During this MSCA fellowship, we have demonstrated in vitro that the charge provided by the N-terminal capping of PSMα3 alters its local interactions with model membranes of controlled lipid composition, without compromising its intrinsic ability to form amyloid fibrils, yet with eventual distinct kinetics as probed by Nuclear Magnetic Resonance (NMR) and fluorescence spectroscopy. Atomic Force Microscopy (AFM) studies have revealed that the N-ter capping indeed eventually dictates PSMα3-membrane binding via electrostatic interactions with the lipid head groups: while N-formylated and N-acetylated peptides only bind membranes containing zwitterionic lipids (e.g. DOPC), N-deformylated peptides also deposit on negatively charged membranes (e.g. DOPG/DOPE). Furthermore, PSMα3 insertion within the lipid bilayer is favoured by hydrophobic interactions with the lipid acyl chains only in the fluid phase of membranes and not in the gel-like ordered domains when present. Strikingly, real-time AFM imaging has provided compelling evidence supporting the role of intermediate protofibrillar entities, formed along PSMα3 self-assembly, in its formylated and acetylated forms, in the loss of membrane integrity. Those entities tend to accumulate and insert within the lipid membranes, subsequently leading to membrane thinning in a peptide concentration and lipid-dependent manner. Polarized Attenuated Total Reflectance (ATR) FTIR spectroscopy has further allowed structural characterization of those entities, that, promoted at the membrane interface, induced lipid depletion in specific membranes. They are enriched in α-helical content compared to their monomeric forms, thus suggesting their transition to cross-α (proto)fibrils, as previously demonstrated for PSMα3 on its own. Interestingly, nanoinfrared spectroscopy, a coupling between AFM and IR that enables structural investigation on individual nano-objects, has additionally emphasized that those protofibrils formed at the lipid membrane interface differ in their secondary structures from the mature fibrils of PSMα3. To get closer to physiological relevance, in cellulo experiments were also performed to explore this “oligomeric hypothesis”, i.e. to investigate if intermediate entities could be responsible for cytotoxic activities, as supported by associated lipid membrane damage in vitro. We first revealed that PSMα3 can only fibrillate in a cellular medium deprived of serum, suggesting lipoproteins eventually inhibit the formation of amyloid structures. In such minimal medium, PSMα3 induces a concentration and time-dependent cytotoxicity towards HEK cells, that is higher for the formylated form compared to the acetylated one, likely due to their distinct kinetics of self-assembly. Importantly, cells treated with the monomeric and fibrillary forms of the peptides remained mostly viable over the same timescales, emphasizing the role of fibrillation, and intermediate entities, in mediating lytic activities. Overall, this multiscale and multimodal approach has shed new light on the key roles of N-terminal capping and intermediate self-assembling entities in dictating deleterious interactions of PSMα3 with membrane lipids and living cells, likely underscoring its ultimate cellular toxicity in vivo, and in turn S. aureus pathogenesis.
