Application of optogenetic Amyloid-β distinguishes between metabolic and physical damages in neurodegeneration

The brains of Alzheimer’s disease patients show a decrease in brain mass and a preponderance of extracellular Amyloid-β plaques. These plaques are formed by aggregation of polypeptides that are derived from the Amyloid Precursor Protein (APP). Amyloid-β plaques are thought to play either a direct or an indirect role in disease progression, however the exact role of aggregation and plaque formation in the aetiology of Alzheimer’s disease (AD) is subject to debate as the biological effects of soluble and aggregated Amyloid-β peptides are difficult to separate in vivo. To investigate the consequences of formation of Amyloid-β oligomers in living tissues, we developed a fluorescently tagged, optogenetic Amyloid-β peptide that oligomerizes rapidly in the presence of blue light. We applied this system to the crucial question of how intracellular Amyloid-β oligomers underlie the pathologies of A. We use Drosophila, C. elegans and D. rerio to show that, although both expression and induced oligomerization of Amyloid-β were detrimental to lifespan and healthspan, we were able to separate the metabolic and physical damage caused by light-induced Amyloid-β oligomerization from Amyloid-β expression alone. The physical damage caused by Amyloid-β oligomers also recapitulated the catastrophic tissue loss that is a hallmark of late AD. We show that the lifespan deficit induced by Amyloid-β oligomers was reduced with Li+ treatment. Our results present the first model to separate different aspects of disease progression.


Introduction
Alzheimer's disease (AD) is a debilitating, age-associated, neurodegenerative disease which affects more than 46.8 million people worldwide and represents the 6th leading cause of death in the United States of America (Ahn et al., 2001;De-Paula et al., 2012;Hawkes, 2016;Kumar et al., 2015;Zhang et al., 2011). Despite extensive efforts over the last 50 years, no disease-modifying therapy has been found and, most recently, several high-profile Phase III clinical trials have failed (Anderson et al., 2017;Lim and Mathuru, 2018;Park et al., 2018). The AD clinical trial landscape has largely been dominated by the amyloid cascade hypothesis, with more than 50% of the drugs targeting amyloid beta (Ab) in Phase III trials alone . The amyloid cascade hypothesis, in its original form, posits the deposition of Ab, in particular the extracellular Ab plaque, as the main driver of AD (Hardy and Higgins, 1992). However, the causative role of Ab plaques has recently been challenged, because of the failure of numerous interventions targeting Ab plaques in Phase III trials and the observation of Ab plaques in brains of non-AD symptomatic individuals (Cummings, 2018). Model organisms ranging from nematodes to mice have further shown that AD pathology can be modeled in the absence of obvious Ab plaques (Duff et al., 1996; Teo et al., 2019). These observations point to a different mechanism for Ab neurotoxicity perhaps through Ab's intracellular accumulation (LaFerla et al., 2007).
Amyloid plaques are macroscopic extracellular protein aggregates, but intracellular Ab must function in smaller units of peptides, oligomers or aggregates. Of these, soluble Ab oligomers appear to be the main toxic species in AD (Ferreira and Klein, 2011). For example, Ab oligomers have been shown to induce neurotoxic effects including memory loss in transgenic animal models reviewed in Mroczko et al. (2018). These studies, however, mostly rely on in vitro injection of synthetic Ab oligomers or oligomers extracted from AD brains (Mroczko et al., 2018). There is currently a lack of tools that can directly control Ab oligomerization in vivo, which would allow direct examination of the effects of Ab oligomerization.
Here, we describe an optogenetic method to study Amyloid-b (Ab) protein oligomerization in vivo in different model organisms and apply it to delineating mechanisms of disease progression. Optogenetics hinges on light responsive proteins that can be fused to genes of interest, allowing for the spatial and temporal regulation of proteins in a highly precise manner. Temporal-spatial control is achieved simply by the exposure of the target system to lights of a specific wavelength, without the need to introduce other external agents (Fenno et al., 2011;Mö glich and Moffat, 2010). One such optogenetic protein, a modified version of the Arabidopsis thaliana cryptochrome 2 (CRY2) protein, oligomerizes into photobodies quickly and reversibly in the presence of blue light at 488 nm (Más et al., 2000). In cultured cells, expression of the photolyase homology region of CRY2 fused to mCherry led to puncta within 10 s of exposure to blue light and these puncta dispersed within minutes (Bugaj et al., 2013). This basic unit of CRY2-mCherry attached to a variety of proteins have, for instance, been used to study cortical actin dynamics in cell contractility during tissue morphogenesis, the dynamics of Wnt and EGF signaling, plasma membrane composition and transcriptional regulation (Bugaj et al., 2018;Bugaj et al., 2015;Guglielmi et al., 2015;Huang et al., 2017;Idevall-Hagren et al., 2012;Johnson et al., 2017;Kaur et al., 2017;Kennedy et al., 2010).

