HIV-1 accessory proteins VPR and Vif modulate antiviral response by targeting IRF-3 for degradation (original) (raw)

Virology. Author manuscript; available in PMC 2009 Mar 30.

Published in final edited form as:

PMCID: PMC2312338

NIHMSID: NIHMS44100

Atushi Okumura

1The Sidney Kimmel Comprehensive Cancer Center

Tim Alce

1The Sidney Kimmel Comprehensive Cancer Center

Barbora Lubyova

1The Sidney Kimmel Comprehensive Cancer Center

Heather Ezelle

3USA University of Maryland, School of Medicine, Baltimore, MD, 21231, USA

Klaus Strebel

4Laboratory of Molecular Microbiology, National Institutes of Health, Bethesda, MD, 20892, USA

Paula M. Pitha

1The Sidney Kimmel Comprehensive Cancer Center

2Department of Molecular Biology and Genetics and Department of Biology, The Johns Hopkins University, Baltimore MD, 21231

1The Sidney Kimmel Comprehensive Cancer Center

2Department of Molecular Biology and Genetics and Department of Biology, The Johns Hopkins University, Baltimore MD, 21231

3USA University of Maryland, School of Medicine, Baltimore, MD, 21231, USA

4Laboratory of Molecular Microbiology, National Institutes of Health, Bethesda, MD, 20892, USA

5Present address: Institute of Immunology and Microbiology, 1st Medical Faculty of Charles University, Prague, Czech Republic

#To whom correspondence should be addressed: email: ude.imhj@eworap

Abstract

The activation of IRF-3 during the early stages of viral infection is critical for the initiation of the antiviral response; however the activation of IRF-3 in HIV-1 infected cells has not yet been characterized. We demonstrate that the early steps of HIV-1 infection do not lead to the activation and nuclear translocation of IRF-3; instead, the relative levels of IRF-3 protein are decreased due to the ubiquitin associated proteosome degradation. Addressing the molecular mechanism of this effect we shown that the degradation is independent of HIV-1 replication and that virion associated accessory proteins Vif and Vpr can independently degrade IRF-3. The null mutation of these two genes reduced the capacity of the HIV-1 virus to down modulate IRF-3 levels. The degradation was associated with Vif and Vpr mediated ubiquitination of IRF-3 and was independent of the activation of IRF-3. N-terminal lysine residues were shown to play a critical role in the Vif-and Vpr-mediated degradation of IRF-3. These data implicate Vif and Vpr in the disruption of the initial antiviral response and point to the need of HIV-1 to circumvent the antiviral response during the very early phase of replication.

Introduction

The innate immune response has developed as a rapid and regulated defense mechanism in which the recognition of an invading pathogenic organism induces multiple signaling pathways leading to the activation of transcription factors that control the expression of a diverse set of genes involved in the coordination of the immune responses (Akira, Uematsu, and Takeuchi, 2006). Antiviral cytokines (interferons), activated as an early response to infection, play an important role both in the outcome of the viral infection and in its virulence (Samuel, 2001). Two families of transcription factors play a major role in the transcriptional activation of Type I interferon genes (IFN A and B): the well-characterized family of NFκB factors and a newly emerging family of interferon regulatory factors (IRF) (Paun and Pitha, 2007) where IRF-3 and IRF-7 play a critical role in the Type I IFN induction. In fact, genetically modified mice, in which expression of these genes is suppressed, are highly susceptible to viral infection (Sato et al., 2000). IRF-3 plays a key role in the induction of IFNB and amplification of the antiviral response. IRF-3 activity is controlled both by phosphorylation (Au et al., 1995), (Juang et al., 1998), (Fitzgerald et al., 2003), (Sharma et al., 2003) and proteasome mediated degradation (Lin et al., 1998), (Saitoh et al., 2006).

A new dimension of complexity to the virus-mediated innate immune response was recognized by the discovery that many viruses have evolved mechanisms allowing them to overcome the host-induced antiviral response either by direct interference with the pathways leading to the transcriptional activation of interferon genes, or with the IFN-induced antiviral state. Thus, some of the viral genes, not directly required for viral replication, such as the NS1 protein of influenza A (Smith et al., 2001), E3L protein of poxviruses (Hornemann et al., 2003) or the Vp35 protein of Ebola virus (Hartman et al., 2006) are essential for viral pathogenicity in vivo (Hengel, Koszinowski, and Conzelmann, 2005).

The role of the innate antiviral response in HIV-1 infection remains largely unexplored; however, the cellular cytidine deaminase APOBEC3G and several other members of the APOBEC3 family show a potent anti HIV-1 activity (Sheehy et al., 2002). APOBEC3G is packaged into HIV-1 virions and blocks viral replication by inducing cytidine to uracil deamination in the newly transcribed single stranded proviral DNA. Most of the uracil containing transcripts are degraded before integration and viral replication (Mariani et al., 2003), (Bishop, Holmes, and Malim, 2006). However APOBEC3 mediated inhibition of HIV-1 was also shown to be independent on the acitivity of cytidine deaminase and editing (Chiu et al., 2005) (Newman et al., 2005). The HIV-1 encoded accessory protein Vif counteracts function of APOBEC3G by targeting it for degradation by the ubiquitin proteasome pathway (reviewed in (Cullen, 2006; Ehrlich and Yu, 2006). Another accessory protein, Vpr, encapsulated in large amounts in HIV-1 virions, induces degradation of cellular uracil DNA glycosylase (UNG) and its variant SMUG, which removes uracil both from single and double stranded DNA (Schrofelbauer et al., 2005). Thus both Vif and Vpr may inhibit APOBEC3G-induced antiviral effects.

