Regulation of proteostasis by sleep through autophagy in Drosophila models of Alzheimer’s disease

Similar to humans, Drosophila follows a 12-h day/12-h night rhythm (25). Sleep in flies is measured as 5 min or more of inactivity (26, 27). To study Drosophila locomotor activity and sleep we used a Drosophila Activity Monitor (TriKinetics) using a locally built detection system that includes an infrared (IR) beam to record the locomotor activity of solitary flies placed in glass tubes (Fig S1A and B), as commonly used in measuring fly sleep (28). Using the activity monitoring system, we first determined the sleep patterns of tauopathy models. Drosophila models for tauopathy, established by expressing either wild-type (Tau^WT) or mutant (Tau^R406W) human Tau protein in neurons, recapitulate adult-onset, progressive neurodegeneration with altered lifespan, and accumulation of pTau (29). We expressed either UAS-CD8-GFP, UAS-Tau^WT, or UAS-Tau^R406W pan-neuronally using elav-GAL4 driver or in the photoreceptors using GMR-GAL4 driver as previously reported (29, 30), and monitored Drosophila activity behavior for 7 d, which included 2 d of acclimation and 4 d of sleep monitoring (Fig 1A–F). This experiment was designed to examine the early effects of sleep modulation before the onset of severe neurodegeneration in tauopathy models (30). Collectively, combining pan-neuronal (elav-GAL4) or cell-type specific expression, in this case, photoreceptor expression of Tau (GMR-GAL4) models with successful sleep modulation allows the mechanistic dissection of the impact of sleep modulation on neurodegeneration in Drosophila models of tauopathy.

Pan-neuronal expression of Tau^WT or Tau^R406W with elav-GAL4 caused significant sleep behavior changes (Fig 1A), specifically in reduced average sleep time per 24 h of 117 min less in Tau^WT, and 197 min less in Tau^R406W (Fig 1B); increased number of sleep bouts (Fig 1C), and decreased length of each sleep bout (Fig 1D). Combined, these results show increased sleep fragmentation and decreased total sleep length, suggesting that Tau^WT and Tau^R406W expression in all neurons causes sleep disturbances, consistent with previous reports that have observed sleep disturbances caused by pan-neuronal expression of Aβ or Tau (31, 32, 33). Such significant change in sleep behavior makes it challenging to dissect the direct effects of sleep modulation on neurodegeneration. To focus on the effects of sleep modulation on neurodegeneration, and minimize the influence of Tau expression on sleep disturbance, we incorporated an additional tauopathy model, by expressing Tau in a subset of neurons with minimum impact on sleep behavior. To that end, we analyzed whether expressing Tau in the photoreceptors using GMR-GAL4 caused sleep disturbances by monitoring their sleep behavior (Fig 1E). When comparing Tau^WT or Tau^R406W to controls (GFP), there was no significant difference in either the amount of average sleep per 24 h (Fig 1F), the number of sleep bouts (Fig 1G), or the average length of each sleep bout (Fig 1H). Detailed analyses of sleep divided into night and day showed that expression of Tau in the photoreceptors had minimum effects on their innate sleep behavior (Fig S2A–C) and no significant change in number of sleep bouts (Fig S2C) or bout length (Fig S2D). Together, these data suggest that expression of Tau in the photoreceptors causes minimum effect on sleep, making it a feasible model for us to examine the molecular impact of sleep modulation on Tau aggregation in a system with relatively normal sleep behavior. Moreover, in Drosophila, gaboxadol induces sleep through the GABA_A receptors, specifically, Ligand-gated chloride channel homolog 3 (Lcch3) and the GABA and glycine-like receptor (Grd) (34). Previous research has shown these receptors are present in interneurons postsynaptic to photoreceptors (35, 36). Therefore, incorporating the expression of Tau in the photoreceptors, using GMR-GAL4 as a driver, also allows the confirmation that any synaptic effects observed when feeding gaboxadol are because of increased sleep and not caused by direct activation of GABA receptors.

