RNA-based therapeutics have demonstrated remarkable efficacy and hold great promise for future applications in personalized medicine. The most common delivery systems for these drugs are lipid-based nanoparticles (LNPs), which incorporate ionizable cationic lipids (ICLs) as key components. Among other, ICLs are believed to facilitate endosomal escape of the cargo by interacting with anionic lipids in the endosomal membrane, although the underlying molecular mechanism remains unclear. One prevailing hypothesis suggests that membrane destabilization is mediated by cone-shaped complexes formed between ICLs and endosomal anionic lipids. However, no clear evidence of stable co-localization of anionic and cationic lipids has been presented so far. To address this gap, equilibrium and nonequilibrium molecular dynamics simulations of model membrane systems containing DODMA (ICL), DOPS, or PI3P (anionic lipid) and DOPE or cholesterol (helper lipid) are performed. The results confirm the absence of co-localization at equilibrium but reveal transient formation of cone-shaped complexes during lamellar-to-inverted-hexagonal phase transitions, which considerably accelerates the transition process. These findings suggest that transient lipid-lipid interactions, rather than stable complexes, may play a critical role in facilitating endosomal escape. This mechanistic insight may inform the rational design of ICLs tailored to interact with specific endosomal anionic lipids, thereby enabling more effective and targeted delivery strategies for RNA-based therapeutics.

Messenger RNA-based pharmaceuticals have recently demonstrated their po- tential with the successful COVID-19 vaccination campaign. They are now undergoing clinical trials for a broad range of therapeutic indications, including cancer and genetic diseases. While in the COVID-19 vaccines selected lipid-based nanoparticles (LNPs), have proven to be successful, there is a need of alternative delivery systems matching efficacy and safety requirements in a broad range of applications. Polyplexes, formed by self-assembly of cationic polymers with the anionic nucleic acids, constitute a valuable extension to the technological space represented by lipid-based systems. For medical applications, it is crucial to pre- cisely control both the number of encapsulated RNA molecules in the polyplexes and their shape. Here, we use molecular dynamics simulations of a coarse-grained model of the system to show that the single most important factor characterizing it is the total ratio of charges on polyelectrolytes and RNA during the prepara- tion of the nanoparticles. Close to the isoelectric point, the polyplexes are large, whereas further away, their size decreases and the preparation of polyplexes con- taining only one RNA copy becomes possible. Our results are consistent with recent experimental work on poly-ethylenimine (PEI) and self-amplifying RNA (saRNA).

RNA-based pharmaceuticals proved successful with the COVID-19 vaccines and are now undergoing clinical trials for a broad range of therapeutic indications. Lipid-based nanoparticles (LNPs) have been used so far as delivery systems, although alternatives are still needed to meet efficacy and safety requirements across a broader range of applications. Polyplexes, formed by the self-assembly of cationic polymers with the anionic nucleic acids, constitute a valuable substitute, especially if precise control of the number and shape of the encapsulated RNA chains is possible. Here, we use molecular dynamics simulations of a coarse-grained polyplex model to show that the most important factors controlling it are the charge ratio between polyelectrolytes and RNA and their concentration during assembly. Close to the isoelectric point, the polyplexes are large, whereas in large excess of cationic polymer, their size decreases, allowing one RNA copy per nanoparticle. Our results are consistent with recent experimental work on polyethylenimine polyplexes.

RNA-based therapeutics have demonstrated remarkable efficacy and hold great promise for future applications in personalized medicine. The most common delivery systems for these drugs are lipid-based nanoparticles (LNPs), which incorporate ionizable cationic lipids (ICLs) as key components. Among other, ICLs are believed to facilitate endosomal escape of the cargo by interacting with anionic lipids in the endosomal membrane, although the underlying molecular mechanism remains unclear. One proposed model suggests that the membrane is destabilized by cone-shaped complexes formed between ICLs and endosomal anionic lipids. However, no clear evidence of stable co-localization of anionic and cationic lipids has been presented, so far. Here, we re-examine the problem through equilibrium and nonequilibrium simulations of model membrane systems containing DODMA (ICL), DOPS (anionic lipid) and DOPE (helper lipid). Our results confirm absence of co-localization at equilibrium, but reveal a transient formation of cone-shaped complexes during lamellar-to-inverted-hexagonal phase transitions, which considerably accelerates the transition process. These findings may open new ways for controlling endosomal escape through the rational design of ICLs optimized to interact with cell- or stage-specific endosomal anionic lipids.