While the expression of APOBEC3G is stimulated by Type I interferon (Chen et al., 2006), analysis of the overall immune response genes expressed during the early steps of HIV-1 infection does not detect expression of IFNA or IFNB genes (Vahey et al., 2002). The infection of human peripheral mononuclear cells with T cell-tropic HIV-1 resulted in an overall increase of genes associated with antiviral immune responses (Vahey et al., 2002), and enhanced expression of chemokines and cytokines was also detected in HIV-infected macrophages; however, induction of Type I IFN or interferon-stimulated genes (ISG) was not observed in either cell type in this study (Vazquez et al., 2005). In contrast, Woelk et al. observed an up-regulation of ISG in HIV-1-infected macrophages, but also did not detect expression of Type I IFN genes (Woelk et al., 2004). Similarly, expression of Type I IFN genes was not detected in HIV-1 infected immature dendritic cells (iDC) (Izmailova et al., 2003). It is not clear how to reconcile the induction of the ISG in the absence of stimulation of interferon synthesis, but it is likely that some of these genes can be directly targeted by HIV-1-stimulated signaling pathways.

Thus, viral control of IRF-3 activity limits not only expression of IFN genes, but also of some ISG. However the activation of IRF-3 by HIV-1 has not yet been defined. The objective of this study was to analyze the activation of IRF-3 during the early steps of HIV-1 infection. While HIV-1 infection did not lead to the activation and nuclear translocation of IRF-3, the relative levels of IRF-3 protein in cytoplasm were decreased during the early stages of HIV-1 infection by ubiquitin-associated proteosome-mediated degradation. Addressing the molecular mechanism of this effect we show that two HIV-1 accessory proteins, Vpr and Vif, target IRF-3 for degradation. Null mutations in these two genes decreased the capacity of the virus to down modulate IRF-3. Thus, Vpr and Vif disrupt an early antiviral response by mediating proteasome degradation of IRF-3.

Results

IRF-3 is not activated during early stages of HIV-1 infection

Binding of HIV-1 virions to its receptors induces several signaling pathways (Popik and Pitha, 1998) (Popik and Pitha, 2000). Binding of HIV-1 envelope glycoprotein gp120 from R5 and R4 viruses to chemokine receptors triggers tyrosine phosphorylation of Pyk2, a focal adhesion component (Davis et al., 1997) and induces calcium signaling (Weissman and Fauci, 1997). In addition HIV-1 mediated signaling activates NFκB and AP-1, which results in aberrant expression of pro-inflammatory genes (Popik and Pitha, 1998), (Briant et al., 1998). The C-terminal phosphorylation of IRF-3 by two IκB kinases, TBK1 and IKKε, is essential for the induction of Type I IFNs and some ISG (Fitzgerald et al., 2003), (Sharma et al., 2003). To investigate whether binding of HIV-1 to its receptors is able to stimulate a signaling pathway leading to the activation of IRF-3, Jurkat T cells were incubated with HIV-1 for 1 h. Since Jurkat cells do not induce significant levels of IFN upon infection with paramyxoviruses such as Sendai virus (SeV) we used HeLa cells as a positive control since SeV induces high levels of Type I IFN in these cells. SeV infection induced various levels of IRF-3 phosphorylation compared to uninfected HeLa cells, here labeled IRF-3PI, PII, and PIII (Fig. 1A). It was previously shown that PII and PIII represent C-terminal serine phosphorylation, whereas PI results from phosphorylation at the N-terminus (Servant et al., 2003), the latter being insufficient for the nuclear translocation of IRF-3. HIV-1 infection of Jurkat cells, while clearly inducing slowly migrating PI form of IRF-3, did not lead to a complete phosphorylation of IRF-3. However, the degree to which SeV and HIV-1 induce IRF-3 phosphorylation could depend on incubation time with the virus. To discount this possibility an infection time course was analyzed (Fig. 1B). Furthermore, in order to ensure that the difference in the IRF-3 phosphorylation pattern following SeV and HIV-1 infection was not a consequence of analyzing two different immortalized cell lines, peripheral blood mononuclear cells (PBMC) were infected with either SeV or HIV-1. PBMC were used rather than CD4+ HeLa cells, since in these cells HIV-1 infection was not very effective and therefore would not allow to differentiate between inhibition of IRF-3 phoispohorylation and lack of activation due to low infection. Furthermore as HeLa cells are not a natural host of HIV-1, results obtained in these cells could not be generalized. In PBMC the majority of the constitutively expressed IRF-3 was present in the PI form; however IRF-3 PII and PIII forms were readily detected within 1–2 h post SeV infection (Fig. 1B). The PII and PIII forms of IRF-3 could be also detected in in HeLa cells at 4 h post SeV infection, however the levels of IRF-3 were significantly lower than at 1 h post infection (Fig1A) due to the virus induced degradation (Lin et al, 1998). In contrast TPA treatment of HeLa cells did not induce IRF-3 phosphorylation beyond the PI stage. HIV-1 infection of PBMC did not induce phosphorylation of IRF-3 beyond the PI form even at 4h post infection. Phosphoserine specific antibody recognizing phosphorylated serine in the 396–405 IRF-3 cluster (Servant et al., 2003) recognized IRF-3 in the SeV infected, but not in the HIV- 1 infected cells. Furthermore, HIV-1 infection did not induced formation of IRF-3 dimers that are specific for the C terminal phosphorylated IRF-3 (Yoneyama et al., 1998) (data not shown). These results indicate that the binding of HIV-1 to its receptors and the early steps of HIV-1 infection does not result in the C terminal phosphorylation of IRF-3. Accordingly, the early phase of HIV-1 infection did not stimulate expression of Type I IFN genes in PBMC and did not activate the synthesis of biologically active IFN (data not shown).

An external file that holds a picture, illustration, etc. Object name is nihms44100f1.jpg

IRF-3 is not activated and translocated to the nucleus during the early stages of HIV-1 infection

(A) Jurkat cells were incubated on ice with HIV-1(NL4-3) (moi 10) for 1 h prior to incubation at 37°C for 1h. HeLa cells were incubated with Sendai virus (SeV) (moi 5) for 2h at 37°C. Moi was calculated as described in Methods. Cells were then washed with PBS and harvested in NP40 lysis buffer. Soluble proteins were separated on 7.5% SDS PAGE, transferred to PVDF membrane, blotted with anti-IRF-3 polyclonal antibody and detected by ECL.

