Nuclear Factor 90, a cellular dsRNA binding protein inhibits the HIV Rev-export function

Abstract

Background: The HIV Rev protein is known to facilitate export of incompletely spliced and unspliced viral transcripts to the cytoplasm, a necessary step in virus life cycle. The Rev-mediated nucleo-cytoplasmic transport of nascent viral transcripts, dependents on interaction of Rev with the RRE RNA structural element present in the target RNAs. The C-terminal variant of dsRNA-binding nuclear protein 90 (NF90ctv) has been shown to markedly attenuate viral replication in stably transduced HIV-1 target cell line. Here we examined a mechanism of interference of viral life cycle involving Rev-NF90ctv interaction.

Results: Since Rev:RRE complex formations depend on protein:RNA and protein:protein interactions, we investigated whether the expression of NF90ctv might interfere with Rev-mediated export of RRE-containing transcripts. When HeLa cells expressed both NF90ctv and Rev protein, we observed that NF90ctv inhibited the Rev-mediated RNA transport. In particular, three regions of NF90ctv protein are involved in blocking Rev function. Moreover, interaction of NF90ctv with the RRE RNA resulted in the expression of a reporter protein coding sequences linked to the RRE structure. Moreover, Rev influenced the subcellular localization of NF90ctv, and this process is leptomycin B sensitive.

Conclusion: The dsRNA binding protein, NF90ctv competes with HIV Rev function at two levels, by competitive protein:protein interaction involving Rev binding to specific domains of NF90ctv, as well as by its binding to the RRE-RNA structure. Our results are consistent with a model of Rev-mediated HIV-1 RNA export that envisions Rev-multimerization, a process interrupted by NF90ctv.

Figure 1

Schematic representation of (A) pCMV128 (CAT-RRE reporter), and (B) the NF90ctv regions investigated here. (C) The NF90ctv domains were cloned into pCI-neo and used in this study. The vertical arrows correspond to the SH3 motifs present in RG- region.

Figure 2

Effect of NF90ctv protein domains on the level of expression of CAT. HeLa cells were cotransfected with CAT-RRE, and with different constructs that code for Rev protein and the NF90ctv domains. After cell lysis, ¹⁴C-labelled chloramphenicol and acetyl CoA were added to the extracts and the CAT activity determined. The construct included in each assay is indicated above the lanes. Vertical arrows highlight the constructs leading to the strongest decrease in acetylation of chloramphenicol. Similar results were obtained in three independent experiments. When the cells were transfected only with the empty expression vector for NF90 (pCI-neo), acetylation of chloramphenicol was not observed (not shown).

Figure 3

Immunoprecipitation. HeLa cells were cotransfected with pRSV/Rev and pOZ-Flag/NF90; 24 h post-transfection, the cells were lysed, α-Flag antibodies were added and the proteins precipitated by adding protein A agarose. The samples were analyzed by SDS-PAGE, and the proteins transferred to a PVDF membrane were immunodetected using α-Rev_31–50 antibodies. Lane 1, extracts from HeLa cells cotransfected with both constructs. Lane 2, extracts from untransfected HeLa cells. Lane 3, same amount of HeLa cell extract prior to immunoprecipitation (input control). Arrow indicates the protein band (~28 kDa) corresponding to Rev.

Figure 4

Interaction between the NF90ctv RG- domain and Rev by affinity chromatography. (A) The GST/RG- recombinant protein was expressed in E. coli, the cell extracts coupled to a glutathione-Sepharose 4B column was used in pull-down assays with HeLa cell extracts previously transfected with pRSV/Rev (lane 4) or the control lysate (lane 3). A protein band corresponding to Rev (arrowhead, lane 4) was detected when extracts from HeLa cells that expressed Rev were added to the GST/RG- protein bound column; such a band was absent from control HeLa cell extract (lane 3). Lane 1, GST/RG- purified from E. coli induced by IPTG, and lane 2, purified from non induced E. coli cells. (B) Similar assays performed with purified Rev protein from E. coli in place of HeLa cells extracts. Arrow in lane 5 indicates the position of Rev. This band was not observed in absence of Rev (lane 6).

Figure 5

Subcellular localization of the Rev and NF90ctv proteins observed by fluorescence microscopy in confocal optical sections. HeLa cells were transfected with pGFP/NF90 or pGFP/Rev and GFP was visualized with an Argon laser at 488 nm. NF90 localized in the nucleus and nucleolus; Rev localized in the nucleus, nucleolus and cytoplasm (panels a and b). To study the effect of NF90ctv on Rev localization, HeLa cells were cotransfected with both expression vectors and the proteins were observed in confocal optical sections; the mRFP was visualized using a Krypton laser at 568 nm. Both protein colocalized in the similar subcellular compartments (c'-c''; d'-d'' are the same as c'-c'''). In the presence of 50 ng/ml of LMB and both proteins localized mostly in the nucleolus. Bar = 5 μm.

Figure 6

NF90ctv affects structured-RNA translocation. HeLa cells were cotransfected with pRev/mRFP and pSGT-5(SDM/RRE/CM-GFP) or with pNF90/mRFP and pSGT-5(SDM/RRE/CM-GFP). The constructs represent HIV-1 Gag and GFP genes linked to RRE, making their expression dependent on Rev. The reporter gene (GFP) linked to RRE is expressed either in presence of Rev or NF90ctv (a and b, respectively). As controls HeLa cells were transfected with each construct alone. In presence the RRE, Rev localized mainly in the cytoplasm (a), colocalized with the GFP reporter (a''). However, NF90ctv in the presence of RRE, preferentially localized in the nucleus/nucleoli (b), and GFP reporter was distributed in the cytoplasm and in the nucleus.

Figure 7

Model for regulation of Rev protein by NF90. The HeLa cells are cotransfected with the respective plasmids. In (a), the transcript of the gene linked to RRE is recognized by Rev and forms a complex with CRM1, it is exported to the cytoplasm, where it is translated by the ribosome. In (b) based on the data presented herein, we suggest that NF90 is able to recognize the transcript of the gene linked to RRE, and then it is exported to the cytoplasm and translated. In (c) we suggest that this process is CRM1-dependent. Because our immunoprecipitation results show that NF90 interacts with Rev and their colocalization in the cytoplasm is CRM1-dependent, In (d) we propose that the affinity of Rev or NF90 for RRE, is affected by protein-protein interaction (NF90-Rev), with as a consequence low expression of the gene linked to RRE; another possibility would be that the interaction of NF90 with Rev, alters the conformation of Rev, allowing decreased access to RRE.