Lipid based nanomaterials are being considered as promising nanocarriers for a variety of therapeutic approaches including for example drug delivery, siRNA gene silencing, mRNA vaccines (as those for COVID-19). When the active substances to be delivered are nucleic acids, which are negatively charged in water, the lipid formulation includes typically a ionizable cationic lipid which is believed to interact with the nucleic acid and help protect it from degradation. However only limited information is available about the internal arrangement of the mRNA inside the nanoparticle. Here we report the results of multiscale molecular dynamics simulations used to characterize the structure of a lipid-based nanomaterial composed of messenger RNA and lipid formulations containing ionizable cationic lipids. In particular, the role of pH in the formation of the mRNA-lipids complex has been investigated. The simulations have shown that the interaction between the ionizable cationic lipids and RNA is attractive only at low pH and effectively repulsive at high pH. In addition, the ionizable lipid, when uncharged, also shows a significantly increased flip-flop probability, which may have consequences in the process of fusion of the nanoparticle with the endosomal membrane.

Summary The cavity-creating p53 cancer mutation Y220C is an ideal paradigm for developing small-molecule drugs based on protein stabilization. Here, we have systematically analyzed the structural and stability effects of all oncogenic Tyr-to-Cys mutations (Y126C, Y163C, Y205C, Y220C, Y234C, and Y236C) in the p53 DNA-binding domain (DBD). They were all highly destabilizing, drastically lowering the melting temperature of the protein by 8–17 °C. In contrast, two non-cancerous mutations, Y103C and Y107C, had only a moderate effect on protein stability. Differential stabilization of the mutants upon treatment with the anticancer agent arsenic trioxide and stibogluconate revealed an interesting proximity effect. Crystallographic studies complemented by MD simulations showed that two of the mutations, Y234C and Y236C, create internal cavities of different size and shape, whereas the others induce unique surface lesions. The mutation-induced pockets in the Y126C and Y205C mutant were, however, relatively small compared with that of the already druggable Y220C mutant. Intriguingly, our structural studies suggest a pronounced plasticity of the mutation-induced pocket in the frequently occurring Y163C mutant, which may be exploited for the development of small-molecule stabilizers. We point out general principles for reactivating thermolabile cancer mutants and highlight special cases where mutant-specific drugs are needed for the pharmacological rescue of p53 function in tumors.

Lipid-based nanomaterials are used as a common delivery vehicle for RNA therapeutics. They typically include a formulation containing ionizable cationic lipids, cholesterol, phospholipids, and a small molar fraction of PEGylated lipids. The ionizable cationic lipids are considered a crucial element of the formulation for the way they mediate interactions with the anionic RNA as a function of pH. Here, we show, by means of molecular dynamics simulation of lipid formulations containing two different ionizable cationic lipids (DLinDMA and DLinDAP), that the direct interactions of those lipids with RNA, taken alone, may not be sufficient to determine the level of protection and packaging of mRNA. Our simulations help and highlight how the collective behavior of the lipids in the formulation, which determines the ability to envelop the RNA, and the level of hydration of the lipid-RNA interface may also play a significant role. This allows the drawing of a hypothesis about the experimentally observed differences in the transfection efficiency of the two ionizable cationic lipids.

Molecular dynamics (MD) simulations provide an increasingly important instrument to study protein-materials interaction phenomena, thanks to both the constant improvement of the available computational resources and the refinement of the modeling methods. Here, we summarize the results obtained along two different research directions within our project. First, we show how MD simulations of several cancer mutants of the DNA-binding domain of the tumor suppressor protein p53 allowed to establish the destabilizing effect of the mutations as well as the stabilizing effects of bound ligands. Second, we report on the development of a new reweighting technique for metadynamics simulations that speeds up convergence and may provide an advantage in the case of simulation of large systems.