(B). Time course of IRF-3 phosphorylation following virus infection. PBMC were incubated with HIV-1 (moi 5) or SeV (moi 5) in duplicate experiments to that described above. Samples were taken at the times indicated, lysates were prepared and IRF-3 detected as described in A. Samples from control (+), SeV and TPA treated HeLa cells are shown for comparison of gel migration. (C). HIV-1 infection does not induce nuclear translocation of IRF-3. PM1 cells were infected with HIV-1 (AD8) (moi 10) as described in methods. Cells were harvested at the indicated times post infection and the levels of IRF-3 in the nuclear and cytoplasmic fractions were determined by immunoblotting. The equal amounts (15 µg) of protein were loaded on the gel. Sp1 is a nuclear protein therefore its levels in nucleus were determined as a control for equal loading. Levels in cytoplasm were determined as a control for the purity of nuclear fraction and absence of leaking of nuclear protein to cytoplasm. The levels of β actin were determined as controls for equal protein loading.

Translocation of the IRF-3 to the nucleus is another indication of its activation and it is critical for the transcriptional activation of IRF-3 stimulated genes (Lin et al., 1998), (Yoneyama et al., 1998), (Weaver, Kumar, and Reich, 1998). To determine whether IRF-3 is activated at the later stages of HIV-1 infection, we have analyzed the relative levels of nuclear and cytoplasmic IRF-3 in HIV-1 infected monocytic PM1 cell line up to 48 h post HIV-1 infection. Fig 1C shows that while low levels of nuclear IRF-3 could be detected in uninfected cells, no IRF-3 could be detected in the nucleus of HIV-1 infected cells. As the control for the purity of isolated nuclei and protein levels analyzed we show that the levels of the nuclear transcription factor Sp-1 remain constant within 48h post infection and that this factor is detected only in the nucleus and not in the cytoplasm. The majority of the IRF-3 both in infected and unifected cells was present in the cytoplasm (Fig. 1C), however its relative levels were reduced within 2 h and did not reach baseline until after 12 h post infection. The constant levels of β-actin in the cytoplasmic extracts indicate that equal levels of protein were analyzed. Altogether these data indicate that during the early stages of HIV-1 infection, IRF-3 is not activated and does not translocate to the nucleus.

HIV-1 infection induces degradation of IRF-3

Analysis of the relative expression levels of IRF-3 protein in cells infected with HIV-1 (moi10) has shown a decrease in the first 2–24 h post infection (Fig. 1C and ​2A). The quantitative normalization of IRF-3 to β-actin shows that the decrease in the IRF-3 levels is not due to the variation in levels of protein analyzed (Fig.2 insert).The degree of inhibition was dependant on the levels of the input virus infection, when cells were infected with HIV-1 moi1 the inhibition was insignificant (data not shown). No decrease in IRF-3 levels was seen upon binding of HIV-1 to the cells (within 1 h p.i. at 4°C) and was used as a control point.

An external file that holds a picture, illustration, etc. Object name is nihms44100f2a.jpg

An external file that holds a picture, illustration, etc. Object name is nihms44100f2b.jpg

An external file that holds a picture, illustration, etc. Object name is nihms44100f2c.jpg

HIV-1 infection induces degradation of endogenous IRF-3

(A). PM1 cells were pre- treated with polybrene (5ug/ml) for 1 h and infected with HIV-1, (AD8) (moi 10) for 1 h at 4° C. Cells were washed with PBS and incubate at 37C for indicated times. Cells were then collected and lysates were analyzed by Western blot with IRF3, ISG15, HIV-1 and β-actin antibodies. Insert: The density of the IRF-3 bands was scanned, normalized to the density of the actin band and the values of the controls were designated as 100%. The normalized integrated density values were plotted as a function of time. (B) Semi-quantitative RT-PCR of IRF3 transcripts in HIV-1 infected PM1 cells. RNA was isolated at indicated time post HIV-1 infection (moi 10), and amplified using IRF3 specific primers; the cDNA were amplified in a linear region of the PCR curve. The nucleotide sequences of the primers are given in Methods. (C). Degradation of IRF-3 is inhibited by the proteasome inhibitor (MG132). PM1 cells were treated with the proteasome inhibitor MG132 (5uM) for 2h, infected with HIV-1 (moi 10) and then incubated in the presence of MG132 for indicated times. Cells were then collected, lysed and analyzed by Western blot for the presence of IRF-3, HIV-1 proteins and β actin as described in Methods. (D). Azidothymidine (AZT) treatment does not prevent degradation of IRF3 in HIV-1 infected cells. PM1 cells were pre-treated with AZT (10 µM) for 12 h and then infected with HIV-1. At indicated time post infection, cell lysates were immunoblotted with IRF3, HIV-1 and β actin antibodies. Control samples were pre-treated with AZT for 12h and than treated with AZT for additional 12 hr without HIV-1 infection. The 24KDa band represent the in input virion p24.

The expression of the ubiquitin like protein ISG15 was increased as early as at 6 h post infection and its conjugation to cellular proteins was clearly seen at 24 h post infection. (Fig. 2A). Since ISG15 is induced both by IFN and activated IRF-3, these data indicate that HIV-1 induced expression of ISG is not mediated by activated IRF-3 or type I IFN (Haas et al., 1987). The analysis of HIV-1-encoded proteins in infected cells has shown decreasing levels of the virion p24 protein, while newly synthesized Pr55 was detected only at 48 h post infection. These data indicate that the decrease in the IRF-3 levels occurred prior to de novo synthesis of Gag protein (Fig. 2A).

The decrease in relative levels of IRF-3 could be due to the inhibition of IRF-3 transcription or an increase in protein turnover. There was no decrease in the relative levels of IRF-3 transcripts in HIV-1 infected cells early after infection, as determined by semi-quantitative PCR analysis (Fig. 2B). Instead, an increase of IRF-3 transcripts was observed at 24 h post infection coinciding with the expression of the early non structural HIV-1 protein nef (data not shown). We then tested whether the decrease in IRF-3 was due to proteosomal degradation by determining the effect of the proteosome inhibitor MG132 on the HIV-1-induced reduction of IRF-3 levels. There was no degradation of IRF-3 in HIV-1-infected cells treated with MG132 (at a concentration that did not affect viability of the cells) (Fig. 2C), while the overall relative level of IRF-3 was increased, indicating that constitutively expressed IRF-3 in the cells is regulated by proteosome degradation and that HIV-1- mediated reduction of IRF-3 is also mediated though proteosome degradation. The analysis of HIV-1 proteins shows a significant inhibition of Pr55 synthesis (compare Fig. 2A and 2C), which indicates that HIV-1 replication is inhibited in the presence of proteasome inhibitors (28).

