Stimulation of CD40 on APCs through CD40L expressed on helper CD4+ T cells activates and “licenses” the APCs to prime CD8+ T cell responses. While other stimuli, such as TLR agonists, can also activate APCs, it is unclear to what extent they can replace the signals provided by CD40-CD40L interactions. In this study, we used an adoptive transfer system to re-examine the role of CD40 in the priming of naïve CD8+ T cells. We find an approximately 50% reduction in expansion and cytokine production of TCR-transgenic T cells in the absence of CD40 on all APCs, and on dendritic cells in particular. Moreover, CD40-deficient and CD40L-deficient mice fail to develop endogenous CTL responses after immunization and are not protected from a tumor challenge. Surprisingly, the role for CD40 and CD40L are observed even in the absence of CD4+ T cells; in this situation, the CD8+T cell itself provides CD40L. Furthermore, we show that although TLR stimulation improves T cell responses, it cannot fully substitute for CD40. We also investigated whether CD40-CD40L interactions are involved in the generation, maintenance, and function of memory CD8+ T cells. Using a virus infection system as well as a dendritic cell immunization system, we show that the presence of CD40 on DCs and other host APCs influences the survival of activated effector cells and directly affects the number of memory CD8+ T cells that are formed. In addition, memory CD8+ T cell persistence is slightly impaired in the absence of CD40. However, CD40 is not required for reactivation of memory CD8+ T cells. It seems that CD40 signals during priming also contribute to memory CD8+ T cell programming but this function can be independent of CD4+T cells, similar to what we showed for primary responses. Altogether, these results reveal a direct and unique role for CD40L on CD8+ T cells interacting with CD40 on APCs that affects the magnitude and quality of primary as well as memory CD8+ T cell responses.

In response to external stimuli, many intracellular signaling proteins undergo dynamic changes in localization to the plasma membrane. Using the Saccharomyces cerevisiaemating pathway as a model, I investigated the molecular interactions that govern plasma membrane localization of signaling proteins, and how the plasma membrane compartmentalization of a signaling complex influences the overall signaling behavior of the pathway. Signaling proteins often consist of multiple interaction domains that collectively dictate their localization and function. Ste20 is a p21-activated kinase (PAK) that functions downstream of the Rho-type GTPase Cdc42 to activate several mitogen-activated protein (MAP) kinase pathways in budding yeast, including the mating pathway. I identified a short domain in Ste20 that directly binds to membrane lipids via electrostatic interaction. A mutation in this domain abolishes both the localization and function of Ste20. Thus, the previously known Cdc42 binding is necessary but not sufficient; instead, direct membrane binding by Ste20 is also critical. By replacing this domain with heterologous membranebinding domains, I demonstrated that phospholipid specificity is not essential in vivo. Functionally important short membrane-binding domains were also found in the Cdc42 effectors Gic1 and Gic2, indicating that generic membrane binding can work in concert with the CRIB domain to regulate activation of Cdc42 targets. These results underscore the importance of cooperation between protein-protein and protein-membrane interaction in achieving proper localization of signaling proteins at the cell cortex. At the system level, MAP kinase cascades can be graded or switch-like. The budding yeast mating pathway exhibits a graded response to increasing levels of pheromone. Previously the scaffold protein Ste5 was hypothesized to contribute to this graded response. To test this idea, I activated the pathway in a variety of ways and measured the response at the single cell level. I found that the graded response is not perturbed by the deletion of negative regulators of the pathway whereas the response became switch-like when the pathway was activated by a crosstalk stimulus that bypasses the upstream components. Interestingly, activation of the pathway in the cytoplasm using the graded expression of MAPKKK resulted in an ultrasensitive response. In contrast, activation of the pathway at the plasma membrane using the graded expression of membranetargeted active pathway components remained graded. In these settings, the scaffold protein Ste5 increased ultrasensitivity when limited to the cytosol; however, if Ste5 was allowed to function at the plasma membrane, signaling was graded. The results suggest that, in the mating pathway, the inherently ultrasensitive MAPK cascade is converted to a graded system by the scaffoldmediated assembly of signaling complexes at the plasma membrane. Therefore, the plasma membrane localization of Ste5 helps shape the input-output properties of the mating MAPK pathway in a manner that is suitable for the biology of mating. Taken together, this thesis underscores the importance of plasma membrane localization during mating pathway signaling in yeast. The examples described here provide further appreciation of how multiple interaction domains can function together to achieve specific targeting of the signaling proteins, as well as advances in understanding the role of scaffold proteins in modulating signaling behavior to promote graded signaling at the plasma membrane.

