Home, Research, People, Contacts, Downloads, Links, Pubs/Presentations/Posters
New Stable-Isotopic Labeling Strategies for Quantitative Proteomics.
It has historically been very difficult to obtain meaningful quantitative comparisons from MS analyses. These difficulties largely stem from differences in ionization and detection as a function of chemical species, but are also compounded by uncontrolled differences in sample preparation, fractionation and ion suppression. The isotope-assisted methods provide a means of skirting these obstacles by comparing all of the chemical species directly in a single analysis from two premixed differentially labeled samples. Isotopically labeled samples are chemically identical but distinct in mass, and thus are directly comparable within single mass spectral analyses.
While it is possible to fully substitute heavy label amino acids at 85-95% in rats and mice, the cost of using the required amount of 15N Spirulina is expensive. These costs limit the length of experiments and the numbers of biological replicates. In order to reduce costs and increase the utility of quantitative proteomics approaches for animal studies, the Sussman lab has adapted existing metabolic labeling methods to compare natural abundance and 6% 15N substitution based on the work of Whitelegge and coworkers (Whitelegge et al., 2004). Examples of this are shown in Figure 1. Full metabolic labeling produces two completely resolved isotopic envelopes whose intensities can be compared, in the case of partial incorporation the light and heavy envelopes overlap, but have distinct shapes. When the 6% labeled and unlabeled samples are combined, the result is a composite envelope whose shape is the sum of the labeled and unlabeled envelopes. If the peptide responsible for a given MS isotopic envelope is known, the relative contributions of the heavy and light envelopes required to produce the observed composite envelope may be determined based on the predicted shapes of the light and heavy envelopes alone.
Isotope-assisted quantification strategies in MS and MS/MS. The need to identify proteomic changes in abundance or structural modification resulting from an experimental treatment (be it genetic, chemical or physical) is made evident by the diversity of high throughput stable-isotope mass spectrometric strategies employed in proteomics (Tao and Aebersold, 2003). Fundamentally, existing comparative proteomics approaches all rely on the incorporation of a stable isotopic label into peptides so that one can observe differences in a control versus test samples just by comparing the intensities of matched pairs of mass spectral peaks. These peaks correspond to chemically identical species that co-elute in all chromatographic steps, and share physical properties pertaining to ionization and detection in the spectrometer, yet are distinguishable by mass. This avoids problems associated with changes in ionization or matrix effects that may skew results in independently run samples.
Approaches to stable-isotope incorporation are separated into two basic groups. The first is in vitro labeling, where the proteins or metabolites are extracted, proteins are proteolysed, and then peptides or metabolites are derivatized with chemical reagents reactive to specific functional groups such as sulfhydryls in cysteines (ICAT, or isotope coded affinity tag; Gygi et al., 1999)), amino groups (iTRAQ (Applied Biosystems, Inc.), and phosphates (PHIAT, (Goshe et al., 2002). Modification strategies aimed at specific reactive functional groups in metabolite populations is the subject of on-going research in the Smith laboratory. 18O can also be incorporated directly into the carboxy terminus of peptides released during trypsinization in the presence of 100% H218O (Heller et al., 2003). The second is in vivo labeling or metabolic labeling. This approach incorporates isotopically labeled metabolites (e.g. 13CO2 or 15NO3) by providing them to living tissues. This has been widely used in animal cell lines via feeding labeled amino acids (SILAC, stable isotope labeling with amino acids in cell culture) (Ong et al., 2003).
Each of the above methods has distinct advantages and disadvantages. A major advantage of the metabolic labeling approach over other stable isotopic labeling strategies is that two biological samples are combined into one prior to extraction and fractionation, providing an almost ideal internal control for finding real proteomic and metabolomic perturbations independent of variability from processing steps. Major disadvantages apply to data processing steps that complicate peptide identification and quantification protocols. These disadvantages have been largely circumnavigated via data processing scripts using Perl and Mathematica that have been written in the Sussman lab to automate data processing, based on the strategy of MacCoss and coworkers (MacCoss et al., 2003). The Sussman laboratory has been performing metabolic labeling of Arabidopsis, taking advantage of the fact that unlike animals, plants can utilize nitrate as its sole nitrogen source. Using inexpensive 98%-[15N] KNO3, Arabidopsis plants have been grown to maturity with near 98% incorporation 15N, with no deleterious growth effects. A similar finding that 15N is easily tolerated, even at near 100% levels, has recently been reported in rats (Wu et al., 2004), using 15N-grown algae (Spirulina) as the sole source of nitrogen and we have confirmed similar observations with mice. In the past year, Sussman has collaborated with Prof. William F. Dove in producing mice containing nearly 90% 15N, in order to identify proteins whose abundance is altered in a mouse genetic model for human colon cancer. Data processing scripts have also been modified to accommodate 90% and lower 15N incorporation as observed in the mouse samples and have been successfully applied to obtain ratios. Figure 1 shows examples of MS spectra of peptides from digested protein isolated from various organisms (including Arabidopsis and mouse) showing pairs of chemically identical natural abundance and 15N labeled peptides and their % incorporation. Further experiments and results relating to this strategy will be described in more detail below.