Off-campus UMass Amherst users: To download campus access dissertations, please use the following link to log into our proxy server with your UMass Amherst user name and password.

Non-UMass Amherst users: Please talk to your librarian about requesting this dissertation through interlibrary loan.

Dissertations that have an embargo placed on them will not be available to anyone until the embargo expires.

Author ORCID Identifier


Open Access Dissertation

Document Type


Degree Name

Doctor of Philosophy (PhD)

Degree Program


Year Degree Awarded


Month Degree Awarded


First Advisor

Jianhan Chen

Second Advisor

Scott Auerbach

Third Advisor

Min Chen

Fourth Advisor

Gregory Grason

Subject Categories

Biochemistry, Biophysics, and Structural Biology | Computational Chemistry


Intrinsically disordered proteins (IDPs) are functional proteins that lack stable tertiary structures in the unbound state. They frequently remain dynamic even within specific complexes and assemblies. IDPs are major components of cellular regulatory networks and have been associated with cancers, diabetes, neurodegenerative diseases, and other human diseases. Computer simulations are essential for deriving a molecular description of the disordered protein ensembles and dynamic interactions for mechanistic understanding of IDPs in biology, diseases, and therapeutics. However, accurate simulation of the heterogeneous ensembles and dynamic interactions of IDPs is extremely challenging because of both the prohibitive computational cost and demanding force field accuracy. In this dissertation, we developed a set of enhanced sampling and multi-scale simulation methods to overcome these limitations, and successfully applied them to study the structure, interaction and phase separation of IDPs. We have first applied the state-of-the-art explicit solvent atomistic simulations to study the inhibitory mechanism of the disordered N-terminal domain of Staphylococcal peroxidase inhibitor (SPIN). We performed high-temperature simulations to study the coupled binding and folding process during SPIN inhibition of the host myeloperoxidase (MPO) enzyme. The results showed that differences in dynamics may provide a physical basis of the ability of different SPIN homologs to inhibit innate immunity. Recognizing the need for enhanced sampling methods for IDP simulation, we have developed a new replica-exchange with solute tempering (REST) protocol to achieve more efficient explicit solvent sampling of disordered protein ensembles. We proposed that the scaling of protein-water interactions in REST is a free parameter that could be optimized to better control how the protein conformational properties (e.g., chain expansion) at different effective temperatures and achieve more effective sampling. Specifically, we developed a REST3 protocol that rebalances the protein-protein and protein-water interactions and eliminates the unanticipated chain collapse at high temperature conditions in the previous REST2 protocol. Application to model IDPs demonstrated that REST3 prevented the conformational segregation during exchanges, leading to an effective temperature random walk across all conditions and accelerating the simulation of the protein conformational space. Even with enhanced sampling, accurate description of disordered conformations at atomistic level remains extremely challenging for complex IDPs. Alternatively, coarse-grained simulations can provide an effective strategy for overcoming the length- and time-scale limitations. Here, we refined a hybrid-resolution coarse-grained model (HyRes) for accurate simulation of disordered protein ensembles and dynamic protein interactions. HyRes contains atomistic backbone and coarse-grained sidechain beads, to provide semi-quantitative description of residual secondary structures and long-range interactions. Specifically, we introduced a surface area-based implicit solvation energy term, and iteratively re-optimized the effective interaction strength potentials. The new model, referred as HyRes II, provides near quantitative descriptions of IDP long-range non-specific interactions and secondary structures, at a level comparable to the latest atomistic protein force fields. Applied to the disordered N-terminal transactivation domain (TAD) of tumor suppressor p53, HyRes II faithfully recapitulates various nontrivial structural properties to a level of accuracy that is comparable to a99SB-disp, one of the best atomistic protein force fields. Moreover, we demonstrate HyRes II’s efficacy in accurately simulating the dynamic interaction between TAD and the DNA-binding domain of p53, generating structural ensembles that align closely with existing NMR data. Encouraged by successes of HyRes II for probing dynamic interactions of IDPs, we further investigated its suitability for simulating IDP-mediated phase separation, which underlies the formation of biomolecular condensates and has attracted intensive interests. Compared to the popular single-bead models, HyRes has the potential to describe backbone-mediated interactions and capture the role of residual structures in phase separation. Reimplemented on GPU, our simulations showed that HyRes is efficient enough to directly simulate the spontaneous phase separation of IDPs and at the time balanced enough to capture the effects of mutational and structural perturbations. For peptide GY-23, HyRes simulations reveal increased ��-structures in condensates, which are consistent with experimental observations. For the conserved region (CR) of TDP-43, HyRes simulations successfully recapitulate the apparent correlation between helical propensities and condensate stability. In depth analysis, however, revealed that residual helices did not directly mediate interpeptide interactions to stabilize the condensed phase. Instead, it is the balance between backbone and sidechain-mediated interactions, as modulated by residual structures, that actually determines phase separation propensity. Finally, we have applied the HyRes II model to study the dynamic interaction of West Nile virus (WNV) NS2B/NS3 proteases with the ClyA protein nanopore. Nanopore tweezers provide a powerful approach for label-free detection of protein dynamics at the single-molecule level, by capturing the protein analyte in the lumen of the nanopore. From the steered-MD and standard MD simulations, we discovered that the protease could bind dynamically to a middle region of the ClyA nanopore, mediated mainly by electrostatically interactions. In particular, we identified a key Glu residue within the ClyA lumen, mutation of which to Ala or Lys could further stabilize the protease/nanopore interaction. This led to the design a modified ClyA nanopore tweezer that can stably capture the protease and resolve the dynamics between NS2B/NS3 open and closed conformations.


Creative Commons License

Creative Commons Attribution 4.0 License
This work is licensed under a Creative Commons Attribution 4.0 License.