Research ArticleAlzheimer’s Disease

Activation of CaMKIV by soluble amyloid-β1–42 impedes trafficking of axonal vesicles and impairs activity-dependent synaptogenesis

+ See all authors and affiliations

Sci. Signal.  11 Jul 2017:
Vol. 10, Issue 487, eaam8661
DOI: 10.1126/scisignal.aam8661

Amyloid-β and intersynaptic trafficking

Synaptic loss and dysfunction as well as neuronal accumulation of amyloid-β (Aβ) are classic features of Alzheimer’s disease (AD). Synaptic components are transported along axons in actin- and synapsin-associated vesicles to adjust synaptic strength in response to activity and to promote the formation of new synapses. Using hippocampal neurons isolated from rats and mouse models of AD, Park et al. found that a soluble form of Aβ impedes Ca2+ clearance from neurons, which led to activation of the kinase CaMKIV. CaMKIV-mediated phosphorylation of synapsin caused its dissociation from synaptic vesicles and actin, thereby impairing vesicular transport. Targeting this pathway might suppress the pathological effects of Aβ in patients with AD.

Abstract

The prefibrillar form of soluble amyloid-β (sAβ1–42) impairs synaptic function and is associated with the early phase of Alzheimer’s disease (AD). We investigated how sAβ1–42 led to presynaptic defects using a quantum dot–based, single particle–tracking method to monitor synaptic vesicle (SV) trafficking along axons. We found that sAβ1–42 prevented new synapse formation induced by chemical long-term potentiation (cLTP). In cultured rat hippocampal neurons, nanomolar amounts of sAβ1–42 impaired Ca2+ clearance from presynaptic terminals and increased the basal Ca2+ concentration. This caused an increase in the phosphorylation of Ca2+/calmodulin-dependent protein kinase IV (CaMKIV) and its substrate synapsin, which markedly inhibited SV trafficking along axons between synapses. Neurons derived from a transgenic AD mouse model had similar defects, which were prevented by an inhibitor of CaMK kinase (CaMKK; which activates CaMKIV), by antibodies against Aβ1–42, or by expression a phosphodeficient synapsin mutant. The CaMKK inhibitor also abolished the defects in activity-dependent synaptogenesis caused by sAβ1–42. Our results suggest that by disrupting SV reallocation between synapses, sAβ1–42 prevents neurons from forming new synapses or adjusting strength and activity among neighboring synapses. Targeting this mechanism might prevent synaptic dysfunction in AD patients.

INTRODUCTION

Neurofibrillary tangles and amyloid plaques, caused by hyperphosphorylated tau and amyloid-β (Aβ) protein, respectively, are hallmarks of Alzheimer’s disease (AD) (1). Previous studies have indicated that the prefibrillar soluble form of Aβ oligomers (sAβ), rather than insoluble amyloid fibrils or plaques, is the major cause of the synaptic dysfunction and cognitive defects associated with AD (2, 3). sAβ1–42, a 42-residue major neurotoxic sAβ fragment, was found in elevated amounts and in association with synapses in the brains of AD individuals with dementia (4, 5). Injection of natural sAβ1–42 to rats specifically disrupted cognitive function (6), and sAβ1–42 was shown to alter synaptic plasticity by inhibiting long-term potentiation (LTP) (7) and to facilitate long-term depression (8). Furthermore, blocking de novo sAβ1–42 production reversed synapse loss and memory impairment in amyloid precursor protein transgenic mice (9).

Most studies so far have focused on the postsynaptic effects of sAβ1–42, and observations of presynaptic effects have been contradictory. For example, injection of sAβ1–42 into the presynaptic terminal of the squid giant axon inhibited synaptic transmission without affecting endocytosis (10), whereas others found that sAβ1–42 disrupted synaptic vesicle (SV) endocytosis (11). Moreover, in one study, endogenously released sAβ1–42 increased the release probability but did not alter postsynaptic function (12), whereas others reported that sAβ1–42 had parallel effects on pre- and postsynaptic mechanisms (13). Nevertheless, the above findings raise the possibility that sAβ1–42 has distinctive effects on presynaptic terminals. In a previous study, sAβ1–42 hampered SNARE (soluble N-ethylmaleimide–sensitive factor attachment protein receptor)–mediated exocytosis through a direct interaction with syntaxin 1a (14). We also found that sAβ1–42 severely impaired SV endocytosis and recycling at presynaptic terminals (15).

Contrary to the conventional view that SVs are synapse-specific, some SVs leave their host synapse and diffuse laterally along the axon (1618). Transport of SVs between synapses enables reallocation of functional SV pools and adjustment of synaptic strengths, thus dynamically regulating presynaptic behavior (1719). In addition, recruitment of intersynaptic vesicles to particular sites could generate new functional synapses, which might be important for producing synaptic plasticity (17, 20). It thus appears that regulatory changes of intersynaptic vesicular trafficking influence synaptic efficacy.

To explore how intersynaptic vesicular trafficking may be involved in neurodegenerative disease, we used live-cell imaging techniques to monitor SV trafficking in hippocampal neurons cultured from embryonic day 18 Sprague-Dawley (SD) rat embryos and found that sAβ1–42 strongly inhibited intersynaptic vesicular trafficking along the axon. We further showed that phosphorylation of Ca2+/calmodulin-dependent protein kinase IV (CaMKIV) and synapsin was a key mechanism involved. Functionally, the inhibitory effect on intersynaptic vesicular trafficking blocked SV reallocation, leading to a failure to form new synapses upon forskolin-induced LTP. These findings lead us to propose that sAβ1–42-induced defects in intersynaptic vesicular trafficking identify a novel cellular mechanism underlying presynaptic dysfunction in early AD.

