The host protein TRIM5á inhibits retroviral infection at an early post-penetration stage by binding to the incoming viral capsid. Interaction with TRIM5á results in a failure of the virus to complete reverse transcription through an unknown mechanism. I have discovered that infection of cells with a restricted retrovirus leads to degradation of TRIM5á in the target cells. This degradation strongly correlates with restriction, is dose dependent, and inversely correlates with infection. Viral-induced degradation of TRIM5á is dependent on the presence of the RING domain of TRIM5á and a functional cellular proteasome. I have also found evidence linking the UbcH5 family of E2 ubiquitin conjugases to a role in post-entry restriction of HIV-1 by rhesus macaque-derived TRIM5á. These findings establish a connection between post-entry restriction by TRIM5á and the ubiquitin-proteasome pathway and indicate a mechanism of retroviral restriction by TRIM5á.

The cathepsin family of endosomal proteases is required for proteolytic processing of several viruses during entry into host cells. Mammalian reoviruses utilize cathepsins B, L, and S for disassembly of the virus outer capsid and activation of the membrane-penetration machinery. Studies using mice deficient for cathepsin B, L, or S revealed that these enzymes influence reovirus pathogenesis. Furthermore, pharmacologic modulation of cathepsin activity attenuates reovirus disease. These findings have broad implications for the role of cathepsin proteases in viral infections and suggest that cathepsin inhibitors should be developed as therapeutics.

Colonization of the human stomach with Helicobacter pylori leads to gastric inflammation and is associated with an increased risk for development of peptic ulceration and gastric cancer. An important H. pylori virulence determinant is a toxin known as VacA. VacA is secreted by an autotransporter pathway, and the secreted VacA protein is comprised of two domains, designated p33 and p55. We have identified specific amino acids within the p55 domain that are essential for assembly of VacA into functional oligomeric complexes, and we show that a VacA mutant protein lacking these amino acids can inhibit the activity of wild-type VacA in a dominant-negative manner. In addition, we show that within a â-helical region of the p55 domain, there are regions of plasticity that tolerate large deletions without detrimental effects on protein secretion or activity, as well as a region that is required for proper folding and secretion of the toxin. These results broaden our understanding of VacA structure-function relationships, and also are relevant to understanding the role of â-helical folds in other bacterial proteins that are secreted by an autotransporter pathway.

Staphylococcus aureus is an important human pathogen and its ability to cause disease is absolutely dependent on iron acquisition from the host. S. aureus obtains iron during infection from vertebrate hemoglobin via the iron-regulated surface determinant (Isd) system. The cytoplasmic components of this system, IsdG and IsdI, are paralagous heme oxygenases which degrade heme to release nutrient iron. I have found that although IsdG and IsdI share 64% amino acid sequence identity, they are differentially regulated depending on the microenvironment experienced by the bacterium. This may represent a strategy by which S. aureus fine-tunes the expression of heme oxygenase activity during infection, as bacteria lacking isdG or isdI exhibit differential virulence defects. In addition, these studies have identified a post-transcriptional regulatory mechanism in which IsdG is specifically stabilized in the presence of the substrate heme. Finally, we have determined the structure of the IsdG- and IsdI-catalyzed heme degradation products and shown that this family of heme oxygenases degrades heme to a novel chromophore which we have named staphylobilin. My studies have significantly advanced our understanding of heme catabolism in S. aureus, a process that I have shown is required for staphylococcal pathogenesis. Moreover, the unique structure and mechanism of the IsdG-family of heme oxygenases indicates that these enzymes may be therapeutic targets.

APOBEC3G is a cytidine deaminase that inhibits HIV replication at a post-entry step in replication. APOBEC3G must be incorporated into the virus particle during assembly in order to inhibit HIV during the next round of replication. We defined the highly basic region between the two zinc binding domains in the nucleocapsid subunit of the viral Gag protein as a crucial packaging domain for APOBEC3G. Using a novel fluorescent resonance energy transfer (FRET) assay to detect APOBEC multimers, we found that APOBEC3G is packaged as a multimer bound to packaged RNA. We demonstrate for the first time that APOBEC multimers complexed with RNA are specifically recruited to the plasma membrane site of assembly by Gag.

Helicobacter pylori is a gram-negative bacterium that contributes to the pathogenesis of peptic ulcer disease and gastric cancer. A key virulence factor for the organism is vacuolating cytotoxin A (VacA), a pore-forming toxin that causes multiple alterations in human cells. The toxin is secreted by an autotransporter pathway as an 88 kDa protein which can be divided into two structural domains; a 33 kDa domain (p33) responsible for pore formation and a genetically diverse 55 kDa domain (p55) associated with receptor binding. We hypothesize that structural, functional and phylogenetic analyses will allow us to assess the mechanism by which VacA interacts with host cells. Using X-ray crystallography, we found that the VacA m1 p55 domain adopts a β-helix fold, a feature that is characteristic of autotransporter passenger domains but unique among known bacterial protein toxins. Notable features of VacA p55 include disruptions in β-sheet contacts that result in five β-helix sub-domains and a novel C-terminal domain. The elongated β-helical structure of VacA presents a unique opportunity for studying the pressures of positive and negative selection in a structural context. We have demonstrated that amino acid substitutions in the VacA p55 domain are under strong, diversifying selection. Docking the p55 structure into a 19 Å cryo-EM map of a VacA dodecamer allows us to propose a model for how VacA monomers assemble into oligomeric structures capable of membrane channel formation. We have recently been able to reconstitute this activity from a mixture of p55 and p33 added together in trans. This advance provides a tractable avenue for future studies aimed at obtaining high resolution structural information for VacA p33.

