Main

Programmed cell death is critical to multi-cellular animal development and homeostasis, and misregulation of cell death can lead to disorders including autoimmunity and cancer.1, 2, 3 Three prominent morphological forms of programmed cell death occur during development, including apoptosis, programmed necrosis, and autophagic cell death.4, 5 Autophagic cell death is characterized by the induction of autophagy in dying cells that are degraded and removed without known phagocytic uptake.4 Although our understanding of autophagic cell death has recently increased, how autophagy and other catabolic processes promote cell death and degradation remains poorly understood.

Cells use two general catabolic processes to degrade and recycle cellular components, the ubiquitin proteasome system (UPS) and autophagy. During autophagy, cytoplasmic contents are sequestered by the autophagosome and delivered to the lysosome for degradation. Proteins to be degraded by the UPS are polyubiquitinated and targeted to the proteasome, an ATP-dependent 26S multisubunit protease complex.6 Whereas the UPS is used to degrade short-lived proteins, autophagy is generally thought to degrade long-lived proteins. Although the fundamental processes of autophagy and the UPS are understood, less is known about the function of these processes during programmed cell death.

The UPS has a key role in multiple cellular processes, including cell cycle regulation7 and antigen processing.8 Studies have also suggested a role for the UPS in coordinating cell death.9 Studies in cell lines suggest that caspase-dependent proteasome impairment may facilitate cell death.10, 11 Furthermore, changes in ubiquitination and proteasome composition occur during hawkmoth development in dying intersegmental muscle cells.12, 13, 14, 15 Interestingly, these cells die with autophagic cell death morphology, suggesting a possible role for the UPS in autophagic cell death, but this has not been tested empirically.

Several studies suggest that ectopic proteasome impairment induces cell death in cell lines,16, 17 and drug-induced proteasome impairment in tumor cells results in caspase activation and cell death.18 In addition, proteasome inhibitors are used in cancer therapies.19 The reversible proteasome inhibitor Bortezomib (Velcade, Millennium Parmaceuticals, Cambridge, MA, USA) is used to treat myeloma and lymphoma.20 Evidence in cell lines suggests that proteasome inhibitors lead to cell death by downregulating NF-κB signaling21, 22 and/or by inducing endoplasmic-reticulum stress.23, 24 However, little is known about the effects of proteasome inhibitors on cellular processes, and how proteasome inhibition leads to cell death has not been investigated under physiological conditions in vivo.

Larval salivary gland cells undergo synchronized programmed cell death requiring both caspases and autophagy (Atg) genes during Drosophila development.25 In wild-type Drosophila, the onset of salivary gland cell death occurs 12 h after puparium formation (APF), and DNA fragmentation and caspase activity can be detected at this time. Between 13 and 14 h APF, autophagy is induced and salivary glands become smaller in size, or condensed. By 16 h APF, wild-type salivary glands are completely degraded.

Recent developments in high-throughput mass spectrometry (MS)-based proteomic technologies prompted us to conduct a new proteomic analysis of dying salivary glands in Drosophila. We found that among the groups of proteins identified, UPS components were the top category enriched in wild-type dying salivary glands, suggesting a possible role for the UPS in larval salivary gland cell death. We monitored UPS function in salivary gland cells and found that flux through the UPS is reduced during wild-type salivary gland cell death. Significantly, premature proteasome impairment using an inducible genetic proteasome mutant leads to the onset of early salivary gland condensation (an overall reduction in salivary gland size) and DNA fragmentation, as evidenced by TUNEL staining. Surprisingly, we find that the expression of the caspase inhibitor p35 during proteasome impairment does not suppress early salivary gland condensation. This suggests the involvement of caspase-independent factors in the cellular response to proteasome impairment leading to ectopic salivary gland condensation. To identify factors that function in the response to proteasome impairment during autophagic cell death, we conducted proteomic analyses of salivary glands following ectopic proteasome impairment. We used a genetic approach to screen the top candidates identified in our proteomics data sets for defects in salivary gland degradation. Here, we show that trol, CG11880, a gene of unknown function, and the cop9 signalsome component csn6, are required for salivary gland cell degradation.

