1. Introduction
Viability of all forms of life is crucially reliant on homeostasis of protein interactome. Therefore, elucidation of networks involving protein-protein or protein-ligand interactions is a cornerstone for understanding the complexity of biological processes. Additionally, comprehensive informatics of protein interactions serves as a useful template for drug discovery and drug repositioning (Payandeh and Volgraf 2021; Scott et al., 2016; Wang et al., 2021). A multitude of methods have been developed to study the protein interactome with high throughput and accuracy (Rao et al., 2014). While the protein interactome illustrates the global network and relation of proteins and ligands, mechanistic insight at molecular level into each set of protein interactions further requires high-resolution technologies capable of probing an array of critical motifs and residues, defined herein as an interface, that are collectively responsible for accurate molecular recognition and energetic stabilization.
Elementary bioanalytic methods readily available in most molecular biology laboratories are size exclusion chromatography (SEC) and native PAGE which examine if biomolecules under investigation do recognize each other, albeit with little information of binding interfaces. These methods macroscopically detect a change in mobility of biomolecules through a matrix, attributed by size increase and charge variation upon molecular association. Quantitative measurements of kinetics and thermodynamics parameters involved in molecular interactions require real-time techniques. Surface plasmon resonance (SPR) and bio-layer interferometry (BLI) employ sophisticated optical biosensors that monitor minute changes of resonance signals, or light interference quantitatively dependent upon intermolecular binding events (Dysinger and King 2012; Rich and Myszka 2011). Data processing from repetitive measurements in different concentrations of an analyte yields the rates of association and dissociations as well as the strength of the interaction in terms of affinity (KD). Microscale thermophoresis (MST) technique is quick to derive KDfrom variations of a molecular movement of a capillary-filled, fluorescently labeled ligand throughout microthermal gradient, affected by its association with an analyte and subsequent changes in physical properties (Seidel et al., 2013). Isothermal titration calorimetry (ITC) is a unique method that investigates the interaction between biomolecules without labeling or immobilization (Pierce et al., 1999). By measuring differential changes in the heat supplied to the sample cell, as compared to the blank cell, in which two binding partners are incrementally mixed, ITC provides useful parameters related to the bimolecular interaction including binding stoichiometry, enthalpy and entropy changes as well as affinity. Each quantitative method has its advantages, and might serve as a standard technique depending on the amount, purity, and physicochemical properties of biomolecules under investigation. Despite being more informative than SEC and PAGE, these methods are still far from defining the binding interface which is mainly responsible for various interaction parameters obtained in the measurements. Structural information of a binding interface at atomic resolution can be derived by implementing X-ray crystallography, nuclear magnetic resonance spectroscopy (NMR), or cryo-electron microscopy (Cryo-EM). While X-ray crystallography is the most predominant in accessibility and practice thus far, the size of the crystallographic interface is usually smaller and the interface water content is higher in general compared to the interface under native environment, referred to herein as ‘biological interface’, due to non-physiological but crystallization-favored conditions inflicted on a protein complex (Schreiber 2020). Recent advances of NMR and Cryo-EM in instrumentation and software reliably extract structural information of protein complexes in pseudo-physiological conditions which can then be processed to reconstruct preexisting interfacial crystal contacts into a near-native state or even newly visualize the binding interfaces not resolved in X-ray crystallography. Cyro-EM and NMR provide unique and additional advantages of facile sample preparation and delivery of structural dynamics data, respectively. Still, variable and dynamic natures of binding interfaces embedded in protein complexes complicates the interpretation of spectroscopic data with reasonable resolution and fidelity, rendering mechanistic details of many protein interactions as yet to be explored.
Aside from the high resolution, another critical attribute expected from structural information of protein complexes is the high fidelity which presents the most probable and representative snapshot of biological interfaces under in-solution native environment, contrasting to a static condition in which X-ray crystallography and electron microscopy are generally best suited. Despite high resolution provided by static methods, however, a static interface features non-physiological water content and accessibility due to non-specificity caused by invasive conditions such as a wide range of pH, high ion strength, high protein concentration, all of which are detrimental to obtaining high-fidelity interfacial landscape. In addition to such empirical artifacts, a static interface often loses fidelity due to inherent natures of certain protein interactions which could be dynamic, transient, weak, or mediated by an intrinsically disordered domain that adopts no static conformation. Thus, in combination of X-ray crystallography and electron microscopy, supplementary methods to obtain structural data from the native solution-phase are of great necessity to probe a biological interface with both high resolution and fidelity.
Chemical crosslinking (CLMS) and hydrogen-deuterium exchange (HDMS), both of which are coupled to mass spectrometry, have become unique and complementary analytic methods to address abovementioned unmet needs in structural biology. CLMS and HDMS are particularly valuable for unveiling hot spots in interfaces that otherwise would be cryptic or of poor fidelity, thus enabling the construction of a new interfacial landscape of an unknown protein complex or reconstitution of a preexisting protein interface, and subsequently aid in the structure-based design of novel modulators of binding interfaces with a high success rate (Arkin et al., 2014; Chen et al., 2012). In this review, we briefly describe how CLMS and HDMS map the interfaces of protein complexes with high fidelity and their integration with other structural information to yield native-like high-order structures assembled via biological interfaces. We then discuss recent efforts that successfully employed CLMS and/or HDMS to reveal dynamic, transient, or intrinsically disordered interfaces found in various protein complexes that had been challenging to probe via X-ray crystallography and electron microscopy.
Mass spectrometry-coupled techniques, CLMS and HDMS, provide novel means of landscaping a biological interface. In contrast to abovementioned spectroscopic techniques focusing on electronic or electromagnetic properties of respective atoms, CLMS and HDMS delve into reciprocal behavior of residues with other residues or matters surrounding them under a native condition. The resultant mass changes are detected to calculate corresponding inter-residue distances or solvent accessibility. Although structural information obtained from CLMS and HDMS is generally intermediate in resolution, it provides improved fidelity by repositioning known residues in a static interface or identifying unknown residues residing in a dynamic, transient, or disordered interfacial region.