Molecules often work together in complexes to carry out their functions in the cell. But how do they get together in such a dynamic environment? A structural study follows proteins as they meet their partners.
Living cells, particularly during growth and proliferation, need regulatory processes of great sensitivity and high specificity. To achieve this, signal-to-noise ratios must be high when information is received and transmitted between the cell surface, the cytoplasm and the nucleus. Just like electrical and engineering control systems, living cells have complex signalling pathways that are moderated by feedback mechanisms. It is becoming increasingly clear that most switches, transducers and adaptors in living systems are created by the assembly and disassembly of multi-component complexes of proteins, nucleic acids and other molecules1. On page 383 of this issue, Tang et al.2 follow the formation of such complexes, and provide structural evidence for the transient interactions necessary to build them Footnote 1.
How do the molecular assemblies in cells achieve the required sensitivity and specificity? Efficient signal transduction must maintain fidelity and decrease noise while amplifying the signal. So the solution cannot be explained in terms of tightly bound, enduring molecular complexes, because the signals could not then be turned off. Rather, it seems to lie in first assembling weak binary complexes, and then using cooperative interactions to produce multi-component complexes in which the weak interactions are replaced by much stronger and more specific interactions3.
Although weak, nonspecific, transient complexes could give rise to a noisy system, such 'encounter complexes' might be exploited so that interaction partners do not have to be found afresh in the busy milieu of the cell, thus increasing the rate of formation of specific binary and higher-order complexes. Essentially, the partners bump into one another and are held loosely, allowing them time to become reorientated and repositioned on the surface or to adjust their shape to fit together more tightly. Recent studies are beginning to describe the dynamics of the assembly processes and to show that nonspecific, transient collisions play an important role in macromolecular associations.
Tang et al.2 study a bacterial signalling system called the phosphotransferase system, and describe evidence for rare encounter complexes involving the amino-terminal domain of enzyme I and the phosphocarrier protein HPr. The authors mutated one of the proteins at specific sites, introducing a paramagnetic atom that causes local perturbations in the magnetic field. Using a technique that is sensitive to these perturbations (intermolecular paramagnetic relaxation enhancement), they show that a variety of transient, nonspecific encounter complexes of the two proteins are present under equilibrium conditions. Neither the final, specific complex structure (previously solved by NMR) nor any other single alternative conformation alone can account fully for their data.
This paramagnetic approach has already proved helpful for detecting rare equilibrium intermediates in protein–DNA interactions4. But Tang and colleagues2 go farther, and use computation-based structural models to build an ensemble of conformations that satisfy the experimental data. They thus provide structural evidence for the rare encounter complexes, and suggest that electrostatic interactions have a major role in forming these intermediates (Fig. 1; see also Fig. 3a of the paper2 on page 385).
Interactions between proteins are highly prevalent in biology, and the sorts of transient interaction revealed by Tang et al. are not captured in the crystals that are the basis of many structural studies, so this dynamic approach will be useful in many other cases. Indeed, populations of alternative encounter complexes will probably be even more evident in highly transient protein–protein interactions5, such as those involved in the inter-protein electron-transfer processes that occur in cellular metabolism. For instance, in the interaction between the CuA domain of the ba3 oxidase and its interaction partner cytochrome c552, a single complex cannot fully account for the restraints on the amino acids observed in NMR-based experiments, leaving open the possibility that an ensemble of conformations coexist in a weakly complexed state6.
Many fundamental questions remain about the structure and energetics of these encounter complexes. The role of long-range electrostatic forces in bringing molecules together before they collide has been studied from both experimental and theoretical viewpoints7,8, and it is clearly a major driving force in the formation of the rare encounter complexes described by Tang and colleagues. Less studied, however, is the role of short-range desolvation effects — where the protein–protein interactions force solvent molecules away from the proteins — during and after collisions. It has been suggested that hydrophobicity is involved in reorientating the molecules to form the final, productive complex5. Computational simulations also show that desolvation energy plays a part in orientating the encounter complexes' interacting subunits around the final, specific complex state9.
Tang et al.2 describe structural features of the encounter complexes (assuming that they interact as rigid bodies), and their results are consistent with the general idea of a funnel-shaped binding-energy well that narrows as the two proteins approach one another. This implies that there are many possible routes for arriving at the final complex at the bottom of the energy well, and that these are determined by transient interactions between the partners in the encounter complexes, with the pathways converging as they get lower in energy and closer to the final complex. Indeed, rigid-body docking calculations based on optimization of the binding energy of the interacting molecules9,10,11 already describe a pool of alternative encounter complexes on the way to forming the functional complex. Whether these ensembles of orientations reflect the true binding-energy landscape will depend on the accuracy of the energy description of these computer models and the efficiency of the sampling method, an area of current debate. Molecular-dynamics simulations show that some encounter complexes could be sufficiently long-lived for their side chains to acquire a variety of conformational states, some of which are similar to those in the final, functional complex12. But it remains to be seen how many of the minor species are true productive encounter complexes, and which are the preferred paths to the specific binding mode of the final complex.
It is now apparent that rare encounter complexes might control not only the kinetics of the assembly process, but also the way the complex is put together and hence its cooperativity. Furthermore, the population of non-specific complexes can be restricted by the order in which the different subunits are assembled. Greater understanding of the route to longer-term relationships between molecules will no doubt emerge from integrating a wide variety of experimental data with theoretically sound computer modelling13,14 of their brief encounters.
This article and the paper concerned2 were published online on 15 October 2006.
About this article
Theoretical Chemistry Accounts (2010)