Most work was towards both objectives together. For example, initial work focused on demonstrating the utility of combining computational biophysics and CryoEM image processing for understanding dynamic biological macromolecular systems and included a study on the effects of SARS-CoV-2 spike mutations originally found in the Spanish variant that we published in PLOS Pathogens together with our collaborators in Barcelona as well as others in Madrid and Valencia through the Spanish interdisciplinary thematic platform for global health (Ginex, Marco-Marín, Wieczór, Mata, Krieger, Ruiz-Rodriguez, et al. 2022). This demonstrated that classical CryoEM image processing has the potential to reveal subtle differences that can be explained by molecular simulations, showing that the spike mutation A222V triggered increased dynamics of the open spike that enabled further adaptation in later variants including a particularly successful Delta subvariant.
Another component of the early work was an exploration of a new way of describing molecular motions as deformations called Zernike3D that we alluded to in our proposal with first signs of usefulness for continuous image classification. We first confirmed its utility in capturing structural changes using reconstructed volumes and atomic structures including those based on molecular simulations in a study published in the International Union of Crystallography Journal (IUCrJ, also for CryoEM; Herreros et al. 2021). We later refined our pipeline for using it for continuous classification of 2D images and published the complete method in Nat Communications (Herreros et al. 2023).
These starting points enabled the main work towards objective 1: the creation of a plugin within the Scipion workflow engine software of the host lab for the ProDy Python package for protein dynamics described in detail in a paper in the International Journal of Molecular Sciences (IJMS; Krieger et al., 2023) and the integration of Zernike3D, ProDy and other continuous heterogeneity analysis methods into a shared framework called the Scipion Flexibility Hub described in a publication in Acta Crystallogr D Struct Biol (Herreros et al. 2023). The ProDy plugin enabled the creation of more interpretable landscapes through principal component analysis of refined structures from Flexibility Hub including comparison to existing structures in the protein data bank and connection to additional computational biophysics approaches. We are also exploring better methods for simulations and landscape analysis with our collaborators in Barcelona and new collaborators Dr Pilar Cossio at the Flatiron Institute in New York City and Dr Erik Thiede at Cornell University in New York State.
The main novel results towards objective 2 were insights into mechanisms for the unique slowness of AMPARs containing the GluA1 subunit and its activity dependence for synaptic integration, which has been published in Nature with the researcher as co-first author (Zhang, Ivica, Krieger et al., 2023). These properties along with its calcium permeability enable the integration of rather different receptors into synapses under certain conditions that change their signal processing properties, driving long-term plasticity which is essential for learning and memory. Understanding the principles behind this process at the level of macromolecular structure and dynamics will guide further studies into their role and behaviour in healthy and disease conditions at the cell and tissue levels, and facilitate the development of better and more specific therapeutics against diseases such as Alzheimer’s and Schizophrenia.