Results
To address the role of aggregation of intracellular Ab in the pathophysiological effects of AD, we developed a light-inducible system for use in model organisms. Specifically, we generated an in vivo, light-dependent, oligomerization switch for the formation and dissolution of Ab oligomers in Drosophila melanogaster, Caenorhabditis elegans and Danio rerio (Figure 1a). This optogeneticallydriven model allowed the visualization of microscopic Ab oligomerization and addressed the question of their relationship to Ab pathogenesis with spatiotemporal precision.

Generation of transgenic models with light-inducible intracellular Ab oligomerization in model organisms
We generated transgenic animals expressing the 42-amino-acid human Ab peptide (Ab1-42) fused to Cryptochrome 2 and the fluorescent protein mCherry (Ab-CRY2-mCh, Figure 1a, please note the relative sizes). Drosophila lines were under the control of the GAL4/UAS system (Brand and Perrimon, 1993) and expressed either ubiquitously via AD-Gal4 driver (Figure 1b  Drosophila embryos using a lightsheet microscope with a 488 nm laser activating CRY2 and the 561 nm laser imaging mCh over time. We observed clusters of mCh fluorescence forming in embryos exposed to 488 nm light, but fewer in embryos that were not exposed to blue light (Figure 1-figure supplement 3, Videos 1-2). The intensity of the clusters was quantified using Imaris quantification software to show the intensity of oligomer formation (Figure 1c, note color scale for cluster intensity).
We proceeded to test the effectiveness of light induced clustering of Ab-CRY2-mCh in C. elegans. As this construct is driven by a heat shock promoter, worms were heat shocked at 35˚C for 90 min followed by resting at 20˚C before being immobilized and imaged using confocal microscopy ( Figure 1f-h). We observed an increase in clusters in worms exposed to blue light over their siblings not exposed to blue light. The results were quantified as fluorescence intensity and represented in (Figure 1h) showing the formation of Ab clusters.
Induction of chaperones from heat-shock can impact the dynamics of Ab. In particular, heat-shock treatment in another transgenic C. elegans strain, CL4176, reduced Ab-mediated paralysis and Ab oligomerization (Winblad, 1989;Wu et al., 2010). Overexpression of chaperone protein suppressed Ab-toxicity in C. elegans (Fonte et al., 2008). The use of hsp-16 promoter in the transgenic C. elegans, which requires heat-shock induction, may hence be a limitation of our current approach. In view of this limitation, we performed several other controls to understand the impact of heat-shock treatment on Ab dynamics, and whether or not heat-shock treatment itself contributes to the metabolic and phenotypic detriments observed in the transgenic animals.
To understand how heat-shock treatment may impact the dynamics of Ab in our transgenic strains, we compared the dynamics of Ab between the transgenic Ab worms with and without heat shock. We found that transgenic animals in both heat-shock (Ab HS L) and non-heat-shock (Ab -HS L) condition expressed detectable Ab-CRY2-mCh protein (Figure 1-figure supplement 2), however animals that had been heat-shocked had significantly higher Ab levels compared to non-heatshocked animals, suggesting that heat-shock, as intended, drove higher level of Ab expression and that this was not compensated for by secondary induction of chaperones.
Even though the CRY2 system predominantly oligomerizes, we also observed the formation of intracellular Ab aggregates, using a direct stain for Ab aggregates (Figure 1-figure supplement 4). Importantly, we observed that Ab aggregation was not reversible as had previously been observed for signaling molecules Johnson et al., 2017;Kaur et al., 2017), rather CRY2 appeared to initiate clustering leading to Ab bundles that did not come apart once blue light was turned off. We further confirmed that turning on blue light alone without the transgene expression did not lead to any Ab expression or cluster formation (Figure 1-figure supplement 2b, Ab HS L vs Ab -HS L), suggesting that the oligomerization process was CRY2-specific and not an artifact of Video 1. Elav-Gal4/UAS-Ab-CRY2-mCh embryos kept in the dark and imaged for neurons in red mCherry and glial cells Repo-QF2 >QUAS GFP. Blue laser power kept low to allow imaging of glial cells, but not high enough to activate aggregation. https://elifesciences.org/articles/52589#video1 Video 2. Elav-Gal4/UAS-Ab-CRY2-mCh embryos exposed to light and imaged for neurons in red mCherry and glial cells Repo-QF2 >QUAS GFP. Blue laser power at higher setting to allow imaging of glial cells and activate aggregation.
https://elifesciences.org/articles/52589#video2 light. We performed FRAP experiments to establish the stability of the Ab clusters ( Figure 2). When compared to an unrelated CRY2 construct, we observed a very slow recovery for Ab-CRY2-mCh compared to Arm-CRY2-mCh ( Figure 2; Kaur et al., 2017). Together, our results showed that the Ab-CRY2-mCh transgene was functional in both organisms, with a high specificity to blue light inducing irreversible oligomerization of Ab.

