Inter-laboratory variability in mass spectrometry-based proteomics workflows

Backgrounds: Large variability in protein quantification by a plethora of proteomics workflows has been observed.
Aims: The aim was to find causes to variability in membrane protein quantification.
Methods: Membrane proteins essential for drug disposition (CYPs, UGTs, SLCs and ABCs) were quantified in ten human liver samples, by six mass spectrometry-based proteomics workflows. Proteins were quantified with five different targeted and one label free approach, using whole cell lysates or isolated subcellular fractions obtained by differential centrifugation. Proteins in subcellular fractions were scaled to protein levels in whole cell to enable comparison across the labs.
Results: Protein levels of the higher expressed drug metabolizing enzymes were overall in good agreement across laboratories using whole tissue lysate, with two-fold average difference. In general, lower levels (four to seven-fold) were obtained in subcellular fractions. Larger variation was seen for lower expressed transporting proteins across the labs. The transporters were in general quantified with a three-fold average difference across five of the laboratories, while scaled protein levels from plasma membrane fractions were consistently more than ten-fold lower. This indicates that incomplete enrichment of isolated membrane proteins is a major contributor to the observed differences. Further, the use of targeted or label-free approach and the choice of peptide sequence did not have a large influence on the quantification.
Summary/Conclusion: We demonstrate that different proteomics workflows give variable results in the quantification of membrane proteins. This is of great importance to consider when choosing proteomic workflows for protein quantification.