To determine whether replication of the input viral genome is required for IRF-3 degradation, PM1 cells were pretreated with AZT before HIV-1 infection and the relative levels of IRF-3 were analyzed for 48h post infection. AZT treatment effectively inhibited HIV-1 replication and no synthesis of Pr55 could be detected (Fig. 2D) while in the absence of AZT Pr55 was detected as early as 48 h post infection (Fig. 2A). The AZT treatment did not prevent IRF-3 degradation instead, blocking of provirus formation and HIV-1 replication appeared to enhance reduction of IRF-3. The reason for enhancement of IRF-3 degradation in AZT treated cells is not clear, but it is not a consequence of toxic effect of AZT, since the relative levels of IRF-3 were not affected by AZT in uninfected controls. (Fig.2D). These data indicate that the proteosomal degradation of IRF3 is mediated by a component of HIV-1 virion, and does not require virus replication.

Vif and Vpr independently target IRF-3 for ubiquitin- mediated degradation

Two HIV-1 accessory proteins, Vif and Vpr, were previously shown to induce degradation of distinct cellular proteins. Vif targets APOBEC3G and 3F for proteasome degradation though E3 ubiquitin ligase complex (SCF) containing Vif, Cul5, and Elongin B/C (Yu et al., 2003) (Wiegand et al., 2004) (Marin et al., 2003), (Mehle et al., 2004a). To determine whether Vif can also induce IRF-3 degradation, IRF-3 was expressed in the presence of increasing levels of Vif (Fig. 3A). The steady–state levels of IRF-3 were reduced by an increasing concentration of Vif; however, even high levels of Vif were not able to completely eliminate IRF-3 expression. Vif-mediated IRF-3 degradation was inhibited in the presence of MG132. When the levels of IRF-3 were normalized to the levels of β actin, the inhibition show a concentration dependence on the input Vif.

An external file that holds a picture, illustration, etc. Object name is nihms44100f3a.jpg

An external file that holds a picture, illustration, etc. Object name is nihms44100f3b.jpg

HIV-1 accessory proteins Vif and Vpr independently target IRF3 for ubiqutin mediated degradation

(A) 293T cells were co-transfected with IRF3 (0.5µg) and indicated amounts of Vif plasmid (0–10µg). The total amount of transfected DNA was held equivalent in all samples by using cDNA3 plasmid DNA. 12 h after the transfection, cell lysates were analyzed by Western blot with IRF-3, Vif and β actin polyclonal antibodies. The IRF-3 bands were scanned, normalized to a density of the β actin band and the values are expressed as % of the normalized IRF-3 expressed in the absence of Vif that was nominated as 100%. (B) 293T cells were co-transfected with IRF3 and the Vpr plasmid of indicated concentrations. The samples were analyzed for the presence of IRF-3 and Vpr and the relative levels of normalized IRF-3 were determined as described above Controls were transfected with the empty vector, pcDNA3 only. Western blots show representative data of one from three independent experiments. The error bars on the scan graph show variations between the individual experiments. Neither Vif nor Vpr induced IRF-3 degradation in the presence of proteasome inhibitor MG132.

Vpr was shown to bind to UNG and SMUG and induce their proteosome-mediated degradation by interacting with E3 ligase components and forming complexes with Cul4 (Schrofelbauer et al., 2005). Like Vif, expression of Vpr induced degradation of ectopic IRF-3. The normalization of the IRF-3 levels to β actin levels has shown that Vpr induced degradation of IRF-3, however the degradation was less effective than the degradation by Vif. (Fig. 3B). In the presence of proteosome inhibitor MG132 neither Vif nor Vpr degraded IRF-3.

The Vif and Vpr mediated degradation was not limited to ectopic IRF-3 as the levels of endogenous IRF-3 were also decreased in cells expressing Vif or Vpr (Fig. 4A). The Vif-and Vpr-mediated degradation of APOBEC3G or UNG respectively, was shown to be associated with the polyubiquitination of these proteins. We have therefore examined whether Vif and Vpr-induced ubiquitination of IRF-3. 293T cells were co-transfected with HA-tagged ubiquitin, IRF-3, and Vif or Vpr expression plasmids. 24 h after the transfection cell lysates were immunoprecipitated with IRF-3 antibodies and the presence of IRF-3 and ubiquitinated IRF-3 was detected by immunoblotting (Fig. 4B). While in the absence of Vif or Vpr, IRF-3 was detected as single band with mobility of 55 KDa, in cell expressing either Vif or Vpr an additional strong, slowly moving high molecular band and weak 65 KDa bands were detected (Fig. 4B). To confirm that the high-molecular forms are ubiquitinated IRF-3, the blots were reacted with anti HA antibodies. The ubiquitin analysis detected high molecular form of IRF-3 but no 50 KDa IRF-3 indicating that both Vif and Vpr induce polyubiquitination of IRF-3 (Fig. 4B).

An external file that holds a picture, illustration, etc. Object name is nihms44100f4a.jpg

An external file that holds a picture, illustration, etc. Object name is nihms44100f4b.jpg

Vif and Vpr induce degradation of endogenous IRF-3 and its ubiquitination

(A). Vif and Vpr induce dose dependant degradation of the endogenous IRF3. 293T cell were transfected with Vif or Vpr expressing plasmids and 12 h post transfection cells were harvested and cell lysates analyzed by immuno blotting with IRF-3, Vif or Vpr and β actin polyclonal antibodies (B). 293T cells were transfected with IRF-3 (1µg) and Vif or Vpr(2 µg) and HA tagged ,ubiquitin expressing plasmids-UbHA (1µg). The total amount of transfected DNA was adjusted by pcDNA3 plasmids DNA. 24 h after the transfection, cell lysates were immunoprecipitated with IRF-3 polyclonal antibodies and the precipitates were immuno blotted with monoclonal IRF-3 and HA antibodies. Controls were transfected with pcDNA3 only.

