Probing native protein structures by chemical cross-linking, mass spectrometry, and bioinformatics

Abstract

Chemical cross-linking of reactive groups in native proteins and protein complexes in combination with the identification of cross-linked sites by mass spectrometry has been in use for more than a decade. Recent advances in instrumentation, cross-linking protocols, and analysis software have led to a renewed interest in this technique, which promises to provide important information about native protein structure and the topology of protein complexes. In this article, we discuss the critical steps of chemical cross-linking and its implications for (structural) biology: reagent design and cross-linking protocols, separation and mass spectrometric analysis of cross-linked samples, dedicated software for data analysis, and the use of cross-linking data for computational modeling. Finally, the impact of protein cross-linking on various biological disciplines is highlighted.

Fig. 1.

Nomenclature of common products of chemical cross-linking reactions. The terminology, cross-link, loop-link, and monolink, used in this article is shown.

Fig. 2.

Structures of most commonly used amine-reactive cross-linking reagents: DSS, BS³, DSG, and bis(sulfosuccinimidyl) glutarate (BS²G).

Fig. 3.

Comparison of cross-linking using DSS and DSG. Shown are the distances (N_ε-N_ε) between two cross-linked lysine residues in different model proteins determined from published three-dimensional structures for DSG (a) and DSS (b) and the theoretically possible (“virtual”) (c) cross-links. Chemical cross-linking was performed as described previously (26); details about the calculations can be found in the supplemental method details for simulations.

Fig. 4.

Identification of cross-linked peptides from MS/MS spectra using information of isotopically coded cross-linkers and xQuest (26). 1, identification of isotopically shifted MS¹ feature pairs. 2, comparison of light and heavy MS/MS spectra of isotope pairs and subsequent sorting of fragment ions into common ions (identical; indicated in green) and cross-link (x-link) ions (shifted; red). 3, preparation of either an “enumeration” mode database for small sequence databases of up to 100 proteins or an “ion-tag” mode database for large sequence databases. 4, xQuest search: retrieving all candidate cross-links filtered by mass if an enumeration mode database is used or retrieving all candidate peptides from the ion index utilizing information of common ions. All possible cross-link candidates are recombined and filtered by mass in a second step. 5, xQuest scoring: the software assigns scores to all candidate cross-links.

Fig. 5.

Influence of distance constraints from Lys-Lys interprotein cross-links in benchmark protein complex data set. Approximate theoretical constraints for DSS (21 Å), DSG (15 Å), and zero-length cross-linking (9 Å) are highlighted in the legend. a, histogram of theoretically possible interprotein cross-links for various distance constraints. b, average r.m.s.d. between native and docked conformations for different numbers of virtual cross-links as constraints. c, fraction of 10,000 calculated structures that can be excluded dependent on the number and maximum length of distance constraints.

Fig. 6.

Refinement of protein-protein docking using distance constraints obtained from cross-linking data. a, conformational space of 10,000 calculated conformations for the bovine trypsin-inhibitor complex. Black, all conformations without use of distance constraints from cross-links; red and yellow, conformations remaining after consideration of one or two distance constraints from cross-links, respectively; green, largest cluster of conformations after consideration of three constraints and clustering. b–e, schematics of trypsin-inhibitor complex structures showing a superposition of 30 systematically sampled conformations from the respective groups in a. All three-dimensional structures were generated using PyMOL.