INTRODUCTION
The term “allostery” was coined in 1961 by two illustrious pioneers of molecular biology, Monod and Jacob,1 who discussed the inhibition of an enzyme by a compound that is not a substrate. The term was later adopted to encompass all cases where coupling between distant sites of a biomacromolecule, related to either tertiary or quaternary structure elements, affects the biological function.2 The topic thus has a long history. However, thanks to the progress in both experimental and theoretical methodologies, the past decade has seen revolutionary changes in our understanding of the scope and nature of allosteric interactions. This Special Issue covers many aspects of this impressive progress.
The Special Issue includes new theoretical models of allostery addressing how local deformation in the floppy and disordered mechanical network of a protein propagates to create long-range coupled response and functionality of the system. It also discusses computational methods for the identification and mapping of allosteric networks, as well as novel experimental approaches to study allosteric mechanisms, including time-resolved and single-molecule studies. The Special Issue also covers approaches to engineering allosteric regulation to enhance function and facilitate the design of sensors and drugs, allostery and the function of molecular machines (natural and synthetic) and protein complexes (conformational spread), the design of synthetic chemical networks that use allostery in feedback mechanisms, directed evolution of allostery employing enhanced sampling, Markov state models, graph analysis, and machine-learning.
The Special Issue explores the most recent advances in elucidating the molecular mechanisms underlying allostery as well as a broad range of applications involving molecular systems. Many different aspects of allostery are covered in this issue, and they are grouped according to certain themes as follows:
Allostery: historical overview, its ubiquitous nature, and remaining challenges
In his comprehensive retrospective, Eaton clarifies the remaining challenges for describing allostery using statistical mechanical approaches by surveying models that explain both the cooperative binding of oxygen molecules to hemoglobin and allosteric regulation of heme oxygen binding.3
Turan et al.4 exploit MD simulations to study the migration of xenon atoms within the important model protein myoglobin and their occupation of specific internal cavities. They observe a non-additive inter-dependence of Xe binding events in two such cavities, which allows them to declare that myoglobin is an allosteric protein, echoing the ubiquitous nature of allostery in all proteins.
Atomistic MD and analysis of communication pathways—structure, dynamics, and energetics
Konovalov et al.5 perform allosteric pathway analysis based on molecular dynamics simulations and combine it with biochemical experiments to investigate the autoinhibition of protein phosphatase 2A. To facilitate the allosteric pathway analysis, they introduce a path clustering algorithm for lumping pathways into channels. The computational findings for several mutants (E198K and E200K) agree with biochemical data and provide insight into how disease mutations in spatial proximity alter the enzymatic activity by surprisingly different mechanisms.
To study the dynamics of ABL kinase, Krishnan et al.6 use multiscale simulations that span the range from atomistic molecular dynamics simulations to Markov state models. Dynamic network methods are also employed to probe the conformational landscape and allosteric dynamics of the protein. This integrative study identifies conserved regulatory hotspots that modulate kinase activity, and interactions between different sites, such as the allosteric pocket and the ATP binding site.
Poudel and Leitner7 address the microscopic origin of allosteric communication. They calculate the change in the energy current between non-polar residue pairs that remain intact during the allosteric transition of β2AR, and they further associate it with the inverse mean square fluctuation and the entropy change between the residues.
Alternative analysis of coupling pathways: graph-based algorithms, augmented with evolutionary data
Kelly et al.8 use methods from evolutionary information, graph-based networks, machine learning, and atomistic molecular dynamics simulations to characterize the functionally important allostery in spastin, a microtubule-severing protein, in the presence or absence of certain ligands.
Madan et al.9 provide a review of the evolution of the field of allostery from a focus on conformations to a broader view that includes dynamics. They then turn their attention to the latter and use the example of protein kinase A to demonstrate modern, graph-theory based analysis methods that can identify ligand-induced correlated motions.
Aguti et al.10 test and challenge three computational approaches for the estimation of correlation and apply them to three proteins of pharmaceutical interest: the androgen A2A receptor, the androgen receptor, and the EGFR kinase domain. The results demonstrate a partial agreement among the methods.
Software packages for communication pathways, coupled conformational changes
Maschietto et al.11 introduce a Python software package, MDiGest, that provides a comprehensive toolkit for studying allostery from MD simulations of biochemical systems. MDiGest allows for comparisons of networks and community structures that capture physical information relevant to allostery, such as correlations of atomic displacements or dihedral angles, as well as the implementation of an approach based on the correlation of Kabsch–Sander electrostatic couplings.
Crean et al.12 develop the Python package “Key Interactions Finder (KIF),” which enables users to identify non-covalent interactions that are strongly associated with a conformational change of interest. Methods from statistics or machine learning can be applied to identify and rank the interactions and residues distributed throughout the protein.
Coarse-grained models for protein allostery: elastic network models for the analysis of couplings
Mugnai and Thirumalai13 construct an elastic network model, a computationally simple but widely used approach to model allosteric communication, and use the structural perturbation method to analyze allosteric coupling in the complex between ACE2 and receptor-binding domain (RBD) of the S protein of SARS-CoV-2. The key result is that complex dissociation opens the ACE2 substrate-binding cleft located away from the interface and that fluctuations of the ACE2 binding cleft are facilitated by RBD binding.
Guarnera and Berezovsky14 present a network-based model with a sequence dependence added in consideration of allosteric communication by combining the structure-based statistical mechanical model of allostery with the Miyazawa–Jernigan residue–residue potential. Applying the model in the analysis of five classical allosteric proteins, they find that it is necessary to consider two major determinants: (i) the free energy exerted by the allosteric site on the regulated site and (ii) the background (average) change in the dynamics of the overall structure.
