Does chemical cross-linking with NHS esters reflect the chemical equilibrium of protein-protein noncovalent interactions in solution?

Chemical cross-linking in combination with mass spectrometry has emerged as a powerful tool to study noncovalent protein complexes. Nevertheless, there are still many questions to answer. Does the amount of detected cross-linked complex correlate with the amount of protein complex in solution? In which concentration and affinity range is specific cross-linking possible? To answer these questions, we performed systematic cross-linking studies with two complexes, using the N-hydroxysuccinimidyl ester disuccinimidyl suberate (DSS): (1) NCoA-1 and mutants of the interacting peptide STAT6Y, covering a K(D) range of 30 nM to >25 μM, and (2) α-thrombin and basic pancreatic trypsin inhibitor (BPTI), a system that shows a buffer-dependent K(D) value between 100 and 320 μM. Samples were analyzed by matrix-assisted laser desorption/ionization mass spectrometry (MALDI-MS). For NCoA-1•STAT6Y, a good correlation between the amount of cross-linked species and the calculated fraction of complex present in solution was observed. Thus, chemical cross-linking in combination with MALDI-MS can be used to rank binding affinities. For the mid-affinity range up to about K(D) ≈ 25 μM, experiments with a nonbinding peptide and studies of the concentration dependence showed that only specific complexes undergo cross-linking with DSS. To study in which affinity range specific cross-linking can be applied, the weak α-thrombin•BPTI complex was investigated. We found that the detected complex is a nonspecifically cross-linked species. Consequently, based on the experimental approach used in this study, chemical cross-linking is not suitable for studying low-affinity complexes with K(D) >> 25 μM.