Evaluation of Drosophila Metabolic Labeling Strategies for in Vivo

Mar 21, 2013 - Quantitative Proteomic Analyses with Applications to Early Pupa ... larval−pupa transition, while several subunits of the V-ATPase co...
1 downloads 0 Views 2MB Size
Download PDF
Article pubs.acs.org/jpr

Evaluation of Drosophila Metabolic Labeling Strategies for in Vivo Quantitative Proteomic Analyses with Applications to Early Pupa Formation and Amino Acid Starvation Ying-Che Chang,†,§,∥ Hong-Wen Tang,†,§ Suh-Yuen Liang,‡ Tsung-Hsien Pu,‡ Tzu-Ching Meng,†,§ Kay-Hooi Khoo,*,†,§ and Guang-Chao Chen*,†,§ †

Institute of Biological Chemistry, Academia Sinica, Taipei, Taiwan NRPB Core Facilities for Protein Structural Analysis, Academia Sinica, Taipei, Taiwan § Institute of Biochemical Sciences, National Taiwan University, Taipei, Taiwan ∥ Clinical Proteomics Center, Chang Gung Memorial Hospital, Taoyuan, Taiwan ‡

S Supporting Information *

ABSTRACT: Although stable isotope labeling by amino acids in cell culture (SILAC)-based quantitative proteomics was first developed as a cell culturebased technique, stable isotope-labeled amino acids have since been successfully introduced in vivo into select multicellular model organisms by manipulating the feeding diets. An earlier study by others has demonstrated that heavy lysine labeled Drosophila melanogaster can be derived by feeding with an exclusive heavy lysine labeled yeast diet. In this work, we have further evaluated the use of heavy lysine and/or arginine for metabolic labeling of fruit flies, with an aim to determine its respective quantification accuracy and versatility. In vivo conversion of heavy lysine and/or heavy arginine to several nonessential amino acids was observed in labeled flies, leading to distorted isotope pattern and underestimated heavy to light ratio. These quantification defects can nonetheless be rectified at protein level using the normalization function. The only caveat is that such a normalization strategy may not be suitable for every biological application, particularly when modified peptides need to be individually quantified at peptide level. In such cases, we showed that peptide ratios calculated from the summed intensities of all isotope peaks are less affected by the heavy amino acid conversion and therefore less sequence-dependent and more reliable. Applying either the single Lys8 or double Lys6/Arg10 metabolic labeling strategy to flies, we quantitatively mapped the proteomic changes during the onset of metamorphosis and upon amino acid deprivation. The expression of a number of steroid hormone 20-hydroxyecdysone regulated proteins was found to be changed significantly during larval−pupa transition, while several subunits of the V-ATPase complex and components regulating actomyosin were upregulated under starvation-induced autophagy conditions. KEYWORDS: SILAC, proteomic, Drosophila melanogaster, metamorphosis, starvation

■

INTRODUCTION Stable isotope labeling by amino acids in cell culture (SILAC)1 is the most commonly used metabolic labeling method to introduce isotope labels into cellular proteins for a straightforward mass spectrometry (MS)-based quantitative proteomic analysis. Compared with the 15N-metabolic labeling technique,2 the SILAC approach has an advantage that the introduced mass differences between labeled and unlabeled peptides are fixed.3 In contrast, amino acid sequence dependent mass shifts of 15N labeled peptides due to variable number of nitrogen atoms in each peptide can complicate its identification.4 Furthermore, the incorporation rate of heavy labels in SILAC experiments is typically very high because only one or two amino acids in each peptide is substituted,5 whereas for 15N-labeling, all nitrogen atoms within a peptide must be replaced, and incomplete labeling is often observed. Such incomplete 15N-labeling results in © 2013 American Chemical Society

additional isotope peaks occurring at m/z lower than the theoretical monoisotopic peak, which can reduce the overall sensitivity of peptide identification and also compromise the quantification accuracy.4 To perform SILAC experiments, a general assumption is that auxotroph for the particular labeled amino acid is required, which may not necessarily be the case.4,6 It has been demonstrated that if the labeled amino acid is sufficiently supplied, cells will utilize the heavy amino acid fed rather than spending resources on synthesizing the amino acid. This observation paves the way for applying SILAC on a wider range of biological sources by simply manipulating their respective diet feed. It is initially and primarily applied to mammalian cell culture or single cell organism such as Received: December 13, 2012 Published: March 21, 2013 2138

dx.doi.org/10.1021/pr301168x | J. Proteome Res. 2013, 12, 2138−2150

Journal of Proteome Research

Article

yeast7,8 but recently has also been extended to multicellular animals such as mice,9 fruit flies10 and nematodes.11,12 In most cases, isotope labeled lysine and arginine are used so that each tryptic peptide except the C-terminal peptide of the protein would be labeled. Further facilitated by powerful computational tools such as MaxQuant3 for the data analysis, the advantages over 15N-labeling are clear and significant. The only caveat that will compromise quantification accuracy is the well-appreciated metabolic conversion of arginine to proline,13−15 for which several corrective measures have been introduced.5,13,16,17 The fruit fly, Drosophila melanogaster (hereafter referred to as Drosophila), remains an indispensable model organism for myriad biological studies, particularly in the area of developmental sciences taking advantages of powerful genetic tools available. More recently, proteomic studies have been applied18−24 but mostly nonquantitative, particularly for those performed at the level of whole organism. To advance further in the direction of systems and computational biology, global quantitative mapping of the differentially expressed proteomes among various developmental stages or mutants are needed. Although raising flies with heavy lysine-labeled yeast cells has been reported recently,10,25 it remains unclear whether a triplet labeling strategy incorporating [ 2H4 ]-lysine (Lys4) and [13C6,15N2]-lysine (Lys8) can also be applied to Drosophila in vivo. Similarly, the technical feasibility of introducing heavy labeled arginine into flies to perform Lys/Arg-double labeling quantitative experiments has also not been established. In view of the added advantages afforded by multiplexed SILAC for quantitative proteomics studies, we have now systematically evaluated the characteristic features of [2H4]lysine (Lys4), [13C6]-lysine (Lys6), [13C615N2]-lysine (Lys8), and [13C615N4]-arginine (Arg10) incorporation into Drosophila and the respective quantification accuracy using heavy lysine, arginine or a combination of both. We achieved near complete labeling with our homemade fly food comprising only 5% SILAC-labeled yeast instead of feeding exclusively with the prohibitively expensive labeled yeast as used in a previous study.10 Unexpectedly, we observed a significant but undesirable in vivo conversion of heavy lysine or heavy arginine into several other amino acids, which could lead to severely underestimated quantification at individual peptide level. However, in cases when normalization on a protein level can be applied, our refined metabolic labeling approach in Drosophila remains practically useful, simple and economical, as demonstrated here with our mapping of the proteomic changes at different developmental stages and in dissected organ, fat body, upon amino acid starvation.

■

labeling, the YNB media were supplemented with 40 mg/L of [12C614N2]-lysine (Lys0) and 40 mg/L of [13C615N4]-arginine (Arg10). For labeling with both heavy lysine and heavy arginine, the YNB media were supplemented with 40 mg/L of [13C6]lysine (Lys6) and 40 mg/L of [13C615N4]-arginine (Arg10). Yeast cells were grown at 30 °C to O.D. 600 equal to 1.2 and then harvested and lyophilized. Lyophilized yeast powder was further used to make fly food. Briefly, lyophilized yeast (2.5 g, light or heavy labeled) was mixed with 0.4 g of agar, 2.5 g of sucrose, and 1 g of starch in 50 mL of water and then cooked. After cooling down for several minutes, 0.15 mL of propionic acid and 0.15 mL of p-hydroxy-benzoic acid methyl ester were added to the fly food. Metabolic Labeling and Sample Preparation of Drosophila

Wild-type Drosophila strain Oregon R was maintained by the standard method at 25 °C. For SILAC labeling, embryos (first filial generation; F1) were collected from apple juice agar plates and transferred to light or heavy SILAC food. After ∼10 days, hatched flies were transferred to fly cages and supplied with SILAC-labeled yeast paste. To facilitate synchronization of developmental stage at pupa formation, embryos (second filial generation; F2) were collected from apple juice plate within 3−4 h period and transferred to stable isotope-labeled fly food. The pupal stages were determined on the basis of the formation of white prepupa (0 h after puparium formation). For protein extraction, adult flies, embryos, (dechorionated with 50% bleach for 3 min and then washed with PBS twice) or pupae were dounce-homogenized in RIPA lysis buffer. Protein concentration was determined by BCA method (Pierce) after removal of debris by centrifugation. Fat bodies were dissected from F2 larvae at 72 h after egg laying and then boiled in 2% SDS buffer for 15 min. The insoluble fraction after centrifugation was removed. In-Gel Digestion

