Human Immunodeficiency Virus Type 1 (HIV-1) is the causative agent of Acquired Immunodeficiency Syndrome (AIDS), currently the leading cause of death from infectious diseases. Since HIV-1 co-opts the host cellular machinery, the study of cellular factors involved is a rational approach in discovering novel therapeutic targets for AIDS drug development. In this thesis, we present studies on two such proteins. APOBEC3G is from the family of cytidine deaminases known to keep endogenous retroviruses and retrotransposons at bay to maintain stability of the human genome. APOBEC3G targets Vif-deficient HIV-1 particles and renders them noninfectious, partially through deaminase-dependent hypermutation of the provirus during reverse transcription. APOBEC3G largely localizes in mRNA processing (P) bodies, cytoplasmic structures involved in RNA metabolism. Here we explore the significance of APOBEC3G localization in P bodies. We found that disrupting P bodies does not affect virion incorporation of endogenous APOBEC3G, implying that the APOBEC3G fraction in P bodies is not directly involved in the production of nascent, non-infectious particles. We also study UPF1, another host protein encapsidated by HIV-1. It is an essential protein mainly studied for its role in nonsense-mediated decay (NMD) pathway and belongs to the same helicase superfamily as MOV10, a recently identified antiviral factor. We found that UPF1 is incorporated in HIV-1 virions in a nucleocapsid-dependent manner and is required for single-cycle infectivity at an early, post-entry step of the viral life cycle. This novel function of UPF1 most likely does not involve NMD since depletion of UPF2 does not affect viral infectivity.

The RNA interference (RNAi) pathway in Caenorhabditis elegans is a two-step, small RNA-mediated silencing pathway. Unlike in other organisms, Dicer processing of double-stranded RNA into small interfering (si) RNAs is not sufficient in worms to induce gene silencing. The activity of cellular RNA-dependent RNA polymerase (RdRP) is necessary to synthesize a secondary pool of siRNAs, which interact with a unique class of Argonaute proteins to form the functional effector complexes that mediate silencing. The aims of this thesis were to: 1) characterize the role of RdRP family members in endogenous small RNA biogenesis; 2) identify the Argonaute proteins that interact with RdRP-dependent small RNAs; and 3) investigate the biological function of RdRP-dependent small RNA pathways in C. elegans. In this thesis, I describe genetic, deep sequencing, and molecular studies, which identify 22G-RNAs as the most abundant class of endogenous small RNA in C. elegans. The 22G-RNAs resemble RdRP-dependent secondary siRNAs produced during exogenous RNAi, in that they possess a triphosphorylated 5’ guanine residue and exhibit a remarkable strand bias at target loci. Indeed, I show that 22G-RNAs are dependent on the activity of the RdRPs RRF-1 and EGO-1 and function in multiple distinct endogenous small RNA pathways. Interestingly, I have found that RRF-1 and EGO-1 function redundantly in the germline to generate 22G-RNAs that are dependent on and interact with members of an expanded family of worm-specific Argonaute (WAGO) proteins. The WAGO/22G-RNA pathway appears to be a transcriptome surveillance pathway that silences coding genes, pseudogenes, transposons, and non-annotated, or cryptic, transcripts. In contrast, I have found that EGO-1 alone is required for the biogenesis of a distinct class of 22G-RNAs that interact with the Argonaute CSR-1. Surprisingly, the CSR-1/22G-RNA pathway does not appear to silence its targets transcripts. Instead, the CSR-1/22G-RNA pathway is essential for the proper assembly of holocentric kinetochores and chromosome segregation. Lastly, I show that a third endogenous small RNA pathway, the ERI pathway, is a two-step silencing pathway that requires the sequential activity of distinct RdRPs and Argonautes. In the first step of this pathway, the RdRP, RRF- 3, is required for the biogenesis of 26G-RNAs that associate with the Argonaute, ERGO-1. In the second step, RRF-1 and EGO-1 generate 22G-RNAs that associate with the WAGO Argonautes. This work demonstrates how several C. elegans small RNAs pathways utilize RdRPs to generate abundant populations of small RNAs. These distinct categories of small RNAs function together with specific Argonaute proteins to affect gene expression, to play essential roles in development, and in the maintenance of genome and transcriptome integrity.

