Identification of a dominant negative inhibitor of human zinc finger antiviral protein reveals a functional endogenous pool and critical homotypic interactions

Abstract

The zinc finger antiviral protein (ZAP) is a host factor with potent antiviral activity when overexpressed in cells. ZAP blocks replication of the prototype alphavirus Sindbis virus (SINV) at a step at or before translation of the incoming viral genome. The mechanism of ZAP anti-SINV activity and the determinants of its antiviral function, however, have not been defined. Here, we have identified a dominant negative inhibitor of human ZAP. Rat ZAP with a cysteine-to-arginine mutation at position 88 (rZAPC88R), previously reported as a nonfunctional form of ZAP, increases SINV growth in cells. These results led us to discover a previously undetectable pool of endogenous functional ZAP within human cells. Investigation of the mechanism of dominant negative inhibition, combined with a comprehensive mutational analysis of the antiviral factor, revealed that homotypic associations are required for ZAP function in limiting SINV propagation.

FIG. 1.

Expression of ZAP mutant construct rZAPC88R enhances SINV replication. (A) T-REx-rZAP or -rZAPC88R cells were left untreated or induced with doxycycline (Dox) for 12 h as indicated and then infected for 1 h with SINV Toto1101 (MOI = 0.1). At the indicated time points, the medium was harvested; separate wells were utilized for each time point. Virus titers were determined in duplicate by plaque assay on BHK-J cells. The mean and standard error of the mean of duplicate wells are shown. Asterisks indicate a significant difference (P < 0.05, unpaired Student t test) compared to the noninduced sample from the same time point. (B) Induction efficiency of T-REx-rZAP and -rZAPC88R cells. Cells were left untreated or induced with doxycycline (DOX) for 12 h and then Western blotted for myc-tagged rZAP or rZAPC88R, as well as β-actin. The values to the left are molecular sizes in kilodaltons. (C) rZAPC88R expression enhances translation of the incoming SINV genome. Induced T-REx-rZAP, -rZAPC88R, or -LacZ cells were infected with temperature-sensitive SINV encoding luciferase, Toto1101/Luc:ts6 (MOI = 10). Infection and incubation were carried out at the restrictive temperature (40°C). Translation was measured by luciferase activity. Means and standard deviations of triplicate infections are shown. Similar results were obtained in an independent experiment. P values (unpaired Student t test) for a subset of the data are shown. RLU, relative light units.

FIG. 2.

Wild-type and mutant ZAP levels dictate the cellular antiviral state. (A) T-REx-rZAPC88R cells were transfected with IRR siRNA or hZAP-specific siRNA, as indicated. Cells were then left uninduced (−) or treated with doxycycline (+) to induce expression of rZAPC88R and infected 12 h later with Toto1101/Luc (MOI = 10). Replication, as monitored by firefly luciferase activity, was normalized to cell viability at the time of infection. RLU, relative light units. Means and standard deviations of triplicate samples are shown. Similar results were obtained in an independent experiment. P values (unpaired Student t test) are shown. Dox, doxycycline. (B) T-REx cell lines were induced to express rZAP, rZAPC88R, or LacZ and subsequently transfected with plasmids expressing HA-tagged rZAP (rZAPHA) or rZAPC88R (rZAPC88RHA). After infection with EGFP-expressing SINV TE/5′2J/GFP (MOI = 10) for 12 h, cells were analyzed for HA (transfection) and EGFP (infection) levels using flow cytometry. HA− (nontransfected) and HA+ (transfected) populations of cells were separately analyzed. The percentage of cells infected (%GFP positive) is shown. Mean fluorescence intensities (MFI) of the infected populations are shown for nontransfected (HA−) cells in T-REx-rZAPC88R and -LacZ cells transfected with HA-tagged rZAPC88R (rZAPC88RHA). Similar results were obtained in three independent experiments. Asterisks indicate significant differences (P < 0.05, unpaired Student t test) between transfected cells and their nontransfected counterparts.

FIG. 3.

hZAP interacts with itself and with rZAPC88R. (A) 293T cells were cotransfected with plasmids expressing myc-tagged hZAP or rZAPC88R together with GST, GST-tagged hNZAP, or rNZAPC88R. Two days posttransfection, GST or GST fusion proteins were precipitated from cell lysates using glutathione agarose beads. The levels of GST or GST-tagged proteins were monitored by Western blotting; copurified myc-tagged proteins were detected by stripping the blot and reprobing with anti-myc antibodies. Expression of myc-tagged hZAP or rZAPC88R in lysates of transfected cells is shown. IP, affinity purification. (B) ZAP interaction was measured by determining reconstituted luciferase activity (relative light units, RLU) in 293T cells using the split-Gaussia assay. The assay was conducted in the absence (black) or presence (gray) of a competing construct encoding myc-tagged hZAP without a split-luciferase fragment. ZIP, leucine zipper GCN4. HCV5A, cytoplasmic fragment of HCV RNA-binding protein NS5A. Similar results were obtained in two other independent experiments. For each transfection pair, asterisks indicate statistically significant differences (P < 0.05, unpaired Student t test) obtained upon addition of the competing hZAPmyc construct.

FIG. 4.

Identification of sequences important for ZAP-ZAP interaction and antiviral function. (A) hNZAP constructs containing quintuple-alanine mutations, expressed as a fusion with the carboxyl-terminal Gaussia luciferase fragment (GLuc2), were cotransfected with a plasmid encoding full-length wild-type hZAP fused to the amino-terminal Gaussia luciferase fragment (GLuc1). Reconstituted luciferase activity was normalized to mutant hNZAP protein levels (relative light units, RLU). Means and standard deviations of triplicate transfections are shown. For both panels A and B, a schematic of hNZAP is shown as a white box; gray shading indicates the position of the putative CCCH zinc finger motifs. Similar results were obtained in another independent experiment. (B) 293T cells were transfected with plasmids encoding alanine mutant hNZAP proteins fused to GST. Relative antiviral activity was determined by a flow cytometry based assay; GST-hNZAP inhibitory activity was defined as 100%, and GST was defined as having 0% antiviral activity. Means and standard deviations of duplicate experiments are shown. (C) The ZAP-ZAP interaction (A) and the antiviral activities (B) of the alanine mutant proteins were plotted. The black square represents the data point of wild-type hNZAP. Gray lines mark the activity level equivalent to 50% of that of wild type hNZAP in the antiviral activity and the ZAP-ZAP interaction. The fitted local regression curve (LOESS curve) is shown. Six mutant proteins (16-20, 111-15, 126-30, 171-75, 221-25, and 231-35) were omitted, as they showed <10% transfected cells in the antiviral assay, presumably due to reduced stability of the protein or poor transfection efficiency.