Extracted proteins were separated by 10% SDS-PAGE. The gels were stained with Coomassie Blue and excised into 5−10 small slices. Each gel slice was further cut into 1 mm3 cubes and separately incubated with 50 mM dithioerythritol (DTE) in 25 mM NH4HCO3 for 1 h at 37 °C, followed by 100 mM iodoacetamide (IAM) in 25 mM NH4HCO3 for another 1 h at RT in the dark, and then washed with 25 mM NH4HCO3 in 50% acetonitrile (ACN) and dehydrated with 100% ACN. For trypsin or Lys-C digestion, the gel pieces were rehydrated with trypsin or Lys-C solution and then digested overnight at 37 °C. For Arg-C digestion, excess IAM in the gels was first neutralized by 50 mM DTE, and the gels were dehydrated with 100% ACN before completely dried in a speed vacuum. The gel pieces were then soaked in Arg-C (Clostripain; Protea or Roche) solution (25 mM NH4HCO3, pH 7.6, 5 mM DTE, 8.5 mM CaCl2) and incubated for 16 h at 37 °C. After digestion, the supernatants were transferred to fresh tubes, and the resulting peptides were extracted by 5% trifluoroacetic acid (TFA) in 50% ACN. The extracts were combined, reduced in volume by a speed vacuum and further purified by using stage tips.

EXPERIMENTAL PROCEDURES

Labeled Fly Food Preparation

Lysine and arginine double auxotroph yeasts (ho::hisG, Δlys2, leu2::hisG, arg4-bgl, Δura3) from a stationary YPD culture were inoculated (dilution 1:100 000) into YNB medium (6.7 g/L of yeast nitrogen base without amino acid, containing 2% glucose, 200 mg/L of adenine sulfate, 50 mg/L of uracil, 100 mg/L of leucinine, 25 mg/L of histidine, 250 mg/L of proline, and 250 mg/L of glutamic acid; pH adjusted to 6.5 with 1 N NaOH). For light (normal) labeling, the YNB media were supplemented with 40 mg/L of [12C614N2]-lysine (Lys0) and [12C614N4]-arginine (Arg0). For heavy lysine labeling, the YNB media were supplemented with 40 mg/L of [13C615N2]-lysine (Lys8) and 40 mg/L of [12C614N4]-arginine (Arg0). For heavy arginine

LC−MS/MS Analysis

Peptides were diluted with 1% formic acid and analyzed by nanospray LC−MS/MS on an LTQ-Orbitrap (XL or Velos) mass spectrometer (Thermo Scientific) coupled to a nanoACQUITY UPLC system (Waters) via a PicoView nanospray interface (New Objective,). Peptide mixtures were loaded onto a 75 μm × 250 mm nanoACQUITY UPLC BEH130 column packed with C18 resin (particle size 1.7 μm, pore size 130 Å; 2139

dx.doi.org/10.1021/pr301168x | J. Proteome Res. 2013, 12, 2138−2150

Journal of Proteome Research

Article

Figure 1. Near-complete incorporation of heavy lysine and arginine with distorted isotope patterns. The MS profiles of a representative peptide each from Lys8 (A, left panel) and Arg10 (B, left panel) labeled adult flies in F1 generation are shown. The monoisotopic peaks for the nonlabeled and labeled peptides are marked with filled and open dots, respectively. The triangle marks the monoisotopic peak of Arg10 labeled peptide containing a heavy proline, which leads to an additional mass shift of 6 Da. In comparison with the theoretical isotope patterns (right panels), the actual isotope patterns of Lys8 or Arg10 labeled peptide showed a shift toward higher m/z region.

Waters) and were separated at a flow rate of 300 nL/min using a linear gradient from 5 to 40% acetonitrile (in 0.1% formic acid) within 90 min. The spray voltage was set to 2.0 kV, and the temperature of the heated capillary was set to 35 °C. The mass spectrometer was operated in the data-dependent mode. Survey full scan MS spectra were acquired in the Orbitrap with a resolution of 60 000 at m/z = 400 after accumulation of 1 × 106 ions in the Orbitrap. Up to 10 (for XL) or 20 (for Velos) most intense ions detected in the survey scan were further sequenced by CID (collision energy 35%) in the LTQ after accumulation of 7000 ions. The Orbitrap measurements were performed with the lock mass option enabled to improve mass accuracy.

Trypsin/P for comparison. The search parameter Lys-C/P, ArgC/P and Trypsin/P allow the cleavage to occur at their target sites, even if these sites were followed by a proline residue. The mass accuracy tolerance for precursor ion was 6 ppm and 0.5 Da for fragment ions. Shared peptide sequences were reported as protein grouped accessions. To achieve reliable identifications, the maximal false discovery rate (FDR) of peptides and proteins was set at 0.01. The minimal peptide length accepted was 6 amino acids and protein group identification required at least 1 unique peptide. All quantified proteins have at least 2 peptides quantified (razor and unique). The final MaxQuant-computed ratios of heavy labeled peptides/proteins to nonlabeled peptides/proteins were referred to as MQ-H/L ratios here. The manually calculated ratios of heavy to light peptides based on the respective H and L intensity values found in the evidence table output by MaxQuant were defined as XIC intensity-H/L ratios. To reduce the system error in each experiment, the median normalization was applied.

Processing of MS Data

The raw files were processed by MaxQuant software (version 1.2.2.5), and Andromeda26 was used to search the acquired CID spectra. The protein database consisted of Drosophila (FlyBase version 5.17; 21753 entries), Saccharomyces cerevisiae sequence (download from NCBI in Mar 30, 2009; 5880 entries) and commonly observed contaminants. The search parameters were as follows: carbamidomethylation of cysteine as a fixed modification, methionine oxidation and protein N-terminal acetylation as variable modifications. Up to 2 missed cleavages were allowed. For trypsin and Lys-C digested samples, enzyme specificity was set as trypsin and Lys-C/P, respectively. For ArgC digested samples, enzyme specificity was set as Arg-C/P or

■

RESULTS

SILAC Diet and Incorporation of Heavy Labeled Lysine and Arginine

A precondition underpinning the success of a SILAC approach is the ability to achieve near complete incorporation of the heavy amino acids into the proteome of the biological subject without 2140

dx.doi.org/10.1021/pr301168x | J. Proteome Res. 2013, 12, 2138−2150

Journal of Proteome Research

Article

Figure 2. Distribution of quantified peptide pairs incorporating different numbers of isotope labels and their respective median MQ-H/L ratios before and after normalization. Peptides were derived from equal mixtures of (A) Lys8 and Lys0 labeled F1 adult males, (B) Lys8 and Lys0 labeled F2 pupae, (C) Arg10 and Arg0 labeled F2 pupae, and (D) Lys6Arg10 and Lys0Arg0 labeled F2 larval fat body. The LC−MS/MS data were processed using MaxQuant, and the resulting quantified peptides were then sorted into different bins according to the total number of heavy isotope contaminated amino acids present in their sequences (see Figure S3, Supporting Information, for the heavy isotope contaminated amino acids). The peptides sorted in bin 0 contain no heavy isotope contaminated amino acid in sequence, and so on. The peptide counts in each bin were plotted as histograms to show the overall distribution. In each bin, the non-normalized and normalized median MQ-H/L peptide ratio were calculated and plotted.

