As critical regulators of vesicular trafficking, Rab proteins comprise the largest GTPase family, with thirty-eight functionally distinct members and another twenty isoforms in the human genome. Activated Rab GTPases interact with effector proteins involved in vesicle formation, transport, tethering, docking and fusion. The specificity of Rab interactions with effectors and regulatory factors plays a central role with respect to the fidelity of membrane trafficking. Rab recognition determinants and the mechanisms underlying interactions with structurally diverse regulatory factors and effectors are complex and poorly understood. Using Rab5 mediated endocytic transport as a model system, the work described in this thesis provides insight into the structural basis underlying the interaction of effectors and regulatory factors with Rab GTPases. In addition, structural and biochemical approaches have been used to define how specific Rab5 interacting proteins function in the endocytic and recycling pathways. These results establish novel structural and functional concepts that can be tested using family wide analyses of Rab GTPase recognition determinants and regulatory roles in the cell.

A clear definition of the mechanisms involved in synaptic transmission is of paramount importance for the understanding of the processes governing synaptic efficacy. Despite decades of intense study, these mechanisms remain poorly understood. The work contained in this thesis examines several such mechanisms using the hypothalamic-neurohypophysial system (HNS), a classical preparation for the study of Ca2+-dependent neuropeptide release. The first portion of this thesis is comprised of my efforts to define the cellular machinery essential for the exocytosis of secretory granules isolated from peptidergic neurohypophysial terminals of the HNS. Here, using the planar lipid bilayer model system, I have been able to show that syntaxin alone in the target membrane is sufficient to elicit fusion of modified neurohypophysial secretory granules. Surprisingly, SNAP-25 does not appear to be necessary for this process. This suggests that syntaxin may be able to substitute for SNAP-25 to form functional non-cognate fusion complexes. Additionally, the coupling of amperometric detection with the planar lipid bilayer system has allowed me to confirm these results using native, unmodified secretory granules, and also provides some insight into the kinetics of release in this reconstituted system. This model system should provide a convenient means for the study of additional regulatory factors believed to be involved in secretory vesicle exocytosis. The second and third sections of this thesis involve my examination of the role of presynaptic Ca2+ stores in neuropeptide secretion from isolated peptidergic neurohypophysial terminals (NHT). I initially examined the source of recently discovered ryanodine-sensitive Ca2+ stores in this system. Using Immuno-electron microscopy I have found that ryanodine receptor (RyR) labeling appears to co-localize with large dense core granules. Additionally, I have shown that a large conductance cation channel, with similarities to the RyR, found in the membrane of these granules has the same characteristic response to pharmacological agents specific for the RyR. Further, application of RyR agonists modulates basal neuropeptide release from NHT. These results suggest that the large dense core granules of NHT serve as the source of a functional ryanodine-sensitive Ca2+store. Recent work has revealed that spark-like Ca2+ transients, termed syntillas, can be observed in NHT. These syntillas arise from ryanodine-sensitive intracellular stores. In other neuronal preparations, similar Ca2+ transients have been suggested to affect spontaneous transmitter release. However, such a role for syntillas had yet to be examined. To assess if syntillas could directly trigger spontaneous release from NHT, I used simultaneous Ca2+imaging along with amperometric detection of release. Amperometry was adapted to this system via a novel method of false-transmitter loading. Using this approach I have found no apparent correlation between these two events, indicating that syntillas are unable to directly elicit spontaneous transmitter release. As this finding did not rule out an indirect modulatory role of syntillas on release, I additionally present some preliminary studies examining the ability of ryanodine-sensitive Ca2+ release to modulate vesicular priming. Using immunocytochemistry, I have shown that RyR agonist treatment shifts the distribution of neuropeptides toward the plasma membrane in oxytocinergic NHT, but not in vasopressinergic NHT. RyR antagonists have the opposite affect, again only in oxytocinergic NHT. Further, I have found that application of RyR agonists result in a facilitation of elicited release in NHT using membrane capacitance recording. This facilitation appears to be due primarily to an increase in recruitment of vesicles to the readily-releasable pool. These findings suggest that ryanodine-sensitive Ca2+stores may be involved in vesicular priming in NHTs. Taken together, the work presented in this thesis provides some new and interesting insights into the underlying mechanisms and modulation of transmitter release in both the HNS and other CNS terminals.

