2′-5′-Oligoadenylate Synthetase-Like Protein Inhibits Respiratory Syncytial Virus Replication and Is Targeted by the Viral Nonstructural Protein 1

Abstract

2′-5′-Oligoadenylate synthetase-like protein (OASL) is an interferon-inducible antiviral protein. Here we describe differential inhibitory activities of human OASL and the two mouse OASL homologs against respiratory syncytial virus (RSV) replication. Interestingly, nonstructural protein 1 (NS1) of RSV promoted proteasome-dependent degradation of specific OASL isoforms. We conclude that OASL acts as a cellular antiviral protein and that RSV NS1 suppresses this function to evade cellular innate immunity and allow virus growth.

TEXT

Cellular innate immunity against virus infection is primarily mediated by type I interferons (IFNs). In turn, the IFNs exert their pleiotropic effects through the induction of a variety of IFN-stimulated genes (ISGs) (1–4). Although the general antiviral roles of several ISGs have been demonstrated, the roles of individual ISGs and their effect on specific viruses have remained largely unidentified (5–9). On the other hand, the coevolution of the host and the virus has resulted in viral strategies to evade the host IFN response by targeting ISGs and other IFN pathway proteins (10). Oligoadenylate synthetases (OAS) are a family of ISGs characterized by their ability to synthesize 2′-5′-oligoadenylate (2-5A), which induces RNA degradation by activating RNase L (11, 12). Human oligoadenylate synthetase-like protein (OASL) is related to the OAS family by its N-terminal OAS-like domain but is devoid of 2-5A synthetase activity. Additionally, OASL contains two tandem ubiquitin-like (UBL) domains in the C terminus, which are absent in any of the other members of the OAS family (12–15). OASL is directly and rapidly induced by virus infection via interferon regulatory factor 3 (IRF3) as well as by IFN signaling (1, 12, 16, 17). Unlike in humans, two OASL isoforms have been identified in the mouse, Oasl1 and Oasl2. We have recently described the antiviral activities of human OASL and mouse Oasl2, which are mediated through the enhancement of RIG-I signaling (18). In the present study, we examined the antiviral activities of various human and mouse OASL proteins against respiratory syncytial virus (RSV), the founding member of the clinically significant Pneumovirus genus of the Paramyxoviridae family. We show that human OASL and mouse Oasl2 strongly inhibit RSV replication and that the RSV nonstructural protein 1 (NS1) specifically targets these two isoforms to promote viral replication.

We first conducted a detailed analysis of the antiviral activity of human OASL against RSV. Recombinant OASL, stably expressed in HEK293 cells, strongly reduced RSV growth, measured by the reduction of progeny viral titer, which paralleled the reduction of intracellular viral RNA and proteins (Fig. 1), confirming our earlier findings (18) and establishing inhibition at the viral genome level. Similar antiviral activity of OASL was also observed using a different cell line, HCT116 (Fig. 2A). In complementary studies, enhanced RSV replication was seen when OASL expression was silenced by stable expression of the short hairpin RNA (shRNA) described previously (18) (Fig. 2B). To further corroborate these observations, we created OASL-deficient HCT116 cells by genome editing using the TALEN technology; RSV replication in these cells was also found to be highly elevated (Fig. 2C). Together, these results fully established the antiviral activity of OASL against RSV.

FIG 1.