KCNE1 (E1) peptide is the founding member of the KCNE family (1-5), which is a class of type I transmembrane ß-subunits. KCNE1 peptides assemble with and modulate the gating, ion conducting properties and pharmacology of a variety of voltage-gated K+ channel a-subunits, including KCNQ1 (Q1). Mutations that interfere with the function of either E1 and/or Q1 and disrupt the assembly and trafficking of KCNE1- KCNQ1 channel complexes give rise to diseases such as Romano-Ward (RW) and Jervell Lange Nielsen Syndrome (JLNS), two different forms of Long QT Syndrome (LQTS). Using enzymatic deglycosylation assays, immunofluorescence techniques and quantitative cell surface labeling, we showed that KCNE1 peptides are retained in the early stages of the secretory pathway as immaturely N-linked glycosylated proteins. KCNE1 co-assembly with KCNQ1 leads to E1 progression through the secretory pathway and glycan maturation, resulting in cell surface expression. N-linked glycosylation of some membrane proteins is critical for proper folding, co-assembly and subsequent trafficking through the biosynthetic pathway. Previous studies have shown that genetic mutations that disrupt one of the two N-linked glycosylation sites on KCNE family members lead to LQTS (T7I, KCNE1 and T8A, KCNE2) (Schulze-Bahr et al., 1997; Sesti et al., 2000a; Park et al., 2003). Having confirmed that KCNE1 proteins acquire N-linked glycans, we examined the kinetics and efficiency of N-linked glycan addition to KCNE1. We showed that KCNE1 has two distinct N-linked glycosylation sites. The N-terminal sequon is a traditional co-translational site. The internal sequon (which is only ~ 20 residues away from the N-terminal sequon) acquires N-linked glycans primarily after protein synthesis (post-translationally). Surprisingly, mutations that prevent N-glycosylation at the cotranslational site also reduce the glycosylation efficiency of post-translational glycosylation at the internal sequon, resulting in a large population of unglycosylated KCNE1 peptides that are retained in the early stages of the secretory pathway and do not reach the cell surface with their cognate K+ channel. We showed that KCNE1 post-translational N-glycosylation in the endoplasmic reticulum is a cellular mechanism that ensures E1 proteins acquire the maximal number of glycans needed for proper channel assembly and trafficking. Our findings provide a new biogenic mechanism for human disease by showing that the JLNS mutation, T7I, not only inhibits glycosylation of the N-terminal sequon, but also indirectly prevents the glycosylation of the internal sequon, giving rise to a large population of assembly incompetent hypoglycosylated KCNE1 peptides. To further investigate the two N-linked glycosylation sites on KCNE1, we generated structure-function deletion scans of KCNE1 and performed positional glycosylation scanning mutagenesis. We examined the glycosylation pattern of glycosylation mutants in an effort to define the glycosylation window important for proper KCNE1 assembly and trafficking. Our findings suggested a nine amino acid periodicity to serve as a desirable glycosylation site and a better substrate for N-glycosylation. Appendix II shows work on the characterization of the C-terminally HA-tagged KCNE1 protein, which was used throughout the experiments presented in Chapter II, Chapter III and Chapter IV. Analysis of the C-terminally HA-tagged KCNE1 protein revealed that in heterologous expression systems KCNE1 had an internal translational start site, a methionine at position 27. A proteolytic cleavage site was also identified at the arginine cluster spanning residues 32 through 38 bearing the two known Long QT mutations (R32H and R36H) (Splawski et al., 2000; Napolitano et al., 2005). My work in Professor Craig C. Mello’s lab during the first four years of my graduate study is presented in Appendix I. The highly conserved Wnt/Wingless glycoproteins regulate many aspects of animal development. Wnt signaling specifies endoderm fate by controlling the fate of EMS blastomere daughters in 4-cell stage Caenorhabditis elegans embryos. A suppressor genetic screen was performed using two temperature sensitive alleles of mom-2/Wnt to identify additional regulators of the Wnt/Wingless signaling pathway during C. elegans endoderm specification. Five intragenic suppressors and three extragenic suppressors of mom-2/Wnt embryonic lethality were identified. We cloned ifg-1, eIF4G homologue, as one of the extragenic suppressors suggesting an intriguing connection between the Wnt signaling pathway and the translational machinery.