Results

UPS components are enriched in dyingDrosophila larval salivary glands

During Drosophila development, autophagic programmed cell death of larval salivary glands is triggered by a rise in the steroid hormone 20-hydroxyecdysone (ecdysone) 12 h after APF, and requires both caspases and autophagy.25, 26, 27 In wild-type animals, salivary glands begin to die 12 h APF, and DNA degradation and caspase activation can be detected at this time in salivary gland cells. By 13–14 h APF, autophagy is induced and salivary glands become smaller in size, or condensed. Thereafter, salivary glands are rapidly degraded and are absent by 16 h APF.

To identify new proteins required for larval salivary gland cell degradation, we isolated proteins from purified larval salivary glands before (6 h APF), during (12 h APF), and after (13.5 h APF) the onset of salivary gland cell death (Supplementary Table S1). In the 6-h samples, we identified an average of 12 487 distinct peptides that mapped to 4157 distinct proteins with >95% confidence (Supplementary Table S2). Similarly, in the 12-h samples, we identified an average of 11 530 distinct peptides that mapped to 3566 distinct proteins, and in the 13.5-h samples, we identified an average of 13 012 distinct peptides that mapped to an average of 4184 distinct proteins with >95% confidence (Supplementary Table S2).

Using this proteomic approach, proteins previously shown to be involved in salivary gland programmed cell death were identified, including the ecdysone receptor component Usp, the ecdysone response protein Eip93F (E93), the nuclear receptor competence factor βFtz-f1, the Drosophila caspase protease Ice (DrIce), and several autophagy proteins (Supplementary Table S3). These data are consistent with published DNA microarray and proteomics studies,28, 29 as well as genetic studies showing that these factors are required for salivary gland cell death.25, 27, 30, 31

We sought to identify groups of proteins that were enriched in dying salivary glands that might suggest the involvement of either a cellular process or a genetic pathway in salivary gland cell death. We used Ingenuity Pathway Analysis to analyze the 6-, 12-, and 13.5-h proteomics data sets, and this approach identified the protein ubiquitination pathway as the top category enriched in dying salivary glands. In two-way comparisons, we identified groups of proteins that increased in detection values between 6 and 13.5 h APF (Supplementary Table S4). The enrichment of a large number of proteasome components and ubiquitination regulators in dying salivary glands suggested a possible role for the UPS in salivary gland cell death (Table 1). These findings suggest several possibilities; UPS components may be enriched in dying salivary glands because the UPS has an increased function during salivary gland cell death. Alternatively, if the UPS has reduced function during cell death, it is possible that dying salivary gland cells attempt to compensate by increasing the levels of UPS components.

Table 1 List of UPS components enriched in dying wild-type Drosophila larval salivary glands
Full size table

UPS function is reduced in dying salivary gland cells

Previous studies in cell lines suggest that inhibition of proteasome function may be part of a normal cell death program.10, 11 To monitor UPS function during Drosophila salivary gland cell death, we used transgenic fly lines containing UPS reporter substrates that consist of fusions between GFP and degron sequences that target proteins for degradation by the UPS.32, 33, 34 During proper UPS function, GFP is targeted to the proteasome for degradation, whereas GFP accumulates during decreased flux through the UPS pathway.