Dengue virus (DENV) infects 50-100 million people worldwide every year and is the causative agent of dengue fever (DF) and the more severe dengue hemorrhagic fever/dengue shock syndrome (DHF/DSS). There are four genetically and immunologically distinct DENV serotypes (DENV-1, DENV-2, DENV-3, and DENV-4). Evidence suggests that an increased risk for DHF/DSS during secondary infection with a heterologous DENV serotype is due to an immunopathological response mediated by serotype-cross-reactive memory T cells from the primary infection. Furthermore, epidemiological studies have shown that the sequence of infection with different DENV serotypes affects disease severity. Though much has been learned from human studies, there exist uncontrollable variables that are intrinsic in this system such as genetic factors and unknown infection histories. These factors can skew experimental results, making interpretations difficult. Therefore, a murine model to study the immunologic aspects of sequential dengue infections would be an asset to the field of dengue research. To examine the effect of sequential infection with different DENV serotypes on the CD8+ T cell response, we immunized Balb/c mice with a primary DENV infection on day 0 and subsequently challenged with a heterologous secondary DENV infection on day 28. We tested all possible sequences of infection with the four serotypes. We analyzed the T cell response to two previously defined epitopes on the DENV E (Ld-restricted) and NS3 (Kd-restricted) proteins. Using ELISPOT and intracellular cytokine staining, we measured the frequency of T cells secreting IFNγ and TNFα in response to stimulation with these epitopes during three phases: acute primary, acute secondary, and the memory phase after primary infection. We found that the T cell response in heterologous secondary infections was higher in magnitude than the response in acute primary infection or during the memory phase. We also found that the hierarchy of epitope specific responses, as measured by IFNγ secretion, was influenced by the sequence of infections. The adoptive transfer of immune serum or immune splenocytes suggested that memory T cells from the primary infection responded to antigens from the secondary infection. In vitroexperiments with T cell lines generated from mice with primary and secondary DENV infections suggested the preferential expansion of crossreactive memory T cells. In testing all of the different possible sequences of infection, we observed that two different sequences of infection (e.g., DENV-2 followed by DENV-1 versus DENV-2 followed by DENV-3) resulted in differential CD8+ T cell responses to the NS3 peptide even though both secondary infection serotypes contain the identical peptide sequence. To investigate this phenomenon, we examined the role of CD4+ T cell help on the memory CD8+ T cell response. We found that CD4+ T cell cytokine responses differ depending on the sequence of infection. In addition, it was also shown that crossreactivities of the CD4+ T cell response are also sequence-dependent. Moreover, denguespecific memory CD4+ T cells can augment the secondary CD8+ T cell response. Taken together, we demonstrated that this serotype sequence-dependent phenomenon is the result of differential help provided by cross-reactive memory CD4+T cells. The findings in this novel mouse model support the hypothesis that both CD4+ and CD8+ serotype-cross-reactive memory T cells from a primary dengue virus infection alter the immune response during a heterologous secondary dengue virus infection. These data further elucidate potential mechanisms whereby the specific sequence of infection with different dengue virus serotypes influences disease outcomes in humans.

