Biology subjects

Vendruscolo, M.

Publications and source records attributed to Vendruscolo, M..

At least 37 records · Page 2Linked to original sources

bioRxiv2025

Destabilization of Helix III Initiates Early Serum Amyloid A Misfolding by Exposing Its Amyloidogenic Core

Serum amyloid A (SAA) is the principal precursor of AA amyloidosis, yet the early molecular steps triggering its pathological misfolding remain unclear. Here, we combine harmonic linear discriminant analysis (HLDA) and parallel-tempering metadynamics (PT-MetaD) to dissect the earliest conformational transitions of the disease-relevant SAA1-76 fragment. By constructing an optimized one-dimensional collective variable (sHLDA) from inter-helix contacts and helical root-mean-square deviations, we perform 4 {micro}s of enhanced sampling across 79 replicas (300-450K). Free-energy surfaces reveal a misfolding trajec-tory where helix III destabilizes first, preceding loss of helices II and I while global com-pactness persists. Solvent-accessible surface-area analysis reveals transient exposure of the aggregation-prone core (residues 42-48) within specific intermediates, implicating localized core exposure rather than wholesale unfolding as the trigger for misfolding. Temperature-dependent secondary-structure profiling confirms SAA1-76 behaves as a folded bundle with disordered loops. These findings highlight helix III stabilization and amyloidogenic segment masking as potential therapeutic strategies. TOC Graphic O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=107 SRC="FIGDIR/small/656333v2_ufig1.gif" ALT="Figure 1"> View larger version (29K): org.highwire.dtl.DTLVardef@167fd1borg.highwire.dtl.DTLVardef@19fa62org.highwire.dtl.DTLVardef@1857066org.highwire.dtl.DTLVardef@1966fb0_HPS_FORMAT_FIGEXP M_FIG C_FIG

biochemistry↗

bioRxiv2025

Prediction of Antibody Non-Specificity using Protein Language Models and Biophysical Parameters

The development of therapeutic antibodies requires optimizing target binding affinity and pharmacodynamics, while ensuring high developability potential, including minimizing non-specific binding. In this study, we address this problem by predicting antibody non-specificity by two complementary approaches: (i) antibody sequence embeddings by protein language models (PLMs), and (ii) a comprehensive set of sequence-based biophysical descriptors. These models were trained on human and mouse antibody data from Boughter et al. (2020) and tested on three public datasets: Jain et al. (2017), Shehata et al. (2019) and Harvey et al. (2022). We show that non-specificity is best predicted from the heavy variable domain and heavy-chain complementary variable regions (CDRs). The top performing PLM, a heavy variable domain-based ESM 1v LogisticReg model, resulted in 10-fold cross-validation accuracy of up to 71%. Our biophysical descriptor-based analysis identified the isoelectric point as a key driver of non-specificity. Our findings underscore the importance of biophysical properties in predicting antibody non-specificity and highlight the potential of protein language models for the development of antibody-based therapeutics. To illustrate the use of our approach in the development of lead candidates with high developability potential, we show that it can be extended to therapeutic antibodies and nanobodies.

bioinformatics↗

bioRxiv2025

Linking Aβ and tau in the amyloid cascade through the intersection of their proteostasis networks

Amyloid plaques and neurofibrillary tangles are molecular hallmarks of Alzheimers disease. According to the amyloid cascade hypothesis, aberrant A{beta} and tau behaviours contribute synergistically to accelerate the Alzheimers pathology. However, the complex molecular mechanisms linking A{beta} and tau dysregulation remain to be fully characterised. To address this problem, we investigated the connection between A{beta} and tau through the protein homeostasis (proteostasis) network. We asked whether A{beta} proteostasis is linked to tau proteostasis. To answer to this question, we first mapped the proteostasis networks of A{beta} and tau, and then studied the interplay of these two networks, identifying the molecular chaperone HSP90 as a central hub. To test this hub role of HSP90, we observed in a cell model that HSP90 and its co-chaperone SUGT1 mediate tau phosphorylation via GSK-3{beta} in an A{beta}42-dependent manner. Furthermore, we also observed that in turn HSP90 and SUGT1 increase the intracellular concentration of A{beta}42. These results suggest that the HSP90/SUGT1 system may act as a hub in the amyloid cascade by lying at the intersection of the A{beta} and tau proteostasis networks.

neuroscience↗

bioRxiv2025

Serial amplification of tau filaments using Alzheimer's brain homogenates and C322A or C322S recombinant tau