Nucleic acid-based therapies have shown enormous effectiveness as vaccines against the recent COVID19 pandemics and hold great promises in the fight of a broad spectrum of diseases ranging from viral infections to cancer up to genetically transmitted pathologies. Due to their highly degradable polyanionic nature, nucleic acids need to be packed in sophisticate delivery vehicles which compact them up, protect them from early degradation and help delivery them to the right tissue/cells. Lipid-based nanoparticles (LNP) represent, at present, the main solution for nucleic acid delivery. They are made of a mixture of lipids whose key ingredient is an ionizable cationic lipid. Indeed, the interactions between the polyanionic nucleic acids and the ionizable cationic lipids, and their pH-dependent regulation in the life cycle of the nanoparticle, from production to cargo delivery, mostly determine the effectiveness of the therapeutic approach. Notwithstanding the large improvements in the delivery efficiency of LNPs in the last two decades, it is estimated that only a small fraction of the cargo is actually delivered, stimulating further research for the design of more effective LNP formulations. A rationally driven design would profit from the knowledge of the precise molecular structure of these materials, which is however still either missing or characterized by poor spatial resolution. Computational approaches have often been used as a molecular microscope either to enrich the available experimental data and provide a molecular-level picture of the LNPs or even simulate specific processes involving the formation and/or the molecular mechanisms of action of the LNP. Here, I review the recent literature in the field.

Lipid-based nanoparticles and lipoplexes containing ionizable lipids are among the most successful nanocarriers for mRNA-based therapies. The molecular structure of these assemblies is still not fully understood, as well as the role played by the ionizable lipids. SAXS experiments have shown that lipoplexes including the ionizable lipid 2-dioleyloxy-N,N-dimethyl-3-aminopropane (DODMA), under specific conditions, have a lamellar structure, where lipid bilayers are separated by mRNA-rich layers, with an overall spacing between 6.5 and 8.0 nm and a complex pH-dependence. Here, the structure and dynamics of these lipoplexes are investigated at varying pH and mRNA concentration using multiscale molecular dynamics simulations. It is observed that the interaction between DODMA and RNA is slightly attractive only at low pH levels, while it becomes effectively repulsive at high and intermediate pH. This results into a pH-dependent relocation of the RNA inside the multilayers, from the lipid head groups at low pH to a more uniform distribution inside the hydrophilic slabs of the multilayers at high pH. It is also observed that at high pH, DODMA lipids shift toward the hydrophobic part of the bilayer, consequently increasing their leaflet-flipping rate, a phenomenon which may ultimately affect the fusion process of the lipoplex with the endosomal membrane.

Hydrophilic polymers are being investigated as possible coating agents for therapeutic nanoparticles because of their capacity to reduce immune response and increase circulation life time. The mechanism of action of these coatings is not well understood although it is clear that they unspecifically reduce the amount of proteins adsorbing on the nanoparticle surface coming in contact with biological fluids. Here we have investigated, using state-of-the-art atomistic molecular dynamics simulations, the equilibrium and kinetic properties of the interactions forming between human serum albumin, the most abundant protein in the blood stream, and two different and promising polymers poly(ethylene glycol) and poly-sarcosine and we have compared the results with a polymer which is an isomer of poly-sarcosine but has a totally different behavior in terms of adsorption, poly-alanine, because of its well-known aggregation propensity. The results show how the two hydrophilic polymers have a very similar behavior in terms of the amount of polymers adsorbed on the protein surface, pattern of interactions and the kinetics of the adsorption process, with differences emerging due to the different flexibility of the two molecules. In contrast poly-alanine adsorbs significantly more strongly on the protein surface, with a slower kinetics, and a quantitatively, but not qualitatively, different interaction pattern with the surface amino acids with respect to the hydrophilic polymers.