Nitric oxide is a ubiquitous signaling molecule with both physiological and pathological functions in biological systems. Formed by the enzymatic conversion of arginine to citrulline, NO, has known roles in circulatory, immune and nervous tissues. In the nervous system nitric oxide has been implicated in long-term potentiation, neurotransmitter release, channel function, neuronal protection and neuronal degeneration. Much of our work has focused on yet another role for nitric oxide in cells, namely, neuronal differentiation. During development, neuronal differentiation is closely coupled with cessation of proliferation. We use nerve growth factor (NGF)-induced differentiation of PC12 pheochromocytoma cells as a model and find a novel signal transduction pathway that blocks cell proliferation. Treatment of PC12 cells with NGF leads to induction of nitric oxide synthase (NOS). The resulting nitric oxide (NO) acts as a second messenger, activating the p21(WAF1) promoter and inducing expression of p21(WAF1) cyclin-dependent kinase inhibitor. NO activates the p21(WAF1) promoter by p53-dependent and p53-independent mechanisms. Blocking production of NO with an inhibitor of NOS reduces accumulation of p53, activation of the p21(WAF1) promoter, expression of neuronal markers, and neurite extension. To deternine whether p21(WAF1) is required for neurite extension, we prepared a PC12 line with an inducible p21(WAF1) expression vector. Blocking NOS with an inhibitor decreases neurite extension, but induction of p21(WAF1) with isopropyl-1-thio-beta-D-galactopyranoside restored this response. Levels of p21(WAF1) induced by isopropyl-1-thio-beta-D-galactopyranoside were similar to those induced by NGF. Therefore, we have identified a signal transduction pathway that is activated by NGF; proceeds through NOS, p53 and p21(WAF1) to block cell proliferation; and is required for neuronal differentiation by PC12 cells. In further studies of this pathway, we have examined the role of MAP kinase pathways in neuronal nitric oxide synthase (nNOS) induction during the differentiation of PC12 cells. In NGF-treated PC12 cells, we find that nNOS is induced at RNA and protein levels, resulting in increased NOS activity. We note that neither nNOS mRNA, nNOS protein nor NOS activity is induced by NGF treatment in cells that have been infected with a dominant negative Ras adenovirus. We have also used drugs that block MAP kinase pathways and assessed their ability to inhibit nNOS induction. Even though U0126 and PD98059 are both MEK inhibitors, we find that U0126, but not PD98059, blocks nNOS induction and NOS activity in NGF-treated PC12 cells. Also, the p38 kinase inhibitor, SB 203580, does not block nNOS induction in our clone of PC12 cells. Since the JNK pathway is not activated in NGF-treated PC12 cells, we determine that the Ras-ERK pathway and not the p38 or JNK pathway is required for nNOS induction in NGF-treated PC12 cells. We find that U0l26 is much more effective than PD98059 in blocking the Ras-ERK pathway, thereby explaining the discrepancy in nNOS inhibition. We conclude that the Ras-ERK pathway is required for nNOS induction. The activation of soluble guanylate cyclase and the production of cyclic GMP is one of the best characterized modes of NO action. Having shown that inhibition of NOS blocks PC12 cell differentiation we tested whether nitric oxide acts through soluble guanylate cyclase to lead to cell cycle arrest and neuronal differentiation. Unlike NOS inhibition, the inhibition of soluble guanylate cylcase does not block the induction of neuronal markers. Moreover, treatment of NGF-treated, NOS-inhibited PC12 cells with a soluble analog of cyclic GMP was unable to restore differentiation of those cells. Hence, cGMP is not a component of this pathway and we had to consider other mechanisms of NO action. It has become increasingly evident that another manner by which NO may exert its effects is by S-nitrosylation of cysteine residues. We tested, in vitro whether nitric oxide may control p53 by S-nitrosylation and inactivation of the p53 negative regulator, Hdm2. Treatment of Hdm2 with a nitric oxide donor inhibits Hdm2-p53 binding, the first step in Hdm2 regulation of p53. The presence of cysteine or DTT blocks this inhibition of binding. Moreover, nitric oxide inhibition of Hdm2-p53 binding was found to be reversible. Sulfhydryl-sensitivity and reversibility are consistent with nitrosylation. Finally, we have identified a critical cysteine residue that nitric oxide modifies in order to disrupt Hdm2-p53 binding. Mutation of this residue from a cysteine to an alanine does not interfere with binding but rather eliminates the sensitivity of Hdm2 to nitric oxide inactivation.

