|Home | About | Journals | Submit | Contact Us | Français|
Stable incorporation of labeled amino acids in cell culture is a simple approach to label proteins in vivo for mass spectrometric quantification. Full incorporation of isotopically heavy amino acids facilitates accurate quantification of proteins from different cultures, yet analysis methods for determination of incorporation are cumbersome and time-consuming. We present QTIPS, Quantification by Total Identified Peptides for SILAC, a straightforward, accurate method to determine the level of heavy amino acid incorporation throughout a population of peptides detected by mass spectrometry. Using QTIPS, we show that the incorporation of heavy amino acids in baker’s yeast is unaffected by the use of prototrophic strains, indicating that auxotrophy is not a requirement for SILAC experiments in this organism. This method has general utility for multiple applications where isotopic labeling is used for quantification in mass spectrometry.
Quantitative mass spectrometry analyses have revolutionized the ability to interrogate the proteome. The stable incorporation of labeled amino acids in cell culture (SILAC)  is a common method for quantitatively assaying proteomes in many systems and identifying functionally relevant, condition-specific sub-proteomes [2–4]. The SILAC method utilizes cells cultured in the presence of isotopically normal or heavy amino acids enabling relative quantitative proteomics of multiple experimental conditions.
While the use of SILAC labeling to quantify proteomic differences between samples does not necessitate the complete incorporation of heavy labels, incomplete incorporation confounds downstream analysis. Long-term growth in SILAC media would help ensure complete incorporation, but it is not economically practical and is problematic because of time dependent arginine-to-proline interconversion [5–7], which is readily detected in yeast after long-term culturing in SILAC media . Efficient label incorporation and the total growth time in SILAC media are therefore, important considerations that must be balanced to achieve the greatest label efficiency in the minimum amount of time.
Several methods of determination of SILAC incorporation have been described [1, 6, 9, 10]. Often, estimates of incorporation are based on manually curated data for a limited number of proteins and/or peptides, which may introduce artifacts resulting from variable protein turnover rates [11, 12]. We hypothesized that the global level of SILAC incorporation could be directly and accurately inferred by comparing the number of heavy peptide identifications to all peptide identifications, both heavy and light, in a given protein sample for a given heavy labeled amino acid. Such a method would be advantageous as it would require only routine shotgun MS/MS analysis and obviate the need for area under the curve (AUC) analysis of the MS1 trace. We used the technique described herein to test parameters in the yeast Saccharomyces cerevisiae, exploring the number of doubling times required for incorporation of isotopically heavy amino acids and the effect of intact versus disabled amino acid biosynthesis pathways.
All experiments were performed using the Saccharomyces cerevisiae strain BY4741 (MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0) and a previously described derivative of BY4742 auxotrophic for both lysine and arginine . Complete synthetic growth media, CSM, [20 g/L D-glucose (VWR), 0.67 g/L Complete Supplement Mixture minus arginine, lysine and histidine (Sunrise Science Products), 1.7 g/L Difco yeast nitrogen base w/o amino acids and ammonium sulfate (BD), 5 g/L ammonium sulfate (Fisher), 30 mg/L L-Histidine (J.T.Baker), 45 mg/L L-Lysine (Fisher) and 50 mg/L L-Arginine (Fisher)] was used for standard, light-labeled preparations. Heavy preparations were obtained by growth in SILAC-KR complete synthetic media (identical to standard CSM formulation above, except that L-Lysine-13C6, 15N2·HCl (Isotec) and L-Arginine-13C6, 15N4·HCl (Isotec) were included in place of their isotopically light variants. All samples were obtained from log phase cultures grown at 30 °C with shaking at 200 rpm at the indicated generation number based on the empirically derived 105 minute doubling time. For each sample, approximately 1 X 108 cells were harvested in 50 mL conical tubes by centrifugation at 7,000·g for 5 minutes, washed with 1 mL of sterile water in a 1.7 mL tube, centrifuged for 15 seconds at 10,000·g and the cell pellet was frozen in liquid nitrogen and stored at −80 °C.