To establish a sleep modulation paradigm that induces significant changes in sleep time, we incorporated both sleep deprivation and sleep induction protocols. Sleep disruption was achieved using a mechanical “deprivator” that rotates 90° for a determined time (37). During mechanical deprivation, we can observe significant reduction in sleep, followed by an earlier sleep onset and increased daytime sleep. These changes in sleep profile are because of the sleep homeostat attempting to reduce the amount of sleep need accumulated during a deprivation event. Sleep induction was achieved by gaboxadol feeding following previously published methods (20). Flies were subjected to either mechanical sleep deprivation for two nights, 9 h each night (hours 111–120 and hours 135–140) or gaboxadol feeding at 0.1 mg/ml concentration for 4 d (hours 48–156) (Fig 2A). Sleep studies showed significant sleep behavior alterations after sleep deprivation or sleep induction in all groups (Fig 2B). On average, sleep deprivation resulted in a reduction in total nighttime sleep by 160, 160, and 95 min for GFP, Tau^WT, and Tau^R406W, respectively; whereas sleep induction resulted in an increase in nighttime sleep of 174, 185, and 145 min, respectively. Daytime activity showed similar changes by sleep deprivation or induction (Fig 2C). Detailed sleep bout analyses showed that the number of sleep bouts was decreased by sleep deprivation, when increased by sleep induction (Fig 2D), without significant changes in sleep bout length (Fig 2E). Taken together, these results showed successful sleep modulation after mechanical deprivation or sleep induction for all groups and demonstrated the feasibility of the optimized sleep modulation paradigm.

It is important to note that all experiments were started using adult flies of 2 DAE (days after eclosion) and cellular and molecular analyses were performed 4 h after the conclusion of sleep modulation (160 h) on 8 DAE (Fig 2A). This paradigm was designed to examine the early effects of sleep modulation before the onset of severe neurodegeneration in tauopathy models (30).

Synaptic loss is one of the early events of neurodegeneration and main lesions presented in human AD patients (38). We have previously reported that our Tau expression models recapitulate the phenotype of Tau-induced synaptic degeneration (29). To evaluate how sleep modulation affects synaptic integrity and function we expressed Tau^WT or Tau^R406W in the Drosophila visual system to take advantage of the highly organized parallel axons of the compound eye: the R1–R6 photoreceptors axons cross the lamina cortex and make synaptic connections at the lamina neuropil, whereas R7–R8 extended their axons beyond the lamina and project to the medullar neuropil (39, 40). The lamina synapses can be visualized in the x-y and x-z planes (Fig 3A). Drosophila lamina synapses are organized into repetitive structures called “cartridges” reflecting the compound eye organization. Each cartridge contains an assemble of synapses formed between six photoreceptor neurons (one of each R1-R6) and the lamina monopolar cells (39). Analyzing synaptic morphology at the x-y and x-z planes provides a three-dimensional view of the synaptic structure. Using Bruchpilot (BRP), an endogenous active zone-associated cytomatrix protein as a marker, we analyzed synaptic structure under different sleep modulation conditions with confocal imaging (Fig 3B).

The synaptic architectural integrity as indicated by tissue thickness of the lamina neuropil was altered by sleep modulation. Whereas sleep modulation had no significant effect on tissue thickness in normal control (GFP) group, sleep deprivation caused an ∼11.5% reduction, and sleep induction an ∼9.5% increase in tissue thickness of Tau^WT. In Tau^R406W, sleep deprivation caused a ∼4.6% reduction, whereas sleep induction showed a ∼5.7% increase in tissue thickness, on average (Fig 3C).

Measurement of total BRP intensity of the lamina neuropil in the x-z plane, in the normal control (GFP) group showed no change after sleep deprivation but a ∼56% increase after sleep induction, suggesting a physiological effect of sleep induction on active zone BRP homeostasis. In tauopathy groups (Tau^WT and Tau^R406W), the BRP level was reduced by 32% after sleep deprivation and increased by 46–55% after sleep induction (Fig 3D). Moreover, we quantified the average BRP intensity per lamina cartridge and found a Tau expression-induced, significant disruption of cartridge integrity and reduced BRP intensity per lamina cartridge. This reduction was exacerbated with sleep deprivation and significantly improved after sleep induction for both Tau^WT and Tau^R406W (Fig 3E). Taken together, these results suggest sleep deprivation exacerbated Tau-induced loss of active zone structures in the presynaptic terminals, whereas sleep induction significantly improved the localization of endogenous BRP at the synapses.