This thesis examines the effects of nerve growth factor (NGF) and lithium on the regulation of neurotensin/neuromedin N (NT/N) gene expression in PC12 pheochromocytoma cells. In PC12 cells, the expression of the rat NT/N gene is strictly dependent on simultaneous exposure to combinations of NGF, glucocorticoids, activators of adenylate cyclase, and lithium. Transient transfection experiments indicated that a consensus AP-1 site located within the NT/N promoter is the principal target of NGF and lithium action. NGF rapidly, but transiently, induces the expression of several AP-1 genes in PC12 cells, suggesting that the effect of NGF on NT/N gene expression results from increased AP-1 activity. These results led to the prediction that the induction of NT/N gene expression should be rapid, transient and dependent on de novoprotein synthesis. These experiments also suggested that the NT/N gene is principally regulated through the initiation of transcription. However, post-transcriptional mechanisms may also be involved. Experiments in this thesis were designed to examine the regulatory mechanisms responsible for increased NT production in PC12 cells when treated with different inducer combinations and whether AP-1 factors could act as mediators in responses to NGF and lithium. Results described in this thesis indicate that the principal mechanism by which NGF and lithium regulate NT biosynthesis is by activating NT/N gene transcription. Comparison of NT/N mRNA, pro NT/N synthetic rates, proNT/N proteins and mature NT levels in induced PC12 cells, demonstrated that NGF and lithium had no effect on the translation of NT/N mRNA and had only a modest effect on post-translational processing. Nuclear run-on assays showed that NT/N transcription is transicntly activated in maximally induced cells. A rapid RNase protection assay was developed to examine both the kinetics of NT/N gene activation and whether activation requires newly synthesized proteins. Quantitation of nuclear NT/N precursor RNA. using a probe spanning the junction between exon onc and intron one, provides a sensitive measure of NT/N gene activity and by several criteria provides an accurate measure of NT/N transcription. When either NGF or lithium was combined with dexamethasone and forskolin, nuclear NT/N precursor RNA transiently accumulated, although each inducer displayed different kinetics, rapid and delayed, respectively. De novo protein synthesis was not required for activating NT/N transcription when NGF was used as the permissive agent, although newly synthesized proteins secm to be needed for subsequent down-regulation. The response to lithium displayed a marked requirement for new protein synthesis, consistent with the involvement of newly synthesized AP-1 factors. RNA blot analysis showed that lithium either alone or in combination with dexamethasone and forskolin induced c-jun and fra-1 gene expression with delayed kinetics, consistent with c-Jun/Fra-1 complexes mediating the effects of lithium on NT/N gene transcription. The pathway identified by lithium does not activate or require protein kinase C. This pathway is also active in neuronally-differentiated PC12 cells suggesting that it could be involved in the regulation of NT/N gcne exprcssion in the intact nervous system. These results and order of addition experiments demonstrate that NGF and lithium activate distinct pathways required for NT/N gene induction.