Die folgende Dissertation befasst sich mit der Entwicklung und Anwendung dreier neur Ansatze fur die Erzeugung und Untersuchung von Energy Landscapes mit dem Ziel, die Proteinfaltung zu charakterisieren und besser zu verstehen. Das Paradigma der Energy Landscapes hat sich in den letzten zehn Jahren bei der Untersuchung des Faltungsverhaltens von Proteinen bewahrt. Dieses Paradigma fuhrt dazu, dass die Energy Landscape eines Proteins vereinfacht als Trichter dargestellt werden kann. Projektionen dieser Energy Landscape auf Ordnungsparameter, wie zum Beispiel die Projektion auf die freie Energie zur Untersuchung von Stabilitat, sollten eine vereinfachte Sicht der Zustande und Energiebarrieren wiedergeben, wobei oftmals versteckte Annahmen getroffen werden, die nicht erfullt sind. Basierend auf der Theorie komplexer Netzwerke wird als erstes ein neur Ansatz zur Untersuchung der Proteinzustande vorgestellt. Dabei sind die Konformationen, welche im Verlauf einer molecular dynamics (MD) Simulation angenommen werden, die Knoten und die zeitlichen Uebergange die Verbindungen. Dieses Netzwerk reprasentiert die vollstandige multi-dimensionale Energy Landscape ohne Projektion auf willkurlich gewahlte Ordnungsparameter. Die Anwendung von Netzwerken enthullt ein komplexes Zusammenspiel von Minima, Basins und Super-Basins, die in der Energy Landscape die Rolle von Attraktoren spielen. Als erstes wird die MD-Simulation eines strukturierten Peptides (Beta3s) untersucht. Dabei werden Basins mit kleiner Entropie/ kleiner Enthalpie, sowie Basins mit grosser Entropie/grosser Enthalpie gefunden. Hierbei ist interessant, dass Beta3s im gefalteten Zustand (im Unterschied zu einer kinetischen Trap) nicht nur enthalpisch, sondern auch entropisch stabilisiert wird. Free Energy Basins entsprechen Subgraphen (Communities) des gesamten Netzwerks, wobei die Knoten in einen Basin dicht miteinander verbunden sind. Das Finden dieser Communities und die Validierung einer Unterteilung des gesamten Netzwerkes in solche ist trotz der Fulle von existierenden Algorithmen noch nicht zufriedenstllend gelost. In dieser Dissertation wird dazu ein Mass, die Goodness Deviation, vorgeschlagen. Zweitens ist die Kinetik der Proteinfaltung ebenfalls eng mit der Topographie von Energy Landscapes verknupft. Die schwer fassbare Natur des Transition State Ensemble (TSE) erlaubt kein einfaches Bild fur Faltunsubergange. Zur Charakterisierung des TSE wird oft die Faltungswahrscheinlichkeit pfold einer Proteinstruktur benutzt. Obwohl g fur die Analyse von Faltungsubergangen sehr nutzlich ist, erfordert dessen Berechnung einen grossen rechnerischen Aufwand, was die praktische Anwendungen limitiert. Die kinetische Homogenitat von strukturell ahnlichen Snapshots erlaubt jedoch einen statistischen Zugang fur die Analyse einzelner Konformationen. Die Idee des sogenannten Cluster-pfold ermoglicht es in dieser Dissertation, g fur Strukturen einer MD-simulation approximativ zu berechnen. Die Anwendung dieser Methode auf Beta3s zeigt ein breites, heterogenes TSE, sowie zwei Haupt-Pathways. Diese Ergebnisse decken sich interessanterweise mit der Netzwerk-Analyse aus vorhergehenden Untersuchungen. Der marginale Rechenaufwand fur den Cluster-pfold ermoglichte die Analyse des TSE von vielen Beta3s-Mutanten und die Berechnung von Phi-Werten. Mit herkommlichen Methoden ware dies fur unsere riesige Menge Simulationsdaten (0.65ms) unmoglich gewesen. Drittens sind in dieser Dissertation alle Simulationen mit konstanter Temperatur bei einer Temperatur T gemacht worden, welche hoher ist als die physiologische Temperatur. Es sei erwahnt, dass die Hohe der Energie-Barrieren proportional zu exp((DeltaE/kT) ansteigt, wobei k die Boltzmann-Konstante ist. Daher bleiben MDSimulationen bei physiologischer Temperatur leicht in lokalen Minima stecken und ergeben somit kein korrektes Sampling der Energy Landscape innert nutzlicher Rechenzeit. In dieser Dissertation wird die Replica-exchange Methode (REM) benutzt, um die fruhen Stadium der Peptid-Aggregation eines amyloidogenetischen Peptides des Prionen-Proteins Sup35 befasst. abcxyz Summary The present thesis is concerned with the development and application of three novel approaches for the generation and the analysis of energy landscapes characterizing protein folding. In the past decade, the energy landscape paradigm