A set of 460 genes which included known transcription factors, cell death and autophagy genes, as well as genes associated transcriptionally with ecdysone-induced cell death of the Drosophila salivary gland were chosen for this functional study. In our initial screen, cells were treated with ecdysone and dsRNA corresponding to each of these genes. A viability assay (WST-1) was used to identify genes with potential pro-death or pro-survival functions, and microscopy was used to visualize cell morphology. For the reproducible positive hits, a second dsRNA was used to confirm the viability phenotype. The screen identified 7 potential pro-death and 18 pro-survival genes. The RNAi/WST-1 screen was repeated with and without ecdysone to determine ecdysone dependency for the 25 identified genes. TUNEL/DAPI and BrdU assays were performed to identify genes with cell death and/or cell proliferation related effects.

DAPI staining (magenta) and TUNEL assays (green) were performed 3 days following addition of dsRNA and ecdysone to l(2)mbn cells. The cells with no ecdysone and no dsRNA (No treatment) showed a background level of 11+/−4% TUNEL positivity (Panel A). Ecdysone (not shown) and ecdysone plus the human dsRNA NM_138278-negative control resulted in an increase in TUNEL positive cells (56+/−5%; Panel B). dsRNA corresponding to diap1 (control gene) increased TUNEL positive cells without (not shown) or with ecdysone treatment (77+/−6%; Panel C). Ecdysone plus RNAi of Sox14 (Panel D) or EcR (Panel E) resulted in decreased TUNEL positive cells (37+/−6% and 9+/−0.5%, respectively), whereas ecdysone plus RNAi of a novel gene CG32016 (Panel F) increased TUNEL positive cells (72+/−8%) compared to the negative control.

Control wt late 3rd instar larva show normal trachea (left, arrow) but Sox14-RNAi late 3rd instar larva displayed tracheal defects detectable as blackened regions (right, arrow). (B) Dorsal tracheal trunks of Sox14-RNAi animals (right) showed severely distorted taenidial folds, collapse of the tracheal cuticle and blackening of the cuticle (arrow). Branching of the trachea in Sox14-RNAi animals appears to occur normally (arrowhead) as in control wt animals (left). (C) Sox14-RNAi pharate adults dissected from the pupal case showed defects in eyes, notum, and bristles (right, arrow). A control wt pharate adult with normal eyes, notum and bristles (arrows) is shown on the left for comparison. (D) In the 0 hr APF larval midgut (LM) from both control wt (left) and Sox14-RNAi animals (right), the proventriculus (PV) and gastric caecae (GC) are evident. (E) In wt animals, the midgut condensed by 4 hrs APF (not shown) and the gastric caecae were not detectable by 7 hrs APF. Condensed wild type larval midgut (LM) was found within the adult midgut by 7–12 hrs APF (left). Sox14-RNAi animals showed some condensation of the midgut but the proventriculus (PV) and remnants of gastric caecae (GC) can still be observed even after 7–12 hrs APF (right). In both wt and Sox14-RNAi animals, a layer of adult midgut (AM, arrow) can be observed after 4 hrs APF.

(A) l(2)mbn cells treated with 10 µM ecdysone showed a >70% reduction in live cells after 72 hours. Cells were exposed to 10 µM ecdysone for the times indicated. Surviving cells were counted by Trypan blue exclusion and percent live cells were calculated by comparing ecdysone-treated cells with untreated cells (100%). (B) QRT-PCR expression profiling showed that the ecdysone induced genes E75, BR-C, E93, rpr, hid and dronc had elevated levels of expression (at least 4 fold increase) in ecdysone treated (10 µM) l(2)mbn cells relative to untreated control l(2)mbn cells. The dotted lines above the bars for E93 indicate that the fold change in expression exceeded the existing scale (48 hrs, 126 fold change; 72 hrs, 371 fold change). The autophagy genes, DmAtg4-like (CG6194), DmAtg6(CG5429), and DmAtg5(CG1643), did not show differential expression following ecdysone treatment. (C) QRT-PCR analysis of gene transcripts following treatment with the indicated dsRNAs and ecdysone. As shown here, the knockdown ranged between 63% and 90% for the representative gene transcripts tested following 72 hr dsRNA and ecdysone treatment compared to ecdysone treatment alone. (D) Cells treated with dsRNA corresponding to BR-C, EcR, dronc, rpr, and hid showed significantly (p≤0.05) increased levels of cell viability and those treated with dsRNA corresponding to diap-1 showed significantly (p≤0.05) reduced levels of cell viability compared to cells treated with ecdysone and human dsRNA NM_138278 (negative control). Cell viability was measured by the WST-1 assay (A₄₅₀–A₆₅₀). The dotted lines above the bar for EcR indicate the A₄₅₀–A₆₅₀ value ( = 1.0) exceeded the existing scale. For (A)–(D), the error bars represent the SD of triplicate samples.

