in cells overexpressing cypin.

In addition, to determine whether cypin-promoted changes to protein ubiquitination are dependent on guanine deaminase (GDA) activity, we cotransfected the cells with plasmids encoding HA-tagged ubiquitin and either mRFP, cypin-mRFP, or a mutant version of cypin lacking GDA activity (cypinΔ76-84-mRFP) (28). We found that cypin-promoted increases in the level of ubiquitinated proteins are not dependent on cypin’s GDA activity (fig. S2C).

Cypin promotes K63 polyubiquitination of proteins in developing neurons

Cypin plays an important role in neuronal function, including regulation of dendritic patterning, synaptic targeting of PSD-95 family members, spine development, neuronal activity, and circuit properties (28–31); thus, we overexpressed or knocked down cypin in primary rodent cortical cultures by transduction with lentiviral particles encoding GFP or cypin-GFP for cypin overexpression (3.30 ± 0.4-fold increase in endogenous levels; n = 12 replicates) or scrambled or cypin short hairpin RNA (shRNA) for cypin knockdown (0.17 ± 0.01 of endogenous levels; n = 12 replicates) (Fig. 1A). We overexpressed cypin since we previously observed a 10-fold increase in expression when cultured neurons are activated with KCl (28) and a 25% increase in cypin levels at 1 day posttraumatic brain injury in mice (32). We also knocked down cypin to assess whether overexpression could have dominant negative effects; however, hippocampal neuronal cultures do not express high levels of cypin (29, 33). We unexpectedly find that cypin had the opposite effect in functional neurons, in contrast to unrelated HEK293T cells, as the levels of ubiquitinated proteins were unaffected when cypin was overexpressed (Fig. 1B). However, cypin knockdown resulted in a significant decrease in the levels of ubiquitinated proteins [fold change: GFP: 1.00 ± 0.01; cypin-GFP: 1.00 ± 0.02, nonsignificant; scrambled: 1.00 ± 0.01; cypin shRNA: 0.88 ± 0.03; P < 0.001, one-way analysis of variance (ANOVA) followed by Tukey’s multiple comparisons test; n = 12 culture wells] (Fig. 1B). Cypin overexpression resulted in increased levels of free-Ub (fold change: GFP: 1.00 ± 0.04; cypin-GFP: 1.35 ± 0.04, P < 0.001; one-way ANOVA followed by Tukey’s multiple comparisons test; scrambled: 1.00 ± 0.03; cypin shRNA: 0.96 ± 0.08; not significant; n = 9 culture wells) (Fig. 1C), similar to that observed in HEK293T cells (fig. S2A). As neurons are functionally and morphologically different from HEK293T cells, we further investigated whether cypin overexpression regulates different types of ubiquitin linkages with distinct functions in neurons. Cypin overexpression decreased the levels of K48-polyUb–linked proteins (fold change: GFP: 1.00 ± 0.01; cypin-GFP: 0.90 ± 0.02; scrambled: 1.00 ± 0.01; cypin shRNA: 1.10 ± 0.03, both P < 0.01, one-way ANOVA followed by Tukey’s multiple comparisons test; n = 12 culture wells) (Fig. 1D) and increased the level of K63-polyUb–linked proteins (fold change: GFP: 1.00 ± 0.01; cypin-GFP: 1.18 ± 0.03, P < 0.01; scrambled: 1.00 ± 0.02; cypin shRNA: 0.87 ± 0.05, P < 0.05, one-way ANOVA followed by Tukey’s multiple comparisons test, n = 12 culture wells) (Fig. 1E). Cypin knockdown had the opposite effect (Fig. 1, C and D). Our data demonstrate that assessment of bulk Ub levels may result in the lack of detection of critical changes in pools of different Ub linkages. Furthermore, as a control, we probed the blots for β-actin and observed no changes (Fig. 1B). We further validated these results by performing ubiquitin enrichment studies using glutathione S-transferase (GST)–tagged ubiquitin-binding domains that preferentially bind to either K48-polyUb chains or K63-polyUb chains. K63-polyUb chains increased with cypin overexpression (fold change: GFP: 1.00 ± 0.10; cypin-GFP: 1.52 ± 0.20, P < 0.05, Student’s t test, n = 12 culture wells) (Fig. 1F) and decreased with cypin knockdown (fold change: scrambled: 1.00 ± 0.05; cypin shRNA: 0.74 ± 0.06, P < 0.01, Student’s t test, n = 9 culture wells) (Fig. 1G); however, we did not observe effects on K48-polyUb linkages with cypin overexpression (fig. S3A) or knockdown (fig. S3B). Our results suggest that cypin specifically promotes K63 polyubiquitination of proteins in neurons.

Fig. 1. Cypin promotes K63-linked polyubiquitination in developing neurons.

(A) Schematic of the experimental design (B and C) Western blot analysis of total ubiquitin and free ubiquitin levels in extracts of primary rat cortical cultures with cypin overexpression (n = 9 to 12 culture wells) or knockdown (n = 9 to 12 culture wells). Levels of total Ub, free ubiquitin, and β-actin normalized to total protein stain (TPS) and GFP (control for overexpression) or scrambled shRNA (scramble; control for knockdown). GST-UBD, glutathione S-transferase–tagged ubiquitin-binding domains. (D and E) Western blot analysis of K48-polyUb and K63-polyUb levels in extracts of primary rat cortical cultures with cypin overexpression (n = 12 culture wells) or knockdown (n = 12 culture wells). Levels of K48-polyUb or K63-polyUb are normalized to TPS and GFP or scrambled. (F and G) K63-polyUb chain capture and subsequent Western blot analysis of total Ub in extracts of primary rat cortical neuronal cultures with cypin overexpression (n = 9 to 12 culture wells) or knockdown (n = 9 to 12 culture wells). K48-polyUb or K63-polyUb levels normalized to TPS and GFP or scramble. *P < 0.05, **P < 0.01, and ***P < 0.001 compared to control as determined by one-way ANOVA followed by Tukey’s multiple comparisons test [(A) to (D)] or two-tailed Student’s t test [(E) and (F)]. Note that experiments in (E) and (F) were performed and run separately and, thus, analyzed separately. Wells from neuronal cultures from the same mother are represented as black, red, green, or blue. Mean ± SEM is shown for all graphs. LV, lentivirus; IB, immunoblot. Schematic created in BioRender. Gandu, S. (2025) https://BioRender.com/g92x639.

Cypin alters the levels of proteasome subunits in developing neurons

Since cypin modulates K63-polyUb linkages, which can promote proteasome turnover (34), we investigated whether cypin contributes to proteasome degradation. We transduced neuronal cultures with lentiviral particles to overexpress or knock down cypin. We found that cypin overexpression significantly reduces the levels of 20S α (1, 2, 3, 5, 6, and 7) proteasome subunits, whereas cypin knockdown has no impact on these subunits (fold change: GFP: 1.00 ± 0.05; cypin-GFP: 1.16 ± 0.04, P < 0.01; scrambled: 1.00 ± 0.03; cypin shRNA: 0.88 ± 0.04, not significant, one-way ANOVA followed by Tukey’s multiple comparisons test, n = 12 culture wells) (fig. S4A). Cypin overexpression or knockdown does not affect 19S regulatory particle subunits (fig. S4A). As a control, we probed the blots for β-actin and observed no changes.

Next, we transfected HEK293T cells with plasmids encoding either mRFP or cypin-mRFP and performed proteasome complex enrichment to determine whether cypin interacts with the proteasome complex. We found that cypin does interact with the proteasome complex (fig. S4B). Our findings support the idea that cypin inhibits proteasome activity by a mechanism that includes 20S subunits and, potentially, proteasome assembly.