RESULTS

sAβ1–42 inhibits chemical LTP-induced new synapse formation

To investigate the effect of sAβ1–42 on presynaptic regulation, we first assessed whether sAβ1–42 affected activity-dependent synaptogenesis during chemically induced LTP (cLTP) by forskolin, which is one of the most obvious readouts for presynaptically expressed synaptic enhancement (21, 22). First, using FM 1-43 [N-(3-triethylammoniumpropyl)-4-(4-(dibutylamino)styryl)pyridinium dibromide], a fluorescent styryl dye, to study SV recycling kinetics, we found that forskolin treatment increased FM 1-43 uptake, which represents the size of the SV pool, and accelerated its destaining rate, which identifies the presynaptic release probability in cultured rat hippocampal neurons (fig. S1), indicating that cLTP enhanced presynaptic activity, as previously reported (21, 22). We verified that treatment with sAβ1–42 (200 nM for 2 hours) did not cause cell death or defects in their excitability (fig. S2). In addition, forskolin increased the number of presynaptic boutons in control neurons (Fig. 1, A and B). Using FM 1-43 and retrospective immunostaining of a postsynaptic marker (SHANK), we confirmed that the newly formed presynaptic boutons were as functional as the preexisting boutons (Fig. 1, A to C) and were juxtaposed to or merged with SHANK-positive postsynaptic structures (Fig. 1, D and E), indicating that cLTP increased bona fide functional synapses. The preexisting presynaptic boutons neighboring the newly formed ones became smaller after cLTP induction, suggesting the redistribution of SVs to the newly formed ones (Fig. 1F). Under these conditions, sAβ1–42 blocked the stimulatory effect of forskolin on synaptogenesis (Fig. 1, G and H). Preincubation of sAβ1–42 with antibody 6E10, which recognizes Aβ, abolished the effect of sAβ1–42 (Fig. 1, G and H). These results indicate that sAβ1–42 inhibits cLTP-induced new synapse formation.

Fig. 1 sAβ1–42 inhibits cLTP-induced synaptogenesis in cultured rat hippocampal neurons at DIV16.

(A and B) Experimental protocol (A) and representative fluorescent images (B) of FM 1-43–loaded presynaptic boutons. Solid arrow heads, newly formed synapses; empty arrow heads, stable synapses. Scale bar, 10 μm. (C) Average destaining time course of FM 1-43 fluorescence intensity represented in (B), normalized to the maximal FM 1-43 signal. Decay time constant (τ) was determined by fitting the decay trace with a double exponential function. n = 4 independent assays. FSK, forskolin; APs, action potentials. (D and E) Experimental protocol (D) and representative images (E) of GFP-synaptophysin (green)–expressing neurons were treated with forskolin for 30 min, fixed, and stained with SHANK antibody (red). Empty arrow heads, newly formed GFP-synaptophysin–expressing presynaptic boutons; solid arrow heads, corresponding SHANK-positive postsynaptic structures; ICC, immunocytochemistry. Scale bar, 10 μm. (F) Relative fluorescence intensity of preexisting presynaptic boutons before and after forskolin treatment. (G) GFP-synaptophysin–transfected neurons were preincubated with sAβ1–42 alone or sAβ1–42 with 6E10 antibody for 2 hours and treated with 50 μM foskolin. Solid arrow heads, newly formed presynaptic boutons. Scale bar, 5 μm. (H)Number of presynaptic boutons and relative change (as %) (right) per 100 μm of axon length. Results from each of three independent experiment are shown (gray) along with the means ± SEM; *P < 0.05, **P < 0.01 by Student’s t test (C, F, and H).

sAβ1–42 inhibits intersynaptic vesicular trafficking

Because of the rapid appearance of new presynaptic boutons and a reduction in the size of preexisting boutons after cLTP (Fig. 1, A to F), we suspected that the numerous vesicles at new synapses are likely to have been recruited from neighboring synapses rather than to have been transported from the soma. Lateral sharing of SVs among neighboring synapses is critical to presynapse formation (1618). To test the effects of sAβ1–42 on intersynaptic vesicular trafficking, we used quantum dot (QD)–based single-SV tracking in cultured rat hippocampal neurons (Fig. 2A). Using this approach, we previously characterized the diffusive behavior of individual SVs (23), and we confirmed that, as reported previously (23), a single SV is labeled by a single QD and showed that this method does not affect the normal physiology of SV recycling (fig. S3). We found that intersynaptic vesicular trafficking was significantly reduced in sAβ1–42-treated neurons (Fig. 2, B to F), and preincubation with 6E10 largely blocked this effect (Fig. 2, B to F). Measurement of both mean square displacement (MSD) and diffusion coefficient over time displayed the inhibitory effect of sAβ1–42 on intersynaptic vesicular trafficking (Fig. 2, D to F). Subchronic treatment with sAβ1–42 (for 96 hours) had a similar effect (fig. S4, A and B), and the effect of sAβ1–42 was confirmed using an independent presynaptic probe, green fluorescent protein (GFP)–synaptophysin (fig. S4, C and D), as well as by using other Aβ1–42-specific antibodies (fig. S4, E and F).

Fig. 2 sAβ1–42 suppresses the intersynaptic vesicle trafficking.

(A to F) Cultured rat hippocampal neurons coexpressing synaptoPHluorin (sPH)–AP and endoplasmic reticulum–retained biotin ligase (BirA-ER) were treated as indicated and labeled with streptavidin-conjugated QD 605 (1 nM) at DIV14 to DIV16. (A) Experimental protocols for QD labeling and (top to bottom) (B) representative images of QD 605–loaded SVs, raw kymographs of QD 605–loaded SVs between stationary boutons (vertical lines), difference kymographs [DIFF(N − 1)], and cumulative QD flux (Σ-flux,; flattened image), with an overlay of QD 605–loaded SVs (green) and Σ-flux (red). Vertical scale bar, 20 s; horizontal scale bar, 5 μm. Imaging data were analyzed for the cumulative QD flux (a.u., arbitrary fluorescence units) (C), MSD versus time (D), average diffusion coefficient of QD 605–loaded SVs (E), and cumulative fraction of QD 605–loaded intersynaptic SVs versus diffusion coefficients (F). Data are means ± SEM; n = 4 independent experiments, each from ~50 boutons per group. (G to I) Cultured mouse hippocampal neurons from 5XFAD TG and LT control mice were transfected with sPH-AP and BirA-ER and labeled with streptavidin-conjugated QD 605. Shown are MSD of QD 605–loaded intersynaptic SVs (G), diffusion coefficients (H), and mean number (±SEM) (I) of mobile SVs in a 200 μm × 200 μm area of a group of cultured mice hippocampal neurons. Data are means ± SEM; n = 4 independent experiments per group. *P < 0.05, **P < 0.01 by Student’s t test (H and I) or analysis of variance (ANOVA) and Tukey’s post hoc test (C and E).

We further found that intersynaptic vesicular trafficking was strongly suppressed in hippocampal neurons derived from postnatal day 1 (P1) transgenic mice expressing five familial AD (FAD) mutations (5XFAD mice) (Fig. 2, G and H) (24) and that this effect was abolished by γ-secretase inhibitor X (fig. S5, A and B). The total number of mobile SVs was also lower in 5XFAD neurons (Fig. 2I). Consistent with this, secreted Aβ1–42 was detected at 14 days in vitro (DIV14) in supernatants of 5XFAD neurons but not in those of littermate (LT) neurons (fig. S5C). We also found that treatment of naturally secreted Aβ in the conditioned medium of 7PA2 cells (25, 26) to cultured rat hippocampal neurons inhibited intersynaptic vesicular trafficking (fig. S5, D to F).

sAβ1–42-induced phosphorylation of synapsin (Ser9) decreases synapsin-actin binding

Previous studies have indicated that Aβ alters cytoskeletal networks (27) and axonal transport (28). However, acute treatment with 200 nM sAβ1–42 affected neither mitochondrial trafficking nor the integrity of microtubule structures in cultured rat hippocampal neurons (fig. S6), suggesting that some other mechanisms were responsible for the inhibitory effect of sAβ1–42 on vesicle trafficking.

