Bicaudal Is a Conserved Substrate for Drosophila and Mammalian Caspases and Is Essential for Cell Survival

Members of the caspase family of cysteine proteases coordinate cell death through restricted proteolysis of diverse protein substrates and play a conserved role in apoptosis from nematodes to man. However, while numerous substrates for the mammalian cell death-associated caspases have now been described, few caspase substrates have been identified in other organisms. Here, we have utilized a proteomics-based approach to identify proteins that are cleaved by caspases during apoptosis in Drosophila D-Mel2 cells, a subline of the Schneider S2 cell line. This approach identified multiple novel substrates for the fly caspases and revealed that bicaudal/βNAC is a conserved substrate for Drosophila and mammalian caspases. RNAi-mediated silencing of bicaudal expression in Drosophila D-Mel2 cells resulted in a block to proliferation, followed by spontaneous apoptosis. Similarly, silencing of expression of the mammalian bicaudal homologue, βNAC, in HeLa, HEK293T, MCF-7 and MRC5 cells also resulted in spontaneous apoptosis. These data suggest that bicaudal/βNAC is essential for cell survival and is a conserved target of caspases from flies to man.


Introduction
Members of the caspase family of cysteine proteases are involved in coordinating the terminal events of apoptosis in organisms as divergent as nematodes and mammals [1][2][3][4]. Caspases typically exist as dormant proenzymes in healthy cells and become activated at the onset of apoptosis through recruitment to activation scaffolds or membrane receptor complexes [5][6][7][8]. Upon activation, caspases target a subset of the proteome for restricted proteolysis and this results in a stereotypical series of events involving nuclear condensation and fragmentation, extensive plasma membrane blebbing, the appearance of ligands for phagocytic cells on the external leaflet of the plasma membrane, and in many cases the collapse of the cell into small fragments; presumably to facilitate removal of the dying cell [9].
While numerous substrates for human and mouse caspases have now been identified [9,10], it is not clear whether the same subset of proteins is targeted for proteolysis by caspases in other organisms. For example, while the nematode worm Caenorhabditis elegans was instrumental in the initial discovery of a role for caspases in cell death control [11,12], substrates for the major worm caspase, CED-3, have only recently been reported [13]. The fly has also proved to be a very instructive model for the study of the role of caspases and their regulators in developmental cell death [14][15][16][17], however, it remains unclear precisely how caspases coordinate apoptosis in this organism.
Because the phenotype of cells undergoing apoptosis in flies and mammals share remarkable similarities [18], it is likely that at least overlapping cohorts of caspase substrates undergo apoptosisassociated proteolysis in both phyla. To date, seven caspases have been identified in Drosophila and, of these, Dronc and DrICE appear to play particularly significant roles in the coordination of programmed cell death in this organism [19]. Dronc is the only CARD-carrying caspase in the fly and can associate with the adaptor molecule, Ark, to form a multi-subunit apoptosome complex in response to developmental triggers of apoptosis as well as toxic stimuli [20,21]. Upon activation within the apoptosome, Dronc can promote activation of other caspases such as DrICE and DCP-1 [22,23], thereby initiating a protease cascade. DrICE appears to represent the major effector caspase in the fly and may be the functional equivalent of mammalian caspase-3 in this organism [22].
At present, three substrates for Drosophila caspases, DIAP1, Lamin DmO, and the Drosophila Apaf-1 homologue, ARK, have been identified [22,24,25]. To gain further insight into how caspases contribute to programmed cell death in Drosophila, and to identify caspase substrates that are conserved between species, we have conducted a proteomics-based screen to search for proteins that undergo caspase-dependent proteolysis during apoptosis of Drosophila D-Mel2 cells. This analysis resulted in the identification of 14 proteins that underwent caspase-dependent alterations to their relative mobilities on two-dimensional (2D) SDS-PAGE gels. We have cloned two of these substrates and confirm that they are efficiently cleaved by the Drosophila caspases DrICE and DCP-1. We also show that the human homologue of one of these substrates, bicaudal, is also cleaved by caspases during apoptosis of mammalian cells. Bicaudal and its human homologue, bNAC, appear to be essential for cell proliferation and cell survival as RNAi-mediated silencing of the expression of these proteins resulted in a block to cell division, followed by spontaneous apoptosis. This suggests that, as well as targeting substrate proteins that contribute to the ordered destruction of the cell, caspases also inactivate key proteins such as bNAC that are essential for cell survival.