Light-inducible Ab oligomerization causes lifespan and behavioral deficits in transgenic models
To test the functionality of Ab-CRY2-mCh, we next looked for phenotypes associated with lightinduced intracellular Ab oligomerization. We performed lifespan studies to assess the effects of Ab-CRY2-mCh by expressing Ab-CRY2-mCh either ubiquitously (AD-Gal4) or specifically in the nervous system (Elav-Gal4). Newly eclosed adult flies were separated into two groups in each case, one of which was exposed to ambient white light and the other kept in the dark. In both cases, flies exposed to light died much more quickly with half the mean and maximum lifespans (Figure 3a-b) compared to those reared in the dark. We extended this analysis to C. elegans, though starting at Day 6 due to the heat-shock manipulation step. Under light conditions, C. elegans showed markedly decreased lifespans as compared to the transgenic C. elegans kept in dark conditions ( Figure 3c). Most significantly, we observed that the lifespan decrease was reversible, as taking animals exposed to light for 24 days and moving them into darkness showed a recovery of their lifespan, or a rescue ( Figure 3a-b). To ensure that these reduced lifespans were physiological, we examined the fitness of the animals. Fitness was severely decreased in light-exposed flies and worms as quantified by fertility and locomotive assays (Figure 3-figure supplement 1). Interestingly, light-induced Ab oligomerization caused further impairment in sensorimotor function of transgenic Ab nematodes ( Figure 3-figure supplement 1c). These results suggesed that light-induced oligomerization of Ab not only affected lifespan and fitness, but also impaired complex sensorimotor function and behavior.

Light-inducible Ab oligomerization leads to physical and metabolic damage in transgenic models
The lethality mediated by light-induced Ab oligomerization led us to examine the morphological effects of Ab-CRY2-mCh clustering during embryonic stages in Drosophila. Embryos expressing Ab in only the nervous system developed at a normal pace and did not show any morphogenetic defects and even hatched in the absence of blue light (Video 1). In contrast, sibling embryos exposed to blue laser light arrested in late neurogenesis stages where the central nervous system stopped developing and the embryos died (Video 2 and Figure  The severity of the physical damage phenotype in embryos exposed to blue light (Figure 4a-d'') led us to investigate the differences in mechanism between the two conditions by exploring inflammatory response, mitochondrial health and oxidative damage. Expression of Ab alone, without lightinduced Ab oligomerization, led to a small increase in inflammatory markers ( Figure 4e; Westfall et al., 2018). In embryos exposed to light, two of these markers increased to a much greater level (Figure 4e). We observed a significant reduction in the number of mitochondria in light-treated Ab animals ( Figure 4f). Markers of oxidative damage in light and dark exposed C. elegans showed differential expression of antioxidant defense genes, specifically Catalase, Trx-2 and Sod-3, suggestive of the occurrence of oxidative stress induced by Ab oligomerization ( Figure 5). These results suggested that light-induced Ab oligomerization phenocopied many of the hallmarks of AD.
We proceeded to look at light-induced Ab oligomerization in metabolic impairments in transgenic C. elegans. Heat-shocked mutants exposed to light (Ab HS L) had lower ATP levels compared to the non-transgenic control (Ctrl -HS D) ( Figure 6). However, heat-shocked mutants in the dark condition (Ab HS D) did not show any significant differences in ATP level compared to the non-transgenic control. This result suggested that presence of Ab alone is insufficient to induce ATP deficits, and that light-induced oligomerization of Ab is required for the defect to manifest. Ab expression alone was also insufficient to affect the nematodes' maximum respiratory capacity, as heat-shocked mutants in the dark condition (Ab HS D) did not display differences in maximum and spare respiration capacity compared to control animals (Ctrl -HS D). However, light-induced Ab oligomerization significantly reduced maximum respiration capacity in C. elegans. Heat-shocked mutants in the light condition (Ab HS L) displayed significantly lower maximum and spare respiration capacity compared to controls (Ctrl -HS D) and to heat-shocked mutants in the dark condition (Ab HS D) (Figure 6b-d).
Together, these results highlighted the metabolic differences between the two conditions and showed that metabolic deficits were mediated by light-induced Ab oligomerization.
To further examine whether the metabolic deficits in transgenic C. elegans were mediated by heat-shock treatment, we included a heat-shocked non-transgenic control in the metabolic and phenotypic assays. We found that heat-shock treatment does not affect ATP levels in the control animals, suggesting that the ATP detriment was a result of Ab expression, but not heat-shock treatment ( Figure 6a). On the other hand, heat-shock treatment appears to reduce the respiratory capacity of the control animals, suggesting that the decline in respiratory capacity was mediated by both heat-shock and Ab expression (Figure 6c). Comparing the heat-shocked transgenic animals in light and dark conditions, we found that light treatment, which induces oligomerization, worsened the phenotypic and metabolic detriments in the transgenic animals. This suggested that regardless of the effects of heat-shock and chaperone induction, oligomerization still negatively impacted the phenotypes.

