NUCLEAR RETENTION OF LEPTIN/UPD2 REGULATES ORGANISMAL RESILIENCE TO NUTRIENT EXTREMES

Adipokines released from the adipocytes function as a systemic adipometer; they impinge on neural circuits to signal nutrient status. On starvation, adipokines must be retained to signal energy deficit; else, it significantly reduces starvation survival. But how fat cells retain adipokines is unclear. Here, we demonstrate that Atg8, a cell-intrinsic autophagy factor, regulates the starvation-induced acute retention of the Leptin Drosophila ortholog Upd2. We show that on starvation, as a direct consequence of Atg8’s lipidation, Upd2 accumulates in the nucleus. We illustrate that Upd2’s nuclear retention is critical to fat mobilization and increased starvation resilience. Furthermore, nuclear Upd2 promotes the expression of a secreted innate immune gene signature. This hints at an unanticipated connection between adipokine nuclear retention and increased innate immunity. In conclusion, we propose that, during starvation, Atg8’s role is not just limited to autophagy but is critical for withholding adipokines in the nucleus to promote starvation resilience. GRAPHICAL ABSTRACT In fed state Upd2 requires Atg8 for nuclear exit and cytosolic localization. Atg8’s lipidation on starvation results in Upd2’s nuclear accumulation. Upd2 nuclear retention on starvation increases fat mobilization and post-starvation hunger. On starvation Upd2 nuclear retention increases expression of a secreted innate immune signature.

Nonetheless, we noted that Upd2's nuclear accumulation inversely correlated with its secretory potential. Altogether, we concluded that acute starvation reduces Upd2's extracellular release and increases Upd2 nuclear accumulation in diverse contexts of 24 Drosophila S2R+ cells and adult fly fat.
UPD2 REQUIRES ATG8 FOR NUCLEAR EXIT: Intriguingly, although Upd2 required CRM1/Emb-based 27 nuclear export ( Figure S1D), we were unable to identify a canonical nuclear export signal (NES) on Upd2, suggesting that Upd2's nuclear export occurs indirectly via other protein adapters (See Discussion). To identify clues to what Upd2's protein partners are likely to be, 30 6 we examined protein motifs in Upd2's protein sequence. We noted that Upd2 had several Atg8-interaction motifs (AIM) [50]. Given Atg8's well-characterized role in nutrient-sensing, we wondered whether Atg8 might interact with Upd2 to control its nutrient-dependent 3 nucleocytoplasmic localization.
We generated transgenic flies with two-point mutations to Upd2's putative AIM-like sequence ( Figure 2A). Multiple AIM-like sequences are found in Upd2, but only one is the 6 canonical 'WXXL' AIM [50]; hence we mutated that. We specifically expressed Upd2-AIM-(WXXLAXXA) in fly fat tissue in an upd2-deletion background (upd2Δ; Lpp-Gal4> UAS-Upd2-AIM-::mCherry ) and examined Upd2's localization in the abdominal fat pads of 7 day-9 old adult male flies, fed ad libitum in relation to control transgene (upd2Δ; Lpp-Gal4> UAS-Upd2-WT::mCherry). Upd2-AIM-expressing flies displayed a higher nuclear accumulation (2Ab) of Upd2 than controls (2Aa). It is to be noted that the Upd2-WT and Upd2-  transgenes are expressed at similar levels ( Figure S7B). Furthermore, corroborating the role of Atg8 in regulating Upd2's localization, fat tissue-specific, acute temporal knockdown of Atg8 (upd2GFP-knockin; ppl-Gal4, tubGal80 ts > Atg8-RNAi) resulted in a significant increase 15 in the endogenous Upd2's mean fly fat nuclear intensity ( Figure 2B). Thus, disruption of Upd2's putative AIM or knockdown of Atg8 increases Upd2's nuclear accumulation and phenocopies acute starvation. 18 Next, we wanted to test whether Upd2's intracellular localization in Drosophila S2R+ cells, like what we observe for fly fat, is regulated by Atg8. In Drosophila S2R+ cells, we observed that Upd2-AIM-expression was significantly nuclear (Figure 2Cb Atg8 are likely to be complex and that Upd2's AIM sequence is required for this interaction. 27 Furthermore, corroborating Atg8's requirement for Upd2's cytosolic localization, Atg8 knockdown in Drosophila S2R+ cells, using two independent dsRNAs, resulted in a 7 significant increase in Upd2-WT::GFP nuclear signal ( Figure S2B). These observations lead us the hypothesize that Atg8 regulates Upd2's cytosolic localization during a fed state.
When we performed quantitative ELISA assays for Upd2 secretion during an Atg8-6 knockdown (KD) or in Upd2-AIM-state, we observed a significant impairment of Upd2 secretion ( Figure S2C and S2D). A previous study identified that Upd2 undergoes GRASPmediated unconventional secretion [27]. In keeping with a role for Atg8 in mediating 9 unconventional GRASP-mediated secretion [40, 51-53], we found using Co-IP experiments in S2R+ cells that Upd2-AIM-ability to interact with GRASP is impaired. In sum, these experiments suggested that Upd2, via its AIM sequence, interacts with Atg8, and this 12 interaction regulates both its cytosolic localization and subsequent extracellular release in fed cells.