Results
Apoptosis in Drosophila D-Mel2 cells shares many features in common with apoptosis in mammalian cells and is caspase-dependent To search for substrates for the fly caspases, we used a subline of the Schneider (S2) cell line, D-Mel2, that can be induced to undergo apoptosis in response to many of the same stimuli that promote apoptosis in mammalian cells. D-Mel2 cells exposed to staurosporine, actinomycin D, etoposide or cycloheximide died rapidly (Fig. 1A), and displayed typical apoptotic features such as cell shrinkage, extensive plasma membrane blebbing and cellular collapse into numerous apoptotic bodies ( Fig. 1B and C). Robust caspase activation, as assessed by hydrolysis of the synthetic caspase substrate DEVD-AFC, was also readily detected in lysates generated from apoptotic D-Mel2 cells but not from control cell lysates (Fig. 1D). All of these events were caspase dependent and could be abolished with the poly-caspase inhibitor z-VAD-fmk ( Fig. 1E and data not shown).

Two-dimensional SDS-PAGE analysis of caspasedependent changes to the D-Mel2 cell proteome during apoptosis
To identify caspase-dependent alterations to the D-Mel2 proteome, we compared two-dimensional protein spot patterns from D-Mel2 cells induced to undergo apoptosis in the presence or absence of z-VAD-fmk (Fig. 2). Approximately 2,000 of the most abundant soluble proteins expressed in D-Mel2 cells could be resolved using 2D gels. We identified numerous alterations to the proteome of apoptotic cells, all of which were abolished in the presence of z-VAD-fmk suggesting that the observed changes were caspase-dependent protein modifications (Fig. 2).
Proteins that underwent caspase-dependent proteolysis during apoptosis of D-Mel2 cells were identified by MALDI-TOF mass spectrometry ( Table 1). The 14 identified substrates include three chaperone proteins, three proteins involved in translation, and proteins involved in a variety of other functions. Two of the putative caspase substrates, bicaudal and stubarista, were chosen for more detailed analysis.

