Combining XL-MS and LiP-MS to gain detailed molecular understanding about protein structural rearrangements upon perturbations
Cathy Marulli1, Alexander Leitner1, Paola Picotti1
1 Department of Biology, Institute of Molecular Systems Biology, ETH Zürich, Zurich, Switzerland
Introduction: Cells quickly adapt to changing environments by modulating protein structure resulting in altered activity or function. Protein structure is regulated by numerous mechanisms such as post-translational modifications, proteolytic cleavage or binding to other proteins, DNA, RNA, and small molecules. Structural proteomics methods such as limited proteolysis (LiP) and cross-linking (XL) coupled to mass spectrometry (MS) capture those events on a proteome-wide scale. While LiP-MS only reports on surface accessibility changes between conditions, XL-MS carries information about distance between residues. We combine for the first time LiP-MS and XL-MS to the glycolytic to gluconeogenic shift in Saccharomyces cerevisiae to gain detailed molecular understanding of how the proteome structurally and functionally adapts to this perturbation.
Methods: Saccharomyces cerevisiae BY4716 was grown in triplicates in glucose and ethanol minimal medium. For LiP-MS, proteinase K (100:1 substrate:enzyme) was added to lysates (1 mg/ml) for 5 min, followed by standard workflows using trypsin. For XL-MS, lysates were fractionated by serial ultrafiltration in decreasing order of molecular weight cutoffs (100 kDa, 50kDa, 30 kDa, 10 kDa). Each fraction was cross-linked at 1 mg/ml protein concentration with 2.25 mM BS3 followed by standard workflows using trypsin. Crosslinks were enriched by size exclusion chromatography (5 fractions). Samples were analyzed by data-dependent and/or data-independent LC-MS/MS on a Thermo Orbitrap Eclipse Tribrid or Fusion Lumos instrument (120 min gradients). LiP-MS data was analyzed in Spectronaut, XL-MS data was searched in xiSearch and quantified in Spectronaut.
Preliminary data: To generate a quantitative XL-MS dataset with high coverage and good reproducibility, we fractionated lysates by serial ultrafiltration prior to cross-linking and show that the number of intra-protein links identified in biological triplicates increases from 1258 to 4292 coming from 306 and 665 proteins, respectively. We use data-dependent acquisition to identify cross-links. From the latter, we build a spectral library and use data-independent acquisition to quantify the cross-links which considerably decreases the variability of cross-link quantifications between replicates and thus increases the statistical power to identify structural changes. Applying the XL-MS workflow in parallel with LiP-MS on yeast grown under glycolytic and gluconeogenic conditions in biological triplicates generates highly informative structural fingerprints for hundreds of proteins that report on protein abundance, intra- and inter-protein-link, mono-link and LiP changes. Those fingerprints allow to classify proteins as being regulated solely by structure, abundance, or both. Parallel identification of structural changes by LiP-MS and XL-MS gives us high confidence and detailed complementary information about the structural rearrangements and allows to generate hypotheses on potential regulatory events. In the case of phosphoglycerate kinase, our data suggests that the closed conformation of the enzyme is more occupied under gluconeogenic conditions and hypothesizes a potential site for allosteric regulation. Thus, we envision that pushing the boundaries of structural proteomics workflows is crucial to produce high throughput data on how protein structure relates to function and therefore essential for a detailed understanding of biological systems.
Novel aspect: Integration of a LiP-MS and a quantitative XL-MS dataset to gain biological understanding of protein structural rearrangements upon perturbation
Event Timeslots (1)