The assembly of tau into amyloid filaments is a defining characteristic of Alzheimers disease (AD) and other tauopathies. Cryo-electron microscopy (cryo-EM) showed that specific tau folds characterise different diseases, and that in vitro models often yield filaments with folds that do not replicate those that form in disease. Here, we investigated the aggregation of full-length recombinant 0N3R tau, using wild-type, or mutations C322A or C322S and a real-time quaking-induced conversion (RT-QuIC) assay with brain homogenate seeding. The assembly of C322A 0N3R tau resulted in filaments with a structure resembling the AD fold in paired helical filaments (PHFs), but with a more open C-shaped core attributed to the C322A mutation. C322S 0N3R tau formed structurally more distinct filaments with respect to PHFs, with an ordered carboxy-terminal region. Both mutant filaments retained the ability to seed a second round of aggregation. Meanwhile, wild-type 0N3R tau exhibited poor reproducibility and formed predominantly unfolded aggregates. Our findings emphasise the need for optimised assembly conditions to obtain disease-relevant filament folds. Refining these methodologies could enhance our understanding of the molecular origins of tauopathies and facilitate the development of targeted therapeutic strategies for these conditions.

neuroscience↗

bioRxiv2025

Structural defects in amyloid-β fibrils drive secondary nucleation

The nucleation of amyloid fibrils from monomeric protein, catalyzed by the surface of existing fibrils, is an important driver of many disorders such as Alzheimers and Parkinsons diseases. The structural basis of this secondary nucleation process, however, is poorly understood. Here, we ask whether secondary nucleation sites are found predominantly at rare growth defects: defects in the fibril core structure generated during their original assembly. We first demonstrate using the specific inhibitor of secondary nucleation, Brichos, that secondary nucleation sites on Alzheimers disease-associated fibrils composed of A{beta}40 and A{beta}42 peptides are rare compared to the number of protein molecules they contain. We then grow A{beta}40 fibrils under conditions designed to eliminate most growth defects while leaving the regular fibril morphology unchanged, and confirm the latter using cryo-electron microscopy. We measure both the ability of these annealed fibrils to promote secondary nucleation and the stoichiometry of their secondary nucleation sites, finding that both are greatly reduced as predicted. Re-analysis of published data for other proteins suggests that fibril growth defects that expose monomer planes or other structural units may also drive secondary nucleation generally, across most or all amyloids. These findings could unlock structure-based drug design of therapeutics that aim to halt amyloid disorders by inhibiting secondary nucleation sites.

biophysics↗

bioRxiv2025

The alpha-Synuclein Proteostasis Network and its Translational Applications in Parkinson's disease

Parkinsons disease (PD) is debilitating neurodegenerative condition that results in the loss of mobility and muscle control. A neuropathological hallmark of PD is the presence aberrant inclusions, known as Lewy pathology, of which -synuclein (-Syn) is a major component. The accumulation of -Syn is a likely consequence of an age-related impairment of the proteostasis system regulating -Syn. To investigate this phenomenon, we map the proteostasis network (PN) of -Syn in the Substania nigra at the proteomic and transcriptomic levels. We then define a -Syn proteostasis activity score (PAS) that quantifies the activity of the PN in regulating -Syn. We thus obtain a PAS signature indicative of the disease state, as well as the age-of-death in PD patients, and the brain regional vulnerability to -Syn aggregation. We then outline a digital twin of the -Syn PN in the Substantia nigra cells by training a model on single-cell data. This digital twin is applied towards target identification for PD. In addition, we further describe the application of the PN to facilitate drug repurposing. Overall, our study highlights the implication of the -Syn PN in PD and how simulations and measurements of its activity can help efforts in translational research for PD.

neuroscience↗

bioRxiv2025

Two sites in the C-terminal β-chain tail mediate interactions of the chaperone clusterin with amyloid beta and other misfolded client proteins

Clusterin (CLU) is a constitutively secreted mammalian chaperone that binds in extracellular body fluids to misfolded client proteins to neutralise their toxicity and mediate their safe disposal by cell uptake and intracellular degradation. However, the regions of CLU critical for its interactions with misfolded proteins remain still largely unknown. To identify binding sites, we expressed a panel of CLU deletion and alanine-stretch mutants in a recently developed mammalian expression system. Mutant CLU molecules lacking detectable structural aberrations were subjected to functional analyses to compare their abilities with that of wild type CLU to bind to misfolded proteins and to inhibit protein aggregation. These analyses implicated two regions in the flexible {beta}-chain C-terminal tail of CLU as being important in the interactions of the chaperone with misfolded proteins, including aggregating the Alzheimers amyloid {beta}-peptide (A{beta}). We then designed in silico sequence-specific single-domain camelid nanobodies to confirm the function of the two putative client protein binding sites. Based on our experimental results and in silico binding site predictions, we suggest that the potent ability of CLU to promiscuously interact with many different misfolded proteins, regardless of their size or structure, arises from the location of multiple client protein binding sites in its flexible tail region.

