Expression of dominant negative Drosophila JNK (Bsk^DN) in GFP-labeled clones of midgut epithelium does not affect unstressed intestinal cells (compare top two panels). However, paraquat-induced autophagy, as visualized by GFP-LC3, is strongly suppressed in Bsk^DN expressing clones (compare bottom two panels). The histograms below to the micrographs show the quantitative analysis of autophagosome abundance per cell. The number of GFP-LC3 labeled autophagosomes or autolysomes per cell was counted and is expressed as follows: high: >10 GFP-LC3 foci/cell; medium: 6-10 GFP-LC3 foci/cell; low: <6 GFP-LC3 foci/cell. At least 30 clones were analyzed per condition. Statistical analysis was done in SPSS 13.0 by Pearson Chi test. p-values: Bsk^DN,GFP-LC3 group: paraquat/control, p=0.038; GFP-LC3 group: paraquat/control, p<0.00001; control: Bsk^DN,GFP-LC3/GFP-LC3, p=0.230; paraquat: Bsk^DN,GFP-LC3/GFP-LC3, p<0.00001.
Genotypes: Control clones: hsFLP/+; UAS-GFP-LC3/+; Act>y>Gal4/+
Bsk^DN clones: hsFLP/+; UAS-GFP-LC3/+; Act>y>Gal4/UAS-Bsk^DN
Scale bars: 18.75 μm.

Ectopic expression of Hep^act in clones of larval gut cells (marked with GFP-LC3) strongly increases the number of Lysotracker/GFP-LC3 marked foci (arrowheads), indicating increased autophagic activity. Out of 71 Hep^act-expressing clones, 43 (61%) showed enhanced autophagy. Non-clonal wild type cells (non-GFP stained cells, outlined by dashed line) show only sporadic Lysotracker Red foci.
The same experiment was also conducted in larval fat body cells. Note that in this case the clones are marked by expression of regular GFP (not fused to LC3) hence the autophagosomes are not labeled in the green fat body panel. Out of 24 fat bodies that clonally expressed Hep^act, 19 (79%) showed enhanced numbers of Lysotracker Red stained foci.
B) Transmission Electron Microscopy of midgut cells undergoing autophagy in response to JNK signaling or starvation. No obvious autophagosome structures were detected in the control midgut cell shown in the upper panel. Conversely, similar cells in which JNK has been ectopically activated by Hep^act expression (middle panel), or midgut cells from starved larvae (lower panel) display frequent autolysosomes. These organelles are discernable as membranous vesicles containing whole organelles (green arrowhead) or electron-dense material, presumably representing degradation products (red arrowhead). “nu” indicates nucleus. Insets show the autophagosomes at higher magnification.
Genotypes:
A) control clones in midgut: hsFLP/+; UAS-GFP-LC3/+, UAS-Hep^act/+; Act>y>Gal4/+ Hep^act clones in midgut: hsFLP/+; UAS-GFP-LC3, UAS-Hep^act/+; Act-Gal4/+
B) control clones in fat body: hsFLP/+; UAS-GFP, Act>y>Gal4/UAS-Hep^act Hep^act clones in fat body: hsFLP/+; UAS-GFP, Act-Gal4/UAS-Hep^act
C) Control and starved: w; T80Gal4, UAS-eGFP/+; tub-Gal80^ts/+ Hep^act : w; T80Gal4,UAS-eGFP/UAS-Hep^act; tub-Gal80^ts/+ Scale bars: A) midgut: 18.75 μm; fat body: 15.87 μm; C) 1μm

A) The pictures in the left hand panel (control) show midgut epithelial cells of mock-treated 3^rd instar larvae. GFP-LC3-marked autophagosomes (green) and Lysotracker Red-labeled lysosomes (red) are rare and do not coincide. 10 hours exposure to food containing 100 mM paraquat (middle panel, paraquat) or 7 hours exposure to food with 1.5% H₂O₂ (right hand panel, H₂O₂) induced the appearance of numerous LC3-GFP and Lysotracker Red double positive large autolysosomes. The bars below each panel show quantitative analyses of GFP-LC3 foci observed in each condition. Autophagy activity was defined as follows: high: >10 GFP-LC3 foci/cell; medium: 6-10 GFP-LC3 foci/cell; low: <6 GFP-LC3 foci/cell. In each group, more than 80 cells were randomly picked for GFP-LC3 focus counting. Quantification was done by Pearson Chi-square test in SPSS 13.0. p-values: in paraquat treatment, treated/control< 0.00001; in H₂O₂ treatment: treated/control< 0.00001.