Light-inducible Ab lifespan decrease is rescued by Li +
The drug Li + is used for treatment of neurological conditions (Licht, 2012). A high drinking water Li + concentration correlated with lower human mortality (Zarse et al., 2011), and Li + increased the lifespan of C. elegans and Drosophila (Castillo-Quan et al., 2016;McColl et al., 2008;Tam et al., 2014;Teo et al., 2020a;Teo et al., 2020b;Zarse et al., 2011). The molecular mechanisms by which Li + functions include inhibition of inositol monophosphotase (IMPA) and glycogen synthase kinase-3 (GSK-3) (Kerr et al., 2018), reduction of oxidative damage (Kasuya et al., 2009;Kerr et al., 2017;Khan et al., 2015) and longevity via a GSK-3/NRF-2 dependent, hormetic mechanism (Castillo-Quan et al., 2016). Among the mechanisms for neuroprotection, the Wnt signaling pathway plays a role in modulating AD pathology and its progression (De Ferrari et al., 2003;Jin et al., 2017;  Kaplan-Meier survival curves of transgenic Ab Drosophila adults driven by AD-Gal4 kept in the dark, exposed to light, exposed to light for 24 days then moved to the dark and exposed to light but fed with a lithiumsupplemented diet. (b) Kaplan-Meier survival curves of transgenic Ab Drosophila adults driven by Elav-Gal4 kept in the dark, exposed to light, exposed to light for 24 days then moved to the dark and exposed to light but fed with Figure 3 continued on next page Parr et al., 2015;Toledo and Inestrosa, 2009). Activation of Wnt signaling through lithium (Li + ) treatment attenuated Ab aggregation and increased the lifespan of Ab models (Toledo and Inestrosa, 2009). By inhibiting the enzyme glycogen synthase kinase-3 (GSK-3), Li + compounds drive the downstream production of intracellular b-catenin activating the Wnt-b-catenin signaling pathway (Klein and Melton, 1996). Li + treatment through Wnt activation also rescued behavioral impairment and neurodegeneration induced by Ab fibrils (De Ferrari et al., 2003). Taken together, these studies point to a putative AD therapeutic intervention through Li + . To test the usefulness of optogenetic Ab for drug testing, we supplemented the food for Drosophila expressing Ab-CRY2-mCh with Li + . Drosophila exposed to light, but consuming Li + had significantly increased lifespans, suggesting both a rescue effect induced by Li + (Figure 3a-b), and that the optogenetic model can be used for inducible expression drug testing. Rapid drug testing can be also be achieved through cell culture. We extended this assay to HEK293 cells transfected with Ab-CRY2-mCh where Ab clusters formed upon stimulation with blue light (Figure 7a). The number of clusters per cell was greatly reduced by the addition of Li + or the specific GSK3 inhibitor CHIR99021, but CHIR99021 additionally reduced the signal intensity of the clusters (Figure 7b-c, quantified 7d). These data demonstrated the potential use of the optogenetics system for drug testing.
Finally, we also extended these findings to a vertebrate system. We generated a zebrafish permanent line with UAS:Ab-CRY2-mCh in the transgenic TgBAC(gng8:GAL4) c416 background (Hong et al., 2013) that limited expression to a few cells. We observed a diffuse mCh fluorescence in a small subset of neurons in the olfactory epithelium, the interpeduncular nucleus and in a few neurons sparsely distributed in the forebrain at 5 dpf. Upon blue light exposure (488 nM) the detectable intensity of the fluorescence increased rapidly (Figure 8-figure supplement 1). Several neurons among those expressing appeared to bleb and began to die within 40 to 50 min of initial exposure to the blue light (Video 4). Such damage was not observable in matched controls expressing Ab-mCh driven by a CMV promotor lacking the CRY2 protein (Video 5). Our approach leads to the expression of Ab-CRY2-mCh in only a small subset of neurons making analysis of AD hallmarks difficult, so we used transient expression of Ab-CRY2-mCh driven by a ubiquitin promotor in 48 hr old embryos to quantify the effects of light-induced Ab oligomerization on mitochondria health and metabolic deficits. Similar to the impairments observed in C. elegans, Ab-CRY2-mCh expressing embryos showed lower maximum and spare respiration capacity compared to un-injected control (Figure 8). In addition, light exposure of Ab-CRY2-mCh expressing embryos caused a significant reduction in ATP levels ( Figure 8a).