The promoter sequence of the mouse high affinity neurotensin receptor, Ntr-1, gene was cloned and characterized, sequences required for positive regulation in N1E-115 cells were localized, and at least two different peptides from these cells were shown to make specific contacts within the most potent positive regulatory element. A mouse neuroblastoma cell line, N1E-115, treated with 1.5% DMSO for 72 hours induces gene expression of both endogenous Ntr-l, and reporter constructs driven by the NTR-1 promoter, by 3 - 4 fold. The sequence ofthe NTR-1 promoter has no canonical TATA box, but is GC rich and contains consensus SP1, CACCC, CRE, and initiator elements. These elements are located within a 193 base positive regulatory region required for DMSO responsive activity and contains the transcriptional start site. Detailed mutational analysis of this region revealed that a CACCC box and the central region of a large GC rich palindrome are crucial cis-regulatory elements for DMSO induction. The SP1 element, an NGFI-A-related element, and the 5' end of the positive regulatory region are required for maintaining basal expression in N1E-115 cells. Cell type differences in the cis-regulatory elements that mediate both DMSO induction and maintenance of basal expression are observed. Characterization of proteins in N1E-115 cells that make specific contacts within the CACCC element identified at least two peptides with predicted sizes of 57 kd and 97 kd. Two dimensional UV crosslinking indicates that these proteins might contribute to inducible gel shift complexes that require the CACCC element. Several previously characterized CACCC binding proteins, belonging to the Kruppel-like family of transcription factors, were tested by supershift analysis for their ability to contribute to NTR-1 CACCC complexes. In fact, a protein closely related to SP1 does bind the CACCC element in N1E-115 cells, but of the other Kruppel-like protein tested, only BKLF contributes to a minor complex in N1E-115 cells. These results provide evidence for the complex regulation of Ntr-1 gene expression mediated by the cooperation of several cis-regulatory elements including a CACCC Kruppel-like binding element.

CP1 (encoded by the gene CEP1) is a sequence-specific DNA-binding protein of Saccharomyces cerevisiae that recognizes a sequence element (CDEI) found in both yeast centromeres and gene promoters. Strains lacking CP1 are viable but exhibit defects in growth, chromosome segregation, and methionine biosynthesis. To investigate the basis of the methionine requirement, a YEp24-based yeast genomic DNA library was screened for plasmids which suppressed the methionine auxotrophy of a cep1 null mutant. The suppressing plasmids contained either CEP1 or DNA derived from the PHO4 locus. PHO4 encodes a factor which positively regulates transcription of genes involved in phosphate metabolism via an interaction with CDEI-like elements within the promoters of these genes. Subcloning experiments confirmed that suppression correlated with increased dosage of PHO4. PHO4c, pho80, and pho84 mutations, all of which lead to constitutive activation of the PHO4 transcription factor, also suppressed cep1 methionine auxotrophy. The suppression appeared to be a direct effect of PHO4, not a secondary effect of PHO regulon derepression, and was dependent on a second transcriptional regulatory protein encoded by PHO2. Spontaneously arising extragenic suppressors of the cep1 methionine auxotrophy were also isolated; approximately one-third of the them were alleles of pho80. While PHO4 overexpression suppressed the methionine auxotrophy of a cep1 mutant, CEP1 overexpression failed to suppress the phenotype of a pho4 mutant; however, a cep1 null mutation suppressed the low-Pi growth deficiency of a pho84 mutant. The results suggest that CP1 functions as a transcriptional regulator of MET genes, and that activation of PHO4 restores expression to those genes transcriptionally-disabled by the cep1mutation. The results also suggest the existence of a network that cross-regulates transcription of genes involved in methionine biosynthesis and phosphate metabolism. A direct molecular approach to investigate CP1's role in MET gene expression was also taken. CDEI sites are associated with the promoter regions of most MET genes, but only MET16, the gene encoding PAPS reductase, has been shown to require CP1 for expression; both PAPS reductase activity, and MET16 mRNA are absent in cep1 mutants. Results of the present study demonstrate that CP1 participates in two systems which regulate expression of MET16, one triggered by methionine starvation and requiring the transactivator MET4 (pathway-specific control), and the other triggered by starvation for many different amino acids and requiring GCN4 (general control). CP1 was shown to mediate its regulatory function through the upstream CDEI site, and to act directly or indirectly to modulate the chromatin structure of the MET16 promoter. In addition, the pho80 mutation was found to partially restore MET16 expression to the cep1 strain, confirming the proposed nature of PHO4 suppression. A second methionine biosynthetic gene MET25, was also analyzed. Like MET16, MET25 was found to be regulated by both pathway-specific and general control mechanisms, but in contrast to MET16, CP1 only participated in the pathway-specific response of this gene. The results demonstrate that CP1, possibly by modulating changes in chromatin structure, assists the regulatory proteins MET4 and GCN4 in activating transcription of MET genes.