labeled yeast, supplemented by 5% (w/v) sucrose, 2% (w/v) starch, and 0.8% (w/v) agar (see Experimental Procedures). Flies supplied with Lys8 or Arg10 labeled food were found to grow normally and showed no obvious difference in morphology, development or fertility between heavy and light labeled ones over three generations. This observation is similar to that

any tangible detrimental effects on its growth. In the case of introducing isotope labeled amino acids into Drosophila, we have adopted a two-step approach similar to a previous report10 but with a slight modification. Instead of feeding Drosophila directly with a diet consisting exclusively of SILAC-labeled yeast, our flies were raised on homemade food comprising only 5% (w/v) 2141

dx.doi.org/10.1021/pr301168x | J. Proteome Res. 2013, 12, 2138−2150

Journal of Proteome Research

Article

not be rectified by the commonly applied “median normalization” (Figure 2A,B) or any other simple means. To determine whether the observed lysine conversion was due to lower content of yeast in the diet, the publicly deposited data from previous study of heavy lysine labeled flies raised with exclusive SILAC-labeled yeast diet10 were re-examined by us using the same data analysis method. A similar result was obtained, which clearly showed that heavy label from Lys8 was transferred to several nonessential amino acids, the isotope pattern was thus distorted, and the generated MQ-H/L ratios were likewise sequence (amino acids) dependently underestimated (data not shown). Collectively, these data suggest that lysine conversion was not due to lower content of yeast in the diet but likely to reflect the unique metabolism of lysine in Drosophila. In search for a potential solution for accurate quantification, we have additionally examined the respective H/L intensity outputs by MaxQuant for individual peptides, which were computed on the basis of the sum of extracted ion current (XIC) of all associated isotope peaks. The H/L peptide ratio manually calculated from these values is defined here as XIC intensity-H/L ratio. Overall, at peptide level, adopting the XIC intensity-H/L values as the basis of quantification appears to fair better than directly using the MQ-H/L values provided by MaxQuant without normalization. The log2 of XIC intensity-H/L ratios approximated a median of −0.5 across the entire peptide range (Figure 3), in contrast to that shown by MQ-H/L ratios, which

reported previously10 indicating that 13C and 15N has no or little side effects. In contrast, higher mortality was observed at larval and pupal stages when the larvae were continually fed on Lys4 labeled food (data not shown), which indicates that deuterium labeled lysine is not suitable for metabolic labeling of Drosophila. To determine the features and incorporation rates of Lys8 or Arg10 into Drosophila fed on our homemade SILAC-labeled yeast food, proteins were extracted from Lys8 or Arg10 labeled F1 adult flies (20 flies each, within one day after hatching) and digested by Lys-C or Arg-C, respectively, to obtain peptides with at least one heavy labeled amino acid in sequence. These were analyzed by LC−MS/MS in the normal data dependent acquisition mode, and the collected raw spectral data sets were processed using MaxQuant version 1.2.2.5.3 The incorporation rates calculated from all quantified proteins were 93.7 and 94.5% for Lys8 and Arg10, respectively (Figure S1A,B, Table S1, Supporting Information), which indicate that a fly food consisting of only 5% labeled yeast is sufficient to achieve nearcomplete labeling for Drosophila in first filial (F1) generation. However, we found that the isotope patterns of most Lys8 or Arg10 labeled peptides showed a shift toward higher m/z values compared to their theoretical patterns (Figure 1). For Arg10 labeled peptides containing a proline, an additional 6 Da mass shift due to heavy proline was also observed (Figure 1B). These mass shifts indicate that a small but significant proportion of heavy label was undesirably transferred from Lys8 or Arg10 into other amino acids. However, such a mass shift was not observed in proteomic analysis of Lys8 or Arg10 labeled yeast used to feed Drosophila (data not shown), suggesting that the unexpected in vivo metabolic conversion of lysine and arginine to other amino acids occurred mainly in Drosophila. Quantification Accuracy Based on Single Lys- and Double Lys/Arg-Labeling

To determine whether the Lys8 or Arg10 conversion will affect the quantification accuracy, a series of quantitative proteomic experiments were performed. Equal mixtures of Lys0 and Lys8 labeled proteins extracted from F1 adult male flies (30 flies each, hatched five days) or F2 pupae (25 pupae each at white pupa stage) were subjected to Lys-C digestion and LC−MS/MS analyses (two technical replicates), followed by data processing and analysis using MaxQuant to calculate the experimentally derived values of quantified heavy labeled proteins relative to nonlabeled proteins. The MQ-H/L ratios thus defined were inspected with histogram plots for overall ratio distribution (Figure S2, Supporting Information). Since all proteins were expected to be of equal abundance, the log2 of their MQ-H/L ratios should theoretically be approximating zero, but more than 30% of the quantified proteins actually gave a log2 MQ-H/L ratio less than −1 before normalization (Figure S2A,B, Supporting Information). This underestimated ratio may be attributed to the isotope pattern shift in Lys8-labeled peptides. Upon further examination, we found that the heavy label from Lys8 could be additionally transferred to several amino acids including alanine, aspartic acid, glutamic acid, glutamine, and proline (Figure S3A,B, Supporting Information). Given that any peptide may contain several of the implicated amino acids in its sequence, the accuracy of the reported MQ-H/L ratios could be severely compromised in certain peptides carrying a higher proportion of these residues. Since the degree of additional heavy label incorporation per peptide is amino acid sequence dependent, the observed quantification defect at individual peptide level could

Figure 3. The median intensity-H/L ratios of quantified peptide pairs incorporating different numbers of isotope labels. The quantified peptides from each SILAC experiment were sorted into different bins according to the total number of heavy isotope contaminated amino acids in their sequences.

can deviate to as much as −2 for certain peptide bin (Figure 2A,B). However, this apparent gain in quantification accuracy at individual peptide level did not extend to protein level, for which the error of underestimation based on the computed MQ-H/L ratios became insignificant after normalization to the median value (Figure 4A,B). Thus for the data set from Lys8/Lys0 labeled F1 adult males, 99.4% of proteins were quantified as giving less than 2 fold change and 96.6% with less than 1.5 fold. For the data set from Lys8/Lys0 labeled F2 pupae, 99.0% of proteins registered less than 2 fold change and 94.6% with less than 1.5 fold (Tables S2 and S3, Supporting Information). These values are well within the common error margin of typical SILAC experiments.9,12 The robustness and computational strengths of MaxQuant in inferring expressed protein ratios based on at least 2 quantified peptides surpasses the gain in quantification 2142

dx.doi.org/10.1021/pr301168x | J. Proteome Res. 2013, 12, 2138−2150

Journal of Proteome Research

Article

Figure 4. Histograms of normalized protein ratios as calculated by MaxQuant from the LC−MS/MS proteomic data sets for the 1:1 mixtures of (A) Lys8 and Lys0 labeled F1 adult males, (B) Lys8 and Lys0 labeled F2 pupae, (C) Arg10 and Arg0 labeled F2 pupae, and (D) Lys6Arg10 and Lys0Arg0 F2 larval fat body tissues dissected at 72 h after egg laying.

accuracy at individual peptide level provided by the alternative manual calculation method. Similar observations and conclusions could be drawn from SILAC experiments with Arg10 labeled Drosophila (Figure S2C, Supporting Information). The undesirable transfer of heavy label into alanine, aspartic acid, and glutamine in addition to the expected proline (Figure S3C, Supporting Information) is an issue at individual peptide level (Figure 2C) but likewise found to be negligible at protein level after normalization (Figure 4C, Table S4, Supporting Information), as demonstrated with the data set of Arg0 and Arg10 (1:1) labeled protein extracts from F2 white pupae (25 pupae each) digested with Arg-C (Clostripain). Finally, for double labeling with both heavy lysine and arginine in a single SILAC experiment, the incorporation of Lys6 and Arg10 at F1 adult (25 flies within one day after hatching) was determined to be also near complete (Figure S1C,D, Table S5, Supporting Information). When tested against equal amounts of fat body tissues dissected from light and heavy F2 larvae (15 larvae each, 72 h after egg laying), the derived MQ-H/L ratios were found to be similarly underestimated (Figures 2D, S2D, Supporting Information) because of heavy isotope transferred into alanine, aspartate, glutamic acid, glutamine, and proline (Figure S3D, Supporting Information), but could be largely corrected at the protein level after normalization, with 99.5% of

proteins showing less than 2 fold change and 96.0% less than 1.5 fold change (Figure 4D, Table S6, Supporting Information). In summary, we showed that 13C- and 15N-labeled lysine and arginine but not 2H-labeled lysine could be efficiently incorporated into Drosophila F1 adult and F2 larval proteome with no adverse effects on growth. However, significant in vivo conversion of heavy lysine and arginine into other amino acids would result in a distorted isotope pattern (Figure 1), leading to an underestimation of the quantity of heavy peptides relative to nonlabeled ones by MaxQuant (Figure S2, Supporting Information). A simple built-in solution is to apply the commonly adopted normalization strategy based on the assumption that the expression level of most proteins in the sample will not be altered by the perturbation under investigation. This would effectively address the aforementioned problem for quantitation on a protein level for most applications (Figure 4). However, such corrective measure is ineffective at individual peptide level (Figure 2) and may not be applicable when the induced changes could affect a board range of protein expression. More importantly, in quantitative analysis of post translationally modified peptide, the MQ-H/L ratios of particular modified sites of interest can be unduly underestimated because of presence of multiple heavy isotope-contaminated amino acids in sequences. We suggest that in such cases, an additional level of manual examination supplemented by an independent calcu2143

dx.doi.org/10.1021/pr301168x | J. Proteome Res. 2013, 12, 2138−2150

Journal of Proteome Research

Article

lation of the XIC intensity H/L ratio, which takes into account all associated isotopic peaks, would provide a more accurate quantification for the particular peptides of interest (Figure 3). In general, the intensity-H/L ratios were less affected by lysine conversion and more accurate especially for the peptides containing multiple isotope contaminated amino acids in sequences. It is therefore more suitable for quantification when the global normalization strategy using the median or mean of all peptides from the sample is not suitable. Comparison of Protein Changes at Early Puparium Formation Using Heavy Lysine Only

