Identification of Interferon-Stimulated Gene 15 as
http://www.100md.com
病菌学杂志 2005年第22期
Departments of Pathology
Immunology, Washington University School of Medicine, St. Louis, Missouri 63110
Division of Infectious Diseases, University of Texas Southwestern Medical Center, Dallas, Texas 75390-9113
Department of Medicine, Columbia University College of Physicians and Surgeons, New York, New York 10032
ABSTRACT
The innate immune response, and in particular the alpha/beta interferon (IFN-/) system, plays a critical role in the control of viral infections. Interferons and exert their antiviral effects through the induction of hundreds of interferon-induced (or -stimulated) genes (ISGs). While several of these ISGs have characterized antiviral functions, their actions alone do not explain all of the effects mediated by IFN-/. To identify additional IFN-induced antiviral molecules, we utilized a recombinant chimeric Sindbis virus to express selected ISGs in IFN-/ receptor (IFN-/R)–/– mice and looked for attenuation of Sindbis virus infection. Using this approach, we identified a ubiquitin homolog, interferon-stimulated gene 15 (ISG15), as having antiviral activity. ISG15 expression protected against Sindbis virus-induced lethality and decreased Sindbis virus replication in multiple organs without inhibiting the spread of virus throughout the host. We establish that, much like ubiquitin, ISG15 requires its C-terminal LRLRGG motif to form intracellular conjugates. Finally, we demonstrate that ISG15's LRLRGG motif is also required for its antiviral activity. We conclude that ISG15 can be directly antiviral.
INTRODUCTION
The innate immune response plays a critical role in controlling viral infections, with the alpha/beta interferon (IFN-/) system acting as an important component of this response. IFN-/ are expressed in response to viral infection and exert their effects through the IFN-/ receptor (IFN-/R), resulting in the activation of JAK/STAT-dependent and STAT-independent pathways and the subsequent induction of hundreds of genes containing interferon-stimulated response elements (9, 45). A subset of interferon-stimulated genes (ISGs) has been shown to directly inhibit viral replication (1, 2, 15, 16, 21, 36, 42, 47, 50). In addition to their antiviral activity, IFN-/ also mediate a variety of immunomodulatory effects that link the innate and adaptive immune responses. These include inducing the maturation of dendritic cells; enhancing NK cell cytotoxicity; increasing the production of cytokines, such as interleukin-15 and IFN-; and upregulating the expression of major histocompatibility complex class I molecules (3). The in vivo protection mediated by IFNs during viral infection is due in part to the direct inhibition of viral replication and in part to these immune-modulatory effects. While studies of protein kinase RNA activated (PKR), RNAse L, and Mx have shown these ISGs to play an important role during certain viral infections, these ISGs do not explain all of the effects mediated by IFN-/ (1, 46, 50, 51), suggesting the existence of additional IFN-induced antiviral pathways.
ISG15 is one of the earliest ISGs to be induced (up to 100-fold) following IFN stimulation (4, 7, 19, 20). It is expressed in a variety of cell types at low levels and induced within 6 h of stimulation with IFN-/, and to a lesser degree by IFN- (26). ISG15 has two ubiquitin-like domains, including a C-terminal LRLRGG motif through which it forms covalent conjugates with intracellular proteins (10, 26, 31). The formation of ISG15-protein conjugates parallels other ubiquitin-like proteins (UBLs), SUMO-1 and RUB-1, in the utilization of conjugating and deconjugating enzymes, many of which are distinct from the ubiquitin pathway (17, 28, 31, 48). The first member of the pathway to be identified was UBP43, a deconjugating enzyme for ISG15 (25, 28). This was followed by identification of UBE1L as an E1 (48) and the identification of UbcM8 as an E2 for ISG15 (17, 49). An E3 has yet to be reported. All members of this cascade, including ISG15, are tightly regulated by IFN-/. To date, only a few of the intracellular targets of ISG15 have been identified, including phospholipase C1, Jak1, Stat1, ERK1, and serpin 2a (8a, 11, 27, 49a). However, the fates of these and other ISGylated proteins remain to be elucidated. An extracellular form of ISG15 can also be found in the media of IFN-stimulated cells and in the sera of patients treated with IFN (6, 18, 35), and human recombinant ISG15 has been reported to exert cytokine activity (5, 18, 38). Despite these findings, the functional significance of ISG15 expression remains unclear.
There are, however, three types of data suggesting that ISG15 may play an important role during viral infection. First, ISG15 is rapidly upregulated by IFN, lipopolysaccharide (LPS), and viral infection. This has been shown both in vitro with microarray analysis (4, 7, 19, 20) and in vivo following viral infection (13, 20, 39). Infection of human fibroblasts with human cytomegalovirus (HCMV) resulted in a 150-fold increase in the expression of ISG15 mRNA (52). ISG15 RNA was also induced by herpes simplex virus infection (33). Infection of neonatal mice with Sindbis virus resulted in a >100-fold increase in ISG15 RNA expression in infected brain tissue (13, 20). ISG15 protein expression has also been localized to the meningeal, ependymal, and choroid plexus cells following lymphocytic choriomeningitis virus (LCMV) infection (39). Therefore the expression of ISG15 RNA and protein is markedly upregulated during viral infection. Second, mice lacking the deconjugating enzyme UBP43 demonstrate an increased resistance to infection with both LCMV and vesicular stomatitis virus (39). UBP43–/– cells have elevated basal levels of ISG15-protein conjugates, which are dramatically increased following IFN stimulation (28, 29). There is also an alteration in interferon signaling, with the UBP43–/– cells being hyperresponsive to interferon stimulation with prolonged STAT-1 phosphorylation, DNA binding, and IFN-mediated gene activation (29). While these data are consistent with a possible antiviral role of ISG15, they are not conclusive, since it is unclear if the antiviral state seen in the UBP43–/– mouse is due to increased levels of ISG15, to an enhanced IFN response that results in the upregulation of multiple ISGs, or to an as-yet-uncharacterized effect of UBP43. Finally, the influenza B protein NS1B binds to ISG15 and inhibits the coupling of UBE1L to ISG15 (48). This inhibits the formation of ISG15 conjugates in virus-infected cells. Therefore, at least one virus has evolved specific mechanisms to disrupt ISG15 function. These findings, in conjunction with the tight regulation of this conjugation cascade by IFN, suggest ISG15 may play an important role during viral infection. However, to date, ISG15 has no known function, and there has been no direct evidence that ISG15 is antiviral in vivo.