Discussion
Aggregation of Ab has been studied extensively, but our results represent the first study to demonstrate that induced oligomerization of intracellular Ab can be used to separate different pathologies (induced by Ab oligomerization vs expression alone). A second approach to making inducible aggregates using a chemically controlled fluorescent protein has been published, but not yet widely tested (Miyazaki et al., 2016). Other previous approaches lacked an inducible oligomerization tool and could only demonstrate Ab oligomers' toxicity through exogenous injection of Ab oligomers (Mroczko et al., 2018). We use a tool to investigate the pathological effects of intracellular Ab oligomerization in nematodes and flies, human kidney cells, and in the vertebrate model organism zebrafish (Figure 8-figure supplement 1, Video 4). Despite the reversibility of CRY2 clustering in previous findings, we find that CRY2 initiated clustering of Ab appeared to be irreversible likely due  to the structure of the aggregates (Antzutkin et al., 2012). Consistent with existing literature showing that Ab oligomers induce oxidative stress and inflammation (Butterfield et al., 2013;Forloni and Balducci, 2018), we show that light-induced oligomerization of Ab, but not Ab expression alone, is necessary for the metabolic defects, loss of mitochondria and inflammation in C. elegans and Drosophila. Light-induced oligomerization also leads to physical damage of the nervous system in D. rerio, resembling the loss of brain tissue in AD (Figure 8-figure supplement  1). The embryonic assay for physical damage phenocopies brain lesions only to some degree as the extent of the damage is enhanced by the forces applied to the developing neural tissue. The irreversibility of Ab aggregates once initiated as seen here, also hints at intrinsic properties that may be unique or unusual to this peptide and warrants further investigation. It is also likely to prove a good model for testing anti-aggregation compounds. We anticipate that our approach will serve as an attractive tool for carrying out drug screens and mechanistic studies for the treatment of AD. This optogenetic strategy will also undoubtedly complement other techniques due to the high level of spatiotemporal specificity. For example, it could be optimally positioned to gain insights into the unresolved mechanism of Ab -whether oligomerization and subsequent accumulation or lack of effective clearance underpins AD pathologies. The separation of two phases of AD progression also suggests that single drug treatments will not suffice, and perhaps combinatorial approaches should be tried.

Crosses and expression of UAS construct
For Drosophila, the transgenes were injected into attP2 (Strain#8622) P[CaryP]attP268A4 by Best-Gene Inc (California) (Groth et al., 2004;Markstein et al., 2008). Expression was driven by Elav-GAL4 the neuronal driver and two ubiquitous drivers armadillo-GAL4 and daughterless-GAL4 Video 3. Elav-Gal4/UAS-MyrTomato embryos exposed to light and imaged for neurons in red (tdTom) and glial cells Repo-QF2 >QUAS GFP. Blue laser power at higher setting to allow imaging of glial cells and control for laser damage. https://elifesciences.org/articles/52589#video3 (Brand and Perrimon, 1993). All additional stocks were obtained from the BloomingtonDro-sophilaStock Center (NIH P40OD018537) were used in this study.
Fly crosses performed were: . Elav-Gal4 x w; UAS-Ab-CRY2-mCh . Arm-Gal4; Da-Gal4 x w; UAS-Ab-CRY2-mCh For C. elegans, 50 ng/ml the construct hsp-16-2 p-Ab-CRY2-mCh was co-injected with 25 ng/ml of pharyngeal-specific fluorescence marker myo-2-gfp into the distal gonads of wild type young adults. Transgenes were maintained as extrachromosomal arrays, hence careful selection of transgenic animals with pharyngeal GFP expression is required prior to all experiments. A myo-2-gfp strain C. elegans was also generated and used alongside as controls.