B)ATG1 is required for starvation- and paraquat-induced autophagy. Midgut preparations from wild type or ATG1 homozygous mutant third instar larvae are shown. The larvae had either been starved for 1 hour in the presence of 5% sucrose only, exposed to food containing 100 mM paraquat or maintained on control food. The two left hand panels show that, as expected, starvation-induced autophagy activity is largely suppressed (fewer Lysotracker Red-positive foci) in an ATG1 mutant background. The right hand panels show that paraquat-induced autophagy is similarly decreased in ATG1 deficient intestinal cells. The bars below each panel show a quantification of Lysotracker Red-positive foci observed in each condition. Autophagy activity was defined as follows: high: >10 Lysotracker Red foci/cell; medium: 6-10 Lysotracker Red foci/cell; low: <6 Lysotracker Red foci/cell. Lysotracker Red-positive foci in 35-61 randomly chosen cells were counted for quantification using Pearson Chi-square test in SPSS 13.0. p-values: in starvation, ATG1^-/-/control< 0.0001; in paraquat treatment: ATG1^-/-/control= 0.0007.
C) Larval midgut treated with 100 mM paraquat for 8 hours (right panel) shows elevated JNK activity as measured by the JNK reporter puc-LacZ. The untreated control shown in the left panel displays basal activity of the reporter in the connection between proventriculus, midgut (peritrophic mill), and foregut ring.
Genotypes: A) UAS-GFP-LC3/+; DaGal4/+; B) ATG1^-/-: ATG1^Δ3D/ ATG1^Δ3D; Control: ATG1^Δ3D /TM3; C) puc^E69, ry/TM3, ry
Scale bars: in A) 10 μm and B) 18.75 μm.

A) In starved larval midgut cells autophagy, as monitored by the frequency of Lysotracker Red stained autolysomes, is prevalent. Suppression of JNK activity in Bsk^DN-overexpressing clones (labeled with GFP, green, and outlined in white on the bottom panel) displays a similar number of autophagic vesicles (arrowheads) as surrounding control cells. In contrast, autophagy is suppressed in clones overexpressing the catalytic subunit of PI3K, Dp110 (green cells).
B) In fat body of early third instar larvae, before the onset of metamorphosis, autophagy can be induced by starvation and the resultant loss of PI3K signaling (Scott et al., 2004). Overexpression of Dp110, but not of UAS-Bsk^DN (GFP positive cell clones) inhibits this starvation-induced autophagy.
C) The prominent autophagy in late third instar fat body that is associated with metamorphosis is not affected in UAS-Bsk^DN-overexpressing clones (marked by GFP expression, green) or in clones homozygous for the JNK null allele bsk^170B (marked by the absence of GFP). The clones, outlined with a white line in the bottom panel, show a similar pattern of Lysotracker red stained autophagosomes (arrowheads) to the surrounding non-clonal cells.