Drosophila metamorphosis is one of the most dramatic morphogenetic processes in fly development. It is controlled by the steroid hormone 20-hydroxyecdysone (ecdysone), which initiates a transcriptional cascade that leads to tissues remodeling during larval−pupal transition. Several studies have used microarrays to analyze genome-wide gene expression changes at the onset of metamorphosis.27,28 However, the changes at the level of protein expression remain unclear. To demonstrate the use of the metabolic labeling (with labeled amino acids) for proteomic studies of Drosophila, we performed a quantitative proteomic comparison of two developmental stages, puparium formation (PF) and 6 h after puparium formation (APF). At the puparium formation stage, which lasts for approximately 15−30 min, the larva shortens, becomes immobile, and everts its anterior spiracles. Following this period, the white puparial cuticle tans and forms a protective case.29 The dynamic morphological changes during early pupal development make it ideal to monitor the differential expression of the proteome. Lys0 labeled pupae were collected at PF (n = 27) stage, and Lys8 labeled pupae were collected at 6 h APF (n = 20) in F2 generation. Equal amount of heavy and light labeled proteins were mixed, separated by SDS-PAGE (10 fractions), digested with Lys-C, and further analyzed by LC−MS/MS (in duplicate). A total of 20 LC−MS/MS runs resulted in 2542 quantified proteins (Table S7, Supporting Information). In total, 65 proteins (∼2.55%) were up-regulated at least 2 fold, and 164 proteins (6.45%) were down-regulated at least 2 fold. Consistent with the DNA microarray data,27,28 our SILAC analysis revealed that protein products of ecdysone-regulated genes such as L71−6, ImpL1, ImpE1, and ecdysone-inducible heat shock genes, Hsp23 and Hsp67, were among the upregulated proteins during early puparium stages (Figure 5A). Although the clear functions of these proteins in metamorphosis are not yet known, L71−6 has been implicated in protecting the animal from bacterial infections.30 ImpL1 and ImpE1 have been shown to be involved in ecdysone-mediated differentiation of imaginal discs.31,32 We also identified a set of down-regulated protein products of ecdysone-repressible genes: the salivary gland secretion genes Sgs3, Sgs4, Sgs7 and Sgs8.33,34 These Sgs glue proteins are known as the components secreted by the larva to attach itself to a solid surface for pupation.34 In agreement with our proteomic data, previous Northern blot and microarray analysis have also shown that the mRNAs of salivary glue secretion family were down-regulated from PF to 6 h APF.28 However, further systematic comparison revealed that the overall correlation between mRNA expression levels and proteins abundance (n = 2312) is quite low (r = 0.25) (Figure 5B, Table S8, Supporting Information). For example, a group of proteins identified by Gene Ontology (GO) analysis as belonging to structural constituents of chitin-based cuticle, including larval cuticle proteins (Lcp3, Lcp4, Lcp9, and

Figure 5. (A) The log2-fold protein changes upon pupa formation based on normalized MQ-H/L ratios versus sum of the peptide intensities detected for each protein. The Lys0 and Lys8 labeled pupae were collected respectively at puparium formation and 6 h after puparium formation. (B) Log2-fold protein changes versus previously published log2-fold mRNA changes, with a Pearson’s correlation coefficient of 0.25.

Lcp65Af),35 cuticular proteins (Cpr11A, Cpr47Eg, Cpr49Ac, and Cpr49Af), and pupal cuticle protein (Pcp), were found here to be significantly down-regulated at 6 h APF. Yet there was no significant change reported in mRNA expression level of Lcp3, Lcp9, Lcp65Af, Cpr11A, Cpr47Eg and Cpr49Af from PF to 6 h APF (Figure 5B, Table S8, Supporting Information).28 These results suggest that a lag exists between mRNA expressions and the corresponding protein abundance during early pupa formation, which can be further delineated by more detailed quantitative proteomic analyses over a period of time. GO analysis also revealed that five members of the cytochrome P450 family proteins and three glutathione S-transferase proteins were down-regulated, while several immune response proteins and three myofibril assembly proteins such as paramyosin, myofilin, and sallimus (mammalian titin) were up-regulated (Table S7, Supporting Information). In addition, several metabolic enzymes including tyrosine 3-monooxygenase (DTH), DOPA decarboxylase (Ddc), and dopamine Nacetyltransferase (Dat) were identified in our SILAC data (Table S7, Supporting Information). These proteins are required for cuticle tanning and sclerotization.36,37 In Drosophila, the 2144

dx.doi.org/10.1021/pr301168x | J. Proteome Res. 2013, 12, 2138−2150

Journal of Proteome Research

Article

Figure 6. (A) The SILAC fly based quantitative proteomic experimental workflow for identification of starvation-induced protein changes. The fat body tissues dissected from Lys0 and Arg0 labeled F2 pupae were collected at 72 h after egg laying. The Lys6 and Arg10 labeled F2 larval were collected at 68 h after egg disposition and soaked in 20% sucrose/PBS solution for another 4 h. The fat body tissues were then dissected from heavy labeled larvae. Equal amount of heavy and light labeled fat bodies were mixed, boiled in 2% SDS buffer, separated by SDS-PAGE and analyzed by LC−MS/MS after in gel tryptic digestion and peptide extraction. (B) The log2-fold protein changes after starvation versus sum of the peptide intensities detected for each protein.

identification of many ecdysone-regulated structural and metabolic proteins during early puparium stages.

tanning of pupal cuticle occurs within 1 h after puparium formation. This process requires conversion of tyrosine to dopamine, which depends on the actions of DTH and Ddc. The N-acylation of dompamine is catalyzed by Dat, and further processing of N-acetyldopamine is required for sclerotization. Although there is only a moderate decrease in protein expression of DTH, Ddc, and Dat (the log2 fold protein changes were −0.88, −0.45, and −0.96, respectively), our data suggests that these proteins were degraded after pupal cuticle tanning and sclerotization.38 Taken together, these results demonstrate that the stable isotope labeling of Drosophila by amino acids could facilitate quantitative proteomic studies leading effectively to the

Identification of Starvation-Induced Protein Changes Using Heavy Lysine and Heavy Arginine

The ability of organisms to adjust themselves in response to changes in nutrient levels is critical for their growth and survival. Although a number of signaling processes involved in regulating gene expression and cellular metabolism under starvation conditions have been identified, the molecules that coordinate these events are still poorly understood. We used double labeling strategy (Lys6 and Arg10) to identify amino acid starvation induced protein changes in larval fat body tissues. Drosophila fat 2145

dx.doi.org/10.1021/pr301168x | J. Proteome Res. 2013, 12, 2138−2150

Journal of Proteome Research

Article

of V-ATPase in regulating membrane fusion and vesicle formation48 may play an important role in autophagic processes.49,50 Furthermore, GO analysis of proteins significantly up-regulated at 4 h starvation also identified proteins that are involved in regulating actomyosin structure organization, including myosin heavy chain (Mhc), myosin light chain 2 (Mlc2), paramyosin (Prm), wings up A (WupA), and upheld (Up) (Figure 5, Table S9, Supporting Information). Interestingly, recent studies have shown that actomyosin activation plays a critical role in the induction of autophagy.51 Together, our results suggest potential roles of V-ATPase and myosin regulators in starvation-induced autophagy. Further investigation on their functions in membrane dynamics may help us to clarify their role in autophagy during nutrient starvation.