In this study, we utilize a recombinant, chimeric Sindbis virus system to overexpress ISG15 in IFN-/R–/– mice, which lack the ability to efficiently upregulate ISGs, including ISG15, in response to IFN-/. We show that expression of ISG15 in IFN-/R–/– mice attenuates Sindbis virus infection, providing in vivo evidence that ISG15 can function as an antiviral molecule.
MATERIALS AND METHODS
Mice. Mice were maintained at Washington University School of Medicine in accordance with all federal and university guidelines. 129, IFN-/R–/–, and IFN-//R–/– mice on the 129/SvPas background were initially obtained from M. Auget and then bred and maintained in our mouse colony (8, 9, 30). Unless otherwise noted, all animals were between 8 and 12 weeks of age and were age and sex matched within experiments.
Generation of ISG15 antibodies. Recombinant murine ISG15 was generated by cloning the open reading frame of murine ISG15 into pET-30a(+) (Novagen, Madison, WI) to express ISG15 with an amino-terminal His6 tag. His-tagged ISG15 protein (recombinant ISG15) was column purified on Ni-NTA His Bind Resin (Novagen) according to the manufacturer's instructions. Polyclonal rabbit antiserum (Cocalico, Reamstown, PA) was generated as previously described (14). Briefly, an initial inoculation of 100 μg recombinant ISG15 in complete Freund adjuvant was followed by 10 inoculations with 50 μg of recombinant ISG15 in incomplete Freund adjuvant 1 month apart. Sera were collected 7 days after boosts. Monoclonal antibodies were generated in Armenian hamsters in the Washington University School of Medicine Hybridoma Center using standard procedures (43).
Recombinant virus strains. Recombinant double-subgenomic Sindbis virus, dsTE12Q, was produced from a viral cDNA clone by in vitro transcription and RNA transfection of baby hamster kidney-21 (BHK) cells (12). ISG15-expressing Sindbis viruses were generated as follows.
(i) ISG15 FULL. Nucleotides 1 to 486 of murine ISG15 were PCR amplified using a 5' primer that contained a BstEII restriction site and FLAG epitope sequence and a 3' primer containing a BstEII site.
(ii) ISG15 MUTANT. The initial start codon of the ISG15 gene was deleted. The 5' primer contained a BstEII restriction site and a FLAG epitope, and the 3' primer contained a BstEII site.
(iii) ISG15 LRLRGG. Nucleotides 1 to 465 of murine ISG15 were PCR amplified using a 5' primer that introduced a BstEII restriction site but no FLAG epitope and a 3' primer containing GGT GGG TAA sequence and a BstEII site.
(iv) ISG15 LRLRAA. Nucleotides 1 to 465 of murine ISG15 were PCR amplified using a 5' primer that introduced a BstEII restriction site but no FLAG epitope and a 3' primer containing GCG GCG TAA sequence and a BstEII site.
The correct sequence for each virus was confirmed by sequencing. Viral stocks were generated by in vitro transcription of linearized cDNA templates, followed by transfection of the transcripts with Lipofectamine (Gibco-BRL, Gaithersburg, MD) into BHK cells. Supernatants were harvested after 24 to 48 h, clarified by centrifugation, and stored at –80°C. Titers of the stocks were determined by standard plaque assay on BHK cells. To examine the growth characteristics of these viruses, BHK cells were infected at a multiplicity of infection of 5.0 in a volume of 0.150 ml at 37°C for 1 h. The cells were then washed two times with 1x phosphate-buffered saline, and then 1 ml of Dulbecco's modified Eagle's medium (DMEM) supplemented with 1% fetal calf serum was added to each well. Samples were freeze-thawed three times and then titered by plaque assay on BHK cells. To determine the expression of ISG15 from the recombinant viruses, BHK cells (2 x 105) were infected at a multiplicity of infection of 5.0 with mock supernatant, dsTE12Q, ISG15 LRLRGG, or ISG15 LRLRAA. Total cell lysates were harvested 18 h postinfection and analyzed by Western blotting.
In vivo induction of ISG15. 129 or IFN-/R–/– mice were infected at 7 days of age with 1,000 PFU dsTE12Q intracranially (i.c.) into the right cerebral hemisphere. The right cerebral hemisphere was harvested into 1 ml of DMEM and then homogenized with 100 μl of 1.0-mm-diameter zirconia-silica beads (Biospec, Inc., Bartlesville, OK) at 3,200 rpm for 2 min with a Mini-Bead-Beater-8 (Biospec, Inc.). The brain homogenate was mixed with sodium dodecyl sulfate loading buffer and boiled for 30 min prior to Western blot analysis.
293T transfection system. Plasmids encoding UBE1L (pCAGGS-HA-UBE1L) and UbcM8 (pFlagCMV2-UbcM8) were kindly provided by Dong-Er Zhang (The Scripps Research Institute, La Jolla, California) (17). To generate ISG15 LRLRGG, nucleotides 1 to 465 of ISG15 were PCR amplified with a 5' primer containing a Lumiotag (6-amino-acid sequence) (New England Biolabs, Beverly, MA) and a HindIII site and a 3' primer containing a TAA sequence and a KpnI site. To generate ISG15 LRLRAA, nucleotides 1 to 465 of ISG15 were PCR amplified with the same 5' primer containing a Lumio tag and a HindIII site and a 3' primer containing GCA GCA TAA sequence and a KpnI site. Both products were cloned into pcDNA3.1/Hygro (Invitrogen, Carlsbad, California). The correct sequences were confirmed by sequencing, and expression of ISG15 was confirmed by Western blot analysis with anti-ISG15 antibodies.
For transfections, 293T cells were plated on 12-well plates and transfected using Lipofectamine 2000 (Invitrogen) following the manufacturer's instructions. Forty-eight hours following transfection, samples were harvested into 100 μl Laemmli sample buffer, boiled, separated on a 4 to 15% gradient gel (Bio-Rad, Hercules, CA), and subjected to Western blotting.
Western blot analysis. Samples were subjected to protein electrophoresis on 4 to 15% Tris gradient gels (Bio-Rad). The gels were then transferred to a polyvinylidene difluoride membrane. For ISG15 expression, the blots were either incubated with a rabbit anti-ISG15 polyclonal serum diluted 1:3,000 and then developed with a horseradish peroxidase (HRP)-conjugated goat anti-rabbit antiserum (Jackson Immunoresearch, West Grove, Pennsylvania) diluted 1:5,000 or probed with 17 μg of anti-ISG15 monoclonal antibody (MAb) (3C2) and then developed with goat anti-Armenian hamster HRP antibody diluted 1:5,000 (Jackson Immunoresearch). For loading controls, parallel blots were probed with an anti--actin MAb (clone AC-74) and then developed with a goat anti-mouse HRP-conjugated secondary antibody (Jackson Immunoresearch). All blots were developed with ECL PLUS chemiluminescent reagent (Amersham, Piscataway, NJ).