Bicaudal and stubarista are bona fide substrates for Drosophila caspases
To confirm that bicaudal and stubarista were genuine substrates for Drosophila caspases, we generated polyclonal antibodies against these proteins and used these to probe cell lysates prepared from healthy versus apoptotic D-Mel2 cells. As illustrated in Fig. 3A, bicaudal underwent rapid proteolysis upon induction of apoptosis and was cleaved with similar kinetics to lamin DmO. Proteolysis of stubarista was less efficient, relative to bicaudal and lamin DmO, being detected later and only a fraction of the total cellular complement of stubarista appeared to be cleaved under the conditions employed (Fig. 3A). Proteolysis of all of these substrates was completely inhibited by z-VAD-fmk, confirming that these proteolytic events were caspase-dependent (Fig. 3A). Similar results were also observed when in vitro transcribed and translated forms of bicaudal and stubarista were incubated in cell-free extracts generated from apoptotic versus untreated D-Mel2 cells (Fig. 3B).
Bicaudal is cleaved at a single caspase cleavage motif at Asp141 Bicaudal is a homologue of human bNAC, involved in binding to nascent chains as they emerge from the ribosome, and mutations in both Drosophila and murine bNAC lead to embryonic lethality [26][27][28]. In addition, mutations in the C. elegans homologue of bicaudal, ICD-1, are associated with elevated levels of programmed cell death during embryonic development [29]. We therefore decided to explore this substrate in further detail. We mapped the caspase cleavage site within bicaudal by in vitro transcribing and translating the protein and incubating with recombinant DCP-1 and DrICE, both of which have been implicated as effector caspases in the fly [22][23]. As shown in Fig. 3C, both DCP-1 and DrICE generated a bicaudal fragment identical in size to the cleavage product seen when 35 S-labeled bicaudal was incubated with lysates generated from apoptotic D-Mel2 cells. We then incubated recombinant bacteriallyexpressed bicaudal with DCP-1 and analysed the resulting cleavage products by mass spectrometry. The peptide coverage of the major fragment of bicaudal indicated that the caspase cleavage site lay within the C-terminus of the protein, within the last 46 amino acids (Fig. 3D). Using this information, we inspected the bicaudal protein sequence and mutated three candidate aspartic acids at positions 139, 140 and 141. As Fig. 3 E illustrates, only one of these mutants, D141A, fully resisted proteolysis by recombinant DrICE and similar results were found using DCP-1 (data not shown). To further confirm this as the site of proteolysis during apoptosis, we then incubated wildtype 35 S-labeled bicaudal, alongside the D141A bicaudal mutant, in extracts generated from apoptotic D-Mel2 cells. This experiment confirmed D141 as the site cleaved by caspases during apoptosis within the motif 138 GDDD 141 (Fig. 3F). This is a somewhat unusual motif for executioner caspases, as these typically prefer a DXXD consensus. However, little information concerning optimal cleavage site motifs for Drosophila caspases is available for useful comparisons to be made.

RNAi-mediated knockdown of bicaudal expression blocks proliferation and results in spontaneous apoptosis
To explore the importance of Bicaudal for cell viability, we silenced expression of the bic transcript using double stranded RNA targeted against this gene. As controls, we also silenced expression of dronc and diap1, both of which are well-established regulators of programmed cell death in the fly. RNAi-mediated silencing of all of these genes was confirmed by RT-PCR analysis (Fig. 4A), and knockdown of bicaudal protein was also confirmed by Western blot analysis (Fig. 4B). Previous studies have shown that loss of diap1 expression results in spontaneous apoptosis in the fly and in cell culture [30,31], and we confirmed this observation ( Fig. 4C and D). In contrast, silencing of dronc expression had no effect on cell viability or rates of proliferation ( Fig. 4C and D), but did block apoptosis induced by exposure to cytotoxic drug treatment (Fig. 4C). However, bic knockdown resulted in a profound block to proliferation and cells also underwent spontaneous apoptosis but with slower kinetics to that seen due to loss of diap1 ( Fig. 4C-E). Furthermore, cells lacking bicaudal were also sensitized to apoptosis triggered by actinomycin D whereas knockdown of dronc expression protected from this stimulus, as expected (Fig. 4C). These data suggest that bicaudal expression is important for the sustained viability of Drosophila cells and that caspase-dependent proteolysis of bicaudal during apoptosis may contribute to cellular demise.

bNAC, a human homologue of bicaudal, is a substrate for caspases and granzyme B
To further explore the significance of bicaudal for apoptosis, we then explored whether the human homologue of this protein, bNAC, the b-subunit of the nascent polypeptide-associated complex, was also cleaved by caspases during apoptosis. A previous study has indicated that bNAC may be cleaved in a caspase-dependent manner during apoptosis [32], however, the functional consequences of this was not examined. An alignment between bicaudal and bNAC revealed very extensive sequence identity between these two proteins of approximately 57% (Fig. 5A). We then explored whether bNAC underwent caspasedependent proteolysis using a cell-free system based upon homogenates generated from human Jurkat cells [33]. Caspase activation in this system can be readily initiated either through introduction of the apoptosome co-factors, cytochrome c and dATP, or by introduction of the CTL/NK protease, granzyme B [34]. As, shown in Fig. 5B and C, bNAC was very efficiently cleaved upon activation of caspases by cytochrome c or granzyme B. Interestingly, granzyme B also produced an additional bNAC cleavage product, suggesting that the latter may also be a direct substrate for this granzyme. This was confirmed upon incubation of bNAC with granzyme B in the presence of z-VAD-fmk (to block caspase activity) where only the faster mobility bNAC cleavage product, seen only in the presence of granzyme B, was now observed (Fig. 5C).

