CysDB: A Human Cysteine Database based on Experimental Quantitative Chemoproteomics

Abstract

Cysteine chemoproteomics studies provide proteome-wide portraits of the ligandability or potential ‘druggability’ of thousands of cysteine residues. Consequently, these studies are enabling resources for closing the druggability gap, namely achieving pharmacological manipulation of ~99% of the human proteome that remains untargeted by FDA approved small molecules. Recent interactive dataset repositories, such as OxiMouse and SLCABPP, have enabled users to interface more readily with cysteine chemoproteomics studies1,2. However, these databases remain limited to single studies and therefore do not provide a mechanism to perform cross-study analyses. Here we report CysDB as a curated community-wide repository of cysteine chemoproteomics data that incorporates high coverage data derived from nine studies generated by the Backus, Cravatt, Gygi, Wang, and Yang research groups. CysDB is a SQL relational database that is publicly available at https://backuslab.shinyapps.io/cysdb/ and features chemoproteomic measures of identification, hyperreactivity, and ligandability for 62,888 cysteines (24% of all cysteines the human proteome). The CysDB web application also includes annotations of functionality (UniProtKB/Swiss-Prot, Pfam, Panther), known druggability (FDA approved targets, DrugBank, ChEMBL), disease-relevance and genetic variation (ClinVar, Cancer Gene Census, Online Mendelian Inheritance in Man), and structural features (Protein Data Bank). Showcasing the utility of CysDB, here we report the discovery and enrichment of ligandable cysteines in undruggable classes of proteins, the observation that a subset of cysteines showed marked preference for specific classes of electrophiles (chloroacetamide vs acrylamide), and that ligandable cysteines are present in numerous undrugged disease-relevant proteins. Most importantly, we have designed CysDB for the incorporation of new datasets and features to support the continued growth of the druggable cysteineome.

Assigning functionality to cysteines by base editing of cancer dependency genes

“…ABPP and related chemical proteomic methods have discovered a wide array of cysteine residues in diverse protein classes that can be targeted by electrophilic small molecules 10, [16][17][18]58 . It remains challenging, however, to determine the functional consequences of such covalent liganding events, especially for proteins lacking established biochemical or cellular assays.…”

“…Here, we describe a genome editing approach for globally assessing the functionality of ligandable cysteines in proteins required for cancer cell growth. Key to the implementation of our strategy was the knowledge afforded by -i) ABPP of site-level resolution of covalent compound-cysteine interactions across the human proteome 10,[16][17][18]58 ; and ii) the Cancer Dependency Map of protein-encoding genes required for cancer cell proliferation 5which, together, enabled focused, base editing-mediated mutational screens of ligandable cysteines and surrounding regions to assess their essentiality for cell growth.…”

Mapping the chemical space of active-site targeted covalent ligands for protein tyrosine phosphatases

“…[25] However, published proteome-wide ligandability screens have not yielded any substantial leads for PTP active site cysteine residues. [30] Rather, these screens suggest that PTP catalytic cysteines are not very reactive relative to the entire proteome, despite their low pKa values, [23,24] and they are apparently unliganded by most fragments used in existing screens. Consequently, it is difficult to extract information from these global ligandability screens on how different structures impact covalent inhibition of PTPs.…”

Proteome-wide structure-based accessibility analysis of ligandable and detectable cysteines in chemoproteomic datasets

“…In recent years, increasingly large-scale fragment library screens have generated a substantial body of publicly available datasets. 2,15,[18][19][20] Whilst the most comprehensive cysteine profiling (CP) studies consistently profile 10,000 or more reactive cysteines in parallel, the structural distribution of these residues across the 261,260 cysteine residues (from the UniProt one-protein-per-gene database) present in the human proteome has yet to be systematically analysed. Similarly, the extent to which CP approaches sample potentially ligandaccessible residues remains undefined but has significant implications for the efficiency of TCI discovery platforms and assessment of proteome-wide selectivity.…”

“…1,2 Subsequent technical developments, including novel cleavable reagents, multiplexing strategies, chemical enrichment and mass spectrometric (MS) acquisition approaches, have together markedly improved both depth and throughput, as reviewed comprehensively elsewhere. 14,15 Weerapana and co-workers further developed this concept for live-cell ( in situ ) labelling, using photo-uncaging of protected α-haloketones in live cells to permit labelling concentrations of up to 200 μM without significant cell death (Figure 1b, bottom). 16,17 Such reactive amino acid profiling platforms have been applied extensively to covalent ligand discovery, primarily through quantifying cysteine reactivity in lysates treated with libraries of electrophilic fragments compared to DMSO-treated controls (Figure 1c).…”