The Imaging Sciences Laboratory is involved in a major collaborative research effort with the NIH Institutes involving the use of image processing techniques and advanced computational techniques in structural biology to analyze electron micrographs and NMR spectra with the goal of determining macromolecular structures and dynamics. Recent efforts have concentrated on the 3D reconstruction, analysis and interpretation of the structures of icosahedral virus capsids in addition to structure determination and analysis of isolated and complexed proteins and nucleic acids. Ongoing research involves analyses of structures related to papillomavirus and RNA containing virus capsids. In addition, we have also been developing computational tools for the study of the structure and dynamics of biological macromolecules using Nuclear Magnetic Resonance (NMR) and other data. We develop and maintain the Xplor-NIH software package for structure determination, which is used in the NMR labs in the Institutes, and also worldwide. Human papillomavirus (HPV) has been implicated as the causative agent in cervical and other epithelial cancers. The major capsid protein (L1), initially forms a loosely connected procapsid which, under in vitro conditions, condenses over several hours into the more familiar 60 nm-diameter papillomavirus capsid. Current work involves collaborations with a goal to engineer a better HPV vaccine. HPV also contains a minor capsid protein (L2). The L1 sequence is not highly conserved across all types of HPV. L2, on the other hand, is highly conserved. Current (L1 containing) vaccines are roughly 85% effective because they contain a mixture of HPV subtypes. However, since L2 is highly conserved, a vaccine based on L2 could provide close to 100% efficacy. Our goal is to study the structure of a better vaccine candidates (i.e. to try to achieve closer to 100% effectiveness). This is an example of Hi-Risk, Hi-Reward research. We have been looking at a chimera papillomavirus, which contains L2 inserted into L1 to produce pseudocapsids. To date, only low-resolution 3D reconstructions have been achieved due to heterogeneity in the sample. Hopefully, future samples will permit high resolution analysis. Another long-term structural project studies double-stranded RNA containing viruses. We published that infectious bursal disease virus (IBDV), a non-enveloped, bipartite dsRNA virus with a T=13 icosahedral capsid, has a virion assembly strategy that initiates with a precursor particle, the procapsid, which is based on an internal scaffold shell similar to that of tailed dsDNA viruses. The VP3 protein was found to act as a scaffold protein in the empty IBDV capsids. In cryo-electron microscopy, VP3 appears as a 40 -thick internal shell without the icosahedral symmetry of the virion capsid. Analysis of IBDV procapsid mechanical properties indicated a continuous VP3 layer beneath the icosahedral shell, which increased effective capsid thickness. Whereas scaffolding proteins are discharged in tailed dsDNA viruses, VP3 is a multifunctional protein; after capsid maturation, it changes its role and binds to the dsRNA genome, which is organized as ribonucleoprotein complexes. Our results show that IBDV is an amalgam of dsRNA viral ancestors and traits from dsDNA and positive and negative ssRNA viruses. We recently determined the cryo-EM structure of Rosellinia necatrix quadrivirus 1 (RnQV1), a fungal double-stranded (ds)RNA virus. The T=1 RnQV1 capsid is built of P2 and P4 protein heterodimers, each with more than 1000 residues. Both proteins have a similar helical domain, the structural signature shared with the dsRNA virus lineage. Double-stranded RNA research is ongoing. Proteins take multiple, biologically important conformations in solution. While NMR and small angle solution scattering data (SASS) are good probes of solution structures, simultaneously determining multiple structures along with their populations is a difficult problem, one which we have been addressing over the past 15 years where we have developed methodology to determine structures of multiple members of an ensemble simultaneously. The ensemble approach has been applied to the Enzyme I (EI) protein of the Phosphotransferase System. This protein forms the first component in a cascade of transfers of a phosphate group which leads to the transportation of specific sugars across the cell membrane. We had previously characterized the wildtype state of free-EI, where the distance between the phosphate binding site on the C-terminal domain is quite distant (roughly 25 angstrom) from the transfer site on the N-terminal domain. The large-scale protein rearrangements necessary for the phosphoryl transfer was one of the key features we were interested in characterizing. More recently we studied a variant complexed with phosphoenolpyruvate for which the transfer site had been mutated (H189A) so that phosphoryl transfer could not occur. We found that a mixture of two states was required to fit the RDC and SAXS data: a closed state, previously seen in a crystal structure in equilibrium with a novel partially-closed state, the presence of which had been previously hypothesized. Because EI is large compared with proteins typically studied by NMR, and because ensemble calculations have added computational expense, during this project I implemented an enhancement to the SAXS computational features present in Xplor-NIH such that SAXS amplitude contributions from rigid bodies could be computed a single time and then rotated in reciprocal space with a phase contribution as the rigid body rotated and translated in real space relative to other portions of the protein, thus avoiding expensive computation involving all constituent atoms at each molecular dynamics time step. In the course of these projects characterizing ensembles of structures it was noted that it would be advantageous to allow optimization of ensemble member population along with the structure determination. This capability was implemented within Xplor-NIH and, we developed two features which should be present in such calculations such that the populations do not go to zero, or oscillate rapidly: 1) a strict limit on the lower bound of a population and 2) a cost function weighting deviations of population from some constant, initially equal populations. This approach was key in determine the structures of an ensemble of conformations of a different EI mutant (H189Q) which takes a mixture of more open states than the H189A mutant. Another work where this joint structure/ population weight determination protocol was applied was to a mixture of states of the riboswitch RNA studied by mix-and-inject X-ray free electron laser serial crystallography in addition to SAXS. In collaboration with researchers at the University of California, San Diego and the Sanford Burnham Prebys Medical Discovery Institute, we have undertaken the development an application of an implicit solvent model for use in conjunction with experimental data in the structure determination of solution-phase and membrane protein structures. Normally, NMR structure calculations are carried out in the absence of solvent. Including solvent effects in an implicit model has been shown to improve the calculated structures of these types of proteins.
Showing the most recent 10 out of 65 publications
