|Home | About | Journals | Submit | Contact Us | Français|
The evasion of host innate immunity by Rabies virus, the prototype of the genus Lyssavirus, depends on a unique mechanism of selective targeting of interferon-activated STAT proteins by the viral phosphoprotein (P-protein). However, the immune evasion strategies of other lyssaviruses, including several lethal human pathogens, are unresolved. Here, we show that this mechanism is conserved between the most distantly related members of the genus, providing important insights into the pathogenesis and potential therapeutic targeting of lyssaviruses.
The principal host response to viral infection in humans is the expression of type I interferons (IFNs) (alpha IFN [IFN-α] and IFN-β), which activate intracellular signaling via phosphorylation of signal transducers and activators of transcription 1 (STAT1) and STAT2 at tyrosines Y701 and Y690, respectively. This results in the generation of STAT1/2 heterodimers, the principal mediators of IFN-α/-β signaling, as well as STAT1 homodimers, which translocate into the nucleus, where they induce gene transcription essential to the establishment of the antiviral state (10). Viruses have evolved powerful countermeasures to inhibit STAT signaling through the activity of virus-expressed IFN antagonist proteins (38), including the phosphoprotein (P-protein) of Rabies virus (RABV) (17, 44), the best-characterized member of the genus Lyssavirus. There are currently 14 viruses in this genus (8, 21, 26), with all characterized members documented to infect mammals (4, 29), including a broad range of livestock such as cattle, sheep, and horses (34). Human infection by RABV alone is estimated to cause >55,000 deaths/year, and six other lyssavirus species have been reported to cause lethal rabies disease in humans (4, 29), the incidence of which is almost certainly underreported (19, 22). Thus, emerging zoonotic lyssaviruses pose important threats to human health and agriculture.
Inhibition of IFN-dependent signaling in RABV infection is achieved through the interaction of the globular C-terminal domain (CTD; residues 173 to 297) of P-protein with STAT1 and STAT2 (3, 27, 45). This interaction is reported to require the C-terminal 10 amino acid residues of P-protein, as deletion of the 10 or 30 C-terminal residues abrogates both STAT1/2 binding and inhibition of IFN-dependent transactivation (3, 27, 45). The P-protein–STAT1/2 interaction inhibits IFN-activated nuclear translocation of STAT1/2 via the activity of a nuclear export sequence (NES) in the N-terminal region of P-protein (residues 49 to 58), which mediates active nuclear export of the P-protein–STAT1/2 complex, dependent on the activity of the cellular nuclear export protein CRM1 (37). The IFN antagonist proteins of several other viruses, including henipaviruses, Mapuera virus, measles virus, Ebola virus, and hepatitis C virus, also cause mislocalization of STAT1/2 (9, 13, 28, 40, 41), but RABV appears to be unique in that it selectively and directly targets the IFN-activated form of STAT1 (3, 27, 45). The evolution of this highly specific mechanism may relate to the limited coding capacity of the RABV genome (36), such that P-protein, which has several vital roles in genome replication in addition to IFN antagonism, is used to target STAT1/2 only when required. Importantly, defects in the NES of P-protein of the RABV strain Nishigahara-chicken embryo (Ni-CE) correlate with impaired nuclear export and IFN antagonist function of P-protein and with attenuated pathogenicity in mice (17). Thus, specific STAT1/2 interaction and active nuclear export by P-protein appear to be essential to the inhibition of STAT1/2 signaling, thus representing a critical mechanism in pathogenic RABV infection and a potential target for attenuated vaccine strain development and/or therapeutics. However, the immune evasion mechanisms of other highly pathogenic lyssaviruses and the role(s) therein of their P-proteins have not been investigated.
To examine the IFN antagonist functions of lyssaviruses, we selected several viral strains and species representative of lyssavirus diversity, including the RABV strains Challenge virus standard (CVS), silver-haired bat rabies virus (SHBRV), 8743THA (THA), 9001FRA (FRA), and 9704ARG (ARG), the closely related Australian bat lyssavirus (ABLV) and European bat lyssavirus 1 (EBLV-1), and the lyssaviruses most distantly related to RABV, Lagos bat virus (LBV) and Mokola virus (MOKV) (1, 5–7, 11, 12, 16, 24, 25, 43, 47). HeLa cells infected with these viruses were treated 24 h postinfection with IFN-α (1,000 U/ml, 0.5 h) or not treated, before fixation (31), immunostaining for Y701-phosphorylated STAT1 (pY701-STAT1), and fluorescence microscopic analysis (20). IFN treatment resulted in clear nuclear accumulation of pY701-STAT1 in mock-infected cells (Fig. 1A), but no nuclear accumulation of pY701-STAT1 was observed in cells infected with any of the lyssaviruses tested (Fig. 1B), indicating that the capacity to inhibit IFN-dependent STAT1 nuclear translocation is conserved in the genus.