Mapping of the IRF-3 lysine residues that are important for the Vif and Vpr mediated degradation

Human IRF-3 contains 14 lysine residues, eleven of which are located in the N-terminal half of the protein. To determine whether the amino terminal peptide is susceptible to Vif and Vpr mediated degradation, we co-transfected Vif or Vpr with IRF-3 or its deletion mutants encoding aa residues 1-269 or 1-329, respectively. All IRF-3 plasmids were well expressed in transfected cells and co-transfection with Vif or Vpr resulted in reduction of relative levels of full size ectopic IRF-3 as well as the two N terminal peptides (Fig. 5A). These data indicate that it is the N terminal region of IRF-3, which contains signals required for the proteosomal degradation. It should be noted that in these experiments the degradation of full length IRF-3 was insignificant while the N terminal peptide were degraded very effectively. These data indicate that unphosphorylated full length IRF-3 is degraded less effectively than the short IRF-3 peptides. The β-actin levels were comparable in all samples indicating equal loading.

An external file that holds a picture, illustration, etc. Object name is nihms44100f5a.jpg

An external file that holds a picture, illustration, etc. Object name is nihms44100f5b.jpg

Mapping of IRF3 ubiquitination domain by Vif and Vpr

(A) 293T cell were co-transfected with IRF-3 (1ìg) or its deletion plasmids encoding 1– 269 and 1–329 aa long IRF-3 peptides and Vpr or Vif expressing plasmids at the 1:1 or 1:10 IRF3: Vif or Vpr ratio. The total amount of transfected DNA was kept constant. Cells were harvested at 24 h post transfection and the relative levels of IRF-3, Vif and Vpr (left- central panel) in cell lysates were determined by immune blotting with specific antibody. The relative levels of actin were determined as controls for equal protein loading (B) IRF-3 (1µg) and its respective K-R mutants were co-transfected together with Vif plasmid (2 µg) to 293T cells. Cell lysates were prepared 24 h later and relative levels of IRF-3 and Vif determined by western blotting. (C) Transfection of IRF-3, its K-R mutants and Vpr was done as described in B. The levels of β- actin were determined as controls for protein loading.

To further identify the lysine residues that are critical for the Vif and Vpr mediated IRF-3 degradation we mutated 6 N terminal lysines (residues 39, 77, 87, 105, 193) and one C terminal lysine (K409) to arginine. All mutated IRF-3 plasmids were well expressed in transfected cells (Fig. 5B). When the mutated plasmids and wt IRF-3 were co-transfected together with the Vif expressing plasmid, all but the K87R and K105R mutants were degraded (Fig. 5B). The immunoblot with anti-Vif antibodies shows that expression of Vif was comparable in all transfected samples. These data indicate that the K87 and K105 play an important role in the Vif mediated proteasome degradation. Analysis of the β-actin levels shows no major differences in protein l loading between the samples.

Co-transfection of the IRF-3 mutants with Vpr gave a slightly different picture. Only the levels of wt IRF-3 and its K193R and K409R mutants were effectively reduced by Vpr while mutation of K39, K77R, K87, and K105 made IRF-3 resistant to Vpr-mediated degradation. These data suggest that although both Vif and Vpr target IRF-3 for proteosomal degradation, the IRF-3 lysines targeted by the Cul4 and Cul5 ubiqitin ligases may be distinct. The analysis of the relative levels of β-actin in all samples show that the observed differences are not consequence of a difference in protein loading.

HIV-1 mutants containing disrupted Vif and Vpr

To determine whether Vif and Vpr are essential for IRF-3 degradation in the context of viral infection we used mutants of the pNL4-3HSA HIV-1 containing stop codon mutations in either Vif or Vpr ORFs prematurely terminating Vif and Vpr translation. (Sakai, Dimas, and Lenardo, 2006). Jurkat cells were infected with wt HIV-1, VSV-G pseudotype and the HIV-1 mutants and the HIV-1 proteins were detected by immune detection. It can be seen in Fig.6a that while both Vif and Vpr can be detected in cells infected with wt HIV-1 or HIV-1VSV pseudotype, the insertion of the stop codon into the respective accessory genes eliminats their expression (Fig. 6A). The relative levels of IRF-3 were determined in Jurkat cells infected with wt HIV-1 (pNL4-3HSA), HIV-1-VSV pseudotype or the HIV-1-VSV mutants lacking expression of Vif, Vpr, or both of these genes (Fig. 6B). Western blot of the representative infection shows that the inactivation of either Vif or Vpr gene alone is not sufficient to inhibit IRF-3 degradation. The silencing of both Vif and Vpr genes in the HIV-1 genome slowed down IRF-3 degradation in the initial stages of infection. There was no degradation in cells infected with the double deleted mutants of HIV-1 at 2h and small level of degradation at 6h post infection, while at 12 h post infection the IRF-3 levels were similar to those detected in cells infected with wt HIV-1 or the signle mutants. To determine whether the observed difference between the effect of the double deleted mutant HIV-1 and the wt or single deleted mutants was significant and reproducible, the relative intensity of IRF-3 from three independent infections was normalized to the intensity of β actin and expressed as function of time. The analysis has shown that the IRF-3 degradation was slow down during the initial stages of infection with the Vif and Vpr deleted mutant, when compare to the infection with wt HIV-1 or single deleted mutants. These data indicate that while over expressed ectopic Vif or Vpr alone are sufficient to induce proteasome degradation of IRF-3, in the context of viral particle the levels of Vif and Vpr are much lower and only mutation in of both Vif and Vpr can stabilize IRF-3, but the stabilization effect diminishes with the progression of viral infection. We have also tested whether the Vif and Vpr deleted HIV-1 mutants can activate IRF-3 and its nuclear translocation. However our results indicate that the infection of PM1 cells with the Vif and Vpr deletion mutants neither induces IRF-3 dimerization or nuclear translocation. These data suggest that the lack of IRF-3 activation in HIV-1 infected cells is not due to the Vif and Vpr induced degradation of IRF-3.