Finally, we sought to test whether reconstituting Upd2-AIM-interaction with Atg8 15 would be sufficient to localize Upd2-AIM-to the cytosol. To test this, we took advantage of the genetically encoded protein binder tag VHH that acts as a nanobody recognizing GFP (vhhGFP4) [54]. Functional studies performed by fusing VHH to different proteins and then 18 asking how it affects the localization of a GFP-tagged protein have been used to dissect molecular players controlling protein localization [55][56][57]. Akin to the experimental design in the studies mentioned above, which utilized nanobody fusions to probe protein 21 localization, we tested whether the interaction between Upd2-AIM-::GFP and VHH tagged was reconstituted Atg8 is sufficient to restore the cytosolic localization of Upd2-AIM-in fed cells (See Schematic in Figure 2D). 24 To this end, in Drosophila S2R+ cells, we co-expressed with Upd2-AIM-::GFP either the control VHH::Myc (control) or Atg8-WT was fused to VHH::Myc (VHH::Myc-Atg8-WT). As would be expected of a GFP-nanobody, in that it would bind to GFP-tagged proteins, we 27 observed that VHH::Myc showed increased nuclear accumulation, overlapping with Upd2-AIM-::GFP (See arrow; Figure 2Da). However, co-expression of the control nanobody did not result in localization change of Upd2-AIM-::GFP. Strikingly, in the presence of VHH-tagged 30 8 Atg8-WT (VHH::Myc-Atg8-WT), we observed that Upd2-AIM-GFP is now cytosolic and colocalized with Atg8-WT in punctate structures (See Arrow; Figure 2Db), and reflecting this, Upd2-AIM-nuclear accumulation was significantly reduced (See quantification in Figure   3 3B'). Hence, reconstituting the physical interaction between Atg8 and Upd2-AIM-is sufficient to localize Upd2-AIM-to the cytosol. This pivotal result demonstrated that Upd2 requires Atg8 for its nuclear exit and cytosolic localization. We posited that Atg8's lipidation is likely to reduce its nuclear availability; 18 consequently, Upd2's nuclear exit is impeded, resulting in Upd2's nuclear accumulation.
To test this hypothesis, we generated lipidation-defective Atg8 by mutating to the Cterminal lipid acceptor moiety glycine of Atg8 (Atg8-GA) [35]. We performed tests on Atg8 21 GA to validate that lipidation was defective by using accepted assays in the field to test Atg8 lipidation [62]. This included: i) doublet on western blots of starvation lysates ( Figure   S4A); ii) recruitment to cytosolic vesicular structures on starvation ( Figure S4B). From this, 24 we inferred that Atg8 GA mutation rendered its lipidation defective (hereafter, it is referred to as Atg8-PE-for simplicity). We fused the lipidation defective Atg8-PE-to an N-terminal VHH tag ( VHH::Myc-Atg8-PE-; see methods). In subsequent experiments, we utilized 27 VHH::Myc-Atg8-PE-to test the role of Atg8 lipidation in Upd2's starvation-induced nuclear accumulation. 9 We hypothesized that lipidation of Atg8 reduces its capacity to participate in Upd2 nuclear exit. Hence, we predicted that reconstituting Upd2-WT::GFP's interaction with lipidation-defective Atg8 should 'rescue' its nuclear accumulation defect ( Figure  Atg8-WT is localized to cytosolic structures on starvation ( Figure S4Bb), but Atg8-PE-does not ( Figure S4Bd). Collectively, these experiments ( Figure 3A) suggest that, on starvation, Atg8's lipidation reduces its nuclear availability, in turn resulting in increased Upd2 nuclear 12 accumulation.
This observation also suggested that Atg8's lipidation is not required for Upd2's nuclear exit in fed cells. Accordingly, VHH nanobody-based reconstitution of Upd2-AIM- . Therefore, we predicted that consistent with mammalian data, our results in Drosophila S22R+ cells ( Figure S4B), Atg8's nuclear amount in vivo in adult fly 24 fat cells will be depleted on starvation. Flies expressing GFP tagged Atg8 in fly fat (Lpp-Gal4> UAS-GFP::Atg8-WT) were subjected to an acute starvation diet (4-6 hours 0% sucrose agar) or ad libitum fed 30% high sugar diet (HSD) for 14 days. We assessed the total nuclear 27 signal of GFP::Atg8-WT in 3D volume between the flies fed ad libitum normal lab food versus starvation and HSD ( Figure 4A; Also see Study Design section). As predicted, acute starvation significantly reduced the total nuclear signal of GFP::Atg8-WT, in conjunction 30 with an increase in cytosolic punctate GFP::Atg8-WT (Figure 4Ac, 4A'). Conversely, a surplus diet (30% high-sugar) significantly increased the total nuclear signal of GFP::Atg8-WT ( Figure 4Ab, 4A'). In sum, in Drosophila adult fly fat Atg8, nuclear signal increases on a 3 surplus diet, whereas Atg8 is less nuclear on starvation.