In mammals a single receptor protein binds both insulin-like growth factor II (IGF-II) and mannose 6-phosphate (Man 6-P) containing ligands, most notably lysosomal enzymes. However, in chick embryo fibroblasts IGF-II binds predominantly to a type 1 IGF receptor, and no IGF-II/Man 6-P receptor has been identified in this species. In order to determine if chickens possess an IGF-II/Man 6-P receptor, an affinity resin (pentamannosyl 6-phosphate (PMP) Sepharose) was used to purify receptors from chicken membrane extracts by their ability to bind mannose 6-phosphate. Then 125I-IGF-II was used to evaluate their ability to bind IGF-II. These experiments demonstrate that nonmammalian Man 6-P receptors lack the ability to bind IGF-II, suggesting that the ability to bind IGF-II has been gained recently in evolution by the mammalian Man 6-P receptor. The second area of study involves the serum form of the IGF-II/Man 6-P receptor. This receptor had been detected in the serum of a number of mammalian species, yet its structure, function, regulation, and origin were unknown. Initial studies, done with Dr. R. G. MacDonald, showed that the serum receptor is truncated such that the C-terminal cytoplasmic domain of the cellular receptor is removed. These studies also demonstrate a regulation of serum receptor levels with age, similar to that seen for the cellular receptor, and that the serum form of the receptor existed in several forms which appeared intact under nonreducing conditions, but as multiple proteolytic products upon reduction. Finally, these studies demonstrated that both the cellular and serum IGF-II/Man 6-P receptors are capable of binding IGF-II and Man 6-P simultaneously. In studies on the serum form of the IGF-II/Man 6-P receptor that I have conducted independently, the regulation of the serum IGF-II/Man 6-P and transferrin receptors by insulin has been demonstrated. In these studies, insulin injected into rats subcutaneously resulted in a time and dose dependent increase in serum receptor levels. Finally, to investigate the relationship of the serum IGF- II/Man 6-P receptor to the cellular form of the receptor, pulse chase experiments were performed. These experiments demonstrate that the soluble (serum form released into the medium) receptor is a major degradation product of the cellular receptor. Furthermore, the lack of detectable amounts of the lower Mr soluble receptor intracellularly and the parallel relationship of cell surface and soluble receptor suggest that the proteolysis is occurring from the cell surface. Finally, a number of experiments suggest that the degradation rate depends upon the conformation state of the receptor: binding of IGF-II or Man 6-P makes the receptor more susceptible to proteolysis while the presence of lysosomal enzymes prevents receptor proteolysis. In summary, the serum form of the IGF-II receptor is a proteolytic product of the cellular form of the receptor. The rate of release depends upon the number of receptors at the cell surface and the binding state of the receptor. In circulation, the receptor retains the ability to bind both types of ligands, it thus may serve as an IGF binding protein and/or a lysosomal enzyme binding protein. These results suggest a model whereby the cellular receptor is proteolytically cleaved by a plasma membrane protease to produce a short membrane anchored fragment and the serum receptor. In vivo this pathway serves as the major degradative pathway of the IGF-II/Man 6-P receptor, with the serum form being cleared from circulation by further degradation and reuptake.