An external file that holds a picture, illustration, etc. Object name is nihms44100f6a.jpg

An external file that holds a picture, illustration, etc. Object name is nihms44100f6b.jpg

Degradation of IRF3 is slow down in cell infected with HIV-1 mutant lacking both Vif and Vpr

(A) Analysis of the viral proteins expression encoded by pNL4-3HSA modified by site-directed mutagenesis in Vif and Vpr genes. The pNL43 DNA or pNL4-3HSA DNA (2 µg) and its Vif and Vpr deletion mutants were co-transfected together with the pMDG plasmid (1 µg) encoding VSV-G protein to 293T cells and 36 h post transfection, cell lysates were analyzed by Western blotting with HIV-1 antiserum and Vif or Vpr polyclonal antibodies. The Env- Vector cotransfected with pMDG served as a negative control. (B) Jurkat cells were infected with Env- HIV-1 (NL43HSA) VSV pseudotype and VSV psedotyped HIV-1 mutants (moi 10) lacking expression of either Vif, or Vpr or both of these genes. Cells were harvested at indicated times post infection and the lysates were immunoblotted with polyclonal IRF3 antibody. Levels of β actin were determined as 26 controls for protein loading. Western blot of a representative experiment is shown. The density of the IRF-3 bands was scanned, normalized to the density of the actin band and the normalized integrated density values were plotted as a function of time. The data represent analysis of three independent infections.

Impairment of IRF-3 mediated activation

Our results indicate that Vif and Vpr induce proteasome mediated degradation of constitutively expressed IRF-3. In cells infected with paramyxoviruses or treated with dsRNA, activated IRF-3 is part of a multi-component transcriptional complex-enhanceosome containing IRF-3, IRF-7, NFκB, and AP-1 assembled on the promoter of the IFNB gene (Wathelet et al., 1998). To determine whether Vif and Vpr inhibit the IRF-3 mediated activation of the IFNB promoter, we have analyzed the IRF-3-mediated stimulation of the IFNB promoter in NDV infected cells expressing Vif or Vpr. 293T cells were co-transfected with IRF-3 and a reporter plasmid expressing the luciferase gene under the control of IFNB promoter in the presence or absence of Vif or Vpr expressing plasmids. Transfected cells were infected with NDV and the relative levels of IRF-3 and the luciferase activity was measured 16 h after the infection. As shown previously, NDV-mediated activation of the IFNB promoter was increased by ectopic IRF-3 (Schafer et al., 1998), (Juang et al., 1998) but this increase was nearly abolished in the presence of Vif or Vpr (Fig. 7A). These results indicate that Vif and Vpr disrupted the transcriptional activity of IRF-3. However, it is unlikely that the Vif- or Vpr-mediated inhibition of the _IFN_B promoter is entirely due to the degradation of IRF-3 since both the ectopic and endogenous IRF-3 could still be detected in cells transfected with Vif or Vpr (Fig. 7A). We therefore examined whether in HIV-1 infected cells the NDV mediated activation of IRF-3 pathway is impaired. NDV, like SeV, activates IRF-3 by the RIG I pathway (Yoneyama et al., 2004), which leads to activation of IRF-3 and IRF-7 and increases expression of the _IFN_B gene. Since HIV-1 does not infect CD4 negative cells, IRF-3 activation was analyzed in 293T cells transfected with HIV-1(NL4-3) proviral DNA. Transfection of HIV-1 provirus into these cells leads to a single cycle infection, expression of HIV-1 proteins and release of virions (Okumura et al., 2006). The results in Fig. 7B show that NDV mediated activation of IRF-3 signaling pathway was impaired in cells transfected with HIV-1. Interestingly, while HIV-1 alone did not stimulate the _IFN_B promoter in these cells, the luciferase levels were higher in NDV-infected cells expressing NL4-3 than in NDV infected cells. Thus the previously demonstrated activation of NFκB and AP-1 by HIV-1 infection may contribute to the NDV mediated activation of the IRF-3 pathway.

An external file that holds a picture, illustration, etc. Object name is nihms44100f7a.jpg

An external file that holds a picture, illustration, etc. Object name is nihms44100f7b.jpg

Vif and Vpr inhibit IRF3 mediated activation of IFNB promoter

(A) 293T cells were co-transfected with IRF-3 (1µg) and IFNB promoter–luciferase reporter plasmids (0.5 µg) in the presence or absence of Vif or Vpr expressing plasmids (1 µg). Cells were infected with NDV (moi 5) 12 h after the transfection and harvested for luciferase activity 16 h post infection. Renila construct was used as control for transfection efficiency. Luciferase activity was analyzed as described in Methods. (B) 293T cells transfected with indicated amounts of NL4-3 proviral DNA (1 µg), were transfected with IRF-3 expressing plasmid (1ug) and _IFN_B luciferase reporter plasmid (0.5µg). When indicated, cells were also infected with NDV (moi 5) 12 h after IRF-3 transfection and harvested and analyzed as described in A.

Discussion

The activation of IRF-3 during the early steps of viral replication is a critical step in the initiation of the cellular antiviral response. The results of this study show that the early steps of HIV-1 infection do not lead to the activation of IRF-3 and its nuclear translocation, but instead target the endogenous cytoplasmic IRF-3 for ubiquitin associated proteosome degradation. The virion associated HIV-1 accessory proteins Vpr and Vif play a central role in IRF-3 degradation and the elimination of Vpr and Vif genes in an HIV-1 provirus decreased the capacity of the virus to down modulate IRF-3 levels at the early stages of viral infection. We have shown in this study that Vpr and Vif disrupt the initial antiviral response by mediating the proteasome-dependent degradation of IRF-3. Two specific lysine residues (K87 and K105) in the N-terminal region of IRF-3 play a critical role in Vif mediated degradation, while Vpr-mediated degradation was dependant on K39, K77, K87, and K105. Further analysis will be needed to determine whether this N-terminal region of IRF-3 represents the IRF-3 degron that targets activated IRF-3 for ubiquitin mediated degradation (Varshavsky, 1997). It was shown recently that Pin1 induces a selective degradation of the activated, but not the constitutive IRF-3 (Saitoh et al., 2006). In contrast, Vif- and Vpr-mediated degradation of IRF-3 was not dependent on its activation by C-terminal phosphorylation.

