Genetic screen of a library of chimeric poxviruses identifies an ankyrin repeat protein involved in resistance to the avian type I interferon response

Abstract

Viruses must be able to resist host innate responses, especially the type I interferon (IFN) response. They do so by preventing the induction or activity of IFN and/or by resisting the antiviral effectors that it induces. Poxviruses are no exception, with many mechanisms identified whereby mammalian poxviruses, notably, vaccinia virus (VACV), but also cowpox and myxoma viruses, are able to evade host IFN responses. Similar mechanisms have not been described for avian poxviruses (avipoxviruses). Restricted for permissive replication to avian hosts, they have received less attention; moreover, the avian host responses are less well characterized. We show that the prototypic avipoxvirus, fowlpox virus (FWPV), is highly resistant to the antiviral effects of avian IFN. A gain-of-function genetic screen identified fpv014 to contribute to increased resistance to exogenous recombinant chicken alpha IFN (ChIFN1). fpv014 is a member of the large family of poxvirus (especially avipoxvirus) genes that encode proteins containing N-terminal ankyrin repeats (ANKs) and C-terminal F-box-like motifs. By binding the Skp1/cullin-1 complex, the F box in such proteins appears to target ligands bound by the ANKs for ubiquitination. Mass spectrometry and immunoblotting demonstrated that tandem affinity-purified, tagged fpv014 was complexed with chicken cullin-1 and Skp1. Prior infection with an fpv014-knockout mutant of FWPV still blocked transfected poly(I·C)-mediated induction of the beta IFN (ChIFN2) promoter as effectively as parental FWPV, but the mutant was more sensitive to exogenous ChIFN1. Therefore, unlike the related protein fpv012, fpv014 does not contribute to the FWPV block to induction of ChIFN2 but does confer resistance to an established antiviral state.

Fig 1

Sensitivity of viruses to recombinant ChIFN1. Plaque reduction assays were performed using CEFs infected with Semliki Forest virus, FWPV FP9, or VACV VTN or MVA in the presence of recombinant ChIFN1. (A) Photograph of crystal violet-stained CEFs showing titration of ChIFN1 inhibition of FP9 and MVA plaque formation. (B) Graph showing plaque production as a function of ChIFN1 concentration. The dotted line indicates the 50% reduction endpoint. Plaque formation is expressed as a percentage of plaques formed in mock-treated wells (mean of n = 3). Each point represents an individual well, except for FP9, where points are the mean (4 wells from 2 independent experiments) ± 1 SD. Symbols: squares, FP9; circles, VACV VTN; closed triangles, MVA II/85; open inverted triangles, SFV A7 (27).

Fig 2

Characterization of the FWPV chimeric MVA library. (A) Distribution on the FWPV FP9 genome of fragments inserted in the chimeric MVA library. The integrity of FWPV inserts was examined by PCR, using internal primers from the overlapping fragments with primers flanking the insertion site. Inserts that were shorter than expected were labeled “unstable.” Where duplicate viruses containing a particular fragment were recovered, the fragment is classified here as a full-length insert if at least one derived recombinant virus was of the anticipated size. Regions of the FP9 genome that were not present in the recombinant MVA library are highlighted as “uncloned genome.” The drawing is to scale. (B) Expression of fpv191 by MVA chimeras MVA-f50a and MVA-f50c. Lysates of uninfected or infected CEFs were subject to 10% SDS-PAGE and then immunoblotted using primary monoclonal antibody DE9 (38) at 1 in 200 and secondary goat anti-mouse IgG (Sigma-Aldrich) at 1 in 25,000. Lane 1, uninfected CEFs; lane 2, FP9-infected CEFs; lane 3, MVA-LZ-infected CEFs; lane 4, MVA-f50a-infected CEFs; lane 5, MVA-f50b-infected CEFs; lane 6, MVA-f50c-infected CEFs; lane 7, MVA-fC2b-infected CEFs. The amount of cell lysate loaded is expressed relative to that for the FP9-infected lysate (lane 2).

Fig 3

Screening of FWPV chimeric MVA for enhanced resistance to ChIFN1. CEFs were pretreated in triplicate with ChIFN1 (10 U ml⁻¹) or mock treated with DMEM–2% FBS for 18 h. The medium was aspirated, and cells were infected with chimeric viruses at 100 to 300 (A) or 50 to 200 (B) PFU/well. Virus was aspirated after 90 min, and semisolid overlay was added. A second overlay containing X-Gal was added at 3 days postinfection. Plaque formation for each virus in ChIFN1-treated wells is expressed as the mean (n = 3, ± SEM) percentage relative to that for mock-treated wells. Separate experiments are shown in panels A and B. The screen shown in panel A used third-bulk-passage, mycophenolic acid-selected MVA-f3La (subsequently found to contain residual MVA-LZ), but that shown in panel B used fourth-passage, plaque-purified MVA-f3La (found to be free of parental MVA-LZ). rChIFN1, recombinant ChIFN1.

Fig 4

qRT-PCR analysis of the expression of FWPV insert-specific genes by first- and second-stage FWPV chimeric MVA. Expression of mRNA specific for fpv012 and fpv014 in FP9-infected and control (MVA-fC2a) or FWPV chimeric MVA-infected CEFs was assayed by qRT-PCR and normalized against that for chicken GAPDH. MVA-f3La is a first-stage chimera carrying the FIR1 locus; MVA-F.012 and MVAF.014 are second-round recombinants carrying only fpv012 and fpv014, respectively. Expression of mRNA specific for MVA A12L and the FWPV equivalent of A12L (fpv176) in FP9- and chimeric MVA-infected CEFs as vector-specific controls was assayed by qRT-PCR and normalized against that for chicken GAPDH. (A) fpv012; (B) fpv014; (C) fpv176 bars; (D) A12L. The mean and standard deviation at 8 and 16 h postinfection (hpi) are plotted from three independent experiments. Note the differences in the y-axis scales.