How proteins undergo conformational changes to bind a ligand is one of the most fundamental questions of protein biology. MD simulations provide a useful computational tool for studying the theoretical movements of protein in solution on nanosecond timescales. The results of these simulations can be used to guide experimental design. By correlating the theoretical models with the results of experimental studies, we can obtain a significant amount of information about protein dynamics. This study represents the application of both computational and traditional experimental techniques to study protein dynamics in HIV-1 protease. The results provide a novel mechanism for the conformational changes in proteins and address the role of residues outside the active site in protein dynamics. Additionally, these results are applied to the complex role of non-active site mutations in the development of drug resistance. Chapter II examines an invariant Thr80 at the apex of the P1 loop of HIV-1, HIV-2, and simian immunodeficiency virus protease. Sequence variability associated with human immunodeficiency virus type 1 (HIV-1) is useful for inferring structural and/or functional constraints at specific residues within the viral protease. Positions that are invariant even in the presence of drug selection define critically important residues for protease function. Three protease variants (T80V, T80N, and T80S) were examined for changes in structure, dynamics, enzymatic activity, affinity for protease inhibitors, and viral infectivity. While all three variants were structurally similar to the wild type, only T80S was functionally similar. T80V significantly decreased the ability of the enzyme to cleave a peptide substrate but maintained infectivity, while T80N abolished both activity and viral infectivity. Additionally, T80N decreased the conformational flexibility of the flap region, as observed by simulations of molecular dynamics. Taken together, these data indicate that HIV-1 protease functions best when residue 80 is a small polar residue and that mutations to other amino acids significantly impair enzyme function, possibly by affecting the flexibility of the flap domain. Chapter III focuses on residues within the hydrophobic core of each monomer in HIV-1 protease. Many hydrophobic residues located in the core of this dimeric enzyme frequently mutate in patients undergoing protease inhibitor therapy. The mechanism by which these mutations aid the development of drug resistance is not well understood. Using MD simulations, this study suggests that the hydrophobic residues outside the active site facilitate the conformational change that occurs in HIV-1 protease upon binding substrates and inhibitors. In these simulations, the core of each monomer significantly rearranges to assist in the expansion of the active site as hydrophobic core residues slide by each other, exchanging one hydrophobic contact for another. Such hydrophobic sliding may represent a general mechanism by which proteins undergo conformational changes. Mutation of these hydrophobic core residues would alter the packing of the hydrophobic core. Thus, these residues could facilitate drug resistance in HIV-1 protease by altering dynamic properties of HIV-1 protease preferentially affecting the relative affinity for inhibitors versus substrates. Chapter IV concentrates on a residue in the flap region, Ile54, which is significantly correlated with the development of drug resistance. A series of patient sequences containing the mutation I54A were evaluated for the most frequently occurring co-mutations. I54A was found to occur with mutations that were previously correlated with I54V mutations, including L10I, G48V, and V82A. Based on the results of this evaluation, the binding properties of five variant proteases were investigated: MDI54V, MDRI54A, I54V, I54A, and G48V. MDRI54V and MDRI54Aeach contained the mutations L10I, G48V, and V82A, and either I54V or I54A, respectively. The other variants contained only the mutation indicated. Mutations at Ile54 were able to significantly impact the thermodynamics of binding to saquinavir, amprenavir, and the recently approved darunavir. The magnitude of this impact depended on the presence or absence of other drug resistance mutations, including another mutation in the flap region, G48V. Therefore, while residues 48 and 54 are not in contact with each other, mutations at both sites had a cooperative effect that varies between inhibitors. The results demonstrate that residues outside the active site of HIV-1 protease are clearly important to enzyme function, possibly through their role in the dynamic properties of protease. Mutations outside the active site of protease that are known to cause drug resistance could alter the conformational flexibility of protease. While the role of protein dynamics in molecular recognition is still not fully understood, the results of this study indicate that altering the dynamic properties of a protein affects its ability to recognize ligands. Therefore, to design better inhibitors we will have to develop a more thorough understanding of protein dynamics.