Cellular phenotypes were visualized 3 days after dsRNA treatment in the presence or absence of ecdysone. The l(2)mbn cells with no ecydsone and no dsRNA (No treatment) were round and uniform in size and shape. l(2)mbn cells treated with ecdysone (not shown) or ecdysone+human dsRNA NM_138278-negative control changed in shape, from round to spindle forms with extensions () (19+/−5%). Large cells with phagocytosed material () and apoptotic bodies () were also observed. RNAi of EcR, dronc and Sox14 each increased viability of the ecdysone treated cells, but their resulting morphologies were distinct. The % of observed spindle shaped cells (top right corner) was quantitated for dsRNAs corresponding to the genes indicated. RNAi of EcR inhibited spindle shape formation (1+/−2% of the cells were spindle shaped), cells remained rounded, and no apoptotic bodies were found. RNAi of dronc inhibited apoptotic body formation, but cells became spindle shaped (20+/−8%). Also, signs of necrosis such as inflated and seemingly empty cells and cell fragments were observed (). RNAi of Sox14 showed few apoptotic bodies and 30+/−8% of the cells were spindle-shaped. RNAi of diap1 resulted in formation of numerous apoptotic bodies within 24 hr and no spindle shaped cells were found. RNAi of Rpn2 showed numerous apoptotic bodies in the presence and absence of ecdysone. while RNAi of Kap-a3 showed a dramatic increase in apoptotic bodies only in the presence of ecdysone (see Table 2 for quantitation of TUNEL positive cells).

(A) QRT-PCR expression profiling of l(2)mbn and S2 cells treated with 10 µM ecdysone for 72 hrs showed a 5 fold and 4 fold increase, respectively, in Sox14 transcripts compared to untreated control cells. Expression of Sox14 was normalized to the housekeeping gene rp49. (B) l(2)mbn cells transfected with either N- or C-terminal FLAG tagged Sox14 constructs showed decreased cell viability after 48 hrs compared to cells transfected with the control vector only (p≤0.001). (C) Sox14 expression in wild-type and Sox14-RNAi whole animals. In control wild type (wt; strain w¹¹¹⁸) whole animals, increased transcript levels of Sox14 were observed at 0 hrs APF compared to wandering larvae. In Tub-Sox14-RNAi (Tubulin-GAL4/Sox14-RNAi) animals, Sox14 transcripts were dramatically reduced at both stages compared to the control (89+/−2% and 91+/−2% reduction in transcript levels at the wandering larva and 0 hrs APF stages, respectively). Error bars in (A)–(C) represent the SD of triplicate samples.

A. In Tub-Sox14-RNAi pupae, intact salivary glands persisted past the stage of wild-type salivary gland cell death (15–18 hrs APF at 25 C). Shown is a segment of an intact salivary gland from a Tub-Sox-14-RNAi pupa that is devoid of TUNEL positive nuclear staining (DAPI = magenta; TUNEL = green). Phalloidin staining of retinas dissected from the same animal in (B) was used as an independent morphological marker to help determine developmental age. The ommatidial arrangement shown (far right) indicates that the animal developed to at least 30 hrs APF (25 C). B. In control pupae (D59-GAL4/D59-GAL4; MKRS/TM6B), extensive nuclear TUNEL staining (DAPI = magenta; TUNEL = green) and degradation (DIC) was evident at stages equivalent to 16–17 hrs APF at 25 C (or 30–32 hrs APF at 18 C). C. In D59-Sox14-RNAi pupae, salivary gland cell death was delayed. Shown is a salivary gland dissected at the same stage as the animal in (B). There is no evidence of TUNEL positive (green) nuclei (DAPI = magenta) and no indications of degradation (DIC). Scale bars for fluorescent and DIC images in A–C = 100 microns. D. Quantitation of salivary gland degradation and persistence in control and D59-Sox14-RNAi pupae raised at 25 C and 18 C. In control animals at 25 C, no salivary glands were detected by 18–19 hrs APF. In D59-Sox14-RNAi animals dissected at 18–19 hrs APF, completely intact salivary glands (persisting salivary glands) were found in 50% of the animals, and some degradation was observed (partial degradation) in the remaining 50%. By 21 hrs APF (25 C), salivary glands were not detected in this strain. In control animals at 18 C, 57% showed partial salivary gland degradation and 43% showed complete degradation by 30–32 hrs APF; salivary glands were not detected by 40 hrs APF. In D59-Sox14-RNAi pupae at 18 C, 91% of the animals had persisting salivary glands at 30–33 hrs APF. At 40–45 hrs APF, one-third of the animals had intact salivary glands, one-third had partially degraded salivary glands, and the remaining third had no detectable salivary glands. Salivary glands were examined from at least 10 animals at each time point and at each temperature. E. QRT-PCR analysis of gene expression in control wild-type (w¹¹¹⁸) and Tub-Sox14-RNAi (Tubulin-GAL4/ Sox14-RNAi) animals. In Tub-Sox14-RNAi animals, Sox14 transcripts were reduced at both stages compared to the control. In control animals, apoptosis genes rpr and dronc and the ecdysone regulated transcription factor gene E93 showed a 4 to 5 fold increase in expression at 0 hrs APF compared to wandering larvae. However, the transcript levels corresponding to these genes showed a reduction at both stages in the Tub-Sox14-RNAi strain compared to the control strain. The ecdysone regulated transcription factor BR-C did not show expression level differences between wild-type and Tub-Sox-14-RNAi animals at the stages examined. Error bars represent the SD of triplicate samples.