Proteome-wide identification of ubiquitin interactions using UbIA-MS

Abstract

Ubiquitin-binding proteins play an important role in eukaryotes by translating differently linked polyubiquitin chains into proper cellular responses. Current knowledge about ubiquitin-binding proteins and ubiquitin linkage-selective interactions is mostly based on case-by-case studies. We have recently reported a method called ubiquitin interactor affinity enrichment-mass spectrometry (UbIA-MS), which enables comprehensive identification of ubiquitin interactors for all ubiquitin linkages from crude cell lysates. One major strength of UbIA-MS is the fact that ubiquitin interactors are enriched from crude cell lysates, in which proteins are present at endogenous levels, contain biologically relevant post-translational modifications (PTMs) and are assembled in native protein complexes. In addition, UbIA-MS uses chemically synthesized nonhydrolyzable diubiquitin, which mimics native diubiquitin and is inert to cleavage by endogenous deubiquitinases (DUBs). Here, we present a detailed protocol for UbIA-MS that proceeds in five stages: (i) chemical synthesis of ubiquitin precursors and click chemistry for the generation of biotinylated nonhydrolyzable diubiquitin baits, (ii) in vitro affinity purification of ubiquitin interactors, (iii) on-bead interactor digestion, (iv) liquid chromatography (LC)-MS/MS analysis and (v) data analysis to identify differentially enriched proteins. The computational analysis tools are freely available as an open-source R software package, including a graphical interface. Typically, UbIA-MS allows the identification of dozens to hundreds of ubiquitin interactors from any type of cell lysate, and can be used to study cell type or stimulus-dependent ubiquitin interactions. The nonhydrolyzable diubiquitin synthesis can be completed in 3 weeks, followed by ubiquitin interactor enrichment and identification, which can be completed within another 2 weeks.

Robust summarization and inference in proteome-wide label-free quantification

“…The second type of preprocessing is imputation of missing values, and this too can be beneficial (Figure 3). However, because several different imputation methods exist, and because each of these applies to different sources of missingness, best results are typically achieved when using a mixed imputation, where randomly missing values and values missing under low abundance are imputed differently Zhang et al [2018]. It should be noted, however, that the robust modelling in MSqRobSum can safely omit imputation altogether (Supplementary Figure 8).…”

“…The recent Differential Enrichment analysis of Proteomics data (DEP) software package greatly improved MaxLFQ based analysis by adopting a mixed imputation strategy for missing protein intensities that infers whether random missingness or missingness due to low abundance occurs Zhang et al [2018]. It also provides a more robust downstream DE analysis using proteinwise linear models combined with empirical Bayes statistics (through the limma package Ritchie et al [2015]).…”

“…MaxLFQ values are analysed with the standard workflow in the Bioconductor software package DEP version 1.2.0. Zhang et al [2018] Proteins that are contaminants or that originate from reversed sequences are removed from the data set. Only proteins with no missing values in at least one treatment group are kept.…”

“…Taken together, the identification issue and the peptide specific effects on quantification have a severe impact on the downstream summarization of peptide intensities towards protein abundances. Goeminne et al [2015] Indeed, because of these issues, simple summarization methods such as the mean or median peptide intensity are known to give unreliable protein abundance estimates Goeminne et al [2015] and more advanced summarization strategies have therefore been proposed for LFQ data in the literature Silva et al [2005], Cox et al [2014], Zhang et al [2018], Choi et al [2014]. In Figure 1 panel A we show the performance of these different data analysis strategies on a benchmark dataset.…”

proDA: Probabilistic Dropout Analysis for Identifying Differentially Abundant Proteins in Label-Free Mass Spectrometry

“…The advantage of using limma or similar approaches for LFQ-MS has been advocated only rather recently (e.g., Kammers et al (2015)). For example, the DEP package (Zhang et al 2018) performs imputation followed by a limma analysis to infer differentially abundant proteins. As stated above, the use of imputation may compromise the validity of limma's statistical inference, and hence, the purpose of the present work is to adapt limma-style inference to account for values missing not at random and so improve power and reliability of differential abundance analysis for LFQ-MS.…”