Next, we wondered what happens to nuclear levels of lipidation defective Atg8 in adult fly fat cells on starvation. Whereas the GFP::Atg8-WT (Figure 4Ba) was observed in 6 cytosolic puncta on starvation, Atg8-PE-in the adult fly fat continued to be nuclear on starvation ( Figure 4Bb). Specifically, we noted that the GFP::Atg8-PE-nuclear signal was the same between fed and starved states ( Figure 4B'; p=0.98). Hence, when Atg8 is lipidation 9 defective, Atg8 is nuclear on both fed and starved states.
Atg8-PE-expression prevented Upd2 nuclear accumulation ( Figure 4C), and we wondered whether this impacted organismal starvation resilience. For conducting whole 24 animal physiology experiments, we could not recover any viable progeny when we expressed Atg8-PE-in an Atg8 mutant background ( Figure S5A). Since the expression of the Atg8-WT and Atg8-PE-transgenes were comparable on starvation ( Figure S5B Next, we went one step further and predicted that the inability of flies with fatspecific Atg8-PE-over-expression to mobilize TAG on starvation is at least in part due to 6 Upd2's cytosolic localization. If this prediction were to hold up, we should observe that the negative effects of Atg8-PE-over-expression are mitigated in the absence of Upd2. Consequently, we performed the TAG mobilization and starvation survival experiments in Atg8-WT transgene expression. In sum, these results support a model that Atg8's lipidation, in addition to its critical role in autophagy, is required for Upd2 retention; this, in turn, determines how an organism survives prolonged starvation. nutrient state and adipokine nuclear accumulation is genetically perturbed. We subjected flies expressing control transgene, Upd2-WT or Upd2-AIM-specifically in fly fat cells, to a high sugar diet (HSD) regime (See Methods and Study Design sections). We found that over-21 expression of Upd2-AIM-renders flies sensitive to an HSD regime ( Figure 5A), suggesting holding Upd2 in the nucleus during a surplus diet where Upd2 is less nuclear ( Figure S6A) reduces HSD survival. Conversely, we predicted that genetically 'holding' Upd2 in the 24 nucleus during starvation, a state where endogenous Upd2's nuclear accumulation increases ( Figure 1A), will allow flies to survive starvation. Consistent with our prediction, the over-expression of Upd2-AIM-in fly fat cells (upd2Δ; Lpp-Gal4> UAS-Upd2-AIM-) on 27 starvation provided resilience; Upd2-AIM-flies survived significantly longer than control or Upd2-WT overexpression ( Figure 5B). Overall, this suggested that the inverse correlation between adipokine nuclear accumulation and nutrient state is important. 30 We wondered whether retention of Upd2 within the cell, but not specifically in the nucleus, would be sufficient to have the same systemic effects on starvation resilience ( Figure 5B). To tease apart this nuclear versus cytosolic retention, we examined a state in 3 which Upd2 was withheld in the cytosol but not released. To achieve this, we ectopically localized Upd2-AIM-to the cytosol by appending a constitutive nuclear export signal (NES) to Upd2-AIM-(See Methods). Expressing cytosolic Upd2-AIM-in the fly fat (upd2Δ; Lpp-6 Gal4>Upd2-NES+-AIM-), as predicted, was sufficient to re-localize Upd2-AIM-to the cytosol ( Figure S6B; See arrows), whereas Upd2-AIM-expressing flies displayed no cytosolic localization ( Figure S6B & Figure 2). Consistent with the role of Atg8 in mediating 9 Upd2's secretion, we found that, despite being localized to the cytosol ( Figure S6D), Upd2-NES+-AIM-was not secreted ( Figure S6E). Significantly, however, we noted that fat-specific expression of Upd2-NES+-AIM-(upd2Δ; Lpp-Gal4>Upd2-NES+-AIM-) did much worse on 12 starvation than control flies ( Figure S6C). These results suggested that it is not just retention within the cell, but specifically Upd2's retention in the nucleus critical for exerting the systemic effects on starvation resilience. 15 We had previously shown that overexpression of Upd2-WT transgene in fly fat renders it incapable of mobilizing TAG reserves on starvation ( Figure 2G in Rajan and Perrimon, 2012). Hence, we wanted to assay TAG mobilization in flies over-expressing  Next, we assayed whether Upd2-AIM-displayed different feeding behavior poststarvation than Upd2-WT over-expression flies. Expressing Upd2-WT transgene, specifically in fly fat (upd2Δ; Lpp-Gal4> UAS-Upd2-WT), displayed no significant increase in hunger-27 driven feeding, as measured by the FLIC assay ( Figure 5D, also see methods); this is consistent with Upd2's role as a satiety signal. Conversely, Upd2-AIM-fat-specific overexpression flies (upd2Δ; Lpp-Gal4> UAS-Upd2-AIM-) displayed a significant increase in 30 13 hunger-driven feeding ( Figure 5D; Upd2-AIM-Fed vs. Starved; p=0.002). These findings suggest that nuclear accumulation of Upd2 on starvation exerts systemic effects; it promotes starvation resilience, fat-mobilization, and hunger-driven feeding. resilience. We hypothesized that Upd2 nuclear retention, during starvation, has some function within the nucleus. We wondered whether one such function might be for nuclear  Figure 6B). Significantly, these two genes -IBIN and SPH93 -were highly upregulated in Upd2-AIM-over-expressing flies ( Figure 6B). Overall, this suggested 24 that nuclear Upd2 localization on starvation is required to express certain immune genes that encode secreted products (See Discussion).