body is a nutrient-sensing and storage organ analogous to the mammalian liver and adipose tissues. The fat body tissues of Lys6-labeled (light) F2 early third instar larvae were dissected at 72 h after egg deposition (AED) under nutrient rich conditions. For amino acid starvation, the Arg10-labeled (heavy) F2 larvae were grown to 68 h AED and transferred to 20% sucrose/PBS for additional 4 h. The fat body tissues collected from light and heavy labeled larvae (each 15 larvae) were mixed, processed and analyzed by LC−MS/MS (Figure 6A). As shown in Table S9 (Supporting Information), a total of 1470 proteins were quantified. We found that 1245 out of 1470 proteins have previously been identified by the Drosophila whole genome microarray assay.39 The distribution of normalized protein MQH/L (log2 scale) ratios against the sum of the peptide intensities was shown in Figure 6B. Among 1470 proteins, 57 proteins (4.6%) were down-regulated at least 2 fold, and only 12 proteins (1%) were up-regulated at least 2 fold in amino acid starvation condition (Table S9, Supporting Information). SILAC analysis revealed that components of the larval serum protein-1 complex (Lsp1α, Lsp1β, Lsp1γ) and fat body protein2 (Fbp2) were down-regulated at least 2 fold under amino acid starvation conditions. Lsp1 complex is enriched in larval fat body cells and functions as a nutrient reservoir for pupal development.40 Consistent with our findings, the mRNAs encoding Lsp1 family proteins were also found to be down-regulated in a microarray assay.41 The marked decrease in the amounts of Lsp-1 and Fbp2 proteins reflects the need to utilize the internal amino acid from storage pool in response to nutrient depletion. Notably, several proteins involved in the macromolecular complex assembly such as Nap1, Mcm7 and Ranbp9 were found to be down-regulated under starvation conditions. Among the proteins that were up-regulated in response to nutrient depletion, we found that enzymes involved in the metabolic pathways such as astray (aay) and phosphoenolpyruvate carboxykinase (pepck) were up-regulated for more than 2fold at 4 h starvation. Astray encodes a phosphoserine phosphatase, which may catalyze the ATP and drive the final step of biosynthesis of L-serine from carbohydrate.42 Pepck catalyzes the reaction that converts oxaloacetate to phosphoenolpyruvate, which can be further converted to pyruvate for gluconeogenesis43,44 or biosynthesis of alanine.45,46 Since glucose was sufficiently supplied in our conditions, we therefore suggest that the metabolic pathway may favor conversion of pyruvate to alanine. We suspect that Astray and Pepck may participate in the de novo synthesis of certain amino acids under starvation conditions. Interestingly, both aay and pepck mRNA were also found to be up-regulated in the microarray assay.41 In addition, we found that six subunits of the V1 complex of vacuolar H+-ATPase (V-ATPase), including Vha26, Vha36−1, Vha44, Vha55, Vha68−2, and VhaSFD, were markedly upregulated (log2 fold change from 0.95 to 0.82; Table S9, Supporting Information). Moreover, several components of the V-ATPase V0 complex including VhaAC39−1, Vha100−1, and Vha100−2 were also up-regulated (log2 fold change from 0.79 to 1.00; Table S9, Supporting Information) upon amino acid starvation. V-ATPase functions as a proton pump that acidifies intracellular organelles and is essential for protein sorting and degradation of endocytic cargoes. Recent studies have implicated a role of V-ATPase in autophagy.47 However, the underlying mechanism remains elusive. Autophagy is a major catabolic process by which cells self-digest cytosolic macromolecules or intracellular organelles to promote cell survival in response to nutrient deprivation. It has been speculated that the involvement

■

DISCUSSION Recognizing the analytical strengths of a SILAC-based quantitative proteomics enabled at whole fly level, we have systematically investigated all experimental aspects associated with introducing the heavy lysine and arginine labels and ensuing data analysis. First, we have successfully formulated a yeast− sugar−starch medium to replace the exclusive diet of SILAClabeled yeast. In general, 5 mL of heavy labeling medium containing 5% w/v labeled yeast could support the growth of 50−70 larvae into adults while achieving near complete labeling. Using a different culture setting, Xu et al. also found that a reduced amount of labeled yeast in their culture condition is enough for Drosophila to achieve near complete labeling.25 Both approaches demonstrated that the consumption of SILAClabeled yeast, and thus the cost of heavy labeled amino acid, can be significantly reduced without adverse effect on labeling efficiency, making in vivo stable isotope labeling of Drosophila by amino acids more affordable for a large scale proteomic study. The use of heavy lysine, in general, is more advantageous than heavy arginine because of the often observed arginine to proline conversion in SILAC experiments.52 However, in the case of Drosophila, our results have shown that heavy labels from both lysine (Lys6, Lys8) and arginine (Arg10) could be transferred in vivo to multiple amino acids. This resulted in distorted isotope pattern exhibited by the labeled peptides, which in turn affected the overall quantification accuracy. Similar shift in isotopic pattern is also evident from the exemplary MS spectra shown in previous work using heavy lysine labeled fruit flies derived from feeding exclusively with heavy lysine labeled yeast diet, although the problem was neither explicitly pointed out then nor addressed.10 Neither was it addressed in the more recent work by Xu et al., and it is unclear whether the conversion rate is reduced in their system since no raw data is available.25 Interestingly, neither lysine nor arginine conversion was observed in SILAC S2 cell culture system even when the concentration of heavy lysine and heavy arginine in the medium are 1600 and 400 mg/L, respectively (unpublished data and previous report53,54). This may be attributed to adequate nonessential amino acids supply in cell culture medium and that these free amino acids may be used directly rather than synthesized de novo or through metabolic conversion. To minimize the effect of heavy labels conversion, a simple solution for cell culture system is to empirically reduce the concentration of the heavy labeled amino acid5 or to supply sufficient target amino acids in medium.13 However, reducing the supply of heavy lysine or arginine is not applicable for the twostep labeling approach employed here. As noted above, reverting back to the use of exclusive labeled yeast diet10 instead of 5% used 2146

dx.doi.org/10.1021/pr301168x | J. Proteome Res. 2013, 12, 2138−2150

Journal of Proteome Research

Article

conclusion. Despite the noted disadvantage in using stable isotope labeled amino acids to label Drosophila, it is still a very powerful approach for select biological problems to be investigated such as the 2 case examples reported here. Compared to other methods, the experimental errors arising from the multiple steps of sample preparation can be significantly reduced since heavy and light labeled samples are combined at the earliest stage. Furthermore, with the whole animal being isotope-labeled, quantitative proteomic changes in response to physiological challenge can be analyzed for many different tissues or organs all at once. Autophagy is a catabolic process that degrades intracellular proteins or organelles via the lysosome. Under nutrient starvation conditions, the autophagy−lysosomal system degrades and recycles cytoplasmic components for cell survival.70 It has been shown that autophagy-related proteins (Atg) play critical roles in the activation of autophagy. However, we did not identify any Atg in our quantitative fly analysis. Instead, we found a marked increase in subunits of V-ATPase complex upon nutrient starvation. V-ATPase acts as a proton pump and is required for the acidification of the lysosome.71 This finding is consistent with the demand for increased lysosomal system during induction of autophagy. In addition to the proton pumping activity, recent studies have suggested that V-ATPase may be involved in regulating membrane dynamics.72 Further study is necessary to delineate the molecular function of VATPase during autophagic process. We believe that SILAC flies will greatly advance our understanding of these biological processes in vivo.

in this study will not alleviate the problem. We have also tried to add extra amino acids such as proline and glutamic acid into fly food to block conversion but without success. Moreover, a high dose supply of certain amino acids seems to be harmful for Drosophila larvae.55 Recently, RNAi has been utilized to eliminate the problem of arginine-to-proline conversion in metabolic labeling of Caenorhabditis elegans,12 but such feeding methods for RNAi delivery did not appear to work in Drosophila.56 To delete or mutate the genes participating in lysine or arginine catabolic pathway may be an alternative strategy,15 but whether it may compromise other biological processes or functions needs to be carefully examined. Short of an immediate solution and given that conversion occurs with both Lys and Arg, our results indicated that either Lys6 or Lys8 but not Lys4 is preferable to Arg10 for single labeling. Lys4 is also commonly used in mammalian cell culture for SILAC-based quantitative proteomics, and no cytotoxicity effect has ever been reported.57−63 Unexpectedly, a high mortality was observed when Drosophila was fed with Lys4labeled yeast. This may be correlated with the formation of deuterium monoxide (D2O), which has been reported to inhibit cell division and resulting in cell death.64,65 The deuterium might be released from Lys4 through catabolic pathway and converted to D2O in Drosophila in vivo. It remains to be investigated if Lys4 is suitable for other model organisms, for example, to derive labeled mice and nematode. The shortcoming of Arg10 compared to Lys6/Lys8 in the case of metabolic labeling in flies is more to do with the general problem associated with the use of Arg-C (clostripain) for proteolytic digestion rather than the Arg to Pro conversion. Although Arg-C cleaves primarily at the C-terminus of arginine residue, cleavage at the lysine residue was also reported66 and similarly observed in our experiments. Many additional peptides ending with a C-terminal Lys instead of Arg were identified if the enzyme specificity was set to trypsin/P instead of Arg-C/P in the search criteria (Figure S4, Supporting Information). These non-Arg10-labeled peptides will reduce the overall sensitivity for quantification and also complicate spectral features. Nevertheless, this is less a problem other than increasing the cost in the case of double labeling using both Lys6/Lys8 and Arg10, since trypsin would have been used. Irrespective of using single or double labeling, the Lys8 and Arg10 conversion reported here could result in underestimated quantification. In the case of MaxQuant, which first determines all corresponding peak intensities for both heavy and light forms, the peptide ratios calculated using least-squares regression will be underestimated due to reduced intensity of the monoisotopic peak of heavy labeled peptides. We have also evaluated Proteome Discoverer (vr 1.3) and observed similar underestimation (data not shown), probably since not all isotopic peaks were quantified because of pattern distortion. Other quantification softwares such as MSQuant, which utilizes monoisotopic peak to calculate the peptide ratio;4,67 ASAPRatio, which sums up the ion intensities covering the first three theoretical peaks for quantification;68 and RelEx, which sums up the ions current within an m/z range of theoretical isotopic distribution and then uses least-squares regression to evaluate the peptide ratio69 may all have the same problem. Normalization can largely overcome this defect at protein level, but the problem remains with experiments in which global normalization cannot be applied. In particular for the case of modified peptides to which normalization at protein level is not applicable, additional calculation of XIC-intensity as introduced here should serve as a second tier check to avoid erroneous