To further extend the analysis of the effects of sleep modulation on synaptic integrity, we analyzed the synaptic localization of endogenous cysteine-string protein (CSP), a synaptic vesicle-associated chaperone critical for neurotransmission (41) (Fig S3A). Quantification of CSP intensity per lamina cartridge revealed a significant reduction in CSP per lamina cartridge in both Tau^WT and Tau^R406W groups. Furthermore, after sleep deprivation, a significant reduction in CSP was observed in the Tau^WT group. Lastly, a significant increase was observed after sleep induction in the Tau^R406W group (Fig S3B). Consistent with BRP levels, these results suggest that sleep deprivation exacerbated overall synaptic loss and integrity and sleep induction provided significant improvement of synaptic structures.

To examine the functional consequence of sleep modulation, we carried out electroretinogram (ERG) study of synaptic function. These electrical responses caused by light stimulation provide information about phototransduction (receptor potential, RP) and pre- and postsynaptic responses of the laminar monopolar cells (On and Off transients) (23, 24, 42). Consistent with previous reports, tauopathy flies showed reduced On and Off transients and receptor potential under control conditions, indicating synaptic degeneration (43). Notably, these defects were partially ameliorated by sleep induction, which showed significant increases in the amplitude of On and Off transients, in all groups (Fig 4A–D). On the other hand, sleep deprivation showed no significant ERG changes. The minimal change caused by sleep deprivation might indicate that a stronger or more prolonged deprivation effect is required to observe substantial functional degeneration at this early stage of the disease. The improvements in On and Off transients after sleep induction suggest improved synaptic function, which is consistent with the overall improvement in synaptic morphology that was previously described. Similarly, deprivation did not result in RP amplitude deterioration, but sleep induction showed a slight RP amplitude increase in normal control GFP flies (Fig 4D), indicating that sleep induction has a general effect on all synapses.

Taken together, our morphological and functional study of the synapse demonstrates that sleep modulation is altering the progression of Tau-induced synaptic degeneration. Specifically, sleep deprivation accelerates synaptic degeneration as shown by exacerbated loss of BRP and CSP and overall synaptic structural integrity, whereas sleep induction improves overall synaptic physiology and structural integrity.

To dissect the molecular underpinnings of the effect of sleep modulation on protein homeostasis we first analyzed how sleep modulation affected general protein processing using the reporter GFP protein. In our tauopathy model, we co-expressed mCD8-GFP, a membrane-targeted GFP as a reporter to monitor the neuronal membrane integrity (Fig 5A). In the control group where only GFP was expressed, normal photoreceptor morphology is marked by GFP fluorescence, showing the neuronal membrane of photoreceptor neurons R1–6, which extend into lamina neuropil, and R7–R8 terminals which are observed in the medulla neuropil (Fig 5B). Importantly, the GFP intensity and localization pattern were largely unchanged by sleep modulation, suggesting the minimal effect of sleep modulation on protein homeostasis in normal healthy neurons. However, in Tau-expressing flies, after sleep deprivation there was a significant increase in GFP in the lamina for the Tau^R406W group, and a significant reduction after sleep induction (Fig 5C). In addition, quantification of the intensity at the R7/R8 terminals in the medulla neuropil shows a significant increase after sleep deprivation in the Tau^WT group and a significant reduction after sleep induction in the Tau^R406W group (Fig 5D). Taken together these results show that sleep modulation had minimal effects on reporter GFP homeostasis in normal conditions, whereas significantly alters protein homeostasis under pathological Tau-expressing conditions. The increase in reporter GFP accumulation after sleep deprivation could indicate a block in protein processing and toxic protein buildup in tauopathy conditions, whereas the decrease in reporter GFP after sleep induction could suggest enhanced clearance of misfolded GFP proteins and reduction in toxic protein burden.

