GLYAT regulates JNK-mediated cell death in Drosophila

Cell death is a fundamental progress that regulates cell number, tissue homeostasis and organ size in development. The c-Jun N-terminal kinase (JNK) pathway has been evolutionarily conserved from fly to human, and plays essential roles in regulating cell death. To characterize additional genes that regulate JNK signaling, we performed a genetic screen in Drosophila and identified dGLYAT, a novel gene whose function was previously unknown, as a modulator of JNK-mediated cell death. We found that loss of dGLYAT suppressed JNK activation and cell death triggered by over-expression of Egr or Hep, or depletion of puc or lgl in development, suggesting dGLYAT regulates both ectopic and physiological functions of JNK pathway. Furthermore, we showed that loss of dGLYAT inhibits JNK-mediated ROS production, suggesting dGLYAT regulates multiple functions of JNK signaling in vivo.

N-acyltransferase (GLYAT), and is referred to as dGLYAT hereafter. The mutant, PBac{PB}CG34010 c02982 , has a piggyBac insertion into the second exon and generates a truncated protein that deletes most of the coding region including the critical Gcn5-related N-acetyltransferases (GNAT) domain. Thus, the mutant is most probably a null allele for dGLYAT. Interestingly, the mutant is homozygously viable, and does not produce any discernible phenotype, suggesting it is not essential for normal development. Furthermore, RNAi-mediated depletion of dGLYAT also suppressed the GMR > Egr induced small eye phenotype (Fig. 1e), compared with the expression of a UAS-GFP transgene that served as a negative control (Fig. 1c), confirming that dGLYAT is required for ectopic Egr-triggered morphological defect. Expression of RNAi-mediated depletion of Bsk, the Drosophila JNK ortholog, served as a positive control (Fig. 1f). Consistently, GMR > Egr-triggered cell death, indicated by acridine orange (AO) staining, posterior to the morphogenetic furrow (MF) in 3 rd instar eye discs (Fig. 1h), was significantly impeded by loss of dGLYAT or Bsk (Fig. 1j-l), but remained unaffected by the expression of GFP (Fig. 1i). The statistics of adult eye sizes (Fig. 1m) and apoptotic cell numbers in larval eye discs (Fig. 1n) were shown. Taken together, the above data suggest that dGLYAT is physiologically required for ectopic Egr-induced cell death in eye development.
Loss of dGLYAT impedes ectopic Hep-induced cell death in eye development. GMR > Egr triggers cell death via two independent pathways, the caspase pathway and the JNK pathway 34 . To examine whether dGLYAT is required for caspase-mediate cell death, we overexpressed Drosophila p53 (Dp53), a pro-apoptotic gene that triggers caspase-mediated cell death [36][37][38][39] , in the eye by GMR-Gal4. We found that GMR > Dp53-triggered small eye phenotype was not suppressed by loss of dGLYAT ( Figure S1), suggesting dGLYAT is not involved in caspase-mediated cell death. To investigate the role of dGLYAT in JNK-mediated cell death, we expressed a constitutive active form of the Drosophila JNK kinase Hemipterous (Hep) in the developing eye. GMR > Hep CA induces JNK-mediated cell death in eye discs (Fig. 2h) and produces a small eye phenotype in adults (Fig. 2b) 31,33 . Both phenotypes were significantly suppressed by loss of dGLYAT or depletion of Bsk, but not the expression of GFP ( Fig. 2c-f,i-n). Thus, dGLYAT is necessary for ectopic Hep-induced JNK-mediated cell death in eye development. Loss of dGLYAT inhibits JNK activation in eye discs. The above data suggest that dGLYAT is necessary for JNK-mediated cell death in eye development, yet it remains unknown whether dGLYAT is required for JNK pathway activation. To address this, we checked the expression of puc-LacZ, a well-known readout of JNK signaling in Drosophila 25,40,41 . We found that GMR > Egr induced strong puc-LacZ expression posterior to the morphogenetic furrow (MF) in the eye disc (Fig. 3b), which was remarkably inhibited by loss of dGLYAT.  Again, expression of Bsk and GFP were served as a positive and negative controls, respectively ( Fig. 3c-f). Hence, dGLYAT is necessary for JNK signaling activation in eye development.
dGLYAT modulates JNK-mediated cell death in other tissues. To investigate whether dGLYAT modulates JNK-mediated cell death in other tissues, we examined the interaction between dGLYAT and JNK signaling in the developing wing, another important tissue frequently used for genetics studies. Ectopic expression of Egr driven by ptc-Gal4 (ptc > Egr) was able to induce extensive cell death along the anterior/posterior (A/P) compartment boundary in 3 rd instar larval wing discs (Fig. 4h) and produce loss of the anterior cross vein (ACV) phenotype in adult wings (Fig. 4b) 22,42 . We found that both phenotypes were significantly suppressed by loss of dGLYAT or Bsk, but not the expression of GFP (Fig. 4c-f,i-n). Thus, dGLYAT is also required for ectopic Egr-triggered cell death in wing development.
( Fig. 6c-f). Collectively, these data indicate that dGLYAT modulates JNK-mediated cell death in a non-tissue specific manner.
To investigate whether expression of dGLYAT is able to trigger JNK activation and cell death, we drove dGLYAT expression in the developing eye or wing by GMR-Gal4 or ptc-Gal4, respectively. We found that ectopic expression of dGLYAT did not trigger JNK signaling activation ( Figure S2f Figure S2h) in the adult. Thus, expression of dGLYAT by itself is not sufficient to trigger JNK activation and cell death. Consistently, expression of Bsk, the fly JNK ortholog, or dTRAF2 that acts upstream of dTAK1, is not sufficient to induce JNK activation and cell death 23 . It remains to be explored whether expression of an activated form of dGLYAT, or co-expression of dGLYAT with its co-factor(s), is able to induce JNK activation and cell death in development.
dGLYAT is required for physiological activation of JNK signaling. The above data suggest that dGLYAT is important for ectopically activated JNK signaling-induced cell death, yet it remains unclear whether dGLYAT modulates the physiological functions of JNK signaling. To address this question, we knocked down puc, a negative regulator of JNK signaling, by the ptc-Gal4 driver. Depletion of puc induced robust cell death in third i, n = 7; j, n = 9; k, n = 10; l, n = 6). n.s., P > 0.05; ****P < 0.0001; **P < 0.01.
instar larval wing discs, as detected by AO staining (Fig. 7a and b). Intriguingly, the phenotype was significantly impeded by expressing a dGLYAT RNAi, but not GFP (Fig. 7c and d), suggesting dGLYAT is essential for physiologically activated JNK-induced cell death.
It has been reported that loss of cell polarity in wing disc epithelial results in JNK-mediated cell death 43,44 . Consistently, knockdown the cell polarity gene lethal giant larva (lgl) by ptc-Gal4 promotes JNK-mediated cell death along the A/P boundary in third instar larval wing discs (Fig. 8b) 30,31 . This phenotype was significantly suppressed in heterozygous dGLYAT mutant or by the expression of a dGLYAT RNAi (Fig. 8c-e). Therefore, dGLYAT modulates the physiological function of JNK signaling in development.    29 . To examine whether dGLYAT is crucial for JNK-mediated ROS activation, we detected ROS level in third instar larval eye discs by CellROX staining 29 . Consistent with previous study 28,29,45 , ectopic expression of Egr (GMR > Egr) promoted abundant ROS production ( Fig. 9a and b), which was considerably suppressed by mutation or RNAi-mediated depletion ( Fig. 9c and d) of dGLYAT, suggesting dGLYAT regulates JNK-mediated ROS activation in vivo.
Intriguingly, PFAMs, the catalytic products of hGLYATs, were reported to play a role in ROS activation 13,14,46 . However, the underlying mechanism has remained unknown, and a connection with JNK signaling has not been examined. Given the fact that both JNK pathway and its role in ROS have been highly conserved from Drosophila to human, it is plausible that GLYATs also regulate JNK signaling in mammals.

