Benchmarking Cleavable Biotin Tags for Peptide-Centric Chemoproteomics

Abstract

Click chemistryspecifically the copper-catalyzed azide-alkyne cycloadditionhas enabled the development of a wide range of chemical probes to analyze subsets of the functional proteome. The "clickable" proteome can be selectively enriched by using diverse cleavable biotin tags, but the direct identification of probe/tag-modified peptides (or peptide-centric analysis) remains challenging. Here, we evaluated the performance of five commercially available cleavable biotin tags in three most common chemoproteomic workflows, resulting in a comparative analysis of 15 methods. An acidcleavable biotin tag with a dialkoxydiphenylsilane moiety, in combination with the protein "click", peptide "capture" workflow, outperforms all other methods in terms of enrichment efficiency, identification yield, and reproducibility, although its performance may be slightly compromised by the formation of an unwanted formate product revealed by blind search. Despite being inferior, photocleavable, and reduction-cleavable, biotin tags can also find their applicable sceneries, especially when dealing with acid-sensitive probes or probe-derived modifications. Furthermore, the systematic comparison of LC−MS/MS characteristics of tag-modified peptides provides valuable information for the future development of cleavable biotin reagents. Taken together, our data provides a much-needed practical guidance for researchers interested in developing and/or applying an ideal cleavable biotin tag to peptide-centric chemoproteomics.

CysDB: A Human Cysteine Database based on Experimental Quantitative Chemoproteomics

“…We prioritized studies that reported high coverage datasets that measured one or more of the following parameters: (1) total number of cysteines identifiable by the pan-cysteine reactive probes, (2) measurement of cysteine intrinsic reactivity towards iodoacetamide alkyne (IAA, 1) and (3) assaying cysteine ligandability (Figure 1A). We collected a total of nine datasets that fulfilled our criteria (Figure 1B for all datasets used) 2,[4][5][6][7][8][9][10][11] .…”

“…Cysteine proteomics experiments can be generally classified into four main categories: (1) identification, (2) measuring hyperreactivity, (3) measuring ligandability and (4) measuring redox state. (1) We consider identification studies as those aiming to increase coverage of cysteine containing peptides using label free quantification [4][5][6] . (2) Hyperreactivity experiments measure the intrinsic reactivity of cysteines towards highly electrophilic probes [7][8][9][10] , while (3) ligandability experiments measure the intrinsic ligandability or potential 'druggability' of cysteines using libraries of drug-like electrophilic molecules, natural products, and lipid derived electrophiles 2,11,[15][16][17][18][19] .…”

“…Reasons for this gap include protein abundance and restricted expression profiles, location of cysteines in very long or very short tryptic peptides, which are not detected in standard trypsin digests, and unreactive cysteines, such as those buried in the protein core or located in structural disulfides. Despite these technical limitations, the cysteinome continues to grow, with the addition of multiple high coverage new studies in this year alone 6,10,14 .…”

“…First, we aggregated all non-redundant cysteines published by all studies, using the unique identifier UniProtKBID_CYS#. For some studies 2,[4][5][6][7][8][9]11 residue positions and protein identifiers were provided in the publication's supporting information. For a subset of studies, the supporting tables instead provided labeled peptide sequences and protein IDs 7,10 .…”

“…Thereby, limiting the potential impact of cell-state dependent differences of cysteine reactivity as a potential confounder to our downstream analyses. Aggregation of all datasets, including results from using multiple cell lines 2,[4][5][6][7][8][9][10][11] , resulted in the chemoproteomic identification of 62,888 unique cysteines and 11,621 proteins (Figure 1C & 1D), which to our knowledge represents the most comprehensive cysteinome dataset reported to date.…”