| Virus capsids & host interactions with antibodies & cellular receptors. Our main focus is Adenoassociated Virus. This is basic research that we hope to put to practical tests in the design of improved viral vectors for in vivo human gene therapy. Details... |
| Enzymes - How different are two-substrate enzymes from better-understood unimolecular paradigms? Arginine kinase is our model for elucidating how protein dynamics and alignment of the substrates contribute to catalytic enhancement and specificity. Details... |
| Methodology: High resolution structures are not always available - here we are developing computer methods to optimize models against crystallographic, electron microscopic and ssNMR data. Details... |
Capsids are roughly spherical protein shells surrounding the DNA or RNA of many viruses and are responsible for transporting it to infect new hosts and cells. Capsids contain many copies of 1 or more proteins arranged with icosahedral symmetry. Our studies of viral assemblies are directed towards visualizing the interactions that target a virus to particular host cells, and that lead to its recognition and neutralization by the host's immune system.
Adenoassociated virus (AAV) is a widely-used vector in experimental gene therapies being developed to cure genetic diseases such as cystic fibrosis, cancers, etc.. In 2002 we published the first atomic structure, the culmination of efforts to culture large quantities of the virus in human cells and to solve one of the largest ever crystal structures at ~ 1 million atoms. AAV has prominent surface protrusions. They are formed by intertwined loops from adjacent subunits that bring together different regions of the primary sequence which had previously been implicated genetically in cell receptor binding. Calculations of the electrostatic potential from the atomic coordinates revealed that one side of each protrusion was positively charged and a candidate site for binding of the heparan sulfate proteoglycan receptor. This general location has since been confirmed by mutagenesis in other laboratories.
We
continue to study AAV-receptor interactions through electron microscopy of
virus-heparan complexes (in collaboration with
Ken Taylor) and crystallography of
virus complexed with small heparan analogs. These should give us
complementary low resolution over-view and detailed visualizations of
virus-receptor interactions. We are also investigating virus-antibody
interactions in three ways: (1) electron microscopic imaging of complexes
with selected monoclonal antibodies; (2) in vitro selection of
neutralizing escape mutant viruses, to map sites of mutation to the structure;
(3) structure determinations of other AAV serotypes (strains), to see how
natural variants have evolved in response to immune challenge.
Together, these basic structure-function studies will allow us and others to pursue a more rationale approach to the engineering of improved viral vectors for gene therapy. We hope that our studies will lay the groundwork for the retargeting of AAV to the tissues of choice and for the delivery of gene therapies without immune interception.
Human rhinovirus 50 (HRV50): We have recently completed a 2.0 Å structure determination of this common cold virus, complexed with a clinically-tested drug. Serotype 50 was chosen as the most typical in response to drugs. Serendipitously, the structure is at higher resolution than hitherto possible. We are beginning analyses of the virus-drug interactions and comparisons to other serotypes.
Arginine kinase is the arthropod homologue of creatine kinase that catalyzes a reaction which maintains the ATP concentration. We use it as a model for understanding how bimolecular (two-substrate) enzyme reactions differ from the predominant unimolecular paradigms in their catalysis and specificity. Our structure was the first of a bimolecular enzyme transition state analog complex, and it is this enzyme's amenability to high resolution crystallographic structure, NMR dynamics and biochemical kinetics that makes it such an appealing system.
Using
structure-assisted mutagenesis and kinetics we have shown that the expected
general base catalysis can account for only part of the catalytic enhancement.
Our refinement at 1.2 Å resolution gave an unprecedented view
of precise substrate pre-alignment in a two-substrate enzyme - within 3º of optimal.
Structures with substrate analogs show that alignment is needed for reactivity.
We are now working towards a quantitative understanding of the various
contributions to the catalytic effect through hybrid quantum mechanical /
molecular mechanical calculations in collaboration with
Jeff Evanseck.
Our structure of the substrate-free enzyme showed that there were large substrate-induced changes in the protein structure, so the precise alignment must be the result of a dynamic process. In collaboration with Jack Skalicky, we have completed an NMR backbone resonance assignment, and are beginning to measure exchange rate constants for the conformational changes of important residues. We are planning to correlate the time-dependent structural transitions with the biochemical kinetics of the enzyme.
Crystallographic model refinement. Our methods
development started with "real-space" refinement of the fit of atomic models
into the experimentally-derived electron density. Generally, this is a
good prelude to a more conventional refinement against diffraction amplitudes,
widening convergence and reducing over-fitting. When the phases are of
excellent quality, as in virus structure determination, real-space methods can
out-perform the conventional methods.
Hybrid electron microscopy (EM) / crystallographic structure refinement. We are now adapting real-space refinement for the fitting of known crystallographic component structures into lower resolution EM images of assemblies. In collaboration with Ken Taylor and Joachim Frank, we are applying these methods to understanding the conformational changes that occur during muscle contraction and protein synthesis at the ribosome. One important result, highlighted in our Cell (2003) paper, is that ribosomal proteins are not just a passive structural scaffold for the RNA, but are the hinge points of conformational dynamics.
Solid-state NMR. In collaboration with Mathematicians Richard Bertram & Jack Quine, and spectroscopist Tim Cross, we have derived objective functions with which to refine atomic models against ssNMR data by analogous methods. This is facilitating the determination of membrane protein structures that are very challenging by other methods.
Stereochemical restraining force fields. Common to our refinement efforts is the need to obtain accurate models, even when high resolution data are not available. To reduce over-fitting, we have been investigating stereochemical restraints additional to those applied in high resolution refinements. To date, we have considered hydrogen bonding and electrostatics, neither of which are usually restrained, because their prior treatments were usually deleterious. By accounting for the partial covalent nature of hydrogen bonds, and by using a Poisson continuum description of the electrostatic potential, we have developed restraints that are beneficial.
By combining these approaches, we plan to support investigations of large assemblies at increasing detail. Currently, elucidation of the conformational changes within domains generally requires crystallographic resolutions, but we plan to make it possible at the sub-atomic resolutions with which large complexes and membrane proteins can be visualized by EM, NMR and other biophysical techniques.
For further information see: Publications; Software distributed
The
research projects offer a variety of opportunities for training in
structure-function for doctoral students in
biochemistry or
molecular biophysics and post-doctoral
fellows. Many group members are jointly mentored in collaborative
projects, offering a broad experience in interdisciplinary science.
Facilities for cell culture and
protein expression, x-ray
diffraction, and physical
biochemistry and computing offer not only equipment, but training by
their managers who hold Ph.D.s in their specialist areas. Highlights
include BL2 facilities with hoods and incubators for cell/virus culture, two
x-ray crystallographic data collection systems, and access to synchrotron
beam-time through our membership in the SERCAT consortium based at the Argonne
National Laboratory. Our research laboratories are located in the Kasha
Laboratory of Biophysics which underwent a complete $7.1M renovation in 2002-3.