Inhibition of RSV replication in cells expressing human OASL. (A) HEK293 cells, stably transfected with V5-tagged OASL expression plasmid or the empty vector (pcDNA), as described before (18), were grown in monolayers on coverslips and infected with RSV Long at a multiplicity of infection of 3. At 18 h postinfection, cells were fixed and immunostained with mouse anti-RSV nucleoprotein (N) antibody (Abnova clone B023), followed by Alex Fluor 610-conjugated donkey anti-mouse IgG (Life Technologies). Images were captured in a Nikon AIRSI spectral confocal microscope system. (B) (Top) The same cell lines were infected as described above, and the total cell lysates were analyzed by immunoblotting using the same primary antibody described above and horseradish peroxidase (HRP)-conjugated secondary antibody, followed by ECL (enhanced chemiluminescence) detection. Actin is the loading control. (Bottom) Total RNA isolated from parallel cultures was subjected to quantitative reverse transcription-PCR (qRT-PCR), as described previously (38). The primers, synthesized by Integrated DNA Technologies (Coralville, IA), were as follows. RSV N gene, forward 5′-TGCAGGGCAAGTGATGTTAC-3′, and reverse, 5′-TTCCATTTCTGCTTGCACAC-3′; actin, forward, 5′-AGAAAATCTGGCACCACACC-3′, and reverse, 5′-GGGGTGTTGAAGGTCTCAAA-3′. A portion of the PCR sample was analyzed on 1.5% agarose gel, and the ethidium bromide-stained bands are shown. (C) (Left) Plot of the PCR results described in panel B. An average from three data sets with error bars is shown. (Right) The liberated virus in the medium at 24 h postinfection was assayed for PFU on HEp-2 monolayers as described previously (39). The asterisks indicate significance (P < 0.001).

FIG 2.

RSV replication in OASL-expressing and OASL-deficient cells. HCT116 cells, in which OASL was either expressed recombinantly or silenced by shRNA, have been described (18). The cells were infected with RSV, viral growth was assayed by quantitative immunoblot as described in the legend to Fig. 1B, and the data were plotted. (A) Inhibition of RSV growth in OASL-expressing cells. (B and C) Enhanced RSV growth in OASL-deficient cells compared to control cells. The asterisks indicate significance (P < 0.001).

As human OASL lacks the enzymatic activity of OAS, it is incapable of generating the second messenger 2-5A and thereby does not activate RNase L. It is thus likely that OASL exerts its anti-RSV effect through a novel mechanism that is 2-5A independent. The existence of two mouse orthologs—the catalytically inactive Oasl1 and catalytically active Oasl2 (18, 19)—provided us with an opportunity to test this. However, whereas Oasl1 expression failed to inhibit RSV activity, Oasl2 did (Fig. 3A); moreover, primary bone marrow-derived macrophages (BMDMs), isolated from Oasl2^−/− mice, supported much more robust RSV growth than cells from isogenic wild-type mice (Fig. 3B). To abolish the catalytic activity of Oasl2, we then mutated two of the three conserved Asp residues in its core catalytic triad (19) to create the mutant designated “DADA.” As shown (Fig. 3C), these mutations did not reduce the anti-RSV activity of Oasl2. Thus, although mouse Oasl2 is enzymatically active, its antiviral activity is clearly not due to the generation of 2-5A.

FIG 3.

Interrogation of the antiviral role of two mouse Oasl isoforms. HEK293 cells were transfected with the indicated expression plasmids (and empty pcDNA vector controls) and then infected with RSV as in Fig. 1B. (A) Replication of RSV was measured by qRT-PCR assay of N gene mRNA, as in Fig. 1B, and by plaque assay as in Fig. 1C. The asterisks indicate significance (P < 0.0001). (B) Bone marrow-derived macrophages from Oasl2^−/− and wild-type mice were tested for RSV growth as in panel A. *, P < 0.05. (C) Human OASL, mouse Oasl1 and Oasl2, and the indicated mutant versions of these proteins were expressed by transfection in mouse embryonic fibroblasts, and RSV growth was assayed as in panel A. ***, P < 0.001.

We have recently shown that the unique UBL domain of OASL, absent in classic OAS enzymes, is important for antiviral activity against vesicular stomatitis virus (VSV) and Sendai virus (SeV) (18). We found this is also true for RSV, since loss of the UBL domain of OASL also led to the total loss of anti-RSV activity (Fig. 3C). Overall, we conclude that the antiviral activity of OASL proteins is independent of 2-5A but requires the UBL domain.