■

ASSOCIATED CONTENT

S Supporting Information *

Tables S1−S11 and Figures S1−S4. This material is available free of charge via the Internet at http://pubs.acs.org. The acquired raw data is available on Tranche proteome. Lys8/Lys0-(1:1)-F1adult male: rV1erMcLaAW/AlF6L01yJtogFotmOIZBvAak +PX6nMmXhOzGj/PKL74+iNPjmnrIDsrFeWFZz0ay3skWHCn8YIdBXpAAAAAAAAAJEg==. Passphrase: pcLeovfxntaUE2WgYstV. Lys8/Lys0-(1:1)-F 2 -pupa: 5sQX25YmBaYLCQYFiRwDzdovqnrUow0QH5DY2B5SJ26b7Nbxmrz12xSjQCuTJkWArFxSH9LwQmVUMpnKRQbtBnHWLmIAAAAAAAAI8g==. Passphrase: 6RhECW7k81oYQQRxjy26. Arg10/Arg0-(1:1)-F 2 -pupa: 1KqMxT7Nzx4IApPxdll11s5DOTLZJ7lH348/3ee/HwSBhitK6YCLH63UJE7Fq7n2G4OpdkR2vvpWEXcOcoM88Hl3q4sAAAAAAAAQ4w==. Passphrase :O1iXQIs9ubGEWILDcNGf. Lys8/Lys0-(6hAPF-vs-PF) F2-pupa: 1SALOpFw+gnW80d +ah8I4KbpJnw/7OSPYpzbJVEm/zazNKP89CiE75K0Uzzdnf9Jp41XAmm0kMQyk5ddqFtOGsXKGJEAAAAAAAAR8Q==. Passphrase: g7pIcyHRQhTYxZcJnPt2. K6R10/K0R0-(fed vs fed)-F2-larval fat body: IKrKhQuz8RIjSCxJYCOcUYdvWtJugrMrhiCR6lIECJ5xmLNdgn +vdGEQtToUWbV8qnX6gQ2E3J5pQ+iCiL/ybb5GDQ4AAAAAAAAGQQ==. Passphrase: xsCp8b58fkboDXHmFp4Y. K6R10/K0R0 (starvation vs fed)-F2-larval fat body: +b3vmKtndEZGjKTV1AbekqMLZQFxtfj9CPA5IcS8fM55jXwENCbskP8DA+zr8ZYjawhar+4gnoAwoByWMxxE6C3d7GwAAAAAAAAK5Q==. Passphrase: FyRGGAYOCZlK0d1F1VXT. 2147

dx.doi.org/10.1021/pr301168x | J. Proteome Res. 2013, 12, 2138−2150

Journal of Proteome Research

■

Article

(11) Fredens, J.; Engholm-Keller, K.; Giessing, A.; Pultz, D.; Larsen, M. R.; Hojrup, P.; Moller-Jensen, J.; Faergeman, N. J. Quantitative proteomics by amino acid labeling in C. elegans. Nat. Methods 2011, 8 (10), 845−7. (12) Larance, M.; Bailly, A. P.; Pourkarimi, E.; Hay, R. T.; Buchanan, G.; Coulthurst, S.; Xirodimas, D. P.; Gartner, A.; Lamond, A. I. Stableisotope labeling with amino acids in nematodes. Nat. Methods 2011, 8 (10), 849−51. (13) Bendall, S. C.; Hughes, C.; Stewart, M. H.; Doble, B.; Bhatia, M.; Lajoie, G. A. Prevention of amino acid conversion in SILAC experiments with embryonic stem cells. Mol. Cell. Proteomics 2008, 7 (9), 1587−97. (14) Blagoev, B.; Mann, M. Quantitative proteomics to study mitogenactivated protein kinases. Methods 2006, 40 (3), 243−50. (15) Bicho, C. C.; de Lima Alves, F.; Chen, Z. A.; Rappsilber, J.; Sawin, K. E. A genetic engineering solution to the “arginine conversion problem” in stable isotope labeling by amino acids in cell culture (SILAC). Mol. Cell. Proteomics 2010, 9 (7), 1567−77. (16) Van Hoof, D.; Pinkse, M. W.; Oostwaard, D. W.; Mummery, C. L.; Heck, A. J.; Krijgsveld, J. An experimental correction for arginine-toproline conversion artifacts in SILAC-based quantitative proteomics. Nat. Methods 2007, 4 (9), 677−8. (17) Park, S. K.; Liao, L.; Kim, J. Y.; Yates, J. R., 3rd A computational approach to correct arginine-to-proline conversion in quantitative proteomics. Nat. Methods 2009, 6 (3), 184−5. (18) Brunner, E.; Ahrens, C. H.; Mohanty, S.; Baetschmann, H.; Loevenich, S.; Potthast, F.; Deutsch, E. W.; Panse, C.; de Lichtenberg, U.; Rinner, O.; Lee, H.; Pedrioli, P. G.; Malmstrom, J.; Koehler, K.; Schrimpf, S.; Krijgsveld, J.; Kregenow, F.; Heck, A. J.; Hafen, E.; Schlapbach, R.; Aebersold, R. A high-quality catalog of the Drosophila melanogaster proteome. Nat. Biotechnol. 2007, 25 (5), 576−83. (19) Zhai, B.; Villen, J.; Beausoleil, S. A.; Mintseris, J.; Gygi, S. P. Phosphoproteome analysis of Drosophila melanogaster embryos. J. Proteome Res. 2008, 7 (4), 1675−82. (20) Bodenmiller, B.; Malmstrom, J.; Gerrits, B.; Campbell, D.; Lam, H.; Schmidt, A.; Rinner, O.; Mueller, L. N.; Shannon, P. T.; Pedrioli, P. G.; Panse, C.; Lee, H. K.; Schlapbach, R.; Aebersold, R. PhosphoPepa phosphoproteome resource for systems biology research in Drosophila Kc167 cells. Mol. Syst. Biol. 2007, 3, 139. (21) Goetze, S.; Qeli, E.; Mosimann, C.; Staes, A.; Gerrits, B.; Roschitzki, B.; Mohanty, S.; Niederer, E. M.; Laczko, E.; Timmerman, E.; Lange, V.; Hafen, E.; Aebersold, R.; Vandekerckhove, J.; Basler, K.; Ahrens, C. H.; Gevaert, K.; Brunner, E. Identification and functional characterization of N-terminally acetylated proteins in Drosophila melanogaster. PLoS Biol. 2009, 7 (11), e1000236. (22) Tan, D. J.; Dvinge, H.; Christoforou, A.; Bertone, P.; Martinez Arias, A.; Lilley, K. S. Mapping organelle proteins and protein complexes in Drosophila melanogaster. J. Proteome Res. 2009, 8 (6), 2667−78. (23) Cammarato, A.; Ahrens, C. H.; Alayari, N. N.; Qeli, E.; Rucker, J.; Reedy, M. C.; Zmasek, C. M.; Gucek, M.; Cole, R. N.; Van Eyk, J. E.; Bodmer, R.; O’Rourke, B.; Bernstein, S. I.; Foster, D. B. A mighty small heart: the cardiac proteome of adult Drosophila melanogaster. PLoS One 2011, 6 (4), e18497. (24) Chang, Y. C.; Lin, S. Y.; Liang, S. Y.; Pan, K. T.; Chou, C. C.; Chen, C. H.; Liao, C. L.; Khoo, K. H.; Meng, T. C. Tyrosine phosphoproteomics and identification of substrates of protein tyrosine phosphatase dPTP61F in Drosophila S2 cells by mass spectrometrybased substrate trapping strategy. J. Proteome Res. 2008, 7 (3), 1055−66. (25) Xu, P.; Tan, H.; Duong, D. M.; Yang, Y.; Kupsco, J.; Moberg, K. H.; Li, H.; Jin, P.; Peng, J. Stable isotope labeling with amino acids in Drosophila for quantifying proteins and modifications. J. Proteome Res. 2012, 11 (9), 4403−12. (26) Cox, J.; Neuhauser, N.; Michalski, A.; Scheltema, R. A.; Olsen, J. V.; Mann, M. Andromeda: a peptide search engine integrated into the MaxQuant environment. J. Proteome Res. 2011, 10 (4), 1794−805. (27) White, K. P.; Rifkin, S. A.; Hurban, P.; Hogness, D. S. Microarray analysis of Drosophila development during metamorphosis. Science 1999, 286 (5447), 2179−84.