The histograms below to the micrographs show the quantitative analysis of autophagosome number per cell in clones expressing the UAS constructs indicated, or in the case of the right hand panel of C, clones that are homozygous for the JNK loss of function allele bsk^170B. The autophagosome count in non clonal control cells is represented in the same manner. The number of Lysotracker Red foci per cell was counted and is categorized as follows: high: >10 Lysotracker Red foci /cell; medium: 6-10 Lysotracker Red foci /cell; low: <6 Lysotracker Red foci/cell. At least 16 clonal cells were analyzed per genotype per condition. Statistical quantification was performed in SPSS 13.0 using Pearson Chi test. p-values are: in A) Bsk^DN: control/starvation, p=0.947, Dp110: control/starvation, p<0.00001; in B) Bsk^DN: control/starvation, p=0.310, Dp110: control/starvation, p=0.0095; in C) control/Bsk^DN, p=0.252; control/bsk^170B, p=0.157;.
Genotypes:
A), B), and C) Bsk^DN clones: hsFLP/+; Act>CD2>Gal4,UAS-eGFP/+; UAS-Bsk^DN/+ and Dp110 clones: hsFLP/+; Act>CD2>Gal4, UAS-eGFP/UAS-DP110
C) bsk^170B clones: hsFLP/+; ubi-GFP, FRT40A/ bsk^170B, FRT40A, and FRT40A clones: hsFLP/+; ubi-GFP, FRT40A/FRT40A.
Scale bars: in C) bsk^170B clone, 25.4 μm, and all other clones: 18.75 μm.

A) ATG1, ATG18.2, and puc mRNA levels rise transiently after adult flies are exposed to food containing paraquat. Total RNAs of 3-5 day old flies treated with paraquat, or mock-treated, for the time points indicated were prepared, and the mRNA level of the above-mentioned three genes were quantitated by real time RT PCR. For each time point, six independent biological repeats were carried out in triplicate each. The ratio between paraquat-treated and control samples is shown as fold induction (mean ± standard error). The confidence in the changes of ATG1, ATG18.2, and puc mRNA levels of upon paraquat-treatment was determined by Student's t-test. The p values for the comparison of the paraquat-treated group at the 10.5 hour time point to control is: puc: p<0.00001; ATG1: p<0.00001. ATG18.2: p<0.00001.
B) Elevated expression of multiple ATG genes in response to JNK activation by the constitutively active JNKK, Hep^act, expressed in early 3^rd instar larvae using the inducible Gal80^ts system (McGuire et al., 2003). Larvae of the genotype T80Gal4, UAS-eGFP/UAS-Hep^act; Tub Gal80^ts/+ express Hep^act at 29°C, but not at 21°C. The reference strain is T80Gal4, UAS-eGFP/+; tub-Gal80^ts/+. The mRNA levels of six ATG genes and puc, a known JNK target gene, were measured by real time RT PCR before and after induction of Hep^act expression for six hours or mock induction in the control strain. The histogram shows that ratio of RNA levels between the control and the Hep^act expressing strain. Before induction (0 hrs Hep^act), this ratio is close to 1 as depicted in the left histogram. After transgene activation ATG and puc mRNA levels in the UAS-Hep^act strain exceeded those in the reference strain by a factor of 2 to 5 (right panel, 6 hrs Hep^act). The differential activation of the ATG genes was statistically evaluated by Student's t-test. Single asterisks indicate that the expression level of ATG genes in the JNK activation group was significantly up-regulated (p<0.05), and double asterisks signify p<0.01. p values are ATG1: p=0.006, ATG4: p=0.006, ATG6: p=0.003; ATG8a (CG32672): p=0.019; ATG18.1 (CG7986): p=0.002; ATG18.2 (CG8678): p=0.005. The fold induction is shown as mean ± standard deviation of three independent experiments, each in triplicate.