The nonstructural proteins of RSV are known to suppress the host cell's IFN response (20–27), thus promoting robust virus growth and pathogenesis. Although the exact molecular mechanism of the suppression is still being actively pursued, we and others have shown that NS proteins target multiple members of the IFN induction and response pathways, such as RIG-I, IRF3, IRF7, and STAT2 (28–37). We also showed that NS1 and/or NS2 promotes the degradation of these substrates, in part by recruiting a proteasomal activity (37). It was, therefore, logical to investigate whether the NS proteins may also target OASL for degradation, in an attempt to counteract the antiviral function of OASL. Indeed, when coexpressed, recombinant NS1 strongly reduced OASL protein levels (Fig. 4A), whereas NS2 was ineffective. This activity of NS1 could be inhibited by lactacystin, indicating proteasomal degradation as the mechanism (Fig. 4B). Furthermore, a pulse-chase experiment showed that the in vivo half-life of OASL is reduced from >4 h to about 30 min in the presence of NS1, confirming a posttranslational effect (Fig. 4C). Finally, a mutated OASL from which the UBL domain was deleted (ΔUBL OASL) failed to be targeted by NS1 (Fig. 4D), suggesting that this ubiquitin-like domain plays a cardinal, potentially novel, role in the proteasomal targeting of OASL. Like human OASL, mouse Oasl2 was targeted by NS1, but interestingly, mouse Oasl1 was not (Fig. 4D). The enzymatically defective DADA mutant of Oasl2 was also NS1 sensitive (Fig. 4D). Based on these results, it appears that the viral NS1 protein has evolved to specifically target the RSV-inhibitory OASL and Oasl2 proteins and that such targeting is dependent on the presence of the ubiquitin-like domain in these proteins. However, our study for the first time points out unique properties in the UBL domains of Oasl1 and Oasl2, which render one resistant to NS1 while making the other sensitive. Finer mapping of the specific residues in the UBL domains responsible for this differential sensitivity will shed light on the molecular mechanism of the NS1-mediated targeting and help create clinically useful inhibitors of NS1.

FIG 4.

Reduction of OASL expression by RSV NS1. (A) (Top) Recombinant V5-tagged human OASL (18) and pCAGGS plasmids expressing Flag-tagged NS1 or NS2 (34) were cotransfected into HEK293 cells, and the levels of the proteins were determined by immunoblotting at 24 h posttransfection. Actin serves as the loading control. (Bottom) qRT-PCR of OASL and actin mRNA from the cells indicated above was conducted, and the products were analyzed as in Fig. 1B. Primers for human OASL were as follows: forward, 5′-AAAGAGAGGCCCATCATCC-3′; reverse, 5′-ATCTGGGTAACCCCTCTGC-3′. (B) The experiment was done as in panel A using either full-length OASL or OASL with UBL deleted cotransfected with NS1 plasmid. Where indicated, lactacystin (LC [10 μM]) was added to the medium. Actin is the loading control. (C) Pulse-chase experiment to determine the effect of NS1 on the half-life of OASL protein. Cells were transfected with NS1 expression plasmid (or empty pCAGGS vector) as in panel A and 18 h later pulse-labeled with [³⁵S]-Met-Cys (Trans-³⁵S-label; MP Biomedicals) in Met-Cys-free Dulbecco's modified Eagle's medium (DMEM) (Life Technologies) for 1 h, followed by chase with complete DMEM supplemented with Met and Cys (10 mg/ml each). At the indicated time points, the cells were processed for standard immunoprecipitation with V5 antibody and protein A-Sepharose, followed by SDS-PAGE and autoradiography. (Top) Autoradiograph. (Bottom) Plot of the band intensities in the autoradiograph (○, no NS1; △, with NS1). (D) The experiment was done as in panel B using the indicated V5-tagged human and mouse Oasl homologs and their mutants cotransfected with NS1. Actin was the loading control, as used before.