The ultimate goal of our studies is to gain a deeper understanding of the molecular basis for important human diseasessuch as sudden death, myocardial infarction, rotavirus infection and HIV infection that cause substantial mortality and suffering. The structural details revealed by our work may provide clues for the design of more effective and safer medicines.

At the basic science level, we are intrigued by biological questions at the interface between cell biology and structural biology. How do membrane channels open and close? How are signals transmitted across a cellular membrane when an extracellular ligand binds to a membrane receptor? How do viruses attach and enter host cells, replicate and assemble infectious particles?

In our laboratory we use high resolution electron cryo-microscopy (cryo-EM) and image processing to explore the molecular design of large, multicomponent supramolecular assemblies. Biological specimens are quick frozen in a physiological state to preserve their native structure and functional properties. A special advantage of this rapid-freezing method is that we can trap and image dynamic states of functioning macromolecular assemblies, such as open and closed states of membrane channels and viruses actively transcribing RNA. Three-dimensional density maps are obtained by digital image processing of the high-resolution electron micrographs. The rich detail in the maps reveals the structural organization of complex biological structures that can be related to the functional properties of such assemblies.

Research projects underway include the structure analysis of:

(1) Membrane proteins involved with cell-to-cell communication (gap junctions), water transport (aquaporins), ionic transport (potassium channels), transmembrane signaling (integrins), and viral recognition (rotavirus NSP4)
(2) RNA viruses responsible for significant human disease (rotavirus, astrovirus, retroviruses)
(3) RNA viruses used as model systems to understand mechanisms of pathogenesis (reovirus, nodaviruses, sobemoviruses, nudaurelia capensis omega Greek symbol virus, rice yellow mottle virus).

Summarized below are selected projects that exemplify the themes and direction of our research program.

Cardiac gap junction membrane channels: Cardiac gap junctions electrically couple adjacent cells, thereby playing a critical role in the normal coordinated depolarization of heart muscle as well as cardiac arrhythmias causing sudden death. Our goal is to determine the structure of these intercellular channels at a level of detail that will allow us to understand the molecular mechanism for channel gating. These channels are formed by protein molecules called connexins. Gap junction channels in the heart are formed by a 43 kDa protein called a1 connexin, also called Cx43. In our previous work, we used site-specific peptide antibody labeling, protease cleavage, CD spectroscopy and electron microscopy of 2D crystals to delineate the membrane topology and quaternary structure of rat heart gap junctions. We have now expressed a C-terminal truncation mutant of a1 connexin and grown 2D crystals of recombinant gap junctions. A projection density map at 7-Å resolution revealed for the first time a ring of a-helices that line the aqueous pore of the channel and a second ring of a-helices in close contact with the membrane lipids. Our 3D structure provided the highest resolution thus far achieved for a gap junction channel. In fact, to our knowledge, this is the first example where a mammalian polytopic membrane protein has been expressed and examined by crystallographic methods. A novel technical accomplishment of this work was that the structure analysis was performed using only microgram amounts of material. The dodecameric channel is formed by the end-to-end docking of two hexamers, each of which displays 24 rods of density in the membrane interior, consistent with an a-helical conformation for the four transmembrane domains of each connexin subunit. The transmembrane a-helical rods merge with a double layer of protein density in the extracellular vestibule, providing a tight seal to exclude exchange of substances with the extracellular milieu. We anticipate that this basic molecular design will be a common folding motif for gap junction channels. High priority projects include the comparison of open and closed channels and crystallization of full-length a1 connexin in order to examine the structure of the C-terminal regulatory domain.


Integrins: Cardiovascular disease is the major cause of mortality in the United States, primarily due to myocardial infarction resulting from coronary atherosclerosis. Myocardial infarction almost always results from formation of a thrombus at the site of a coronary artery stenosis. A key event that stimulates thrombus formation is platelet aggregation, which is mediated by the prototypical integrin aIIbb3. Integrins are a large family of heterodimeric transmembrane receptor proteins, which modulate cell adhesion, such as platelet aggregation, as well as other important biological processes such as development, angiogenesis, wound healing, and neoplastic transformation. Integrins accomplish these diverse functions by mediating dynamic linkages between extracellular adhesion molecules and the intracellular environment. Integrin functions are regulated by transmembrane signaling, which can occur as a consequence of binding extracellular ligands (so-called “outside-in” signaling), as well as the binding of molecules to the cytoplasmic domains (so-called “inside-out” signaling). We used cryo-EM and single particle image reconstruction to derive a three-dimensional structure at 20 Å resolution of the unliganded, low-affinity state of the human platelet integrin aIIbb3. The large ectodomain and small cytoplasmic domains are connected by a rod of density that we interpret as two parallel transmembrane a-helices. The docking of the X-ray structure of the aVb3. ectodomain into the electron cryomicroscopy map of aIIbb3. requires hinge movements at linker regions between domains in the crystal structure. Comparison of the putative high- and low-affinity conformations reveals dramatic conformational changes associated with integrin activation. The structural details revealed by these studies will provide insight into the molecular basis of integrin activation that will be relevant for the rational design of drugs to modulate integrin functions.