We then performed deeper analyses for all the upregulated genes in Upd2-AIM-fat-27 specific expressing flies on starvation (upd2Δ; Lpp-Gal4> UAS-Upd2-AIM-; Figure 6C). We noted that in addition to one gene (mAchR-C) implicated in promoting feeding in mammals [63], most of the genes that were upregulated included secreted antimicrobial peptides 30 14 ( Figure 6C). When we analyzed the GO terms for genes upregulated in Upd2-AIM-during starvation, the key GO terms ( Figure 6D) were all related to defense response to the bacterium. Hence, the over-expression of Upd2-AIM-during starvation upregulates a 3 secreted immune peptide gene signature that provides a potential explanation for how it exerts systemic effects (See Discussion). Notably, overexpression of Upd2-WT transgene, which differs from AIM only in two amino acid sequences, and is expressed at relative levels 6 to Upd2-AIM-transgene in baseline ( Figure S7Ba), and starved states ( Figure S7Bb), does not result in upregulation of this immune GO signature ( Figure S7Ab). Hence, this supports the idea that increased nuclear retention of Upd2 on starvation promotes starvation 9 resilience by upregulating a secreted innate immune peptide signature. In future work, directed experiments will be required to ask whether the upregulation of secreted immune genes contributes to the increased TAG mobilization, feeding behavior, and starvation 12 resilience observed in Upd2-AIM-(See Discussion).
In conclusion, our results suggest that we have revealed a specific intersection node between cell-intrinsic and extrinsic mechanisms in regulating an organism's response to 15 nutrient extremes. We demonstrate that Atg8 regulates nutrient state-dependent localization of the Drosophila Leptin, Upd2 in both fed and starved cells (Summary of findings in Figure 7). Upd2's nuclear accumulation on starvation, as a direct result of Atg8's 18 lipidation, is important for organismal adaptation to nutrient deprivation. Thus, we propose that Atg8's role in the starvation response is not just limited to autophagy but also is required for adipokine retention. This process is critical for an organism to sense and adapt to only permit proteins <40kDA to pass through [64], Upd2 is likely to require active nucleocytoplasmic (NC) transport mechanisms that enable transit through the NPC [65].

27
Consistent with the role for classic nuclear export being required, we find that Embargoed/CRM1 is required for Upd2's nuclear exit ( Figure S1C). But, using the NETNES mapper tool [66] did not identify a canonical NES in Upd2. Hence, we hypothesized that other adapter proteins interacting with Upd2 might promote its nuclear export.
Unexpectedly, we found that mutations to Upd2's AIM or KD of Atg8 phenocopies 3 knockdown of nuclear export protein Emb (Figure 2, S2). This suggested to us that Atg8 is required for Upd2's cytosolic localization in fed cells. Consistent with this nanobody-based reconstitution of the physical interaction between Upd2-AIM-, that is only localized to the 6 nucleus, and Atg8 is sufficient to re-localize Upd2 to the cytosol ( Figure 2D, 3B, 3B').
Our study does not yet resolve how Upd2 localizes to the nucleus in the first place. But, based on the observations that Upd2 is nuclear when Atg8 is knockdown and when its AIM 9 is disrupted, we conclude that Atg8 does not regulate Upd2's nuclear entry. Analysis of Upd2's sequence using a bioinformatic tool-classic NLS (cNLS) mapper [67][68][69], predicted that Upd2 contains an NLS at residues 141-150 (GRVIKRKHLE) with a score of 10.5.

12
According to the cNLS mapper, a GFP reporter protein fused NLS, with a score of 10, is exclusively localized to the nucleus. Hence, we reasoned that Upd2 is likely to be nuclear, but when Upd2 is ' licensed' by cofactors like Atg8, it will Upd2 transit to the cytosol. 15 We considered the alternate hypothesis that Atg8 cytosolically tethers Upd2 to prevent nuclear entry. But if that were the case, we would expect that Atg8, which is highly cytosolic on starvation, would tether Upd2 to the cytosol. Given that Atg8 nuclear depletion, within 4 18 hours of acute starvation ( Figure 4A), coincides with increased Upd2 nuclear accumulation ( Figure 1A), we favor the hypothesis that the nuclear pool of Atg8 in fed cells facilitate Upd2 nuclear exit.