biochemistry↗

bioRxiv2025

Design of Tau Aggregation Inhibitors Using Iterative Machine Learning and a Polymorph-Specific Brain-Seeded Fibril Amplification Assay

The aggregation of tau into amyloid fibrils is associated with Alzheimers disease (AD) and related tauopathies. Since different tauopathies are characterised by the formation of distinct tau fibril morphologies, it is important to combine the search of tau aggregation inhibitors with the development of in vitro tau aggregation assays that recapitulate aggregation as it may occur in the brain. Here we address this problem by reporting an in vitro tau aggregation assay in which AD brain homogenates are used to seed the generation of first-generation tau fibrils in a polymorph-specific manner under quiescent conditions. These fibrils are then used to create amyloid seed libraries from which second-generation kinetic assays can be readily performed. Using this strategy, we illustrate an iterative machine learning method for the identification of small molecules for the polymorph-specific inhibition of the in vitro formation of tau fibrils. We further show that the small molecules selected by this procedure are potent inhibitors in a Drosophila tauopathy model.

biophysics↗

bioRxiv2025

Design of Protein Sequences with Precisely Tuned Kinetic Properties

Recent advances in computational biology have enabled solutions to the inverse folding problem - finding an amino acid sequence that folds into a target structure. An open question concerns the design of proteins that in addition to having the correct target structure also have precisely tuned kinetic properties, such as folding and unfolding rates. To address this problem, we formulate the inverse folding problem as a quest for a sequence with a target free energy landscape. To propose a procedure to address this problem, here we describe the Inverse Folding Molecular Dynamics (IF-MD) method, which combines inverse folding with enhanced sampling molecular dynamics and Bayesian optimization. IF-MD leverages ensemble averages from molecular dynamics simulations, reweighted according to a Bayesian framework, to guide the design of sequences exhibiting specific kinetic properties. We demonstrate the methodology by optisising the binding kinetics of H11, a nanobody against the SARS-CoV-2 spike receptor-binding domain (RBD), thus identifying nanobody variants with slower unbinding kinetics than H11. Mechanistic analysis reveals that this kinetic property arises from a shift towards configurations closer to the bound state and increased free energy barriers for dissociation. These findings highlight the power of IF-MD for efficiently navigating the vast sequence space to design proteins with a tailored free energy landscape.

bioinformatics↗

bioRxiv2024

Proteostasis Signatures in Human Diseases

The protein homeostasis (proteostasis) system maintains the proteome in a healthy state. Although this system has been comprehensively mapped, its perturbations in disease remain incompletely characterised. To address this problem, here we define the proteostasis signatures, which represent the characteristic patterns of change in the proteostasis system associated with disease. We performed a large-scale, pandisease analysis across 32 human diseases spanning 7 disease types. We first identified unique proteostasis perturbations in specific disease states. We then uncovered distinctive signatures differentiating disease types, pointing to a range of proteostasis mechanisms in disease development. Next, we tracked the temporal evolution of proteostasis signatures, revealing shifts in proteostasis disruption over the course of disease progression. Finally, we demonstrated how smoking, a major risk factor of many diseases, impairs proteostasis in a manner similar to disease, potentially creating a predisposed environment for disease onset. These results illustrate the opportunities offered by the study of human diseases from the perspective of proteostasis signatures.

bioinformatics↗

bioRxiv2024

Immuno-moodulin is Differentially Expressed in T Cells and Plasma in Obsessive-Compulsive Disorder Patients

Immuno-moodulin (Imood), a recently discovered protein expressed in T cells, is associated with anxiety-like behavior in mice. However, its mechanism of action in modulating neuroimmune interactions remains unclear. To investigate this problem, we characterized Imood in human blood and immune cells using neutralizing monoclonal antibodies, revealing its nature as an intrinsically disordered protein (IDP) with unique expression patterns. Our findings indicate that Imood is predominantly expressed intracellularly in peripheral blood mononuclear cells (PBMCs), particularly T lymphocytes, but is absent in polymorphonuclear cells. Upon T-cell activation, Imood exhibits distinct mobilization patterns with increased surface expression. Bioinformatics analysis identified a strong propensity for oligomerization and liquid-liquid phase separation. We also found that T cells from patients with Obsessive-Compulsive-Disorder (OCD) displayed significantly elevated surface Imood expression compared to healthy controls, as well as an altered level of Imood polymerization in the plasma. Taken together, these results elucidate the expression patterns and structural properties of Imood in human immune cells, which open new avenues for OCD diagnostics, and prompt further study for understanding the aetiology of OCD and related disorders.