Synapsin is an SV-associated phosphoprotein that serves as a linker between actin and SVs (17, 29, 30). Protein kinase A (PKA)– and CaMKIV-mediated phosphorylation of synapsin dissociates it from actin and SVs (3032), thus making it a candidate regulator of intersynaptic vesicular trafficking. We found that application of sAβ1–42 to cultured rat hippocampal neurons increased synapsin phosphorylation on Ser9 (Fig. 3A) and that phosphosynapsin was more abundant in cultured 5XFAD hippocampal neurons than in LT neurons (Fig. 3B). To directly assess whether sAβ1–42 inhibited actin-synapsin binding, we performed immunoprecipitation assays using synapsin I antibody (Fig. 3C). As expected, pretreatment with sAβ1–42 reduced synapsin-actin binding, and this effect was countered by Aβ1–42-specific antibody (Fig. 3, C to E). These results suggest that sAβ1–42 dissociates an SV-synapsin-actin ternary complex that is important for intersynaptic vesicular trafficking.

Fig. 3 sAβ1–42 inhibits the synapsin and actin binding through synapsin phosphorylation at Ser9.

(A) Relative amount of phosphosynapsin (P-syn) to synapsin I (syn I) in cultured rat hippocampal neurons at DIV16. Blots are representative of four experiments. (B) Phosphosynapsin relative to synapsin I in cultured mice hippocampal neurons (DIV16) from LT and 5XFAD. Blots are representative of four experiments. (C) Representative blot from three independent experiments assessing coimmunoprecipitation of actin with synapsin I from cultured rat hippocampal neuron lysates at DIV16. Aβ1–42 ab (1), Aβ1–42-specific antibody (Aβ1–42-1); IgG, immunoglobulin G. (D) Mean amount of actin coimmunoprecipitated with synapsin I, relative to control group; n = 3 blots. (E) Relative amount of phosphosynapsin to synapsin I. Data were normalized to the control group; n = 3 blots. (F) Immunoblots of transfected human embryonic kidney (HEK) 293T cells to confirm expression of wild-type (WT) or mutant (S9A) synapsin Ia. (G and H) Neurons were cotransfected with GFP-synaptophysin and mCherry–synapsin Ia (WT or S9A) and treated with either vehicle or sAβ1–42 (200 nM). MSD curves versus time (G) and diffusion coefficients (H) were then assessed from more than or equal to seven independent experiments. Data are means ± SEM; *P < 0.05, **P < 0.01 by Student’s t test (A, B, D, and E) or ANOVA and Tukey’s post hoc test (H).

To prove the direct relationship between synapsin phosphorylation and intersynaptic vesicular trafficking, we cotransfected cultured rat hippocampal neurons with an mCherry-tagged phosphodeficient (S9A) mutant of synapsin Ia and GFP-synaptophysin (Fig. 3F). Overexpression of the phosphodeficient protein abrogated the inhibitory effect of sAβ1–42 on intersynaptic vesicular trafficking (Fig. 3, G and H). Together, these results show that altering synapsin I phosphorylation has a direct impact on intersynaptic vesicular trafficking.

Inhibition of the CaMKK-CaMKIV pathway prevents sAβ1–42-induced hyperphosphorylation of synapsin and CaMKIV

Both PKA and CaMKI/IV phosphorylate synapsin Ia at Ser9 (30), and an increased abundance of phospho-CaMKIV (Thr196) has been reported in AD models, such as the triple-transgenic AD mice and FAD mutant–expressing cells (33). We also observed a substantial increase in the abundance of phospho-CaMKIV (Thr196) in sAβ1–42-treated rat hippocampal neurons in culture (Fig. 4, A and B). In addition, in vitro kinase activity assays of CaMKIV purified from sAβ1–42-treated rat hippocampal neurons in culture revealed a more than twofold increase in CaMKIV phosphorylation in sAβ1–42-treated neurons compared with control neurons (fig. S7).

Fig. 4 The inhibition of CaMKK prevents the sAβ1–42-induced hyperphosphorylation of synapsin and CaMKIV.

(A) Representative Western blots of phosphosynapsin or phospho-CaMKIV in cultured rat hippocampal neurons at DIV16 treated as indicated for 2 hours. (B) Average relative amount of phosphosynapsin or phospho-CaMKIV relative to the respective total protein, each normalized to the control group. n = 6 blots. (C to E) Cultured rat hippocampal neurons were cotransfected with sPH-AP and BirA-ER, treated as indicated, and labeled with streptavidin-conjugated QD 605. (C) Representative [DIFF(N − 1)] kymographs showing the trafficking of QD 605–loaded SVs (diagonal lines) in each group; scale bars, 20 s (right) and 10 μm (bottom). (D) MSD curves versus time (D) and diffusion coefficients (E) were calculated from 1 min of tracking of QD-loaded intersynaptic vesicles. n = 4 independent experiments per group. (F) Immunoblot assay confirming the expression of Flag-tagged CaMKIV and dominant inhibitory CaMKIV (K75E) in HEK-293T cells. (G and H) Cultured rat hippocampal neurons were transfected with GFP-synaptophysin and treated as indicated. Representative kymographs (G) and diffusion coefficients (H) assessing the trafficking of GFP-synaptophysin–expressing SVs (diagonal lines) in each group. Scale bars, 10 s (right) and 20 μm (bottom). Data are means ± SEM; n = 4 independent experiments per group; *P < 0.05, **P < 0.01 by ANOVA and Tukey’s post hoc test. Asterisks at the bottom of the bars in (B) and (E) represent the significance compared to the control.

CaMK kinase (CaMKK) is the only kinase known to phosphorylate CaMKIV on Thr196, which increases CaMKIV activity (34, 35). We treated rat hippocampal neurons with a potent and specific CaMKK inhibitor, STO-609, for 2 hours and found that synapsin and CaMKIV phosphorylation was decreased, indicating that some CaMKIVs are phosphorylated by CaMKK under basal conditions (Fig. 4, A and B). We found, however, that simultaneous treatment with sAβ1–42 and STO-609 did not completely prevent the sAβ1–42-induced increase in synapsin and CaMKIV phosphorylation but allowed it to increase close to the control level (Fig. 4, A and B). This may be due to some CaMKKs activated by rapid sAβ1–42-induced Ca2+ influx. Accordingly, when neurons were pretreated with STO-609 before the addition of sAβ1–42, phosphorylation of synapsin and CaMKIV was inhibited in a similar extent to that observed with STO-609 alone (fig. S8, A to C).