Viruses and viral studies. For some experiments (see Fig. 2), 8- to 10-week-old mice were infected with either 0.1 or 1.0 PFU of recombinant chimeric virus diluted in 10 μl of Hank's balanced salt solution and injected i.c. into the right cerebral hemisphere. For subsequent experiments, 8- to 12-week-old mice were infected subcutaneously (s.c.) with either 1 x 106 or 5 x 106 PFU of recombinant chimeric virus diluted in 50 μl of Hank's balanced salt solution in the left hind footpad. Viral titers were determined in organs harvested in 1 ml of DMEM without fetal bovine serum and homogenized with 100 μl of 1.0-mm-diameter zirconia-silica beads at 3,200 rpm for 2 min with a Mini-Bead-Beater-8 prior to plaque assay on BHK cells. The limit of detection of the assay is 50 PFU for all organs except the liver, for which it is 500 PFU.
Histopathology. Mice were infected s.c. with 5 x 106 PFU, and organs were harvested 5 days postinfection and fixed with 4% paraformaldehyde. The left cerebral hemisphere was embedded in paraffin, and a series of 4-μm sagittal sections were cut from medial to lateral. Sequential sections were stained with hematoxylin and eosin to detect histopathology, and immunostaining was done to detect Sindbis antigen. All slides were evaluated and scored blindly by one of the authors (R.E.S.) for the presence of pathological changes consistent with meningoencephalitis. Sindbis virus antigen was detected using a polyclonal anti-Sindbis antibody kindly provided by Dianne Griffin (22). Briefly, anti-Sindbis antibody was diluted 1:400 in phosphate-buffered saline with 2% goat sera, and tissue sections were incubated overnight at 4°C. HRP-conjugated goat anti-rabbit antibody (Jackson Immunoresearch) was diluted 1:250 in blocking buffer, and sections were incubated overnight at 4°C. Antigen was visualized by a 2-min staining with a solution of 3,3'-diaminobenzidine (Vector Laboratories, Burlingame, CA). No staining was observed in infected tissue incubated with control rabbit sera or mock-infected tissues incubated with immune sera. To determine the percentage of sections positive for Sindbis virus antigen, four independent sections from each animal were stained for anti-Sindbis virus and scored for the presence or absence of staining in the cerebellum, cerebral cortex, and hippocampus. The number of sections scored as positive was then divided by the number of total sections counted for each group and was reported as the percentage of positive sections.
Statistical methods. All data were analyzed with Prizm software (GraphPad Software, San Diego, CA). Survival data were analyzed by the Mantel-Haenzsel test, with death as the primary variable. Acute titer data were analyzed using the Mann-Whitney test. Error bars in figures represent the standard error of the mean.
RESULTS
Induction of ISG15 following in vivo Sindbis virus infection is markedly decreased in IFN-/R–/– mice compared to wt mice. Previous work has shown that ISG15 is induced in vitro by IFN-/, LPS, and viral infections (20, 26, 39). It has also been shown that ISG15 protein is induced in vivo during LCMV infection in the meningeal, ependymal, and choroid plexus cells following i.c. infection (39). Following infection of neonatal mice with Sindbis virus (dsTE12Q), ISG15 RNA is upregulated by >100-fold over mock-infected mice (13, 20). We wanted to determine if ISG15 protein was upregulated in vivo during Sindbis virus infection and whether this induction was dependent upon IFN-/. We therefore examined the expression of ISG15 in wild type (wt) and IFN-/R–/– mice following intracerebral Sindbis virus infection. In uninfected wt and IFN-/R–/– mice, there was no detectable ISG15 in brain lysates (Fig. 1). Following Sindbis virus infection of wt mice, free ISG15 and a small amount of ISG15 conjugates were detected 1 day postinfection. In contrast, neither free ISG15 nor ISG15 conjugates were detected in IFN-/R–/– mice 1 day after Sindbis infection. By 3 days postinfection, ISG15 and its conjugates could readily be detected in the brains of infected wt mice, and this response was further increased at 6 days postinfection (Fig. 1). Expression could not be analyzed after 1 day postinfection in IFN-/R–/– mice, as they succumbed to infection within 26 h. In adult IFN-/R–/– mice infected with Sindbis virus by the subcutaneous route, barely detectable amounts of ISG15 and ISG15 conjugates were found within the brain at 5 days postinfection (data not shown). These results demonstrate that ISG15 is induced following Sindbis virus infection, and this induction is markedly attenuated in animals lacking the IFN-/ receptor.
Expression of ISG15 by recombinant chimeric Sindbis virus in IFN-/R–/– mice attenuates infection. To evaluate the ability of ISG15 to function as an antiviral molecule in vivo, we utilized a recombinant chimeric Sindbis virus to express ISG15 in IFN-/R–/– mice and then evaluated these animals for attenuation of infection. The recombinant chimeric Sindbis virus system has been well characterized and was previously used to overexpress a variety of genes both in vitro and, more importantly, in vivo (12, 13, 22-24), providing a system to identify the potential antiviral effects of specific genes.
To examine the ability of ISG15 to function as an antiviral molecule, we first generated two recombinant Sindbis viruses. The first virus expressed a FLAG-tagged full-length ISG15 molecule (ISG15 FULL), and the second virus contained the ISG15 open reading frame with the initial ATG deleted (ISG15 MUTANT) (Fig. 2A). These viruses, as well as the parental virus dsTE12Q (no insert), were used to infect IFN-/R–/– mice, and attenuation of infection was determined by following the animals for lethality. While animals infected with either dsTE12Q or ISG15 MUTANT rapidly succumbed to the infection, IFN-/R–/– mice infected with ISG15 FULL were protected from lethality at both 0.1 PFU i.c. (P = 0.0020) (Fig. 2B) and 1.0 PFU i.c. (P = 0.0145) (Fig. 2C). One- to 2-day-old CD1 mice infected with ISG15 FULL were also partially protected from lethality, with 20% of the mice infected with ISG15 FULL surviving at 21 days postinfection compared to 2% survival in the mice infected with ISG15 MUTANT (data not shown). These results provide the first direct evidence that ISG15 can function as an antiviral molecule in vivo when expressed in infected cells.