This dissertation research was focused on elucidating mechanisms at the reovirus-host interface that allow successful completion of the initial replicative steps and lead to viral protein synthesis. During the course of the study, I discovered new functional domains in the reovirus attachment protein σ1. Sigma 1 is an elongated trimer with head-and-tail morphology that engages cell-surface carbohydrate and junctional adhesion molecule-A (JAM-A). This protein is comprised of three domains partitioned by two flexible linkers termed inter-domain regions (IDRs). To determine the importance of σ1 length and flexibility at different stages of reovirus infection, I generated viruses with mutant σ1 molecules of altered length and flexibility and tested these viruses for the capacity to bind the cell surface, internalize, uncoat, induce protein synthesis, assemble, and replicate. I reduced the length of the α-helical σ1 tail to engineer mutants L1 and L2 and deleted midpoint and head-proximal σ1 IDRs to generate ∆IDR1 and ∆IDR2, respectively. Decreasing length or flexibility of σ1 resulted in delayed reovirus infection and reduced viral titers. L1, L2, and ∆IDR1 but not ∆IDR2 displayed reduced cell attachment, but altering σ1 length or flexibility did not diminish the efficiency of virion internalization. Replication of ∆IDR2 was hindered at a post-disassembly step. Differences between wild-type and σ1 mutant viruses were not attributable to alterations in σ1 folding, as determined by experiments assessing engagement of cell-surface carbohydrate and JAM-A by the L and IDR viruses. However, ∆IDR1 harbored substantially less σ1 on the outer capsid. Taken together, these findings suggest that σ1 length is required for reovirus binding to cells. In contrast, IDR1 is required for stable σ1 encapsidation, and IDR2 is required for a post-uncoating replication step. Thus, I discovered that σ1 functions are not restricted to mediating reovirus attachment.

The pathogenesis of Clostridium difficile is dependent on two large homologous toxins, TcdA and TcdB. These toxins contain glucosyltransferase domains that inactivate host Rho proteins by glucosylation. Delivery of the glucosyltransferase domain into the target cell is achieved by binding to the target cell, pH-dependent translocation of the glucosyltransferase domain across the membrane, and release of the domain by autoproteolysis. These steps in delivery are mediated by receptor-binding, pore-forming, and autoprotease domains. Structures existed for a portion of the TcdA receptor-binding domain and the TcdB glucosyltransferase domain, but there had previously been no structural information for the protease and pore-forming domains or for the holotoxins. I have visualized the TcdA and TcdB holotoxins by negative stain electron microscopy, determined the TcdA three-dimensional structure by random conical tilt, and mapped the TcdA functional domains within this structure. A second structure, obtained at acidic pH, revealed significant structural changes involved in delivery of the glucosyltransferase domain across the membrane. In addition, I have elucidated the structure of the TcdA autoprotease domain and used complementary functional assays to uncover key aspects of its function. Finally, I have determined the structure of the TcdA glucosyltransferase domain. The activity of the TcdA glucosyltransferase domain was compared with that of TcdB and shown to be a potent enzyme with a broader range of substrates than TcdB. These data provide a structural framework for understanding the molecular mechanisms by which TcdA and TcdB deliver their cytotoxic cargo into the target cell and disrupt host processes.

Several viruses are known to induce stress granules (SGs), however, a role for these structures during viral infection remains unknown. We first showed that respiratory syncytial virus (RSV) induces SGs starting approximately 12 hours after infection in cultured cells. We then generated a cell line with limited capacity to form stress granules by knocking down expression of the SG assembly protein, G3BP. When G3BP knockdown cells were infected with RSV, we observed a significant decrease in viral titer in comparison to control cells. Since SGs are known to contain RNAs, we performed extensive studies with sensitive novel viral RNA probes in live cells to determine the location of production of viral RNA in infected cells. We showed that at the earliest time points in viral replication, viral RNAs appear in discrete granules containing viral proteins (termed inclusion bodies) that were distinct from SGs. These data suggested that inclusion bodies and not SGs were the primary site of replication of viral RNA. We then determined the mechanism by which RSV induces SGs. We showed that RSV infection activates PKR, which subsequently phosphorylates eIF2α and results in SG formation. In addition, when PKR expression was inhibited, we observed a decrease in viral-mediated SG formation, although we found no change in viral titer. In contrast, a PKR inhibitor 2-AP that also may affect other cellular kinases potently inhibited RSV replication. These data indicate that RSV specifically induces a stress response including activation of PKR and eIF2α with formation of host cell SGs, which appear to play an enhancing role in viral replication.

Uncoating of the viral core following penetration into the target cell represents a fundamentally obscure step in the HIV-1 life cycle. Our laboratory has previously reported that mutations in the CA protein that positively and negatively alter the stability of the viral capsid shell result in impaired HIV-1 infectivity, suggesting that the HIV-1 core is optimally balanced for proper uncoating in target cells. To identify additional determinants contributing to optimal HIV-1 capsid stability in CA, I identified second-site suppressors of HIV-1 uncoating mutants by serial passage. My results indicate that CA determinants outside the CypA-binding loop can modulate the dependence of HIV-1 infection on CypA. Further, based on the recently reported full-length HIV-1 hexamer structure, my data provide novel genetic evidence of HIV-1 CA direct intersubunit interactions. My results also suggest that cellular factors are involved in HIV-1 uncoating. A detailed understanding of the mechanism and determinants of HIV-1 uncoating should facilitate the development of novel antiviral therapies targeting this key step in HIV-1 infection.