On June 9 , 2026, Dr. Meng-Qiu Dong’s laboratory at the National Institute of Biological Sciences, Beijing (NIBS) published a research article in Molecular Systems Biology. This article—To cleave or not to cleave: a systemic evaluation of DSS versus DSSO for cross-linking mass spectrometry analysis—bears witness of a 3-way collaboration between the Dong lab, the NIBS Chemistry Center, and Dr. Si-Min He’s research team at the Institute of Computing Technology, Chinese Academy of Sciences (ICT, CAS).

The authors conducted a thorough evaluation of two most widely used cross-linking mass spectrometry (CXMS or XL-MS) approaches, one based on gas-phase non-cleavable cross-linkers such as DSS and the other on cleavable cross-linkers such as DSSO. Through detailed, head-to-head comparisons, the team identified key factors affecting cross-link identification, clarified long-standing debates, and offered actionable, evidence-based guidance for CXMS experiments.

CXMS is a facile tool for determining three-dimensional structures of proteins and protein complexes. It is also a potentially powerful tool to map protein-protein interactomes.

Meng-Qiu Dong and collaborators have been advancing CXMS since 2008. Over the years they have developed the pLink search engine  (Nat Methods 2012, Nat Commun 2019), various cross-linkers (eLife 2016, Nat Commun 2019, Nat Commun 2022), and new applications (Nat Methods 2015, Biophys Rep 2015, Anal Chem 2016, J Biol Chem 2017, MCP 2025).

In recent years, the push to map protein-protein interactomes infused immense interest in gas-phase cleavable cross-linkers such as DSSO. Two symmetric C-S bonds in the spacer arm of DSSO are equally susceptible to cleavage by CID, resulting in a pair of ions with a characteristic mass difference for each of the released peptides. Such doublet peptide ions help reduce the cross-link search space from n² to 2n, with n being the number of peptides in the database, which is up to 10⁶ in protein-protein interactome mapping. Thus, search space reduction is believed to be the primary benefit of DSSO. However, a study from Juri Rappsilber’s lab (Analytical Chemistry, 2022) challenged this belief, arguing that DSSO's true advantage lies instead in producing more backbone fragmentation for peptides linked by it.

To resolve this debate, the authors compared DSSO against non-cleavable DSS through a rigorous benchmarking study across different sample complexities, from purified proteins to mammalian whole-云顶集团(中国) lysates. Key findings from this study are as follows:

1.   DSS, with its longer and more flexible spacer arm, probes a greater spatial volume and yields more cross-link identifications than DSSO from purified protein complexes to yeast ribosome and bacterial lysate.

2.   When sample complexity increases to that of a mammalian whole-云顶集团(中国) lysate, more cross-links are identified using DSSO than DSS. Such a turnaround is due to a slightly lower fragment ion coverage of DSS cross-links, not the absence of doublet ions. This supports Rappsilber's conclusion in 2022.

3.   High fragment ion coverage of 85% or above maintains identification sensitivity, irrespective of search space magnitude or cross-linker type.

Dr. Yong Cao, the first author of this study, resolved another debate earlier (Journal of Proteome Research, 2023)—on whether NHS ester-based cross-linkers react with serine, threonine, and tyrosine residues (STY). Using amino acids of non-reactive side chains such as glycine, valine, and leucine as a negative control, he demonstrated that most STY cross-links identified are unreliable—some of them are false-positive matches and some are actually lysine-lysine (K-K) cross-links.

Based on the above findings, the authors have three practical recommendations for CXMS experiments.

1. For Low-to-Medium Complexity Samples: DSS is better than DSSO for samples up to the complexity of a bacterial lysate.
2. For High-Complexity Samples: Current cross-linkers are not yet up to the task of mapping mammalian interactomes; a cleavable, enrichable cross-linker paired with extensive fractionation of the cross-linked sample is the best of what can be done.
3. For Data Analysis: avoid designating STY as cross-linkable residues.

Dr. Yong Cao (NIBS), Peng-Zhi Mao (ICT), Dr. Zhen-Lin Chen (ICT), and Qing-Cui Wu (NIBS Chemistry Center) are co-first authors of the above Molecular Systems Biology paper. Drs. Meng-Qiu Dong (NIBS), Xiang-Bing Qi (NIBS), and Si-Min He (ICT) are co-corresponding authors. Other co-authors include Drs. Jincai Yang, Yue Zhao, and Adalet Memetimin of NIBS, and Dr. Hao Chi (ICT).

This research was generously supported by the National Key R&D Program of China, the National Natural Science Foundation of China, the Beijing Natural Science Foundation, and intramural funds of NIBS and TIMBR.