Intrinsically disordered proteins (IDPs) are abundant in cells and have central roles in protein-protein interaction networks. Many are involved in cancer, aging, and neurodegenerative diseases. IDPs do not possess a single native structure but an ensemble of possible conformations related to interactions with binding partners. Due to this inherent flexibility, conventional methods for determining the structure of globular proteins may not be directly applicable to IDPs. A primary aim of this work was to use Enhanced Sampling Molecular Dynamics simulations in conjunction with Neural Network algorithms and FRET biophysical techniques to achieve a deeper understanding of the structure and dynamics of a specific IDP, namely Heat Shock Protein B8 (HSPB8), a human chaperone involved in preventing the aggregation of misfolded proteins, whose mutation leads to neurodegenerative diseases. HSPB8 consists of a conserved structure called the α-crystallin domain and two intrinsically disordered regions (IDRs). At present, no experimental 3D structures are available for wt HSPB8 and its K141E variant. To establish suitable force field parameters and a computational workflow for simulating IDPs, extensive Temperature Replica Exchange (TREMD) simulations of HSPB8 variants were analyzed with a Neural Network algorithm called EncoderMap, and the results were compared to available experimental data. The findings allowed us to determine key factors affecting the structure of the HSPB8 K141E variant. Indeed, its structures tend to be more compact, and the hydrophobic area exposed to the solvent is reduced. These features likely play a role in the K141E reduced chaperone activity. Because IDPs are highly sensitive to environmental conditions, the effects of high ionic strength on the dynamics of HSPB8 IDRs were also investigated. The results show that ionic strength might have differential effects on the conformational propensities of distinct regions of HSPB8. This information will help to understand the mutant's behavior and determine suitable conditions for accurately studying it. The effect of small-molecule binding to IDPs was investigated by simulating HSPB8 variants in the presence of paroxetine, a small molecule commonly found in antidepressant drugs that was shown to have a high affinity for HSPB8 and is able to partially restore the chaperone activity in the mutated K141E variant. Our results show that the paroxetine effect is stronger on the K141E variant and affects the compactness of its conformations. To directly compare computational results with smFRET experiments, we developed FRETpredict, a Python program to calculate FRET efficiency from protein structures and trajectories based on the Rotamer Library Approach. This software is freely available on GitHub (KULL-Centre/FRETpredict/) and as a Python PyPI package. Moreover, Molecular Dynamics simulations were employed to design a novel FRET sensor for the detection of Sars-Cov-2 spike protein. The results obtained in this Ph.D. thesis provide the first 3D structural characterization of HSPB8 and reveal the effects of the pathogenic K141E mutation on its conformational ensembles. These results offer the possibility of rationalizing the pathogenic effects of the K141E mutation in terms of conformational changes. This work has helped to improve the understanding of how HSPB8 structure and dynamics are related to environmental changes such as ionic strength and the presence of small molecules. The FRETpredict software, efficiently operating on large protein conformational ensembles, facilitates the validation or refinement of molecular models and the interpretation of experimental data.

Le proteine intrinsecamente disordinate (IDP) sono abbondanti nelle cellule e hanno ruoli centrali nelle reti di interazione proteiche. Molte sono coinvolte nel cancro, nell'invecchiamento e nelle malattie neurodegenerative. Le IDP non possiedono una singola struttura nativa, ma un insieme di possibili conformazioni determinate dai binding partners. A causa di questa flessibilità, i metodi convenzionali per la determinazione della struttura non sono generalmente applicabili alle IDP. L'obiettivo principale di questo lavoro è stato quello di utilizzare simulazioni di dinamica molecolare Enhanced Sampling in combinazione con algoritmi di reti neurali e tecniche FRET per ottenere una comprensione più profonda della struttura e della dinamica di una specifica IDP, ovvero la Heat Shock Protein B8 (HSPB8), una chaperona umana che previene l'aggregazione di proteine misfolded e la cui mutazione porta a malattie neurodegenerative. La HSPB8 è costituita da un dominio α-cristallino e da due regioni intrinsecamente disordinate (IDR). Al momento non sono disponibili strutture 3D sperimentali per HSPB8 wt e la sua variante K141E. Per stabilire i parametri del force field e un workflow computazionale adatti per la simulazione delle IDP, estese simulazioni di Temperature Replica Exchange Molecular Dynamics (TREMD) delle varianti di HSPB8 sono state analizzate con un algoritmo di Neural Network chiamato EncoderMap e i risultati sono stati confrontati con i dati sperimentali disponibili. Ci è stato possibile determinare i fattori chiave che influenzano la struttura della variante HSPB8 K141E e le caratteristiche che possono portare ad una ridotta attività chaperonica. Le strutture della K141E tendono infatti a essere più compatte e la superficie idrofobica esposta al solvente ridotta. Poiché le IDP sono altamente sensibili alle condizioni cellulari, sono stati studiati gli effetti di un'elevata forza ionica sulla dinamica delle IDR di HSPB8. I risultati mostrano effetti sulle propensioni conformazionali dipendenti dall’IDR, informazione che aiuterà a determinare le condizioni adatte per uno studio accurato di HSPB8. L'effetto del legame di molecole alle IDP è stato studiato simulando HSPB8 in presenza di paroxetina, una molecula presente in farmaci antidepressivi che ha dimostrato di avere un'elevata affinità per HSPB8 ed è in grado di ripristinare parzialmente l'attività chaperonica nella variante K141E. I nostri risultati mostrano che l'effetto della paroxetina è più forte sulla variante K141E e influisce sulla compattezza delle sue conformazioni. Per confrontare direttamente i risultati computazionali con gli esperimenti smFRET, è stato sviluppato FRETpredict, un programma Python per calcolare l'efficienza FRET da strutture e traiettorie proteiche. Questo software è disponibile gratuitamente su GitHub (KULL-Centre/FRETpredict/) e come pacchetto Python PyPI. Inoltre, le simulazioni di dinamica molecolare sono state impiegate per progettare un nuovo sensore FRET per la rilevazione della proteina spike Sars-Cov-2. I risultati ottenuti in questa tesi di dottorato forniscono la prima caratterizzazione strutturale 3D di HSPB8 e rivelano gli effetti della mutazione patogena K141E sui suoi ensemble conformazionali. Questi risultati permettono di razionalizzare gli effetti patogeni della mutazione in termini di cambiamenti conformazionali. Questo lavoro ha contribuito a migliorare la comprensione di come la struttura e la dinamica di HSPB8 siano correlate a fattori come la forza ionica e la presenza di piccole molecole. Il software FRETpredict, che opera in modo efficiente su grandi ensemble conformazionali di proteine, facilita la validazione o il perfezionamento dei modelli molecolari e l'interpretazione dei dati sperimentali.