This thesis is comprised of studies of proteins involved in class I and class II major histocompatibility complex (MHC) antigen procressing. In class I MHC processing, structural and functional studies were conducted of an aminopeptidase, ERAP1, that mediates the final step in antigen processing to understand how it is particularly suitable for cleavage of antigenic peptides for class I MHC presentation. In the class II MHC antigen presentation pathway, structural studies were conducted to characterize a fluorogenic peptide that can be used to understand peptide loading events in vivo and in real time. Also structural studies of class II MHC and peptide complexes were conducted to understand the nature of an unique C-terminal secondary structure element exhibited by an HIV derived peptide in the peptide binding groove of class II MHC. The studies discussed in this thesis provide insights into the proteins involved in the class I and class II MHC antigen presentation pathway. The endoplasmic reticulum (ER) aminopeptidase, ERAP1, is a 941 amino acid member of the M1 family of zinc metalloaminopeptidases. Unlike other aminopeptidases, ERAP1 has a length and C-terminal preference for its substrates. Interestingly, ERAP1 has been shown to trim antigenic peptides to lengths of 8 or 9 amino acids long. This length matches the length required to bind into the peptide binding groove of class I MHC molecules. In addition, ERAP1 is upregulated in the ER of cells treated with interferon gamma (IFN-γ). Knock-down of ERAP1 by siRNA results in less overall antigenic presentation during IFN-γ treatment, although the knock-down does not affect all class I MHC epitopes equally. Knock-out studies show that ERAP1 effects the antigen repertoire at the cell surface. These and other data implicate ERAP1 as an important player in class I MHC antigen presentation. A chapter of this thesis will describe the crystallographic work describing the structures of ERAP1 with an aminopeptidase inhibitor, bestatin, and ERAP1 without an inhibitor that suggest possible peptide binding site in ERAP1 that will allow it to generate suitable substrates for a subset of class I MHC alleles. Class II MHC plays a key role in the immune response by presenting antigenic peptides on CD4+ cytotoxic cell surfaces for T-cell response. The binding of peptides onto the MHC is an important step in creating an immune response. Structures of peptide bound MHC class II show conserved side chain binding pockets within the overall peptide-binding groove. In HLA-DR1, a common human class II MHC, the P1 pocket shows a preference for large hydrophobic side chains. Development of environmentally sensitive peptide analogs, that can bind into the class II MHC the same way as native peptides, can assist in visualizing the antigen binding process. A chapter in this thesis describes the crystallographic work showing that (4-DAPA)-HA can be used to study antigen-presenting processes in a cell by visualizing the changes in fluorescence of the synthesized peptide upon antigen loading. Crystallographic analysis of MHC class II, HLA-DR1, in complex with HIV gag-derived peptide, GagP16(PEVIPMFSALSEGATP), and superantigen, SEC3- 3B2, reveals the conventional polyproline conformation up to MHC binding pocket residue, P9, while the C-terminus of GagP16 adopts an unusual β- hairpin loop structure. Additionally, interactions between the leucine at P8 (LeuP8) and other residues on the loop such as ThrP16 and AlaP14 of the hairpin loop, was observed. Importantly, GagP16 requires the last 4 amino acids (P13-P16), which is part of the hairpin loop, for T-cell recognition. Understanding what dictates the C-terminal hairpin loop and the interaction motif of HLA-DR1/GagP16 complex with its TCR will provide insights on why it is important for T cell activation. A chapter in this thesis discusses the structural investigation conducted to understand the determinants of the loop at the C-terminus of GagP16 using designed peptides. It will also discuss work involving HLA-DR1 with the T cell receptor, AC25, that was cloned from T cells that are specific to HLA-DR1 in complex with the GagP16 peptide.