To dissect the molecular underpinnings of the effect of sleep modulation on Tau-induced synaptic degeneration, we set out to examine the biochemical changes of Tau. Expression of Tau^R406W results in Tau hyperphosphorylation and synaptic aggregation of hyperphosphorylated pTau (30). We have optimized a high-resolution imaging approach that allows the characterization of the properties of pTau and the dynamic interaction with cellular organelles (30, 44). Using pTau-specific antibody AT8 (for Ser202, Thr205) that is specific for Tau pared helical fragments (PHF-tau), we quantitatively analyzed overall optic lobe pTau intensity and individual size distribution and intensity of pTau clusters to examine changes in the aggregational state of Tau caused by sleep modulation (Fig 6A). The total intensity of pTau showed a remarkable alteration under sleep modulation with a significant increase after sleep deprivation and a modest reduction after sleep induction in Tau^R406W (Fig 6B). When comparing individual pTau clusters a significant increase in intensity was observed after sleep deprivation for Tau^WT and a slight increase in Tau^R406W (Fig 6C). The size analysis of pTau cluster revealed an interesting effect of sleep modulation, where sleep deprivation significantly increased medium and large clusters in Tau^R406W and Tau^WT, respectively, whereas sleep induction decreased medium and large-sized clusters in Tau^WT (Fig 6D).

Taken together, these results show that when compared with controls, tauopathy flies subjected to sleep deprivation have a significantly higher accumulation of pTau deposition in the brain. In contrast, sleep induction reduced the accumulation of hyperphosphorylated pTau specifically in medium and large clusters and dampened pTau cluster accumulation in axons. When combined with the results observed in Fig 5, these results suggest altered protein processing after sleep modulation.

To uncover the mechanism underlying the reduction in pTau accumulation and the protective effects of sleep induction, we assessed the impact of sleep induction on Tau protein levels. To enable biochemistry analysis, we expressed Tau pan-neuronally using elav-GAL4. For the following experiments expressing Tau pan-neuronally, we first analyzed the sleep pattern and confirmed that these flies presented sleep disturbances as shown in Fig 1. After 4 d of sleep monitoring (Fig S4A), flies expressing Tau showed significant reduction in average sleep (Fig S4B), increased sleep fragmentation (Fig S4C) and decreased sleep length (Fig S4D). We then subjected the flies to sleep modulation and analyzed their sleep profiles (Fig 7A). Quantification of sleep per 12 h showed significant sleep reduction after sleep deprivation for the normal control GFP group but showed no significant deprivation for Tau-expressing groups. This is likely because of the already significant sleep reduction and fragmentation because of Tau expression and the additional deprivation has reached the ceiling of the sleep disruption. In contrast, with sleep induction treatment, all groups showed significant sleep increase after gaboxadol feeding (Fig 7B), allowing detailed biochemical analysis of the potential neuroprotective effects of sleep induction. Western blot analysis of Tau expression levels (Fig 7C) showed no changes in protein levels in the deprived group but a significant decrease in total Tau (5A6) (Fig 7D) and hyperphosphorylated pTau, using S262 probe (Fig 7E) after sleep induction, consistent with behavioral data. The significant reductions in both total Tau and pTau levels through sleep induction suggest Tau protein clearance as a key mechanism underlying the beneficial effects of sleep increase.

Misfolded Tau proteins are usually degraded through multiple pathways such as the autophagy-lysosomal system and the ubiquitin-proteosome system (5). A common early step in these degradation pathways is the ubiquitination of Tau to mark the protein for degradation. Our previous study has shown the different oligomeric states of ubiquitinated Tau and identified the strong toxicity associated with low molecular weight Tau oligomers (29). We next examined the effects of sleep induction on the oligomerization of Tau species using biochemical approaches. Tau was immunoprecipitated by human Tau-specific antibody from lysates of Tau^WT or Tau^R406W expressing brains and probed from ubiquitin. Western blot analysis identified ubiquitinated Tau oligomeric species and found a reduction in low molecular weight toxic oligomers after sleep induction, when compared with control (Fig 7F). These results suggest that sleep induction alters Tau aggregation and promotes Tau clearance. These results are consistent with previous reports showing that enhanced autophagy can suppress Tau accumulation (45).

