4. Probing biological interfaces by integrative approaches
While CLMS and HDMS may stand alone as the techniques to map the biological interface, their integration aided by computational modeling with the dataset obtained from X-ray crystallography, EM, or NMR can significantly improve spatial resolution and coverage of binding interfaces to provide comprehensive illustrations. For example, Zhang et al. first implemented CLMS and HDMS together to extract previously unknown information on the epitope and paratope interface of a programmed cell death-1 (PD-1) and the corresponding antagonistic antibody. In the next round of refinement, the suggested critical binding residues and distance restraints were utilized to build high-confidence binding models through molecular docking onto the crystal structure (Zhang et al., 2020). While the epitope-paratope relationships revealed by the approach were generally comparable with those assigned by the crystal structure, one cryptic loop in PD-1 that had not been crystallographically resolved due to its flexible nature was found to be a non-epitope. The study demonstrated a complementary role of CLMS and HDMS for accurate and detailed examination of a binding interface. The same research group described a similar integrated approach combining CLMS, HDMS, and molecular docking to probe the binding interface of interleukin 7 (IL-7) complexed with its receptor IL-7Rα. While the predicted model was generally in accordance with the crystal structure, the approach newly discovered the C-terminal binding region of IL-7, highlighting the value of integrative approaches to obtain a high-confidence structural model (Zhang et al., 2019).
Interpretation of HDMS data using a predetermined crystal structure as a template provides greater insight into reversible changes in regional flexibility of binding interfaces. A junctional epitope antibody VHH6, specifically recognizing a neo-epitope created only at the junction in which IL-6 and gp80 are interlocked, was considered a molecular clamp as shown in the ternary crystal structure. To understand the effect of VHH6 clamping on the structural flexibility of the junctional interface of IL-6-gp80, Adams et al. compared the amounts of deuterium exchange therein in the presence or absence of VHH6 (Adams et al., 2017). The presence of VHH6 increased the rigidity in the local region spanning the junction, stabilizing a transient interface between IL-6 and gp80.
Likewise, advances in CLMS techniques are better refining a low-resolution interface solved by other methods into medium- to high-resolution details. By applying multiple orthogonal crosslinking chemistries to a target protein complex, Mintseris and Gygi could attain a higher crosslinking density and improved sequence coverage (Mintseris and Gygi 2020). Self-consistent analytic results could be mapped onto cryo-EM models to define the interaction interface with high resolution and confidence.
A dramatic contribution of HDMS in harmony with CLMS to modeling the interface of an unstable macromolecular complex was presented by Shuka et al (Shukla et al., 2014). While the human β2adrenergic receptor (β2AR) and β-arrestin-1 had crystal structures reported individually, the β2AR-β-arrestin-1 complex was intolerant to experimental conditions in X-ray crystallography or EM with only low-resolution map for the overall conformation available. Remarkably, constraints provided by HDMS and CLMS were mapped onto the preexisting data in an integrative manner, resulting in the three-dimensional reconstruction of the complex that displayed an unexpectedly extensive interface and a crucial involvement of the finger loop in β-arrestin-1 in the complex formation. More recently, such a combined approach investigated the mechanism of client binding of the periplasmic chaperone SurA and identified dynamic inter-domain interfaces that underwent substantial structural reorganization in response to the substrate binding interface being occupied by a client OmpX (Calabrese et al., 2020).