Zheng15 use the normal modes of an elastic network model to predict cryptic binding sites for allosteric modulators in four classical examples of allosteric regulation, i.e., the GluR2 receptor, GroEL, a G-protein coupled receptor, and myosin.
Factors that contribute to allostery at multiple scales
Ito et al.16 investigate the first stage of the reaction of the enzyme tryptophan synthase during catalysis by carrying out minimum-energy pathway searches based on a hybrid quantum mechanics/molecular mechanics (QM/MM) model, which leads them to identify a particular side-chain near the ligand that plays an essential role in the allosteric regulation.
Salehi et al.17 employ MD simulations to analyze the local hydration around hemoglobin and its change in the T-to-R transition, illuminating the role of water at subunit interfaces in the conformational transition and allostery of hemoglobin.
Xie et al.18 investigate the connection between allosteric regulation and protein topology. They suggest a dependence of the locations of allosteric sites, including cryptic ones, on topology and further develop a novel approach, TopoAlloSite, to predict the location of allosteric sites.
Connection between allostery and folding/stability, roles of disordered regions/domain linkers
In an attempt to merge the pictures of allosteric communication with that of the cooperativity of protein folding, Prabhakar et al.19 replace the concept of the active-site reaction by the reaction pathway that leads to protein folding, and the presence or absence of an effector by mutant-vs-wild type changes in key residues.
By carrying out single-molecule FRET measurements on the nuclear coactivator binding domain (NCBD) of the protein CBP interacting with its binding partners, Buholzer et al.20 demonstrate that the modulation of NCBD in its folding and binding capacity with its partners is highly complex, affected by diverse factors, such as the peptidyl-prolyl cis/trans isomerization, and by phosphorylation.
Dey and Zhou21 use molecular dynamics simulations to study allosteric interactions in the intrinsically disordered protein N-WASP, which is involved in actin polymerization. They are interested in the release of the acidic tail (A tail) from autoinhibition through binding to PIP2 in membranes. Interestingly, they find that the fully open, uninhibited conformation of the A tail is relatively rarely populated even after binding to PIP2. However, a semi-open conformation involving only a partial release of the A tail seems to be enough to promote binding of the protein Arp2/3 to N-WASP and initiate actin polymerization.
Cruz et al.22 use replica exchange discrete molecular dynamics simulations, single-molecule fluorescence experiments, and other biophysical tools to understand how domain tethering in the protein FoxP1 impacts dimerization at its ZIP and FKH domains and how DNA binding allosterically regulates their dimerization. The results suggest that upon DNA binding, the interdomain linker plays a crucial role in the gene regulatory function of FoxP1.
Dayananda et al.23 probe by molecular dynamics simulations the allosteric communication underlying the conformational transitions between a disordered conformation in the “closed” pore state and an ordered hairpin in the “open” pore state of the double-ring protease ClpP. Community network analysis reveals a switch between intra- and inter-protomer coupling in the open–closed pore transition.
The power of integrating experiment and computations for better defining conformational ensembles, transitions, and time scales
VanSchouwen et al.24 use the backbone hydrogen-bond strengths within protein kinase G (PKG), measured with the help of hydrogen/deuterium fractionation factors, as indicators of the perturbations arising from both ligand binding and allostery. The results suggest that frustration may contribute to the reversibility of allosteric conformational shifts by avoiding over-rigidification that may otherwise trap a protein domain in its active state.
Belato et al.25 employ NMR and computation to offer an atomic level understanding of Cas9 domain dynamics. In particular, they study the effect of mutation on the allosteric crosstalk and identify an allosteric hotspot in the HNH domain of the thermostable GeoCas9. A mutation of the hotspot disrupts a salt-bridge network, which in turn alters the structure, the time scale of allosteric motions, and the thermostability of the HNH domain.
Heckmeier et al.26 design photoswitchable versions of peptides derived from two of the protein binders of the Myeloid Cell Leukemia 1 (MCL-1) protein. They then use transient infrared spectroscopy to probe the dynamics induced in the peptides and in MCL-1 following switching, identifying a rich spectrum of conformational changes on time scales from nanoseconds to microseconds.
Samanta et al.27 combine experiment and computation to study allostery in bacterial biotin ligases. Intriguingly, mutations in the catalytic sites of these enzymes affect their ability to homodimerize and bind to DNA. Community network analysis is used to identify the networks of interactions that facilitate this complex allosteric behavior.
Finally, Liu et al.28 use molecular dynamics simulations to assist in the identification of subtle conformational changes of the periplasmic ferric binding protein A from small-angle x-ray scattering (SAXS) experiments. Fitting conformations from the MD simulations to SAXS data allows them to quantify the populations of conformational substates of the protein.
CONCLUSIONS
This Special Issue thus highlights the latest advancements in theoretical models of allostery, which focus on understanding how localized deformations within the flexible and disordered mechanical network of a protein lead to long-range coordinated responses and enhanced functionality within the system. In addition, cutting-edge computational methods are being utilized to identify and map out allosteric networks, providing valuable insights into the intricate communication pathways that govern allosteric regulation. Furthermore, novel experimental approaches are being developed to investigate allosteric mechanisms in greater detail, offering new opportunities to unravel the complex dynamics underlying protein function. Overall, this Special Issue sheds light on the diverse strategies being employed to unravel the mysteries of allostery and to pave the way for future breakthroughs in this field.
