Here I describe my work in developing a computer-modeling platform for simulating cellular organization and kinetics in a more realistic fashion than other existing technologies, and apply this technology to two large-scale biological systems, microtubule assembly and cytoplasmic crowding. This platform facilitates the multiscale approach in which the consequence of atomic details is measured on a bulk scale.

In the first application, I created a coarse-grain model for microtubule assembly derived from atomic-scale calculations. Specifically, I use high-resolution, all-atom Molecular Dynamics simulations to quantify critical interaction strengths and conformational dependencies thought to be central to the assembly process. I then incorporate these parameters into coarse-grain Brownian Dynamics simulations and kinetic simulations that can handle the large timescales required for MT assembly. My results show that the unassembled GDP-tubulin heterodimer exists in a continuum of conformations ranging between straight and bent, but, in agreement with existing structural data, suggests that an intermediate bent state has a lower free energy (by ~1 kcal/mol) and thus dominates in solution. In agreement with predictions of the lattice model of microtubule assembly, lateral binding of two tubulins strongly shifts the conformational equilibrium towards the straight state, which is then ~1 kcal/mol lower in free energy than the bent state. Finally, calculations of colchicine binding to a single tubulin dimer strongly shifts the equilibrium toward the bent states, and disfavors the straight state to the extent that it is no longer thermodynamically populated.

In the second application, I created a coarse-grain model of the bacterial protein cytoplasm. I developed automated ways of coarse- graining macromolecules based on experimental structures or homology models, and simulated proteins at a domain-level of resolution. My results show that an intermediate resolution model is sufficient to achieve the results of the high-resolution model. In addition, we show that incorporation of shape, while of no consequence in dilute solution, significantly transforms the energetic properties of crowded media via weak, nonspecific interactions. Providing a more complete understanding of excluded volume in vivo, our model is then applied to address the open questions of the effects of crowding on effective diffusion and association reactions.

More accurate comparative (homology) models would enable greater biological understanding through structural genomics efforts as well as aid in biological and small-molecule drug development. However, improving the accuracy of comparative models beyond that of the homologous template protein has proven extremely difficult for many years. The primary aim of this work is to develop more accurate, molecular mechanics-based, computational methods for refining loops in comparative models. My approach is two-fold:

A. Create a set of protein "model systems" that exhibit specific types of modeling error as found surrounding loops in comparative models

B. Develop new loop-sampling methods that optimize atoms outside the loop

The types of structural errors found in comparative models can be divided into bins based on sampling degrees-of-freedom: 1. side-chain error, 2. backbone error, and 3. larger-scale structural errors, such as helix or domain orientations. In chapter 1, we perturbed crystal structures to contain side-chain errors exclusively (error type 1.) We then augmented our previous loop prediction method to simultaneously optimize side-chains surrounding the loop. Results show that our new method can recover the native state in most of the cases where our previous method failed. In chapter 2, we chose homology models of antibodies as a test system to investigate loop prediction when the surrounding backbone atoms are incorrect (error type 2.) We predict the antibody H3 hyper-variable loop ab initio while the remaining five, hyper-variable loops are modeled using loop templates whose backbone atoms tend to deviate slightly from native. By increasing H3 loop sampling and performing optimizations on the surrounding loops iteratively, we were able to increase accuracy over previous methods. In chapter 3, through collaboration with Arjun Narayanan, we have taken initial steps in analyzing the determinants of variation in antibody light and heavy domain orientation (error type 3.) In chapter 4, through collaboration with Sergio Wong, I applied these new loop prediction methods to investigate a previous hypothesis that antibody H3 loops rigidify during affinity maturation which occurs within B-cells upon antigen encounter. In summary, this work constitutes a significant step towards a general method of comparative model refinement.

Allostery describes altered protein function at one site due to a perturbation at another site. One mechanism of allostery involves correlated motions, which can occur even in the absence of substantial conformational change. I present two novel information-theoretic molecular dynamics simulation analysis methods, one based on the mutual information and another based on the Kullback-Leibler divergence, to identify statistically significant correlated motions from equilibrium molecular dynamics simulations and statistically significant torsional population shifts when comparing sets of simulations under different conditions. I describe applications of these methods, and then novel experiments and kinetic modeling regarding the enzymology of caspase-1. These showed caspase-1's robust cooperativity requires substrate or inhibitor-assisted dimerization, and that even within the dimer, the enzyme is more active when two substrates are bound than when one substrate is bound.

Transcription is the process in which our genetic information in the form of DNA is first utilized to create the machinery, structural integrity, and regulatory components that allows for cells to simply survive. While there are many, many, many regulatory modes and mechanisms that govern processes from enzyme catalysis to cell differentiation (and everything in between), the regulation of transcription is the initial decision to ‘put the players on the pitch.’ Transcriptional regulation is immensely complex, as it rightfully should be, as it is required for both simple procedures, for example simply increasing the mRNA copy number of a single species in response to a cellular cue, to intricately coordinated events such as decisions regarding cell fate. Transcription factors are the protein components that interplay between environmental cues and gene expression. In particular, nuclear receptors are ligand dependent transcription factors that are modulated by a small molecule ligand which provides information about a particular metabolic pathway. The focus of my research has been a biochemical and structural understanding of how this subset of transcriptional machinery is implicated in the etiology of disease.

A multiscale, modular approach to protein sampling with novel Monte Carlo algorithms is is presented. The systems studied use an all atom forcefield with a Generalized Born implicit solvation model. The multiscale approach addresses 3 degrees of freedom: 1) the solvation terms, 2) the sidechain degrees of freedom, and 3) the backbone degrees of freedom. The goal of the work is to identify the special design issues surrounding these degrees of freedom, and create an overall sampling approach that optimizes all of these, while generating coherent trajectories that obey detailed balance. This design is expected to sample challenging, highly constrained systems that may be exceedingly difficult using standard molecular dynamics methods. The work presents present developments with regard to algorithmic approaches, design features, and applications to protein systems. Future directions in these areas are also discussed.

Cells use a variety of different mechanisms to sense and respond to the constantly changing environment. Single celled organisms use signaling pathways to find food, escape toxins, and protect themselves in dangerous or overcrowded environments. The failure of cell signaling in these organisms usually leads to the death of the cell. In a multi-celled organism, individual cells rely on signaling pathways to cooperate with the rest of the organism. A failure in cell to cell signaling in higher organisms may lead to the brief survival of the cell but the eventual failure and death of the organism as a whole. We chose to apply various methods to study cell signaling and communication in both prokaryotic and eukaryotic organisms. The first two approaches we took to this problem were entirely computational. We developed parameters for phosphorylated amino acids and used these parameters to predict structural changes as a function of phosphorylation. In addition, we showed that both phosphate charge and the geometry by which this phosphate interacts with other residues determine the energy gained or lost as a result of this interaction. Next, we joined computation and experiment to successfully predict agonists and antagonists for a G-protein coupled receptor. In a follow up study, we used the information gained on the G-protein coupled receptor to investigate selectivity among a set of similar receptors from different mammals. In the final section of this work, we use a mixture of computation and experiment to show that two component signaling pathways do not interfere with one another in vivo. We then apply this knowledge to deconstruct the bacterial chemotaxis pathway into two separate, orthogonal signaling systems.