2. Mapping biological interfaces by CLMS
A crosslinking reagent makes a short covalent bridge between two biomolecules in proximity, i.e. freezing them, that occurs when they are bound through specific molecular contacts, thereby usefully aiding the identification of an unknown binding partner interacting with a host protein, especially in proteome-wide studies (Figure 1) (Götze et al., 2019; Holding 2015; Piersimoni and Sinz 2020). Due to poor site-specific reactivity and/or lack of highly sensitive analytical tools, residue-level elucidation of a binding interface via chemical crosslinking had not been possible until controlled crosslinking chemistry and high-sensitivity mass spectrometry emerged into practice (Mishra et al., 2020; Yang et al., 2016).
Among a variety of chemical reactivity available, to date, a photo-activated functionality such as diazirine provides the most effective resolution of crosslinks due to in situ initiation and the fast kinetics of chemical steps. Additionally, a carbene formed from diazirine upon UV irradiation shows an equivalent crosslinking reactivity to all proteinogenic amino acids, avoiding biased estimation in the following mass spetrometric analyses. An unnatural amino acid containing a photo-activated residue (UAA) can be genetically and site-specifically inserted into a protein or a photo-activated moiety can be synthetically incorporated into a peptide or a small molecule (Chen et al., 2020; Pham et al., 2013; Yang et al., 2016). Alternatively, a natural amino acid residue such as amine or thiol in a protein is derivatized with N-hydroxysuccinimide (NHS) ester in a heterobifunctional linker that also contains a photo-activated group for a subsequent crosslinking reaction to a binding partner.
A conventional homobifunctional linker bearing NHS esters is still widely applicable not only for its ease of manipulation but also for newly introduced functionalities like MS-cleavable moiety, isotope labeling, or well-defined spatial constraints which, to a greater extent, facilitate automated mass data analyses and computational modeling with high confidence (Iacobucci et al., 2018; Ihling et al., 2020; Rappsilber 2011). For more details about the workflow of CLMS, crosslinking chemistries, and data processing which are beyond the scope of this review, interested readers are referred to several recent reviews (Chavez et al., 2019; Holding 2015; Piersimoni and Sinz 2020).