RNA interference (RNAi) is an evolutionarily conserved, sequence-specific gene silencing pathway found in eukaryotes, in which 21-nucleotide, small interfering RNAs (siRNAs) guide destruction of a corresponding target mRNA. RNAi is a natural mechanism for both genome surveillance and gene regulation. Moreover, siRNAs can be transfected into cultured mammalian cells, causing the sequence-specific ‘knock down’ of an mRNA. My work in the Zamore lab has centered around the Drosophilain vitro system and cultured mammalian cells to study the RNA interference (RNAi) pathway. small interfering RNAs (siRNAs) are incorporated into the RNA-induced silencing complex (RISC), which culminates in the cleavage of a complementary target mRNA. Previous work proved that certain structural features of siRNAs are essential for RNAi in flies, including the requirement for 5´ phosphates and 3´ hydroxyl groups. In cultured mammalian cells, the requirement for a 5´ phosphate also holds true, but we found no evidence to support the necessity for 3´ hydroxyls in either system. In addition, siRNAs can act as single strands entering the pathway downstream of double-stranded siRNAs, both of which are competent in directing the cleavage of its cognate mRNA at a single site. While these key features are a requirement for functional siRNAs, alone they do not determine the efficiency to which an siRNA can enter the RISC. In fact, both strands of an siRNA can enter RISC to a different degree as determined by the stabilities of the 5´ ends of the siRNA strand, a phenomenon termed ‘functional asymmetry’. This characteristic is also reflected in another class of small RNAs involved in gene silencing known as microRNAs (miRNAs), which are processed from long hairpin RNA structures into mature, single-stranded non-coding RNAs. The asymmetric loading of siRNAs suggests that miRNAs are initially generated from siRNA-like duplexes cleaved from the stem of the hairpins. The strand whose 5´ end is less tightly paired will be processed into the mature miRNA, while the other strand is destroyed. By applying the rules of siRNA asymmetry it is possible to predict which side of the stem will be processed into the mature miRNA, a finding verified experimentally by our lab and others. This discovery also has additional implications in designing highly effective siRNAs and in reducing siRNA off-target effects. We used these results to design siRNAs that target the single nucleotide polymorphism in superoxide dismutase that causes the familial form of amyotrophic lateral sclerosis (ALS), but leave the wild-type mRNA intact and functional. Our experiments have helped define the ‘rules’ for creating SNP-specific siRNAs. In particular, we found that only siRNAs with a purine:purine mismatch to the allele not intended for destruction show good discrimination. The placement of the mismatch in a tiled set of siRNAs shows that mismatches located in the 5´ region of the siRNA, a region shown to be responsible for siRNA binding, can not discriminate between alleles. In contrast, mismatches in the 3´ region of the siRNA, the region contributing to catalysis, discriminate between wild-type and mutant alleles. This work is an important step in creating allele-specific siRNAs as therapeutics for dominant negative genetic diseases. But how does RISC cleave its target? By isolating both the 5´ and 3´ cleavage products produced by RISC in the Drosophila in vitro system, we discovered that RISC acts as a Mg2+-dependent endonuclease that cleaves a single phosphodiester bond in the mRNA target, leaving 5´ phosphate and 3´ hydroxyl groups. These findings were a critical step in the demonstration that Argonaute, a protein known to be a component of RISC, is the RNAi endonuclease.

ATP plays a critical role in the regulation of many enzyme processes. In this work, I have focused on the ATP mediated regulation of the recombination processes catalyzed by the E. coliRecA and the human Rad51 proteins. The RecA protein is a multifunctional enzyme, which plays a central role in the processes of recombinational DNA repair, homologous genetic recombination and in the activation of the cellular SOS response to DNA damage. Each of these functions requires a common activating step, which is the formation of a RecA-ATP-ssDNA nucleoprotein filament. The binding of ATP results in the induction of a cooperative, high affinity ssDNA binding state within RecA (Menetski & Kowalczykowski, 1985b; Silver & Fersht, 1982). Data presented here identifies Gln194 as the NTP binding site "γ-phosphate sensor", in that mutations introduced at this residue disrupt all ATP induced RecA activities, while basal enzyme function is maintained. Additionally, we have dissected the parameters contributing to cooperative nucleoprotein filament assembly in the presence of cofactor. We show that the dramatic increase in the affinity of RecA for ssDNA in the presence of ATP is a result of a significant increase in the cooperative nature of filament assembly and not an increase in the intrinsic affinity of a RecA monomer for ssDNA. Previous work using both mutagenesis and engineered disulfides to study the subunit interface of the RecA protein has demonstrated the importance of Phe217 for the maintenance of both the structural and functional properties of the protein (Skiba & Knight, 1994; Logan et al., 1997; Skiba et al., 1999). A Phe217Tyr mutation results in a striking increase in cooperative filament assembly. In this work, we identify Phe217 as a key residue within the subunit interface and clearly show that Phe217 is required for the transmission of ATP mediated allosteric information throughout the RecA nucleoprotein filament. The human Rad51 (hRad51) protein, like its bacterial homolog RecA, catalyzes genetic recombination between homologous single and double stranded DNA substrates. This suggests that the overall process of homologous recombination may be conserved from bacteria to humans. Using IAsys biosensor technology, we examined the effect of ATP on the binding of hRad51 to ssDNA. Unlike RecA, we show that hRad51 binds cooperatively and with high affinity to ssDNA both in the presence and absence of nucleotide cofactor. These results show that ATP plays a fundamentally different role in hRad51 vs.RecA mediated processes. In summary, through the work presented in this dissertation, we have defined the critical molecular determinants for ATP mediated allosteric regulation within RecA. Furthermore, we have shown that ATP is not utilized by Rad51 in the same manner as shown for RecA, clearly defining a profound mechanistic difference between the two proteins. Future studies will define the requirement for ATP in hRad51 mediated processes.