Obesity has increased globally in epidemic proportions and as have the associated disorders. Insulin resistance that could further lead to type 2 diabetes is a major obesity associated dysfunction. Studies using insulin resistant mouse models and observations from human subjects exhibiting insulin resistance provide evidence for ectopic lipid deposition in organs like liver, muscle and heart as one of the major risk factors for developing insulin resistance. These observations suggest that deregulated adipose function to sequester and store excess energy as fat, could lead to insulin resistance. Furthermore, several studies have demonstrated adipose tissue dysfunction leading to inflammation and related syndromes. Interestingly, a mouse model with transgenic expression of glucose transporter in the adipose tissue exhibited improved glucose tolerance and increased insulin sensitivity despite development of obesity, upon high fat feeding. Thus mechanisms that improve adipose function could alleviate insulin resistance and associated diseases. Mitogen activated protein kinase kinase kinase kinase 4 (MAP4K4) was identified in our laboratory as a negative regulator of adipocyte function. Interestingly, siRNA mediated knockdown of MAP4K4 promoted PPARγ protein expression. Additionally, silencing of MAP4K4 increased adipocyte triglyceride content. Because MAP4K4 is a negative regulator of PPARγ expression and adipocyte function, understanding the mechanism by which MAP4K4 regulates PPARγ expression is of interest. Thus, for the first part of this thesis, I characterized the signaling pathways utilized by MAP4K4 to regulate PPARγ expression in cultured adipocytes. Here I show that MAP4K4 regulates PPARγ expression through regulation of its protein translation. siRNA mediated MAP4K4 gene silencing stimulated PPARγ protein synthesis without changing its mRNA transcription or its protein degradation. This increase in PPARγ protein translation was due to an increase in the activity of mammalian target of rapamycin (mTOR). The increase in PPARγ protein expression mediated by mTOR activation was a specific effect of the 4E-BP1 phosphorylation that leads to its inactivation and was not a general increase in mTOR activity towards all of its substrates. Finally, adenovirus mediated over expression of MAP4K4 inhibited mTOR activation, and suppressed PPARγ protein translation. For the second part of this thesis, I assessed the role of MAP4K4 in adipocytes in vivo. To accomplish this, a lentivirus mediated shRNA construct was generated to attenuate MAP4K4 expression selectively in the mouse adipose tissue. First we demonstrate that the MAP4K4 shRNA construct is able to efficiently silence the expression of MAP4K4 in vitro when co-expressed with Cre recombinase. Furthermore, we show that following modification of the lentiviral conditional vector that was introduced into a mouse embryo at one cell stage, and crossing the resulting founders with aP2-Cre mice, adipose tissue specific MAP4K4 gene silencing was achieved. Moreover, shRNA mediated gene silencing is a faster and an inexpensive means of achieving tissue specific gene knockdown relative to the available traditional gene knockout approaches. Utilizing these adipose specific MAP4K4 gene knockdown mice, I reveal that MAP4K4 silencing enhanced fat mass as well as PPARγ expression significantly. This is accompanied by improved whole body insulin sensitivity. Furthermore, when challenged with high fat diet, adipose-specific MAP4K4 silenced mice exhibit enhanced adiposity with decreased lean mass. Moreover, adipocyte cell size and triglyceride content are significantly increased. Interestingly, despite increased adiposity, hepatic insulin sensitivity is significantly improved leading to decreased glucose output. Thus MAP4K4 is an important regulator of adipocyte function that mediates whole body glucose homeostasis, through a mechanism that is yet to be identified.

Runt related transcription factors (Runx) play an important role in mammalian development by regulating the expression of key genes involved in cell proliferation, differentiation and growth. The work described in this thesis details the mechanisms by which the activity of two members of this family are regulated in human cells. Chapter One provides a brief introduction of Runx transcription factors. Chapter Two describes the regulation of Runx2 protein by the PI3 kinase/Akt pathway in human breast cancer cells. The PI3 kinase/Akt pathway is one of the major signal transduction pathways through which growth factors influence cell proliferation and survival. It is also one of the most frequently dysregulated pathways in human cancers. We identify Runx2 protein, a key regulator of breast cancer invasion as a novel substrate of Akt kinase and map residues of Runx2 that are phosphorylated by Akt in breast cancer cells. Our results show that phosphorylation by Akt increases the binding of Runx2 protein to its target gene promoters and we identify the phosphorylation events that enhance DNA binding of Runx2. Our work establishes Runx2 protein as a critical effecter downstream of Akt that regulates breast cancer invasion. In Chapter Three we describe the subnuclear localization of the tumor suppressor protein Runx3 during interphase and mitosis. We find that similar to other Runx family members, Runx3 protein resides in nuclear matrix associated foci during interphase. We delineate a subnuclear targeting signal that directs Runx3 to these nuclear matrix associated foci. Our work establishes that this association of Runx3 protein with the nuclear matrix plays a vital role in regulating its transcriptional activity. Chromatin immunoprecipitation results show that Runx3 occupies rRNA promoters during interphase. We also find that Runx3 remains associated with chromosomes during mitosis and localizes with nucleolar organizing regions (NORs), reflecting an interaction with epigenetic potential. This thesis provides novel insights into various mechanisms by which cells regulate the activity of Runx proteins.