Cypin overexpression regulates UBE4A, a K63 polyubiquitination enzyme, in developing neurons

To identify the mechanism through which cypin overexpression increases K63-polyUb linkages, we performed a proteomic screen. We identified proteins that are altered with cypin overexpression in vitro and focused on those that play a role in the protein ubiquitination process. We transduced neuronal cultures on day in vitro 19 (DIV9) with lentivirus encoding GFP or cypin-GFP and processed the cultures on DIV14 (Fig. 2A). As expected, principal components analysis grouped the identified proteins into clusters of similarity within replicates, but the clusters differed between control and cypin overexpression (Fig. 2B). Proteins that are significantly altered when cypin is overexpressed (Fig. 2, C and D) were used for functional analysis using QIAGEN Ingenuity Pathway Analysis (IPA). Upon filtering the pathways for neuronal-specific process and protein metabolism, we found that the synaptogenesis signaling and protein ubiquitination pathways are regulated by cypin overexpression (Fig. 2E). Upon further investigation, we identified ubiquitination factor E4A (UBE4A) as significantly increased with cypin overexpression (Fig. 2F). UBE4A promotes K63-polyUb linkages (35), which are specifically increased when cypin levels are increased. To validate these results, we performed Western blot analysis on extracts from neuronal cultures with cypin overexpression or knockdown. Cypin overexpression significantly increased UBE4A but had no impact on the levels of UBE2J1, a ubiquitin-conjugating enzyme (fold change: GFP: 1.00 ± 0.06; cypin-GFP: 1.26 ± 0.06, P < 0.01; scrambled: 1.00 ± 0.03; cypin shRNA: 0.97 ± 0.05, not significant, one-way ANOVA followed by Tukey’s multiple comparisons test, n = 12 culture wells) (Fig. 2G), whereas cypin knockdown had no effect on the levels of UBE4A and Ube2j1 (Fig. 2G). As a control, we probed the blots for β-actin and observed no changes. These data suggest that cypin-mediated up-regulation of UBE4A may be responsible for increased K63 polyubiquitination of proteins.

Fig. 2. Cypin increases UBE4A levels and promotes K63 polyubiquitination in developing neurons.

(A) Schematic representation of experimental workflow used to identify changes in the proteome with cypin overexpression in primary rat cortical cultures transduced with lentiviral particles encoding GFP (control) or cypin-GFP. (B) Principal components analysis shows clusters of similarity within replicates but not between experimental conditions. Red circles (1 and 2), untransfected; blue dash (3 to 5), epoxomicin treated; red asterisks (6 and 7), GFP; blue squares (8 to 10), cypin-GFP. (C) Volcano plot showing significantly regulated molecules with cypin overexpression versus GFP (Ctrl). Four hundred twenty-nine proteins were identified to be significantly different with cypin overexpression. (D) Heatmap of cypin overexpression (OE)–mediated changes in the proteome. (E and F) Functional analysis of proteins in (D) using IPA uncovers the synaptogenesis signaling and protein ubiquitination pathways as the top two altered pathways with cypin overexpression. UBE4A is up-regulated with cypin overexpression. MAGUK, membrane-associated guanylate kinase; ATPase, adenosine triphosphatase. (G) Western blot analysis of UBE4A and UBE2J1 levels in extracts of primary rat cortical cultures with cypin overexpression (n = 12 culture wells) or knockdown (n = 12 culture wells). UBE4A and UBE2J1 levels are normalized to GFP or scrambled shRNA (scrambled). **P < 0.01 compared to control as determined by one-way ANOVA followed by Tukey’s multiple comparisons test. Wells from neuronal cultures from the same mother are represented as black, red, green, or blue. Mean ± SEM is shown for all graphs. β-Actin serves as a negative control. Schematic created in BioRender. Gandu, S. (2025) https://BioRender.com/ldrwnhp.

Cypin regulates NMDA receptor levels and PSD-95 K63 polyubiquitination

Pathway analysis revealed that the synaptogenesis signaling pathway is activated with cypin overexpression. To identify changes to proteins involved in this pathway, we transduced neurons with lentivirus to overexpress or knock down cypin. GluN2A (fold change: GFP: 1.00 ± 0.12; cypin-GFP: 1.40 ± 0.11, P < 0.05; scrambled: 1.00 ± 0.05; cypin shRNA: 0.98 ± 0.06, not significant, one-way ANOVA followed by Tukey’s multiple comparisons test, n = 9 culture wells) and PSD-95 (fold change: GFP: 1.00 ± 0.02; cypin-GFP: 1.25 ± 0.04, P < 0.0001; scrambled: 1.00 ± 0.01; cypin shRNA: 1.05 ± 0.04, not significant; one-way ANOVA followed by Tukey’s multiple comparisons test, n = 12 culture wells) levels significantly increase with cypin overexpression (Fig. 3A). These results suggest that cypin regulates glutamate receptor signaling at the synapse by altering the levels of GluN2A receptor subunits. As a control, we probed the blots for β-actin and observed no changes. Cypin overexpression also increased the levels of PSD-95 (Fig. 3A), a key scaffolding protein at the synapses (36), and we asked whether this may be due to increased K63-polyUb. As a preliminary study for our in vivo experiments, we performed coimmunoprecipitation and K63-polyUb pull-down assays in HEK293T cells cotransfected with plasmids encoding mRFP or cypin-mRFP, PSD-95–GFP, and HA-Ub. Cypin overexpression led to the accumulation of ubiquitinated PSD-95 (Fig. 3B), and since PSD-95 is not endogenously expressed in HEK293T cells, our results suggest that cypin regulates total PSD-95 protein levels via a posttranslational mechanism.

Fig. 3. Cypin regulates synaptic protein levels and K63 polyubiquitination of PSD-95 in vitro.

(A) Western blot analysis of GluN2A, GluN2B, GluR1, PSD-95, and β-actin levels in extracts of primary rat cortical cultures with cypin overexpression (n = 12 culture wells) or knockdown (n = 12 culture wells). Protein levels are normalized to TPS and GFP or scrambled shRNA (scrambled). β-Actin serves as a negative control. (B) Coimmunoprecipitation (co-IP) of HA-Ub and PSD-95–GFP from extracts of HEK293T cells expressing mRFP (n = 3 independent replicates) or cypin-mRFP (n = 3 independent replicates). (C) Western blot analysis of eluates from K63-polyubiquitin chain capture from HEK293T cells coexpressing mRFP (n = 3 independent replicates) or cypin-mRFP (n = 3 independent replicates) and PSD-95–GFP. K63-polyUb levels are normalized to TPS, where the immunoglobulin G (IgG) bands of the antibody used in the co-IPs is detected and then to mRFP control. (D) Western blot analysis of PSD-95 levels in lysates from HEK293T cells overexpressing mRFP (n = 3 independent replicates) or cypin-mRFP (n = 3 independent replicates) and PSD-95–GFP (WT), PSD-95–GFP–K544R (impairs K48- and K63-polyUb), PSD-95–GFP–K558R (impairs K63-polyUb), or PSD-95–GFP–K703R (impairs K48-polyUb). Protein levels are normalized to TPS and mRFP. (E) Intracellular Ca²⁺ levels were measured using detection of Fluo-4 fluorescence in neurons overexpressing mCherry (n = 30 independent replicates) or cypin-mCherry (n = 30 independent replicates). *P < 0.05, **P < 0.01, and ****P < 0.0001 versus respective control as determined by one-way ANOVA followed by Tukey’s multiple comparisons test (A) or two-tailed Student’s t test [(C) and (D)]. Wells from neuronal cultures from the same mother are represented black, red, green, or blue. Mean ± SEM is shown for all graphs.

Since cypin overexpression promotes K63-polyUb over K48-polyUb linkages and results in the accumulation of ubiquitinated PSD-95, we investigated whether this accumulation reflects an increase in K63 polyubiquitination of PSD-95. We used HEK293T cells to express PSD-95 at levels high enough to detect K63 polyubiquitination and inform later K63-polyUb enrichment studies from dorsal hippocampi from animals with altered cypin. We transfected the cells with plasmids encoding mRFP or cypin-mRFP and PSD-95–GFP and subjected protein extracts to K63-polyUb enrichment. Western blot analysis revealed that cypin overexpression significantly increases K63-polyUb linkages on PSD-95 (fold change: mRFP: 1.00 ± 0.16; cypin-mRFP: 2.15 ± 0.18, P < 0.01, Student’s t test, n = 3 independent replicates) (Fig. 3C), in line with increased levels of total ubiquitination in neurons (Fig. 1D). Furthermore, we coexpressed cypin and mutants of PSD-95 that impair K48-polyUb (K703R), K63-polyUb (K558R), or both (K544R) as previously described (20). Cypin overexpression resulted in increased PSD-95 protein levels [fold change: PSD-95–wild type (WT): 1.38 ± 0.06, P < 0.0001], and mutations did not block this increase (PSD-95–K544R: 1.32 ± 0.07, P < 0.001; PSD-95–K558R: 1.28 ± 0.08, P < 0.01; PSD-95–K703R: 1.21 ± 0.02, P < 0.05; determined by one-way ANOVA followed by Dunnett’s multiple comparisons test, n = 6 culture wells) (Fig. 3D). These results suggest that cypin-promoted increases in PSD-95 are independent of K63-polyUb and K48-polyUb on these lysines.

