α-Factor receptors from Saccharomyces cerevisiae are G-protein-coupled receptors containing seven transmembrane segments. The ability of α-factor receptors to form oligomeric complexes with each other and with other proteins was investigated. Both in vivo and in vitroevidence was obtained that suggests homo-oligomerization of receptors in the plasma membrane. When the membranes from cells coexpressing two differentially-tagged receptors were solubilized with detergent and subjected to immunoprecipitation, the antibodies specific for either epitope tag resulted in precipitation of both tagged species. Treatment of cultures with α-factor had little effect on the extent of oligomerization as judged by the sedimentation behavior of the receptor complexes and by the efficiency of coimmunoprecipitation. The ability of receptor complexes to undergo ligand-mediated endocytosis was evaluated by using membrane fractionation and fluorescence microscopy. Mutant receptors that fail to bind α-factor (Ste2-S184R) or lack the endocytosis signal (Ste2-T326) became competent for ligand-mediated endocytosis when they were expressed in cells containing wild-type receptors. Coimmunoprecipitation experiments indicated that the C-terminal cytoplasmic domain and intermolecular disulfide bonds were unnecessary for oligomer formation. Therefore, α-factor receptors form homo-oligomers and that these complexes are subject to ligand-mediated endocytosis. A crosslinking and immunoprecipitation strategy was used to capture and characterize the transient complexes that contain the α-factor receptor Ste2. Tagged receptors were crosslinked to form at least three high molecular weight complexes and the complexes were immunoprecipitated with antibodies against the tag. Western blotting analysis of the precipitated material revealed the presence of β and γ subunits of the heterotrimeric G protein, Ste4 and Stel8. Similar results were obtained when the cultures had been treated with α-factor prior to analysis. A truncated receptor missing most of the cytoplasmic C-terminal tail was also active in binding Ste4. Overall, these results constitute the first biochemical evidence for a physical association between the α-factor receptor and its cognate G-protein. Endocytic signals in the C-terminal tail (residues 297-431) of the α-factor receptor were analyzed. One signaling element, SINNDAKSS, (residues 331-339) is known to be sufficient (but not necessary) for endocytosis. Internal deletions of the STE2 gene were constructed that remove sequences encoding SINNDAKSS and selected regions of the C-terminal tail. Strains containing these alelles were then assayed for endocytosis in the presence and absence of α-factor. Residues from 360 to 431 were sufficient to mediate both constitutive and ligand-mediated endocytosis of the receptor even though 63 residues including the SINNDAKSS motif had been removed. Structural features of this region that were investigated further were the highly-ubiquitinated Lys374, the neighboring Lys387, and the GPFAD motif (residues 392-396). Lys374 and Lys387 were unnecessary for the element to promote exit from the plasma membrane; however, Lys374 may play some role in intracellular trafficking. The GPFAD motif was not sufficient to promote endocytosis, since the residues 360-399 provided no detectable endocytic activity. Overall, these results suggest that a new region in the C-terminal of the α-factor receptor, redundant with the SINNDAKSS motif, is sufficient to mediate the constitutive endocytosis as well as the ligand-mediated endocytosis of the receptor.

