Robust MS quantification method for phospho-peptides using (18)O/(16)O labeling

BACKGROUND: Quantitative measurements of specific protein phosphorylation sites, as presented here, can be used to investigate signal transduction pathways, which is an important aspect of cell dynamics. The presented method quantitatively compares peptide abundances from experiments using (18)O/(16)O labeling starting from elaborated MS spectra. It was originally developed to study signaling cascades activated by amyloid-β treatment of neurons used as a cellular model system with relevance to Alzheimer's disease, but is generally applicable. RESULTS: The presented method assesses, in complete cell lysates, the degree of phosphorylation of specific peptide residues from MS spectra using (18)O/(16)O labeling. The abundance of each observed phospho-peptide from two cell states was estimated from three overlapping isotope contours. The influence of peptide-specific labeling efficiency was removed by performing a label swapped experiment and assuming that the labeling efficiency was unchanged upon label swapping. Different degrees of phosphorylation were reported using the fold change measure which was extended with a confidence interval found to reflect the quality of the underlying spectra. Furthermore a new way of method assessment using simulated data is presented. Using simulated data generated in a manner mimicking real data it was possible to show the method's robustness both with increasing noise levels and with decreasing labeling efficiency. CONCLUSION: The fold change error assessable on simulated data was on average 0.16 (median 0.10) with an error-to-signal ratio and labeling efficiency distributions similar to the ones found in the experimentally observed spectra. Applied to experimentally observed spectra a very good match was found to the model (<10% error for 85% of spectra) with a high degree of robustness, as assessed by data removal. This new method can thus be used for quantitative signal cascade analysis of total cell extracts in a high throughput mode.