Fig 5

Screening and characterization of second-round chimeric MVA to identify the FIR1 gene responsible for enhanced resistance to ChIFN1. CEF cells (in 9-cm² dishes) were incubated with 1 ml DMEM–2% FBS with or without ChIFN1 at 27 (A), 13 (B), or 6 (C) U ml⁻¹ for 18 h. The medium was aspirated, and cells were infected at 90 to 180 PFU/well. All viruses except control MVA-fC2a were plaque purified. MVA-f3La is a first-stage chimera carrying the FIR1 locus; MVA-F.012, MVA-F.013, and MVAF.014 are second-round recombinants carrying only fpv012, fpv013, and fpv014, respectively. Except in panel A, which is an example of several experiments, the numbers of plaques formed in the ChIFN1-treated wells were expressed as a percentage of the mean (n = 3) number of plaques formed in mock-treated wells. (D) Plaque reduction assays were performed on FWPV chimeric MVA over a range of ChIFN concentrations from 5 to 200 U ml⁻¹. IC₅₀ values for ChIFN1 with each of the viruses were derived in Prism software (GraphPad) using nonlinear fits of the log-converted ChIFN1 concentrations against the percentage of normalized plaque responses. IC₅₀ values are plotted with 95% confidence intervals. P values were determined by two-way analysis of variance: *, P < 0.05; **, P < 0.01.

Fig 6

Characterization of expression of fpv014 in FWPV. (A) Analysis of the kinetics of expression of mRNA specific for fpv014 in wild-type FWPV was performed by qRT-PCR, using as controls FWPV genes fpv100 (ortholog of VACV E4L; RNA polymerase subunit RPO30) and fpv168 (ortholog of VACV A4L). Expression in the absence and presence of poxviral DNA replication inhibitor AraC was normalized against that for chicken GAPDH. Means and standard deviations are plotted from three independent experiments. (B) Expression of TAP-tagged fpv014, either FL or CtDel, inserted back into the native locus in FWPV FP9 under the control of its cognate promoter, detected by immunoblotting of SDS-polyacrylamide gels with anti-FLAG antibody (Sigma) and anti-mouse secondary antibody (LI-COR) per the manufacturers' protocols. The immunoblots were imaged using a LI-COR Odyssey infrared imaging system. Samples were obtained at 24 h postinfection at an MOI of 3. Molecular mass markers (M) are shown, as are the predicted (Pred.) masses of FL (53 kDa) and CtDel (50 kDa) TAP-tagged fpv014.

Fig 7

Characterization of an fpv014-knockout mutant of FWPV. (A) Like parental FWPV FP9, but unlike one defective in fpv012 (Del012), an FP9 mutant defective in fpv014 (Del014) retains the full ability to block poly(I·C)-mediated induction of the ChIFN2 promoter. Chicken DF1 cells were transfected with the ChIFN2 promoter reporter (pChIFN2lucter) and the constitutive lacZ reporter plasmid pJATlacZ. Following recovery for 24 h, cells were either left uninfected or infected with parental FWPV FP9 or single-gene mutants of FP9 at an MOI of 10. Following infection for 4 h, cells were either left untreated or transfected with poly(I·C) (10 μg ml⁻¹) and incubated for 16 h. Luciferase assays were carried out, and data were normalized using β-galactosidase measurements. The result for each sample was compared to that for the uninfected, poly(I·C)-treated control to calculate percent induction. Results show the mean (n = 3) + SD. (B) Plaque reduction assays were performed on parental and fpv014-knockout (Del014) FWPV FP9 over a range of ChIFN1 concentrations from 1 to 10,000 U ml⁻¹. IC₅₀ values for ChIFN1 with each of the viruses were derived in Prism software (Graphpad) using nonlinear fits of the log-converted ChIFN1 concentrations against the percentage of normalized plaque responses. IC₅₀ values are plotted with 95% confidence intervals.

Fig 8

fpv014 is an ANK/PRANC protein that associates with Skp1 and cullin-1. (A) Domain structure of fpv014 showing N-terminal ankyrin repeats (ANKs) as well as the C-terminal F-box motif and the larger, encompassing PRANC domain. The structure is to scale. (B) Chicken cullin-1 amino acid sequence (GenBank accession number XP_418878) aligned (dots indicate identical residues) with that of human cullin-1 (GenBank accession number NP_003583), showing peptides (underlined) identified by mass spectrometric sequence analysis of proteins copurified with TAP-tagged fpv014. (C) Immunoblot detection of Skp1 and cullin-1 co-affinity purified with TAP-tagged fpv014. DF1 control cell lysate (Control lysate) or the final eluate of TAP affinity-purified lysate from a DF-1 cell line inducibly expressing TAP-tagged fpv014 (TAP-affinity-purified eluate) were subject to nonreducing 15% SDS-PAGE and then immunoblotted using mouse anti-FLAG (Sigma-Aldrich), mouse anti-Skp1, or mouse anti-cullin-1 (BD Transduction Laboratories) at 1 in 1,000, followed by goat anti-mouse IgG (Sigma) at 1 in 25,000. Molecular mass markers (M) are shown, as are the predicted masses of TAP-tagged fpv014 (53 kDa), ChSkp1 (19 kDa), and chicken cullin-1 (90 kDa).