Session W32: Self-Assembly of Biomacromolecules: From Simulations to Experiments

Sequence Effects on Thermodynamic Interactions in Protein Polymers

Biological polymers offer an as-yet unparalleled level of sequence control in polymer design, enabling them to achieve a wide variety of properties in natural systems. A key question is how tuning protein sequence changes coarse-grained interactions in protein-solvent and protein-protein systems.  To explore this effect, we have studied the solubility of a wide variety of elastin-like proteins (ELPs) in cosolvent systems that include water, alcohols, and salts, providing tuning of the hydrophobicity and ionic strength of the solvent.  We observe for the first time cononsolvency effects in ELPs, mapping in detail the phase diagram for one commonly studied sequence as a function of solvent conditions.  To understand the effect of single residue mutations, we explored a variety of different hydrophobic, hydrophilic, and charged amino acid mutations within the ELPs, showing how the cononsolvency behavior changes as a function of sequence.  Interestingly, proteins can have different phase behavior even if they are compositional isomers, showing that specific sequence effects can play a relatively large role in the thermodynamics of these systems in a way that remains poorly understood.  Block copolymers are further explored to understand protein-protein interactions.   

Electrostatics-driven peptide-directed encapsulation of nanoparticles into protein cages

Experiments have shown the gold nanoparticles (NPs) which are densely decorated by the short positively-charged ligands and sparsely decorated by longer uncharged cargo-loading peptides (CLPs) could be encapsulated into negatively-charged encapsulin protein cages with high efficiency. We build an experimentally-informed coarse-grained model with explicit ions and a Martini representation of water, and use it to perform molecular dynamics simulations to probe the mechanisms responsible for the encapsulation as the favorable co-assembly product. Experimental observations and simulations reveal three co-assembly states as a function of salt concentration c: for high salt (c > 0.5 M), the binding energy between protomer and NP is weak (~5–10 kBT) which leads to empty cages, for low salt (c < 0.2 M) the protomer-NP attraction is very strong yielding kinetically trapped protomer-NP aggregates, for 0.2 M < c < 0.5 M, NP-encapsulated protein cages are observed which correspond to a binding energy of ~15–20 KBT. Simulations reveal that NP encapsulation at high efficiency is the result of a two-step process: the electrostatic attraction between NPs and protomers driving the NPs into the recruitment zone of protomers, which is followed by the CLPs directing the NPs into the favorable binding region of the protomer.   

Characterization of biomimetic membranes formed from amphiphilic block copolymers assembled at oil-water interfaces

The droplet interface bilayer (DIB) is a model membrane technique that leverages aqueous compartments submerged in oil to form lipid membrane-divided functional materials, with comparable hierarchal architectures and diverse functionalities to natural cells and tissues. Instead of using lipids, amphiphilic block copolymers (ABCs) that can assemble into lamellar structures enable alternative membrane compositions that have greater mechanical and chemical tunability. Poly(butadiene -b- poly(ethylene oxide) (PBPEO) with molecular weights c.a. 1 kDa is of specific interest as it forms biomimetic membranes capable of hosting integral membrane proteins. Yet, its use in DIBs and tissue-like materials has been limited to membranes containing both lipids and ABCs. Thus, the effects that ABC molecular weight, concentration, and oil selection have on the self-assembly of PBPEO-stabilized DIBs are not well understood. Preliminary results indicate that varying the molecular weights of the polymer and the alkane oils in which they are dissolved affects bilayer thinning and therefore membrane adhesion due to differing amounts of polymer chain swelling. The objectives of this work thus aim to characterize the stability, adhesion, morphology, and selective transport properties of non-lipid amphiphile interfacial membranes through pendant drop interfacial tensiometry, electrophysiology characterization, and imaging of PBPEO-coated adhesive droplets.    

Predicting and Simulating the Self-Assembly of Sequence-Specific Peptoids

Polymer synthesis has grown increasingly sophisticated, allowing for precise sequence-controlled polymers with tailored properties and routes to new materials with highly tuned structure and functionality. However, the enormous number of possible sequences requires robust and efficient modeling to understand and predict how sequence impacts macromolecular self-assembly. Here, we use sequence-specific polypeptoids (a biomimetic of polypeptides) as a platform for developing design rules for relating chemical sequence to polymer conformation. Our earlier atomistic simulation studies of small polypeptoid systems with advanced sampling molecular dynamics methods examined changes in the local and global structure of short chains in response to the number and location of the hydrophobic and chiral monomers, with excellent agreement with experiments. In this work, we developed a bottom-up coarse-grained peptoid model, which allows access to longer and multiple peptoid systems. With this simulation workflow we study the effect of sequence on broader chain shape effects and self-assembly behavior. Moreover, we leverage inverse design methods based on genetic optimization to suggest sequences with unique folding and self-assembly properties into varied structures and phases. These new computational methods provide a molecular-level understanding of the factors governing polymer conformation and offer new in silico screening tools to guide the development of sequence-specific polymeric materials with tunable properties.   

