Figure 1.
Summary of the Gal4-Driven Phenotypic Comparison Screen
(A) Gal4 driver lines were crossed to various UAS alleles at 25 °C in order to identify novel dominant phenotypes. Phenotype modification was compared relative to control chromosome w1118. Crosses lethal to progeny are indicated in black, while viable crosses are indicated in white. Viable “escaper” flies from lethal crosses are indicated (E). Phenotypes generated in viable or escaper progeny are also indicated for each cross (P).
(B) Expression of rpr and dE2f1 resulted in significant lethality in the majority of Gal4 lines tested.
Figure 2.
Generation and Characterization of Tissue-Specific dE2F1-Dependent Apoptotic Phenotypes in Drosophila
The Act88F-Gal4 driver was crossed to either a wild-type chromosome (w1118; +) (A), UAS-dE2f1 (B), UAS-rpr (C), or UAS-CycE (D). Expression of dE2f1 and rpr, but not CycE, was sufficient to induce a ventrally curved, blistered, and gnarled wing phenotype. In order to visualize expression patterns, Act88F-Gal4 was crossed to either a wild-type chromosome (Canton-S) (E and F) or UAS-EGFP transgene (G and H). Act88F drives expression in the newly eclosed adult fly wing blade cells, and coexpression of dE2f1 removes many EGFP-positive cells (I and J). EGFP expression was quantitatively determined in single transgenic flies by fluorescent spectrophotometry (K). Expression of dE2f1 significantly reduced the levels of EGFP and was rescued by coexpression of Rbf, dDpDN, or p35, but not p21. Newly eclosed wild-type wings (L and N) or dE2f1-expressing wings (M and O) were stained with AO to visualize apoptotic cells. Cells expressing dE2f1 had strong, punctuate AO staining in the distal wing region where EGFP is mainly expressed early after eclosion.
DIC, differential interference contract.
Figure 3.
Genetic Characterization of a Recombinant Act88F-Gal4,UAS-dE2f1 Transgenic Stock
Various alleles were analyzed for modification of the dE2f1-dependent phenotype in trans. The wild-type wing phenotype is depicted in Act88F-Gal4,UAS-EGFP/+ (w1118) recombinant stock as control (A). The Act88F-Gal4,UAS-dE2f1/+ (w1118) recombinant stock phenotype (B) is strongly phenocopied by caspase expression (C). Coexpression of Rbf (D) or dDpDN (E) completely suppressed the dE2f1 phenotype. The dE2f1-dependent phenotype was also suppressed by various apoptotic regulators including coexpression of the caspase inhibitor baculovirus p35 (F) or dIAP1 (G) or a heterozygous dominant allele of dArk (H), the Drosophila APAF1 homolog.
(I) Summary of the genetic interactions with Act88F-Gal4,UAS-dE2f1. The Act88F-Gal4,UAS-dE2f1 recombinant chromosome was crossed at 25 °C to various transgenic and mutant alleles and phenotypes analyzed in transheterozygous progeny. Modification of the dE2F1 phenotype was compared relative to control chromosome w1118. See Materials and Methods for mutant alleles used in this study.
Figure 4.
Aac11 Loss-of-Function Specifically Enhances Multiple dE2f1-Dependent Phenotypes In Vivo
Through genetic screening, the P-element insertion l(2)k06710 was identified as a strong enhancer of the Act88F-Gal4;UAS-dE2f1 apoptotic wing phenotype (A). The arrow in (A) indicates enhanced ventral wing curvature and blistering in the l(2)k06710-enhanced dE2f1-dependent wing phenotype. In secondary screening, l(2)k06710 was found to enhance multiple dE2f1-dependent phenotypes in different tissues including a nos-Gal4;UAS-dE2f1 notum bristle phenotype (B–D) and a GMR-Gal4,UAS-dE2f1,UAS-dDp rough eye phenotype (E–H). The l(2)k06710 insertion enhanced bristle degeneration (compare arrows in [B] and [C]) and bristle stubble (arrowheads in [C]) induced by dE2f1 without inducing bristle phenotypes under heterozygous conditions alone (D). Expression of a UAS-Aac11 RNAi allele also enhanced the dE2f1-induced rough eye resulting in a blackened phenotype (I–K). Expression of the UAS-Aac11 RNAi allele under engrailed results in dominant posterior wing blistering which is enhanced by the l(2)k06710 P-insertion (L and M).
Figure 5.
Aac11 Is a Member of the Api5 Protein Family
(A) P-element l(2)k06710 insertion in the Drosophila Aac11 gene. The genomic locus of Aac11 on Chromosome 2L depicting P-insertion l(2)k06710 (+187 nucleotides) in Aac11 exon1. This transposon failed to complement Df(2L)H20 covering region 36C9 but not the nearby deficiency Df(2L)VA18.
