Stable Isotope Labeling with Amino Acids in Cell Culture (SILAC) is a metabolic labeling technique for mass spectrometric (MS)-based quantitative proteomics. It was developed in the Center for Experimental Bioinformatics (CEBI) at the University of Southern Denmark (Ong SE, et al, MCP, 2002). In SILAC, differentially labeled samples are mixed early in the experimental process, and analyzed together by LC-MS/MS. Since the labeling does not affect the chemical properties of the molecules, they co-elute from the LC column and analyzed together in the mass spectrometer. The peptide peaks of the differentially labeled samples can be very accurately quantified relative to each other to determine the peptide and protein ratios.
Typical SILAC labeling uses lysine and arginine, which in combination with trypsin digestion results in labeling of every peptide in the mixture (except for the c-terminal peptide of the protein). SILAC experiments involve differential labeling of two to three cell types: cells grown with the natural amino acids, with 2H4-lysine and 13C6-arginine and with 15N213C6-lysine and 15N413C6-arginine. The SILAC method relies on the complete incorporation of heavy amino acids during protein turnover. Dialyzed serum is employed so that the added amino acids are the exclusive source. In contrast to labeling with heavy nitrogen or carbon, the heavy amino acids create a distinct and known mass difference between the samples, and only affect the isotope distribution of the peptides in a minor and predictable way, thereby making data interpretation and quantification more accurate and robust. Peptides are quantified from the MS scans allowing multiple scans for each peptide across its elution time. Furthermore, protein quantification is usually based on the median values of multiple peptides, resulting in high accuracy of ratio determination. While classical SILAC relies on full incorporation of amino acids into proteins in cultured cells, SILAC has expanded to whole organisms and to human tissues, by using SILAC as a ‘spike-in’ standard.
Recipe of SILAC medium
- SILAC medium- medium that is deprived of lysine and arginine. Commercial vendors are: Invitrogen, PAA, Sigma and Pierce.
- Dialyzed serum- Fetal bovine serum dialyzed with a 10 kDa cutoff. Commercial vendor: Invitrogen.
- SILAC amino acids- double labeling experiments use the natural amino acids and 15N213C6-lysine (Lys8) and 15N413C6-arginine (Arg10). Triple labeling experiments use additionally 2H4-lysine (Lys4) and 13C6-arginine (Arg6). Commercial vendors: Silantes, Sigma and Cambridge Isotope Laboratories.
- Penicillin and streptomycin. Should be used in the standard concentrations.
For the preparation of 500 ml of medium add 50 ml of dialyzed serum, antibiotics and amino acids to 450 ml of SILAC medium. The final concentration of the amino acids should be 73 mg per liter for lysine and 42 mg per liter for arginine. The concentration should be titrated for each cell line according to the incorporation efficiency.
Cells are cultured in SILAC medium according to the standard protocols. Every time the cells are sub-cultured with trypsin, they should be centrifuged to eliminate the trypsin, since residual trypsin might reduce incorporation efficiency. After 5-10 cell doublings they usually reach complete incorporation. If a cell type or cell line is labeled for the first time, or when new batches of amino acids are used, it is essential to examine the efficiency of incorporation of the heavy amino acids. Note that some cell lines may convert arginine to proline and vice versa, depending on the arginine concentrations. Arginine concentration has to be titrated to minimize such conversion.
Examination of SILAC labeling efficiency
For incorporation testing a small aliquot of cells should be lysed and digested using standard protocols. Protein amount of 10-50 g is sufficient for the test. After digestion peptides can be purified on StageTips (Rappsilber J, et al, Analytical chemistry, 2003) and analyzed with a single LC-MS/MS run. MaxQuant analysis of the raw file will determine the ratio of heavy labeled peptides to the remaining non-labeled ones. The incorporation efficiency can be calculated as: (1-1/Ratio(H/L)) on the peptide level. Since arginine and lysine may have different labeling efficiencies, it is recommended to perform the calculation separately for lysine- and for arginine-containing peptides. For proper SILAC experiments it is necessary to obtain labeling of more than 95%. To examine arginine to proline conversion, heavy proline can be added as a variable modification in the MaxQuant analysis.