Animal husbandry
Drosophila were maintained at standard humidity and temperature (25˚C) with food containing 6 g Bacto agar, 114 g glucose, 56 g cornmeal, 25 g Brewer's yeast and 20 ml of 10% Nipagin in 1L final volume. Transgenic Drosophila and controls were distributed into either dark or light condition on Day 1. Transgenic flies fed with a lithium-supplemented diet were maintained in the same food as above with the addition of lithium chloride to a final concentration of 5 mM. C. elegans were maintained as previously described at 20˚C (Stiernagle, 2006). Age-synchronized nematodes were generated by hypochlorite bleaching. 250 mM of 5fluoro-2'-deoxyuridine (FUDR) was added to prevent progeny production in all experiments except for fertility assays. For Fertility assay, normal Nematode Growth Media (NGM) agar plates were used. To induce expression of Ab-CRY2-mCh, Day 4 young adult worms were heat shocked at 35˚C for 90 min, and subsequently incubated at 20 0 C for a day (Day 5) for recovery. The heat-shocked transgenic Day 6 C. elegans were then separated into dark or light condition and was maintained at 20 0 C throughout the experimental period. Non-heat shock controls were used for lifespan, fertility and locomotion studies. Adult zebrafish were reared under standard zebrafish facility conditions with a 14 hr light/10 hr dark cycle. Zebrafish embryos injected with different expression vectors were screened for fluorescence and distributed into either dark or light condition 24 hr post fertilization. Embryos were dechorionated at 48 hr post fertilization and exposed to blue light (488 nM) for 1-2 hr at room temperature to induce oligomerization and subsequently used for mitochondria metabolic flux assay or ATP assay. Figure 5. Gene expression of oxidative stress related genes in transgenic Ab C. elegans with and without light-induction. Differential gene expression of catalase (Cat), peroxiredoxin-3 (Prdx3), thiroredoxin-1 (Trx-1), thioredoxin-2 (Trx-2), superoxide dismutase-1 (SOD-1), superoxide dismutase-2 (SOD-2) and superoxide dismutase-3 (SOD-3) in transgenic Ab C. elegans in light condition compared to control dark condition. Light-induced Ab aggregation showed up-regulation in expression of Cat (**p<0.01) and down-regulation in expression of Trx-2 (*p<0.01) and Sod-3 (***p<0.001).

FRAP assay
Adult fly midguts were dissected in 200 ml of 1x PBS in a PYREX Spot Plates concave glass dish (Fish-erScientific). Subsequently, the guts were carefully transferred onto a small droplet of 1x PBS on a 35 mm glass bottom dish. Using fine forceps, the gut was repositioned to resemble its natural orientation. PBS was then removed from the area surrounding the gut, leaving a small amount of excess PBS to hold the gut in place and prevent desiccation. The 3 mm glass bottom dish was then mounted onto the Zeiss LSM800 (Carl Zeiss AG, Germany) for imaging. The super-resolution imaging function on the Zeiss LSM800 was used for the bleaching and fluorescence recovery. A small square within a cell was defined as the bleaching area and this same square was used for all bleaching experiments. Bleaching was performed at 4% laser power using the 488 nm laser for 1 s and the recovery was followed every 30 s for 20 minutes. The fluorescence intensity of the bleaching square (F S ), the entire cell (F C ) and the background (F B ) at each time point was computed by the ZEN 3.1 (Blue edition) software (Carl Zeiss, Germany). For determination of percentage fluorescence recovery, the background fluorescence was subtracted from the fluorescence of the bleaching square (F S -F B ) and the fluorescence of the entire cell (F C -F B ). The (F S -F B )/ (F C -F B ) was computed at each time point and the (F S -F B )/(F C -F B ) before bleaching was set to 100%. This was done for all FRAP assays for both constructs and 3 separate experiments were conducted per construct.