(B) Phylogenetic tree of Api5 family proteins. Nine Homologene-annotated homolog sequences (NCBI) were aligned using MegAlign PhyloTree (DNAStar software, http://dnastar.com), using a Clustal method with PAM250 residue weight table. Additional species expressed-sequence tags are present but not included here.
(C) Schematic of conserved Api5 protein domains.
(D) ClustalW multiple alignment of human, mouse, frog, fly, mosquito, and plant Api5 homologs with gray and black depicting amino acid similarity and identity, respectively.
Table 1.
Genetic Interaction of l(2)k06710 in Secondary Screen Phenotypes
Figure 6.
RNAi of Aac11 Enhances dE2F1-Induced Apoptosis and Is Synthetic-Lethal with RBF1 RNAi
(A–D) Transfection of dE2F1 in Drosophila SL2 cells induces cell death as determined by co-transfected GFP loss.
(E) Quantitative measurements of GFP in dE2F1 transfections demonstrated significant GFP loss from dE2F1 that could be rescued by either RBF1 or p35 cotransfection (*p < 0.05 by t-test). Transfection of dE2F1 induced (F) apoptotic chromatin condensation in DAPI-stained nuclei in GFP-positive cells, (G) caspase-3 activation, and (H) caspase-9 activation 48 h after transfection.
(I) RBF1 or Aac11 RNAi significantly enhanced dE2F1-dependent apoptosis (p < 0.01 by t-test). Cell survival was determined by GFP assay 48 h after transfection.
(J) Aac11 RNAi does not affect dE2F1 transcriptional activation of the Drosophila PCNA promoter.
(K) Aac11 depletion does not alter cell cycle profiles in SL2 cells as determined by flow-cytometry analysis (div = days in vitro after RNAi).
(L) Aac11 RNAi is synthetic lethal with RBF1 RNAi under conditions of low-serum stress. Cells were treated with dsRNA in serum-free media and split into media with and without serum, and cell survival was determined 3 d after RNAi.
Figure 7.
Human Api5 Specifically Abrogates E2F1-Dependent Apoptosis without Generally Affecting E2F1-Dependent Transcription
(A–D) API5 stably expressing Saos-2 cells were generated with a Tet-inducible E2F1 transgene in the background. Following E2F1 induction by Tet treatment, the parental cells undergo rapid widespread apoptosis; however, the Api5-expressing cells are highly resistant to E2F1-induced cell death.
(E) Api5-expressing cells survive and proliferate even following high and sustained levels of E2F1 expression. Cells were grown for 6 d after Tet re-dosing every other day.
(F and G) Api5 reduces the levels of E2F1-mediated caspase-3 and PARP cleavage in both stable and Tet-inducible Api5 Saos-2 cells.
(H) Api5 expression does not inhibit the E2F1-mediated induction of target genes CycE and p14ARF in Saos-2 cells.
(I) Api5 expression blocks death induced by E2F1 (+T) but not by treatment with the DNA-damaging agent camptothecin (CPT) as compared to DMSO vehicle control (Veh). Saos-2 cell survival was assayed at 48 h by MTT.
Figure 8.
Depletion of Human Api5 in p16INK4a-Deficient Squamous Cell Carcinoma Cells Results in Reduced Survival versus Normal Human Fibroblasts
(A) RNAi-mediated depletion of transfected Api5. Various shRNA constructs were tested for their ability to deplete FLAG-Api5 transfected U2OS cells. The AB, CD, and EF (but not Scramble or AB-L) cotransfected constructs strongly depleted Api5 expression after 3 d as determined by anti-FLAG immunoblot.
(B) RNAi-mediated depletion of endogenous Api5 protein. JHU-029 cells were infected with lentiviral encoding scramble or API5-AB shRNAs and selected with puromycin for 3 d. Expression of endogenous Api5 was determined by immunoblotting with affinity-purified anti-Api5 polyclonal antibody (G3162).
(C and D) Endogenous Api5 expression in JHU-029 cells is nuclear and excluded from the nucleolus. JHU-029 cells were stained with the G3162 polyclonal antibody after 4% paraformaldehyde fixation.
(E) Api5 RNAi reduces survival of JHU-029 tumor cells as compared to normal human fibroblasts. JHU-029 cells, as well as normal human WI38 diploid fibroblasts, were infected with lentiviral empty vector or vectors encoding API5-AB shRNAs for 24 h and plated onto culture plates in 10% fetal calf serum–containing media. Cell survival was determined by MTT assay at the indicated days post–lentiviral infection. Api5 and control actin expression was determined from equally loaded protein from day 3 lysates.