A) Heterozygosity for the gene encoding the JNK phosphatase Puckered (puc^E69/+), which causes mildly elevated levels of JNK activity, confers increased resistance against paraquat. Loss of one copy of ATG1 due to heterozygosity for the ATG1^Δ3D null allele, in ATG1^Δ3D/puc^E69 flies, significantly decreases JNK-dependent paraquat resistance to a level approaching that of isogenic control flies (+/+). Loss of one copy of ATG1 in an otherwise wild type background has no significant effect on paraquat resistance in comparison to wild type flies. Similar effects were seen with an independent allele of ATG1, ATG1^EY09216(B) and with a loss of function allele of ATG6, ATG6⁰⁰⁰⁹⁶(C). All stocks were isogenized by extensive backcrossing (at least 10 generations) into the cognate genetic backgrounds (w¹¹¹⁸ and ry⁵⁰⁶). The exact genotypes are listed below. Paraquat exposure times and levels were calibrated to achieve approximately 50% lethality in the wild type controls (15 to 20 mM paraquat for 24 to 48 hours). Data are shown for female flies. Panels A), B), and C) show independently conducted experiments. While the mortality of dietary paraquat is reproducible from vial to vial within one experiment, there is a considerable degree of variability if the experiments are done on different days. This is probably caused by batch variations of food, paraquat or other parameters such slight fluctuations in temperature or humidity. Importantly, the direct comparisons as depicted in Fig 2 are between mutants in isogenic conditions and treated in parallel. The results for the different allele combinations are qualitatively similar between males and females (data not shown).
D) Ubiquitous overexpression of ATG6 from a UAS-ATG6-HA transgene driven by the RU486-inducible tubGSGal4 promotes the survival of flies challenged by paraquat. Induction of ATG6 expression by feeding flies with RU486 for 4 days decreased mortality of a subsequent 20 mM paraquat exposure compared to a mock-treated group. Data for male flies are shown, females display a comparable effect. p values between the death rate of different genotypes were obtained via Student's t-test. The histograms shown above depict a representative experiment out of 2-4 independent repeats for each set of genotypes. Sample size and p-values used in this figure are as follows:
A) +/+, n=120; puc^E69/+, n=98; puc^E69/ATG1^Δ3D, n=69; ATG1^Δ3D/+, n=79; p values: (+/+)/(puc^E69/+), p=0.009; (+/+)/(puc^E69/ATG1^Δ3D), p=0.784; (+/+)/(ATG1^Δ3D/+), p=0.114; (puc^E69/+)/(puc^E69/ATG1^Δ3D), p=0.011; (puc^E69/+)/(ATG1^Δ3D/+), p=0.004; (puc^E69/ATG1^Δ3D)/(ATG1^Δ3D/+), p=0.246; B) ATG1^EY09216/+, n=58; +/+, n=120; puc^E69/ATG1^EY09216, n=59; puc^E69/+, n=104; p values: (+/+)/(puc^E69/+), p= 0.002; (+/+)/(puc^E69/ATG1^EY09216), p=0.002; (+/+)/(ATG1^EY09216/+), p=0.235; (puc^E69/+)/(puc^E69/ATG1^EY09216), p= 0.012; (puc^E69/+)/(ATG1^EY09216/+), p=0.047; (puc^E69/ATG1^EY09216)/(ATG1^EY09216/+), p=0.455; C) ATG6⁰⁰⁰⁹⁶/+, n=88; +/+, n=82; puc^E69/ATG6⁰⁰⁰⁹⁶, n=77; puc^E69/+, n=125; p values: (+/+)/(puc^E69/+), p=0.005; (+/+)/(puc^E69/ATG6⁰⁰⁰⁹⁶), p=0.190; (+/+)/(ATG6⁰⁰⁰⁹⁶/+), p=0.058; (puc^E69/+)/(puc^E69/ATG6⁰⁰⁰⁹⁶), p= 0.002; (puc^E69/+)/(ATG6⁰⁰⁰⁹⁶/+), p=0.011; (puc^E69/ATG6⁰⁰⁰⁹⁶)/(ATG6⁰⁰⁰⁹⁶/+), p=0.395; D) tubGS>w1118: +RU486, n=152, -RU486, n=144, p-value= 0.930; tubGS>UAS-ATG6-HA: +RU486, n=118, -RU486, n=107, p-value=0.012.