Complex-centric proteome profiling by SEC-SWATH-MS for the parallel detection of hundreds of protein complexes

Abstract

Most catalytic, structural and regulatory functions of the cell are carried out by functional modules, typically complexes containing or consisting of proteins. The composition and abundance of these complexes and the quantitative distribution of specific proteins across different modules is therefore of major significance in basic and translational biology. To date, the systematic detection and quantification of protein complexes has remained technically challenging. The chromatographic separation of native protein complexes followed by the mass spectrometric analysis of the proteins contained in sequential fractions results in potentially thousands of protein elution profiles from which, in principle, the presence of specific complexes can be inferred. However, the de novo inference of protein complexes from such datasets has so far remained limited with regard to selectivity and the retrieval of quantitative information. We recently developed a variant of this strategy, complex-centric proteome profiling, which extends the concepts of targeted proteomics to the level of native protein complex analysis. The complexcentric workflow consists of size exclusion chromatography (SEC) to fractionate native protein complexes, DIA/SWATH mass spectrometry to precisely quantify the proteins in each SEC fraction based on a consistent set of peptides, and targeted, complex-centric analysis where prior information from generic protein interaction maps is used to detect and quantify protein complexes with high selectivity and statistical error control via the computational framework CCprofiler. Complex-centric proteome profiling captures the majority of proteins in complex-assembled state and reveals their organization into hundreds of complexes and complex variants observable in a given cellular state. The protocol is applicable to genetically unaltered tissue cultures and adaptable to primary tissue. At present it requires approximately 8 days of wet-lab work, 15 days of MS measurement time and 7 days of computational analysis.

Discovery–Versus Hypothesis–Driven Detection of Protein–Protein Interactions and Complexes

“…The key benefit of the discovery approach is that only limited prior knowledge is necessary (only enough to train the ML algorithm) and that novel protein complexes can be readily identified. In practice, high proteomic coverage coupled to limited biochemical resolution (determined by the resolution of fractionation strategies in CoFrac-MS, sampled temperatures in TPP and the number of individual samples in full proteome co-variation analysis) results in a high degree of random correlation of non-interacting proteins [ 42 ]. This has negative effects on the sensitivity and selectivity of purely discovery-based approaches.…”

Systematic detection of functional proteoform groups from bottom-up proteomic datasets

“…first, there will likely be true proteoform groups in the negative background dataset of the HEK293 sample, which are here treated as false positives. Second, some of the proteins that were grouped into one mixed protein are likely to be involved in protein interactions or having a similar elution profile by chance 29 .…”

“…We applied the COPF strategy to our previously published native complex cofractionation dataset of Hela CCL2 cells synchronized in interphase and mitosis 28 to identify cell-cycle and assembly-specific proteoforms. In this case the dataset consists of native complexes isolated from cells in two cell cycle states, separated by size-exclusion chromatography (SEC), and then analyzed by bottom-up proteomic analysis using data-independent acquisition mass spectrometry (DIA/SWATH-MS) 21,29 . The workflow with individual steps is schematically illustrated in Figure 3A.…”

“…Third, each sampled fraction is separately processed for bottom-up proteomic measurements by SWATH-MS 17 followed by peptidecentric analysis [32][33][34] . In a fourth step, peptide elution profiles across the SEC fractions are evaluated to infer protein complex assemblies by CCprofiler as described before 21,29 and additionally by the COPF method to identify proteoforms differentially associating with complexes within or across cell cycle states.…”

“…Key properties of applicable datasets include both high sequence coverage and quantitative accuracy as well as a sufficient number of replicates with respect to the expected effect size of abundance differences between proteoforms across the investigated biological conditions. We recently presented a novel strategy for detecting and quantifying protein complexes from co-fractionation MS datasets 21,28,29 . The method consists of native cell lysis to extract protein complexes, followed by size-exclusion chromatography (SEC) and analysis by DIA/SWATH-MS. Because of the rich biological context, high sequence coverage and accurate quantitative peptide-level information inherent to SEC-SWATH-MS, we argued that it provides a unique opportunity to evaluate proteoform groups that are relevant for protein complex organization.…”