Various stable isotope labeling (SIL) techniques have recently emerged to improve the efficiency and accuracy of protein quantitation by mass spectrometry (MS). We have developed a mass-tagging strategy to incorporate stable isotope tagged amino acids into cellular proteins in a residue-specific manner during cell growth. In this study, we further extend this residue-specific SIL approach to the accurate quantitation of protein abundances in different cell populations. For proteins whose expression levels are the same in cells grown in the normal and labeled media, the relative areas of the normal (light) and labeled (heavy) isotopic peaks are linearly correlated with the cells mixing ratios. This approach was first used to determine the effect of the zinc-responsive transcription factor Zap1 on the yeast proteome. Ten protein spots from a PAGE gel were chosen randomly and their differential protein expression levels in wild-type and zap1delta cells were readily determined by the isotopic ratio. Methionine synthase (Met6) was identified to be up-regulated more than four times in the zap1delta mutant strain whereas the expression level of other nine proteins remained unchanged. Further, we applied this strategy to study the cellular response to radiation in human skin fibroblast cells. Analyzing one protein band randomly selected from SDS-PAGE, the expression level of a novel protein was found to increase two-fold in response to radiation whereas the expression level of a control protein remained unchanged. This strategy is generally applicable using any particular type of amino acid as the labeling precursors for accurate quantitation of protein relative abundances.