21
Further work will be needed to resolve all the events leading to Upd2's Atg8-mediated nuclear exit. We postulated that Atg8, a ubiquitin-like protein (Shpilka et al., 2011), might operate akin to other ubiquitin and ubiquitin-like proteins that have been shown to function 24 in cellular localization [70][71][72]. Specifically, SUMO (small ubiquitin-like modifier) positions its target protein (RanGAP1) to the NPC to increase the efficiency of its translocation [73].
Similarly, Atg8 might position adipokine Upd2 to enable efficient export via the classical 27 nuclear export pathway or may serve to direct Upd2 to other nuclear export mechanisms such as nuclear budding [74].
How widespread this role for Atg8 in nucleocytoplasmic protein localization remains resolved. But it is quite plausible that other proteins, such as ribophagy receptor NIUFIP1 [75] with 4 AIM sequences, and localizes to cytosol on starvation, might utilize Atg8 for 3 context-dependent NC shuttling. Furthermore, insulin retention on acute starvation is a critical response in flies and mammals [76]. It will be important to explore whether Atg8 mediated retention, like Leptin/Upd2 in fat, also applies insulin in pancreatic beta cells. In In this work, we reveal that the adipokine Upd2 needs to be retained within the nucleus for exerting its positive effect on starvation resilience ( Figure S6B, C). Our bulk transcriptomic analysis hints at some possibilities for why Upd2-AIM-might be able to exert 24 this systemic influence. First, we find very few genes are upregulated in overnight 16-hour starvation in our control background. Of them, two innate immune genes are dependent on upd2 for expression since they are not upregulated in a upd2Δ background. Significantly 27 SPH93 is secreted in the extracellular space and required for gram-positive bacteria response [77][78][79]. IBIN, which encodes a long non-coding RNA (lncRNA) induced upon infection, also required Upd2 to express starved states ( Figure 6B). IBIN enhances the 30 expression of genes required for glucose retrieval [80]. Hence IBIN's upregulation downstream of Upd2 could play in starvation resilience by increasing sugar utilization.
Furthermore, in a background where Upd2 is highly nuclear (Upd2-AIM-), the GO signature 3 of upregulated genes is that of secreted innate immune peptides ( Figure 6D long-term memory [85]. Hence, we postulate that increased AMP expression in Upd2-AIMmay impinge on systemic processes, perhaps by altering brain function, and this might underlie their increased hunger feeding motivation ( Figure 5D). Furthermore, the increased 12 lipolysis we observe with Upd2-AIM over-expression flies ( Figure 5C) aligns with the emerging role in mammalian studies for AMPs in TAG mobilization [86,87]. In sum, our current working hypothesis is that the secreted immune peptide signature downstream of 15 Upd2 nuclear retention enables the organism to adapt to starvation and promotes hungerdriven feeding. Though it is beyond the scope of the current study, testing these specific hypotheses will be refined and developed in future work. fine-tunes gene expression by binding to other DNA and chromatin binding factors will need to be explored in future work. The latter hypothesis is supported by preliminary findings of 27 our lab's proteomic datasets. In multiple independent proteomic surveys of Upd2 immunocomplexes, we identified ten bonafide DNA binding and chromatin binding factors that complexed with Upd2 (Kelly and Rajan, unpublished results); whether these are bona fide interactors and if they have a functional significance will be tested in future work.
An intriguing observation is that Upd2-AIM-upregulates the innate immune gene 3 signature only on starvation. Why does Upd2-AIM-not upregulate starvation responsive genes even in the fed state? One possibility is that factors required for nuclear Upd2 to engage in gene expression become available only on starvation. Another possibility is that 6 nuclear factors in the fed state might prevent Upd2-AIM-from accessing its gene expression partners. Could Atg8 itself be such a factor that prevents Upd2 from engaging in gene regulation during the fed state? Atg8 signal in the nucleus is directly correlated with the 9 nutrient state. It is upregulated on an HSD regime and significantly reduced in fat nuclei within 4 hours of starvation ( Figure 4A). Hence, we speculate that Atg8, though it cannot bind Upd2-AIM, could sequester factors required for Upd2-AIM's access to gene regulatory 12 function in fed cells. This hypothesis is consistent with recent reports that Atg8 can regulate gene expression in Drosophila fat by working with transcription factors [92]. Even if we discount the Upd2-AIM-effects on gene expression as a gain-of-function effect, we note 15 that Upd2 is required to upregulate gene expression at endogenous levels in starved flies ( Figure 6B). Upd2 is likely to be nuclear localized in the endogenous expression state because of a strong NLS. Therefore, an active mechanism to prevent improper gene 18 regulation by Upd2 in fed cells is likely to be in place. Hence, Atg8 itself, in addition to promoting Upd2's nuclear exit ( Figure 2), might prevent improper gene regulation by Upd2.