The GFP reporter for flux through the UPS pathway (UAS-GFP-CL1) was expressed specifically in the salivary glands of experimental animals using the salivary gland-specific driver forkhead-GAL4 (fkh-GAL4), whereas a stable GFP reporter (UAS-Ub-MGFP) was expressed in salivary glands of control animals. To monitor UPS function in salivary gland cells, glands were dissected and examined by fluorescence microscopy to assay for GFP fluorescence at stages before (6 h APF), during (12 h APF), and after the onset of salivary gland cell death (13.5 h APF). All of the control animals possess fluorescence in salivary gland cells 6, 12, and 13.5 h APF (Figure 1a, top panels). Salivary glands expressing the UPS flux reporter GFP-CL1 exhibit low fluorescence at 6 and 12 h APF (Figure 1a, bottom panels). However, GFP fluorescence increases dramatically in salivary glands by 13.5 h APF (Figure 1a, bottom panels), suggesting that flux through the UPS pathway is decreased in dying salivary glands, perhaps due to a decrease in UPS function. If UPS function is decreased in salivary gland cells during cell death, we would expect ubiquitinated proteins to accumulate in these cells. We analyzed wild-type salivary gland protein extracts by western blot with ubiquitin antibody and found that ubiquitinated proteins increase in salivary gland extracts between 6 and 13.5 h APF (Figure 1b). These results suggest that the flux through the UPS pathway decreases, and the function of the UPS is reduced, during salivary gland cell death.

Figure 1
figure 1

UPS function is reduced during wild-type salivary gland developmental cell death. (a) Control animals (fkh-GAL4/+; UAS-Ub-M-GFP/+) expressing a stable GFP reporter in salivary glands exhibit GFP fluorescence in salivary glands 6, 12, and 13.5 h AFP, n=11. Pupae with salivary gland-specific expression of a GFP reporter for UPS impairment (fkh-GAL4/+; UAS-GFP-CL1/+) exhibit low GFP fluorescence in salivary glands at 6 and 12 h AFP, but intense GFP fluorescence in salivary glands 13.5 h AFP, n=29. (b) Protein extracts from wild-type salivary glands 6 and 13.5 h AFP were analyzed by western blotting with an anti-ubiquitin antibody

Full size image

Ectopic proteasome impairment leads to early salivary gland condensation and DNA fragmentation that is not suppressed by caspase inhibitor p35 expression

Proteasome impairment leads to caspase activation and cell death in cell lines.18 To test whether proteasome impairment leads to early cell death in vivo, we expressed an inducible dominant temperature-sensitive mutant allele of the β2 subunit of the 20S proteasome (UAS-Dts72c)35 specifically in salivary gland cells, using the salivary gland-specific GAL4 driver fkh-GAL4. To impair proteasome function before the onset of salivary gland cell death, animals expressing Dts72c in salivary gland cells were shifted from permissive to restrictive temperature at puparium formation (0 h) to 8 h APF. In wild-type animals, salivary gland cell death commences 12 h APF. As a normal aspect of salivary gland cell death, we observe salivary gland condensation, or a decrease in salivary gland size, at 13–14 h APF. Salivary gland condensation can be measured as a decrease in salivary gland area in histological sections. We found that salivary glands of control animals lacking the GAL4 driver (UAS-Dts7/+) are intact and have a normal size and morphology at 8 h APF. In contrast, we observed that salivary glands of experimental animals, in which the proteasome is ectopically impaired specifically in salivary gland cells (fkh-GAL4/+; UAS-Dts72c/+), are smaller in size, or more condensed, than those of control animals (Figure 2a). To quantify salivary gland size, we measured the area of salivary gland material in histological sections and found a significant difference in salivary gland area between control animals and those with ectopic proteasome impairment in salivary gland cells (Figure 2e). These data suggest that in vivo ectopic proteasome impairment in salivary gland cells leads to the early onset of salivary condensation.