Block Sequence Effects on Phase Behavior, Oligomerization, and Conformation of Racemic Polyampholyte Peptides

Polyampholytes (PAs) are polymers containing both positively and negatively ionizable groups along their backbone. Using solid phase synthesis, we synthesized a set of 5 PA peptides consisting of 16 glutamic acid and 16 lysine residues arranged in increasing blocks sizes from 1 residue (alternating sequence) to 16 residues (diblock sequence). Our previous study on PA peptides consisting of all L-chiral residues demonstrated sequence-dependent secondary structure formation; thus, complicating determination of eletrostatic contribution to PA conformation and phase behavior. To minimize hydrogen bonding, we redesigned our PAs to consist of randomly ordered D,L-chiral residues instead. From FTIR and CD measurements, no secondary structure was detected. Similar to L-chiral PAs, increasing block size increased phase separation; however, liquid-liquid phase separation was observed instead of liquid-solid phase separation. We also observed increasing oligomerization of the PA peptides with increasing block size. Through SAXS, we detail the effect of block size on the conformation of PA assemblies and how block size be used to tune inter- and intra-molecular electrostatic interactions.   

Discovering optimal kinetic pathways for self-assembly using automatic differentiation

During self-assembly of macromolecules ranging from ribosomes to viral capsids, the formation of long-lived intermediates or kinetic traps can dramatically reduce yield of the functional products. Understanding biological mechanisms for avoiding traps and efficiently assembling is essential for designing synthetic assembly systems, but learning optimal solutions requires numerical searches in high-dimensional parameter spaces. Here, we exploit powerful automatic differentiation algorithms commonly employed by deep learning frameworks to optimize physical models of reversible self-assembly, discovering diverse solutions in the space of rate constants for 3-7 subunit complexes. We define two biologically-inspired protocols that prevent kinetic trapping through either internal design of subunit binding kinetics or external design of subunit titration in time. Our third protocol acts to recycle intermediates, mimicking energy-consuming enzymes. Whilst all protocols can produce good solutions, diverse subunits always helps; these complexes access more efficient solutions when following external control protocols, and are simpler to design for internal control.  Our results identify universal scaling in the cost of kinetic trapping, and provide multiple strategies for eliminating trapping and maximizing assembly yield across large parameter spaces.    

Biomolecules for non-biological things: Polymer, 2-d lattice, and liquid crystal construction through peptide ‘bundlemer’ design and solution assembly

A new solution assembled system comprised of theoretically designed coiled coil bundle motifs, also known as ‘bundlemers’, will be introduced.  The molecules and nanostructures are not natural sequences and provide opportunity for arbitrary nanostructure creation with peptides.  With control of the display of all amino acid side chains (both natural and non-natural) throughout the peptide bundles, desired physical and covalent interactions have been designed to produce 1-D polymer nanostructures as well as 2-D assembled lattices.  One-dimensional nanostructures span exotically rigid rod molecules that produce a variety of liquid crystal phases to semi-flexible chains, the flexibility of which are controlled by the interbundle linking chemistry. Computational design is used to design bundlemers with different net charged character to manipulate their interactions in solution. Mixtures of oppositely charged bundlemer particles produce either 2-D lattice assemblies or amorphous aggregates, depending on the specific display of charge on the bundlemer particle surfaces.  Furthermore, bundlemer particles were designed with only positive or negative charges on the bundlemer to produce an overall particle charge as opposed to the typical mix of positive and negative charges on natural proteins that produce a net charge.  The exclusively positively or negatively charged particles produce unique assembly and liquid crystal behavior in solution and when used in mixtures.   

Self-limitation in geometrically frustrated, deformable particle assemblies with finite attraction range

Geometrically frustrated assemblies are an emerging class of systems where inter-subunit misfits propagate to large-scale strain gradients, giving rise to anomalous self-limiting thermodynamics under certain conditions.  Recently, the “curvamer” model was introduced to study self-limitation in 1D stacks of deformable, cylindrical shell-like particles, where an elastic energy emerges from curvature changes in stacks of uniformly spaced particles.  In general, elastic strains will also be borne out of stretching cohesive bonds between particles.   Here, we generalize the curvamer model to consider the effect of inter-particle bond stiffness, or alternatively finite attraction ranges between particles.  From a continuum elastic theory and coarse-grained numerical model, we find stack size is controlled by not only the ratio of inter-particle adhesion to intra-particle stiffness but also the ratio of intra-particle stiffness to inter-particle stiffness, which controls the nature of frustration propagation through the stack and the regimes of self-limiting behavior.  We also introduce a numerical model to explore effects of particle geometry, where we expect to see bond-stretched stacking dominate for hyperbolic and spherical curvamers as large elastic costs suppress curvature changes for these shapes. These predictions provide critical guidance for experimental realizations of frustrated particle systems designed to exhibit self-limitation at especially large multi-particle scales.   