The C-terminal LRLRGG domain of ISG15 is required for ISG15 conjugate formation. Structure-function studies of ubiquitin have previously demonstrated that the LRLRGG motif, and the terminal diglycine in particular, are required for ubiquitin to form intracellular conjugates (44). As ISG15 shares the LRLRGG motif with ubiquitin and other UBLs, such as RUB-1 and SUMO (10), we wanted to ascertain if the LRLRGG domain of ISG15 was also required for intracellular-conjugate formation.
A transient-transfection assay (17) was performed in which an ISG15 E1 (UBE1L) and E2 (UbcM8) and either wild-type ISG15 with an intact LRLRGG domain or a mutant form of ISG15 in which LRLRGG was mutated to LRLRAA were transfected into 293 T cells. Protein ISGylation was assessed by Western blot analysis. Cells transfected with just E1 and E2 did not display any background ISG15 expression or conjugation (Fig. 3A, lanes 1 and 2). ISG15 LRLRGG transfection alone produced free ISG15 expression (Fig. 3A, lane 3), while the addition of E1 and E2 resulted in protein ISGylation (Fig. 3A, lane 4). In contrast, if ISG15 LRLRAA was transfected into the cells, only free ISG15 was detected, even in the presence of E1 and E2 (Fig. 3A, lanes 5 and 6). These data demonstrate that, similar to ubiquitin, the LRLRGG domain of ISG15 is necessary for intracellular conjugate formation.
The C-terminal LRLRGG motif is important for ISG15's antiviral activity. Since the C-terminal LRLRGG motif is critical for ISG15 to form intracellular conjugates, we wanted to determine whether the LRLRGG motif was also important for ISG15's antiviral activity. We generated additional recombinant chimeric Sindbis viruses that expressed ISG15 either with an intact LRLRGG motif (ISG15 LRLRGG) or with a diglycine-to-dialanine mutation (ISG15 LRLRAA) (Fig. 2A). To avoid potential complications, we also removed the FLAG epitope tag from these viruses and detected the expression of ISG15 protein with a monoclonal antibody generated against murine ISG15. Following the generation of these recombinant viruses, we infected BHK cells with either the parental dsTE12Q virus or ISG15 containing Sindbis viruses. There was no evidence of ISG15 expression in either mock- or dsTE12Q-infected BHK cells (Fig. 3B, lanes 1 and 2). In contrast, BHK cells infected with either ISG15 LRLRGG or ISG15 LRLRAA expressed similar levels of ISG15 (Fig. 3B, lanes 3 and 4). All of the viruses (dsTE12Q, ISG15 LRLRGG, and ISG15 LRLRAA) grew with similar kinetics and to similar final titers under single-step growth conditions in both BHK cells and IFN-/R–/– murine embryonic fibroblasts (MEFs) (Fig. 3C).
We next wanted to determine what effect the LRLRAA mutation would have on ISG15's ability to protect IFN-/R–/– mice from Sindbis virus infection. Since previous studies evaluating Sindbis virus pathogenesis in IFN-/R–/– mice had utilized a subcutaneous route of infection, we infected mice by the s.c. route for the remainder of the experiments in this study (9, 40). Similar to the ISG15 FULL virus, ISG15 LRLRGG protected IFN-/R–/– mice from lethality, with 100% of the mice surviving at 1 x 106 PFU s.c. (P = 0.0291) (Fig. 3D) and >80% surviving at 5 x 106 PFU s.c. (P = 0.0106) (Fig. 3E) compared to 60% and 20% survival with dsTE12Q. Mutating the LRLRGG domain to LRLRAA abrogated this protection, as mice infected with ISG15 LRLRAA had only 60% survival at 1 x 106 PFU s.c. (P = 0.9612) (Fig. 3D) and only 30% survival at 5 x 106 PFU s.c. (P = 0.7332) (Fig. 3E). A similar loss of protection was seen when a virus expressing ISG15 lacking the entire C-terminal LRLRGG motif was used (data not shown). We also saw no attenuation of infection when a Sindbis virus expressing a different ISG (ISG12) was used to infect IFN-/R–/– mice (data not shown).
It has been shown that, following subcutaneous infection with Sindbis virus strain TR339, mice with an intact interferon system limited viral replication and dissemination, with clearance of nearly all organs by 72 to 96 h postinfection (40). In contrast, IFN-/R–/– mice develop a rapid dissemination of TR339, with high titers (>104 PFU) found in multiple organs, and eventually succumb to the infection (40). We therefore wanted to determine if the IFN-/R–/– mice that survived infection with ISG15 LRLRGG had limited viral replication and dissemination or still developed a high-titer systemic Sindbis infection. IFN-/R–/– mice were infected with either dsTE12Q, ISG15 LRLRGG, or ISG15 LRLRAA and were sacrificed at either 1, 3, or 5 days postinfection. Viral titers were then determined for various organs. As expected, IFN-/R–/– mice infected with either dsTE12Q or ISG15 LRLRAA developed a systemic infection with titers of >104 PFU detected in the serum, spleen, liver, and lung as early as 1 day postinfection and in the brain and spinal cord by 3 days postinfection (Fig. 4). When viral titers were analyzed in mice infected with ISG15 LRLRGG, we found that they also developed a systemic infection, with equivalent titers being found in all organs at 1 day postinfection. This indicates that ISG15 expression did not alter the initial Sindbis replication and dissemination.
However, by 3 days postinfection, mice infected with ISG15 LRLRGG had at least 10-fold-lower viral titers (compared to mice infected with ISG15 LRLRAA) in the serum (P < 0.05), spleen (P < 0.005), lung (P < 0.05), and liver (P < 0.0005). By 5 days postinfection, these effects were even more dramatic. Viral titers could not be detected in the serum or liver of mice infected with ISG15 LRLRGG. In the brain (P < 0.005), spinal cord (P < 0.0005), spleen (P < 0.0005), and lung (P < 0.005), viral titers were decreased by 10- to 100-fold compared to the titers found in mice infected with dsTE12Q or ISG15 LRLRAA (Fig. 4). Interestingly, in both the brain and spinal cord, there was also a statistically significant (P < 0.005) decrease in viral titers in mice infected with ISG15 LRLRAA compared to those infected with dsTE12Q. Thus, while the overexpression of ISG15 did not limit the initial high-titer, rapid systemic dissemination of Sindbis virus infection observed in IFN-/R–/– mice, it did result in a significant decrease in viral titers in multiple organs at both 3 and 5 days postinfection. Both lethality data and viral titers demonstrate that ISG15's LRLRGG motif plays an important role in its antiviral activity in this in vivo system.