Deoxycytidylate (dCMP) hydroxymethylase (CH) catalyzes the formation of 5-hydroxymethyl-dCMP (Hm5CMP) from dCMP and methylene tetrahydrofolate (CH2THF), analogous to the reaction between dUMP and CH2THF catalyzed by thymidylate synthase (TS), an enzyme of known structure. The amino acid sequence identity between invariant TS residues and CH is at least 50%. Most of the residues which contact the dUMP and CH2THF in TS are conserved in CH. It is hypothesized that CH is homologous to TS in both structure and mechanism. The project described in this thesis tests this hypothesis. In-vitro studies on catalysis by CH variants. The roles of three residues in catalysis by CH have been tested using site-directed mutagenesis. Conversion of Cys148 to Asp, Gly or Ser decreases CH activity at least 105 fold, consistent with a nucleophilic role for Cys148 (analogous to the catalytic Cys in TS). In crystalline TS, hydrogen bonds connect O4 and N3 of bound dUMP to the side chain of an Asn; the corresponding CH residue is Asp179. Conversion of Asp179 in CH to Asn reduces kcat/KM for dCMP by 104 fold and increases kcat/KM for dUMP 60 fold, changing the nucleotide specificity of the enzyme. Other studies have shown that the specificity of TS was changed from dUMP to dCMP by conversion of the appropriate Asn to Asp. Based on the crystal structure of TS, a Glu residue (also conserved in CH) is proposed to catalyze formation of the N5 iminium ion methylene donor by protonation of N10 of CH2THF. In CH and TS, overall turnover and tritium exchange are tightly coupled. Replacement of Glu60 in CH or Glu58 in TS uncouples these catalytic steps. Conversion the Glu60/58 to Gln or Asp results in a 5-50 fold decrease in the ability to catalyze tritium exchange, consistent with an inability to catalyze formation of the N5 iminium ion, but also results in a 104-105 decrease in product formation. This suggests that Glu60/58is also involved in a step in catalysis after nucleotide and folate binding and proton removal from carbon 5 of the nucleotide. Isotope effect studies. The observed value of the α-secondary tritium inverse equilibrium isotope effect (EIE = 0.8) on formation of the complex between FdUMP, CH2THF and both wild-type CH and CH(D179N) indicates that carbon 6 of FdUMP is sp3 hybridized (tetrahedral) in the ternary complex. This is consistent with the hypothesis that that carbon 6 is bonded to Cys148 in the complex. Removal of Cys148in CH prevents complex formation with FdUMP. Lack of an observed α-secondary tritium kinetic isotope effect (KIE) for position 6 of dCMP for both enzymes suggests that the intrinsic KIE is masked by other rate-limiting steps or that rehybridization follows the first irreversible step. An observed KIE on carbon 6 of dUMP by CH(D179N) suggests the rate-limiting steps for the two nucleotide substrates is different. In-vivo studies catalysis by CH variants. In order to prevent recombination between CH deficient T4 phage and plasmid borne copies of CH variants, the gene coding for CH, gene 42, was deleted from the T4 chromosome. The T4Δ42 phage requires wild-type CH expressed from a plasmid to kill their host cell. CH variants C148G, D179N, E60Q, and E60D, all which exhibit at least 2000 fold lower activity in vitro, do not complement the T4Δ42 phage in vivo. Interchanging the functional domains of CH and TS. It is proposed that shortening the C-terminal loop seen in the structure of TS changes the solvent structure of the CH active-site such that it becomes more hydrated. Differences in the solvent structure of the active-site may account for differences in the catalytic specificity between CH and TS, respectively, hydration versus reduction. In order to test the hypothesis that these catalytic differences between TS and CH lie within the C-terminal portion of the enzyme, the N-terminus of the CH(D179N) variant was fused to the C-terminus of the wild-type TS to create a chimeric CH/TS enzyme. The chimeric enzyme was predicted to have specificity for dUMP and a active-site solvent structure similar to that for wild-type TS. However, the resulting protein cannot be overproduced to significant levels and does not have any detectable TS activity in vivo.

