The work has been published in the journal PROTEINS: Structure, Function, and Bioinformatics. ATP synthase utilizes the membrane potential to synthesize adenosine triphosphate (ATP) molecules, the universal "fuel cell" of living systems. This enzyme can be likened to an electric motor: it has a stationary part — the stator, and a rotating part — the rotor, which is also referred to as the c-ring. The ions transported by ATP synthase across the membrane facilitate the rotation of the rotor and the synthesis of ATP by the stator. The rotor itself consists of repeating units (c-subunits), the number of which determines how many ions will pass through the membrane per rotation and what potential is sufficient for the protein to function.
ATP synthases are known to contain between eight and 17 c-subunits, with studies often utilizing X-ray crystallography or cryo-electron microscopy. However, advancements in machine learning methods have led to the development of several groundbreaking approaches, including those for predicting protein structures. The developers of the most well-known of these — AlphaFold — were awarded the Nobel Prize in Chemistry in 2024. Nonetheless, the standard AlphaFold does not perform well with large protein complexes, which include ATP synthases.
In their study, researchers from MIPT proposed a new approach based on AlphaFold that confidently predicts the number of subunits in protein complexes with rotational symmetry. For known data, the correlation between predictions and experimental values exceeded 90 percent. Furthermore, the approach proved to be very fast, allowing for the estimation of the number of subunits in ATP synthases across a wide variety of living organisms.
“The most surprising aspect of our findings is the prediction of the existence of microorganisms in nature with very large c-rings in ATP synthases, which may contain up to 27 repeating subunits! Previously, experiments had found a maximum of 17 c-subunits. Importantly, molecular dynamics simulations confirm the results of these predictions,” says Stepan Osipov, the lead author of the published work and a graduate student at the Landau Institute of Physics and Research at MIPT.
Why does nature require such unusually large c-rings when more "standard" rotors also demonstrate high efficiency? The answer to this question is closely related to the environments in which these organisms inhabit. They may exist in high temperatures or aggressive environments where they constantly lose ions through their membranes. Alternatively, it could be due to a low energy potential across the membrane — in which case, a high repetition rate in the c-ring aims to better "catch" rare protons or sodium ions. Such large rotors will “consume” more protons per rotation but will allow for ATP synthesis at significantly lower transmembrane potentials. To verify these assumptions, experimental studies of real microorganisms, which calculations suggest should have such unusual c-rings in their ATP synthases, will be necessary.
“This also means that new opportunities are opening up for many areas of biotechnology. For instance, if we can learn to design molecular “motors” with specified parameters, we can deliberately adjust the efficiency of energy exchange in cells, creating microorganisms for the production of various compounds in conditions where traditional strains cannot survive,” comments Alexey Vlasov, a senior researcher and acting head of the Laboratory of Molecular Cell Biology and Optogenetics at MIPT.
“Our work highlights the importance of computational algorithms based on machine learning and artificial intelligence in modern structural biology, as well as opens new avenues for the search for interesting proteins in genomic data and the engineering of new proteins with desired properties,” summarizes Ivan Gushchin, executive director of the Center for Research on Molecular Mechanisms of Aging and Age-Related Diseases at MIPT.
The work was supported by the Russian Science Foundation and the Ministry of Science and Higher Education of Russia.