Figure 2
figure 2

Ectopic proteasome impairment leads to early salivary gland condensation that is not suppressed by caspase inhibitor p35 expression. Pupae were shifted from permissive (25°C) to restrictive (28°C) temperature at puparium formation (0 h), fixed 8 h AFP, and processed for histology. (a) Control animals lacking the GAL4 driver (+/UAS-Dts72c), n=9, and those with salivary gland-specific expression of Dts72c to genetically impair proteasome function (fkh-GAL4/+; UAS-Dts72c/+), n=13, were analyzed by histology 8 h after temperature shift for salivary gland material (outlined in red). Gland area is shown in (e). (b) Control animals (+/UAS-Dts72c), n=10, and those with salivary gland-specific expression of Dts72c to impair proteasome function (fkh-GAL4/+; UAS-Dts72c/+), n=12, were in situ TUNEL-labeled to assay for DNA fragmentation. Salivary glands (sg) of control animals lacking the GAL4 driver (+/UAS-Dts72c) do not possess fragmented DNA, as indicated by light TUNEL-negative salivary gland nuclei (arrows), whereas internal control midgut (mg) nuclei possess fragmented DNA (arrowheads), as indicated by dark TUNEL-positive staining. Salivary glands of experimental animals (fkh-GAL4/+; UAS-Dts72c/+) contain fragmented DNA (arrowheads), as indicated by dark TUNEL-positive staining, following UPS impairment. Cns, central nervous system; mg, midgut; Sg, salivary glands. (c) Control animals lacking the GAL4 driver (UAS-p35/+), n=9, and those expressing p35 in salivary glands (fkh-GAL4/+; UAS-p35/+), n=18, were analyzed for salivary gland material (outlined in red). Gland area is shown in (e). (d) Control animals lacking the GAL4 driver (UAS-p35/+; UAS-Dts72c/+), n=12, and those expressing p35 in the salivary glands (fkh-GAL4/+; UAS-p35/+; and UAS-Dts72c/+), n=15, were analyzed for salivary gland material (outlined in red). Gland area is shown in (e). (e) For all slides of genotypes shown in (a, b, and d), the largest salivary gland section per slide was outlined for measurement, and salivary gland area was measured using the Zeiss AxiovisionLE software. Red outlines designate the area measured. Error bars represent S.E.

Full size image

We next tested whether ectopic proteasome impairment in salivary gland cells leads to early caspase activation in vivo. DNA fragmentation is a caspase-dependent process in salivary glands.26 We used the TUNEL assay to detect DNA fragmentation in response to ectopic proteasome impairment. Control animals and those expressing Dts72c in salivary gland cells were shifted from permissive to restrictive temperature at puparium formation (0 h) to 8 h APF, and histological sections were TUNEL stained to assay for DNA fragmentation. We found that whereas the nearby nuclei of the larval midgut of control pupae stain positively for TUNEL at this stage, their salivary gland cells do not (Figure 2b). By contrast, salivary gland cells of pupae expressing Dts72c contain fragmented DNA as indicated by dark TUNEL-positive salivary gland nuclei (Figure 2b). These data suggest that ectopic proteasome impairment in salivary gland cells leads to early DNA fragmentation in vivo.

To test whether the early reduction in salivary gland size upon ectopic proteasome impairment is caspase dependent, we co-expressed the caspase inhibitor p35 with Dts72c in salivary glands. We found that salivary glands of pupae expressing p35 alone are not significantly different in size compared with control salivary glands (Figures 2c and e). Surprisingly, however, we found that salivary glands of pupae co-expressing p35 and Dts72c still undergo significant early condensation (Figures 2d and e). These data suggest that although DNA becomes fragmented following ectopic proteasome impairment in salivary gland cells, the early condensation of salivary glands following proteasome impairment may not be caspase-dependent.

Identification of proteins altered following ectopic proteasome impairment that are required for salivary gland degradation

Drugs targeting the proteasome are used in anticancer therapies.19, 20 To better understand cellular responses to altered proteasome function, we conducted a proteomics study of salivary glands following ectopic proteasome impairment. At least three independent biological replicates for each of the control and proteasome impaired salivary gland samples were analyzed using high-throughput shotgun proteomics (Supplementary Table S1). In the proteasome impairment samples (fkh-GAL4/+; UAS-Dts72c/+), an average of 12 601 distinct peptides that mapped to an average of 4184 distinct proteins were identified (Supplementary Table S2). In control samples that lack the GAL4 driver (+/UAS-Dts72c), an average of 12 094 distinct peptides that mapped to an average of 3922 distinct proteins were identified (Supplementary Table S2). Proteins identified in the proteasome impairment and control samples are listed in Supplementary Table S5. An additional fkh-GAL4/+ control was collected and analyzed by proteomics to control for proteins that may be enriched in response to GAL4 expression (data not shown).