Mapping of the caspase cleavage site within bNAC
The aspartic acid within bicaudal that is targeted for cleavage by Drosophila caspases is altered to a glutamic acid at the equivalent position within human bNAC (Fig. 5A). Therefore we mutated alternative potential caspase cleavage motifs and found that bNAC is instead cleaved at Asp 175. Consistent with this, a D175A point mutant of bNAC completely resisted caspase-dependent proteolysis under conditions where the wild-type bNAC protein was completely cleaved by caspases (Fig. 5D).

Silencing of bNAC expression results in proliferative arrest and spontaneous apoptosis of mammalian cells
To explore whether bNAC expression was also required for survival of mammalian cells, we silenced expression of this protein using siRNA directed against the bNAC coding sequence. To minimize the possibility of non-specific off-target effects, we designed three different siRNAs against the bNAC mRNA. As controls, we employed siRNAs targeted against CASP-9 and P84, an irrelevant gene in this context. siRNA-mediated knockdown of After 7 h, cell lysates were prepared and 500 mg of protein from each reaction was analysed by two-dimensional SDS-PAGE followed by silver-staining. Representative 2D gels are shown with caspaseinduced alterations annotated. (B) Enlarged regions of silver-stained two-dimensional SDS-PAGE gels from control or actinomycin D-treated cell lysates are shown. Proteins reproducibly altered upon caspase activation are indicated. Protein identifications were obtained by MALDI-TOF mass spectrometry. Results are representative of at least three independent experiments. Blue squares represent protein spots that disappear, in a caspasedependent manner, upon induction of apoptosis. Red circles represent new protein spots that appear in apoptotic cells. Spot numbers on the gel in Figure 2B refer to the protein numbers in Table 1. doi:10.1371/journal.pone.0005055.g002 bNAC expression was confirmed by Western blot analysis (Fig. 6A). As Fig. 6B illustrates, knockdown of bNAC expression in MCF7 cells resulted in a dramatic reduction in cell numbers over 72 h and a substantial increase in apoptosis during this time course. Essentially identical results were observed in untransformed MRC-5 human fibroblasts (Fig. 6C), HEK293T cells ( Fig. 6D and E) and HeLa cells (data not shown). Collectively, these data indicated that bNAC serves an essential function in diverse cell types and proteolysis of this protein by caspases may contribute to the irreversible termination of cell viability during apoptosis.