ISG15 expression reduces Sindbis virus-induced pathology in the brain. To further study the effects of ISG15 expression on Sindbis virus infection in vivo, we examined brain sections 5 days after infection for histopathology and for viral-antigen expression as detected by immunohistochemistry (IHC). IFN-/R–/– mice infected with dsTE12Q had clear evidence of meningoencephalitis, with inflammation of the meninges, vascular cuffing, and focal parenchymal involvement characterized by neuronal loss, pyknotic nuclei, and gliosis. The localization of these focal lesions correlated with the presence of viral antigen detected by IHC. Mice infected with ISG15 LRLRAA also demonstrated evidence of meningoencephalitis, although the severity was decreased compared to animals infected with dsTE12Q, consistent with the decreased viral titers detected in the brains of mice infected with ISG15 LRLRAA on day 5 (Fig. 4). When brain sections were analyzed by IHC, foci of viral antigen were detected in the cerebral cortex and cerebellum, similar to dsTE12Q (Fig. 5A, B, and D). In contrast, animals infected with ISG15 LRLRGG not only had 100-fold-lower viral titers in their brains (Fig. 4), but also demonstrated decreased histopathological evidence of meningoencephalitis and decreased viral antigen immunoreactivity in the cerebral cortex and, most strikingly, in the cerebellum (Fig. 5C and D). No viral staining was detected in any of the sections of the cerebellum in 11 out of 13 animals infected with ISG15 LRLRGG. These data, in combination with the lethality and viral-titer data, provide strong evidence that ISG15 functions as an antiviral molecule in this system and that the LRLRGG motif of ISG15 plays an important role in this activity.
DISCUSSION
IFN-/ play a critical role during viral infection through their direct inhibition of viral replication and by mediating a variety of immunomodulatory effects. While PKR, RNAse L, Mx, and inducible nitric oxide synthase have been shown to inhibit the replication of certain viruses, there are several lines of evidence that indicate there are additional IFN-induced antiviral genes yet to be identified. First, the action of the known IFN-induced antiviral genes is virus specific. For example, PKR–/– mice are more susceptible to vesicular stomatitis virus and encephalomyocarditis virus but clear influenza and vaccinia virus infections (1, 46). Pretreatment of triply deficient mice (Mx–/–, PKR–/–, and RNAse L–/–) with IFN protects them from lethal encephalomyocarditis virus infection (51). Furthermore, while IFN-/R–/– mice rapidly succumb to fatal Sindbis virus infection, triply deficient mice develop only subclinical infection (41). In this report, we explore the ability of ISG15 to function as an antiviral molecule. Following Sindbis virus infection, we demonstrate that mice rapidly upregulate the expression of ISG15 and ISG15 conjugates and that this upregulation is markedly decreased in animals lacking the IFN-/ receptor. Reconstituting the expression of ISG15 in virus-infected cells with a recombinant Sindbis virus system attenuated viral infection in IFN-/R–/– mice, as manifested by decreased lethality, decreased viral titers, and decreased viral antigen staining within the brains of infected mice. The antiviral activity was largely dependent upon the LRLRGG motif, since mutating this motif abrogated the protection. These data provide in vivo evidence that ISG15 functions as an antiviral molecule.
Possible mechanism(s) of ISG15's antiviral activity. Previous studies have suggested two mechanisms by which ISG15 may exert its antiviral activity—protein conjugation and cytokine activity (5, 18, 26, 31, 35, 38, 39, 48). Similar to ubiquitin, the LRLRGG motif is required for ISG15 conjugate formation (Fig. 3A). However, only a limited number of ISG15 targets have been defined, and the fates of these ISGylated proteins are still unknown (8a, 11, 27, 49a). Furthermore, the relevance of these known targets to in vivo antiviral activity is currently unclear. Studies are ongoing by several groups to identify additional ISG15 conjugation targets and to understand the fates of these ISGylated proteins. In vivo analysis of Sindbis viruses engineered to express either ISG15 LRLRGG or LRLRAA demonstrated that the LRLRGG motif is important for ISG15's antiviral activity (Fig. 3D and E). These results support the possibility that ISG15's antiviral activity is due to its conjugation to as-yet-unidentified targets. These targets could be host proteins that either directly effect viral replication or in some way modulate the immune response. Alternatively, the targets could be viral proteins that are conjugated to ISG15 and alter various aspects of the viral life cycle or the immune response to the virus. Modification of viral proteins by either ubiquitin or UBLs, such as SUMO, has been seen with both RNA and DNA viruses. However, the functional consequences of these modifications in many cases are unknown. The Gag protein of human immunodeficiency virus is ubiquitinated, and this modification increases its ability to recruit the endosomal complexes required for transport (ESCRT-I) to the plasma membrane for viral budding (37). The IE1-72kDa and IE-2-86kDa proteins of HCMV are sumoylated (32). The generation of a mutant virus in which the sumoylation site of IE1-72kDa was disrupted resulted in a virus with decreased in vitro growth and reduced accumulation of the IE2 transcripts and protein (32). It is therefore possible that ISG15, similar to other UBLs, can modify viral proteins and alter the viral life cycle. One hypothesis is that ISG15 functions by antagonizing ubiquitin's activity by competing for ubiquitin conjugation sites on specific host or viral proteins, for example, those required for human immunodeficiency virus budding or for HCMV IE1 function.
All members of the ISG15 conjugation pathway are regulated by IFN-/. This is evident in IFN-/R–/– cells and mice, in which there is markedly decreased expression of ISG15 and ISG15 conjugates (Fig. 1). Therefore, despite driving expression of ISG15 in infected cells, the efficient induction of the conjugation cascade may not occur in this system. While IFN-/ are potent inducers of the ISG15 conjugation cascade, previous work has also shown that IFN- can induce low-level expression of ISG15 and ISG15 conjugates (10, 19). Analysis of animals 5 days postinfection with any of the recombinant Sindbis viruses demonstrates the presence of IFN- in their sera (data not shown). IFN- induced by the viral infection may allow low-level induction of the ISG15 conjugation cascade. ISG15 and ISG15 conjugate expression in IFN-/R–/– MEFs can also be induced by pI:pC and LPS (D. J. Lenschow and N. V. Giannakopoulos, unpublished results), albeit at much lower levels than in wild-type cells. Therefore, the overexpression of ISG15 in infected cells, which will also be receiving signals via Toll-like receptor ligation or additional cytokine stimuli, such as IFN-, may reconstitute sufficient ISG15 conjugation to mediate an antiviral effect that is dependent upon conjugation of ISG15 to target proteins.
