Session L6: Bring Order from Disorder with Intrinsically Disordered Proteins

$\alpha -$synuclein under the magnifying glass. Insights from atomistic and coarse-grain simulations.

Neurodegenerative diseases are linked to the accumulation of misfolded intrinsically disordered proteins in the brain. Here, we use both all-atom and coarse-grain simulations to explore the intricate dynamics and the aggregation of $\alpha $-synuclein, the protein implicated in Parkinson's disease. We explore the free energy landscapes of $\alpha $-synuclein by using Molecular Dynamics simulations and extract information on the structure of the protein as well as on its binding affinities[1]. Next, to study the aggregation, we proceed with representing $\alpha $-synuclein as a chain of deformable particles that can adapt their geometry, binding affinities and can rearrange into different disordered and ordered structures[2,3]. We use Brownian Dynamics to simulate the translational and rotational motions of the particles[4], as well as their interaction properties[2]. The simulations show valuable insight into the internal dynamics of $\alpha $-synuclein[1] and the formation of ordered and disordered aggregates[2]. In addition, the study is extended to investigate the attachment and folding of a protein to a fiber[3]. [1]I.M.Ilie, D.Nayar, W.K.den Otter, N.F.A.van der Vegt {\&} W.J.Briels, in prep [2]I.M.Ilie, W.K.den Otter and W.J.Briels, JCP 144, 085103 (2016) [3]I.M.Ilie, W.K.den Otter and W.J.Briels, in prep [4]I.M.Ilie, W.J.Briels and W.K.den Otter, JCP 142, 114103 (2015)   

Early-Stage Aggregation of Human Islet Amyloid Polypeptide

Human islet amyloid polypeptide (hIAPP, or human amylin) is implicated in the development of type II diabetes. hIAPP is known to aggregate into amyloid fibrils; however, it is prefibrillar oligomeric species, rather than mature fibrils, that are proposed to be cytotoxic. In order to better understand the role of hIAPP aggregation in the onset of disease, as well as to design effective diagnostics and therapeutics, it is crucial to understand the mechanism of early-stage hIAPP aggregation. In this work, we use atomistic molecular dynamics simulations combined with multiple advanced sampling techniques to examine the formation of the hIAPP dimer and trimer. Metadynamics calculations reveal a free energy landscape for the hIAPP dimer, which suggest multiple possible transition pathways. We employ finite temperature string method calculations to identify favorable pathways for dimer and trimer formation, along with relevant free energy barriers and intermediate structures. Results provide valuable insights into the mechanisms and energetics of hIAPP aggregation. In addition, this work demonstrates that the finite temperature string method is an effective tool in the study of protein aggregation.   

The effects of bound state motion on macromolecular diffusion

The diffusion of macromolecules is modified in crowded environments by both inert obstacles and interaction sites. Molecules are generally slowed in their movement inducing transient anomalous subdiffusion. Obstacles also modify the kinetics and equilibrium behavior of interaction between mobile proteins. In some biophysical contexts, bound molecules can still experience mobility, for example transcription factors sliding along DNA, membrane proteins with some entry and diffusion within lipid domains, or proteins that can enter into non-membrane bound compartments such as the nucleolus. We used lattice and continuum models to study the diffusive behavior of tracer particles which bind to obstacles and can diffuse within them. We show that binding significantly alters the motion of tracers. The type and degree of motion while bound is a key determinant of the tracer mobility. Our work has implications for protein-protein movement and interactions within living cells, including those involving intrinsically disordered proteins.   