Since PSD-95 modulates channel gating and postsynaptic targeting of NMDA receptors, specifically the GluN2A subunit (37), and K63 polyubiquitination leads to proteasome and degradation-independent processes, such as signal transduction (38), we cultured neurons and determined whether cypin overexpression affects intracellular Ca⁺² levels. We found that cypin overexpression leads to an increase in the levels of intracellular Ca⁺² (fold change: mCherry: 1.00 ± 0.02; cypin-mCherry: 1.25 ± 0.03; P < 0.001, Student’s t test, n = 30 replicates) (Fig. 3E). These data suggest that cypin overexpression increases NMDA receptor activity by increasing K63-polyUb linkages on PSD-95 and total levels of both PSD-95 and GluN2A, demonstrating a previously unidentified role for K63 polyubiquitination in modulating synaptic signaling.

Cypin overexpression regulates synaptic content in vivo

To assess whether overexpression of cypin alters protein composition at the synapse, we prepared synaptosome-enriched fractions (fig. S5A) of 3-month-old male mice overexpressing cypin (4.1 ± 0.40-fold increase in WT levels in synaptosomes; n = 10 mice per group) in the dorsal hippocampus and measured the amount of ubiquitinated proteins. We observed an increase in ubiquitinated proteins at synapses with cypin overexpression (fold change: adeno-associated virus (AAV)-GFP: 1.00 ± 0.04; AAV-cypin-GFP: 1.26 ± 0.10, P < 0.05, Student’s t test, n = 10 animals) (Fig. 4B). Consistent with our results in cultured neurons, cypin overexpression led to a significant decrease in the levels of K48-polyUb–linked proteins (fold change: AAV-GFP: 1.00 ± 0.05; AAV-cypin-GFP: 0.74 ± 0.05, P < 0.001, Student’s t test, n = 10 animals) (Fig. 4B); however, we did not observe an effect on the level of K63-polyUb–linked proteins (Fig. 4B). Since we did not assay for all types of polyUb linkages, the opposing changes to K48- and K63-polyUb linkages are not inconsistent with increases in total polyubiquitinated proteins.

Fig. 4. Cypin overexpression regulates synaptic content in vivo.

(A) Schematic representation of experimental workflow. (B) Western blot analysis of levels of total Ub, K48-polyUb, and K63-polyUb in lysates of synaptosomes isolated from the dorsal hippocampus of mice overexpressing GFP (AAV-GFP; n = 5 or 10 animals) or cypin-GFP (AAV-cypin; n = 5 or 10 animals). Protein levels are normalized levels to TPS and GFP. (C) Western blot analysis of GluN2A, GluN2B, GluR1, PSD-95, SNAP25, and β-actin levels in lysates of synaptosomes isolated from the dorsal hippocampus of mice overexpressing GFP or cypin-GFP. Protein levels are normalized levels to TPS where the IgG bands of the antibody used in the co-IPs is detected and then to GFP control. β-Actin serves as a negative control. (D) Cypin overexpression–mediated changes to the proteome analyzed for cellular components and biological functions of synapses (n = 7 animals for each group). *P < 0.05, **P < 0.01, and ***P < 0.001 versus GFP control as determined by two-tailed Student’s t test [(A) and (B)]. Mean ± SEM is shown for all graphs. n = number of mice. Schematic created in BioRender. Gandu, S. (2025) https://BioRender.com/i09q340. ECM, extracellular matrix; ER, endoplasmic reticulum. DCV, dense core vesicles; SV, synaptic vesicles.

Since K48 polyubiquitination targets proteins for degradation via the proteasome (39), we investigated whether cypin overexpression affects the levels of synaptic proteins. We found that cypin overexpression significantly increases the levels of PSD-95 (fold change: AAV-GFP: 1.00 ± 0.06; AAV-cypin-GFP: 1.23 ± 0.06, P < 0.05, Student’s t test, n = 10 animals) but not β-actin (Fig. 4C), consistent with our in vitro results. Overexpression of cypin also significantly increased the levels of GluN2A (fold change: AAV-GFP: 1.00 ± 0.06; AAV-cypin-GFP: 1.18 ± 0.05, P < 0.05, Student’s t test, n = 10 animals) and GluR1 (fold change: AAV-GFP: 1.00 ± 0.04; AAV-cypin-GFP: 1.28 ± 0.07, P < 0.01, Student’s t test, n = 10 animals) subunits, but it did not change the levels of GluN2B subunits at synapses (Fig. 4C).

Since cypin overexpression decreases K48 polyubiquitination, which is responsible for the proteasomal degradation of proteins (38), we performed mass spectrometry analysis of the synaptosome-enriched fractions from our mice. We detected a total of 1957 proteins across all samples. We further processed the data to map the gene IDs that are significantly regulated on the basis of fold change and synaptic location and function. We found that cypin overexpression alters proteins at both presynaptic and postsynaptic sites (Fig. 4D). Gene IDs mapped based on function revealed that cypin overexpression regulates synaptic organization predominantly at presynaptic sites and also at postsynaptic sites (Fig. 4D), with changes observed to PSD-95 and NMDA receptor subunits.

Cypin knockout regulates synaptic content in vivo

To further identify the role of cypin in the ubiquitination of proteins, we injected CamKII-cre virus into the dorsal hippocampus of 3-month-old male cypin^flox/flox (GDA ^flox/flox) mice. This results in cypin knockout in neurons expressing cre-recombinase (0.69 ± 0.07 of WT levels in synaptosomes; n = 6 mice). Unlike our previous results using total lysate of cultured neurons, cypin knockout did not affect the levels of total ubiquitinated proteins at the synapse (Fig. 5A). We also investigated whether cypin knockout affects the levels of K63-polyUb–linked proteins at synapses, and Western blot analysis revealed that cypin knockout decreases the level of synaptic K63-polyUb–linked proteins (fold change: AAV-GFP: 1.00 ± 0.07; AAV-cre: 0.81 ± 0.03, P < 0.05, Student’s t test, n = 6 animals) (Fig. 5A), consistent with our in vitro data.

Fig. 5. Cypin knockout regulates synaptic content in vivo.

(A) Western blot analysis of levels of total ubiquitin and K63-polyUb in lysates of synaptosomes isolated from the dorsal hippocampus of cypin knockout mice, resulting from injection of AAV-cre into cypin^flox/flox mice (AAV-cre). Total ubiquitin and K63-polyUb levels are normalized to TPS and GFP (AAV-GFP; n = 5 to 6 animals per group). (B) Western blot analysis of levels of GluN2A, GluN2B, GluR1, PSD-95, SNAP25, VAMP2, cypin, and β-actin levels in lysates of synaptosomes isolated from the dorsal hippocampus of mice expressing GFP or cre. Protein levels are normalized to TPS and GFP. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001 as determined by two-tailed Student’s t test. Samples where cypin knockout was less than 25%, as described in Materials and Methods, were excluded from the analysis. Mean ± SEM is shown for all graphs. n = number of mice. β-Actin serves as a negative control.

Since K63-polyUb plays a role in targeting proteins to different sites in the cell (38), we then asked whether cypin knockout affects the levels of synaptic proteins. Cypin knockout decreases synaptic PSD-95 (fold change: AAV-GFP: 1.00 ± 0.03; AAV-cre: 0.52 ± 0.06, P < 0.0001, Student’s t test, n = 6 animals), GluN2A (fold change: AAV-GFP: 1.00 ± 0.13; AAV-cre: 0.33 ± 0.05, P < 0.001, Student’s t test, n = 6 animals), GluR1 (fold change: AAV-GFP: 1.00 ± 0.06; AAV-cre: 0.58 ± 0.06), GluN2B (fold change: AAV-GFP: 1.00 ± 0.11; AAV-cre: 0.56 ± 0.08, P < 0.01, Student’s t test, n = 6 animals), and VAMP2 (fold change: AAV-GFP: 1.00 ± 0.06; AAV-cre: 0.81 ± 0.04, P < 0.05, Student’s t test, n = 6 animals) levels (Fig. 5B). In agreement, the opposite effect was observed when cypin was overexpressed. Furthermore, β-actin did not change. We confirmed decreased cypin levels (fold change: AAV-GFP: 1.00 ± 0.08; AAV-cre: 0.69 ± 0.07, P < 0.05, Student’s t test, n = 6 animals) (Fig. 5B) and the presence of cre-recombinase (fig. S5C). These data suggest that cypin modulates the levels of receptors at synapses.

