Biomolecular Hydration and Interactions

The last two decades have seen an incredible revolution in structural biology: over 90,000 protein structures have now been solved at atomic resolution. Translating this wealth of static structural information into a molecular understanding of dynamic intracellular processes represents a grand challenge in molecular biology, with grave implications on our understanding of human health and disease. Progress in this area hinges on our ability to understand, predict, and manipulate the interactions of various proteins, given their precise structure, i.e., out of the veritable menagerie of molecules that a protein encounters in the cellular milieu -- ligands, peptide fragments, nucleic acids, membranes, and other proteins -- which molecules will it interact favorably with, how strong will the interaction be, and what are the strategies for modulating the interaction strength?

Based on recent work in our group and those of others, we hypothesize that the biggest challenge in accurately predicting protein interactions is precisely accounting for the role of water in mediating these interactions. Because all of biology happens in water, every biomolecular binding process requires the dehydration of the binding partners, with direct interactions between them replacing their individual interactions with water. While the direct interactions are straightforward to estimate, the protein-water interactions are difficult to quantify because proteins have incredibly complex surfaces that disrupt the inherent structure of water (strong hydrogen bonds) in countless different ways. So, what makes predicting protein interactions particularly challenging, is quantifying this disruption of the hydration shell water structure, which depends not only on the chemistry of the underlying protein surface, but also on the precise topography and chemical pattern of amino acids.

Using principles of liquid state theory in conjunction with novel simulation techniques, we are working towards developing a framework, which will transform our ability to accurately and efficiently predict protein interactions, and provide a basis for understanding how proteins respond to various perturbations, aggregate into multimeric assemblies, and are able to function.

Computational Material Design

By understanding the molecular underpinnings of solvation (of water and other liquids) adjacent to complex nanoscopic surfaces, that is, surfaces with nanoscopic pattern and/or texture, our goal is to inform the design of the next generation of advanced materials.

One example is the design of robust superhydrophobic surfaces (left). Surface texture can transform hydrophobic surfaces into "superhydrophobic" surfaces, endowed with properties like water-repellency, self-cleaning, and fouling resistance. However, superhydrophobicity is fragile, and depends on water being excluded from the surface texture. Thus, fully realizing its promise requires an understanding of the underpinnings of wetting-dewetting transitions on nanotextured surfaces.

Other examples of interest to our group include the design of superhydrophilic surfaces with superior protein non-fouling properties (center), nanoparticle membranes for desalination (right), and polymer electrolytes for batteries.

Understanding Hydrophobic Effects

The aversion of oil and water towards each other is something we have all experienced. The same driving forces that cause oil and water to phase separate, also drive important molecular level phenomena: protein folding/misfolding and aggregation, formation of micelles and membranes, and colloidal self-assembly. However, molecular-level hydrophobic effects can be manifest in mulifarious and non-trivial ways.

Most famously, the strength of hydrophobic interactions increases temperature, providing an explanation for the peculiar unfolding/denaturation of proteins on cooling. In more recent work, it was shown that extended hydrophobic surfaces can generically act as catalysts for the folding and unfolding of proteins, helping explain why amyloid fibril formation is facilitated by the presence of hydrophobic surfaces.

Examples of interesting questions concerning multi-faceted, molecular-level hydrophobic effects that we grapple with include: What is a good way to define the hydrophobicity of a complex, nanoscopic surface such as a protein? How does surface topography and/or roughness affect its hydrophobicity? What are the pathways to dewetting in hydrophobic environments? What role does the flexibility/rigidity of surface play in determining its hydrophobicity?

Developing Novel Simulation Techniques and Theoretical Models

Many of the problems that are of interest to us, require the use of enhanced sampling techniques. We make extensive use of configurational and path sampling techniques, and when necessary, develop new simulations techniques. We also develop highly efficient coarse-grained theoretical models to study phenomena characterized by large length scales and/or long time scales.

As an example of simulation methods developed by Dr. Patel, is the INDUS (INDirect Umbrella Sampling) method, for estimating the energetics of density fluctuations. INDUS has been applied extensively to characterize density fluctuations in bulk water, in volumes of various shapes and sizes and under a range of temperatures and pressures, as well as in volumes located at interfaces, such as the hydration shells of self-assembled monolayers and proteins.

More recently, Dr. Patel developed a novel simulation technique for estimating cavity hydration free energies that is about two orders of magnitude more efficient than conventional techniques. The method enables the estimation surface hydrophobicity efficiently, and has important implications on characterizing protein hydration and interactions.