We sought to identify genes that are required for salivary gland cell degradation. We queried our proteomics data for proteins that were enriched (indicated by positive ratio2 values) in (1) a two-way comparison between the control (+/UAS-Dts72c) and ectopic proteasome impairment (fkh-GAL4/+; UAS-Dts72c/+) data sets and also in (2) a two-way comparison between the 6- and 13.5-h APF wild-type dying salivary gland data sets (Supplementary Table S6). Of these, proteins detected in the proteasome impairment sample with P-values ≤0.002 that were also not previously shown to function in salivary gland degradation were considered top candidates for genetic analyses. From this list of candidate proteins, eight were identified with available transgenic RNAi lines (CG11880, CG3074, CG9336, CG9917, CG9977, Dbi, Ect3, and Trol), and these were all screened for defects in salivary gland cell degradation.

To knockdown genes specifically in salivary gland cells, transgenic fly lines containing UAS-RNAi constructs were crossed to those expressing fkh-GAL4. Progeny expressing RNAi in salivary glands were staged to 24 h APF; a stage that is 8 h after salivary gland material is cleared in wild-type Drosophila. Of the eight genes tested, two had defects in salivary gland degradation. We found that compared with control animals, those expressing RNAi against the gene encoding Trol possessed persistent salivary gland fragments (Figures 3a and b). Similarly, we found that animals expressing RNAi in salivary glands against CG11880, a gene of unknown function, were deficient in salivary gland degradation compared with controls (Figures 3c and d). All other morphological and developmental events occurred normally in these animals, including the timing and morphology of future adult head eversion. Therefore, the inhibition of salivary gland degradation upon either trol or CG11880 expression in salivary glands was not due to a developmental delay. Trol and CG11880 were identified following proteasome inhibition, which follows activation of caspases in salivary glands. As proteasome inhibition can lead to activation of autophagy,34 we tested if these genes are required for autophagy in dying salivary glands. To our surprise, knockdown of neither trol nor CG11880 had any impact on autophagy (data not shown).To our knowledge, this is the first demonstration that trol and CG1180 are required for salivary gland cell degradation.

Figure 3
figure 3

Trol and CG11880 are required for salivary gland cell degradation. (a) Control animals lacking the GAL4 driver (+/w; +/UAS-trol IR), n=10, and those with salivary gland-specific knockdown of trol (fkh-GAL4/w; UAS-trol IR/+), n=17, were analyzed by histology for the presence of salivary glands (red circles) 24 h AFP. (b) Quantification of data from (a) where 10% of control animals and 71% of experimental animals possess persistent salivary gland fragments. (c) Control animals lacking the GAL4 driver (+/w; +/UAS-CG11880IR), n=10, and those with salivary gland-specific knockdown of CG11880 (fkh-GAL4/w; UAS-CG11880IR/+), n=23, were analyzed by histology for the presence of salivary glands (red circles) 24 h AFP. (d) Quantification of data from (c) where 0% of control animals and 61% of experimental animals possess persistent salivary gland fragments

Full size image

The cop9 signalsome (CSN) functions in salivary gland degradation

We queried our proteomics data for proteins that are part of a pathway or cellular process and were enriched both in response to ectopic proteasome impairment and during wild-type salivary gland cell death. Interestingly, the CSN network was identified as being enriched by Ingenuity Pathway Analysis. A number of CSN subunits were detected in both wild-type dying salivary glands and in salivary glands with ectopic proteasome impairment (Table 2). Although the P-values for the detection of each individual component were relatively high and therefore did not engender high confidence in the detection of any single subunit, the detection of 7 out of 8 known CSN subunits, including 1b, 3, 4, 5, 6, 7, and 8, suggested a possible role for the CSN in salivary gland death (Table 2).

