Computational elucidation of allosteric communication in proteins for allosteric drug design

Highlights

• Allosteric signals underlie the functional mechanisms of allosteric drugs.

• Elucidation of allosteric signals facilitates rational allosteric drug discovery.

• Computational methods supply promising tactics to characterize allosteric signals.

• Bioinformatic tools delineating allosteric signals promote allosteric drug design.

Allosteric modulators target topologically distal allosteric sites in order to modulate orthosteric sites, providing enhanced specificity and physiochemical properties. Harnessing allostery for drug discovery is an emerging paradigm in modern pharmaceutics. Allosteric regulation substantially depends on the propagation of allosteric signaling. Delineating allosteric signaling pathways is therefore one of the leading prerequisites for allosteric drug discovery. Allosteric signal transduction is subtle and dynamic, posing challenges for characterization through traditional experimental techniques, but computational strategies promise to provide a solution to this problem. Here, we comprehensively review bioinformatic methods for elucidating allosteric communication, along with their successful applications in allosteric drug design. Current challenges and future perspectives are also discussed. We aim to provide guidance for the future application and optimization of these computational strategies, thereby promoting rational allosteric drug discovery.

Introduction

Allostery, or allosterism, is defined as the regulation of an orthosteric site by the distal allosteric site.1., 2., 3., 4. Perturbations at allosteric exosites, such as ligand binding,⁵ mutations,⁶ post-translation modifications,⁷ and even light absorption,⁸ act through the propagation of allosteric signals to confer subtle regulatory impacts on orthosteric pockets, finetuning their structural dynamics and biological activities. Allosteric regulation is pervasive in every aspect of life science.4., 6., 9., 10. Current biophysical research efforts have unveiled the omnipresence of allosterism, extending throughout the spectrum of proteins, from single chain monomers^2,9 to huge macromolecule assemblies,10., 11. and even to large-scale molecular machines.¹² Allosteric regulation is of utmost importance in manipulating a plethora of pivotal cellular events, such as enzyme catalysis, gene expression and protein modification, under both physiological and pathological conditions, including different states of energy metabolism,¹³ cancer pathogenesis,14., 15. and drug resistance.6., 16.

Considering its prevalence and significance, allosteric regulation is emerging as a novel paradigm in pharmaceutical research, opening a new avenue for drug discovery.6., 17., 18. Traditional drug development usually deploys the binding of molecules to conserved orthosteric pockets, which inevitably suffers from both compromised drug selectivity and direct competition with the high-affinity endogenous orthosteric ligands, as well as from receptor desensitization originating from orthosteric interactions. These difficulties can be circumvented by the use of allosteric modulators, which tailors protein activities either on its own or in concert with endogenous ligands, to target evolutionarily less conserved allosteric exosites. This approach offers enhanced specificity, improved physiochemical properties, and reduced off-target side effects.2., 9., 17., 19., 20. Thus, the harnessing of allosterism for drug development is an emerging direction in pharmacological research.

Allosteric signals constitute the mechanistic bases for the subtle regulatory effects of allosteric drugs.3., 21., 22., 23. Intrinsically, allosteric signaling propagation refers to the transduction of perturbations through inter-residue steric leverages. Through residue–residue interactions, allosteric signals that are initiated at allosteric sites communicate dynamically throughout the protein structures, further impacting their active sites.

Our understanding of the generic nature of allosteric signals is evolving rapidly, with the significant progresses in biophysical and structural biology. Traditionally, allosteric regulation has been considered to be a two-state system with an on–off switch, or simplified as positive, negative, and neutral modulation. Latest investigations, particularly in the field of G-protein-coupled receptors (GPCRs), have revolutionized such a concept. Allosteric signaling is modulated by the binding of different ligands to different protein conformations, thereby tweaking diverse allosteric communications and resulting in multiple downstream outputs. This theory establishes allosteric proteins as ‘microprocessors’ that monitor various states and multiple pathways of allosterism.24., 25. Put another way, allosteric signals were previously established as unidirectional, stemming from allosteric sites towards orthosteric sites. However, cutting-edge studies have revealed a bidirectional system, in which the dynamic properties of orthosteric and allosteric pockets are correlated and they can each tweak structural and biological events that occur at the other site.26., 27., 28. Collectively, our expanding knowledge of protein allostery promises the possibility of allosteric drug discovery, prompting the prospect of allosteric pharmacology.

As one of the prerequisites for rational allosteric drug design and optimization, untangling allosteric signals is of utmost significance, receiving intense interest from both academia and the pharmaceutical industry. Owing to the dynamic and subtle nature of allosteric signals, however, it can be extremely challenging to probe these pathways using existing experimental approaches. Therefore, computational methodologies are an emerging alternative trend for the analysis of allosteric signaling pathways.²⁹ A plethora of computational methods have been developed to decipher allosteric signal pathways, and according to the underlying theoretical bases, they are categorized into molecular dynamics (MD)-based, normal mode analysis (NMA)-based and topology-based methods. A series of software pipelines, as exemplified by AlloReverse,²⁷ structure-based statistical mechanical model analysis (SBMMSA),³⁰ and CorrSite,³¹ are also entering the field of allosteric signal investigation, with a particular focus on bidirectional allosteric communications. These state-of-the-art computational methodologies contribute to the elucidation of protein allosteric signals, deepening our understanding of allosteric phenomena and shedding light on allosteric drug discovery.

Here, we survey the latest advances in computational studies of allosteric signaling. Their diverse theoretical backgrounds are comprehensively reviewed, and a systematic description of their successful application is also presented. Importantly, we summarize the challenges in their development and optimization. We aim to increase the awareness of computational investigations of allosteric communication and to guide future methodological development.