Because no potent selective inhibitor of CaMKIV is currently available, we performed additional rescue experiments using the general CaMK inhibitor, KN-93 (fig. S8, D to H), which inhibits CaMKI, CaMKII, and CaMKIV with similar Ki values (36). We found that KN-93 alone did not affect basal phosphorylation of synapsin and CaMKIV, but it suppressed the sAβ1–42-induced hyperphosphorylation of CaMKIV and synapsin (fig. S8, D to F).

MSD and diffusion coefficient analysis showed that STO-609 and KN-93 rescued the inhibitory effect of sAβ1–42 on intersynaptic vesicular trafficking (Fig. 4, C to E, and fig. S8, G and H), as did a dominant inhibitory form of CaMKIV (K75E, a kinase-dead mutant) (Fig. 4, G and H), indicating that the abundance of phospho-CaMKIV and phosphosynapsin is inversely correlated with the extent of intersynaptic vesicular trafficking.

sAβ1–42 induces prolonged hyperphosphorylation of synapsin

Forskolin activates several targets including PKA, which leads to synapsin Ser9 phosphorylation (30). We found, however, that phosphorylated synapsin had a detrimental effect on cLTP-induced synaptogenesis, and this seemed to contradict with our other results. However, we then observed that a brief exposure of cultured rat hippocampal neurons to forskolin (5 min) increased synapsin phosphorylation, but this effect declined as treatment was prolonged (fig. S9, A and B). Even in neurons stimulated with forskolin for 30 min, synapsin phosphorylation was markedly higher when sAβ1–42 was present (fig. S9, C and D). Therefore, although the effect of forskolin on synapsin phosphorylation was transient, sAβ1–42 induced a lasting increase in synapsin phosphorylation, accompanied by defects in intersynaptic vesicular trafficking.

sAβ1–42 increases cytosolic Ca2+ concentration and impairs Ca2+ clearance

CaMKIV is activated by Ca2+-dependent calmodulin. We observed a marked increase of cytosolic Ca2+ concentration in presynaptic boutons of cultured hippocampal neurons in response to sAβ1–42 treatment, which was prevented by 6E10 (fig. S10, A and B). Furthermore, when calcium transients (CaTs) were evoked in Fluo-4 AM–loaded neurons by applying a single depolarizing pulse to mimic a single AP, CaT decay was slowed in sAβ1–42-treated neurons (fig. S10, C to E). These results indicate that sAβ1–42 increases basal Ca2+ concentration and perturbs Ca2+ clearance in presynaptic terminals, both of which may contribute to the hyperactivation of CaMKIV.

sAβ1–42 increases the size of the resting pool but not of the recycling pool

SV populations are classified into two groups: a recycling pool (a readily releasable pool and reserve pool) and a resting pool (37). When neurons were transfected with a vesicular glutamate transporter-1 fused with pHluorin, a modified GFP with high pH sensitivity, the amplitude of the response to a train of 1800 APs at 20 Hz in the presence of bafilomycin, a V-type adenosine triphosphatase inhibitor that blocks the acidification of endocytosed SVs, reflects the size of the total recycling pool, whereas the resting pool refractory to stimulation is uncovered by adding NH4Cl, which traps all of the vesicles in an alkaline state (37).

We found that the ratio of the recycling to the resting pool declined in sAβ1–42-treated rat hippocampal neurons in culture, as previously described (15), and both 6E10 and STO-609 prevented this change (Fig. 5, A to D). However, when we analyzed the raw intensity data without normalization (Fig. 5, E to H), we found that sAβ1–42 did not affect the size of the recycling pool but increased that of the resting pool (Fig. 5, I to K).

Fig. 5 sAβ1–42 increases the size of the resting but not that of the recycling pool.

(A) Representative pseudocolor images of vesicular glutamate transporter conjugated with ecliptic GFP (vGlut-pH)–transfected presynaptic boutons in cultured rat hippocampal neurons before and after 1800 APs [20 Hz, with bafilomycin A1 (Baf)], followed by NH4Cl. Scale bar, 10 μm. (B) Average time course of vGlut-pH fluorescence traces. Each pool size was normalized to the total vesicle pool size. (C and D) Average fraction values of recycling pool (C) and resting pool (D). (E to H) Distribution within a single bouton of recycling versus resting pool size. Slope was fitted by the linear plot. (I to K) Mean values of the raw intensity without normalization showing recycling pool size (I), resting pool size (J), and total pool size (K). Data are means ± SEM; n= 4 to 6 independent experiments per group (n > 300 boutons per independent experiment for each group). **P < 0.01 by ANOVA and Tukey’s post hoc test.

STO-609 restores the defects in cLTP-induced synaptogenesis

Finally, we tested whether STO-609 could prevent the defects in cLTP-induced synaptogenesis brought about by sAβ1–42. We found that the addition of STO-609 abolished the inhibitory effect of sAβ1–42 on cLTP-induced synaptogenesis in cultured rat hippocampal neurons (Fig. 6, A and B) and prevented the increase in total SV pool size induced by sAβ1–42 (Fig. 6C).

Fig. 6 Inhibition of CaMKIV phosphorylation rescues the sAβ1–42-induced presynaptic defects.

(A and B) GFP-synaptophysin–transfected rat hippocampal neurons at DIV16 were preincubated for 2 hours with sAβ1–42 alone or sAβ1–42 with STO-609 and then treated with foskolin (50 μM). Shown are time-lapse images (A) and the mean number of (left three graphs) and relative change (far right) in presynaptic boutons per 100-μm axon length (B). All three independent experiments for each group are noted (gray lines) along with the mean (bars; black lines). Solid arrowheads mark newly formed presynaptic boutons. Scale bar, 5 μm. (C) Box-and-whiskers graph of presynaptic boutons stained with synaptophysin antibody. n ≥ 19 independent experiments per group (n = ~70,000 boutons assessed in each group). Data are means ± SEM; *P < 0.05, **P < 0.01 by Student’s t test (B) or ANOVA and Tukey’s post hoc test (C).