To investigate the IFN antagonist functions of P-proteins from different lyssaviruses directly, we generated constructs using the pEGFP-C1 plasmid for the expression in mammalian cells of green fluorescent protein (GFP)-fused P-proteins, including those from the pathogenic laboratory RABV strains CVS and Nishigahara (Ni), which efficiently inhibit IFN-α-dependent STAT1/2 signaling, the highly pathogenic street strain SHBRV, the P-protein of which has not previously been assessed for IFN antagonist function, and the lyssavirus species ABLV and MOKV. We also included the P-protein of the attenuated Ni derivative strain Ni-CE, which is defective in inhibition of IFN-dependent STAT1 nuclear localization and signaling in infected and transfected cells due to mutations that impair the function of its NES (17). The P-proteins of Ni, Ni-CE, SHBRV, ABLV, and MOKV show 91.9%, 90.6%, 90.2%, 85.5%, and 64.0% sequence similarity to CVS P-protein, respectively, and the hydrophobic residues of the NESs are highly conserved throughout the genus (Fig. 2A; data not shown). However, the C termini differ markedly, particularly through their extensions in the P-proteins of several viruses, including MOKV (Fig. 2B; data not shown). Living Cos-7 cells transfected to express the GFP–P-proteins were imaged by confocal laser scanning microscopy (CLSM) to determine the ratio of nuclear to cytoplasmic fluorescence (Fn/c), a quantitative measure of nucleocytoplasmic protein localization (17, 30, 31, 33, 42). All P-proteins were excluded from the nucleus to similar extents, except for Ni-CE P-protein, in which the NES is mutated, resulting in a more nuclear phenotype (17) (Fig. 3A). Treatment of cells with leptomycin B (LMB; 5.2 nM, 3 h), a specific inhibitor of CRM1 (30, 31, 35), significantly increased the nuclear localization of all P-proteins, indicating that NES function is conserved in the genus (Fig. 2A) as the main determinant of P-protein subcellular localization, suggestive of an essential function in infection. Interestingly, the nuclear localization of ABLV and MOKV P-proteins in LMB-treated cells was significantly higher than that of P-proteins from the RABV strains, which showed equivalent nuclear localization (Fig. 3B). Thus, the activity of the nuclear localization signal, predicted to reside in the CTD of RABV P-protein (37), may differ across the genus.
To examine the capacity of lyssavirus P-proteins to inhibit IFN-activated nuclear translocation of STAT1, mock-transfected cells and cells transfected to express lyssavirus P-proteins or control proteins were treated with IFN-α (1,000 U/ml, 0.5 h) or not treated, before immunostaining for STAT1 and CLSM analysis (31). Truncated CVS P-protein lacking the C-terminal 30 residues (CVS PΔC30), which is known to be impaired for STAT1/2 interaction (2, 45), and CVS N-protein, which has no role in inhibiting STAT1/2 signaling (17), did not affect IFN-α-dependent STAT1 nuclear translocation compared with that in mock-transfected cells (Fig. 4). In contrast, STAT1 nuclear accumulation in IFN-α-treated cells expressing full-length P-proteins of different lyssaviruses was significantly inhibited (Fig. 4A). As previously reported, Ni-CE P-protein was impaired in this respect due to defective activity of its NES (17). Importantly, the Fn/c for STAT1 in IFN-treated cells expressing full-length P-proteins was significantly lower than that measured for STAT1 in nontreated cells, except for cells expressing the Ni-CE P-protein, in which the Fn/c of activated STAT1 was slightly increased, as previously reported (17) (Fig. 4B). This suggests that all P-proteins bind selectively to IFN-α-activated STAT1 and, with the exception of the NES-defective Ni-CE P-protein, cause STAT1 to be exported from the nucleus. Latent STAT1, however, is not bound by P-protein and thus can localize diffusely between the cytoplasm and nucleus. In LMB-treated cells, the P-protein–STAT1 complexes showed increased nuclear localization in all cases (data not shown), confirming that STAT1 nuclear exclusion by lyssavirus P-proteins is dependent on CRM1-mediated nuclear export, consistent with previous reports for RABV P-protein (46). Thus, the unique mechanism of selective targeting of IFN-α-activated STAT1 to cause its CRM1-dependent nuclear exclusion appears to be conserved in the lyssavirus genus.