A large body of literature has demonstrated the role of autophagy in neurodegeneration and highlighted the disturbance of autophagic flux in AD and related disorders (46, 47, 48, 49). To assess how sleep induction impacts autophagic flux in tauopathy models, we incorporated an in vivo autophagy dual-color reporter expressing Atg8 tandem tagged with pH-sensitive GFP and mCherry at its amino-terminal (UAS-GFP-mCherry-Atg8a) (45, 50). Because GFP signal is quenched under the acidic environment, combination of GFP and mCherry signal intensity will mark the stages of autophagy flux, where autophagosomes will be high GFP and high mCherry, whereas fusion with the lysosome will be low GFP and high mCherry (Fig 8A).

We established autophagy reporter in tauopathy models by co-expressing UAS-GFP-mCherry-Atg8a with either UAS-lacZ (control) UAS-Tau^WT or UAS-Tau^R406W pan-neuronally using elav-Gal4. We first analyzed the baseline sleep profiles to assess whether reporter expression caused changes in sleep behavior (Fig 8B). Analysis of their activity behavior showed a significant reduction in nighttime and daytime sleep for Tau-expressing groups when compared with lacZ (Fig 8C). The activity profile is similar to the pan-neuronal tauopathy models (Fig 1A–D), suggesting that pan-neuronal Tau expression was the major driver of sleep disturbances, and the Atg8a reporter expression had minimum impact on sleep in tauopathy models.

Using confocal imaging and ratiometric-based quantification (44) we analyzed the fluorescence of Atg8-positive neuronal cell bodies in the fly midbrain region as described previously (44) (Fig 8D). The GFP and mCherry fluorescence intensity plotted is the combined fluorescence of all Atg8-positive organelles present per neuron. The results were plotted in XY for the GFP fluorescence (Y) as a function of mCherry fluorescence (X) (Fig 8E). Quadrants were divided based on intensity, where Quadrant I, high GFP-high mCherry marks autophagosome; Quadrant II, low GFP-high mCherry marks autolysosome; and Quadrant III, low GFP-low mCherry marks lysosomal degradation. The gating thresholds for quadrants were set based on the control group using the following principles. (1) The design of tandem reporter of pH-sensitive GFP and pH-insensitive mCherry defines the stoichiometry of GFP/mCherry as ≤ 1. Therefore, the content of Quadrant IV (GFP/mCherry ≥ 1) is 0. Puncta with high GFP signal and no mCherry signal were likely the result of GFP-mCherry protein truncation and, therefore, were excluded from the analysis. (2) The distribution of Atg8-positive particles reflects the distribution of autophagy-lysosome pathway-related compartments in wild-type neurons (44, 51).

Compared with the lacZ control group, Tau^WT- and Tau^R406W-expressing neurons showed a significant increase in Quadrant I, autophagosome accumulation, and a significant reduction in Quadrant II, fusion of the autophagosome with the lysosome (Fig 8F). Moreover, the significant increase in GFP/mCherry ratio in Tau expression neurons suggests a higher pH level and dysfunction of the cellular organelles with acidic pH (Fig 8G). Together, these results reveal a shift in Atg8a puncta distribution in Tau pathology and further suggest a potential block in autophagy flux at the step of fusion of autophagosome and lysosome (Fig 8H). The finding of reduced fusion and a consequent increase in autophagosome accumulation and an increase in GFP/mCherry ratio would be consistent with Tau^WT or Tau^R406W expression-induced impairment of autophagic flux.

To examine the cellular impact of sleep modulation on autophagy flux, we first evaluated the control condition where lacZ and GFP-mCherry-Atg8 reporter were co-expressed. Sleep profiles confirmed that sleep modulation was successful in Atg8 reporter expressing flies (Fig 9A), with a significant sleep decrease after sleep deprivation and a significant increase in sleep after induction (Fig 9B). Quantitative confocal imaging analysis of Atg8a puncta in the neuronal cell body region in the midbrain (Fig 9C), showed a significant shift in puncta distribution under sleep modulation (Fig 9D and E). Specifically, sleep deprivation caused a significant drop in Quadrant II, but no change in Quadrant III, whereas sleep induction caused a remarkable drop in Quadrant II and a concomitant increase in Quadrant III, suggesting a significant reduction in puncta in the fusion state, and a significant increase in puncta undergoing lysosomal degradation (Fig 9E). Moreover, there was a significant increase in GFP/mCherry ratio after sleep deprivation and induction (Fig 9F). An increase in ratio could be because of either increase in GFP because of increased autophagosome accumulation or a decrease in mCherry because of increased degradation. These results suggest that under normal (non-pathological) conditions, the effects of sleep deprivation on autophagy are minimal, whereas sleep induction appears to promote autophagic flux and may facilitate lysosomal degradation (Fig 9G).