Table 2 CSN subunits detected in Drosophila larval salivary glands by proteomics
Full size table

To test the function of the CSN in salivary gland degradation, we expressed RNAi against the gene encoding CSN subunit 6 (CSN6) in salivary glands. Previous studies have shown that the loss or knockdown of one CSN subunit leads to the destabilization and loss of the entire complex.36, 37, 38 We found that whereas 10% of control animals contained persistent salivary gland fragments, 100% of those expressing csn6-IR in salivary glands contained persistent salivary gland fragments 24 h APF (Figures 4a and b). We identified csn6 as being enriched after proteasome inhibition. We tested if csn6 is required for autophagy in dying salivary glands as altered proteasome function can lead to activation of autophagy in neurons,34 however, knockdown of csn6 did not influence autophagy (data not shown).

Figure 4
figure 4

RNAi knockdown of CSN6 prevents salivary gland degradation. (a) Control animals lacking the GAL4 driver (+/w; +/UAS-csn6IR ), n=10, and those with salivary gland-specific knockdown of csn6 (fkh-GAL4/w; UAS-csn6IR/+), n=22, were analyzed by histology for the presence of salivary glands (red circles) 24 h AFP. csn6IR=VDRC TID no. 22308 on II. (b) Quantification of data from (a) where 10% of control animals and 100% of experimental animals possess persistent salivary gland fragments

Full size image

Discussion

Autophagic programmed cell death appears to be conserved between insects and mammals,25, 39 yet the mechanisms that contribute to this type of cell death are not well understood. We used a shotgun proteomic approach to identify proteins enriched during autophagic programmed cell death in purified dying Drosophila larval salivary glands. We found that a number of UPS components are enriched in dying salivary glands, suggesting a role for the UPS in salivary gland cell death. Significantly, UPS components are enriched in proteomic studies of apoptotic cells.40 We tested the function of the UPS in salivary glands and found that flux through the UPS pathway is decreased during cell death in wild-type salivary glands.

Although the mechanism leading to decreased UPS function during salivary gland cell death remains unknown, it could be that UPS function is reduced by system overload, but other possibilities exist. It is interesting that a large number of UPS components are enriched in dying salivary glands at the time when UPS function is decreased, suggesting a possible compensatory upregulation to deal with a reduction in the function of this important catabolic process. We find that ectopic proteasome impairment leads to the onset of early salivary gland condensation and DNA fragmentation in vivo. Unexpectedly, however, we find that caspase inhibitor p35 expression during proteasome impairment does not suppress ectopic salivary gland condensation. This suggests that proteasome impairment-induced gland condensation may require the activity of caspase-independent processes.

We queried our proteomics data to identify factors present in salivary glands upon proteasome impairment. To identify physiologically relevant factors that might function in salivary gland cell degradation, we compared factors identified following ectopic proteasome impairment in salivary glands to those identified during wild-type salivary gland cell death. The simple presence of proteins in dying salivary glands is not evidence that these proteins function in cell death and degradation. Therefore, we took a genetic approach to screen top candidates for defects in salivary gland degradation by RNAi knockdown. Here, we show that the knockdown of trol and CG11880, a gene of unknown function, prevent proper salivary gland degradation. The gene trol has been shown to function in controlling the onset of stem cell proliferation in the developing Drosophila nervous system by modulating fibroblast growth factor branchless and Hedgehog signaling, and mutations in trol also affect Dpp and Wingless signaling.41 However, neither CG11880 nor trol has previously been shown to function in salivary gland cell degradation.