Alternatively, ISG15's antiviral activity may be completely independent of its ability to modify target proteins. Human ISG15 has been reported to function as a cytokine. Recombinant human ISG15 induces NK cell proliferation, augments lymphokine-activated-killer activity, and stimulates the production of IFN- (5, 38). Additional studies have suggested that ISG15 may induce dendritic cell maturation and neutrophil recruitment (34, 35). Therefore, in addition to its capacity to complex with intracellular proteins, ISG15 may also function as a cytokine whose activity could impact upon either the innate or adaptive immune response. If unconjugated ISG15 either within the cell or functioning as a cytokine is responsible for the antiviral activity, it appears that the LRLRGG motif is also important for this activity. Studies of recombinant human ISG15 demonstrated that the LRLRGG motif had to be exposed for full cytokine activity, since full-length recombinant ISG15 had no activity (5). This motif may be required for processing and/or release of ISG15 from the cell or may be important for receptor binding.
Utilization of recombinant Sindbis virus system as a screen for IFN-induced antiviral molecules. IFN-/ induce hundreds of genes, a subset of which are likely to be antiviral. The recombinant chimeric Sindbis virus has been used to express antiapoptotic genes and antibodies in vivo to evaluate the roles of these proteins during Sindbis virus pathogenesis (12, 23, 24). In this study, we utilized this system to express ISG15 in IFN-/R–/– mice to determine if the expression of this ISG could restore any of the antiviral activity mediated by IFN-/. Expression of individual ISGs in the recombinant Sindbis virus system could be used to screen a variety of ISGs to determine if they exhibited antiviral activity. By utilizing an in vivo approach, we can identify not only those genes that directly inhibit viral replication, but also those genes that can impact upon the immune system. Testing these viruses in the IFN-/R–/– mice allows evaluation of other ISGs in isolation. Finally, since this system is easy to manipulate, structure-function analysis can be readily performed.
There are also limitations to this in vivo genetic screen for IFN-induced antiviral molecules. First, many antiviral genes are virus specific. Since the readout for this screen is the attenuation of Sindbis infection, it may miss antiviral genes that have activity against other viruses. This can be addressed by using other viral systems to express host genes in IFN-/R–/– mice. Second, the initial readout utilized in this screen is lethality, which may be too stringent to identify subtle effects. Using less stringent screens, such as the analysis of viral titers and immunohistochemistry, may be beneficial in this regard. In addition, this screen depends on the overexpression of the IFN-induced molecule, perhaps resulting in nonphysiological effects. In this study, the comparison between LRLRGG and LRLRAA addresses this concern by allowing a comparison between two overexpressed proteins in vivo. Furthermore, since this screen is performed in mice lacking IFN-/ receptors, expressed genes are evaluated in isolation from other maximally induced ISGs. Given the potential redundancy between ISGs, an effect seen in an IFN-/R–/– mouse may be less significant in an intact animal. The ultimate validation of this as an in vivo genetic approach to the identification of novel IFN-induced antiviral genes will be the analysis of mice deficient in selected ISGs and determining if these mice display altered susceptibility to viral infections.
ACKNOWLEDGMENTS
H.W.V. was supported by NIH grant U54 AI057160 and the Pfizer Biomedical Agreement. D.J.L. was supported by a Pfizer Postdoctoral Fellowship in Immunology and NIH Career Development Award K08 AI59390-01. B.L. was supported by NIH grant R01 A144157. The Washington University Hybridoma Core Facility was supported in part by NIH grant P30 AR048335.
We thank Darren Kreamalmayer for his outstanding expertise in animal care.
REFERENCES
Abraham, N., D. F. Stojdl, P. I. Duncan, N. Methot, T. Ishii, M. Dube, B. C. Vanderhyden, H. L. Atkins, D. A. Gray, M. W. McBurney, A. E. Koromilas, E. G. Brown, N. Sonenberg, and J. C. Bell. 1999. Characterization of transgenic mice with targeted disruption of the catalytic domain of the double-stranded RNA-dependent protein kinase, PKR. J. Biol. Chem. 274:5953-5962.
Arnheiter, H., S. Skuntz, M. Noteborn, S. Chang, and E. Meier. 1990. Transgenic mice with intracellular immunity to influenza virus. Cell 62:51-61.
Biron, C. A. 2001. Interferons alpha and beta as immune regulators—a new look. Immunity 14:661-664.
Blomstrom, D. C., D. Fahey, R. Kutny, B. D. Korant, and E. Knight, Jr. 1986. Molecular characterization of the interferon-induced 15-kDa protein. Molecular cloning and nucleotide and amino acid sequence. J. Biol. Chem. 261:8811-8816.
D'Cunha, J., E. Knight, Jr., A. L. Haas, R. L. Truitt, and E. C. Borden. 1996. Immunoregulatory properties of ISG15, an interferon-induced cytokine. Proc. Natl. Acad. Sci. USA 93:211-215.
D'Cunha, J., S. Ramanujam, R. J. Wagner, P. L. Witt, E. Knight, Jr., and E. C. Borden. 1996. In vitro and in vivo secretion of human ISG15, an IFN-induced immunomodulatory cytokine. J. Immunol. 157:4100-4108.
Der, S. D., A. Zhou, B. R. Williams, and R. H. Silverman. 1998. Identification of genes differentially regulated by interferon alpha, beta, or gamma using oligonucleotide arrays. Proc. Natl. Acad. Sci. USA 95:15623-15628.
Dunn, G. P., A. T. Bruce, K. C. F. Sheehan, V. Shankaran, R. Uppaluri, J. D. Bui, M. S. Diamond, C. M. Koebel, C. Arthur, J. M. White, and R. D. Schreiber. 2005. A critical function for type I interferons in cancer immunoediting. Nat. Immunol. 6:722-729.
Giannakopoulos, N. V., J. K. Luo, V. Papoz, W. Zou, D. J. Lenschow, B. S. Jacobs, E. C. Borden, J. Li, H. W. Virgin, and D. E. Zhang. 2005. Proteomic identification of proteins conjugated to ISG15 in mouse and human cells. Biochem. Biophys. Res. Commun. 336:496-506.
Gil, M. P., E. Bohn, A. K. O'Guin, C. V. Ramana, B. Levine, G. R. Stark, H. W. Virgin, and R. D. Schreiber. 2001. Biologic consequences of Stat1-independent IFN signaling. Proc. Natl. Acad. Sci. USA 98:6680-6685.
Haas, A. L., P. Ahrens, P. M. Bright, and H. Ankel. 1987. Interferon induces a 15-kilodalton protein exhibiting marked homology to ubiquitin. J. Biol. Chem. 262:11315-11323.