Bio-Inspired Random Heteropolymers

A new class of bio-inspired PMMA-based random heteropolymers (RHPs) have displayed versatile protein-like properties, including catalytic activities, transporting protons across a bilayer lipid membrane, and protecting protein in a non-native environment. By using atomistic molecular dynamics simulation, these RHP chains collapse into compact globules, and the hydration of monomers is frustrated. The interaction with a hydrophobic substrate (i.e., graphene and its derivatives) and a β-barrel protein is driven by the ethylene glycol side chains, and the adsorption is hindered by the glassy PMMA backbone architecture and substrate hydrophilicity. These protein-like properties, including compactness, hydration frustration, and protein stabilization through heterodimerization are sequence-insensitive, which is not generally held for other similar backbone architectures. The design of PMMA-based RHP provides an alternative orthogonal to the sequence-structure-function paradigm of the protein.   

Intermolecular Protein Interactions and Self-assembly of a Synthetic Therapeutic T-cell Receptor-like Molecule

The emergence of protein engineering has enabled the synthesis of tailor-made therapeutic molecules designed to target specific diseases. One such class of biologics is the immune mobilizing monoclonal T-cell receptor against cancer (ImmTAC). ImmTACs target cancerous cells with high specificity and harness the body’s own immune response. These potentially transformative molecules are unique in the biopharmaceutical space.

The ability to create proteins not found in nature has profound implications for the rational control of phase behaviour, protein-protein interactions (PPIs), and self-assembly. Natural proteins have been finely tuned through evolution to be colloidally stable under physiological conditions, maintaining a delicate balance between solubility and aggregation to ensure proper cellular function. In contrast, engineered proteins like the ImmTACs offer a unique opportunity to explore the effects of deliberate design modifications and the consequences for phase behaviour. These modifications, aimed at enhancing target specificity and immune activation, also impact the intermolecular interactions and hence the propensity of the protein for self-assembly or aggregation.

Here we show that the design of the ImmTAC molecule leads to a highly anisotropic surface charge distribution which in turn results in unexpected phase behaviour. Using a range of analytical tools such as light scattering, analytical ultracentrifugation (AUC), and thermal denaturation, we characterise the protein-protein and protein-excipient interactions and observe the formation of small, but stable oligomers.   

Beyond Antibiotics: Unravelling A Death Mechanism of Superbugs by Nanoengineered Star Peptide Polymers through Molecular Simulation and Experiment

 Multi-drug resistant (MDR) bacteria pose a significant threat to global health and the economy, with staggering statistics revealing the severity of the issue. A recently synthesized class of polymers, so-called Structurally Nanoengineered Anti-Microbial Peptide Polymers (SNAPPs) composed of arms made of amino acid residues have shown to display superior antibacterial performance against gram negative and gram positive bacteria compared to traditional antibiotics. Despite laboratory experiments confirming SNAPP's effectiveness in killing bacteria, the molecular mechanisms responsible for the kill mechanism and therefore corresponding control parameters are yet to be fully understood. Molecular dynamics (MD) simulation is a powerful technique that enables direct studies of molecular scale interactions among SNAPPs and bacterial cell envelope. MD also allows to establish the complex linkages between SNAPPS’s chemical structure parameters and its antibacterial performance indicators, paving the way to optimise SNAPP’s chemical structure for maximum performance through machine learning and statistical optimization.

 Through MD simulations of interaction of SNAPP composed of valine and lysine residues in its arms, we established step-by-step pore forming mechanism which leads to bacteria cell death. The mechansim was further validated through laboratory experiment in vivo.    

Algebra for classical multiparticle complexes

Theoretical studies in biophysics, polymer physics, and other fields often focus on the behavior of large complexes of classically behaving molecules held together by site-specific interactions. Mathematical methods have yet to be developed, however, for handling the combinatorial explosion of complexes that often occur in these systems. We introduce a general mathematical language for describing the formation and thermodynamics of classical multiparticle complexes in terms of algebraically-defined assembly rules. At the heart of this formalism is a Fock space that supports the creation and annihilation of not only individual particles but also multiparticle interactions and binding site occupancies. The action of these rules on component particles occurs through a manifestation of Wick's theorem. To facilitate mathematical computations, we introduce a compact diagrammatic representation that makes the algebra visually intuitive. We demonstrate the formalism's utility in evaluating statistical properties of assembly in static and dynamic environments for a number of models in chemical kinetics and polymer physics, and highlight its efficacy as a basis for deterministic and stochastic analysis of infinitely-extendable systems with varying heterogeneous complexity.