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected] (K.-H.K.); gcchen@gate. sinica.edu.tw (G.-C.C.). Fax: 886-2-27889759. Tel: 88627855696. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We would like to thank Dr. Ting-Fang Wang at Institute of Molecular Biology, Academia Sinica, Taipei, Taiwan, for providing the yeast strain. This work was supported by Academia Sinica (to K.-H.K. and G.-C.C.) and the National Science Council of Taiwan (NSC99-2311-B-001-017-MY3 to G.-C.C.) and was carried out as part of the Research and Development program at the previous NRPGM Core Facilities for Proteomics and Glycomics (NSC 99-3112-B-001-025; 98-3112-B-001-023), and current Core Facilities for Protein Structural Analysis at Academia Sinica, supported under the Taiwan National Core Facility Program for Biotechnology, NSC 100-2325-B-001-029.

■

ABBREVIATIONS SILAC, stable isotope labeling by amino acids in cell culture; Lys0, [12C6,14N2]-lysine; Lys4, [2H4]-lysine; Lys6, [13C6]-lysine; Lys8, [13C6,15N2]-lysine; Arg0, [12C6,14N4]-arginine; Arg10, [13C6,15N4]-arginine; Lys-C, lysyl endoprotease; Arg-C, endoprotease Arg-C; PF, puparium formation; APF, after purparium formation; ADE, after egg deposition

■

REFERENCES

(1) Ong, S. E.; Blagoev, B.; Kratchmarova, I.; Kristensen, D. B.; Steen, H.; Pandey, A.; Mann, M. Stable isotope labeling by amino acids in cell culture, SILAC, as a simple and accurate approach to expression proteomics. Mol. Cell. Proteomics 2002, 1 (5), 376−86. (2) Oda, Y.; Huang, K.; Cross, F. R.; Cowburn, D.; Chait, B. T. Accurate quantitation of protein expression and site-specific phosphorylation. Proc. Natl. Acad. Sci. U. S. A. 1999, 96 (12), 6591−6. (3) Cox, J.; Mann, M. MaxQuant enables high peptide identification rates, individualized p.p.b.-range mass accuracies and proteome-wide protein quantification. Nat. Biotechnol. 2008, 26 (12), 1367−72. (4) Gouw, J. W.; Tops, B. B.; Mortensen, P.; Heck, A. J.; Krijgsveld, J. Optimizing identification and quantitation of 15N-labeled proteins in comparative proteomics. Anal. Chem. 2008, 80 (20), 7796−803. (5) Ong, S. E.; Mann, M. A practical recipe for stable isotope labeling by amino acids in cell culture (SILAC). Nat. Protoc. 2006, 1 (6), 2650− 60. (6) Dilworth, D. J.; Saleem, R. A.; Rogers, R. S.; Mirzaei, H.; Boyle, J.; Aitchison, J. D. QTIPS: a novel method of unsupervised determination of isotopic amino acid distribution in SILAC experiments. J. Am. Soc. Mass Spectrom. 2010, 21 (8), 1417−22. (7) Gruhler, A.; Olsen, J. V.; Mohammed, S.; Mortensen, P.; Faergeman, N. J.; Mann, M.; Jensen, O. N. Quantitative phosphoproteomics applied to the yeast pheromone signaling pathway. Mol. Cell. Proteomics 2005, 4 (3), 310−27. (8) de Godoy, L. M.; Olsen, J. V.; Cox, J.; Nielsen, M. L.; Hubner, N. C.; Frohlich, F.; Walther, T. C.; Mann, M. Comprehensive massspectrometry-based proteome quantification of haploid versus diploid yeast. Nature 2008, 455 (7217), 1251−4. (9) Kruger, M.; Moser, M.; Ussar, S.; Thievessen, I.; Luber, C. A.; Forner, F.; Schmidt, S.; Zanivan, S.; Fassler, R.; Mann, M. SILAC mouse for quantitative proteomics uncovers kindlin-3 as an essential factor for red blood cell function. Cell 2008, 134 (2), 353−64. (10) Sury, M. D.; Chen, J. X.; Selbach, M. The SILAC fly allows for accurate protein quantification in vivo. Mol. Cell. Proteomics 2010, 9 (10), 2173−83. 2148

dx.doi.org/10.1021/pr301168x | J. Proteome Res. 2013, 12, 2138−2150

Journal of Proteome Research

Article

in rat hepatoma cell line, H-4-II-E cells. Cell Struct. Funct. 1998, 23 (1), 33−42. (50) Klionsky, D. J.; Elazar, Z.; Seglen, P. O.; Rubinsztein, D. C. Does bafilomycin A1 block the fusion of autophagosomes with lysosomes? Autophagy 2008, 4 (7), 849−950. (51) Tang, H. W.; Wang, Y. B.; Wang, S. L.; Wu, M. H.; Lin, S. Y.; Chen, G. C. Atg1-mediated myosin II activation regulates autophagosome formation during starvation-induced autophagy. EMBO J. 2011, 30 (4), 636−51. (52) Ong, S. E.; Kratchmarova, I.; Mann, M. Properties of 13Csubstituted arginine in stable isotope labeling by amino acids in cell culture (SILAC). J. Proteome Res. 2003, 2 (2), 173−81. (53) Bonaldi, T.; Straub, T.; Cox, J.; Kumar, C.; Becker, P. B.; Mann, M. Combined use of RNAi and quantitative proteomics to study gene function in Drosophila. Mol. Cell 2008, 31 (5), 762−72. (54) Hilger, M.; Bonaldi, T.; Gnad, F.; Mann, M. Systems-wide analysis of a phosphatase knock-down by quantitative proteomics and phosphoproteomics. Mol. Cell. Proteomics 2009, 8 (8), 1908−20. (55) Zinke, I.; Kirchner, C.; Chao, L. C.; Tetzlaff, M. T.; Pankratz, M. J. Suppression of food intake and growth by amino acids in Drosophila: the role of pumpless, a fat body expressed gene with homology to vertebrate glycine cleavage system. Development 1999, 126 (23), 5275− 84. (56) Perrimon, N.; Ni, J. Q.; Perkins, L. In vivo RNAi: today and tomorrow. Cold Spring Harbor Perspect. Biol. 2010, 2 (8), a003640. (57) Cox, J.; Matic, I.; Hilger, M.; Nagaraj, N.; Selbach, M.; Olsen, J. V.; Mann, M. A practical guide to the MaxQuant computational platform for SILAC-based quantitative proteomics. Nat. Protoc. 2009, 4 (5), 698− 705. (58) Olsen, J. V.; Blagoev, B.; Gnad, F.; Macek, B.; Kumar, C.; Mortensen, P.; Mann, M. Global, in vivo, and site-specific phosphorylation dynamics in signaling networks. Cell 2006, 127 (3), 635−48. (59) Hilger, M.; Mann, M. Triple SILAC to determine stimulus specific interactions in the Wnt pathway. J. Proteome Res. 2012, 11 (2), 982−94. (60) Cambridge, S. B.; Gnad, F.; Nguyen, C.; Bermejo, J. L.; Kruger, M.; Mann, M. Systems-wide proteomic analysis in mammalian cells reveals conserved, functional protein turnover. J. Proteome Res. 2011, 10 (12), 5275−84. (61) Dengjel, J.; Hoyer-Hansen, M.; Nielsen, M. O.; Eisenberg, T.; Harder, L. M.; Schandorff, S.; Farkas, T.; Kirkegaard, T.; Becker, A. C.; Schroeder, S.; Vanselow, K.; Lundberg, E.; Nielsen, M. M.; Kristensen, A. R.; Akimov, V.; Bunkenborg, J.; Madeo, F.; Jaattela, M.; Andersen, J. S. Identification of autophagosome-associated proteins and regulators by quantitative proteomic analysis and genetic screens. Mol. Cell. Proteomics 2012, 11 (3), M111 014035. (62) Henningsen, J.; Rigbolt, K. T.; Blagoev, B.; Pedersen, B. K.; Kratchmarova, I. Dynamics of the skeletal muscle secretome during myoblast differentiation. Mol. Cell. Proteomics 2010, 9 (11), 2482−96. (63) Schreiber, T. B.; Mausbacher, N.; Keri, G.; Cox, J.; Daub, H. An integrated phosphoproteomics work flow reveals extensive network regulation in early lysophosphatidic acid signaling. Mol. Cell. Proteomics 2010, 9 (6), 1047−62. (64) Gross, P. R.; Spindel, W. The inhibition of mitosis by deuterium. Ann. N. Y. Acad. Sci. 1960, 84, 745−54. (65) Gross, P. R.; Spindel, W. Mitotic arrest by deuterium oxide. Science 1960, 131, 37−8. (66) Mitchell, W. M.; Harrington, W. F. Purification and properties of clostridiopeptidase B (Clostripain). J. Biol. Chem. 1968, 243 (18), 4683−92. (67) Mortensen, P.; Gouw, J. W.; Olsen, J. V.; Ong, S. E.; Rigbolt, K. T.; Bunkenborg, J.; Cox, J.; Foster, L. J.; Heck, A. J.; Blagoev, B.; Andersen, J. S.; Mann, M. MSQuant, an open source platform for mass spectrometrybased quantitative proteomics. J. Proteome Res. 2010, 9 (1), 393−403. (68) Li, X. J.; Zhang, H.; Ranish, J. A.; Aebersold, R. Automated statistical analysis of protein abundance ratios from data generated by stable-isotope dilution and tandem mass spectrometry. Anal. Chem. 2003, 75 (23), 6648−57.