Hamerman, J. A., F. Hayashi, L. A. Schroeder, S. P. Gygi, A. L. Haas, L. Hampson, P. Coughlin, R. Aebersold, and A. Aderem. 2002. Serpin 2a is induced in activated macrophages and conjugates to a ubiquitin homolog. J. Immunol. 168:2415-2423.
Hardwick, J. M., and B. Levine. 2000. Sindbis virus vector system for functional analysis of apoptosis regulators. Methods Enzymol. 322:492-508.
Johnston, C., W. Jiang, T. Chu, and B. Levine. 2001. Identification of genes involved in the host response to neurovirulent alphavirus infection. J. Virol. 75:10431-10445.
Kapadia, S. B., B. Levine, S. H. Speck, and H. W. Virgin. 2002. Critical role of complement and viral evasion of complement in acute, persistent, and latent gamma-herpesvirus infection. Immunity 17:143-155.
Karupiah, G., J. H. Chen, C. F. Nathan, S. Mahalingam, and J. D. MacMicking. 1998. Identification of nitric oxide synthase 2 as an innate resistance locus against ectromelia virus infection. J. Virol. 72:7703-7706.
Karupiah, G., Q. Xie, R. M. L. Buller, C. Nathan, C. Duarte, and J. D. MacMicking. 1993. Inhibition of viral replication by interferon-gamma induced nitric oxide synthase. Science 261:1445-1448.
Kim, K. I., N. V. Giannakopoulos, H. W. Virgin, and D. E. Zhang. 2004. Interferon-inducible ubiquitin E2, Ubc8, is a conjugating enzyme for protein ISGylation. Mol. Cell. Biol. 24:9592-9600.
Knight, E., Jr., and B. Cordova. 1991. IFN-induced 15-kDa protein is released from human lymphocytes and monocytes. J. Immunol. 146:2280-2284.
Korant, B. D., D. C. Blomstrom, G. J. Jonak, and E. Knight, Jr. 1984. Interferon-induced proteins. Purification and characterization of a 15,000-dalton protein from human and bovine cells induced by interferon. J. Biol. Chem. 259:14835-14839.
Labrada, L., X. H. Liang, W. Zheng, C. Johnston, and B. Levine. 2002. Age-dependent resistance to lethal alphavirus encephalitis in mice: analysis of gene expression in the central nervous system and identification of a novel interferon-inducible protective gene, mouse ISG12. J. Virol. 76:11688-11703.
Leib, D. A., M. A. Machalek, B. R. Williams, R. H. Silverman, and H. W. Virgin. 2000. Specific phenotypic restoration of an attenuated virus by knockout of a host resistance gene. Proc. Natl. Acad. Sci. USA 97:6097-6101.
Levine, B., J. E. Goldman, H. H. Jiang, D. E. Griffin, and J. M. Hardwick. 1996. Bcl-2 protects mice against fatal alphavirus encephalitis. Proc. Natl. Acad. Sci. USA 93:4810-4815.
Liang, X. H., Jiang, H. H., and B. Levine. 1997. Expression of a biologically active antiviral antibody using a sindbis virus vector system. Molecular Immunology 34:907-917.
Liang, X. H., L. K. Kleeman, H. H. Jiang, G. Gordon, J. E. Goldman, G. Berry, B. Herman, and B. Levine. 1998. Protection against fatal Sindbis virus encephalitis by beclin, a novel Bcl-2-interacting protein. J. Virol. 72:8586-8596.
Liu, L. Q., R. Ilaria, Jr., P. D. Kingsley, A. Iwama, R. A. van Etten, J. Palis, and D. E. Zhang. 1999. A novel ubiquitin-specific protease, UBP43, cloned from leukemia fusion protein AML1-ETO-expressing mice, functions in hematopoietic cell differentiation. Mol. Cell. Biol. 19:3029-3038.
Loeb, K. R., and A. L. Haas. 1992. The interferon-inducible 15-kDa ubiquitin homolog conjugates to intracellular proteins. J. Biol. Chem. 267:7806-7813.
Malakhov, M. P., K. I. Kim, O. A. Malakhova, B. S. Jacobs, E. C. Borden, and D. E. Zhang. 2003. High-throughput immunoblotting: ubiquitin-like protein ISG15 modifies key regulators of signal transduction. J. Biol. Chem. 278:16608-16613.
Malakhov, M. P., O. A. Malakhova, K. I. Kim, K. J. Ritchie, and D. E. Zhang. 2002. UBP43 (USP18) specifically removes ISG15 from conjugated proteins. J. Biol. Chem. 277:9976-9981.
Malakhova, O. A., M. Yan, M. P. Malakhov, Y. Yuan, K. J. Ritchie, K. I. Kim, L. F. Peterson, K. Shuai, and D. E. Zhang. 2003. Protein ISGylation modulates the JAK-STAT signaling pathway. Genes Dev. 17:455-460.
Muller, U., S. Steinhoff, L. F. L. Reis, S. Hemmi, J. Pavlovic, R. M. Zinkernagel, and M. Aguet. 1994. Functional role of type I and type II interferons in antiviral defense. Science 264:1918-1921.
Narasimhan, J., J. L. Potter, and A. L. Haas. 1996. Conjugation of the 15-kDa interferon-induced ubiquitin homolog is distinct from that of ubiquitin. J. Biol. Chem. 271:324-330.
Nevels, M., W. Brune, and T. Shenk. 2004. SUMOylation of the human cytomegalovirus 72-kilodalton IE1 protein facilitates expression of the 86-kilodalton IE2 protein and promotes viral replication. J. Virol. 78:7803-7812.
Nicholl, M. J., L. H. Robinson, and C. M. Preston. 2000. Activation of cellular interferon-responsive genes after infection of human cells with herpes simplex virus type 1. J. Gen. Virol. 81:2215-2218.
Owhashi, M., Y. Taoka, K. Ishii, S. Nakazawa, H. Uemura, and H. Hambara. 2003. Identification of a ubiquitin family protein as a novel neutrophil chemotactic factor. Biochem. Biophys. Res. Commun. 309:533-539.
Padovan, E., L. Terracciano, U. Certa, B. Jacobs, A. Reschner, M. Bolli, G. C. Spagnoli, E. C. Borden, and M. Heberer. 2002. Interferon stimulated gene 15 constitutively produced by melanoma cells induces e-cadherin expression on human dendritic cells. Cancer Res. 62:3453-3458.
Player, M. R., and P. F. Torrence. 1998. The 2-5A system: modulation of viral and cellular processes through acceleration of RNA degradation. Pharmacol. Ther. 78:55-113.