DISCUSSION

Here, we report a novel cellular mechanism underlying presynaptic neuronal dysfunction caused by sAβ1–42 in early AD. Our results indicate that sAβ1–42 increases the Ca2+-dependent phosphorylation of Ser9 of synapsin through CaMKIV, thus releasing SVs and actin from synapsin and subsequently inhibiting intersynaptic vesicular trafficking (Fig. 7). Hence, by disrupting SV reallocation between synapses, sAβ1–42 prevents neurons from forming new synapses during plasticity. To our knowledge, this is the first demonstration of a relationship between intersynaptic vesicular trafficking and the synaptic defects observed in AD.

Fig. 7 Model of the mechanism by which sAβ1–42 impairs the intersynaptic vesicular movements.

Mobile SVs destined to the intersynaptic pool bind to the actin through the SV-synapsin-actin complex. By forming this complex, SVs can move to intersynaptic spaces because actin is suggested to function to convey mobile SVs through microtubule-based axonal transport machinery. Our data suggest that sAβ1–42 impedes Ca2+ clearance, which activates CaMKK-to-CaMKIV pathway, and subsequently phosphorylates synapsin, resulting in dissociation of the SVs-synapsin-actin complex. Hence, sAβ1–42 prevents SVs from entering into the vesicle pool for intersynaptic trafficking, which could prevent new synapse formation and synaptic plasticity.

Synapsin is a substrate for a number of kinases, including PKA, CaMKI, CaMKII, and CaMKIV, extracellular signal–regulated kinases 1 and 2, Src, mitogen-activated protein kinase, and cyclin-dependent kinases 1 and 5 (30). Under resting conditions, dephosphorylated synapsin serves as a linker between actin and SVs, and phosphorylation by PKA and CaMKIV dissociates synapsin from the actin and SVs (3032). We found that phosphorylation of synapsin on Ser9 was markedly increased in neurons treated with sAβ1–42 as well as in 5XFAD neurons. In addition, CaMKIV was significantly activated by sAβ1–42, and this observation agreed with the previous finding that CaMKIV is constitutively active in the 3XFAD brain (33).

Unlike other CaMKs, CaMKIV activity is regulated by rather a complex process. Binding of Ca2+/CaM leads to basal CaMKIV activation. Ca2+/CaM binding also exposes the CaMKIV activation loop, which is then phosphorylated by CaMKK, resulting in increased CaMKIV activity (35). Furthermore, its activation generates an autonomous Ca2+/CaM-independent enzyme that remains active for a period that outlasts the Ca2+ stimulus (35). It is also known that even in the absence of Ca2+/CaM, some fractions of CaMKIV exist in a complex with CaMKK (35, 38).

We found that a specific CaMKK inhibitor, STO-609, applied as a single agent to cultured neurons for 2 hours, decreased the phosphorylation of both synapsin and CaMKIV, indicating that some CaMKIVs are phosphorylated by CaMKK under basal conditions. However, here, we found a significant difference in the abundance of synapsin and the extent of CaMKIV phosphorylation and MSD values between neurons exposed to STO-609 alone and those exposed to sAβ1–42 and STO-609. We reasoned that when we treated neurons with sAβ1–42 together with STO-609, rapid Ca2+ influx induced by sAβ1–42 increased binding of Ca2+/CaM to CaMKIV, which could have led to (i) increased CaMKIV phosphorylation by some CaMKKs activated by rapid Ca2+ influx and (ii) generation of autonomous Ca2+/CaM-independent CaMKIVs. Therefore, it is possible that simultaneous treatment with sAβ1–42 and STO-609 does not completely prevent sAβ1–42-induced synapsin/CaMKIV phosphorylation but rather restricts it. To test this possibility, we pretreated neurons with STO-609 before adding sAβ1–42, expecting CaMKK to be fully inhibited by STO-609. As expected, pretreatment of STO-609 suppressed sAβ1–42-induced phosphorylation of synapsin and CaMKIV to a similar extent as that observed with STO-609 alone. Therefore, we believe that STO-609 not only inhibits basal phosphorylation of CaMKIV by CaMKK but also prevents sAβ1–42-induced phosphorylation. We further showed that overexpression of a dominant inhibitory CaMKIV and a phosphodeficient synapsin Ia mutant restored intersynaptic vesicular trafficking in the presence of sAβ1–42, confirming that CaMKIV-mediated signaling regulates intersynaptic vesicular trafficking.

sAβ1–42 triggers extracellular Ca2+ influx by altering membrane Ca2+ permeability, interacting with voltage-gated Ca2+ channels or forming Aβ pores (39, 40). We have observed that sAβ1–42 increased presynaptic Ca2+ in cultured hippocampal neurons. Because Ca2+ release from endoplasmic reticulum is also involved in the sAβ1–42-evoked Ca2+ rise (41), Ca2+ influx across the plasma membrane and Ca2+ release from internal stores may contribute to the hyperphosphorylation of both CaMKIV and synapsin.

Previous studies indicate that intersynaptic vesicular trafficking can contribute to synaptic strength by modulating the size of the functional SV pools in individual presynaptic terminals (18). Direct recruitment of intersynaptic vesicles can generate new functional synapses in mature neurons without perturbing the integrity of neighboring synapses (16). We found that sAβ1–42 strongly inhibited activity-dependent synaptogenesis and that STO-609 prevented this defect, suggesting that inhibition of intersynaptic vesicular trafficking is one of the molecular mechanisms underlying the sAβ1–42-induced defects in synaptic plasticity (42). Furthermore, we found out that sAβ1–42 specifically increased the size of the resting pools without changing the recycling pool size, indicating that SVs destined for the resting pool may participate in intersynaptic vesicular trafficking.

In conclusion, our results identify defective intersynaptic vesicular trafficking as a novel sAβ1–42-induced defect in presynaptic function relevant to the early stages of AD. Therefore, they may contribute to developing a treatment to prevent sAβ1–42-induced synaptic dysfunction in early-stage AD.

MATERIALS AND METHODS

Plasmid construction

The BirA-ER (43) and sPH (VAMP2 conjugated with ecliptic GFP) (44) plasmids were provided by A. Ting (Massachusetts Institute of Technology) and J. Rothman (Yale University). Acceptor peptide sequence (KKKGPGGLNDIFEAQKIEWH) that is recognized and biotinylated by the BirA-ER was conjugated to the luminal domain of sPH (sPH-AP). When exposed to extracellular space during exocytosis, acceptor peptide tag is labeled with streptavidin-conjugated QD 605, and tethered QD is internalized during endocytosis (23). Synapsin Ia was subcloned into the mCherry-C1 vector (provided by R. Tsien, University of California, San Diego). Flag-CaMKIV and CaMKIV (K75E) plasmids were provided by H.-S. Park (Chungnam National University, Korea). The fidelity of all DNA constructs was confirmed by sequencing.

