- Combining structural and functional techniques
our interests are in three disparate areas:
- Viral-host interactions, relevant to the
development of gene therapy vectors.
- Enzyme structural dynamics and mechanism.
- Computational methods for refining
structures.
Structural Biology of Viral - Host Interactions:
Adeno-associated Virus (AAV) is the simplest of the
viruses being developed as a vector for gene therapy, i.e. to transport DNA into
cells in experimental treatments of inherited and other diseases. In viral
infection, the outer protein shell is responsible for receptor attachment in
cell entry and is the target that antibodies recognize in immune neutralization,
processes that have been imaged at atomic resolution for several viruses.
For gene therapy, there is a need to modulate cell entry (to target therapies to
specific cell types) and to evade immune neutralization. Efforts towards
these medical applications will put our fundamental understanding of the
molecular interactions to practical test and forward the basic science.
 Our role continues to be mapping the sites of the
functionally important interactions on the viral capsid protein. In 2002,
we completed the first AAV atomic structure by x-ray crystallography.
There are 60 copies of a subunit building block protein (top right) assembled
into a near-spherical shell (bottom right), three of which comprised the unique
region of the crystal - a technical challenge at ~ 1 million atoms.
Currently, we are imaging virus complexed with fragments of its primary cellular
receptor through electron microscopy. We are similarly mapping the sites
of antibody recognition, complementing imaging with in vitro biological
selection experiments, and structural studies of AAV variants that have evolved
in nature under the selection pressure of immune neutralization. Future
studies will include co-receptors and interactions key to intra-cellular
transport.
Enzyme Dynamics and Mechanism
Students of biochemistry might be forgiven for
understanding that enzymes have a structure and a mechanism of action.
Our studies of arginine kinase are emphasizing that enzymes can have
multiple structures and multiple mechanisms. Arginine kinase is a
member of a family that is core to energy metabolism catalyzing the
transfer of a phosphoryl to and from a phosphagen as it buffers cellular
ATP levels. It has been chosen as an experimentally accessible
paradigm of two-substrate enzymes that are more common, but less well
understood than their single substrate cousins.
Our transition state analog structure at 1.2
Å (part of which is shown) refuted the prevailing
hypothetical mechanism. Subsequent mutational analysis showed that rate is
enhanced by several distinct mechanisms combining to achieve the overall
catalytic effect. Traditional approaches overlook such nuances, so we seek
a more quantitative testing of mechanism by comparing the experimental effects
of mutation to (quantum mechanical) computer modeling of the reaction. The
preliminaries have thrown up many worthy challenges, including a new
understanding of the chemical basis of biology's "high energy bonds", so our
goal of quantitatively partitioning the different catalytic effects remains
distant.
 Structures of substrate-bound and -free
enzyme show large induced-fit conformational changes (right). Preliminary
indications are that enzyme turnover is limited by a conformational change, the
dynamics of which we are elucidating by nuclear magnetic resonance (NMR)
dispersion and residual dipolar coupling techniques. Our goal is an
unprecedented 4-dimensional picture of 3-D structure and the time constants of
motion as the enzyme steps through the catalytic cycle.
Computational Methods of Refinement
The structures of large biomolecular assemblies are
increasingly studied through electron microscopy (EM) of the whole
complex, combined with detailed crystallography of components. By
fitting the components into the EM maps, it is possible to achieve a
structural resolution that is intermediate between the two techniques,
allowing more detailed interpretation than from the electron microscopy
alone. Using algorithms that we originally derived for
crystallography, we are developing methods to optimize the fit of models
into EM maps, maximizing precision. We are also developing
stereochemical restraints applicable to low resolution models to ensure
that the models are not only in agreement with the experimental data,
but embody sound principles of molecular interactions.
 Our focus is development of the computer methods. We
work with electron microscopists on their projects to test the impact of our
work. Particularly exciting has been our collaboration with Joachim
Frank's group in determining the conformational states of the ribosome during
the peptide elongation cycle.
Collaborators
Much of the research is interdisciplinary. Group
members are often co-mentored by co-PIs/colleagues experts in other techniques:
|