This study shows that Vif-mediated protein degradation is not limited to APOBEC3G but affects another antiviral factor, IRF-3. The degradation of APOBEC3G is mediated by direct binding of Vif to APOBEC3G, which though its zinc-binding motif also binds to the Cullin (Cul)-dependent ubiquitin ligase SCF complex containing Cul5, Elongins B/C and the Ring protein Rbx (Yu et al., 2003), (Mehle et al., 2004b). Vpr binds UNG and induces ubiquitination and proteasome degradation of UNG by forming a complex with Cul4 (Schrofelbauer et al., 2005). However, while Vif and Vpr induced ubiquitination of IRF-3, direct binding of Vif or Vpr to IRF-3 was not detected (data not shown). Both Vpr and Vif are incorporated into HIV-1 virions. Vpr is effectively packaged into the virions via interaction with p6 domain of Gag. According to a recent estimate the ration of Vpr and capsid protein is 1:7 and HIV-1 particle contains about 275 molecules of Vpr and 1800 molecules of capsid (Muller et al., 2000). Vif is packaged into virions with genomic RNA as a nucleoprotein complex. While Vif is efficiently packaged into virions from acutely infected cells (60–100 copies per virion), packaging from chronically infected cells is very low (4–6 copies per virion) (Kao et al., 2003). Thus while the levels of viron associated Vpr are relatively high, the levels of virion associated Vif depend on type of infection and can be substantially lower.

A number of viruses redirect the cellular ubiquitination machinery to target the antiviral defense mechanisms. The specificity of ubiquitin conjugation to cellular proteins is determined by E3 ubiquitin ligase that selects the substrate as well as the multi-protein complexes that contain the cullin Ring ubiquitin ligase. The cullin Ring ligases are activated by nedylation (conjugation of Nedd8) and by the Ring protein Roc1 that interacts with the ubiquitin ligase E2 (reviewed in (Barry and Fruh, 2006). Recruitment of the substrate to cullin ligase occurs though specific linkers and adaptors that interact with cullin. Cullin 5 interacts with elongin B and C that recognize the SOCS box in the substrate adaptors. This mechanism is used by Vif to eliminate APOBEC3G (Yu et al., 2003). Vpr was shown to associate with cullin 4 (Schrofelbauer et al., 2005), which interacts with DDB 1 (DNA damage binding protein 1). The simian virus 5 encoded V protein, which like Vif contains a zinc- binding domain, interacts with cullin 4–DDB1 complex and targets STAT-1 for degradation. On the other hand, paramyxovirus V protein targets STAT2 and mumps V protein induces degradation of both STAT1 and STAT3 (rev in (Horvath, 2004). The fact that both Vif and Vpr target IRF-3 for degradation, presumably by association with cullin 5 and cullin 4 respectively, indicate that the functions of the cullin containing complexes may be multifunctional. Additional experiments are in progress to identify the linker and the adaptor that interact with IRF-3 in cullin containing complexes.

Constitutive over-expression of ectopic IRF-3 in T cells and monocytes prior to infection inhibits de novo HIV-1 replication indicating that the disruption of IRF-3 pathway enhances HIV-1 replication (unpublished data). However, the decrease in IRF-3 was observed only during the first 24 h post infection, suggesting that during the later course of infection cells, counteract the Vif-and Vpr-mediated degradation of IRF-3. The increase in the relative levels of IRF-3 transcripts coincides with the synthesis of HIV-1 proteins, suggesting that IRF-3 transcription may be stimulated by an HIV-1 gene product. We have also shown recently that the ubiquitin-like protein ISG15 inhibits proteosome mediated degradation of cellular proteins including IRF-3 and intervenes with their ubiquitination (Lu et al., 2006). Whether ISG15, which is induced in HIV-1 infected cells and conjugated to cellular proteins, attenuates the Vif- and Vpr-mediated degradation is under investigation.

In conclusion these results indicate that HIV-1, like other RNA viruses developed specific mechanism to down regulate the early antiviral response (Basler and Garcia-Sastre, 2002). IRF-3 constitutes a component of the innate immune system that induces expression number of cellular genes, including Type I IFN genes and Rantes (Lin et al., 1999). Vpr and Vif are conserved in primary virus isolates and are incorporated in virions. The finding that virion proteins Vif and Vpr induce degradation of IRF-3 in a co-operative manner, points out to the need of HIV-1 to circumvent the antiviral response during the very early phase of viral replication. Our results also suggest that the antiviral compounds targeting the Vif- and Vpr- induced ubiquitination, will not only block viral replication, but also restore the basal levels of IRF-3 and rescue the activity of the IRF-3 mediated pathway.

Material and Methods

Cell Culture, Plasmids and Virus

293T cells were cultured in DMEM with 10% FBS. PM1 cell line, which is CD4+CXCr4+CCR+ derivative of Hut79T cells (Lusso et al., 1995), was cultured in RPMI medium 1640 with 10% FBS and 5 mM L-glutamine. The peripheral blood leucocytes (PBMC) were prepared from buffy coats by FicolPaque Plus gradient separation (Amersham Pharmacia Biosciences) and cultured at 1 × 106 cells/ml in RPMI 1640 with 10% FCS, 2 mM glutamine (Invitrogen Life Technologies). They were treated with 25 mM 5 phorbol 12-tetradecanoate 13-acetate (TPA) for 24 h before HIV-1 infection.