Cypin regulates localization of the 19S proteasome at synapses in vivo

The presence of proteasome complexes at synapses (40) and the fact that cypin alters total levels of 20S α (1, 2, 3, 5, 6, and 7) proteasome subunits in cultured neurons prompted us to investigate whether overexpression or knockout of cypin causes changes to synaptic levels of proteasome subunits in vivo. Using extracts from mice transduced with virus to overexpress or knockout cypin, we used Western blot analysis of synaptosomes. Cypin overexpression decreased the levels of Rpt6, a subunit of the 19S regulatory particle (fold change: AAV-GFP: 1.00 ± 0.04; AAV-cypin-GFP: 0.75 ± 0.04, P < 0.001, Student’s t test, n = 10 animals) (fig. S5B). In contrast, cypin knockout increased the levels of Rpt5, a subunit of the 19S regulatory particle (fold change: AAV-GFP: 1.00 ± 0.08; AAV-cre: 1.31 ± 0.09, P < 0.05, Student’s t test, n = 6 animals) (fig. S5C), which suggests that cypin can alter proteasomal composition at the synapse. As a control, β-actin did not change. These results suggest that cypin alters proteasomal composition at synapses, resulting in changes to synaptic content.

Changes to cypin levels result in altered protein abundance at both pre- and postsynaptic sites in vivo

Using an unbiased proteome-wide approach, we investigated cypin-mediated changes in mice with cypin overexpression. We injected AAV-cypin (n = 7 animals) or CamKII-cre (n = 6 animals) virus into the dorsal hippocampus of 3-month-old male C57BL6 or cypin^flox/flox mice, respectively, along with appropriate control AAVs. We extracted proteins from the total lysate and subjected them to mass spectrometry analysis and identified significantly altered proteins by volcano plot (fig. S6A). To identify the functional importance of proteins altered by changes to cypin protein levels in synapse biology, we performed a gene set enrichment analysis (GSEA) and focused on cellular components and biological processes at the synapse using Synaptic Gene Ontologies (SynGO). We found that cypin overexpression and knockout significantly regulate protein abundance in GO cellular components, such as cytosol, membranes, at both postsynaptic and presynaptic sites (Fig. 6A). Furthermore, both postsynaptic specialization and presynaptic active zones were changed with altered cypin levels (Fig. 6A).

Fig. 6. Altered cypin expression results in changes to the synaptic proteome in vivo.

(A and B) Functional analysis of cellular components and biological processes at the synapse identified with significantly regulated genes from mass spectrometry analysis from total lysates of dorsal hippocampus of mice with cypin overexpression (n = 7 animals per group) or knockout (n = 6 animals per group). (C) Comparison (left) and heatmap (right) analyses of genes encoding proteins regulated with both cypin overexpression (OE) and knockout (KO) (missing values are represented by white color). (D) GSEA of genes encoding proteins regulated with both cypin overexpression and knockout and mapped into the cellular component (top) and biological process (bottom) pathways at the synapse. Abundances were normalized to respective controls (GFP for cypin overexpression and AAV-GFP for knockout) to generate the fold change of expression. Significant molecules were identified by P < 0.05 as determined by Student’s t test of fold change values.

To focus on the role of cypin in the regulation of synaptic content, we isolated proteins that change with both cypin overexpression and knockout (Fig. 6C). We identified 440 proteins that only change with overexpression and 1017 proteins that change with knockout (Fig. 6C). There were 343 proteins that changed when cypin was either overexpressed or knocked out, and we performed GSEA on the common genes encoding the proteins to identify synapse biology–related hits. We found that a significant number of these common genes are located at the synapse (Fig. 6D), including calcium/calmodulin-dependent protein kinase II (CAMK2B; postsynaptic), cholinergic muscarinic receptor 1 (CHRM1; postsynaptic), neurexin-3 (Nrxn3; presynaptic), synaptoporin (SYNPR; presynaptic), and synaptosomal-associated protein, 25 kDa (SNAP25; presynaptic). Together, our results suggest that cypin plays a crucial role in regulating synaptic content at both pre- and postsynaptic sites.

Cypin regulates the K63-polyubiquitinated proteome at synapses in vivo

Since we found that cypin alters K63-polyUb of proteins in developing neurons in culture, we asked whether cypin-mediated changes to proteins at the synapse in vivo are due to changes to K63 polyubiquitination. We performed mass spectrometry analysis of synaptosomal proteins (n = 7 animals) and K63-polyUb–enriched proteins (n = 4 animals) from the total lysate of the dorsal hippocampus of 3-month-old male mice overexpressing cypin. Consistent with our GSEA analysis of the total lysate from mice with altered cypin levels, synaptosomal protein abundance changed with cypin overexpression in the synaptogenesis signaling pathway and other pathways associated with synaptic functions (Fig. 7A). Cypin overexpression resulted in changes to the protein ubiquitin pathway (Fig. 7A). These changes may be due to reduced proteasome activity and to increased K63-polyUb, whether in combination or independently.

Fig. 7. Cypin regulates K63 polyubiquitination proteome at synapses in vivo.

(A) IPA showing activation in orange (z-score > 2) or inhibition in blue (z-score < −2) of pathways altered with cypin overexpression in isolated synaptosomes. (B) GSEA of genes encoding proteins changed in both synaptosomes (n = 7 animals per group) and K63-polyUb–enriched (n = 4 animals per group) samples with cypin overexpression. Genes are mapped into the cellular components of the synapse. (C and D) Heatmaps showing fold change of expression of common significantly regulated genes in both synaptosomes and K63-polyubiquitin enrichment fractions (C), and total lysate, synaptosomes, and K63-polyUb–enriched fractions (D) from mice with cypin overexpression (missing values are represented by white color). (E) Synaptic genes annotated against cellular components and biological process ontology terms from cypin overexpression (total lysate, synaptosomes, and K63-polyUb enriched) and cypin knockout samples. SynGo ontology terms with corrected P < 0.05 were included in the analysis. Abundances were normalized to respective controls (GFP for cypin overexpression and AAV-GFP for knockout) to generate fold change of expression. P values were determined by the Student’s t test.

To determine whether the changes to synaptic content reflect K63-polyUb, we compared significantly regulated genes present in synaptosomes and K63-polyUb–enriched samples and mapped them to cellular components at the synapse. We found that most of the genes in common mapped to the presynaptic endosome (Fig. 7B). To more completely understand how cypin regulates K63-polyUb protein abundance at the synapse, we compared the expression profiles in synaptosomes and K63-polyUb–enriched samples (Fig. 7C). There was a high correlation (~68%) in changes to specific proteins between the two samples, indicating that cypin mediates changes to synaptic content by both regulating K63-polyUb linkages and by affecting proteasome activity. Furthermore, we compared the expression fold change profiles of the two samples with total lysate and found a lower number of proteins in common (Fig. 7D). These results highlight the significant contribution of cypin-mediated K63-polyUb linkages to protein modifications at synapses.

To identify the synaptic location where cypin-mediated K63-polyUb occurs, we performed GSEA of genes encoding proteins from the total lysate from mice with cypin overexpression or knockout in the dorsal hippocampus, synaptosomes, and K63-polyUb–enriched samples. This analysis revealed that cypin regulates important biological processes at both pre-and postsynaptic sites and that cypin-mediated changes to K63-polyUb linkages is primarily involved in presynaptic processes, synapse organization, and transport but not metabolism or postsynaptic processes (Fig. 7E).