Next, we analyzed how sleep modulation affects autophagic flux in tauopathy models. In Tau^WT expression (moderate) model, sleep modulation was successful with sleep induction (Fig 10A and B). Imaging analysis of the Atg8a dual reporter expressed in the brain (Fig 10C) showed a significant increase in autophagosome puncta (Quadrant I) after sleep deprivation, accompanied by a decrease in Quadrant II. Moreover, after sleep induction a decrease in Quadrant II, with a concomitant increase in Quadrant III, was observed (Fig 10D and E). In addition, the GFP/mCherry ratio was increased after sleep modulation (Fig 10F). These results suggest that under moderate pathology sleep deprivation leads to increased autophagosome accumulation and decreased fusion with the lysosome, whereas under sleep induction, an increase in fusion and degradation occur (Fig 10G). In the Tau^R406W expression (severe) model, sleep modulation was successful with sleep induction (Fig S5A and B), similar to that in the Tau^WT expression (moderate) model (Fig 10A and B). However, analysis of autophagy flux using the Atg8a reporter revealed differences between the two pathological states (Fig S5C). Specifically, no significant changes were observed after sleep deprivation, whereas after sleep induction, we observed a decrease in Quadrant II (autolysosome) (Fig S5D and E) and an increase in GFP/mCherry ratio after sleep modulation (Fig S5F). These results suggest that under severe tauopathy pathology, sleep modulation had a limited effect on autophagic flux because of the system being severely degenerated. Although we observe minor changes after induction, there is high variability in the results (Fig S5G). Taken together, these results suggest that under moderate Tau pathology, sleep modulation promotes autophagic flux.

Flies were reared on cornmeal-molasses-yeast medium at 22°C, 65% humidity, with 12-h light/12-h dark cycles. The following strains were used in this study: UAS-Tau^WT, UAS-Tau^R406W obtained from Dr. Mel Feany (29). UAS-GFP-mCherry-Atg8a, GMR-GAL4, and elav-GAL4 were obtained from Bloomington Drosophila Stock Center.

Male flies were placed in single glass tubes as previously described (61). Locomotion was recorded using activity monitors (DAM2, Trikineticks Inc.). Monitors were placed in an environmental chamber maintained at 25°C and 70–80% relative humidity. In all experiments, flies were first entrained in 12-h light/12-h dark conditions for 2 d. After entrainment, locomotion was measured with 20 s binning for 4 d. Data from monitors were visualized and processed with custom-written Matlab (MathWorks Inc.) scripts.

Mechanical deprivation was adapted from previously described Sleep Nullifying Apparatus methods (37). Flies were subjected to two episodes of 9 h of deprivation. For sleep induction flies were fed either vehicle or 0.1 mg/ml gaboxadol that was mixed with normal fly food (20).

Fly brains were dissected in PBS (pH 7.4) as previously described (44). Samples were fixed in 4% formaldehyde for 10 min and washed in PBS containing 0.4% vol/vol Triton X-100 (PBTX). Samples were then incubated with primary antibodies diluted in 0.4% PBTX with 5% normal goat serum at 4°C overnight, followed by incubation with secondary antibodies diluted in 0.4% PBTX with 5% normal goat serum at 4°C overnight, and DAPI (1:300, D1306; Invitrogen) staining at room temperature for 10 min. The following primary antibodies were used: mouse anti-BRP (NC82,1:250; Developmental Studies Hybridoma Bank), mouse anti-CSP (6D6,1:250; Developmental Studies Hybridoma Bank), AT8 human PHF-Tau (MN1020, 1:250; Thermo Fisher Scientific). The following secondary antibodies were used: Alexa Fluor 555-conjugated anti-mouse (A21422, 1:300; Invitrogen) and Alexa Fluor 647 anti-mouse secondary antibody (A21235, 1:300; Invitrogen). The samples were mounted on glass slides with VECTASHIELD Antifade Mounting Medium (Vector Laboratories). Slides were imaged using an Olympus IX81 confocal microscope with a 40x or 60x oil immersion objective lens with a scan speed of 8.0 µs per pixel and spatial resolution of 1,024 × 1,024 pixels. Images were processed using FluoView 10-ASW (Olympus). Quantification of the number, area and intensity of fluorescently tagged proteins was carried out using ImageJ/Fiji (1.53q) using previously published methods (44). For quantification of tissue thickness of the lamina neuropil, three regions of interest (ROI) were averaged. For BRP intensity in the x-z plane, the lamina neuropil was chosen as the ROI. For intensity of the x-y plane in BRP and CSP staining, a 3 × 3 ROI was measured and divided by 9 to obtain average intensity per lamina cartridge.