In sum, our study has generated many testable hypotheses regarding Upd2's potential role 21 in gene regulation. We hope that this will stimulate future work. This study uncovered a possible link between adipokine nuclear retention on starvation and the upregulation of antimicrobial peptides (AMPs). A recent study showed 6 that in mice, gut cells upregulate AMP production. It was suggested that increased AMP production in non-feeding states is an anticipatory mechanism to prevent infection.
Significantly, Hooper and colleagues' study found that the increase in gut AMP production 9 is controlled by the transcription factor STAT3 [94]. Strikingly, leptin also activates the STAT3 transcription cascade [95]. Hence, we suggest that, like Upd2, Leptin retention in adipocytes could increase immunity by promoting AMP expression downstream of STAT3. 12 In conclusion, the study on how Drosophila Leptin-Upd2 -is acutely retained in fly fat cells on starvation has led to numerous unanticipated insights raises many questions that should stimulate work both in invertebrate and mammalian systems.

STUDY DESIGN-LIMITATIONS AND CONSIDERATIONS:
Atg8: In Drosophila, clonal analysis of larval fat tissues is the standard for studying how 18 autophagy (Atg) core pathway genes affect cell-intrinsic processes [96]. While clonal analysis addresses is an elegant genetic tool to study cell-intrinsic functions of Atg genes, it is not compatible with experimental design to study whole animal physiology. Therefore, 21 for the physiology assays, in which we assess how Atg8's lipidation in fat affects TAG mobilization and starvation resilience ( Figure 4D, 4E), we have compared the effect of Atg8-WT transgenic over-expression with the Atg8-PE-in fly fat. The reason for this design is two-24 fold. The first is to overcome the hurdle of studying a specific function of Atg8, its lipidation, which is critical to animal survival. We note that lipidation-defective Atg8 flies do not survive ( Figure S5A). Hence the only viable strategy to study the effect of defective Atg8 lipidation 27 is to utilize fat tissue-specific expression of Atg8-PE-and compare its effects to Atg8-WT. It is important to note that we compare the effects of two transgenes -Atg8-PE-and Atg8-WT-20 head-to-head. These transgenes differ in a single amino acid (the penultimate glycine of Atg8) and are expressed at comparable levels even on starvation ( Figure S5B). Furthermore, these results will be viewed in conjunction with the VHH-nanobody based 3 reconstitution of interactions between Atg8-PE-and Upd2, where Atg8-PE-increases Upd'2 cytosolic exit, even on starvation, despite the presence of endogenous Atg8 ( Figure 3A).
Secondly, our goal is to uncover the systemic effects of only altering Atg8's lipidation status 6 in the fat. Since Atg8 has roles in every tissue, the most feasible design is to use the tissuespecific expression of lipidation-defective Atg8 and assess how it impacts whole organism physiology in relation to Atg8-WT. Hence, within the limitations of the system, we have Upd2 exerts systemic effects by signaling from other organs and, specifically, coupling gut 24 metabolic state to the olfactory system [97] or acting in hemocytes to regulate immunity [98]. We have utilized fat-tissue-specific drivers to distinguish Upd2's adipokine-specific function from its other organ-specific roles. In lieu of the ability to perform longitudinal diet- Given that these transgenes are all inserted in the same genome site (attP40) and have 3 similar expression levels ( Figure S7B), it allows for interpretation of the specific mutation to Upd2's Atg8 binding domain means. One valid reservation is that we reveal are gain-offunction effects of Upd2-AIM-. Even if that were to be the case, it is striking that we don't 6 see these gain-of-function effects on starvation resilience when we retain Upd2-AIM-in the cytosol (Upd2-NES+-AIM-) ( Figure S6A, S6B); this suggests that retention of Upd2 in the nucleus on starvation is biologically meaningful. Hence, within the limitations of the 9 system, we have utilized internally controlled experimental approaches to assess the effect of Upd2's nuclear localization on whole animal physiology.

Lead Contact
Requests for further information, reagents, and resources should be directed to and fulfilled by the Lead Contact, Akhila Rajan (akhila@fredhutch.org).

Materials Availability
Drosophila strains generated in this study are available from the corresponding author, Akhila Rajan (akhila@fredhutch.org).

Data and Code Availability
The datasets generated in this study are available from the corresponding author,  The following UAS lines were generated for this study: UAS-Upd2::mCherry, UAS-Upd2-AIM-::mCherry, UAS-GFP::Atg8; UAS-GFP::Atg8-PE-. All transgenes of the same gene were inserted at the same att site to control expression levels-either attp40 or attP2. For RNAi 3 experiments and temporal over-expression in Figure 2B and Figure 4C, flies were generated with the following genotype: Upd2-crGFP; Ppl-Gal4, TubGal80ts was crossed to Atg8-RNAi, Luc-RNAi for 2B; For 4B and 4C and UAS-GFP::Atg8-WT and UAS-GFP::Atg8-PE-. The 6 parents of this cross (F0 generation) were maintained at 18c for both the strains, F1s were allowed to eclose at 18c. A week after eclosion, the F1s were moved to 29c for 5-7 days before dissection and staining.