Antibodies

The following antibodies were used: phospho-Ser9–synapsin (cat. no. 2311, Cell Signaling Technology), synapsin I (cat. no. 106 103, Synaptic Systems), phospho-Thr196–CaMKIV (cat. no. Sc-28443-R, Santa Cruz Biotechnology), CaMKIV (cat. no. ab3557, Abcam), mCherry (cat. no. ab167453, Abcam), β-tubulin (cat. no. ab11367, Abcam), 6E10 (cat. no. SIG-39300, Covance), Aβ1–42-1 (cat. no. 218 721, Synaptic Systems) or Aβ1–42-2 (cat. no. ab10148, Abcam), synaptophysin (cat. no. 101 011, Synaptic Systems), actin (cat. no. A4700, Sigma-Aldrich), and glutathione S-transferase (GST; custom-made). SHANK antibody was provided by E. Kim (Korea Advanced Institute of Science and Technology, Korea).

sAβ1–42 preparation

sAβ1–42 was prepared as previously described (15, 45, 46). Synthetic Aβ1–42 peptide (Bachem) was dissolved in 1 mM HFIP (1,1,1,3,3,3-hexafluoro-2-propanol; Sigma-Aldrich) and incubated at room temperature for 1 hour. Aliquots were evaporated in the fume hood for 2 hours, dried under the vacuum for 10 min, and stored at −20°C. Peptide film was resuspended to 1 mM in dimethyl sulfoxide for 10 min, diluted to 100 μM by adding Ham’s F-12 (Invitrogen), and incubated over 12 hours at 4°C. Before treatment, it was mixed with neurobasal medium to the final concentration of 200 nM. To confirm the synthesis of sAβ1–42, it was loaded onto 4–12% Bis-Tris NuPAGE gels and transferred to polyvinylidene difluoride (PVDF) membranes (Bio-Rad), which were then boiled for 5 min in phosphate-buffered saline (PBS) and incubated with 50% blocking solutions (LI-COR Biosciences) and 50% tris-buffered saline. Membranes were incubated with 6E10 antibody, and the protein bands were visualized with enhanced chemiluminescence reagent (ECL; AbClon). To eliminate any sAβ1–42 effect, sAβ1–42-containing medium was preincubated with 6E10 (Covance) and Aβ1–42-specific antibody 1 (Synaptic Systems) or Aβ1–42-specific antibody 2 (Abcam) for 2 hours at room temperature.

Hippocampal neuron culture, transfection, and image acquisition

Hippocampal neurons were dissociated from embryonic day 18 fetal fetal SD rats, as previously described (23). Briefly, hippocampi were dissociated with papain and plated on poly-d-lysine–coated coverslips and were grown in neurobasal medium with 2% B-27 (Invitrogen), 0.5 mM l-glutamine, and 4 μM 1-β-d-cytosine-arabinofuranoside (Sigma-Aldrich). Neurons were transfected using a modified calcium-phosphate method, as previously described (23). Briefly, 6 μg of complementary DNA (cDNA) and 9.3 μl of 2 M CaCl2 were mixed in distilled water to a total volume of 75 μl, and the same volume of 2×BBS [50 mM BES, 280 mM NaCl, and 1.5 mM Na2HPO4 (pH 7.1)] was added. The cell culture medium was completely replaced by transfection medium [minimum essential medium (MEM); 1 mM pyruvate, 0.6% glucose, 10 mM glutamine, and 10 mM Hepes (pH 7.65)], and the cDNA mixture was added to the cells and incubated in a 5% CO2 incubator for 60 min. Cells were washed twice with washing medium (pH 7.35) and then returned to the original culture medium. All animal experiments were performed in accordance with the guidelines set by the Institute of Animal Care and Use Committee of Seoul National University, Korea.

Time-lapse images were acquired with an Olympus IX-71 microscope (Olympus) with a 40×, 1.0–numerical aperture oil lens using an Andor iXon 897 EMCCD camera (Andor Technologies) driven by MetaMorph Imaging software (Molecular Devices) using an excitation filter (475AF40) and an emission filter (605WB20, Omega Filters). All solutions included 10 μM 6-cyano-7-nitroquinoxaline-2,3-dione to prevent any recurrent excitation.

5XFAD transgenic mice model

5XFAD transgenic mice were provided by I. Mook-Jung (Seoul National University) and J.-S. Han (Konkuk University). 5XFAD hippocampal neurons were cultured from P1 mice, and tail genotyping was performed with PS1 transgene targeting primers [AATAGAGAACGGCAGGAGCA (forward) and GCCATGAGGGCACTAATCAT (reverse)] to distinguish 5XFAD mice from LT mice. Media from cultured LT and 5XFAD neurons were collected at each DIV7 and DIV14, and the concentration of secreted Aβ1–42 in each medium was measured by using human Aβ1–42 enzyme-linked immunosorbent assay (ELISA) kit following the manufacturer’s instruction (KHB3441, Invitrogen).

γ-Secretase inhibitor X treatment

Hippocampal neurons from 5XFAD mice and their LT controls were cultured. At DIV14, 50% of cultured medium was replaced with the equal volume of new complete neurobasal medium with the final concentration of 0.2 μM γ-secretase inhibitor X (Calbiochem), then the neurons were incubated for 48 hours before the assay.

Preparation of 7PA2 conditioning medium

The protocol for medium conditioning was followed as previously described (25). Briefly, 7PA2 cells (provided by D. Selkoe, Harvard Medical School) were grown in Dulbecco’s modified Eagle’s medium (DMEM) with 10% fetal bovine serum (FBS) and G418 (200 μg/ml). Cells were washed with Dulbecco’s PBS when they reached 90% confluence and incubated in a fresh neurobasal medium for 16 hours. Conditioned medium was collected and filtered for debris removal with a 0.22-μm filter. Aβ1–42 concentration was measured by ELISA (Invitrogen). Control Chinese hamster ovary cells were grown in DMEM with 10% FBS without G418, followed by the same conditioning steps with 7PA2.

Cell death assay

Cultured hippocampal neurons were treated with either 200 nM sAβ1–42 for 2 hours or 10% MEM with 90% salt-glucose-glycine medium (114 mM NaCl, 0.22% NaHCO3, 5.3 mM KCl, 1 mM MgCl2, 2 mM CaCl2, 10 mM Hepes, 1 mM glycine, 30 mM glucose, and 0.5 mM sodium pyruvate) for 3 days. After a brief washout, neurons were fixed in 4% paraformaldehyde and then stained with 4′,6-diamidino-2-phenylindole (2 μg/ml) for 15 min at room temperature to measure the number of pyknotic nuclei.