The JNK signal transduction pathway mediates a broad, complex biological process in response to inflammatory cytokines and environmental stress. These responses include cell survival and apoptosis, proliferation, tumorigenesis and the immune response. The divergent cellular responses caused by the JNK signal transduction pathway are often regulated by spatial and cell type contexts, as well as the interaction with other cellular processes. The discovery of additional components of the JNK signal transduction pathway are critical to elucidate the stress response mechanisms in cells. This thesis first discusses the cloning and characterization of two novel members of the JNK signal transduction pathway. JIP1 and JMP1 were initially identified from a murine embryo library through a yeast Two-Hybrid screen to identify novel JNK interacting proteins. Full length cDNAs of both genes were cloned and analyzed. JIP1 represents the first member of the JIP group of JNK scaffold proteins which were characterized. The JNK binding domain (JBD) of JIP1 matches the D-domain consensus of other JNK binding proteins, and it demonstrates JNK binding both in vitro and in vivo. This JNK binding was demonstrated to inhibit JNK signal transduction and over-expression of JIP1 inhibits the JNK mediated pre-B cell transformation by bcr-abl. Over-expressed JIP1 also sequesters JNK in the cytoplasm, which may be a mechanism of the inhibition of JNK signaling. A new, high-resolution digital imaging microscopy technique using deconvolution demonstrated the absence of JNK1 in the nucleus of co-transfected JIP1 and JNK1 cells. The other protein discussed in this thesis is JMP1, a novel JNK binding, microtubule co-localized protein. There is a JBD in the JMP1 carboxyl end and a consensus D-domain within this region. The JMP1 JBD demonstrates an increased association with phospho-JNK from UV irradiated cells compared to un-irradiated cells in vivo. JMP1 also has 12 WD-repeat motifs in its amino terminal end which are required for microtubule co-localization. JMP1 demonstrates a cell cycle specific localization at the mitotic spindle poles. This co-localization is dependent on intact microtubules and the amino-terminal WD-repeats are required for this localization. JMP1 mRNA is highly expressed in testis tissues. Immunocytochemistry on murine testis sections using an affinity purified anti-JMP1 antibody demonstrates JMP1 protein in the lumenal compartment of the seminiferous tubules. JMP1 protein is expressed in primary and secondary spermatocytes, cells which are actively undergoing meiosis. The results obtained from the localization of JMP1 in meiotic spermatocytes led to an investigation of the roles of JNK signal transduction in the testis. The testis is an active region of cellular proliferation, apoptosis and differentiation, which make it an appealing model for studying JNK signal transduction. However, the roles JNK signaling have in the testis are poorly understood. I investigated the reproduction capability of Jnk3-/- male mice and discovered older Jnk3-/- males had a reduced capacity to impregnate females compared to younger animals and age-matched wild type controls. The testis morphology and sperm motility of these animals were similar to wild-type animals, and there was no alteration of apoptosis in the testis. The final section of this thesis involves the study of this breeding defect and investigating for cellular defects that might account for this age-related Jnk3-/- phenotype.

