An experimentally generated peptide database increases the sensitivity of XL-MS with complex samples

Cross-linking mass spectrometry (XL-MS) is steadily expanding its range of applications from purified protein complexes to more complex samples like organelles and even entire cells. One main challenge using non-cleavable cross-linkers is the so-called n² problem: With linearly increasing database size, the search space for the identification of two covalently linked peptides per spectrum increases quadratically. Here, we report an alternative search strategy that focuses on only those peptides, which were demonstrated to cross-link under the applied experimental conditions. The performance of a parallel XL-MS experiment using a thiol-cleavable cross-linker enabled the identification of peptides that carried a cleaved cross-link moiety after reduction and hence were involved in cross-linking reactions. Based on these identifications, a peptide database was generated and used for the database search of the actual cross-linking experiment with a non-cleavable cross-linker. This peptide-focused approach was tested on protein complexes with a reported structural model and obtained results corresponded well to a conventional database search. An application of the strategy on in vivo cross-linked Bacillus subtilis and Bacillus cereus cells revealed a five- to tenfold reduction in search time and led to significantly more identifications with the latter species than a search against the entire proteome. SIGNIFICANCE: Instead of considering all theoretically cross-linkable peptides in a proteome, identification and pre-filtering for a subset of cross-link peptide candidates allows for a dramatically decreased search space. Hence, there is less potential for the random accumulation of false positives ultimately leading to a higher sensitivity in the XL-MS experiment. Using the peptide-focused approach, a cross-linking database search can be conducted in a fraction of time while yielding a similar or higher number of identifications, thereby enabling the cross-linking analysis of samples of mammalian proteome complexity.