has emerged as a consistent and useful framework for the study of many aspects of protein folding leading to the so-called funnel picture for folding. Projections of the energy landscape to order parameters should give a simple and synthetic view of the states and energetic barriers of the system. Free-energy projections are used to estimate the stability and in some cases also the kinetics of the states involved in the folding process. However, this simplistic picture revealed some hidden assumptions on order parameter(s) that in many cases are not fulfilled. First, a novel approach for the study of the states of a protein, based on complex networks, is introduced. Conformations visited during a molecular dynamics (MD) simulation and the transitions between them are the nodes and the links of the network, respectively. The network represents the multi-dimensional free-energy landscape and does not require any projections to arbitrarily chosen order parameters. The network approach revealed a complex scenario of minima, basins and superbasins of attraction in the free-energy landscape. As a first application, a structure peptide, namely beta3s, was investigated by MD simulations and network analysis. Both low-enthalpy/low-entropy and high-enthalpy/high-entropy basins were found in the denatured state. Interestingly, the native state of beta3s is stabilized not only enthalpically but also entropically, the latter with respect to a kinetic trap. In the network framework, free-energy basins are defined as subgraphs (called communities or clusters) of the network where nodes are highly connected. Although many communities detection algorithms exist in literature, it is not yet clear how to judge the quality of a clusterization. A new measure to assess this problem, called goodness deviation, is proposed in the thesis. Second, kinetics of protein folding are strongly connected to the topography and the organization of the energy landscape. The elusive nature of the transition state ensemble (TSE) doesn’t allow an easy and clear picture of the folding energy barrier. A common technique, widely used in literature, to characterize the TSE is the computation of the folding probability pfold which is the probability of a structure to fold before unfolding. Even if pfold is a useful quantity for transition state analysis, it requires a very large computational effort which in practice limits its application to the analysis of small sets of data. The kinetic homogeneity of structurally similar snapshots suggests the use of a statistical approach for the analysis of the kinetic properties of a peptide conformation. In this thesis, a novel approach for the computation of the folding probability, called cluster-pfold, allows the approximate estimation of the pfold for every structure sampled along an MD simulation. The application of the method on the beta3s peptide has shown the presence of a broad and heterogeneous TSE as well as two average folding pathways. Interestingly, these findings are in agreement with the network analysis performed on the peptide in our former studies. The negligible computational demand for the calculation of the cluster-pfold allowed the analysis of the TSE for a large set of beta3s mutants and the calculation of the Phi-values. Given the huge amount of simulation time (0.65 ms), this analysis would have been impossible with traditional methods. Third, all constant temperature simulations presented in this thesis were run at temperatures higher than the physiological temperature. It is worth noting that the height of energy barriers is proportional to exp(-DeltaE/kT), where k and T are the Boltzmann constant and temperature, respectively. For this reason, MD simulations ran at physiological temperatures by conventional MD. One disadvantage of the method is that it cannot be used for the analysis of the kinetics. REM simulations were also used to study the early stages of peptide aggregation of an amyloidogenic peptide from the yeast prion protein sup35. abcxyz Posted at the Zurich Open Repository and Archive, University of Zurich ZORA URL: https://doi.org/10.5167/uzh-163326 Dissertation Published Version Originally published at: Rao, Francesco. Theoretical and computational studies of protein folding energy landscapes. 2005, University of Zurich, Faculty of Science.