To identify additional genes that might function in autophagic cell death, we searched for proteins that are members of a pathway or cellular component and present during wild-type cell death and in response to proteasome impairment in salivary glands. We identified 7 out of 8 known CSN components by proteomics and demonstrated that RNAi knockdown of CSN subunit csn6 prevents salivary gland degradation. The CSN is a conserved multi-subunit protein complex that functions in the ubiquitin/proteasome pathway.42 First discovered in plants, the CSN is essential for Drosophila development and functions in a variety of essential cellular processes, including cell cycle control, the DNA-damage response, and circadian rhythm.42 Composed of eight subunits, CSN1–CSN8, the CSN contains Nedd8 isopeptidase activity, and functions to remove the ubiquitin-like Nedd8 from proteins (deneddylate).42

Our finding that csn6 is required for salivary gland degradation demonstrates a role for the CSN in autophagic cell death, and suggests that proper control of protein neddylation and deneddylation cycles is important to regulate salivary gland degradation. Several studies have demonstrated a role for the CSN in regulating cell death. Csn5 knockout mice exhibit elevated apoptosis.43, 44 In addition, several cullin-RING ligase components negatively regulate Caenorhabditis elegans p53 (CEP-1)-dependent germ cell apoptosis in response to DNA damage.45 Cullin-RING ligase complexes may also function in salivary gland cell death, a subject for future investigation.

Our proteomics analyses led to the discovery of several genes required for proper salivary gland cell degradation. These findings exhibit the power of a high-throughput proteomics approach to identify factors that function in biological processes. Our proteomics data will serve as a foundation for future genetic studies of autophagic programmed cell death and in vivo cellular responses to proteasome inhibition by providing a description of proteins enriched during wild-type cell death as well as in response to ectopic proteasome impairment.

Materials and Methods

Drosophila strains

The Canton-S strain was used as wild-type. For ectopic proteasome impairment studies, UAS-Dts72c on III was used.35 To monitor proteasome impairment, UAS-M-GFP was used as the positive control, and UAS-GFP-CL1 was used as the experimental stock.32, 33, 34 For ectopic p35 expression, UAS-p35 was used.46 The following Vienna Drosophila RNAi Center (VDRC, Vienna, Austria) RNAi stocks were used: CG11880IR VDRC Transpformant ID (TID) 22867 on II, trolIR VDRC TID no. 245498 on II.

Proteomics analysis

To generate the proteomics data sets, at least three independent biological samples for each condition were collected. For each independent sample, 80 pairs of salivary glands were dissected per 100 μl 8 M urea/40 mM Tris-HCl, pH 8. At least two runs of each biological sample were analyzed, using 10 μg of sample per run, by high-throughput MS-based proteomics.47 Proteins were fully trypsinized, and peptides were separated by two-dimensional chromatography; first into 15 fractions by capillary isoelectric focusing, and further resolved by capillary reverse-phase liquid chromatography (RPLC) (nano-RPLC). Two runs of each protein sample were carried out. Fractions were analyzed by ESI-tandem-MS using a Thermo Finnigan LTQ linear ion trap mass spectrometer (San Jose, CA, USA). Following MS analysis, peaklists were generated using extract_msn, and searches to identify MS/MS peptide spectra were carried out using the Open Mass Sepctrometry Search Algorithm48 (NCBI). For each data set, protein and peptide hits were generated assuming a false positive rate of 5%, and proteins present in the samples were identified based upon peptide sequences detected. Comparison of proteomics data sets: Files containing flybase gene identifiers (FBGNs), flybase gene symbols, and RefSeq protein accession numbers were batch downloaded from http://www.flybase.org. Microsoft Access (Seattle, WA, USA) was then used to merge proteomics data files with that containing FBGN, flybase gene symbols, and RefSeq accessions. Using Microsoft Access, separate proteomics data sets containing ratio2 values were then merged based upon their common FBGN numbers. This allowed the comparison of ratio2 values and P-values between proteins identified in more than one proteomics data set.

Histology

Histology was performed as previously described.49

Protein extracts and western blotting

Protein extraction and western blotting were performed as previously described.50 Primary antibodies used were rabbit anti-ubiquitin (1 : 1000), and mouse anti-beta-tubulin (1 : 50; z).

Microscopy

Imaging was performed on a Zeiss (New York, NY, USA) Axiophot II microscope. To quantify salivary gland size, the area of the largest salivary gland section per pupa was measured using the Zeiss AxioVisionLE software.