Lifespan assays
Drosophila and C. elegans were counted daily for the number of dead subjects and the number of censored subjects (excluded from the study). Drosophila that failed to respond to taps were scored as dead and those stuck to the food were censored. C. elegans that failed to respond to plate-tapping were scored as dead, those that burrowed to the side of the plate were censored.

Locomotion assay
Drosophila locomotion was assessed using an established geotaxis assay previous described in Rival et al. (2009). In brief, 25 Drosophila were enclosed in a plastic column (25 cm tall, with a 1.5 cm internal diameter) and were tapped to the bottom. The number of Drosophila at the top (Ntop) of the column, and that at the bottom (Nbot) were counted after 20 s. Three trials with the same sample were performed within 30 s interval. Performance index was defined as (15+Ntop-Nbot)/30.
Heat-shocked C. elegans were assessed at Day 9, by placing worms onto individual NGM plates, pre-spotted with Escherichia coli OP50. C. elegans were placed on the side of the bacteria spot on the NGM plate using a platinum worm picker and were left undisturbed to move freely for 15 min. Distance travelled was determined by imaging and measuring of worm tracks on bacteria using a stereomicroscope.
Fertility assay 25 pairs of male and female Drosophila expressing MD-Ab-Cry2-mCh were kept in a single tube under dark or light conditions. The number of eggs laid were scored after 5 hr 30 min for Days 5, 6, 7, 13, 14 and 20. Day 5 heat-shocked C. elegans were transferred onto individual NGM plates spotted with E. coli. Each worm was transferred to a fresh plate daily from Day 6 to Day 8, allowing 24 hr' egg laying period. The number of new young worms were counted on each plate as the progeny of a single individual.
Adenosine triphosphate (ATP) assay ATP assay using firefly lantern extract (Sigma-Aldrich) was performed as described by Tsujimoto et al. (1970). Each C. elegans sample containing 100 D2 adult worms was freeze-thawed in liquid nitrogen and sonicated in 10% Trichloroacetic acid (TCA) buffer while each zebrafish sample containing five 48hpf embryos was lysed in 10% TCA buffer. The extracted ATP from the different conditions and ATP standards were loaded onto a 96-well plate and injected with firefly lantern extract. Luminescence was measured using a Cytation 3 Imaging Reader (Biotek).

Mitochondrial metabolic flux assay
Mitochondrial metabolic flux assay for C. elegans was performed as described by Fong et al. (2017) using XF96 Extracellular Flux Analyzer. 60 D2 adult worms from each condition 24 hr post light-treatment were transferred to a 96-well Seahorse plate containing M9 buffer. Final concentrations of the drugs used in the OCR experiment were 10 mM FCCP and 50 mM sodium azide. For zebrafish embryos, mitochondrial metabolic flux assay was performed as described by Stackley et al. (2011) with some modifications. Oxygen consumption rate (OCR) measurements were performed using the XF96 Extracellular Flux Analyzer. Each well contained one embryo (48hpf) in 175 mL of E3 medium (fish system water). Final concentrations of the drugs used in the OCR experiment were 9.4 mM oligomycin, 2.5 mM FCCP and 20 mM sodium azide.

Light-sheet microscopy
Drosophila embryos were dechorionated using bleach, rinsed twice with water and dried, and loaded into a capillary filled with 1% low-melting agarose Type VII-A in water (Sigma) (Colosimo and Tolwinski, 2006). C. elegans and D. rerio were embedded directly in the agarose. Samples were imaged with a Lightsheet Z.1 microscope. Cry2 was activated by a dual-side illumination with 10% power 488 nm laser for 29.95 ms for every 2.5 min for 500 cycles. Controls were only exposed to 561 nm. Images were acquired with a water immersion objective at 10x/0.2 Illumination Optics and W Plan-Apochromat 20x.1.0 UV-VIS detection objective (Carl Zeiss, Germany). Image data were processed using the maximum intensity projection function of ZEN 2014 SP software (Carl Zeiss, Germany), and were analyzed with ImageJ (NIH) and IMARIS 9.0 (Bitplane AG, UK).

