Mass spectrometry reveals the chemistry of formaldehyde cross-linking in structured proteins

Formaldehyde is a widely used fixative in biology and medicine. The current chemical model for formaldehyde cross-linking of proteins is the formation of a methylene bridge that incorporates one carbon atom into the link. We present mass spectrometry data that largely refute this model. Instead, the data show that cross-linking of structured proteins mainly involves a reaction that incorporates two carbon atoms into the link. Under MS/MS fragmentation, the link cleaves symmetrically to yield unusual fragments carrying a modification of one carbon atom. If these characteristics are considered, then formaldehyde cross-linking is readily applicable to the structural approach of cross-linking coupled to mass spectrometry. From a cross-linked mixture of purified proteins, a suitable analysis identifies tens of cross-links that fit well with their atomic structures. From in situ cross-linking of human cells in culture, our analysis identified 469 intra-protein and 90 inter-protein cross-links, which also fit well with available atomic structures. Interestingly, many of these cross-links could not be mapped onto a known structure and thus provide new structural insights. An example of such case are formaldehyde cross-links that localize the binding site of βNAC on the ribosome. Also of note are also cross-links of actin with several auxiliary proteins for which the structure is unknown. Based on these findings we suggest a revised chemical model, which has relevance to the reactivity and toxicity of formaldehyde.