The fusion protein of the Newcastle Disease Virus (NDV) contains three hydrophobic domains. To explore the topogenic signals of these domains, mutants were constructed in which each of the hydrophobic domains was deleted. The membrane insertion and topology of these proteins was characterized in a wheat germ cell-free translation system supplemented with canine microsomal membranes. The results indicated that the first 13 amino acids of the fusion protein are necessary to confer translation inhibition by SRP. Translocation of the nascent chains containing all or part of the first hydrophobic sequence resulted in the appearance of a species of higher molecular weight consistent with glycosylation of at least four of the five potential N-linked glycosylation sites. When glycosylation was inhibited with a glycosylation competitor peptide, signal sequence cleavage was detected. Protease digestion of mutants missing the C-terminal hydrophobic domain indicated that the C-terminus has stop transfer activity. A comparison of membrane insertion of the wild-type fusion protein to that of a mutant missing the second hydrophobic domain, the fusion sequence, indicated that the fusion domain has stop-transfer activity when synthesized in vitro. Furthermore, the fusion domain shows little signal sequence activity when positioned near the amino terminus of the fusion protein. The fusion protein has a highly conserved leucine zipper motif immediately upstream from the transmembrane domain of the F1 subunit. In order to determine the role that the conserved leucines have for the oligomeric structure and biological activity of the NDV fusion protein, the heptadic leucines at positions 481,488, and 495 were changed individually and in combination to an alanine residue. Whereas single amino acid changes had little effect on fusion, substitution of two or three leucine residues abolished the fusogenic activity of the protein although cell surface expression of the mutants and sedimentation in sucrose gradients was similar to that of the wild type. Furthermore, deletion of the C-terminal 91 amino acids, including the leucine zipper motif and transmembrane domain resulted in secretion of an oligomeric structure. These results indicate that the conserved leucines do not play a role in oligomer formation but are required for the fusogenic ability of the protein. When the polar face of the potential alpha helix was altered by nonconservative substitutions of a serine-to-alanine (position 473), glutamic acid-to-lysine (position 482) or an asparagine-to-lysine (position 485), the fusogenic ability of the protein was not significantly disrupted. A phenylalanine residue is at the amino terminus of the F1 protein of all paramyxovirus fusion proteins with the exception of the avirulent strains which have a leucine residue in this position. To explore the role of this phenylalanine in the fusion activity of the protein, this residue was changed to leucine (F117L) or to glycine (F117G) by site-specific mutagenesis while maintaining the cleavage site sequence of virulent strains of NDV. Whereas both the wild-type and the F117G proteins were proteolytically cleaved and F1 was detected, the leucine subsitution abolished cleavage. When co-expressed with the HN protein, the fusion protein with either a phenylalanine and glycine residue at position 117, but not a leucine, was shown to stimulate membrane fusion. However, incubation in trypsin activated the fusion activity of the F117L protein. Thus the presence of a leucine at position 117 of the precursor sequence blocks cleavage, but not fusion acitivity, and indicated that the phenylalanine at the amino terminus of the F1 subunit is not conserved for the fusion activity of the protein.

This thesis examines the structure and regulated expression of the gene encoding the neuroendocrine peptides neurotensin and neuromedin N (NT/N gene). Previous studies have shown that expression of NT/N mRNA and NT peptide in PC12 cells are strictly dependent on simultaneous exposure to combinations of nerve growth factor (NGF), glucocorticoids, activators of adenylate cyclase, and lithium ion. My objective was to characterize the cis-regulatory DNA sequences involved in regulated expression of this gene. The initial focus of this study was an analysis of the structure, tissue-specific expression, and exon evolution of the rat NT/N gene. Nucleotide sequence comparisons between the rat gene and the canine and bovine cDNA sequences indicated that the predicted structure of a 170 amino acid precursor protein is highly conserved. Furthermore, the close similarity between the two cDNAs suggested that identical precursor proteins are expressed in neural and endocrine tissues. RNA analysis revealed that the gene is transcribed to yield two distinct mRNAs, 1.0 kb and 1.5 kb in size. The two mRNA species differ only in the size of their 3' untranslated regions. Interestingly, the smaller mRNA is predominant in the gastrointestinal tract, while both mRNAs are equally abundant in all neural tissues examined, except the cerebellum, where no expression was observed. Transient transfection assays were used to delineate the rat NT/N gene cis-regulatory DNA sequences. Progressive deletion of the NT/N 5' flanking region revealed that sequences between -216 and +56 of the NT/N gene are sufficient to confer the full spectrum of responses of the endogenous gene to either of two reporter genes. A detailed mutational analysis of the NT/N control region indicated that it is composed of an array of inducible cis-regulatory elements, including an AP-1 site, two cAMP-responsive elements (CREs), and a glucocorticoid-responsive element (GRE). Specific mutations to the AP-1 site and either CRE suggested that these elements are functionally interdependent. I propose that this array of cis-regulatory sequences in the NT/N transcriptional control region serves to integrate multiple environmental stimuli into a unified transcriptional response. To further examine the role of the AP-1 site and CREs in the NT/N promoter, reporter genes containing either a single or multiple AP-1 or CRE sites were expressed in PC12 cells and protein kinase A-deficient PC12 cells treated with forskolin, NGF, and lithium, either individually, or in combination. The results indicated that lithium and NGF markedly activate promoters containing multiple AP-1 sites, but not a single site, and that these effects were additive. Both agents potentiated forskolin-induced activation of promoters containing a single or multiple CREs, but had no effect, individually. Also, in contrast to the activation of multiple AP-1 sites by lithium and NGF, activation of the NT/N promoter and promoters containing CREs is absolutely dependent on protein kinase A activity. These results suggested that promoters containing multiple AP-1 sites, or a single AP-1 site in the context of nearby active CREs, are selectively activated by lithium and NGF in PC12 cells. Based on the results of this thesis I have proposed a model to account for the complex transcriptional regulation of the NT/N gene in PC12 cells. I have also addressed the relevance of these findings to the mechanisms of phenotypic plasticity of embryonic neural crest cells, NGF-induced neuronal differentiation, and the pharmacological actions of lithium.