DISCUSSION

The current work demonstrates the simultaneous regulation of ubiquitin, specifically K63-polyUb linkages, on pre- and postsynaptic proteins. We have uncovered a previously unidentified pathway by which synaptic content is regulated by K63 polyubiquitination during neuronal development and in adult mice. Cypin binds to the β7 subunit of the proteasome, resulting in decreased proteasome activity. Furthermore, cypin alters the levels of proteasome subunits in developing cultured neurons and regulates synaptic localization of 19S proteasome subunits. The regulation of K63 polyubiquitination by cypin of pre- and postsynaptic synaptic proteins influences their abundance at synapses both in developing neurons and in vivo in the adult mouse, elucidating how synaptic content can change during plasticity events. Our data support the idea that increased K63 polyubiquitination of pre- and postsynaptic proteins explains the role of cypin at the synapse. A role for cypin in proteasome assembly and activity is also evident; however, its significance in synaptic function will be addressed in future studies. Nonetheless, these two discoveries illuminate roles for cypin in proteasome-dependent and proteasome-independent mechanisms, both of which may involve K63-polyUb linkages (Fig. 8).

Fig. 8. Model of the effects of cypin-regulated K63 polyubiquitination of the synaptic proteome.

Model for the effects of cypin based on the data from Western blot analysis and mass spectrometry analysis. Cypin regulates the expression of synaptic proteasome subunits, K63-polyUb of both pre- and postsynaptic proteins, and UBE4A protein levels. In turn, the abundance of synaptic proteins changes, including PSD-95, glutamate receptor subunits, and presynaptic proteins. As a result, synaptic signaling at both pre- and postsynaptic sites is modulated. Schematic created in BioRender. Gandu, S. (2025) https://BioRender.com/c40z366. NMDAR, NMDA receptor; AMPAR, AMPA receptor.

The regulation of K63 polyubiquitination of proteins by cypin has implications in development and after injury. Cypin plays a role in brain development (41, 42), increases with neuronal activity (28), shapes dendrites (28, 31, 43, 44), and regulates formation of spines (28, 31, 43). At the single-cell level, overexpression of cypin increases the frequency of miniature excitatory postsynaptic currents (31). At the circuit level, cypin overexpression results in an increase changes to firing patterns (31) and sensitivity of neural circuits to the AMPA receptor antagonist cyanquixaline (45) by regulating AMPA receptor targeting and changes to network organization, information relay, and how information is encoded (45). In adult mice, cypin protein expression increases after traumatic brain injury (TBI) (32). Furthermore, activation of cypin after TBI restores fear-conditioned memory (32). We observed that cypin regulates the protein levels of glutamate receptors, notably at the synapse.

Loss of a single gene that encodes Ub is lethal, thereby limiting studies in the developing brain (46–49). However, studies on patients with Angelman syndrome including UBE3A mutations point to the importance of this posttranslational modification in developing children and into adulthood (50, 51). Our study in cultured cortical neurons allows us to identify the role of polyubiquitination in the regulation of synaptic content as the neurons develop, which is not possible in developing embryos. In addition, the fact that we manipulate levels of a regulator of K63-polyUb linkages, and do so for a short time, i.e., over a few weeks, bypasses the lethality associated with knocking out Ub genes. However, it should be noted that there are limitations as our cultures include only neurons and astrocytes in a 1:1 ratio and lack oligodendrocytes (52). They also lack the inherent connectivity found in the intact brain. However, the fact that the mechanisms by which cypin regulates K63 polyubiquitination is similar to what we observed in vivo in adult mice gives confidence that our results are biologically relevant.

Protein ubiquitination is a major posttranslational modification at excitatory synapses (53). Specifically, ubiquitination of PSD-95 (54) and the deubiquitinating enzyme USP46 (23) regulate AMPA receptors at the synapse. Changes to K63-polyUb linkages on proteins by cypin in adult mice reflect relevant mechanisms underlying learning and memory. K63-polyUb has been linked to nuclear factor κB activation, which promotes long-term memory (55). DNA damage responses (56) and mitochondrial homeostasis (56), which are regulated by K63-polyUb linkages, play roles in synaptic plasticity. Furthermore, proteasome-independent K63 polyubiquitination regulates fear memory formation in female rats (57, 58). In line with these reports, we found that inhibition of cypin with a small molecule (G5) results in decreased learning in a fear conditioning test (32). Although this small molecule decreases the enzymatic activity of cypin as a GDA (32), we did not observe changes to ubiquitination levels of proteins with expression of cypin mutants lacking GDA activity (fig. S2C). Thus, how cypin mediates changes to K63-polyUb linkages and the resulting neurobehavioral changes are important future questions to pursue in rodent models.

The effects of cypin overexpression on neural circuit dynamics can be explained by a twofold increase in presynaptic activity (45). In addition, we previously reported that knockdown of cypin increases mEPSC amplitude (32), which correlates with decreased postsynaptic strength and inhibition of long-term depression (59). Thus, cypin may modulate synaptic plasticity via both pre- and postsynaptic mechanisms. In addition to postsynaptic pathways in our SynGO and GSEA analyses, we identified presynaptic active zones and presynaptic endosomes as being affected in the synaptosomes and K63-polyUb–enriched fractions in mice with altered cypin expression. Furthermore, one of the proteins that changes with both cypin overexpression and knockout is SNAP25, which mediates synaptic vesicle exocytosis (60). We previously identified snapin, a protein that interacts with SNAP25, as a cypin interactor that regulates postsynaptic dendritic arborization (61). Our new data suggest that this interaction may also play a presynaptic role, demonstrating cypin as a simultaneous regulator of pre- and postsynaptic function.

We did not extend our study to other types of Ub linkages. For example, we found that cypin overexpression leads to a significant decrease in the levels of K48-polyUb–linked proteins in both developing and mature neurons; however, although K63-polyUb linkages in total proteins increase in developing neurons with cypin overexpression, changes are only observed in vivo when we analyze tissue from mice with cypin knocked out. These data suggest that there may be a homeostatic mechanism in mature neurons that regulates the levels of K63-polyUb when cypin is overexpressed in vivo. Furthermore, the fact that the postsynaptic density is not yet mature in the developing neurons may explain differences in cypin action in developing versus mature neurons. In addition, in concert with K63-polyUb, changes to K48-polyUb may play important roles in regulation of synaptic proteins. Since much more is known about K48-polyUb and its role in protein turnover, we chose to study the regulation of K63-polyUb on synaptic proteins in neurons. Future studies will address how the interplay between K48-polyUb and K63-polyUb regulates neuronal function.

There are additional challenges to studying polyubiquitination as the identification and determination of specific ubiquitin linkages requires specialized tools. For example, antibodies that are specific to one type of polyubiquitin linkage are not available, making detection and quantification difficult. Ubiquitination is transient in nature, and thus, capturing and detecting ubiquitinated proteins degradation are often challenging. Multiple E3 ligases can target same substrate, making it difficult to understand the specific roles of these enzymes. As new tools become available, we can tackle important questions about cypin-promoted changes to polyubiquitination, such as resulting neurobehavior from changes to K63-polyUb linkages.

Most of our experiments were performed with cypin overexpression. This is due to the fact that we observed at least a 10-fold increase in cypin protein expression when cultured neurons are activated with KCl (28) and a 25% increase in cypin levels at 1 day posttraumatic brain injury in mice (32). It is also possible that knockdown of cypin does not have effects in culture (Figs. 2G and 3A) due to the fact that cypin protein levels increase during development (30) and that cultures do not express high levels of cypin or levels found in the adult brain (29, 33). In addition, our in vitro studies use cortical cultures, and our in vivo studies target the cortex. This is due to the following reasons. For cultures, the yield of E18 hippocampal neurons is quite low, making it difficult to perform large-scale biochemical studies, and thus, we chose cortical cultures for the higher yield of neurons. Since targeting the hippocampus in vivo allows for higher transduction efficiency [discussed in (62)] and for a more concentrated and restricted area of expression, we chose the hippocampus for in vivo studies. It has been reported that gene expression in the dorsal hippocampus correlates with cortical regions involved in information processing (63), allowing for comparisons between the two brain regions.