Given that Streptococcus pneumoniae can cause life-threatening pulmonary and systemic infection, an apparent paradox is that the bacterium resides, usually harmlessly, in the nasopharynx of many people. Humoral immunity is thought to be the primary defense against serious pneumococcal infection, and we hypothesized that nasopharyngeal colonization of mice results in the generation of an antibody response that provides long-term protection against lung infection. We found that survival of of C57L/6 mice after intranasal inoculation with wild-type serotype 4 strain TIGR4 pneumococci required B cells but not T cells, suggesting that nasopharyngeal colonization elicited a protective humoral immune response. In fact, intranasal inoculation resulted in detectable pneumococcal-specific antibody responses, and protected mice against a subsequent high-dose S. pneumoniae pulmonary challenge. B cells were required for this response, and transfer of immune sera from i.n. colonized mice, or monoclonal antibodies against phosphorylcholine, a common surface antigen of S. pneumoniae, was sufficient to confer protection. IgA, which is thought to participate in mucosal immunity, contributed to but was not absolutely required for protection from pulmonary challenge. Protection induced by i.n. colonization lasted at least ten weeks. Although it was partially dependent on T cells, depletion of CD4+ T cells at the time of challenge did not alter protection, suggesting that T cells did not provide essential help in activation of conventional memory cells. Peritoneal B1b cells and radiation-resistant, long-lived antibody secreting cells have previously been shown to secrete anti-pneumococcal antibodies and mediate protection against systemic infection following immunization with killed bacteria or capsular polysaccharide [1, 2]. We found that peritoneal cells were not sufficient for colonization-induced protection, but sub-lethally irradiated mice largely survived pulmonary challenge. Thus, our results are consistent with the hypothesis that nasopharyngeal colonization, a common occurrence in humans, is capable of eliciting extended protection against invasive pneumococcal disease by generating long-lived antibody-secreting cells.

Of the eleven species of bacteria that comprise the genus Yersinia of the family Enterobacteriaceae, three species are pathogenic for humans. Yersinia pseudotuberculosis and Yersinia enterocolitica usually cause a mild, self-limiting mesenteric lymphadenitis or ileitis. Yersinia pestis causes a highly invasive often fatal disease known as plague. All three elaborate a type three secretion system that is essential for virulence and encoded on closely related plasmids. In Y. pestis, all the effectors, structural components and chaperones are encoded on the 70kb plasmid, pCD1. Of these, LcrV from Y. enterocolitica has been implicated in playing an immunosuppressive role through its interaction with host Toll-like receptor 2 (TLR2) and induction of IL-10. Through expression and purification of recombinant LcrV from Escherichia coliwe show that only high molecular weight species of rLcrV are able to stimulate TLR2. In a highly sensitive subcutaneous mouse infection model we demonstrate no difference in the time to death between TLR2-sufficient or deficient mice. Analysis of cytokine levels between these two genotypes also shows no significant difference between splenic IL-10 and IL-6 or levels of bacteria. We conclusively show that this interaction, if it does occur, plays no significant role in vivo. In a separate set of experiments, we also determined that the expression of F1, a peptide shown to be responsible for 37°C-dependent inhibition of invasion by Y. pestis in vitro, was significantly decreased under high oxygen conditions. This led us to re-examine the invasion phenotype both in vitro and in vivo. These results give new insights into virulence gene expression in Y. pestis by environmental cues other than temperature.