The overall goal of my thesis research was to understand better the mechanisms that control antigen processing and presentation by class II MHC molecules. Towards this goal I investigated ways in which the physical structure and post-translational modifications of the class II MHC alpha and beta chains and associated molecules might serve to regulate antigen processing and presentation. Specifically, I investigated (1) a hypothesis that Ii might aid binding of foreign antigenic peptides to the class II MHC foreign antigen binding site (desetope), and the application of this hypothesis to the prediction of class II-presented peptides; (2) the proteolytic cleavage of Ii to p25; (3) the proteolytic cleavage of the class II MHC alpha and beta chains, and (4) the phosphorylation of Iiand the alpha and beta chains. In exploring the hypothesis that amphipathic alpha helical peptides digested from foreign antigen, bind to the class II MHC desetope, to be presented to T cell receptors, we found such an extended, amphipathic helix in Ii (Phe146-Val164). A hypothesis was developed that this amphipathic alpha helix of Ii bound to the desetope of class II MHC molecules, and remained there from time of synthesis until catalyzing the charging of the desetope with a foreign peptide. This region of Iicould then be considered to be the prototypic T cell-presented peptide and the "strip-of-helix" algorithm was developed to search the sequences of proteins for similar amphipathic alpha helices. Such peptides might bind to the class II MHC desetope and have a high probability to be presented to the T cell. The strip-of-helix algorithm calculated the mean hydrophobicity (from Kyte-Doolittle values; Kyte and Doolittle, 1982) of sets of amino acids in axial strips down sides of helices for 3 to 6 turns, at positions n, n+4, n+7, n+11, n+14, and n+18. Peptides correlating well with T cell responsiveness had: (1) 12 to 19 amino acids (4-6 turns of an alpha helix), (2) a strip with highly hydrophobic residues, (3) adjacent, moderately hydrophilic strips, and (4) no prolines to break the helix. This algorithm predicted 10 of 12 T cell-presented peptides in 7 well-studied proteins. In a study of the post-translational modifications of Ii, an early proteolytic pathway of the destruction of Ii, resulting in the generation of p25, was described. This 25,000 dalton protein, seen in immunoprecipitates with antibodies to class II MHC molecules or to Ii, was shown to be a C-termina1 fragment of a high mannose form of Ii. The evidence for this conclusion includes the following results. [35S]methionine-1abe1ed Ii and associated molecules were immunoprecipitated, denatured, resolubi1ized and subjected to a second immunoprecipitation with various antibodies. Two antisera to C-termina1 peptides of Ii (183-193 and 192-211), but not an antiserum to an N-termina1 peptide (12-28), immunoprecipitated p25. A monoclonal antibody (mAb) to Ii immunoprecipitated [35S]methionine-1abe1ed p25 but not [35S]cysteine-1abe1ed p25, consistent with the loss of a portion of Ii containing the only cysteine in Ii, Cys28. [35S]methionine pulse-chase labeling demonstrated the maximal appearance of p25 at 20-40 min chase times. p25 molecules were reduced to about 10.5 kD by treatment with endoglycosidases F and H. p25 was, therefore, generated from a high mannose form of Ii in the ER or cis-Golgi. This finding could either implicate that site for class II MHC desetope charging with foreign peptides or reflect a mechanism for degradation of "excess" Ii molecules in the ER. Digestion of class II MHC antigen-Ii complexes with various proteases yielded fragments, migrating at and near p25 in 2-D electrophoretic gels, which were relatively resistant to further digestion. This observation was consistent with the presence of relatively protease-resistant secondary structures (domains) and a relatively protease-sensitive (IgG hinge-like) region in Iinear its insertion into the membrane. In a study of the post-translational modifications of the class II MHC alpha and beta chains, well conserved pairs of basic amino acids in the sequences of these molecules were observed. It was hypothesized these could be sites for proteolytic cleavage, as precedented in other systems (i.e.proinsulin processing). These potential cleavage sites fall in significant locations with respect to the deduced structure of the class II MHC desetope, supporting the hypothesis that these cleavages might either aid or destroy antigen presenting functions. To test this hypothesis we looked for remnant polypeptides of the alpha and beta chains. Polypeptides were observed in gels of immunoprecipitated class II MHC complexes. To identify if such polypeptides were derived from the alpha and beta chains, immunoblotting to electrotransferred polypeptides was attempted, with antisera made to synthesized peptides that mimicked eight regions of the alpha and beta chains. These antisera were produced and characterized by dot blotting, ELISA, western blotting, and immunoprecipitation of native and denatured material. One antiserum, to an alpha chain peptide (77-88), blotted to a polypeptide immunoprecipitated by anti-class II MHC antiserum. This observation supported the hypothesis that the alpha and beta chains undergo proteolytic cleavages, possibly in the control of antigen presentation. It was also demonstrated that Ii and the alpha and beta chains can be phosphorylated under varying culture conditions, but this project was not pursued.

Hebert and Carruthers (1992) showed that the human erythrocyte glucose transporter is an allosteric complex of four GLUT1 proteins whose structure and substrate binding properties are stabilized by reductant-sensitive noncovalent subunit interactions. The GLUT1 tetramer dissociates into dimers upon exposure to reductant but subunits are not associated via disulfide bridges. Each subunit of SDS-denatured tetrameric GLUT1 exposes only two thiols while reduced denatured GLUT1 exposes all six sulfhydryl groups. They hypothesized that glucose transporter oligomeric structure and cooperative catalytic function resulted from noncovalent subunit interactions promoted or stabilized by intramolecular disulfide bridges. These interactions give rise to an antiparallel arrangement of substrate binding sites within the transporter complex. In the present studies, we tested aspects of this model. Specifically, we wanted 1) to understand why the native, noncovalent, homotetrameric GLUT1 complex is sensitive to reductant, 2) to determine whether the tetramer is more catalytically efficient than the dimer in situ, and 3) to test the hypothesis that it is the antiparallel arrangement of substrate binding sites between subunits that provides the transporter with its catalytic advantage. We used biochemical and molecular biological approaches to isolate specific determinants of transporter oligomeric structure and/or transport function in purified isolated transporter preparations, in intact red cells and in CHO cells. We have also examined the hypothesis that net sugar transport in the human erythrocyte is rate limited by reduced cytosolic diffusion of sugars and/or by reversible sugar association with intracellular macromolecules. Our findings support the hypothesis that each subunit of the parental glucose transporter contains a single intramolecular disulfide bridge located between cysteine residues 347 and 421. This disulfide seems to be necessary for GLUT1 tetramerization. Our findings suggest that GLUT1 N-terminal residues 1 through 199 provide contact surfaces for subunit dimerization but are insufficient for subunit tetramerization. Our studies also show that in situ disulfide disruption by cell impermeant reductants results in the loss of cooperative subunit interactions and a 3 to 15-fold reduction in the transport efficiency of the transporter. We further find that in situ GLUT1 is susceptible to exofacial proteolysis. Exofacial trypsin cleavage eliminates cooperativity between subunits but does not affect transporter oligomeric structure or transport activity. Thus catalytic efficiency does not derive directly from cooperative interactions between substrate binding sites on adjacent subunits. We have confirmed that 30MG transport in human erythrocytes is a diffusion limited process. We find that steady-state sugar uptake in red cells and K562 cells measures two processes - sugar translocation and intracellular sugar binding. We propose a model for native GLUT1 structure and function.

