Inhibition of filovirus replication by the zinc finger antiviral protein

Abstract

The zinc finger antiviral protein (ZAP) was recently shown to inhibit Moloney murine leukemia virus and Sindbis virus replication. We tested whether ZAP also acts against Ebola virus (EBOV) and Marburg virus (MARV). Antiviral effects were observed after infection of cells expressing the N-terminal part of ZAP fused to the product of the zeocin resistance gene (NZAP-Zeo) as well as after infection of cells inducibly expressing full-length ZAP. EBOV was inhibited by up to 4 log units, whereas MARV was inhibited between 1 to 2 log units. The activity of ZAP was dependent on the integrity of the second and fourth zinc finger motif, as tested with cell lines expressing NZAP-Zeo mutants. Heterologous expression of EBOV- and MARV-specific sequences fused to a reporter gene suggest that ZAP specifically targets L gene sequences. The activity of NZAP-Zeo in this assay was also dependent on the integrity of the second and fourth zinc finger motif. Time-course experiments with infectious EBOV showed that ZAP reduces the level of L mRNA before the level of genomic or antigenomic RNA is affected. Transient expression of ZAP decreased the activity of an EBOV replicon system by up to 95%. This inhibitory effect could be partially compensated for by overexpression of L protein. In conclusion, the data demonstrate that ZAP exhibits antiviral activity against filoviruses, presumably by decreasing the level of viral mRNA.

FIG. 1.

Inhibition of EBOV replication in Rat2-NZAP-Zeo cells. (A) Immunofluorescence analysis of NZAP-Zeo-expressing cells infected with Zaire-EBOV. Rat2-Zeo cells expressing the empty vector or Rat2-NZAP-Zeo cells expressing the N-terminal portion of ZAP fused to the product of the zeocin resistance gene were infected with Zaire-EBOV at an MOI of 1. Five days p.i., cells were fixed and immunofluorescence analysis was performed using a monoclonal antibody directed against Zaire-EBOV NP protein. (B) Growth kinetics of Zaire-EBOV. Rat2-Zeo cells or Rat2-NZAP-Zeo cells were infected with Zaire-EBOV at an MOI of 0.01 or 5. The amount of Zaire-EBOV-specific RNA in the supernatant was quantified using a Zaire-EBOV-specific real-time RT-PCR. The data represent the means and ranges of duplicate infection experiments. Dashed line, detection limit. (C) Determination of infectious virus titer. Rat2-Zeo or Rat2-NZAP-Zeo cells were infected with Zaire-EBOV or Sudan-EBOV at an MOI of 0.01 or 5. Seven days p.i., the amount of infectious virus released into the supernatant was determined by immunofocus assay on Vero-E6 cells. (D) Northern blot analysis of Zaire-EBOV NP-specific RNA. Rat2-Zeo or Rat2-NZAP-Zeo cells were infected with Zaire-EBOV at an MOI of 1. Five days p.i., total RNA was isolated and Northern blot hybridization was performed with a ³²P-labeled probe directed against Zaire-EBOV NP RNA. Noninfected cells served as a control. The methylene blue-stained 28S RNA is shown below the blot as a semiquantitative marker for gel loading and RNA transfer.

FIG. 2.

Inhibition of MARV Musoke replication in Rat2-NZAP-Zeo cells. (A) Rat2-Zeo cells or Rat2-NZAP-Zeo cells were infected with MARV Musoke at an MOI of 0.01 or 5. The amount of MARV-specific RNA in the supernatant was quantified using a MARV-specific real-time RT-PCR. The data represent the means and ranges of duplicate infection experiments. Dashed line, detection limit. (B) Determination of infectious virus titer. Rat2-Zeo or Rat2-NZAP-Zeo cells were infected with MARV Musoke at an MOI of 0.01 or 5. Seven days p.i., the amount of infectious virus released into the supernatant was determined by immunofocus assay on Vero-E6 cells. The data represent the means and ranges of duplicate infection experiments. neg, negative.

FIG. 3.

Inhibition of filovirus replication in 293TRex-ZAP cells. (A) ZAP expression was induced with doxycycline in 293TRex-ZAP cells, and 6 h later cells were infected with Zaire-EBOV or Sudan-EBOV (upper panel) as well as MARV Musoke or MARV Popp (lower panel) at an MOI of 0.1. Noninduced cells served as a control. Four days p.i., the amount of infectious virus released into the supernatant was determined by immunofocus or plaque assay on Vero-E6 cells. The data represent the means and ranges of duplicate infection experiments. (B) Western blot analysis of ZAP induction in 293TRex-ZAP cells. Zaire-EBOV- and MARV Popp-infected noninduced and induced 293TRex-ZAP cells used for the experiments shown in panel A were lysed 4 days p.i. and analyzed by Western blotting. Detection of ZAP was performed using a polyclonal anti-ZAP antibody. As a loading control, GAPDH was detected using an anti-GAPDH monoclonal antibody.

FIG. 4.

Effect of mutations within the zinc finger motifs of ZAP on its antiviral activity. (A) Rat2 cells expressing wild-type NZAP-Zeo or mutants containing single amino acid exchanges (H86K, C88R, C168R, or H191R) were infected with Zaire-EBOV, Sudan-EBOV, or MARV Musoke at an MOI of 0.01. Seven days p.i., the infectious virus titer was determined by immunofocus assay. Data represent means and ranges of duplicate infection experiments. Dashed line, immunofocus assay detection limit. (B) Level of NZAP-Zeo mutant protein expressed by cell lines as detected by Western blot analysis. Rat2-Zeo cells, Rat2-NZAP-Zeo cells, and Rat2 cells expressing NZAP-Zeo mutants containing the single amino acid exchange H86K, C88R, C168R, or H191R were lysed in Laemmli loading buffer. Proteins were separated by SDS-PAGE and transferred onto a membrane. NZAP-Zeo and NZAP-Zeo mutants were detected using a polyclonal anti-ZAP antibody. As a loading control, detection of GAPDH using an anti-GAPDH monoclonal antibody was performed.