Once seen as a disease of wealthy nations, type 2 diabetes mellitus is now showing unprecedented growth throughout the world, fueling increases in microvascular and macrovascular complications. A compelling and growing body of evidence suggests that glucose intolerance and insulin resistance, hallmarks of the diabetic patient, may be driven by chronic inflammation. In particular, a predominance of visceral fat has been associated with enhanced inflammatory cytokine secretion that may contribute to enhanced risk of diabetes and comorbid cardiovascular disease in these individuals. As a function of its potency and wide environmental and biological distribution, we hypothesized that bacterial lipopolysaccharide (LPS, also known as endotoxin) may promote adipose inflammation and concomitant metabolic dysfunction. Indeed, expression of the LPS receptor CD14 is enhanced on visceral adipocytes of ob/ob mice, paralleling enhanced IL-6 secretion ex vivo. Furthermore, rosiglitazonefed ob/obmice demonstrated a reduction in CD14 that coordinated with diminished IL-6 secretion, suggesting a basis for the touted anti-inflammatory effects of this commonly employed type 2 diabetes medication. Mice deficient in components of the LPS signaling cascade, namely CD14, TLR4, and MyD88, yielded adipocytes with markedly attenuated IL-6 secretion, corroborating the central importance of LPS in adipocyte inflammation and supporting the role of this signaling pathway in depot-specific inflammation. Despite the prominent role of LPS signaling in adipocyte inflammation, CD14-, TLR4-, and MyD88-deficient mice failed to show resistance to diet induced obesity. Surprisingly, cd14-/- and tlr4-/- mice had marked glucose intolerance without alteration in total weight or adipose accumulation. In contrast, myd88-/- mice revealed minor glucose intolerance only with high fat diet challenge at an advanced age despite being overtly obese. In cd14-/- and tlr4-/-, but not myd88-/-, mice, an exaggerated rebound to hypoglycemia was associated with enhanced norepinephrine secretion, which could be abrogated by the adrenergic β-blocker propranolol. The overlay of these mouse models reveals a divergence of phenotypes that demonstrate LPS signaling disruption may lead to glucose intolerance and insulin resistance in part due to enhanced sympathoadrenal tone, uncovering an essential role of innate immunity in physiological stress and its impact upon glucose homeostasis.

Animals have evolved complicated and overlapping mechanisms to guard against the development of cancer and infection by pathogenic organisms. ARF, a potent tumor suppressor, positively regulates p53 by antagonizing p53’s negative regulator, MDM2, which in turn results in either apoptosis or cell cycle arrest. ARF also has p53-independent tumor suppressor activity. The CtBP transcriptional co-repressors promote cancer cell survival and migration/invasion. CtBP senses cellular metabolism via a regulatory dehydrogenase domain, and is a target for negative regulation by ARF. ARF targets CtBP to the proteasome for degradation, which results in the up regulation of proapoptotic BH3-only proteins, and p53-independent apoptosis. CtBP inhibition by ARF also up regulates PTEN, reducing cancer cell motility, making CtBP a potential therapeutic target in human cancer. The CtBP dehydrogenase substrate 4-methylthio-2-oxobutyric acid (MTOB) can act as a CtBP inhibitor at high concentrations, and is cytotoxic to cancer cells from a wide variety of tissues. MTOB induced apoptosis was independent of p53, and correlated with the de-repression of the pro-apoptotic CtBP repression target Bik. CtBP over-expression, or Bik silencing, rescued MTOB-induced cell death. MTOB did not induce apoptosis in mouse embryonic fibroblasts (MEFs), but was increasingly cytotoxic to immortalized and transformed MEFs, suggesting that CtBP inhibition may provide a suitable therapeutic index for cancer therapy. In human colon cancer cell peritoneal xenografts, MTOB treatment decreased tumor burden, and induced tumor cell apoptosis. To verify the potential utility of CtBP as a therapeutic target in human cancer the expression of CtBP and its negative regulator ARF was studied in a series of resected human colon adenocarcinomas. CtBP and ARF levels were inversely-correlated, with elevated CtBP levels (compared with adjacent normal tissue) observed in greater than 60% of specimens, with ARF absent in nearly all specimens exhibiting elevated CtBP levels. Targeting CtBP with a small molecule like MTOB may thus represent a useful and widely applicable therapeutic strategy in human malignancies. ARF has long been known to respond to virally encoded oncogenes. Recently, p14ARF was linked to the innate immune response to non-transforming viruses in mice. Therefore a wider role for the ARF pathway in viral infection was considered. Previous studies linking p53 to multiple points of the Human Immunodeficiency Virus-1 (HIV-1) life cycle suggested that ARF may also play a role in the HIV life cycle. In this study the interdependency of ARF and HIV infection was investigated. ARF expression was determined for a variety of cell types upon HIV infection. In every case, ARF levels exhibited dynamic changes upon HIV infection-in most cases ARF levels were reduced in infected cells. The impact of ARF over-expression or silencing by RNAi on HIV infection was also examined. Consistently, p24 levels were increased with ARF overexpression, and decreased when ARF was silenced. Thus ARF and HIV modulate each other, and ARF may paradoxically play a positive role in the HIV life cycle.