Neutrons and supercomputing for Complex Biological Systems

Frontier challenges in biological research increasingly require gaining a predictive understanding of complex dynamic and flexible multi-component systems. Gaining that understanding can come from combining several experimental approaches to inform multi-scale computer models and simulations. Complementary experimental approaches used include electron microscopy, mass spectrometry, X-ray scattering, and NMR, but outstanding challenges remain. Neutron scattering has great potential to address the remaining challenges by providing elusive information that cannot be obtained otherwise. Researchers are now gaining access to new instrumentation on intense neutron beam lines at the Spallation Neutron Source (SNS) and the High Flux Isotope Reactor (HFIR) at Oak Ridge National Laboratory (ORNL). An unprecedented opportunity exists that we are exploiting by developing and broadening the use of neutrons in biological research by leveraging deuterium labeling and high performance computing. In order to develop this innovative integrated approach, and to take full advantage of the increase in availability and capability of neutron beam lines, further technological advances are required. In this talk I will present an overview of neutron facilities at ORNL, and give examples of their growing application in biological research. I will then discuss how the future challenges in biology are driving further technological developments that will lead to new understanding in the emerging areas of dynamic functional assemblies, disorder and flexibility, biological membranes and associated complexes, and biomolecular function and ligand binding. Neutrons can provide unique information that will transform these areas of research, opening up new lines of biological inquiry.

Insights into the phosphoryl transfer catalyzed by cAMP-dependent protein kinase

Protein kinases are involved in a number of cell signaling pathways. They catalyze phosphorylation of proteins and regulate the majority of cellular processes (such as growth, differentiation, lipid metabolism, regulation of sugar, nucleic acid synthesis, etc.). Chemically, protein kinases covalently transfer the gamma-phosphate group of a nucleoside triphosphate (e.g. ATP) to a hydroxyl group of a Ser, Thr or Tyr residue of substrate protein or peptide. The reaction involves moving hydrogen atoms between the enzyme, substrate and nucleoside. The unanswered question is whether the proton transfer from the Ser residue happens before the phosphoryl transfer using the general acid-base catalyst, Asp166, or after the reaction went through the transition state by directly protonating the phosphate group. To address this key question about the phosphoryl transfer, we determined a number of X-ray structures of ternary complexes of catalytic subunit of cAMP-dependent protein kinase (PKAc) with various substrates, nucleotides and cofactors. Importantly, we were able to trap and mimic the initial (Michaelis complex) and final (product complex) stages of the reaction. The results demonstrate that Mg2+, Ca2+, Sr2+, and Ba2+ metal ions bind to the active site and facilitate the reaction to produce ADP and a phosphorylated peptide. The study also revealed that metal-free PKAc can facilitate the phosphoryl transfer reaction; a result that was confirmed with single turnover enzyme kinetics measurements. Comparison of the product and the pseudo-Michaelis complex structures, in conjunction with molecular dynamics simulations, reveals conformational, coordination, and hydrogen bonding changes that help further our understanding of the mechanism, roles of metals, and active site residues involved in PKAc activity.

Combining neutron crystallography, high-performance computing for enzyme design

Enzymes continue to expand their role in industry as a "green" option for the synthesis of value-added products. They are targeted for the design of drugs in pharmaceutical applications and also for protein engineering in industry to improve their efficiency, stability, and specificity. Knowledge of the exact mechanisms of enzymatic reactions may provide essential information for more effective drug design and enzyme engineering. For the first time, we are employing a joint X-ray/neutron (XN) protein crystallographic technique in combination with high-performance computing, including QM and QM/MM calculations, MD and Rosetta simulations, to investigate the mechanisms of several enzymes that are important to renewable energy and chemical synthesis. D-xylose isomerase (XI) is an enzyme which can be used to increase the production of biofuels from lignocellulosic biomass and also to synthesize rare sugars for pharmaceutical industry. XI catalyzes the reversible multi-stage sugar inter-conversion reaction facilitated by the presence of two divalent metal cations in its active site. It primarily catalyzes the isomerization of the aldo-sugar D-xylose to the keto-isomer D-xylulose, but can also epimerize L-arabinose into L-ribose, albeit much less efficiently. The reaction involves moving hydrogen atoms between the protein residues, sugar and water molecules, and can only be understood if hydrogen atoms are visualized at each reaction stage. We have obtained a number of joint XN structures of XI complexes representing snapshots along the reaction path with D-glucose, D-xylose and L-arabinose. The suggested reaction mechanism has been verified by QM calculations using the novel O(N) methodology. We are using this structural and mechanistic information to re-design XI to be more efficient on D-xylose and L-arabinose for biofuels and biomedical applications by employing QM/MM, MD, and Rosetta methodologies.

Ultra-high resolution neutron and X-ray crystallography: structure of crambin

Neutron diffraction data to 1.1 Å was collected on a crystal of the small protein crambin at the Protein Crystallography Station (PCS) at Los Alamos, the highest resolution neutron structure of a protein to date, and a technical benchmark for the instrument. 95 % of the hydrogen atoms in the protein structure were resolved. The data allowed for the refinement of anisotropic temperature factors for selected deuterium atoms within the protein. Hydrogen bonding networks ambiguous in room temperature, ultra-high resolution (0.84 Å) electron density maps are clarified in the nuclear density maps. The ultra-high resolution data also reveals unusual H/D exchange patterns and novel chemistry in the side chains and protein backbone. Complementary X-ray diffraction data was collected at 19-ID at the Advanced Photon Source, with extensive re-configuration of the beamline to allow for operation at higher energy settings.