Discussion
Here we have identified several substrates for the Drosophila cell death-associated caspases. Previously, few substrates for caspases activated during cell death in the fly have been reported and the substrates reported herein provide a basis for further investigations exploring the functional significance of those proteolytic events. Furthermore, three of the substrates identified in this study, bicaudal, a heterogeneous nuclear ribonucleoprotein, and lamin DmO, are homologues of proteins known to be substrates of caspases in human cells [9,10,32]. Although it is commonly assumed that caspases promote the characteristic features of apoptosis by cleaving overlapping sets of substrates in divergent organisms, very few conserved caspase substrates have been identified to date.
One such substrate, the b-subunit of the nascent polypeptideassociated complex, bicaudal/bNAC, was efficiently cleaved by both Drosophila and human caspases. Interestingly, bNAC has also previously been implicated as a negative regulator of programmed cell death in the nematode C. elegans and ablation of this gene product results in massive and unscheduled apoptosis in developing worm embryos [29]. NAC functions to bind short nascent polypeptides as they emerge from the ribosome [26]. The latter event prevents inappropriate interactions with cellular proteins and non-specific binding by the signal recognition particle and consequent targeting to the ER [26,35]. NAC also prevents the targeting of non-translating ribosomes to the ER [26,35]. This fundamental role of NAC is reflected in the catastrophic phenotype of null mutations affecting the bNAC-coding sequence gene in a range of species. Loss of bNAC in developing mice leads to post-implantation lethality and mutation of Drosophila bicaudal promotes developmental arrest, which is associated with duplication of the posterior embryonic regions in the place of the anterior embryonic segments [27,28]. As mentioned earlier, RNAimediated silencing of the C. elegans bNAC homologue, ICD-1, also results in developmental arrest associated with massive cell death [29]. Thus, the disablement of bNAC function through caspase-dependent proteolysis may contribute substantially to cellular demise.
In line with this, here we have shown that interference with bNAC expression levels in Drosophila or human cell lines leads to proliferative arrest, followed by cell death. However, although silencing of bNAC expression may replicate one possible consequence of caspase-dependent inactivation of this protein, it is also possible that caspase-mediated proteolysis of bNAC could generate a cleaved form of this protein with altered function (such as dominant-negative inhibition of native bNAC) that gene silencing may not faithfully reproduce. It is also important to note that expression of non-cleavable b-NAC in human cells did not give rise to any obvious change in the apoptotic phenotype (data not shown). This suggests that caspase-dependent proteolysis of b-NAC does not contribute to any obvious morphological feature of apoptosis. However, we propose that at least two functional categories of caspase substrates may be important in apoptosis. One category of caspase substrates may contribute to specific morphological features of this mode of cell death. Examples of such substrates include ICAD, the inactivation of which results in activation of the CAD endonuclease that contributes to apoptosis-associated DNA hydrolysis [36,37]. Similarly, proteolysis of ROCK I is thought to result in the extensive plasma membrane blebbing and nuclear fragmentation that typifies apoptotic cells [38][39][40]. On the other hand, it is very likely that other proteins are cleaved by caspases in order to terminate the life of the cell, thereby contributing to the irreversibility of this process. We suggest that bicaudal/bNAC may represent a caspase substrate belonging to the latter category.
In summary, here we have provided the first snapshot of the proteins that become targeted for proteolysis by caspases during programmed cell death in a Drosophila cell line. Clearly, numerous additional substrates for the fly caspases are likely to exist but will require techniques that penetrate deeper into the fly proteome to identify these. However, our study provides direct evidence that caspases most likely coordinate apoptosis in divergent organisms through cleaving overlapping cohorts of cellular substrates.

Materials
Antibodies specific to Drosophila Lamin DmO and Actin were purchased from the Developmental Studies Hybridoma Bank (USA). Polyclonal antibodies to Drosophila Bicaudal and Stubarista were generated by repeated immunization of rabbits with recombinant polyhistidine-tagged forms of these proteins. Recombinant polyhistidine-tagged Bicaudal and Stubarista were expressed and purified from bacteria (data not shown). The peptides, z-VADfmk, Ac-DEVD-AFC, Ac-LEHD-AFC, and Ac-YVAD-AFC were all purchased from Bachem (UK). 17 cm IPG strips (pH 3-6 and pH 5-8), easymelt agarose and Bio-Lyte ampholytes TM were purchased from Biorad (UK). Unless otherwise indicated, all other reagents were purchased from Sigma (Ireland) Ltd.

Cell culture
The S2 subline, D-Mel-2, are a subline of Schneider S2 insect cells isolated from late stage Drosophilia melanogaster embryos [41], and were cultured at 28uC/0.5% CO 2 in serum-free medium (Invitrogen, UK). HeLa and MRC-5 cells were cultured at 37uC in RPMI 1640 containing 5% foetal calf serum and HEK293T cells were cultured in DMEM containing 10% foetal calf serum.  35 S-methionine-labeled bicaudal was incubated either with actinomycin D-treated D-Mel2 cell lysate as above, or with recombinant DCP-1 and drICE, and samples taken at the indicated times followed by SDS-PAGE/fluorography analysis. (D) Peptide coverage of the caspase-cleaved bicaudal protein sequence obtained by mass spectrometry is underlined. (E) Individual aspartic acid residues in the bicaudal sequence, D139, D140 and D141 were converted by site-directed mutagenesis to alanine (aa139 and aa141) or glutamic acid (aa140) residues. 35 S-methionine-labelled wild type and mutant proteins were incubated in buffer or with recombinant drICE for 120 min, before analysis by SDS-PAGE and fluorography. (F) 35 S-methionine-labeled wild type and D141A mutant bicaudal proteins were incubated in actinomycin D-treated cell lysate as above and samples taken at the indicated times for analysis by SDS-PAGE/fluorography. doi:10.1371/journal.pone.0005055.g003