We next examined the physical interaction of P-proteins with STAT1 and STAT2 by coimmunoprecipitation from transfected Cos-7 cells (Fig. 5A). Consistent with the CLSM data, all P-proteins bound to STAT1 and STAT2 in an IFN-α-dependent manner, in contrast to the control proteins (Fig. 5A). RABV P-protein also inhibits dephosphorylation of pY701-STAT1, which is thought to be due to the inhibition of pY701-STAT1 nuclear translocation, preventing its interaction with nuclear phosphatases (3, 15). Using pY701-STAT1-specific antibody we could clearly detect pY701-STAT1 at 16 h after IFN-α treatment in lysates from cells expressing CVS, SHBRV, ABLV, and MOKV P-proteins but, as expected, no pY701-STAT1 could be detected in cells expressing control proteins (Fig. 5B). Thus, the capacity of P-protein to interact selectively with pY701-STAT1 and inhibit its dephosphorylation is maintained between distantly related lyssaviruses, indicating that the strategy for targeting STAT1 signaling is conserved across the genus.
To test the capacity of lyssavirus P-proteins to inhibit IFN-α-dependent STAT1/2 signaling, an IFN-α/STAT1/2-specific dual luciferase reporter-gene assay was used (14, 17). Cos-7 cells cotransfected with plasmids to express P-proteins or control proteins and the pISRE-luc/pRL-TK-luc vectors were treated with 1,000 U/ml IFN-α 6 h posttransfection or not treated and incubated for a further 16 h before analysis of IFN-α-specific luciferase activity. With the exception of the Ni-CE P-protein, which is defective for IFN antagonism due to its defective NES (17), all P-proteins suppressed IFN-α/STAT1/2-dependent luciferase transactivation activity to similar extents, compared with the control proteins CVS-PΔC30 and CVS-N protein (Fig. 5C). Thus, the capacity to antagonize IFN-α-dependent signaling is conserved in the lyssavirus genus, correlating with the conservation of IFN-α-dependent interaction with, and NES-dependent nuclear exclusion of, STAT1. Importantly, the conservation of this mechanism of IFN antagonism in distantly related viruses such as MOKV and LBV, the P-proteins of which differ significantly at their C termini compared with other lyssavirus P-proteins (Fig. 2B), suggests that this region may not directly contribute to the STAT1/2 binding site. Indeed, the C-terminal 10 residues of CVS P-protein include residues implicated in stabilization of the core CTD structure (18, 27), such that the impairment of STAT1/2 binding and IFN antagonist functions by deletion of this region (3, 45; this report) may be due to effects on the globular CTD structure that impact on a distinct CTD-localized STAT1/2 binding site.
In conclusion, we have shown that the unique mechanism of RABV P-protein inhibition of IFN-α-dependent STAT1/2 signaling is conserved in the lyssavirus genus. Because RABV P-protein-mediated inhibition of IFN-α-dependent STAT1/2 signaling correlates with the pathogenicity of RABV in infected animals (17), this step in viral infection is likely to represent a key pathogenicity factor and potential therapeutic target not only for RABV but also for emerging lethal human-pathogenic viruses, including ABLV and MOKV.
This research was supported by the National Health and Medical Research Council Australia project grant 1003244 to G.W.M., Senior Principal Research Fellowship 1002486 to D.A.J., Australian Research Council discovery project grant DP110101749 to G.W.M., D.A.J., and L.F.W., and European Union Seventh Framework Programme (FP7/2007-2013) PREDEMICS grant 278433 to F.L. and H.B.
We acknowledge Cassandra David for assistance with tissue culture and the facilities and technical assistance of Monash Micro Imaging, Victoria, Australia. We also acknowledge B. Dietzschold (Thomas Jefferson University, Philadelphia, PA), for providing the full-length cDNA clone of SHBRV strain (SBO) and are grateful to E. Perret and S. Shorte (Plate-forme d'Imagierie Dynamique, Institut Pasteur) for invaluable experimental help, discussion and advice on data processing.
Published ahead of print 27 June 2012