9
The ability of UAS-GFP::Atg8 and UAS-GFP::Atg8-PE-to rescue the loss of Atg8 function was tested by introducing Atg8 variant transgenes into the Atg8-Trojan-Gal4 line [100] in which insertion of the Trojan-Gal4 into the Atg8 locus generates a lethal allele. We sequence of amplicons used in this study can be found in the table below. LacZ dsRNA was used as controls. All dsRNA knockdown experiments were carried out using two independent dsRNAs per gene. We found that this produced a knockdown efficiency of >85% (based on qPCR analysis) in S2R+ cells. S2R+ transfected with dsRNAs were incubated for four days to allow for gene knockdown. On the 4th day, the media was changed, and the ELISA assay was carried out on the 5th day. Note that the data is represented as percent change in Upd2/Leptin secretion normalized to transfection efficiency with 0% change indicating baseline secretion level. See the ELISA assay procedure below.

26
Primer sequences used for dsRNA production: Renilla Luciferase assay: On day 2 of the ELISA assay, after the conditioned medium was transferred for use in ELISA assays, cells were re-suspended in 50μl of PBS and incubated with 50μl/well Renilla-Glo® Luciferase reagent (Cat# E2710, Promega) for 10 minutes and read using a multiwell luminometer.
Quantification for ELISA assays: For ELISA assays, the ELISA signal readings are normalized to transfection efficiency; the data is represented as percent fold change from control used as the baseline.
Specifically, the ratio of TMB readings to Renilla Luciferase readings is calculated. This ratio from the control is used as a baseline, and the data is represented as the percent fold change of experimental conditions with respect to the control. Hunger-driven feeding: For hunger-driven feeding analysis, age-matched W1118 flies were given a normal diet or HSD for 5, 7, 14, 21, 24, and 28 days after an initial 7 days of development on a normal diet.
All other experiments were performed for 7-day or 14-day durations. Sixteen hours before feeding behavior assessment, half of the flies from each treatment were moved to normalized to the control fed group as a percentage. A signal above 40 (a.u.) was considered a feeding event. Analysis of feeding events was performed using R.
Triglyceride Measurements: TAG assays were carried out as previously described in ( To generate the 3D nuclear masks, Lamin or DAPI stacks were first maximally projected along the z-axis. After global thresholding, basic morphological operations, and watershed transform, the locations of the nuclear centroids in the x-y plane were used to scan the z-stacks and reconstruct the nuclear volume. The whole-cell volume was approximated from the nuclear volume by dilation.
Atg8 vesicles were segmented within each cell volume by performing morphological filtering and image opening with defined structural elements. The accuracy of the segmentation was assessed by manual inspection of random cells.
Once these cellular compartments had been defined, they were used as masks to extract intensity values of the signal of interest and to compute either the mean pixel intensity in the nucleus, the ratio of the integral intensities between the nucleus and the whole cell, or between the combined vesicles and the whole cell.
Two-sided Wilcoxon rank-sum tests were used to assess the statistical significance of pairwise comparisons between experimental conditions.
We determined that our data-points were normally distributed, based on two measures: i) A GraphPad outlier test did not identify any outliers in our data; and ii) the majority of our data points for a particular condition were relatively similar to one other, with only a small standard error of mean or standard deviation.              ) with GFP tag in the endogenous Upd2 locus (that rescues the upd2Δ mutant-see Figure S1). Fat is stained with GFP antibody (in a, b) in well-fed flies (a), or flies starved for 4 hours. Upd2 shows increased nuclear accumulation on starvation [see look up table (LUT)]. Scale bar is 1µM. (A') Upd2-GFP endogenous nuclear accumulation is quantified using 3D-volumetric segmentation-based calculations (See Methods). Each dot represents a fat cell nucleus, 75-100 fat nuclei were counted per condition per genotype. Statistical significance is quantified by the two-sided Wilcoxon test. Also see Figure S1 for similar experiment performed on an HA-tag knock into the Upd2 genomic locus (Upd2-HA).
(B) Effect of starvation and re-feeding on Upd2's nuclear accumulation in Drosophila S2R+ cells: Confocal micrographs of single optical-slices of Drosophila S2R+ cells transiently transfected with Upd2-WT::GFP (green; anti-GFP) and co-stained with Lamin (red). Scale bar is 5um. In B', the ratio of Upd2::GFP nuclear/whole cell intensity is plotted. Each dot represents a cell, 50-100 cells were counted per time point. Statistical significance is quantified by the two-sided Wilcoxon test.