The data collected along a metadynamics simulation can be used to recover information about the underlying unbiased system by means of a reweighting procedure. Here, we analyze the behavior of several reweighting techniques in terms of the quality of the reconstruction of the underlying unbiased free energy landscape in the early stages of the simulation and propose a simple reweighting scheme that we relate to the other techniques. We then show that the free energy landscape reconstructed from reweighted data can be more accurate than the negative bias potential depending on the reweighting technique, the stage of the simulation, and the adoption of well-tempered or standard metadynamics. While none of the tested reweighting techniques from the literature provides the most accurate results in all the analyzed situations, the one proposed here, in addition to helping simplifying the reweighting procedure, converges quickly and precisely to the underlying free energy surface in all the considered cases, thus allowing for an efficient use of limited simulation data.

We have previously shown that the thermolabile, cavity-creating p53 cancer mutant Y220C can be reactivated by small-molecule stabilizers. In our ongoing efforts to unearth druggable variants of the p53 mutome, we have now analyzed the effects of other cancer-associated mutations at codon 220 on the structure, stability, and dynamics of the p53 DNA-binding domain (DBD). We found that the oncogenic Y220H, Y220N, and Y220S mutations are also highly destabilizing, suggesting that they are largely unfolded under physiological conditions. A high-resolution crystal structure of the Y220S mutant DBD revealed a mutation-induced surface crevice similar to that of Y220C, whereas the corresponding pocket’s accessibility to small molecules was blocked in the structure of the Y220H mutant. Accordingly, a series of carbazole-based small molecules, designed for stabilizing the Y220C mutant, also bound to and stabilized the folded state of the Y220S mutant, albeit with varying affinities due to structural differences in the binding pocket of the two mutants. Some of the compounds also bound to and stabilized the Y220N mutant, but not the Y220H mutant. Our data validate the Y220S and Y220N mutants as druggable targets and provide a framework for the design of Y220S or Y220N-specific compounds as well as compounds with dual Y220C/Y220S specificity for use in personalized cancer therapy.

Nanoparticles coated with hydrophilic polymers often show a reduction in unspecific interactions with the biological environment, which improves their biocompatibility. The molecular determinants of this reduction are not very well understood yet, and their knowledge may help improving nanoparticle design. Here we address, using molecular dynamics simulations, the interactions of human serum albumin, the most abundant serum protein, with two promising hydrophilic polymers used for the coating of therapeutic nanoparticles, poly(ethylene-glycol) and poly-sarcosine. By simulating the protein immersed in a polymer-water mixture, we show that the two polymers have a very similar affinity for the protein surface, both in terms of the amount of polymer adsorbed and also in terms of the type of amino acids mainly involved in the interactions. We further analyze the kinetics of adsorption and how it affects the polymer conformations. Minor differences between the polymers are observed in the thickness of the adsorption layer, that are related to the different degree of flexibility of the two molecules. In comparison poly-alanine, an isomer of poly-sarcosine known to self-aggregate and induce protein aggregation, shows a significantly larger affinity for the protein surface than PEG and PSar, which we show to be related not to a different patterns of interactions with the protein surface, but to the different way the polymer interacts with water.

In the development of new therapeutic agents based on nanoparticles it is of fundamental importance understanding how these substances interact with the underlying biological milieu. Our research is focussed on simulating in silico these interactions using accurate atomistic models, and gather from these information general pictures and simplified models of the underlying phenomena. Here we report results about the interactions of blood proteins with promising hydrophilic polymers used for the coating of therapeutic nanoparticles, about the salt dependent behavior of one of these polymers (poly-(ethylene glycol)) and about the interactions of blood proteins with silica, one of the most used materials for the production of nanoparticles.

We use plasmon rulers made from two connected gold nanoparticles to monitor the conformation and stiffness of single PEG molecules and their response to cations. By observing equilibrium fluctuations of the interparticle distance, we obtain the spring constants or stiffness of the connecting single-molecule tether with pico-Newton sensitivity. We observe a transition of the PEG molecules' extension and stiffness above about 1.2 mM K+ ion concentration which is specific to potassium ions. Molecular dynamics simulations reveal the formation of crown-like structures as the most likely molecular mechanism responsible for this specific effect.