It should be noted that there are other mechanisms that regulate synaptic content. Specifically, local protein synthesis occurs in dendrites and axons of hippocampal neurons (64–70). Long-distance transport of proteins and RNAs play important roles in shaping synaptic content and long-term memory [reviewed in (71)]. For example, RNAs and RNA binding proteins are transported by molecular motors along microtubules (72–75). Most recently, long noncoding RNAs (lncRNAs) have been identified as transcriptional regulators of synaptic function. These lncRNAs regulate genes that play a role in neuronal development, differentiation, and activity-dependent changes to synapses (76) and in shaping neuronal structure (77, 78). Posttranslational modifications, including phosphorylation, O-GlcNAcylation, acetylation, palmitoylation, and nitrosylation, play a role in localization and activation of AMPA receptors (79, 80), gephyrin (81), and PSD-95 (82–86). In addition to our current work, these studies demonstrate the complex regulation of synaptic content that occurs during development and learning.

In summary, we have report a previously unidentified pathway by which K63-polyUb linkages are regulated on synaptic proteins by cypin. We demonstrate changes to the synaptic proteasome composition and activity and additionally demonstrate differences in K63-polyUb and K48-polyUb with changes to cypin expression levels. We also report that cypin overexpression up-regulates UBE4A, which mediates K63-polyUb linkages on proteins. Together, our results suggest that the therapeutic targeting of cypin may have multiple effects at the synapse, including regulation of glutamate receptors, thereby shaping neural circuits to promote synaptic plasticity.

MATERIALS AND METHODS

Animals

All animal experiments were approved by the Rutgers Institutional Animal Care and Use Committee (protocol nos. 999900080 and 202000069). Male C57BL/6J mice used in these experiments were acquired from the Jackson Laboratory, and cypin^flox/flox (GDA^flox/flox) mice were generated as follows. We used a construct from BACPAC Resources Center (Emeryville, CA), with flox sites flanking the promoter, and exon 4 was introduced into the cypin gene. The construct was transfected into R1 embryonic stem cells. DNA was isolated from these colonies, and polymerase chain reaction was used to identify clones that had undergone homologous recombination. After double confirmation using Southern blotting with 3′ and 5′ probes, the ES cells were used to generate chimeras, with one going germ line (fig. S7). These mice were crossed with flippase mice (the Jackson Laboratory), following the breeding scheme, to generate cypin^flox/flox mice (fig. S7). Animals were housed at a maximum capacity of five animals per cage. All animals were provided with food and water ad libitum and subjected to 12-hour light and 12-hour dark cycles. Animals were randomly assigned to experimental groups.

Stereotaxic surgery and AAV injection

Three-month-old male C57BL/6J (WT) or cypin^flox/flox mice were anesthetized with isoflurane gas (3.5% for induction and 1.5% for maintenance), mounted onto a stereotaxic device (KOPF Instruments), and held in place with ear bars. A linear incision was made on the skin on the midline of the head, and a 0.5-mm burr was used to drill a hole in the skull 1.9 mm lateral (X) and 1.4 mm anterior (Y) to the midline on both hemispheres using a mechanical drill. pAAV-CMV-GFP (5 × 10⁹ vg) or pAAV-CMV-mGda(cypin) (5 × 10⁹ vg) was injected into WT mice, and pAAV-hSyn-EGFP (3 × 10⁹ vg) or pAAV-hSyn-Cre (3 × 10⁹ vg) was injected into cypin^flox/flox mice. Viral particles were injected into the dorsal hippocampus at 1.6 mm below the surface of the skull using a microneedle prepared from glass capillaries (Sutter Instruments). The injection rate was set at 3 nl/s. After injection, the needle remained in place for an additional 5 min to prevent any leakage before being removed. Analgesic (Carprofen, 5 mg/kg) was administered to animals immediately after surgery.

AAV constructs for overexpression of cypin in vivo

Cypin overexpression was achieved by stereotaxic injection of AAV particles encoding cypin (NM_010266.2) under the cytomegalovirus (CMV) promoter for constitutive expression followed by 3× GGGGS linker and GFP reporter. A construct encoding GFP was used as a control. Sequences were packaged into a virus of the AAV9 serotype (high affinity for central nervous system transduction). pAAV-CMV-GFP (VB010000-9394npt) and pAAV-CMV-mGda (cypin) (VB220207-1188cej) in vivo grade ultra-purified viral preparations were obtained from VectorBuilder (Chicago, IL). Cypin knockout was achieved by stereotaxic injection of AAV particles encoding cre-recombinase or GFP (control) under the synapsin promoter into cypin^flox/flox mice. pAAV-hSyn-EGFP (catalog no. 50465) and pAAV-hSyn-Cre (catalog no. 105553) in vivo grade viral particles were obtained from Addgene (Watertown, MA).

Neuronal and mammalian cell culture

Primary neurons were cultured from cortices isolated from embryos harvested from Sprague-Dawley rats on the 18th day of gestation, as previously described by the Firestein Laboratory (32, 33). Before culturing the cells, all plates and dishes were coated with poly-D-lysine (0.1 mg/ml; Sigma-Aldrich, catalog no. P0899). Neurons were plated at a density of approximately 105,000 cells/cm² and cultured in Neurobasal medium (Gibco, catalog no. 21103049) supplemented with B-27 (Gibco, catalog no. 17504044) and 2 mM l-glutamine (Gibco, catalog no. 25030081). All neuronal culture experiments were conducted in three trials, with three independent biological replicates in each trial. HEK293T cells were plated at a density of 52,500 cells/cm² in Dulbecco’s modified Eagle’s medium (Gibco, catalog no. 11965092) supplemented with 10% fetal bovine serum (Gibco, catalog no. 26140079).

Plasmids and transfection

At 24 hours after plating, HEK293T cells were transfected with plasmids encoding mRFP or cypin-mRFP and HA-Ub for total ubiquitination assays; mRFP or cypin-mRFP, HA-Ub, and PSD-95–GFP for ubiquitin enrichment assays; mRFP or cypin-mRFP, PSD-95–GFP (WT), ΔPSD-95–GFP–K544R (K48-polyUb and K63-polyUb), ΔPSD-95–GFP–K558R (K63-polyUb), and ΔPSD-95–GFP–K703R (K48-polyUb) for mapping ubiquitin sites; and PzsGreen (Takara Bio USA, catalog no. 632425) or UbG76V (Addgene, catalog no. 11941) for proteasome sensor assays using the Lipofectamine 2000 reagent (Invitrogen, catalog no. 11668027) or a CalPhos Mammalian Transfection kit (Takara Bio Inc., catalog no. 631312) following the manufacturer’s protocols as we previously described (87).

Lentiviral preparation and transduction

cDNA encoding cypin was subcloned into FG12 vectors as previously described (31). HEK293T cells were plated at a density of 100,000 cells/cm² in T-75 flasks. Cells were transfected with plasmids encoding GFP (gift from C. Pröschel, University of Rochester) or cypin-GFP or mCherry (Addgene, catalog no. 36084) or cypin-mCherry along with the packaging plasmid (psPAX2) and envelope plasmid (VSVg) using Lipofectamine 2000 transfection reagent following the manufacturer’s protocol. At 24 hours posttransfection, the transfection medium was replaced with a fresh medium. Medium-containing viral particles were collected at 48 and 72 hours posttransfection and concentrated using PEG-it virus precipitation solution (System Biosciences, catalog no. LV810A-1) as per the manufacturer’s protocol. Neuronal cultures were transduced with a lentiviral concentrate on DIV9. Successful transduction was confirmed by fluorescence microscopy and Western blot analysis for each trial.

Synaptosome isolation from brain tissue

Flash-frozen brain tissue was briefly homogenized with a glass Dounce homogenizer by applying 15 to 20 gentle strokes in 500 μl of Syn-PER Synaptic Protein Extraction Reagent (Thermo Fisher Scientific, catalog no. 87793) supplemented with protease inhibitors (Roche, catalog no. 11836170001) and phosphatase inhibitors (Roche, catalog no. 4906837001), to separate cytosolic and synaptosome-enriched fractions. The homogenate was centrifuged at 1200g for 10 min at 4°C. The resulting supernatant was collected and centrifuged at 15,000g for 20 min at 4°C. The supernatant was collected and saved for further processing. The pellet, corresponding to the synaptosome-enriched fraction, was resuspended in 100 μl of Syn-PER Synaptic Protein Extraction Reagent with inhibitors.