Expression of the cell cycle dependent FO10S human H4 histone gene is regulated at both the transcriptional and post-transcriptional levels. We have investigated the 5' promoter elements mediating the transcriptional aspects of its regulation. A detailed in vivo and in vitro transcriptional analysis of promoter deletion mutants from this gene has identified three positive regulatory elements and two potentially negative regulatory elements within the first 1000 base pairs upstream of the transcription initiation site. In addition, the minimal promoter located within the first 70 base pairs is required for accurate transcription initiation and contains one of two in vivo identified protein-DNA interactions, site II. Binding of the nuclear factor HiNF-D to this region was correlated with the turn-on of histone gene transcription following stimulation of quiescent normal diploid fibroblasts to re-enter the proliferative phase. The most proximal positive regulatory element contains the other in vivo identified protein-DNA interaction, site I. Results from a series of in vitroprotein-DNA interaction studies revealed the binding of two nuclear factors to this element. One protein, HiNF-C, is indistinguishable from the transcription factor Sp1 while the other, HiNF-E, is a novel, potentially histone-specific member of the ATF transcription factor family. Binding of HiNF-C was required to stabilize the interaction of HiNF-E and together this region stimulated transcription 5 fold. The near-distal transcription activator region lies between -418 and -213 base pairs and forms a single protein- DNA complex, H4UA-1. The interaction domain for H4UA-1 contains recognition sequences for both the thyroid hormone receptor and the nuclear factor CTF/NF-1. The far-distal activator region (-730 and -589 base pairs) was the strongest positive regulatory element identified in the H4 promoter. This region increased transcription 10 fold and contains three protein-DNA interactions. One of the factors, H4UA-2, is an ATF transcription factor closely related to the HiNF-E interaction in the proximal positive element. These studies have defined the functional human H4 histone promoter to be a complex, modular structure extending at least 1000 base pairs.

The questions outlined in this thesis dissertation were proposed in order to provide insight regarding the mechanism by which the Drosophila Son of sevenless (dSOS) protein activates Ras. Ras proteins are GTP-binding proteins which bind guanine nucleotides very tightly and cycle between the inactive GDP-bound state and the active GTP-bound state. To address the mechanism by which the dSOS proteins activates Ras, a structure-function analysis of the dSOS protein was performed using truncation and deletion mutants of dSOS. In vivo Ras activation experiments using transiently transfected cells revealed that the NH2-terminal domain of dSOS is required in order for the catalytic domain of dSOS to exhibit exchange activity in cultured mammalian cells. The COOH-terminal GRB2 (Growth Factor Receptor Binding Protein) binding domain on the otherhand was insufficient to confer Ras exchange activity to the dSOS catalytic domain. Further analysis of the NH2-terminal domain of the dSOS protein demonstrated that the function of promoting catalytic domain activity could be localized by mutational analysis to the pleckstrin (PH) and DBL (Diffuse B-cell Lymphoma) homology sequences. Fractionation studies of cells transiently transfected with various dSOS mutant proteins demonstrated that the NH2-terminus of dSOS is also necessary for membrane association. These findings suggested that the model proposing that the recruitment of SOS via the adaptor protein GRB2 to the membrane is the main mechanism by which SOS activates Ras is unlikely to be the only mechanism by which SOS can activate Ras. From our data, a model can be proposed which postulates that SOS can activate Ras as a consequence of at least two steps. One step involves the SOS/GRB2 interaction and the second step involves the NH2-terminal domain of SOS associating with unidentified cellular elements.