Summary. dGLYAT contains a conserved GNAT domain and is supposed to function as an Acyl-CoA
N-acyltransferase, yet its in vivo function has not been previously explored. In the present study, we identified dGLYAT as a crucial modulator of JNK pathway in vivo by using Drosophila as a model organism. We showed that loss of dGLYAT blocks not only ectopic Egr-or Hep-induced JNK activation and cell death, but also depletion-of-puc or lgl-triggered physiological JNK activation and cell death in development. In addition, loss of dGLYAT impedes JNK-dependent ROS activation. Thus, dGLYAT regulates multiple physiological functions of JNK signaling in vivo, yet the molecular mechanism by which dGLYAT regulates JNK pathway remains unknown, and should be addressed by further investigations.

Materials and Methods
Drosophila Genetics and Stocks. All stocks were raised on a standard cornmeal and agar medium at 25 °C unless otherwise indicated. For experiments involving tub-Gal80 ts , eggs were allowed to develop at 25 °C for 2-3 days, then transferred to 29 °C for 2 days to inactivate Gal80.
X-gal staining. Wing and eye discs were dissected from third instar larvae in PBST (1 × PBS pH 7.0, 0.1% Triton X-100) and stained for β-galactosidase activity as described 49 . Microscopy and phenotype analysis. Flies of indicated genotypes were collected and frozen in −80 °C.
Wings were dissected and mounted on the slide in the alcohol/glycerol (1:1) medium, and flies were mounted in the alcohol on 3% agarose plate. Image of wings were captured with Olympus microscope BX51, and light image of eyes were captured with Olympus stereo microscope SZX16 32 .

Statistical analysis.
Results are presented in bar graphs and box graphs created with GraphPad Prism 6. For loss-of-ACV phenotype, statistics were analyzed by chi-square test. For AO staining and area of eye size, one-way ANOVA with Bonferroni's multiple comparison tests are used to calculate statistical significance. P-values are included in the relevant figure legends.