The structure-function relationship explains how the human erythrocyte glucose transport protein (GLUT1) catalyzes sugar transport across the plasma membrane. This work investigates the glucose transport mechanism, the structural arrangement and dynamics of GLUT1 membrane-spanning α-helices, the molecular basis for glucose transport regulation by ATP, and how cysteine accessibility contributes to GLUT1 structure. A rapid kinetics approach was applied to examine the conformational changes GLUT1 undergoes during the transport cycle. To transition from a global to molecular focus, a novel mass spectrometry technique was developed to resolve GLUT1 sequence that is associated either with membrane embedded GLUT1 subdomains or with water exposed domains. By studying accessibility changes of specific amino acids to covalent modification by a Sulfo-NHS-LC-Biotin probe, specific protein regions associated with glucose transport modulation by ATP were identified. Finally, mass spectrometry was applied to examine cysteine residue accessibility under native and reducing conditions. This thesis presents data supporting the isolation of an intermediate, occluded GLUT1 conformational state that temporally bridges import and export configurations during glucose translocation. Our results confirm that amphipathic α-helices line the translocation pathway and promote interactions with the aqueous environment and substrate. In addition, we show that GLUT1 is conformationally dynamic, undergoes reorganization in the cytoplasmic region in response to ATP modulation, and that GLUT1 contains differentially exposed cysteine residues that affect its folding.

Enteropathogenic Escherichia coli (EPEC) is one of many bacterial and viral pathogens that can exploit the eukaryotic actin cytoskeleton for its own purposes. EPEC injects its own receptor, Tir, into the host cell plasma membrane where, upon binding the bacterial adhesin, intimin, can trigger actin assembly beneath bound bacteria resulting in characteristic actin "pedestals". The formation of these lesions is thought to be critical for bacterial colonization; and can also provide insight into actin dynamics of mammalian cells. EPEC Tir stimulates multiple signaling pathways converging on a central actin nucleation promoting factor, N-WASP. The best-characterized pathway of actin pedestal formation also involves the eukaroytic adaptor protein, Nck, but at least two Nck-independent signaling cascades have also been identified. Multiple aspects of Tir-mediated signaling cascades remain unclear. For example, although Nck can directly bind and activate N-WASP, current models of Tir-mediated, Nck-dependent actin signaling postulate an indirect interaction between Nck and N-WASP mediated by one or more unidentified host factors. Additionally, the relationship of this pathway to the Nck-independent pathways is unknown. To better understand Tir-mediated actin assembly, a detailed and quantitative analysis of the domain requirements of Nck and N-WASP for pedestal formation was conducted. The results indicate that, contrary to previously favored models, Nck is unlikely to require additional host factors to bind N-WASP during pedestal formation, but apparently directly stimulates this nucleation promoting factor. In addition, the results show that the Nck-dependent and -independent pathways target distinct regulatory domains of N-WASP.

Modulating chromatin structure is an important step in maintaining control over the eukaryotic genome. SWI/SNF, one of the complexes belonging to the growing family of ATP-dependent chromatin remodeling enzymes, is involved in controlling the expression of a number of inducible genes whose proper regulation is vital for metabolism and progression through mitosis. The mechanism by which SWI/SNF modulates chromatin structure at the nucleosome level is an important aspect of this regulation. The work in this dissertation focuses on how the Saccharomyces cerevisiae SWI/SNF complex uses the energy of ATP-hydrolysis to alter DNA-histone contacts in nucleosomes. This has been approached in a two part fashion. First, the three-dimensional structure and subunit composition of SWI/SNF complex has been determined. From this study we have identified a potential region of the SWI/SNF complex that might [be] a site for nucleosomal interaction. Second, functional analysis of the ATPase domain of Swi2p, the catalytic subunit of SWI/SNF, has revealed that a specific conserved motif is involved in coupling ATP hydrolysis to the mechanism of chromatin remodeling. These results provide a potential model for the function of the SWI/SNF chromatin remodeling complex at the nucleosome level.