(28) Beckstead, R. B.; Lam, G.; Thummel, C. S. The genomic response to 20-hydroxyecdysone at the onset of Drosophila metamorphosis. Genome Biol. 2005, 6 (12), R99. (29) Chihara, C. J.; Silvert, D. J.; Fristrom, J. W. The cuticle proteins of Drosophila melanogaster: stage specificity. Dev. Biol. 1982, 89 (2), 379− 88. (30) Wright, L. G.; Chen, T.; Thummel, C. S.; Guild, G. M. Molecular characterization of the 71E late puff in Drosophila melanogaster reveals a family of novel genes. J. Mol. Biol. 1996, 255 (3), 387−400. (31) Natzle, J. E. Temporal regulation of Drosophila imaginal disc morphogenesis: a hierarchy of primary and secondary 20-hydroxyecdysone-responsive loci. Dev. Biol. 1993, 155 (2), 516−32. (32) Natzle, J. E.; Robertson, J. P.; Majumdar, A.; Vesenka, G. D.; Enlow, B.; Clark, K. E. Sequence and expression of IMP-L1, an ecdysone-inducible gene expressed during Drosophila imaginal disc morphogenesis. Dev. Genet. 1992, 13 (5), 331−44. (33) Crowley, T. E.; Meyerowitz, E. M. Steroid regulation of RNAs transcribed from the Drosophila 68c polytene chromosome puff. Dev. Biol. 1984, 102 (1), 110−21. (34) Hansson, L.; Lambertsson, A. Steroid regulation of glue protein genes in Drosophila melanogaster. Hereditas 1989, 110 (1), 61−7. (35) Snyder, M.; Hirsh, J.; Davidson, N. The cuticle genes of drosophila: a developmentally regulated gene cluster. Cell 1981, 25 (1), 165−77. (36) Stathakis, D. G.; Burton, D. Y.; McIvor, W. E.; Krishnakumar, S.; Wright, T. R.; O’Donnell, J. M. The catecholamines up (Catsup) protein of Drosophila melanogaster functions as a negative regulator of tyrosine hydroxylase activity. Genetics 1999, 153 (1), 361−82. (37) Wright, T. R. The genetics of biogenic amine metabolism, sclerotization, and melanization in Drosophila melanogaster. Adv. Genet. 1987, 24, 127−222. (38) Davis, M. M.; O’Keefe, S. L.; Primrose, D. A.; Hodgetts, R. B. A neuropeptide hormone cascade controls the precise onset of posteclosion cuticular tanning in Drosophila melanogaster. Development 2007, 134 (24), 4395−404. (39) Chintapalli, V. R.; Wang, J.; Dow, J. A. Using FlyAtlas to identify better Drosophila melanogaster models of human disease. Nat. Genet. 2007, 39 (6), 715−20. (40) Massey, H. C., Jr.; Kejzlarova-Lepesant, J.; Willis, R. L.; Castleberry, A. B.; Benes, H. The Drosophila Lsp-1 beta gene. A structural and phylogenetic analysis. Eur. J. Biochem. 1997, 245 (1), 199−207. (41) Zinke, I.; Schutz, C. S.; Katzenberger, J. D.; Bauer, M.; Pankratz, M. J. Nutrient control of gene expression in Drosophila: microarray analysis of starvation and sugar-dependent response. EMBO J. 2002, 21 (22), 6162−73. (42) Kong, E. C.; Allouche, L.; Chapot, P. A.; Vranizan, K.; Moore, M. S.; Heberlein, U.; Wolf, F. W. Ethanol-regulated genes that contribute to ethanol sensitivity and rapid tolerance in Drosophila. Alcohol.: Clin. Exp. Res. 2010, 34 (2), 302−16. (43) Hanson, R. W.; Garber, A. J. Phosphoenolpyruvate carboxykinase. I. Its role in gluconeogenesis. Am. J. Clin. Nutr. 1972, 25 (10), 1010−21. (44) Rognstad, R. Rate-limiting steps in metabolic pathways. J. Biol. Chem. 1979, 254 (6), 1875−8. (45) Newsholme, E. A.; Williams, T. The role of phosphoenolpyruvate carboxykinase in amino acid metabolism in muscle. Biochem. J. 1978, 176 (2), 623−6. (46) Snell, K.; Duff, D. A. The release of alanine by rat diaphragm muscle in vitro. Biochem. J. 1977, 162 (2), 399−403. (47) Mijaljica, D.; Prescott, M.; Devenish, R. J. V-ATPase engagement in autophagic processes. Autophagy 2011, 7 (6), 666−8. (48) Hiesinger, P. R.; Fayyazuddin, A.; Mehta, S. Q.; Rosenmund, T.; Schulze, K. L.; Zhai, R. G.; Verstreken, P.; Cao, Y.; Zhou, Y.; Kunz, J.; Bellen, H. J. The v-ATPase V0 subunit a1 is required for a late step in synaptic vesicle exocytosis in Drosophila. Cell 2005, 121 (4), 607−20. (49) Yamamoto, A.; Tagawa, Y.; Yoshimori, T.; Moriyama, Y.; Masaki, R.; Tashiro, Y. Bafilomycin A1 prevents maturation of autophagic vacuoles by inhibiting fusion between autophagosomes and lysosomes 2149

dx.doi.org/10.1021/pr301168x | J. Proteome Res. 2013, 12, 2138−2150

Journal of Proteome Research

Article

(69) MacCoss, M. J.; Wu, C. C.; Liu, H.; Sadygov, R.; Yates, J. R., 3rd A correlation algorithm for the automated quantitative analysis of shotgun proteomics data. Anal. Chem. 2003, 75 (24), 6912−21. (70) He, C.; Klionsky, D. J. Regulation mechanisms and signaling pathways of autophagy. Annu. Rev. Genet. 2009, 43, 67−93. (71) Mindell, J. A. Lysosomal acidification mechanisms. Annu. Rev. Physiol. 2012, 74, 69−86. (72) Wada, Y.; Sun-Wada, G. H.; Tabata, H.; Kawamura, N. Vacuolartype proton ATPase as regulator of membrane dynamics in multicellular organisms. J. Bioenerg. Biomembr. 2008, 40 (1), 53−7.

2150

dx.doi.org/10.1021/pr301168x | J. Proteome Res. 2013, 12, 2138−2150