For many paramyxoviruses, including Newcastle disease virus (NDV), syncytia formation requires the expression of both surface glycoproteins (HN and F) in the same cell, and evidence suggests that fusion involves a specific interaction between the HN and F proteins (23, 73). Because a potential interaction in paramyxovirus infected cells has never been clearly demonstrated, such an interaction was explored in Chapter 2 using coimmunoprecipitation and crosslinking. Both HN and F proteins could be precipitated with heterologous antisera after a five minute radioactive pulse as well as after a two hour chase in non-radioactive media, but at low levels. Chemical crosslinking increased detection of complexes containing HN and F proteins at the cell surface. After crosslinking, intermediate as well as high molecular weight species containing both proteins were precipitated with monospecific antisera. Precipitation of proteins with anti-HN after crosslinking resulted in the detection of complexes which electrophoresed in the stacker region of the gel, from 160-300 kD, at 150 kD and at 74 kD. Precipitates obtained with anti-F after crosslinking contained species which migrated in the stacker region of the gel, between 160-300 kD, at 120 kD and at 66 kD. The 3-4 discrete complexes ranging in size from 160-300 kD contained both HN and F proteins when precipitated with either HN or F antisera. That crosslinking of complexes containing both HN and F proteins was not simply a function of overexpression of viral glycoproteins at the cell surface was addressed by demonstrating crosslinking at early time points post infection, when levels of viral surface glycoproteins are low. Use of cells infected with an avirulent strain of NDV showed that chemically crosslinked HN and F proteins were precipitated independent of cleavage of F0. Furthermore, under conditions that maximized HN protein binding to its receptor, there was no change in the percentages of HN and F0 proteins precipitated with heterologous antisera, but a decrease in F1protein precipitated was observed upon attachment. These data argue that the HN and F proteins interact in the RER. Upon attachment of the HN protein to its receptor, the HN protein undergoes a conformational change which causes a subsequent change in the associated F protein, releasing the hydrophobic fusion peptide into the target membrane and initiating fusion. Chapter 3 explores the stalk region of the NDV HN protein, which has been implicated in both fusion promotion and virus specificity of that activity. The NDV F protein contains two heptad repeat motifs which have been shown by site-directed mutagenesis to be critical for fusion (7, 51, 57). Heptad repeat motifs mediate protein-protein interactions by enabling the formation of coiled-coils. Upon analysis of the stalk region of the NDV HN protein, we identified two heptad repeats. Secondary structure analysis of these repeats suggested the potential for these regions to form alpha-helices. To investigate the importance of this sequence motif for fusion promotion, we mutated the hydrophobic "a" position amino acids of each heptad repeat to alanine or methionine. In addition, hydrophobic amino acids in other positions were also changed to alanine. Every mutant protein retained levels of attachment activity that was greater than or equal to the wild-type protein and bound to conformation-specific monoclonal as well as polyclonal antisera. Neuraminidase activity was variably affected. Every mutation, however, showed a dramatic decrease in fusion promotion activity. The phenotypes of these mutant proteins indicate that individual amino acids within the heptad repeat region of the stalk domain of the HN protein are important for the fusion promotion activity of the protein. These data are consistent with the idea that the HN protein associates with the F protein via specific interactions between the heptad repeat regions of both proteins.

Although much progress has been made in understanding the events that lead to successful cell division, many details of this process remain a mystery. This dissertation presents findings which help to explain events that occur in the latest stages of cytokinesis, with an emphasis on the role of centrosome proteins. The first chapter introduces the novel centrosome protein centriolin. We show that this protein is localized specifically to the subdistal appendages of the maternal centriole in interphase, and it localizes to the midbody during cytokinesis. Disruption of this protein results in a unique cytokinesis defect in which cleavage furrow formation and ingression appear normal, but the cells remain connected by a thin intracellular bridge for extended periods of time. These results lead us to the conclusion that centriolin has an important function in cytokinesis. The second chapter describes our attempt to identify centriolin interacting partners. A yeast two hybrid screen was performed, and the results of this screen revealed an interaction between centriolin and proteins involved in vesicle target specificity and fusion. Further studies of these proteins revealed a novel localization to the midbody in cycling cells and a novel function in the final stages of cytokinesis, similar to centriolin. The third chapter discusses my attempts to clone and characterize a novel GTPase Activating Protein (GAP), which was also discovered in the screen for centriolin interacting proteins.