Protein extraction and Western blot analysis

Cells plated in six-well plates were washed twice with ice-cold phosphate-buffered saline (PBS). Proteins were extracted by adding 100 μl of ice-cold radioimmunoprecipitation assay (RIPA) lysis buffer [50 mM tris (pH 7.4), 1% NP-40, 0.25% sodium deoxycholate, 150 mM NaCl, and 1 μM EDTA] containing protease inhibitors, phosphatase inhibitors, and 100 μM phenylmethylsulfonyl fluoride. Lysates were sonicated for 20 s (32) and incubated at 4°C for 1 hour, followed by centrifugation at 18,000g for 15 min. The supernatant was collected, and protein concentrations were determined using the Pierce BCA protein assay. Proteins (10 μg) were resolved on NuPAGE 4 to 12% bis-tris gels (Invitrogen, catalog no. NP0335BOX) for 1 hour at 120 V by electrophoresis in Mops running buffer. Proteins were later transferred to a polyvinylidene difluoride membrane in NuPAGE transfer buffer (Invitrogen, catalog no. NP00061) containing 10% methanol for 1 hour at 20 V. Membranes were stained for total protein using Revert total protein stain (TPS; LI-COR Biosciences, catalog no. 103546-310) and imaged in the 700 channel using the LI-COR Odyssey imaging system. Following TPS, membranes were destained and blocked in 5% bovine serum albumin in TBS-T [50 mM tris, 150 mM NaCl, and 0.1% Tween 20 (pH 7.6)], referred to as blocking solution, for 1 hour at room temperature. Membranes were then incubated with primary antibodies diluted in blocking solution at 1:500 (anti-Ub, Cytoskeleton, catalog no. AUB01), 1:1000 (anti-HA, Cell Signaling, catalog no. 3724; anti-PSD-95, Neuromab, catalog no. 75-028; anti-K63-polyUb, Cell Signaling, catalog no. 5621; anti-K48-Ub, Cell Signaling, catalog no. 8081; anti–β-actin, Cell Signaling, catalog no. 3700; anti-GluN2A, Cell Signaling, catalog no. 4205; anti-GluN2B, Cell Signaling, catalog no. 4207; anti-GluR1, Cell Signaling, catalog no. 13185; anti-UBE4A, Santa Cruz, catalog no. sc-365904; anti-PSMC2, Santa Cruz, catalog no. sc-166972; anti-UBE2J1, Santa Cruz, catalog no. sc-377002; anti-proteasome 20S α1, 2, 3, 5, 6, and 7 subunits, Enzo, catalog no. BML-PW8195; anti-Rpt5, Enzo, catalog no. BML-PW8770; anti-Rpt6, Enzo, catalog no. BML-PW9265; anti-cypin, Firestein lab BF6 clone) overnight at 4°C. Membranes were washed three times with TBS-T and incubated with respective horseradish peroxidase–conjugated secondary antibodies (1:2500; Rockland) or fluorescently labeled secondary antibodies (1:2500; LI-COR) in blocking solution for 1 hour at room temperature. Membranes were again washed three times with TBS-T, and blots were developed with Immobilon ECL Ultra (Millipore, catalog no. WBULS0100) or fluorescently labeled secondary antibodies, which were imaged directly in the 700 channel. Images were acquired using the LI-COR Odyssey imaging system, and band optical density was quantified using Image Studio Lite software.

Calcium assay

Intracellular Ca⁺² levels were measured using the Fluo-4 Direct Calcium Assay Kit (Invitrogen, catalog no. F10471) following the manufacturer’s instructions with minor adjustments. Primary neurons were plated at a density of 78,000 cells/cm² in black-walled 96-well plates. On DIV9, neurons were transduced with lentiviral particles encoding mCherry or cypin-mCherry. On DIV17, cells were treated with 50 μl of assay buffer containing Fluo-4 reagent, 2.5 mM probenecid, and 50 μl of conditioned medium. To determine the ideal incubation time, fluorescence was measured (excitation at 494 nm and emission at 516 nm) every 10 min for a period of 2 hours using the Varioskan LUX multimode microplate reader (Thermo Fisher Scientific). Data were obtained by averaging the fluorescence measured from 13 different locations in each well. Fluorescence was normalized against control cultures expressing mCherry

Coimmunoprecipitation

HEK293T cells were cotransfected with plasmid encoding mRFP or cypin-mRFP, HA-Ub, and PSD-95–GFP. Protein extraction was performed in lysis buffer [25 mM tris-HCl (pH 7.5), 150 mM NaCl, 1% NP-40, and 2 mM EDTA] containing protease and phosphatase inhibitors. Protein extract (500 to 1000 μg) was incubated with 3 μg of rabbit anti-HA antibody overnight at 4°C. Pierce Protein A agarose beads (25 to 50 μl; Thermo Fisher Scientific, catalog no. 20333) were added to the extracts and incubated for 2 hours at room temperature. Extracts were centrifuged at 2500g for 3 min to pellet the beads, and the supernatant was collected. Beads were then washed three times with lysis buffer without protease or phosphatase inhibitors. Immunoprecipitated proteins were eluted in 2× Laemmli sample buffer (Bio-Rad, catalog no. 1610747), boiled for 5 min at 90°C, followed by centrifugation. The supernatant was collected, and proteins were resolved by SDS–polyacrylamide gel electrophoresis (SDS-PAGE).

Proteasome sensor assays

HEK293T cells were plated at a density of 25,000/cm² in 24-well plates containing glass coverslips. At 24 hours after plating, cells were cotransfected with plasmids encoding mRFP or cypin-mRFP and PzsGreen or UbG76V plasmid plasmids (27, 88). At 24 hours after transfection, cells were fixed in 4% paraformaldehyde in PBS and stained for nuclei with Hoechst 33342 (Thermo Fisher Scientific, catalog no. H3570). ZsGreen or GFP signal intensity was quantified using ImageJ software (National Institutes of Health).

Proteasome activity assay

Primary cortical neurons were plated in 12-well plates at a density of 145,000/cm². Cells were lysed in proteasome activity assay buffer [50 mM Hepes (pH 7.5), 5 mM EDTA, 150 mM NaCl, 1% Triton X-100, 2 mM adenosine triphosphate (ATP), and 1 mM dithiothreitol] to measure proteasome activity. ATP was included in the lysis buffer to improve the recovery of the intact 26S proteasome. Cell extract (10 μg) was used to measure chymotrypsin-, trypsin-, and caspase-like activity in the presence of fluorogenic substrate 50 μm of Suc-LLVY-AMC (Enzo, catalog no. BML-P802), 50 μm of Boc-LRR-AMC (Boston Biochem, catalog no. S-300), and 50 μm of Z-LLE-AMC (Boston Biochem, catalog no. S-230), respectively. The release of free AMC was measured at excitation/emission wavelengths of 346/442 nm. The amount of AMC released is directly proportional to proteasome activity.

Proteasome enrichment

HEK293T cells were plated at a density of 105,000 cells/cm² in a 100-mm dish. At 24 hours after plating, cells were transfected with plasmids encoding mRFP or cypin-mRFP using Lipofectamine 2000. At 24 hours after transfection, cells were washed once with ice-cold PBS and lysed in 400 μl of lysis buffer [25 mM Hepes (pH 7.4), 10% glycerol, 1 mM ATP (pH 7.4), and 5 mM MgCl₂] containing protease (Roche, catalog no. 11836170001) and phosphatase (Roche, catalog no. 4906837001) inhibitors. Cells were lysed by sonication at an amplitude of 10% for 10 s three times. Cellular debris was cleared by centrifugation at 10,000g for 10 min at 4°C, and the supernatant (S10) was collected and centrifuged at 100,000g for 30 min at 4°C. The supernatant (S100) was further centrifuged at 150,000g for 60 min at 4°C. The supernatant (S150) was collected, and the pellet was washed twice with wash buffer (lysis buffer without inhibitors) and resuspended in a 30- to 50-μl lysis buffer (89).