Definitive evidence that natural killer (NK) cells mediate an antiviral effect in vivo was obtained using murine cytomegalovirus (MCMV) as a model system. Adoptive transfer studies using a variety of physical and immunochemical techniques to enrich and deplete NK cell activity showed that the cell population capable of mediating resistance (as assayed by enhanced survival and reduction in spleen virus titers) had the phenotype of an NK cell: a nylon wool nonadherent, asialo GM1+, NK 1.2+, ly 5+, Thy-1-, Ia-, low-density lymphocyte. Adoptive transfer of IL-2-dependent cloned NK cells (but not T cells) also provided resistance. NK cells did not provide resistance to lymphocytic choriomeningitis virus (LCMV). Selective depletion of NK cell activity by injection of mice with antibody to anti-asialo GM1 lowered resistance to MCMV, mouse hepatitis virus, and vaccinia virus but not to LCMV. NK cell depletion resulted in up to 1000-fold increases in spleen and liver virus titers, correlating with more severe pathology in these organs. NK cells were found to have antiviral effects early (0-3 days) but not late (6-9 days) postinfection. NK cell depletion resulted in markedly increased MCMV-induced suppression of T cell function, which is probably responsible for the delayed clearance of virus seen in these mice. NK cell depletion resulted in increased virus synthesis during persistent MCMV infection, but had no effect on the course of persistent LCMV infection, despite elevated NK cell and interferon (IFN) levels found in these LCMV-infected mice. The reason why NK cells play a role against MCMV but not LCMV infection was not due to differences in NK cells induced by these 2 viruses, but more likely due to target cell susceptibility. IFN pretreatment of MCMV-infected cells failed to protect them against NK cell-mediated lysis, whereas uninfected and LCMV-infected cells were almost totally protected. These IFN-pretreated, LCMV-infected cells were not resistant to cell-mediated lysis in general, as this treatment increased their sensitivity to virus-specific T cell-mediated lysis by 2- to 3-fold. This enhanced sensitivity to lysis correlated with increased surface expression of H-2 antigens, but not viral antigens. In summary, these studies provide compelling evidence that NK cells can mediate antiviral effects in vivo, and provide some insights into their mode of action and consequences of their disfunction.

The KIL gene of vaccinia virus encodes for a host range protein; in the absence of which, the virus is unable to grow in certain cell lines (RK-13 and some human cell lines). KIL function can be complemented in RK-13 cells by the cowpox host range gene product CP77 despite a lack of homology between the two proteins except for ankyrin repeats. We investigated the role of ankyrin repeats ofthe K1L gene in the host-range restriction of growth in RK-13 cells. The growth of a recombinant vaccinia virus, with the K1L gene mutated in the most conserved ankyrin repeat, was severely impaired as evidenced by lack of plaque formation and reduction in viral titers. Infection of RK-I3 cells with the mutant recombinant vaccinia virus resulted in total shutdown of both cellular and viral protein synthesis early in infection, indicating that the host restriction mediated by the ankyrin repeat is due to a translational block. A comparison of the cellular localization of the K1L wild type and mutated forms showed no difference, as both localized exclusively in the cytoplasm of RK-I3 cells. We also investigated the interaction of the vaccinia virus K1L protein with cellular proteins in RK-13 cells and co-immunoprecipitated a 90 kDa protein identified as the rabbit homologue of human ACAP2, a GTPase-activating protein with ankyrin repeats. Our result suggests the importance of ankyrin repeat for host-range function of K1L in RK-13 cells and identifies ACAP2 as a cellular protein which may be interacting with K1L.