immunology↗

bioRxiv2024

Language Models for Molecular Dynamics

Molecular Dynamics (MD) simulations provide accurate descriptions of the motions of molecular systems, yet their computational demands pose significant challenges in applications in molecular biology and materials science. Given the success of deep learning methods in a wide range of fields, a timely question concerns whether these methods could be leveraged to improve the efficiency of MD simulations. To investigate this possibility, we introduce Molecular Dynamics Language Models (MDLMs), to enable the generation of MD trajectories. In the present implementation, an MDLM is trained on a short classical MD trajectory of a protein, where structural accuracy is maintained through kernel density estimations derived from extensive MD datasets. We illustrate the application of this MDLM in the case of the determination of the free energy landscape a small protein, showing that this approach makes it possible to discover conformational states undersampled in the training data. These results provide initial evidence for the use of language models for the efficient implementation of molecular dynamics.

bioinformatics↗

bioRxiv2024

Organization of the Proteostasis Network of Membraneless Organelles

Membraneless organelles (MLOs) are dynamic macromolecular condensates that act as crucibles to modulate biological process within the cell. Since MLOs form in the absence of lipid membranes, it is important to understand how their effective regulation is achieved by the protein homeostasis (proteostasis) system. To address this question, we report a comprehensive mapping of the proteostasis network (PN) of MLOs, comprising over 220,000 protein-protein interactions. The analysis of this PN reveals how the regulatory proteins (PN proteins) occupy central roles in the overall protein-protein interaction network of MLOs. We then investigate which branches of the PN are most important in the regulation of MLOs, and find that the anabolic component, which makes up [~]30% of the PN, is more closely involved with MLOs than the catabolic component, which makes up the remaining [~]70% of the PN. We also find that translation-related PN proteins and molecular chaperones play central roles in the regulation of MLOs. We then analyse the similarity between the PNs of different MLOs, revealing that stress granules and P-bodies share large portions of their PNs. Finally, to understand how specificity is achieved despite these similarities, we further analyse the PN organization, showing that HSP70 chaperones are generic in regulating MLOs, while client-specific HSP70 co-chaperones confer specificity to the chaperone action. Taken together, these results identify the composition of the PN of MLOs, rationalize its organization, and reveal a central role of molecular chaperones in the regulation of protein conformations within MLOs.

bioinformatics↗

bioRxiv2024

AlphaFold-Metainference: Prediction of Structural Ensembles of Disordered Proteins

Deep learning methods of predicting protein structures have reached an accuracy comparable to that of high-resolution experimental methods. It is thus possible to generate accurate models of the native states of hundreds of millions of proteins. An open question, however, concerns whether these advances can be translated to disordered proteins, which should be represented as structural ensembles because of their heterogeneous and dynamical nature. To address this problem, we introduce the AlphaFold-Metainference method to use AlphaFold-derived distances as structural restraints in molecular dynamics simulations to construct structural ensembles of ordered and disordered proteins. The results obtained using AlphaFold-Metainference illustrate the possibility of making predictions of the conformational properties of disordered proteins using deep learning methods trained on the large structural databases available for folded proteins.

biophysics↗

bioRxiv2024

Global kinetic model of lipid-induced alpha-synuclein aggregation and its inhibition by small molecules

The aggregation of -synuclein into amyloid fibrils is a hallmark of Parkinsons disease. This process has been shown to directly involve interactions between proteins and lipid surfaces when the latter are present. Despite this importance, the molecular mechanisms of lipid-induced amyloid aggregation have remained largely elusive. Here, we present a global kinetic model to describe lipid-induced amyloid aggregation of -synuclein. Using this framework we find that -synuclein fibrils form via a two-step primary nucleation mechanism and that lipid molecules are directly involved in both the nucleation and fibril elongation steps, giving rise to lipid-protein coaggregates. To illustrate the applicability of this kinetic approach to drug discovery, we identify the mechanism of action of squalamine, a known inhibitor of -synuclein aggregation, finding that this small molecule reduces the rate of lipid-dependent primary nucleation. Our work will likely guide the rational design of -synuclein aggregation inhibitors. Significance StatementAmyloid aggregation is a hallmark of a diverse range of diseases including Parkinsons Disease, where the protein -synuclein is a major con-stituent of proteinaceous deposits found in patients. It is well established that interactions between -synuclein and lipids modulate aggregation. However, the molecular mechanisms driving this lipid-induced aggregation have remained largely elusive, which has frustrated so far the discovery of drugs that prevent lipid-induced aggregation. In this work, we present the first global kinetic model describing lipid-induced aggregation and by integrating this theoretical framework with in vitro experimental data of lipid-induced -synuclein aggregation, we reveal the role of lipid membranes in the aggregation process and uncover the mechanism by which small molecule inhibitors interfere with this process.