Cell pellets were resuspended in 500 μL of 100 mM ammonium bicarbonate, pH 9, and 300 μL of acid-washed glass beads (425–600 μm, Sigma) were added. Tubes were mixed vigorously for 10 minutes at 4 °C and cellular debris was removed by centrifugation at 15,000·g for 5 minutes at 4 °C. The supernatants were transferred to fresh 1.7 mL centrifuge tubes and the protein concentration was determined for each lysate using the Warburg-Christian equation: protein concentration (mg/mL) = 1.55·A280 − 0.76·A260. We have performed QTIPS analyses on data obtained by shotgun MS/MS using as little as 5 μg of total yeast cell lysate. In general, 50 μg of protein was brought to 100 μL with 100 mM ammonium bicarbonate, pH 9. The sample was reduced with tris(2-carboxyethyl)phosphine, alkylated with iodoacetamide and then digested with sequencing grade modified porcine trypsin (Promega). Digested samples were desalted using Vydac C18 Silica MicroSpin columns (The Nest Group) by manufacturer’s instructions. Purified peptides were diluted in load buffer (0.1% Formic acid, v/v) prior to MS/MS analysis.
Mass spectrometry was performed on Thermo Electron LTQ and LTQ-Orbitrap mass spectrometers equipped with an electrospray ionization source and an Agilent HP1100 liquid chromatography system. The gradient was run from 10% to 35% acetonitrile over 60 minutes at a constant flow rate of 0.350 μl/min. XCalibur Raw files were converted to mzXML format using ReAdW (v4.2.0) using the readw profile and default parameters. Spectral searches were done for tryptic fragments using X!Tandem  (v2007.07.01.3 with the k-score plugin ). Searches were done against a non-redundant S. cerevisiae reference protein database (the union of the SGD, Ensembl, NCI, and GenBank databases, plus keratin and trypsin) containing 13616 entries. Parent tolerance was ± 3 Da, the fragment mass tolerance was ± 0.4 Da, with no missed cleavages. A wide mass window was used to increase peptide identifications . Modifications included in the searches are as follows: static modification of cysteine 57.021464 Da (C[143.13]), and variable modifications of methionine 15.994915 Da M[147.19], glutamine -17.0306 Da Q[111.10], glutamic acid -18.01056 Da E[111.10], lysine 8.014199 Da, K[136.19] and arginine 10.008269 Da R[166.20].
Individual search results were processed using the Trans Proteomic Pipeline . Search result validation was done using Peptide Prophet (v3.0) , and relative quantification of isotopically labeled peptides was reported by ASAP  and Xpress , both of which integrate the AUC of an ion chromatogram. Scores of -1, 0 and 999 (no information) were excluded from the final distribution analyses. A probability of 0.9 was used as the cutoff for peptides. This cutoff returned a total of 17,318 spectra containing 124 decoy peptides for an overall false discovery rate of 0.7% for peptides.
Tab-delimited Peptide Prophet text files were imported into the QTIPS application (available at: http://qtips.systemsbiology.net), which identifies informative peptides and calculates the frequency of heavy arginine or heavy lysine containing peptides in the population of informative peptides. Informative peptides are those that contain either a single K or a single R, not both, that also meet user defined Peptide Prophet probability cut-offs (default 0.9, 90% confidence), thresholding based on the total number of peptide identifications for each protein (default 0, no threshold) and peptide ID search limits (default set to analyze yeast proteins only). The QTIPS application also determines the average Xpress and ASAP percent incorporation values for informative peptides and presents histograms of their distributions. Heat maps of ASAP and Xpress distributions were generated using MultiExperiment Viewer v4.4.