Intrinsically Disordered Proteins and the Origins of Multicellular Organisms

In simple multicellular organisms all of the cells are in direct contact with the surrounding milieu, whereas in complex multicellular organisms some cells are completely surrounded by other cells. Current phylogenetic trees indicate that complex multicellular organisms evolved independently from unicellular ancestors about 10 times, and only among the eukaryotes, including once for animals, twice each for green, red, and brown algae, and thrice for fungi.\\ \\Given these multiple independent evolutionary lineages, we asked two questions: 1. Which molecular functions underpinned the evolution of multicellular organisms?; and, 2. Which of these molecular functions depend on intrinsically disordered proteins (IDPs)? Compared to unicellularity, multicellularity requires the advent of molecules for cellular adhesion, for cell-cell communication and for developmental programs. In addition, the developmental programs need to be regulated over space and time. Finally, each multicellular organism has cell-specific biochemistry and physiology. Thus, the evolution of complex multicellular organisms from unicellular ancestors required five new classes of functions. To answer the second question we used Key-words in Swiss Protein ranked for associations with predictions of protein structure or disorder. With a Z-score of 18.8 compared to random-function proteins, “differentiation” was the biological process most strongly associated with IDPs. As expected from this result, large numbers of individual proteins associated with differentiation exhibit substantial regions of predicted disorder. For the animals – for which there is the most readily available data – all five of the underpinning molecular functions for multicellularity were found to depend critically on IDP-based mechanisms and other evidence supports these ideas. While the data are more sparse, IDPs seem to similarly underlie the five new classes of functions for plants and fungi as well, suggesting that IDPs were indeed crucial for the evolution of complex multicellular organisms.\\ \\These new findings necessitate a rethinking of the gene regulatory network models currently used to explain cellular differentiation and the evolution of complex multicellular organisms.   

Probing phase transitions in dynamic biopolymer complexation

In nature, biopolymers partition into dynamic compartments to facilitate and regulate their interactions. These dynamic compartments are referred to as membraneless organelles and consist of biopolymer rich interiors that rapidly assemble and disassemble to form liquid droplets, hydrogels or fibril structures. However, the physical interactions that affect the formation, dissolution, and regulation of these assemblages are poorly understood. Interestingly, polyelectrolyte complexes produced using simple homopolymers are similar to membraneless organelles. Polyelectrolyte complexation is an entropically driven process, where electrostatic attraction between oppositely charged polymers results in a release of bound counterions and rearrangement of water molecules. Under defined conditions, oppositely charged polyelectrolytes can form complexes consisting of a dense polymer rich phase in a polymer depleted aqueous phase. Whether complexation results in a liquid or solid precipitate depends on the strength of electrostatic interactions, which are mediated by salt concentration, acidity/basicity of the monomers and their distribution along the polymer backbone. In this work, we investigate the forces that govern membraneless organelle formation by engineering model polypeptide analogs of intrinsically disordered sequences and study their phase transition behavior. Our work holds the potential to develop a basic understanding of polyampholyte/polyelectrolyte complexation.   

Random-phase-approximation theory for sequence-dependent behaviors of intrinsically disordered proteins in liquid-liquid phase separation

The amino acid sequences of intrinsically disordered proteins (IDPs) have few hydrophobic but many more polar, charged, and aromatic residues. IDPs do not fold into a unique structure in isolation and can remain disordered while performing key biological functions. Recently, it was discovered that some IDPs can undergo liquid-liquid phase separation in an aqueous milieu. When in the cell, this process underlies membraneless organelles of condensed IDPs that often incorporate other biomolecules such as RNA and DNA. Without a membrane, these organelles can rapidly respond to environmental stimuli and play critical roles in many biological functions. Here, we present a polymer physics approach for this phenomenon by applying the random phase approximation (RPA) theory of polyampholytes to a collection of RNA helicase Ddx4 proteins with a charged and aromatic-enriched IDP region. Our theory predicts that Ddx4 phase behavior is significantly influenced by its specific charged sequence as well as pi-electron interactions associated with aromatic rings, consistent with recent experiments on wildtype and a charge-scrambled mutant of Ddx4. Our theory is applicable to any charged biopolymers and thus provides a general analytical framework for studying biological phase separation.   

FT-IR Study Reveals Intrinsically Disordered Nature of Heat Shock Protein 90

Heat shock protein 90 (Hsp90) is a highly conserved chaperone protein that enables the proper folding of a large number of structurally diverse proteins (a.k.a., clients) in the crowded cytosolic environment and plays a key role in regulating the heat shock response. A long standing open question is how Hsp90 accommodates the structural diversity of a large cohort of client proteins? We report ATR FTIR study on structural properties of Hsp90 C-terminal domain (CTD) and their temperature dependences. Effects of temperature on Hsp90 structure are dissected into the C-terminal domain (CTD) and the N-terminal/middle domain (NTMD). One of our major findings reveals that within a narrow temperature window across the physiological temperatures (35 to 45 C), Hsp90CTD exhibits significant increases in protein aggregation and increases in unordered structures. Despite the intrinsically disordered nature of Hsp90CTD, it retains a protected hydrophobic core at 40 C. Implications of these results will be discussed in the light of the structural dynamics and client diversity of Hsp90.   