biophysics↗

bioRxiv2024

A human neuronal model of sporadic Alzheimer's disease induced by FBXO2 downregulation shows A β aggregation, tau hyperphosphorylation and functional network impairment

Sporadic Alzheimers disease (sAD) arises from a complex interplay between genetic and environmental factors that remains poorly understood, making it challenging to develop accurate cell models. To address this problem, by hypothesing that the early disease sAD states can be characterised by transcriptomic fingerprints, we assessed the effect on A{beta} aggregation in human neuroblastoma cells a set of genes obtained by analysing snRNA-seq data from post-mortem AD patients. We then validated the most effective genes in human iPSC-derived cortical neurons, and selected FBXO2, a gene encoding a subunit of the ubiquitin protein ligase complex SCF, for further analysis. We found that early downregulation of FBXO2 in human iPSC-derived cortical neurons resulted in A{beta} aggregation, tau hyperphosphorylation, and structural and functional neuronal network impairment. Based on these results, we report a neuronal sAD model (FBXO2 KD sAD) that recapitulates a set of molecular hallmarks of sAD. We suggest that this strategy can be expanded towards the generation of panels of preclinical stem cell-derived models that recapitulate the molecular complexity of the broad spectrum of AD patients.

neuroscience↗

bioRxiv2024

In situ generation of Aβ42 oligomers via secondary nucleation triggers neurite degeneration and synaptic dysfunction in human iPSC-derived glutamatergic neurons

The aggregation of A{beta}42 into misfolded oligomers is a central event in the pathogenesis of Alzheimers disease. In this study, we aimed to develop a robust experimental system that recapitulates A{beta}42 oligomerization in living cells to gain insight into their neurotoxicity and to provide a platform to characterize the effects of inhibitors of this process. Our strategy is based on the in situ generation of A{beta}42 oligomers via secondary nucleation by repeatedly treating the cells with A{beta}42 monomers in the presence of pre-formed A{beta}42 fibrils. This approach enables an accurate control over the levels of on-pathway soluble A{beta}42 oligomers and cell-associated aggregates, as well as the study of their neurotoxic effects. By implementing this approach in human glutamatergic neurons derived from induced pluripotent stem cells (iPSCs), we were able to replicate key aspects of Alzheimers disease, including neurite degeneration and synaptic dysfunction. Using BRICHOS, a molecular chaperone that specifically inhibits secondary nucleation, we confirmed that aggregation in this system occurs through secondary nucleation, and that quantitative parameters for comparing potential A{beta}42 aggregation inhibitors can be obtained. Overall, our results demonstrate that by in situ generation of on-pathway A{beta}42 oligomers, one can obtain translational cellular models of AD to bridge the gap between basic research and clinical applications.

cell biology↗

bioRxiv2024

TNFα-driven Aβ aggregation, synaptic dysfunction and hypermetabolism in human iPSC-derived cortical neurons

Alzheimers disease (AD) patients exhibit an increased load of A{beta} aggregates in the brain parenchyma. The neurotoxic nature of these aggregates has been underscored by recent advances in therapies aimed at reducing their load. To make further progress towards the development of increasingly effective treatments, there is a still largely unmet need for reliable cell models that comprehensively recapitulate aggregate-driven AD pathology. Here, we report a robust and scalable pipeline for generating human iPSC-derived cortical neurons that display A{beta} aggregates in their axonal projections. This phenotype is caused by a repeated dosage of tumour necrosis factor-alpha (TNF) to simulate the chronic inflammatory environment characteristic of AD and enhanced in neurons carrying the Swedish mutation. In association with the increased A{beta} deposits in the cell bodies, this cell model exhibits other key hallmarks of AD, including structural alterations of synapses, electrophysiological asynchronous hyperactivity, and hypermetabolism. Overall, these results illustrate how repeated TNF treatment models central aspects of AD pathology, and provides a platform that could be used for facilitating the translation of potential drugs to clinical applications.

cell biology↗