Protein expression profiles vary considerably between human cell lines and tissues, which is in part a reflection of their specialized roles within an organism. It is of considerable practical use to establish which proteins constitute the primary components of the respective proteomes. When compiled into databases, such information can facilitate the assessment of selectivity and specificity of a wide range of proteomic experiments. Here we describe the major constituents of proteomes of six human immortalized cell lines. By employing a combination of one-dimensional SDS-PAGE and nanocapillary liquid chromatography-tandem mass spectrometry (LC-MS/ MS), we identified up to 1785 non-redundant cytoplasmic and nuclear proteins from a single cell line using 50 and 30 g of total protein from the corresponding fractions. Up to 38 proteins could be identified from a single band in one liquid chromatography-MS/MS experiment. When combined with systematic gridding of gel lanes into 48 slices, a dynamic range for protein identification of ϳ1:2000 can be envisaged for this approach. Identified proteins range from 4 -553 kDa in size, cover the pI range between 3.4 and 12.8, and include 255 proteins with predicted transmembrane domains. Repeated analysis of peptides derived from the same gel band showed that the reproducibility of nanocapillary liquid chromatography-MS/MS of such complex mixtures is about 60 -70% suggesting that a particular analytical experiment would need to be repeated about three times to arrive at a representative estimate of the set of highly abundant proteins in a given proteome. Given its technical simplicity, sensitivity, and wealth of generated information, we have adopted this experimental approach to characterize every cell line and tissue that is the subject of experimentation in our laboratory. The combined dataset for the six cell lines consists of 2341 non-redundant human proteins and thus constitutes one of the largest collections of human proteomic data published to date.
Molecular & Cellular Proteomics 2:1297-1305, 2003.Cells and tissues express many thousands of proteins at any one time whose expression levels differ enormously between different cell types and span at least six orders of magnitude. As a result, a key question in any biochemical experiment aiming at the identification of a set of proteins in a particular biological context is how significant the observation actually is. In other words, the ability to discriminate between specific protein identifications in the experiment and so-called background identifications is of utmost importance. A range of factors can contribute to the often large number of nonspecific or at least functionally irrelevant identifications of a protein in a given experiment. These may include particular functional properties of a detected protein (such as the binding of chaperones to proteins) that result in prominent identifications of such proteins but also biochemical properties such as multiple hydrophobic interactions between proteins (often obser...