Multiple structure-intrinsic disorder interactions regulate and coordinate Hox protein function.

During animal development, Hox transcription factors determine fate of developing tissues to generate diverse organs and appendages. Hox proteins are famous for their bizarre mutant phenotypes, such as replacing antennae with legs. Clearly, the functions of individual Hox proteins must be distinct and reliable \textit{in vivo}, or the organism risks malformation or death. However, within the Hox protein family, the DNA-binding homeodomains are highly conserved and the amino acids that contact DNA are nearly invariant. These observations raise the question: How do different Hox proteins correctly identify their distinct target genes using a common DNA binding domain? One possible means to modulate DNA binding is through the influence of the non-homeodomain protein regions, which differ significantly among Hox proteins. However genetic approaches never detected intra-protein interactions, and early biochemical attempts were hindered because the special features of ``intrinsically disordered'' sequences were not appreciated. We propose the first-ever structural model of a Hox protein to explain how specific contacts between distant, intrinsically disordered regions of the protein and the homeodomain regulate DNA binding and coordinate this activity with other Hox molecular functions.   

Mechanisms of selective transport through nuclear pore complex mimics

Few cellular processes require such intricate active control as transport through the nuclear envelope. The nuclear pore complex (NPC) facilitates all transport, preventing most macromolecules from crossing the envelope while allowing the passage of transport factors (TFs) and their cargo. While the basic biochemical interactions of transport are well-understood, the detailed mechanism remains a topic of significant debate. We create tunable mimics of the NPC using PEG hydrogels filled with FG nucleoporins, the intrinsically disordered proteins that line the NPC channel in vivo. Using fluorescence microscopy and single-molecule fluorescence spectroscopy, we measure TF diffusion through the NPC mimics. Modeling based on our results suggests two possible mechanisms of TF diffusion through the nuclear pore. We aim to distinguish between these possible mechanisms and to tune the mimic's parameters to maximize the rate of passage of TFs while inhibiting the passage of other inert molecules.   

Modeling phase separation in mixtures of intrinsically-disordered proteins

Phase separation in a pure or mixed solution of intrinsically-disordered proteins (IDPs) and its role in various biological processes has generated interest from the theoretical biophysics community. Phase separation of IDPs has been implicated in the formation of membrane-less organelles such as nucleoli, as well as in a mechanism of selectivity in transport through the nuclear pore complex. Based on a lattice model of polymers, we study the phase diagram of IDPs in a mixture and describe the selective exclusion of soluble proteins from the dense-phase IDP aggregates. The model captures the essential behaviour of phase separation by a minimal set of coarse-grained parameters, corresponding to the average monomer-monomer and monomer-protein attraction strength, as well as the protein-to-monomer size ratio. Contrary to the intuition that strong monomer-monomer interaction increases exclusion of soluble proteins from the dense IDP aggregates, our model predicts that the concentration of soluble proteins in the aggregate phase as a function of monomer-monomer attraction is non-monotonic. We corroborate the predictions of the lattice model using Langevin dynamics simulations of grafted polymers in planar and cylindrical geometries, mimicking various in-vivo and in-vitro conditions.   

Fluctuations in protein aggregation

Autocatalytic fibril nucleation has recently been proposed to be a determining factor for the spread of neurodegenerative diseases, but the same process could also be exploited to amplify minute quantities of protein aggregates in a diagnostic context. Recent advances in microfluidic technology allow analysis of protein aggregation in micron-scale samples potentially enabling such diagnostic approaches, but the theoretical foundations for the analysis and interpretation of such data are so far lacking. Here we study computationally the onset of protein aggregation in small volumes and show that the process is ruled by intrinsic fluctuations whose volume dependent distribution we also estimate theoretically. Based on these results, we develop a strategy to quantify in silico the statistical errors associated with the detection of aggregate containing samples. Our work opens a new perspective on the forecasting of protein aggregation in asymptomatic subjects.