Chemical LTP

Neurons at DIV14 to DIV16 were first imaged for 5 min before cLTP induction. Then, 50 μM forskolin was applied to induce cLTP, and the same neurons were imaged for 30 min with 5- or 10-min intervals. The relative bouton number (%) was defined as the number of bouton after forskolin treatment divided by the initial number of bouton.

FM 1-43 loading and unloading

Neurons at DIV16 were stimulated with 600 APs at 10 Hz in the presence of 10 μM FM 1-43 (Invitrogen) in tyrode (136 mM NaCl, 2.5 mM KCl, 2 mM CaCl2, 1.3 mM MgCl2, 10 mM Hepes, and 10 mM glucose), kept in the presence of dye for an additional 30 s after stimulation to label poststimulus endocytosed vesicles, and washed out in low-Ca2+ and high-Mg2+ tyrode (119 mM NaCl, 2.5 mM KCl, 0.5 mM CaCl2, 10 mM MgCl2, 25 mM Hepes, and 30 mM glucose) with 1 mM ADVASEP-7 (Sigma-Aldrich) for an efficient removal of the cell surface–bound FM 1-43. After a 10-min resting period, 1200 APs at 10 Hz were given to unload and measure the amount of loaded FM 1-43. Decay time constant (τ) was calculated by fitting the decay trace with a double exponential function.

SV labeling with QDs

QD labeling of SVs was done, as previously described (23). Briefly, cultured neurons were cotransfected with sPH-AP and BirA-ER at DIV9. At DIV14 to DIV16, neurons were pretreated with 100 nM of free streptavidin in cold tyrode solution for 5 min to block the surface sPH-AP. Then, neurons were washed and incubated with 1 nM streptavidin–conjugated QDs in 90 mM high KCl [31.5 mM NaCl, 90 mM KCl, 2 mM CaCl2, 2 mM MgCl2, 25 mM Hepes, 30 mM glucose (pH 7.4)] for 1 min, followed by 10 min of washing to remove the unbounded QDs. Single QD labeling of a single SV was confirmed as described in fig. S3.

Determination of the number of sPH-AP-QDs in a single SV

Ten picomolar QD was embedded in a 1% agarose gel, and then, this mixture was mounted in an 18 mm × 18 mm square glass coverslip. QDs that show a characteristic blinking behavior were selected, and their intensity was measured. Neurons were labeled with 100 pM of QD-streptavidin, and intensity of their QD photoluminescence was measured. The unitary intensity of QD photoluminescence in an agarose gel and that of QDs in the neuron were compared.

MSD and diffusion coefficient analysis

Single QD tracking was performed with MetaMorph Imaging software, in which the center of the spot fluorescence was determined using a Gaussian fit (23). Analysis and fitting were performed with MetaMorph and Origin 9.0 (OriginLab). MSD was calculated from the following formula:MSD(nτ)=1Nni=1Nn[(x((i+n)τ)x(iτ))2+(y((i+n)τ)y(iτ))2]where xi and yi are coordinates of an object on frame, N is the total number of steps in the trajectory, and τ is the acquisition time. Diffusion coefficients (D) were calculated by fitting the first five points of the MSD curves versus time (τ) with the equation MSD(nτ) ≈ 4Dnτ.

Mitochondria tracking

To measure the mitochondria movement, DsRed-mito (provided by G. Yoon, Ajou University)–transfected neurons were imaged at DIV14 to DIV16 every 5 s for 500 s and analyzed for the number of mobile mitochondria and their diffusion coefficients in 135 μm × 135 μm field of view. Mitochondria were considered as highly mobile if they moved more than 5 μm during 5 min. Diffusion coefficients of mobile mitochondria were calculated in the same way as QD-labeled SV.

Immunocytochemistry

Neurons at DIV14 to DIV16 were fixed in 4% paraformaldehyde in 4% sucrose–containing PBS for 15 min at room temperature and permeabilized for 5 min in 0.25% Triton X-100. Then, neurons were blocked for 30 min with 10% bovine serum albumin (BSA) in PBS at 37°C and incubated with primary antibody diluted in 3% BSA in PBS overnight at 4°C, followed by incubation with secondary antibody in 3% BSA/PBS for 45 min at 37°C. The fluorescence intensity of individual boutons was measured using MetaMorph with double-blinded manner and further analyzed using Origin 9.0. For SHANK staining, neurons were fixed in precooled 10% MES solution [100 mM MES (pH 6.9), 1 mM EGTA, and 1 mM MgCl2] in methanol for 5 min at −20°C.

Microtubule staining

For microtubule staining, neurons at DIV14 to DIV16 were fixed with precooled methanol for 10 min at −20°C, and then, they were blocked in 10% BSA in PBS for 20 min at 37°C. Cells were incubated with β-tubulin antibody diluted in 3% BSA in PBS overnight at 4°C and followed by the incubation with Alexa 488–conjugated secondary antibody (Invitrogen) diluted in 3% BSA in PBS for 45 min at 37°C.

Western blotting

Cultured neurons at DIV14 to DIV16 or transfected HEK-293T cells were lysed with 1% Triton X-100 lysis buffer with 1% serine and threonine phosphatase inhibitors (Sigma-Aldrich). After sonication, lysates were centrifuged at 14,000g at 4°C for 20 min to collect only the supernatant and were loaded onto 10% polyacrylamide gels, transferred to PVDF membranes (Pall Life Sciences), and incubated with primary antibodies for overnight at 4°C. After washing in TBST (tris-buffered saline with Tween 20), PVDF membranes were incubated with horseradish peroxidase–conjugated secondary antibody (Jackson Immunoresearch Laboratories) for 1 hour at room temperature. Immunoreactivity was detected with an ECL using LAS 4000 (GE Healthcare) and quantified with ImageJ.

Immunoprecipitation

Neurons were lysed in a lysis buffer and centrifuged for 20 min at 14,000g at 4°C. Equal amounts of the total cell lysates were incubated with synapsin I antibody (Synaptic Systems) overnight at 4°C, incubated with protein A–Sepharose (GE Healthcare) for 1 hour, pelleted by centrifugation, and analyzed by SDS–polyacrylamide gel electrophoresis (PAGE). Proteins on the gels were transferred onto PVDF membranes and incubated with the antibody against actin (Sigma-Aldrich) for 1 hour at room temperature. Immunoreactivity was detected with an ECL using LAS 4000.