Flow cytometry
To quantify levels of apoptosis-associated cell shrinkage and fragmentation, forward scatter (FSC) and side scatter (SSC) of cell populations were measured by flow cytometery (FACSCalibur, Becton Dickinson, CA).

Preparation of cell-free extracts
Cell-free extracts of Drosophila D-Mel-2 cells were prepared by allowing cells to swell for 20 minutes on ice in CEB containing 250 mM sucrose (CEB: 20 mM HEPES-KOH, pH7.5, 10 mM Kcl, 1.5 mM MgCl 2 , 1 mM EDTA, 1 mM EGTA, 1 mM DTT, 100 mM PMSF, 10 mg/ml leupeptin, 2 mg/ml aprotonin). Cells were then lysed by homogenization with ,20-30 strokes of a Btype pestle and crude extracts centrifuged for 15 mins at 15,000 g to remove nuclei, unbroken cells and other debris. Extracts from human cells were prepared as described previously [33].

Two-dimensional (2D) gel electrophoresis
For 2D gel electrophoresis, cells were lysed in 2D sample buffer (8M Urea, 4% CHAPS, 100 mM DTT, 0.05% SDS, 0.5% ampholyte 3-10 and a trace of bromophenol blue) and were rehydrated into 17cm IPG strips (BioRad). Passive sample rehydration into IPG strips was performed at room temperature overnight. Isoelectric point focussing (IEF) was performed in a BioRad Protean IEF Cell under the following conditions: (1) linear voltage ramp to 500 V over 1 hr, (2) 5 hr at 500 V, (3) linear voltage ramp to 3500V over 5 hr and (4) 12 hr at 3500 V. Following IEF, the IPG strips were reduced and alkylated with 2% DTT and 2.5% IAA, respectively, in 5 min incubations in an equilibration buffer containing 6M Urea, 375 mM Tris HCl, pH 8.8, 2% SDS and 20% Glycerol. Strips were then mounted on 12% SDS-PAGE gels using easymelt agarose (Biorad, UK) and electrophoresed at 37.5 mA per gel in a Biorad Protean IIxi  electrophoresis cell (Biorad, UK). 2D gels were either Coomassieblue stained or stained using a Mass Spectrometry-compatible silver staining protocol that is based on a modification of the EMBL silver staining protocol [42].
In-gel protein digestion and protein identification by MALDI-TOF mass-spectrometry Protein spots were manually excised from 2D gels. Gel pieces were incubated at room temperature on a shaking platform in oxidation buffer (15 mM K 3 Fe(CN) 6 , 50 mM Na 2 S 2 O 3 ) until the spots were completely destained. Gel pieces were then washed 5 times (5-10 min per wash) in 50% methanol/10% acetic acid. Samples were then incubated in 50 mM NH 4 HCO 3 for 5 minutes, prior to dehydrating in 100% acetonitrile. To further dehydrate the pellets, the acetonitrile was aspirated off and samples were spun in a speed-vac (ThermoSavant) for 5 minutes at room temperature. For trypsin digestion, a 100 mg/ml aliquot of Sequencing grade trypsin (Roche) dissolved in 1 mM HCl was diluted 1:10 in digestion buffer (25 mM NH 4 HCO 3, 0.1 n-octyl b-D-Glucopyranoside). Typically, for low abundance silver stained spots, 2 ml of trypsin solution (20 ng) was pipetted directly unto the desiccated gel piece. After allowing the gel piece to rehydrate for 5 minutes, a further 10 ml of digestion buffer was added and samples were incubated overnight at 37uC. Following trypsin digestion, peptides were extracted twice into 40 ml 66% acetonitrile/0.1% trifluoroacetic acid in a sonicating water bath, followed by lyophilization in a speed-vac at room temperature.
For mass-spectrometric analysis, peptides were solubilized by sonication in 5 ml of 5% formic acid. Digested samples (0.5 to 1 ml) were applied to a Teflon-coated 96-well MALDI target plate (Applied Biosciences, UK), followed by the addition of 0.5 to 1 ml of a 10 mg/ml matrix solution of a-cyano-4-hydroxy-cinnamic acid in 60% acetonitrile/0.1% trifluoroacetic acid. Samples were allowed to air-dry at room temperature before analysis in positive reflectron mode in a Voyager DE Pro MALDI mass spectrometer (Applied Biosciences, UK).