For quantification of autophagic flux using tandem reporter previously published methods were followed (44, 51). The plot was divided into four quadrants using gating thresholds of GFP = 750,000 and mCherry = 250,000. The gating thresholds for quadrants were set based on the control group using the following principles. (1) The design of tandem reporter of pH-sensitive GFP and pH-insensitive mCherry defines the stoichiometry of GFP/mCherry as ≤ 1. Therefore, the content of Quadrant IV (GFP/mCherry ≥ 1) is 0. (2) The distribution of Atg8-positive particles reflects the distribution of autophagy-lysosome pathway-related compartments in WT neurons (44, 51).

Previously described methods were followed (42, 62). Flights were anesthetized with CO₂ and immobilized on a glass slide. A recording electrode with 3 M NaCl was placed on the surface of the left eye, and another reference electrode was inserted into the thorax. After 5 min of dark adaptation, flies were given 1-s light stimulation (Digitimer), with 5 s of dark between each stimulus a total of 10 traces were recorded per fly and analyzed by pCLAMP 10 Electrophysiology Data Acquisition & Analysis Software (ver.10.5).

Ten fly heads of each genotype were collected and flash-frozen before extraction. They were homogenized in radioimmunoprecipitation assay buffer (Sigma-Aldrich), protease inhibitor cocktail (Sigma-Aldrich) and a phosphatase inhibitor cocktail (Roche). For immunoprecipitation experiments, 30 8 DAE male fly heads were homogenized in lysis buffer, and incubated with Protein-G beads conjugated with 10 μg of antibody per sample overnight at 4°C; the bead pellets were collected and washed four times with lysis buffer before eluting out the beads fraction with 4x Lamelli buffer, heating the samples at 95°C and centrifuging for 5 min to collect the bead-eluted fraction. Lysates were probed with anti-human tau (5A6, 1:1,000; Developmental Studies Hybridoma Bank). For Western blotting, 10 μl of total protein lysate per sample (equivalent to 1 head) was loaded onto a polyacrylamide gel (PAGE) and resolved by SDS–PAGE and transferred onto nitrocellulose membrane. After blocking at room temperature for 1 h, the membrane was incubated with a primary antibody at 4°C overnight, followed by a secondary antibody for 1 h at room temperature. Imaging was performed on an Odyssey Infrared Imaging System (LI-COR Biosciences) and analyzed using Image Studio (v4.0). Primary antibody dilutions were used as follows: ubiquitin (3936, 1:1,000; Cell signaling), 5A6 anti-human tau (5A6, 1:250; Developmental Studies Hybridoma Bank), Phospho-Tau Ser²⁶² (44-750G, 1:1,000; Thermo Fisher Scientific), anti-β-actin (A1978, 1:5,000; Sigma-Aldrich). Secondary antibodies used were IRDYE800CW anti-mouse (610-145-002, 1:10,000; Rockland) and IRDYE700DX anti-rabbit (611-144-002, 1:10,000; Rockland).

Biological sample size (n) and P-values are indicated in the corresponding figure legends. Either one-way ANOVA with Tukey post hoc test or Brown-Forsythe and Welch ANOVA with Dunnett’s multiple comparison correction was applied to compare multiple groups. P < 0.05 was considered statistically significant. All statistical analyses were performed in GraphPad Prism software (version 10.0).