The relationship between global heavy amino acid incorporation and the relative number of heavy labeled peptides identified by MS/MS was investigated as an unbiased means to quantify heavy amino acid incorporation levels in biological samples. As proof of principle for this method, termed QTIPS (Quantification by Total Identified Peptides for SILAC), we performed a simple mixing experiment in which protein samples of known amounts of heavy arginine and lysine incorporation were analyzed and compared to the expected level of incorporation. Fully heavy labeled (10 generations in SILAC-KR media) and fully light labeled (10 generations in CSM) lysates were prepared and, from these, samples with heavy amino acid incorporation levels of 0%, 25%, 50%, 75%, 95% and 100% were prepared and analyzed by MS/MS. The results, (Figure 1), demonstrate the utility of the QTIPS method to accurately predict the level of heavy lysine and arginine in a protein sample over a wide range of incorporation. Importantly, an analysis of the same data using two established alternate methods, ASAP  and Xpress, clearly demonstrate their limitations when applied without manual curation to complex protein mixtures wherein one isotopic variant predominates (see also Table 1). ASAP and Xpress integrate AUC of MS1 traces for relative quantification. Both perform well within a 2-fold relative abundance range, but become inaccurate at low or high heavy amino acid incorporation levels.
Most commonly, yeast SILAC experiments utilizing heavy arginine and lysine are performed in strains auxotrophic for these amino acids, eliminating the cell’s inherent ability to synthesize light counterparts. This has the potential to be a major hurdle, given that many widely used laboratory strains and libraries are not auxotrophic for one or both of these amino acids; therefore, we used QTIPS to determine if efficient heavy amino acid incorporation in yeast requires auxotrophy.
We hypothesized that, in strains with intact lysine and arginine amino acid biosynthesis pathways, cells would preferentially utilize available free heavy lysine and arginine rather than spend resources on amino acid synthesis. To test this, the incorporation of heavy lysine and arginine was monitored by the QTIPS method in both LYS+ARG+ and lys-arg- strains through nine generations. In agreement with previous findings [1, 6, 9, 10], near complete incorporation was achieved within seven generations for the lys-arg- strain and heavy amino acid incorporation in the prototrophic (LYS+ARG+) strain closely mirrored that observed in the dually auxotrophic background (Figure 2, Table 1). In both cases, full incorporation was achieved within seven generations. Relative to heavy lysine, there was no bias in the incorporation of heavy arginine and, therefore, it is unlikely that arginine-to-proline interconversion is a significant source of error at this number of doublings. Notably, all samples trend along the predicted incorporation based on population doubling, which suggests that the number of generations in SILAC media is the most important factor in determining incorporation levels in yeast. As with mixing experiments, we find QTIPS to be superior to unsupervised ASAP and Xpress in these complex protein mixtures (see also Table 1).
The QTIPS method that we have described is a simple, robust means of determining heavy isotope coded amino acid incorporation. It does not require manual curation of MS quantification data and can be applied to both simple and complex protein mixtures. The accuracy and dynamic range of measurements by QTIPS depend directly on sample size (the total number of peptide identifications). Importantly, because the method includes all detected proteins, there is less likelihood that bias is introduced by variation in protein turnover rates, which can be an issue for MS1 quantification methods in which only a limited number of peptides and/or proteins are interrogated. We have focused our initial characterization of QTIPS using baker’s yeast, but this methodology is applicable to any other experimentally manipulable biological system. In yeast, the incorporation of heavy amino acids is highly efficient in strains with intact amino acid biosynthesis pathways, making the method suitable for quantitative proteomics without the need for additional genetic manipulations. The potential of isotopic pulse-chase to estimate turnover of individual proteins on a global scale has been demonstrated , but is limited by the time required for manual data curation. The combination of QTIPS with large MS/MS data sets allows quantification at single protein resolution, and is widely applicable to methodologies exploiting isotopic labeling for quantitative mass spectrometry-based proteomics.
Traditional MS1-based area under the curve methods of isotope quantification, such as ASAP and Xpress ratios are computationally demanding and require manual curation of output ratios. In contrast, QTIPS utilizes MS2-based peptide identifications to quickly determine relative isotope abundance.
This work was supported by grants GM075152, GMO76547 and RR022220 from the U.S. National Institutes of Health.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.