Plasmids and site-directed mutagenesis
Site-directed mutagenesis was carried out using the Quickchange kit (Stratagene). All plasmids were verified by DNA sequencing.
Expression and purification of Drosophila caspases DCP-1 and DrICE were expressed in BL21 cells as polyhistidine-tagged fusion proteins and were purified over Nickel-NTA agarose (Qiagen, UK) followed by elution with 100 mM imidazole. Both proteases demonstrated robust proteolysis of the synthetic peptide substrate DEVD-AFC and their activities were normalized using this peptide.

Fluorimetry assays
Reactions containing recombinant caspases or Drosophila cellfree extract were typically assembled in a final volume of 100 ml. Following incubation at 37uC for 15 min, 2.5 ml samples were diluted to a final volume of 200 ml in WCEB containing 50 mM Ac-DEVD-AFC. Samples were then measured in an automated fluorimeter (Spectrafluor Plus, TECAN, UK) at wavelengths of 430 nm (excitation) and 535 nm (emission). 35 S-methionine-labelled proteins were generated using the TNT kit (Promega) as described previously [33].

RNAi and siRNA-mediated silencing of gene expression
RNAi-mediated silencing of gene expression in Drosophila was achieved using double-stranded RNA transcripts generated from the coding sequences of diap-1, bicaudal and dronc. All dsRNAs were generated using the Megascript kit (Ambion, UK) followed by purification of RNA using ethanol precipitation. For genesilencing, ,15 mg of dsRNAs were transfected into D-Mel2 cells followed by incubation for 72 h. For siRNA-mediated gene silencing, 21 base long oligonucleotides were generated to match the target sequence and were transfected into mammalian cells using oligofectamine (Invitrogen, UK). siRNA oligos were transfected at 200 nM, followed by incubation for 48-96 h.

RNA analysis
To confirm knockdown following transfection with dsRNA, RNA was extracted from cells by lysis in RNeasy lysis buffer (RTL Buffer, Qiagen) with 1% b-mercaptoethanol, followed by addition of an equal volume of 70% ethanol. Lysates were applied to RNeasy columns (Qiagen) and eluted with ddH 2 O. Specific RNAs were then reverse-transcribed to cDNAs using the Omniscript kit (Qiagen) and analysed by electrophoresis through agarose gels.

Image acquisition and analysis
The images presented in Fig. 1B, Fig. 4E and Fig. 6E were taken on an inverted microscope (IX71; Olympus) using a 406 objective, and images were captured using a Colorview II camera (Soft Imaging System) equipped with Analysis image acquisition software. All images were acquired at room temperature. All twodimensional gel images were captured using a digital camera (CoolPix 5000; Nikon) and were imported into Photoshop (Adobe) for uniform brightness and contrast adjustment. All Western blot and fluorogram images were scanned from x-ray film using a scanner (Epson Perfection) and imported into Photoshop for uniform brightness and contrast adjustment.