FIG. 5.

Mapping of ZAP target sequences in Zaire-EBOV and MARV Musoke genomes. (A) cDNA fragments of Zaire-EBOV- and MARV Musoke-specific genes were cloned into pGL3 or pGL3-B between the firefly luciferase gene and poly(A) site. Inserted sequences as well as the approximate size of each insert (in kilobases) are indicated below each bar. The L genes were divided into four fragments (L1 to L4). SIN corresponds to a SIN genome fragment known to be sensitive to ZAP. The plasmids were transfected into Rat2-Zeo or Rat2-NZAP-Zeo cells. Cotransfection of a plasmid not sensitive to ZAP encoding Renilla luciferase (pRL-TK) was used to normalize firefly luciferase values. Cells were lysed 48 h posttransfection, and luciferase activity was measured. The inhibition (n-fold) was calculated as the normalized luciferase activity in Rat2-Zeo cells divided by the normalized luciferase activity in Rat2-NZAP-Zeo cells. Data are means and ranges of duplicate transfection experiments. (B) The EBOV L4 fragment was further divided into three or four overlapping fragments. A schematic representation of the fragments is shown at the left. Transfection and calculation of inhibition was performed as described for panel A. Numbers in brackets depict the approximate size of the inserts in kilobases. Data represent means and ranges of duplicate transfection experiments. (C) Four overlapping sequences covering the genome of YFV (YF1 to YF4) were inserted into pGL-3. The size of each insert (in kilobases) is shown in brackets. Transfection and calculation of inhibition was performed as described for panel A. Data represent means and ranges of duplicate transfection experiments. (D) The pGL-3 plasmid containing the EBOV L4 fragment was used to analyze the activity of NZAP-Zeo zinc finger mutants in the luciferase reporter assay. Transfection was performed with Rat2 cells expressing NZAP-Zeo mutants (H86K, C88R, C168R, or H191R). Empty pGL3 was used as a control. Transfection and calculation of inhibition was performed as described for panel A. Data represent means and ranges of duplicate transfection experiments.

FIG. 6.

Effect of ZAP on levels of L mRNA, genomic RNA, and antigenomic RNA during Zaire-EBOV infection. (A) Experimental strategy for quantification of Zaire-EBOV L mRNA, genomic RNA, and antigenomic RNA. Reverse transcription was performed with primers (large arrows) specifically binding either mRNA (RT-mRNA), genomic RNA (RT-g), or antigenomic RNA (RT-ag). The synthesized cDNA was quantified by real-time PCR using primers targeting the 3′ end of the L gene (small arrows). UTR, untranslated region; L, L gene. (B to D) 293TRex-ZAP cells were infected with Zaire-EBOV at an MOI of 3 for 1 h. After removing the inoculum, cells were induced with doxycycline. Noninduced cells served as controls. Total cellular RNA was isolated at various time points p.i. Zaire-EBOV-specific L mRNA (B), genomic RNA (C), and antigenomic RNA (D) were quantified using the PCR method as described for panel A. The amount of RNA derived from the inoculum (0 h) was defined as 1 (0 after log transformation). The average cyclic threshold (C_t) values at 0 h were the following: L mRNA, C_t = 29; genome, C_t = 22; antigenome, C_t = 26. Control reactions lacking reverse transcriptase were negative in the real-time PCR (data not shown). Data represent means and ranges of duplicate infection experiments (most error bars are obscured by the symbols).

FIG. 7.

Inhibition of Zaire-EBOV replicon system by ZAP. Huh-T7 cells were transfected with plasmids encoding Zaire-EBOV NP, L protein, VP35, VP30, and minigenome containing a Renilla luciferase gene. Increasing amounts of plasmid encoding ZAP-HA were cotransfected as indicated (−ctrl, negative control without L plasmid; +ctrl, positive control without ZAP-HA plasmid). Empty vector pTM1 was used to keep the amount of transfected DNA constant. Plasmid pTM1-FFluc was included to normalize Renilla luciferase values. Luciferase activities were determined 2 days posttransfection. The expression levels of ZAP-HA and EBOV NP were analyzed by Western blot using anti-HA monoclonal antibody and anti-Zaire-EBOV polyclonal antibody, respectively (bottom). The amount of cell lysate loaded on the protein gel was adjusted according to the firefly luciferase values to ensure comparability with the normalized Renilla values. The levels of GAPDH served as a loading control.

FIG. 8.

Effect of L-protein overexpression on ZAP-induced inhibition of the Zaire-EBOV replicon system. (A) Huh-T7 cells were transfected with plasmids encoding Zaire-EBOV NP, VP35, VP30, increasing amounts of L-protein expression plasmid, and a minigenome containing a Renilla luciferase gene (−ctrl, negative control without L plasmid). Different amounts of ZAP-HA expression plasmid were cotransfected. Empty vector pTM1 was used to keep the amount of transfected DNA constant. Plasmid pTM1-FFluc was included to normalize Renilla luciferase values. (B) Expression levels of ZAP-HA and EBOV NP were analyzed by Western blot using anti-HA monoclonal antibody and anti-Zaire-EBOV polyclonal antibody, respectively. The amount of cell lysate loaded on the protein gel was adjusted according to the firefly luciferase values to ensure comparability with the normalized Renilla values. The levels of GAPDH served as a loading control.