(C) Effect of starvation and re-feeding on Upd2's release into media: Normalized percent fold change in secreted GFP signal detected by GFP sandwich ELISA assay performed on conditioned media of S2R+ cells transiently transfected with Upd2-WT::GFP. Cells were incubated either with complete media (control) or Amino-Acid (AA) free media for the indicated times. The refeeding % fold change is indicated with respect to starvation. Statistical significance is quantified by unpaired two-tailed t-test on 6 biological replicates per condition.   In a, note that Upd2-AIM-::GFP displays nuclear accumulation and so does VHH-Myc, consistent with its role as an antibody to GFP (see arrows). In b, punctate overlap between VHH Atg8 and GFP are observed-see arrow and inset. Note Upd2-AIM-is present in the cytosolic puncta, instead of nuclear accumulation (compare with Figure 2Cb). Also see Figure 3B' for quantification of the nuclear/cytosolic localization of Upd2-AIM-in the presence of VHH-Myc::Atg8-WT compared to control (VHH-Myc empty).    In (A) and (B), schematics illustrate experimental design. VHH tag, a small 15kDa antibody to GFP, is the control and VHH tag fused to the N-terminal of Atg8-WT or Atg8-GA mutation (Atg8-PE-) Also see companion Figure  A' and B' graphs show 3D-volumetric quantification of nuclear to whole cell intensity of GFP signal. Statistical significance is calculated using two-sided Wilcoxon sum rank tests. Each circle represents a cell.    Figure S4: companion to Main Figure 3 A B
(A) Western blot on lysates prepared from cells transfected with mCherry::Atg8-WT or mCherry::Atg8-G A. Cells were either maintained in complete media (Fed-CM) or AA-acid starved for 2 hours (2hr) and 4 hours (4hr). The cell lysates were probed with anti-RFP antibody. Atg8 (15kDa Atg8+ 27kDa mCherry) is recognized at ~42kDa. On 4 hr AA starvation, only in Atg8-WT, a lower doublet band is observed (see arrow), but not in the mCherry::Atg8 GA, suggesting defective lipidation. Loading is shown as a total protein stain for the blot.  GFP::Atg8-WT nuclear mean pixel intensity 10 4 p=7.5e-05 p=0.008 p=0.0008       Genes that are significantly upregulated during starvation are shown as red dots and downregulated genes as blue dots. In the control flies (a), among the six genes upregulated on starvation, 2 genes encode genes involved in innate immunity IBIN and SPH93 are not upregulated (See Main Figure 6B).

anti-Lamin anti-GFP
(B) Upd2 mRNA levels assessed by RNAseq in the indicated genotypes. (a) Each bar represents difference in fold change between fed state control (Lpp-Gal4> UAS-Luciferase) and either Upd2-WT (green) or Upd2-AIM-overexpression in the WT background. Error bars represent the standard deviation of log2 fold change among biological replicates (n=3). Significance was calculated using t-test and p<0.05 were considered significant. *** represents p<0.0001. Note Upd2-AIM-and Upd2-WT expression levels are comparable. (b) Each bar represents difference in fold change between starved and fed state of the indicated genotype. Error bars represent the standard deviation of log2 fold change among biological replicates (n=3).

Atg8-knock down: Fed
Atg8-PE-: Starved Atg8-PE-Atg8-membrane recruitment  Each illustration summarizes the key findings of the study. The upper panel is the key. Top three panels are fed state and lower three panels are in acute starvation (stv). Each panel illustrates a single cell. Notations: N-nucleus, C-cytosol, E-extracellular space.
• Top left: In a normal baseline fed state, the ortholog of human Leptin in flies, Upd2, enters the nucleus. Subsequently, Upd2's Atg8-interaction motif (AIM) mediates interactions with Atg8 that enables Upd2 nuclear exit to the cytosol and then Upd2 is released to the extracellular space.
• Bottom left: During starvation, Atg8 transits to the cytosol and is lipidated. Depletion of the nuclear pool of Atg8 on starvation increases Upd2 nuclear accumulation and causes cells to retain the adipokine on starvation. Thus, Atg8's lipidation state serves as a switch to license adipokine Upd2's cytosolic or nuclear localization.
• Top middle and right: When Atg8's interaction with Upd2 is disrupted (Upd2-AIM-) and when Atg8 levels are reduced by RNAi-mediated knock-down, Upd2 accumulates in the nucleus even in fed cells. This provides evidence for Atg8's role in facilitating Upd2's cytosolic localization in fed cells. We note that constitutive expression of Upd2-AIM-in fly fat renders flies sensitive to a HSD. This suggests that maintaining Upd2 in the nucleus, during a nutrient state where it is usually cytosolic, has a negative impact on resilience to a surplus diet.
• Bottom middle: Upd2-AIM-remains in the nucleus on starvation, while Atg8 translocates to the cytosol. Upd2-AIM-flies are starvation resilient and upregulate a systemic innate immune gene expression signature.
• Bottom right: Lipidation-defective Atg8-PE-continues to be nuclear on starvation. Presence of Atg8-PE-allows Upd2 to be cytosolic even on starvation and renders flies sensitive to starvation.