Water channels: Aquaporins are channels that play a critical role in water transport across membranes. We previously used electron cryo-crystallography to derive a 3D map of AQP1 at 6-Å resolution. This is the first water channel to be visualized in aqueous buffer within a lipid bilayer. Each monomer is composed of six membrane-spanning, tilted a-helices that form a barrel which encloses a vestibule leading to the water-selective channel. The structure has an in-plane, intramolecular pseudo-twofold axis of symmetry located in the hydrophobic core of the bilayer, which is consistent with the sequence-related N- and C-terminal halves of the protein. This folding pattern represents a new motif for the topology and design of membrane protein channels, and is a simple and elegant solution to the problem of bidirectional water transport across the bilayer. We are now focusing on a-TIP (tonoplast intrinsic protein isolated from the membranes of plant vacuoles) because this aquaporin is gated by phosphorylation. Analysis of tubular crystals shows that the a-helical design of aquaporins is conserved between the plant and animal kingdoms. Our aim is to delineate the conformational changes associated with phosphorylation-dependent gating.


Our goal is to understand the molecular design and assembly pathway of rotavirus, which is the major cause of human infant mortality in developing countries. Using cryo-EM and icosahedral image analysis, we showed for the first time that the structural proteins in rotavirus are organized into three layers: an outer capsid shell formed by 780 VP7 molecules and 60 VP4 hemagglutinin spikes; an inner capsid shell formed by 260, pillar-shaped, VP6 trimers; and a core shell formed primarily by VP2 as well as VP1 and VP3. The molecular design of rotavirus is therefore substantially more complicated than simple icosahedral viruses that have a single capsid shell. Essential steps in infection are trypsin cleavage of VP4, the surface hemagglutinin, and attachment to susceptible cells. Using difference map analysis between native and spikeless particles, we showed that VP4 extends ~110Å from the surface of the virus and has a bi-lobed head. The spike penetrates ~90Å beneath the virion surface and interacts with VP7 and VP6, that form the outer and inner capsid shells. The structural features of VP4 revealed by our analysis have implications for interactions of the virus with cell surface receptors as well as for viral morphogenesis. We have now decorated rotavirus with VP4-specific antibodies that neutralize the virus. Difference maps between native and Fab-decorated rotavirus allow us to map the binding "footprint" of the antibodies on VP4. The maps are of sufficient clarity that a canonical Fab molecular structure can be docked within the EM density in order to infer how the hypervariable loops of the Fab interact with the surface of the hemagglutinin domains. By recording and analyzing images of 1000's of virus particles, we aim to delineate the subtle conformational changes that accompany activation of VP4 by trypsin cleavage. Understanding this process will provide specific insight into the pathogenesis of rotavirus and may provide clues for vaccine and drug design.


Reovirus is in the same family as rotavirus and has served as an important model system for exploring viral assembly and pathogenesis. We previously used cryo-EM and image analysis to examine the 3D structure of intact virions, infectious subvirion particles and core particles. A comparison of the 3D structures and inspection of difference maps provided considerable insight into the conformational changes and structural rearrangements related to virus-cell interactions and viral pathogenesis. Double-stranded RNA was discovered in reovirus, and we are using the core particles as a system for exploring RNA transcription. Cryo-EM and image analysis revealed that transcriptionally active reovirus core particles display a rod-like density centered within the channel formed by the pentameric guanylyltransferase, which we attribute to exiting nascent (+) strand RNA. Comparison of active and inactive particles reveal that transcription is associated with dramatic reorganization of the ten dsRNA genomic segments and small protein conformational changes. The approach we have taken should be of general use for exploring the molecular events associated with dynamic transcription/translation processes.


Retroviruses: Our goal is to understand the assembly pathway of retroviruses and the structural rearrangements that occur with cleavage of the Gag polyprotein, a step which confers infectivity. We recently used cryo-EM and image analysis to examine the native structure of immature, protease-deficient Moloney murine leukemia virus (MuLV) and infectious, wild-type MuLV. A common assumption in the literature has been that the core of retorviruses is assembled with icosahedral symmetry. In fact, our analysis disproved this concept. Instead, the Gag subunits in the immature particle are packed in paracrystalline domains, which become further disordered upon cleavage of the polyprotein. Structural details about the transformations that take place in the formation of infectious MuLV will be important for a complete understanding of mechanisms of retrovirus pathogenesis. To this end, our current aim is to grow 2D crystals of expressed constructs of the matrix, capsid and nucleocapsid proteins.

| Basic Research | Clinical Research | Publications | Gallery of Structures
Group | Contact Us | Links | Members Only | Philanthropy