RNA silencing is an evolutionary conserved sequence-specific mechanism of regulation of gene expression. RNA interference (RNAi), a type of RNA silencing in animals, is based on recognition and endonucleolytic cleavage of target mRNA complimentary in sequence to 21-nucleotide (nt) small RNA guides, called small interfering RNAs (siRNAs). Another class of 21-nt small RNAs, called micro RNAs (miRNAs), is endogenously encoded in eukaryotic genomes. Both production of siRNAs from long double-stranded RNA (dsRNA) and biogenesis of miRNAs from hairpin structures are governed by the ribonuclease III enzyme Dicer. Although produced as duplex molecules, siRNAs and miRNAs are assembled into effector complex, called the RNA-induced silencing complex (RISC), as single-strands. A member of the Argonaute family of small RNA-binding proteins lies at the core of all known RNA silencing effector complexes. Plants and animals contain multiple Argonaute paralogs. In addition to endonucleolytic cleavage, Argonaute proteins can direct translational repression/destabilization of mRNA or transcriptional silencing of DNA sequences by the siRNAdirected production of silent heterochromatin. The Schizosaccharomyces pombe genome encodes only one of each of the three major classes of proteins implicated in RNA silencing: Dicer (Dcr1), RNA-dependent RNA polymerase (RdRP; Rdp1), and Argonaute (Ago1). These three proteins are required for silencing at centromeres and for the initiation of transcriptionally silent heterochromatin at the mating-type locus. That only one Dicer, RdRP and Argonaute is expressed in S. pombe might reflect the extreme specialization of RNA silencing pathways regulating targets only at the transcriptional level in this organism. We decided to test if classical RNAi can be induced in S. pombe. We introduced a dsRNA hairpin corresponding to a GFP transgene. GFP silencing triggered by dsRNA reflected a change in the steady-state concentration of GFP mRNA, but not in the rate of GFP transcription. RNAi in S. pombe required dcr1, rdp1, and ago1, but did not require chp1, tas3, or swi6, genes required for transcriptional silencing. We concluded that the RNAi machinery in S. pombecould direct both transcriptional and posttranscriptional silencing using a single Dicer, RdRP, and Argonaute protein. Our findings suggest that, in spite of specialization in distinct siRNA-directed silencing pathways, these three proteins fulfill a common biochemical function. In Drosophila, miRNA and RNAi pathways are both genetically and biochemically distinct. Dicer-2 (Dcr-2) generates siRNAs, whereas the Dicer-1 (Dcr-1)/Loquacious complex produces miRNAs. Argonaute proteins can be divided by sequence similarity into two classes: in flies, the Ago subfamily includes Argonaute1 (Ago1) and Argonaute2 (Ago2), whereas the Piwi subfamily includes Aubergine, Piwi and Argonaute 3. siRNAs and miRNAs direct posttranscriptional gene silencing through effector complexes containing Ago1 or Ago2. The third class of small RNAs, called repeat-associated small interfering RNAs (rasiRNAs), is produced endogenously in the Drosophilagerm line. rasiRNAs mediate silencing of endogenous selfish genetic elements such as retrotransposons and repetitive sequences to ensure genomic stability. We examined the genetic requirements for biogenesis of rasiRNAs in both male and female germ line of Drosophilaand silencing of 8 different selfish elements, including tree LTR retrotransposons, two non-LTR retrotransposons, and three repetitive sequences. We find that biogenesis of rasiRNAs is different from that of miRNAs and siRNAs. rasiRNA production appears not to require Dicer-1 or Dicer-2. rasiRNAs lack the 2´,3´ hydroxy termini characteristic of animal siRNA and miRNA. While siRNAs derive from both the sense and antisense strands of their dsRNA precursors, rasiRNAs accumulate in antisense polarity to their corresponding target mRNAs. Unlike siRNAs and miRNAs, rasiRNAs function through the Piwi, rather than the Ago, Argonaute protein subfamily. We find that rasiRNAs silence their target RNAs posttranscriptionally: mutations that abrogate rasiRNA function dramatically increase the steady-state mRNA level of rasiRNA targets, but do not alter their rate of transcription, measured by nuclear run-on assay. Our data suggest that rasiRNAs protect the fly germ line through a silencing mechanism distinct from both the miRNA and RNAi pathways.