Transfection and live cell confocal microscopy
Human Embryonic Kidney (HEK293T) cells were obtained from ATCC through the local distributor Bio-REV Pte Ltd after Short Tandem Repeat profiling, confirmed mycoplasma-free and maintained in Dulbecco's Modified Eagle's Medium with 10% Fetal Bovine Serum and 1% Penicillin and Streptomycin (Invitrogen). Cells were plated at 80% confluence on 35 mm TC treated glass bottom dish 24 hr prior to transfection with 2.5 mg of pDEST40-Ab-CRY2-mCh using the Effectene Transfection Reagent (Qiagen). A transfection master mix containing the plasmid and transfection reagent was prepared for all treatment plates during each biological replicate to ensure the transfection efficiency was the same across all plates. Two hours after transfection, the control plate was topped up with 250 ml of complete media without antibiotics. For cells treated with drugs, LiCl or CHIR99021 were also added at this point to achieve a final concentration of 2 mM LiCl or 3 mM CHIR99021 respectively. Cells transfected with only the pDEST40-Ab-CRY2-mCh plasmid were used as negative control. Cells were imaged 24 hr after transfection using the Zeiss LSM800 confocal microscope with 63x oil immersion lens.

Immunofluorescence and confocal microscopy
Drosophila embryos 9 hr after deposition were incubated at 25˚C in either light or dark conditions for 13 hr before being dechorionated using bleach. Embryos were then fixed with Heat-Methanol treatment (Müller and Wieschaus, 1996) or with heptane/4% formaldehyde in phosphate buffer (0.1M NaPO4 pH 7.4) (Tolwinski and Wieschaus, 2001). Staining, detection and image processing as described in Colosimo and Tolwinski (2006).
Primary antibodies used were the glial cell marker anti-Repo (mouse mAb, 8D12) and the neuronal cell marker anti-Elav (rat mAb, 7EA810) from Developmental Studies Hybridoma Bank (DSHB developed under the auspices of the NICHD and maintained by The University of Iowa, Department of Biological Sciences, Iowa City, IA 52242). Secondary antibodies used were Alexa Flour 488 antirat and Alexa Flour 647 anti-mouse (Invitrogen).
For detection of Ab aggregates, thioflavin T (ThT) staining were performed as previously described Iijima et al. (2008). Formaldehyde-fixed embryos were incubated in 50% EtOH containing 0.1% ThT (Sigma) overnight. Embryos were destained in 50% EtOH for 10 min, followed by three washes in PBS. Embryos were then mounted on microscope slides using Aquapolymount (Polysciences, Inc).
Images were acquired on the Zeiss LSM 800 (Carl Zeiss, Germany) using the following settings: 1% laser power for 488 nm; 5% laser power for 561 nm; 2% laser power for 647 nm. Images were processed using the ZEN 2014 SP1 software (Carl Zeiss, Germany) and Imaris (Bitplane AG).

RNA extraction, cDNA synthesis and qPCR
Drosophila embryos 9 hr after deposition were incubated at 25˚C in either light or dark conditions for 13 hr were used. Embryos were dechorionated and washed with 100% ethanol prior to RNA extraction using the ISOLATE II RNA Mini Kit's protocol (Bioline, UK). The extracted RNA was quantified using Nanodrop (Thermo Fisher Scientific). cDNA synthesis was done according to the Sensi-FAST cDNA Synthesis Kit's protocol (Bioline). Primers pairs used: AttA, IMD, DptA, Def, Duox, Rel, catalase, Prdx3, Trx1, Trx2, SOD-1, SOD-2, SOD-3 and reference gene Rpl32. Quantitative PCR was performed using SYBR Green. Expression data were normalized to the dark controls. For mitochondrial copy number, the above collection method was used except the following: the embryos were stored in tritonX-100 and the primers used were Mitochondria Cytochrome b (MtCyb) and reference gene RNAse P.

Image data and statistical analysis
Ab aggregates were quantified in Drosophila using ImageJ (NIH) and in C. elegans and HEK cells with IMARIS. For HEK cells, the number of clusters larger than 0.4 mm in diameter were determined for at least 3 cells for the control and drug treated cells to calculate the average number of clusters per cell. The mean intensity per cluster across all these cells was also determined to calculate the average intensity of the clusters for the control and drug treated cells. For statistical analysis of the expression of genes, student's t-test was used for except for cases where the data showed unequal standard deviation (F-test, p<0.05), in which the Mann-Whitney nonparametric test was performed. Statistical analysis of lifespan studies and behavioral assays were performed using OASIS 2 (Han et al., 2016). The number of samples was determined empirically. All graphs were plotted using Graphpad PRISM 6 (Graphpad Software).

Key resources table
New model organisms generated for this paper are available to the community by contacting the authors. Stable lines and stocks will be made available through stock centers.