SP5 - Protein Interaction Mapping

AC Gavin (EMBL)

Here, we aim to establish the protein complex map for about 50 cell cycle genes as well as genes with relevance in drug resistance. We use a combination of TAP tagging and purification of protein complexes followed by an analysis of these complexes by LC-MS/MS as an unbiased approach to dissect the protein components of pathways relevant in IG-CSG. Different cellular systems including normal breast cell lines, tumor cell lines, invasive tumor cell lines and a tamoxifen-resistant MCF-7 cell line are used to unravel specific differences in the composition of complexes that may be associated with the progression of tumors. Protein complexes are isolated by TAP, and constituent proteins identified by liquid chromatography tandem mass spectrometry. Bioinformatic and statistical processing of the raw mass spectrometric data is carried out to establish a consolidated protein interaction map. Specific differences in the composition of complexes in different cell lines are unraveled, providing relevant information for further validation, quantification and structural determination.

Efforts have also been paid to the characterization of the proteome of most abundant proteins in MCF7 cells; this set of most abundant proteins will be used to help filtering out the non-specific binders in the TAP-MS purifications. On the MS front, we also significantly improved protein identification procedure by combining two different search algorithms, ‘Mascot’ and ‘Aldente’; this increases the identification of known complex components by ~20% compared with either method alone (Kuehner et al., 2009).

We significantly refined the protocols for the bioinformatics filtering of false positives based on statistical estimators that assign a significance score to each interaction in the raw dataset. As for the previous system, we measure the frequency with which pairs of proteins were found associated in our set of biochemical purifications. We improved the concept by integrating predicted interactions from databases (for example ‘STRING’) and the relative abundance of a given prey when associated with different baits (i.e., across different purifications). We used the MS scores that measure the probability for a peptide mass fingerprint to characterize each protein based on spectral counting. A reduced score for a prey in purification, when compared to the same prey in other purifications, reflects identifications by a smaller number of peptides (lower spectral counts); this is indicative of a spurious interaction and is therefore downweighted (Kuehner et al., 2009)