The alarming rise of resistant bacteria urgently calls for the design of novel antibiotics that are robust to resistance development. Ideal templates could be peptide-antibiotics that destroy the bacterial cell wall by binding to its essential membrane-anchored precursor lipid II at irreplaceable phosphate groups. Indeed, these peptide-antibiotics kill the most refractory bacteria without detectable resistance and without showing cytotoxicity. However, due to the challenge of studying antibiotic-receptor complexes in membranes, the available structural information on peptide–lipid II interactions is extremely small and native binding modes could, so far, never be visualized. Moreover, it is virtually unknown how native variations of the lipid II structure and the membrane composition across bacteria modulate lipid II binding antibiotics. Altogether, this lack of knowledge critically limits the development and application of lipid II attacking drugs. We propose to study the modes of action of two signature peptide-antibiotics, which are representative for the two major peptide-lipid II binding modes, in physiological conditions and at atomic detail: 1. The lantibiotic nisin kills bacteria via a mechanism called targeted pore formation. Nisin is by far the most studied lipid II binding peptide, however, despite of extensive research, structural models of the nisin-lipid II pore have been lacking hitherto. Solving the nisin pore structure will hence have a major impact on antibiotic design. 2. The defensin plectasin with strong potential to kill multi-drug resistant Streptococci, Staphylococci, and other superbugs. Plectasin is representative for a large number of structurally similar antibiotics, and hence, understanding its binding mode to lipid II will hence have broad relevance for drug design. To solve these highly challenging structures in physiological conditions, we will use a cutting-edge solid-state NMR (ssNMR) approach. We will develop and use ssNMR methods based on the groundbreaking techniques 1H-detection and Dynamic Nuclear Polarization, which both can tremendously increase spectral sensitivity. In addition, we will take advantage of a worldwide unique NMR setup in Utrecht (1200 MHz & 950 MHz magnets, 400 & 800 MHz DNP). Together, the combination of top-notch methods and excellent hardware will provide us the spectral sensitivity to study the mode of action of peptide-antibiotics at highest-accuracy and in native cellular conditions. We will integrate these ssNMR data with complementary MD simulations to fundamentally understand how lipid II is targeted. Our project will mark a pioneering advancement for our knowledge of lipid II binding peptides, which is a critical step towards the pharmaceutical use of these powerful antibiotics. Building-up on this project, in collaboration with expert groups, we will use our high-resolution data for rational drug design.

Protein Detective: An Automated Pipeline for Identifying Unknown Proteins in Cryo-Electron Microscopy Densities. Visualising proteins, ideally in their cellular environment, is a first step toward understanding their function. This can be done using cryo-electron microscopy. The resulting 3D images are, however, not always of sufficient quality to identify the proteins hidden in those. In this project the researchers will build an automated “protein detective” pipeline that will combine all available clues on the system, together with AI-generated protein models to build reliable 3D models of intricate molecular machine in which all components are identified.

Structural biology is vital in understanding details of biological molecules, shaping our knowledge of how cells work with applications in medicine and more. During COVID-19, the WeNMR computational services have been crucial in modelling viral interactions. To continue this work, continued access to the Dutch high throughput computing grid resources is needed, supporting global research and collaborations in structural biology.

Nuclear Magnetic Resonance Spectroscopy (NMR) and Imaging (MRI) exploit the magnetic properties of atoms which allows for non-invasive characterization. In the last project period, most of the technical innovations and upgrades could be realized. These improvements led, for example, to insight into the workings of novel antibiotics as well as energy storage and solar energy conversion materials. These scientific advancements are broadly relevant including fields such as biomedicine and the transition to a circular economy. Lastly, the uNMR-NL network also connected to new user communities and engaged in national and international alliances, in the fields of NMR/MRI and beyond.

Membrane proteins code one third of all genes in all organisms. They are involved in any known metabolic or signalling pathway and constitute half of all drug targets. To date, since membrane proteins are inappropriate for conventional structure determination techniques, only one percent of all known protein structures account for membrane proteins. Moreover, membrane proteins act in dynamic interplay with their membrane environment, which is hitherto poorly understood on a molecular base. This crucial protein?-membrane dialogue shapes protein structure, determines protein orientation relative to the membrane (known as topology), modulates protein?-protein communication by steering protein localisation and assists protein folding. Together, structure, topology and protein-?membrane interplay comprise the supramolecular structure, i.e., the complete picture of a membrane protein. This proposal aims to study the supramolecular structures and functions of large integral membrane proteins using a strongly complementary joint-approach of solid state Nuclear Magnetic Resonance (ssNMR) and Molecular Dynamics (MD) simulations. SsNMR allows investigating membrane proteins in their native environment at atomic-resolution. Large membrane proteins, however, only give access to spectra of low quality. This shall be solved by tailoring ssNMR methods to work at optimal sensitivity and resolution provided by ultra-high magnetic fields and spinning frequencies, proton-detection and dynamic nuclear polarisation. SsNMR information will be complemented by MD simulations to refine and validate protein structure and to establish protein topology by back-calculating ssNMR observables (like the protein water-access) over MD trajectories. I will particularly focus on protein-membrane interactions by virtue of combined ssNMR-MD investigations with ssNMR experiments especially designed to be correlated to MD trajectories. This approach shall be applied to model membrane proteins with known crystal structures (ion channel KcsA, photosensor ASR) and employed to study beta-barrel assembly machinery BamA (88 kDa), which is a key player in membrane protein folding of unknown high-resolution structure.