The goal of this project is to test the hypothesis that the interplay between chain connectivity and hydrophobic clusters of branched aliphatic side chains in two members of the very common Rossmann-fold family of proteins, CheY and dihydrofolate reductase (DHFR), dictates their folding free energy surfaces. Available evidence on both proteins suggests that locally-connected clusters of isoleucine, leucine and valine side chains rapidly collapse via subdomains that can enhance or impede subsequent folding reactions leading to the native conformation. A battery of spectroscopic methods, at equilibrium and interfaced to ultra-rapid mixing systems, will probe the size, shape and pair-wise distances in the chemically-denatured state and in partially-folded states that appear in the microsecond time range after dilution to native-favoring conditions for CheY, complementing previous findings on DHFR. Chemical shift index and paramagnetic relaxation enhancement NMR measurements will probe for nonrandom structure in the chemically denatured state. Complementary pulse-quench hydrogen exchange experiments on CheY will probe the formation of secondary structure at the peptide and the site-specific level in the early intermediates. Mutational analysis will test the role of local and nonlocal ILV clusters in driving these early folding reactions, and permutations of the sequences will test the role of the connectivity of the polypeptide chain in driving the formation of these clusters and the N- and C-terminal subdomains. Appropriate permuted variants of CheY will be subjected to single molecule pulling experiments to study the effect of subdomain connectivity and the ILV cluster integrity on the cooperativity of the unfolding reaction and reveal the stabilization of partially-folded states. The experimental data will be used to validate course-grained MD simulations of the folding reactions of CheY and DHFR, and high-resolution simulations on CheY. It is anticipated that the combined application of experimental and computational methods on the same target will substantially enhance the value of both approaches and expedite the solution of the protein folding problem.

The protein folding problem remains as one of the outstanding challenges in molecular biophysics, and its solution would have a major impact on biology and the biotechnology industry. To expedite a solution to the folding problem, a collaborative network of investigators has been established to generate a comprehensive experimental data set on a single protein target that will validate companion coarse-grained and high-resolution MD simulations of its folding reaction. This collaborative approach will serve as a paradigm for the solution of other complex problems in biology. A micro-channel mixing system has been developed over the course of this work that allows access to microsecond folding reactions and that can be interfaced to a variety of spectroscopic methods. This technology has been shared with colleagues at other institutions, and its dissemination in the open literature has enabled others to study the early folding events in their target systems. Pursuit of these research objectives will also provide training opportunities for high school students, undergraduates, graduate students and postdoctoral fellows, and the scientific advances are being incorporated into a graduate molecular biophysics course.

Project Report

Training: One postdoctoral fellow, one graduate student, 4 summer undergraduate students and 1 high school student were trained during this period. Publications: Four research papers, 1 research proceeding, 2 reviews, 1 methods papers and 1 book chapter have resulted from this research. Five research papers and 2 methods papers are in various stages of preparation. Intellectual Merit: The influence that the sequence and topology of a protein have on its folding free energy landscape is not well understood. The goal of this project was to study the relationships between the sequences and topology and the structures of high-energy states on the folding landscape of CheY, a member of one of the most common families of structure in biology. We have found that this protein, which may be partly structured in the unfolded state, rapidly collapses to a structured intermediate within the microsecond dead time of our experiments. The collapse is native-like in the N-terminal half, but not the C-terminal half, of the protein, and native centric simulations from the Brooks group suggest nonproductive interactions between the two halves result in a trapped intermediate state. By connecting the N- and C- termini with a linker and introducing new termini at different positions throughout the protein, we were able to show that the wild type chain connectivity of the N-terminal half of the protein is essential for folding. Further, it was possible to partially relieve the frustration of the folding landscape by locally connecting a large hydrophobic cluster in the protein. We have also prepared the reagents and developed the technology that will soon enable us to assess the development of structure in specific regions of CheY in the nanosecond to microsecond time range. Our unique advances in technology have generated multiple collaborations focused on the high-energy states in a variety of other proteins. We have found that, similar to CheY and dihydrofolate reductase, unfolded proteins with > 100 amino acids tend to experience heterogeneous collapse to partially-folded states in the microsecond time range. Interestingly, these early intermediates tend to reflect the development of nonnative structure that must melt before productive folding begins. Collaborations with computational biologists are beginning to offer insights into these early misfolding events. Broader Impact Outreach: As members of the Protein Folding Consortium, funded by a Research Coordination Network grant from the National Science Foundation, we have had the opportunity to share our technological developments and establish collaborations with several groups in the US and internationally. Our methods have made a direct impact on training students and postdoctoral fellows from the following labs: Susan Marqusee (UC Berkeley), Daniel Raleigh (SUNY Stony Brook), Tobin Sosnick (Univ. of Chicago), Elisha Haas (Bar Ilan University, Israel), Thomas Irving (Illinois Institute of Technology and BioCAT), and Feng Gai (UPenn). These collaborations also allowed us to disseminate our continuous-flow fluorescence (CF-FL) and small angle x-ray scattering (CF-SAXS) technology to several of these groups. The graduate student funded by this grant, R. Paul Nobrega, has volunteered at a local high school (Wachusett Regional High School) to mentor Nation Honors Society Students with their science fair projects from 2011 to 2013. The topics covered include biology, chemistry, and physics, with the participation of approximately 15 students each year. Additional outreach efforts include judging state and local science fairs by members of the Matthews group. Dr. Bilsel helped organize an Arts and Science Night at a local elementary school, developing experiments on fluid flow, gravity, optics and interference. Dr. Bilsel also performed demonstrations on the states of matter for elementary school students at a local day care center. Mentoring: The research funded by this grant formed the basis for the training of a graduate student, three undergraduate students in the UMass Medical School summer program, one high school student and an undergraduate by an REU supplement. Two of the undergraduate students have gone on to join prestigious graduate programs at University of Pennsylvania and Vanderbilt University. Dr. Bilsel also mentored a physics science project for a high school student. Software: A web-based algorithm to calculate and display clusters of Branched Aliphatic Side Chains has been set up on the Umassmed Biotools web server. ( Software for data reduction and analysis of CF-SAXS data has been implemented at BioCAT at the Argonne National Laboratory. This software is designed to convert, in real time, several gigabytes of 2D SAXS images to 1D data and perform data averaging, buffer correction and report the radii of gyration as a function of folding time. This software has substantially increased the throughput at the beamline and enabled us and others to make the most efficient use of precise time at the facility. A stand-alone version of this software for pre-reduced datasets has been made available at This software can perform several automated routines, such as subtractions, transforms, plotting, SVD, Guinier analysis.

Project Start
Project End
Budget Start
Budget End
Support Year
Fiscal Year
Total Cost
Indirect Cost
University of Massachusetts Medical School
United States
Zip Code