Nonradioactive CaMKIV activity assay

N-terminal 1– to 20–amino acid sequence of synapsin I (MNYLRRRLSDSNYMANLPNG) was cloned into the pGEX4T-1 vector after the amplification by polymerase chain reaction, and the ligation product was transformed into BL21. Protein expression was induced by adding 0.5 mM isopropyl β-d-1-thiogalactopyranoside overnight at 25°C. After induction, cells were sonicated in the hypotonic buffer [20 mM tris-HCl (pH 8), 150 mM NaCl, 1 mM EGTA, 1 mM MgCl2, 1% Triton X-100, and 0.1% sodium deoxycholate] with protease inhibitor and centrifuged for 20 min at 13,200 rpm. The supernatants were collected and incubated with GST bead for 2 hours at 4°C. After washing, GST bead–conjugated proteins were eluted with an elution buffer [20 mM glutathione, 100 mM tris-HCl (pH 8), 120 mM NaCl, and 10% glycerol] overnight at 4°C. To purify the CaMKIV or P-CaMKIV, control- or sAβ1–42-treated neurons were lysed with 1% Triton X-100 lysis buffer with protease inhibitor and phosphatase inhibitor cocktail (Sigma-Aldrich), followed by the sonication and centrifugation. The concentration of proteins was measured by bicinchoninic acid assay kit, and the equal amounts of total cell lysates were each incubated with CaMKIV- or P-CaMKIV–specific antibodies for 1 hour and 30 min at 4°C and subsequently with protein A–bead for 1 hour at 4°C. Purified CaMKIV or P-CaMKIV was incubated with 2 μg of substrate protein (GST–synapsin I; 1 to 20 amino acids) in a kinase buffer (10 mM MgCl2, 1 mM CaCl2, and 0.2 mM adenosine 5′-triphosphate) for 30 min at 30°C. After incubation, reactions were terminated with 5× sample buffer with β-mercaptoethanol, and the products were loaded onto SDS-PAGE gels (12%) after heating at 100°C for 5 min. The abundance of phosphorylated substrate proteins was measured using the P-synapsin (Ser9) antibody.

Ca2+ measurements

Hippocampal neurons were cotransfected with DsRed-VAMP2 and a genetically encoded calcium indicator, GCaMP6f (47). Colocalization images were acquired before and after the pretreatment of sAβ1–42 with or without 6E10. For Ca2+ clearance experiments, the cultured hippocampal neurons were loaded with a 0.5 μM calcium indicator, Fluo-4 AM (Invitrogen), for 15 min at 37°C. After 10 min of washing in tyrode, time-lapse images were taken with a 0.2-s interval during the single action potential (1 AP) stimulation.

SV pool size measurement

vGlut-pH is provided by J. Rubenstein of the University of California, San Francisco. Ecliptic GFP is a modified GFP with high pH sensitivity (44). At rest, the fluorescence of ecliptic GFP is quenched by the acidic intraluminal pH (pH ~5.5). Upon stimulation, vesicles fuse with the plasma membrane, exposing their intraluminal proteins to the basic extracellular medium (pH 7.4), resulting in an increase in fluorescence. During endocytosis, the rapid reacidification of vesicles leads to a decrease in ecliptic GFP fluorescence, as it is quenched once again. To estimate the size of each fraction of the SV pool, vGlut-pH–transfected neurons at 16 days were stimulated with 1800 APs (20 Hz) in the presence of 0.5 μM bafilomycin A1 to release the entire recycling pool (bafilomycin was dissolved in Me2SO to 0.2 mM and diluted to a final concentration of 0.5 μM before the experiments). The change in fluorescence intensity to the plateau reflects the entire recycling pool. The resting pool was uncovered by applying 50 mM NH4Cl to unquench all acidic SVs that have not been released. Fluorescence intensity was normalized to the maximum fluorescence change after NH4Cl treatment (Fig. 5, A to D).

Statistical analysis

Unless otherwise indicated, data are given as means ± SEM, with n indicating the number of independent experiments. Numerical values represented in the graphs are provided in table S1. Statistical comparisons were performed with Origin 9.0 and SPSS (IBM) software. The normality of data distribution was checked with Kolmogorov-Smirnov normality test and normality plot. If normality assumption was satisfied, Student’s two sample t test was performed for comparisons between two independent groups, and one sample t test was done to compare each group with its corresponding control. For multiple group comparison, one-way ANOVA followed by Tukey’s post hoc test was performed to correct type I error inflation. The relevant P values are reported in the figure panels and legends. All statistical comparisons were confirmed by a statistician at the Medical Research Collaborating Center (MRCC) at the Seoul National University Hospital.

SUPPLEMENTARY MATERIALS

www.sciencesignaling.org/cgi/content/full/10/487/eaam8661/DC1

Fig. S1. Forskolin induces presynaptic potentiation (cLTP).

Fig. S2. Acute treatment of nanomolar sAβ1–42 does not induce cell death or alter the neuronal excitability.

Fig. S3. Determination of the QD 605–labeling method.

Fig. S4. Additional confirmation of sAβ1–42-induced inhibition of the intersynaptic vesicle mobility and chronic sAβ1–42 treatment effect.

Fig. S5. The effect of γ-secretase inhibitor and naturally secreted Aβ on intersynaptic vesicular trafficking.

Fig. S6. Nanomolar concentration of sAβ1–42 does not disrupt the mitochondria transport or microtubule structures.

Fig. S7. Nonradioactive CaMKIV activity assay.

Fig. S8. Inhibition of CaMKK-to-CaMKIV pathway rescues the sAβ1–42-induced defects.

Fig. S9. PKA-dependent synapsin Ser9 phosphorylation.

Fig. S10. sAβ1–42 increases presynaptic Ca2+.

Table S1. Numerical data underlying each graph in the figures.

This is an article distributed under the terms of the Science Journals Default License.

REFERENCES AND NOTES

Acknowledgments: We are grateful to numerous colleagues credited above for giving us reagents, to the Biomedical Imaging Center at the Seoul National University College of Medicine for the microscope services, and to the MRCC at the Seoul National University Hospital for their support with statistical analyses. Funding: This research was supported by grants from the Biomembrane Plasticity Research Center (no. 20100029395) to S.C. funded by the Ministry of Science, ICT and Future Planning and from the Brain Research Program (NRF-2015M3C7A1028790) to S.C. through the National Research Foundation of Korea. This work was also supported by the Education and Research Encouragement Fund of the Seoul National University Hospital. Author contributions: D.P. and S.C. designed the experiments. D.P., M.N., J.A.K., U.L., E.C., and M.J. performed the experiments. D.P. and S.C. analyzed the data, and D.P., J.A.K., and S.C. wrote the manuscript. All authors read and approved the final manuscript. Competing interests: The authors declare that they have no competing interests.
View Abstract

Navigate This Article