The Vif and Vpr expression plasmid were described (Mehle A and Gabudza, D 2004, Paxton ,W and Landau N 1993). The hemagglutinin (HA)-tagged Ub plasmid (Ub- HA) was obtained from Dr. H. Gottlinger (Harvard Medical School, Boston) and pCMV IRF3 was described (Au et al 1995). pNL4–3HSA Vif-, pNL4-3HSA Vpr- and pNL4-3HSA Vif- Vpr- plasmids were kindly obtained from Dr. M. Leonardo (National Institute of Health, Bethesda) (Sakai, Dimas, and Lenardo, 2006). The macrophage-tropic HIV-1 AD8 was obtained from Dr Martin (National Institutes of Health, Bethesda).

Infectious virus was produced and purified as described previously (Okumura et al., 2006). The pLN43 virus was titrated on Jurkat cells and AD8 was titrated on PM-1 cells. Moi of infection was calculated as number of cpm (determined by RT assay)/ cell or by number of TCID50/cell. For the HIV-1 infection, cells were incubated with HIV-1 (moi 10, unless specified otherwise)) for 1 h at 4° C, than extensively washed and incubated for an indicated time at 37° C.

NL4-3HSA Vif-, pNL4-3HSA Vpr- and pNL4-3HSA Vif- Vpr- VSV Pseudotype virus

NL4-3HSA Vif-, pNL4-3HSA Vpr- and pNL4-3HSA Vif- Vpr- pseudotyped virus stocks were produced in 293T cells and pseudo typed by using vesicular virus (VSV) G protein to replace Env (Burns et al., 1993). To this effect HIV-1 plasmids were co-transfected with pMDG plasmid encoding VSV G protein, to 293T cells and the virus containing supernatants were harvested 48 h after transfection. Cellular debris was removed by centrifugation and virus was purified as described before (Okumura et al 2006). Virus titers were determined by RT assay (Vicenzi and Poli, 1994) and moi calculated as described above.

Transfection, Immunoprecipitation, and Western Blot Analysis

For transfection, 293T cells were transfected with IRF-3 expressing plasmid (0.5–1.0µg) and indicated amounts of Vif or Vpr plasmids (1– 10 µg) and when indicated ubiquitin expressing plasmid (2 µg) using Lipofectamine Plus (In vitrogene). Total amount of plasmid DNA transfected was kept constant in all samples by using empty pcDNA3 vector. At indicated time post transfection cells were collected and cell lysate analyzed by western blot or immunoprecipitated using specific antisera. The protein concentration in cell lysates was determined by the Biorad protein assay.

The IRF3 and hemagglutinin (HA) antibodies were purchased from Santa Cruz Biotechnology. Human serum, which detects HIV-1 proteins, was a generous gift from Dr. M. Martin. The Western blot analysis and immunoprecipitations of cell lysates were described recently (Okumura et al., 2006). Briefly, cells were lysed in lysis buffer (20 mM Tris-HCl/150 mM NaCl/10 mM EDTA/0.5% Nonidet P-40/PMSF, and proteinase inhibitor mixture). For the immunoprecipitation, 1 mg of protein was incubated for 30 min on ice with 20 µl of protein G-agarose (Invitrogen) in a total volume of 1 ml. The mixture was centrifuged at 1,000 g for 5 min to preclear the lysate of proteins binding nonspecifically to the matrix. The precleared supernatants were incubated with specific antibody (2 mg/ml) for 1 h on ice, and then 20 µl of protein G plus-agarose was added to the immune complexes for an additional 1 h on ice. The bound proteins were pelleted at 1,000 g for 5 min and washed with lysis buffer thee times. The final pellets were resuspended in 40 µl of sample buffer and boiled for 5 min before loading on a 10%–13% SDS/PAGE gel. The co-precipitated proteins were identified by Western blotting. For Western blot analysis, cells were lysed in buffer (20 mM Tris-HCl, pH 7.5,150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1% Triton, 2.5 mM sodium pyrophosphate and protease inhibitor mixture) and the lysates were clarified by centrifugation at 10,000 g for 10 min. Proteins (10–50 µg) were separated on 10 or 13% SDS-polyacrylamide gels and electrotransferred to the membranes (Bio-Rad). The membranes were blocked in 5% nonfat milk for 30 min at room temperature and then subsequently incubated with primary antibody for 30 min in blocking buffer and with the horseradish peroxidase-conjugated secondary antibody (1:10,000 dilutions; Amersham Pharmacia). Immunodetection was realized by ECL reagents (Amersham Biosciences) and autoradiography on hyperfilm MP (Amersham Biosciences).

Luciferase assay and RT PCR analysis

For the luciferase assay 293T cells were transfected with IRF-3 expressing plasmid (1 µg) using Lipofectamine 2000 (Invitrogen) and the IFNB promoter reporter – luciferase reporter plasmid(0.5 µg) and Vpr or Vif expressing plasmid (1µg). Alternatively cells were transfected with IRF-3 (0.2 µg), _IFNB_-luciferase reporter(0.5 µg) and pNL43 proviral DNA at 1:1 or 1:10 IRF3: pLN43 ratio. Renila plasmid was used as control for transfection efficiency The final concentration of transfected DNA was kept constant in all co-transfection assays. Transfected cells were infected with NDV 12 h post transfection and cells were lysed, 16 h later. Luciferase activity was measured as described previously (Lu et al., 2000), all samples were normalized to the constant level of Renila luciferase.

Semi-quantitative RT-PCR analysis was described previously (Au, Yeow, and Pitha, 2001).The IRF-3 specific primers used were: 5’ primer CTGAAGCGGCTGTTGGTG and 3’ primer ACCATGAGGAGCGAGGGC

Acknowledgements

We thank H. Gottlinger and M. Leonardo for providing us with the UbHA plasmid and the mutants of pNL4-3HSA provirus DNA respectively and M. Martin for the anti HIV-1 serum and the AD8 virus. We also wish to thank Dr. A. Paun for the preparation of the figures for publication. This research was supported by the NIH-NIAID AI054276-01A1 grant to P.M.P.

Abbreviations

SeV Sendai virus
NDV Newcastle Disease virus
ISG interferon stimulated gene
AZT azidothymidine
UNG uracil DNA glycosylase
PBMC peripheral blood leucocytes

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Conflict of interest statement. No conflicts declared.

References