Combinazione di simulazioni Enhanced Sampling, reti neurali e tecniche FRET per lo studio della struttura e della dinamica di small Heat Shock Proteins / Daniele Montepietra , 2023 Mar 30. 35. ciclo, Anno Accademico 2021/2022.

Combinazione di simulazioni Enhanced Sampling, reti neurali e tecniche FRET per lo studio della struttura e della dinamica di small Heat Shock Proteins

MONTEPIETRA, DANIELE
2023

Abstract

Intrinsically disordered proteins (IDPs) are abundant in cells and have central roles in protein-protein interaction networks. Many are involved in cancer, aging, and neurodegenerative diseases. IDPs do not possess a single native structure but an ensemble of possible conformations related to interactions with binding partners. Due to this inherent flexibility, conventional methods for determining the structure of globular proteins may not be directly applicable to IDPs. A primary aim of this work was to use Enhanced Sampling Molecular Dynamics simulations in conjunction with Neural Network algorithms and FRET biophysical techniques to achieve a deeper understanding of the structure and dynamics of a specific IDP, namely Heat Shock Protein B8 (HSPB8), a human chaperone involved in preventing the aggregation of misfolded proteins, whose mutation leads to neurodegenerative diseases. HSPB8 consists of a conserved structure called the α-crystallin domain and two intrinsically disordered regions (IDRs). At present, no experimental 3D structures are available for wt HSPB8 and its K141E variant. To establish suitable force field parameters and a computational workflow for simulating IDPs, extensive Temperature Replica Exchange (TREMD) simulations of HSPB8 variants were analyzed with a Neural Network algorithm called EncoderMap, and the results were compared to available experimental data. The findings allowed us to determine key factors affecting the structure of the HSPB8 K141E variant. Indeed, its structures tend to be more compact, and the hydrophobic area exposed to the solvent is reduced. These features likely play a role in the K141E reduced chaperone activity. Because IDPs are highly sensitive to environmental conditions, the effects of high ionic strength on the dynamics of HSPB8 IDRs were also investigated. The results show that ionic strength might have differential effects on the conformational propensities of distinct regions of HSPB8. This information will help to understand the mutant's behavior and determine suitable conditions for accurately studying it. The effect of small-molecule binding to IDPs was investigated by simulating HSPB8 variants in the presence of paroxetine, a small molecule commonly found in antidepressant drugs that was shown to have a high affinity for HSPB8 and is able to partially restore the chaperone activity in the mutated K141E variant. Our results show that the paroxetine effect is stronger on the K141E variant and affects the compactness of its conformations. To directly compare computational results with smFRET experiments, we developed FRETpredict, a Python program to calculate FRET efficiency from protein structures and trajectories based on the Rotamer Library Approach. This software is freely available on GitHub (KULL-Centre/FRETpredict/) and as a Python PyPI package. Moreover, Molecular Dynamics simulations were employed to design a novel FRET sensor for the detection of Sars-Cov-2 spike protein. The results obtained in this Ph.D. thesis provide the first 3D structural characterization of HSPB8 and reveal the effects of the pathogenic K141E mutation on its conformational ensembles. These results offer the possibility of rationalizing the pathogenic effects of the K141E mutation in terms of conformational changes. This work has helped to improve the understanding of how HSPB8 structure and dynamics are related to environmental changes such as ionic strength and the presence of small molecules. The FRETpredict software, efficiently operating on large protein conformational ensembles, facilitates the validation or refinement of molecular models and the interpretation of experimental data.
Combining Enhanced Sampling simulations, Neural networks, and FRET technique to study the structure and dynamics of small Heat Shock Proteins
30-mar-2023
BRANCOLINI, Giorgia
CECCONI, CIRO
File in questo prodotto:
File Dimensione Formato  
PhD_thesis_Daniele_Montepietra.pdf

Open access

Descrizione: Tesi definitiva Montepietra Daniele
Tipologia: Tesi di dottorato
Dimensione 30.03 MB
Formato Adobe PDF
30.03 MB Adobe PDF Visualizza/Apri
Pubblicazioni consigliate

Licenza Creative Commons
I metadati presenti in IRIS UNIMORE sono rilasciati con licenza Creative Commons CC0 1.0 Universal, mentre i file delle pubblicazioni sono rilasciati con licenza Attribuzione 4.0 Internazionale (CC BY 4.0), salvo diversa indicazione.
In caso di violazione di copyright, contattare Supporto Iris

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11380/1301098
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus ND
  • ???jsp.display-item.citation.isi??? ND
social impact