Classical SILAC experiments
In SILAC, 2-3 samples are differentially labeled and directly compared in the MS analysis. Equal amounts of cells or cell lysates of the different samples are combined and treated as a single sample in all subsequent steps. This enables the use of any method of protein or peptide enrichment without introducing quantification errors into the final quantitative analysis. Typical SILAC experiments compare two or three samples, but it is also possible to compare five samples by combining two or more triple-SILAC experiments that have at least one common sample. This is frequently done in time course experiments. In the case of double-SILAC, MaxQuant determines the peptide ratios of the ‘heavy’ to ‘light’, and for triple-SILAC it determines the ratio of ‘heavy’ to ‘medium’, ‘medium’ to ‘light’ and ‘heavy’ to ‘light’. The median peptide ratios are then taken as the protein ratio.
Spike-in SILAC standard
As an alternative to the classical SILAC approach, labeled samples can be used as an internal standard for quantification. In such cases, the SILAC sample is produced separately and is spiked into each of the experimental samples, which are then processed and analyzed together. Importantly, the experimental samples are prepared without any restriction on experimental design and in particular, normal media and amino acids are used. The samples are then analyzed with the experimental sample considered ‘light’ and the spike-in standard as ‘heavy’. Peptide ratios between sample and the heavy spike-in standard are determined for each of the samples. Since the standard is identical in all cases, the peptide fold change ratio between the samples is then the ‘ratio of ratios’. Spike-in SILAC can be applied to cultured cells, to primary cultures and to tissues from various origins.
SILAC labeling of whole organisms
Application of SILAC to whole organisms is valuable when in-vivo experiments are performed that analyze whole tissues rather than cultured cells. SILAC labeling of whole organisms requires preparation of food that contains the SILAC amino acid as the sole source and that is compatible with growth of the organism. Full incorporation requires the turnover of the organism’s entire proteome, which can be achieved by feeding them for more than one generation with food that contains heavy amino acids in place of normal ones. These then represent the real complexity of the organism and the specific organs and tissues of interest, including extracellular proteins and body fluids. Since arginine is not an essential amino acid in some tissues, most model organisms are labeled only with lysine. Drosophila can be labeled by feeding the flies with heavy labeled yeast. SILAC-labeled mice have been generated by feeding them with 13C6-Lys containing diet. After two generations they are fully labeled and can serve as a spike-in standard for the analysis of mouse tissues and body fluids. After establishing a mouse colony, it can be maintained by exclusively feeding the mice with SILAC food. We recommend using the SILAC model organisms as standards rather than as the experimental system themselves, since the SILAC food might have metabolic effects. Moreover, with the spike-in standard approach the same SILAC organism can be used for experiments with strains of various genetic backgrounds.
Super-SILAC for tissue quantification
Tissue proteomics has posed a great challenge in proteomics due to the great complexity of the sample and the limitation on metabolic labeling of non-proliferating cells. Use of SILAC-labeled cells as a ‘spike-in’ standard solves the labeling problem because the quantification of each of the tissue samples can be performed relative to a standard. This approach has been demonstrated for the first time by spiking labeled Neuro2A cell lysates into mouse brain tissue lysates (Ishihama Y, et al, Nature biotechnology, 2005). Depending on the complexity of the tissue, however, using a single cell line as representative of the tissue may be insufficient. Furthermore, when the samples of interest are diverse (e.g. tumor samples) a single cell line may not represent this variability. To overcome these challenges it is recommended to mix several SILAC-labeled cell lines that together serve as the spike-in standard. This creates a super-set of SILAC labeled cells, termed ‘super-SILAC’ mix. The super-SILAC mix is more representative of the tissue, resulting in smaller and more accurate ratios between the proteins in the standard and the tissue of interest. In our initial description of the super-SILAC approach we used five cell lines to represent the breast cancer proteome. We are currently expanding super-SILAC to multiple other systems.