Polyubiquitin chain capture

Primary cortical neurons were plated at a density of 105,000 cells/cm² on PDL-coated six-well plates. On DIV9, cells were transduced with lentiviral particles encoding GFP or cypin-GFP, and on DIV17, the cells were lysed in lysis buffer [50 mM tris (pH 7.6), 50 mM NaCl, 10% glycerol, 1% Triton X-100, and 5 μM EDTA] supplemented with protease and phosphatase inhibitors. Cell extract (100 μg) was incubated with either GST-Rad23 (K48-polyUb capture) or GST-TAB2 (K63-polyUb capture; UBPBio, catalog no. J4440) overnight at 4°C followed by incubation with Pierce glutathione magnetic agarose beads (Thermo Fisher Scientific, catalog no. 78601) for 1 hour at 4°C. Beads were washed three times with wash buffer (lysis buffer without inhibitors), resuspended in 1× Laemmli sample buffer (Bio-Rad, catalog no. 1610747), and boiled for 5 min at 90°C, followed by centrifugation at 2500g for 3 min. The supernatant was collected, and proteins were resolved by SDS-PAGE.

HEK293T cells were plated at a density of 52,500 cells/cm² in 60-mm dishes. At 24 hours after plating, cells were cotransfected with plasmids encoding mRFP or cypin-mRFP and PSD-95–GFP. At 24 hours after transfection, cells were lysed for K63 polyubiquitin capture using GST-TAB2 recombinant protein. The supernatant was collected, and proteins were resolved by SDS-PAGE.

Mass spectrometry

Primary neuronal cells were lysed in 1× RIPA lysis buffer, and protein extract (150 μg) was digested using a standard gel-plug protocol. Half of each sample was labeled with TMT10 plex reagent (Thermo Fisher Scientific, catalog no. PI90110) according to the manufacturer’s instructions. Tandem mass tag (TMT)-labeled samples were desalted on a SPEC C18 column, and the proteins were solubilized in 200 μl of buffer A [20 mm of ammonium (pH 10)] and separated on an Xbridge column (Waters; C18; 3.5 μm, 2.1 mm by 150 mm) using a linear gradient of 1% B min^–1 and then from 2 to 45% B [20 mm of ammonium in 90% acetonitrile (ACN) (pH 10)] at a flow rate of 200 μl min⁻¹ using Agilent HP1100. Fractions were collected at 1 min intervals and dried under vacuum. Fractions from 31 to 44 min (14 fractions) were chosen for total proteome analysis. Nano–liquid chromatography–tandem mass spectrometry (LC-MS/MS) was performed using a Dionex rapid-separation liquid chromatography system interfaced with Orbitrap Eclipse (Thermo Fisher Scientific). Samples were loaded onto an Acclaim PepMap 100 trap column (75 μm by 2 cm, Thermo Fisher Scientific). The scan sequence began with an MS1 spectrum Orbitrap analysis, a resolution of 120,000, a scan range from 350 to 1600 mass/charge ratio (m/z), and a maximum injection time of 100 ms. For traditional MS2 analysis, the MS1 scan was set at a mass range from 375 to 1575 m/z, a scan resolution of 120,000, and a maximum injection time of 50 ms, and precursor ions were isolated with an m/z window of 0.7 Th (90). All LC-MS data were analyzed using Maxquant (version 1.6.2.6) with the Andromeda search engine. The type of LC-MS run was set to reporter ion MS2 with 6plex TMT as isobaric labels. Reporter ion mass tolerance was set at 0.003 Da. LC-MS data were searched against the The Arabidopsis Information Resource database with the addition of potential contaminants. The protease was set as trypsin/P, allowing two missed cuts. Carbaidomethylation of cysteine was set as a fixed modification, and N-terminal acetylation, oxidation at methionine, and acetylation at lysine were set as variable modifications. Proteins with false discovery rate (FDR) < 1% were reported. For quantification, spectra were filtered by minimum reporter PIF set at 0.5 (spectra purity) (90–92). All samples submitted for analysis were coded for experimenter blinding.

Tissue extract (10 μg of the total protein content) from each sample was buffer exchanged into 25 mM ammonium bicarbonate (AMBIC) using Amicon Ultra 0.5-ml centrifugal filters. Samples were dried to completion, resuspended in 50 μl of 25 mM AMBIC 50% ACN, and digested overnight with 1 μg of trypsin and 500 ng of Lys-C at 37°C. Samples were dried to completion and resuspended in 50 μl of 2% ACN and 0.1% formic acid (FA). Nano–LC-MS/MS was performed using an UltiMate 3000 UHPLC System (Thermo Fisher Scientific) with in-line desalting interfaced with a Q Exactive HF-X (Thermo Fisher Scientific). Each sample was analyzed via 5-μl injection onto a C18 EasySpray column (75 μm by 25 cm, Thermo Fisher Scientific). The linear separation gradient was from 4% ACN 0.1% FA to 40% ACN 0.1% FA over 100 min at a flow rate of 350 nl/min. The Q Exactive HF-X mass spectrometry analysis consists of one full-scan MS followed by top 20 ddMS2. The MS1 spectrum resolution was 60,000 at 200 m/z, the scan range was 300 to 1500 m/z, the maximum injection time was 45 ms, and the automatic gain control (AGC) was 3 × 10⁶. The MS2 spectrum resolution was 30,000 at 200 m/z, the maximum injection time was 54 ms, the AGC was 1 × 10⁵, and the isolation width was 1.3 m/z. Higher-energy collisional dissociation fragmentation was NCE 28, dynamic exclusion was 30 s, and singly charged and greater than +6 charged ions were excluded from selection. Database mining and protein identification were performed using Proteome Discoverer 2.5.0 via SEQUEST against the Mus musculus database (UP000000589), including trypsin and Lys-C sequences. Parameters were set as follows: at least two peptides (minimum length = 6, minimum precursor mass = 350 Da, maximum precursor mass = 5000 Da), tolerance of 10 parts per million for precursor ions and 0.02 Da for fragment ions (b and y ions only), up to two missed cleavages, and Percolator FDR (strict minimum value of 0.01). Oxidation (M), acetylation (protein N-term), and methionine loss (protein N-term) were set as dynamic modifications. Label-free quantitative values were determined by non-nested analysis with the precursor ion quantifier node and normalization via total peptide amount.

Bioinformatic analysis of proteomic data

Protein group results for in vitro samples were analyzed using Perseus (version 1.6.10.43) (93). The data were first filtered for reverse and contaminant hits. The reporter ion intensity data were further log₂ transformed and normalized to the column median value. For group comparisons, statistical significance between groups was analyzed using Student’s t test with equal variance on both sites, and the q value was calculated with permutation. Generated P values (Raw. P) and fold changes (Raw. FC) for each identified protein were based on preprocessed raw intensity data. The principal components analysis, volcano plot, and heatmap were generated in Perseus. Protein group results of normalized abundances for in vivo samples were log₂ transformed to calculate the P value and expression fold change. Further, P values were used to calculate the FDR (q value). Significantly altered proteins were identified using the P value of 0.1 and expression fold change of <−0.1 and >+0.1. Proteins that were significantly altered were isolated and subjected to GSEA (94), using QIAGEN IPA (QIAGEN Inc., https://digitalinsights.qiagen.com/IPA), to identify signaling pathways that were changed by cypin overexpression. Pathway analysis was restricted specifically to synapse and protein posttranslational modification only. Pathways that were activated are indicated in orange, those inhibited are in blue, those unchanged are in white, and those that cannot detect an activity pattern due to insufficient data are in gray color (95). The synapse biology–related gene set was prepared using the SynGO portal (https://syngoportal.org), which maps various protein IDs into gene identifiers and symbols (96).

Statistical analysis

Data were analyzed using GraphPad Prism software. Statistical significance was determined as P < 0.05 using either Student’s t test or one-way ANOVA, followed by Tukey’s post hoc comparison test, and all values are presented as mean ± SEM. All cell culture experiments were derived from three to four pregnant animals, and n values represent the number of individual culture wells. Cultures derived from the same pregnant animal were labeled in the same color (black, red, green, or blue). All in vivo experiments had at least n = 5 to 10 mice in each group. Means and SEs of means were reported for all results unless otherwise specified. Animals that did not have significant transgene expression (>150% gene expression of WT value for overexpression and <75% gene expression of WT value for knockout) were excluded from the study. To enhance rigor and reproducibility, we performed a two-way ANOVA on data from three independent neuronal cultures, each derived from a separate dissection. In this analysis, each well was treated as an individual unit, and wells from the same dissection were color-coded to indicate their shared biological origin (black, red, green, or blue), as shown in figs. S8 and S9. This approach allowed us to assess potential variation across biological replicates while preserving within-culture resolution.

Supplementary Materials

This PDF file includes:

Supplementary Text

Figs. S1 to S9