Pornillos, O., J. E. Garrus, and W. I. Sundquist. 2002. Mechanisms of enveloped RNA virus budding. Trends Cell Biol. 12:569-579.
Recht, M., E. C. Borden, and E. Knight, Jr. 1991. A human 15-kDa IFN-induced protein induces the secretion of IFN-gamma. J. Immunol. 147:2617-2623.
Ritchie, K. J., C. S. Hahn, K. I. Kim, M. Yan, D. Rosario, L. Li, J. C. de la Torre, and D. E. Zhang. 2004. Role of ISG15 protease UBP43 (USP18) in innate immunity to viral infection. Nat. Immunol. 10:1374-1378.
Ryman, K. D., W. B. Klimstra, K. B. Nguyen, C. Biron, and R. E. Johnston. 2000. Alpha/beta interferon protects adult mice from fatal Sindbis virus infection and is an important determinant of cell and tissue tropism. J. Virol. 74:3366-3378.
Ryman, K. D., L. J. White, R. E. Johnston, and W. B. Klimstra. 2002. Effects of PKR/RNase L-dependent and alternative antiviral pathways on alphavirus replication and pathogenesis. Viral Immunol. 15:53-76.
Sen, G. C. 2001. Viruses and interferons. Annu. Rev. Microbiol. 55:255-281.
Sheehan, K. C., N. H. Ruddle, and R. D. Schreiber. 1989. Generation and characterization of hamster monoclonal antibodies that neutralize murine tumor necrosis factors. J. Immunol. 142:3884-3893.
Sloper-Mould, K. E., J. C. Jemc, C. M. Pickart, and L. Hicke. 2001. Distinct functional surface regions on ubiquitin. J. Biol. Chem. 276:30483-30489.
Stark, G. R., I. M. Kerr, B. R. Williams, R. H. Silverman, and R. D. Schreiber. 1998. How cells respond to interferons. Annu. Rev. Biochem. 67:227-264.
Stojdl, D. F., N. Abraham, S. Knowles, R. Marius, A. Brasey, B. D. Lichty, E. G. Brown, N. Sonenberg, and J. C. Bell. 2000. The murine double-stranded RNA-dependent protein kinase PKR is required for resistance to vesicular stomatitis virus. J. Virol. 74:9580-9585.
Yang, Y. L., L. F. Reis, J. Pavlovic, A. Aguzzi, R. Schafer, A. Kumar, B. R. Williams, M. Aguet, and C. Weissmann. 1995. Deficient signaling in mice devoid of double-stranded RNA-dependent protein kinase. EMBO J. 14:6095-6106.
Yuan, W., and R. M. Krug. 2001. Influenza B virus NS1 protein inhibits conjugation of the interferon (IFN)-induced ubiquitin-like ISG15 protein. EMBO J. 20:362-371.
Zhao, C., S. L. Beaudenon, M. L. Kelley, M. B. Waddell, W. Yuan, B. A. Schulman, J. M. Huibregtse, and R. M. Krug. 2004. The UbcH8 ubiquitin E2 enzyme is also the E2 enzyme for ISG15, an IFN-/-induced ubiquitin-like protein. Proc. Natl. Acad. Sci. USA 101:7578-7582.
Zhao, C., C. Denison, J. M. Huibregtse, S. Gygi, and R. M. Krug. 2005. Human ISG15 conjugation targets both IFN-induced and constitutively expressed proteins functioning in diverse cellular pathways. Proc. Natl. Acad. Sci. USA 102:10200-10205.
Zhou, A., J. Paranjape, T. L. Brown, H. Nie, S. Naik, B. Dong, A. Chang, B. Trapp, R. Fairchild, C. Colmenares, and R. H. Silverman. 1997. Interferon action and apoptosis are defective in mice devoid of 2',5'-oligoadenylate-dependent RNase L. EMBO J. 16:6355-6363.
Zhou, A., J. M. Paranjape, S. D. Der, B. R. Williams, and R. H. Silverman. 1999. Interferon action in triply deficient mice reveals the existence of alternative antiviral pathways. Virology 258:435-440.
Zhu, H., J. P. Cong, G. Mamtora, T. Gingeras, and T. Shenk. 1998. Cellular gene expression altered by human cytomegalovirus: global monitoring with oligonucleotide arrays. Proc. Natl. Acad. Sci. USA 95:14470-14475.(Deborah J. Lenschow, Nadi)
Immunology, Washington University School of Medicine, St. Louis, Missouri 63110
Division of Infectious Diseases, University of Texas Southwestern Medical Center, Dallas, Texas 75390-9113
Department of Medicine, Columbia University College of Physicians and Surgeons, New York, New York 10032
ABSTRACT
The innate immune response, and in particular the alpha/beta interferon (IFN-/) system, plays a critical role in the control of viral infections. Interferons and exert their antiviral effects through the induction of hundreds of interferon-induced (or -stimulated) genes (ISGs). While several of these ISGs have characterized antiviral functions, their actions alone do not explain all of the effects mediated by IFN-/. To identify additional IFN-induced antiviral molecules, we utilized a recombinant chimeric Sindbis virus to express selected ISGs in IFN-/ receptor (IFN-/R)–/– mice and looked for attenuation of Sindbis virus infection. Using this approach, we identified a ubiquitin homolog, interferon-stimulated gene 15 (ISG15), as having antiviral activity. ISG15 expression protected against Sindbis virus-induced lethality and decreased Sindbis virus replication in multiple organs without inhibiting the spread of virus throughout the host. We establish that, much like ubiquitin, ISG15 requires its C-terminal LRLRGG motif to form intracellular conjugates. Finally, we demonstrate that ISG15's LRLRGG motif is also required for its antiviral activity. We conclude that ISG15 can be directly antiviral.
INTRODUCTION
The innate immune response plays a critical role in controlling viral infections, with the alpha/beta interferon (IFN-/) system acting as an important component of this response. IFN-/ are expressed in response to viral infection and exert their effects through the IFN-/ receptor (IFN-/R), resulting in the activation of JAK/STAT-dependent and STAT-independent pathways and the subsequent induction of hundreds of genes containing interferon-stimulated response elements (9, 45). A subset of interferon-stimulated genes (ISGs) has been shown to directly inhibit viral replication (1, 2, 15, 16, 21, 36, 42, 47, 50). In addition to their antiviral activity, IFN-/ also mediate a variety of immunomodulatory effects that link the innate and adaptive immune responses. These include inducing the maturation of dendritic cells; enhancing NK cell cytotoxicity; increasing the production of cytokines, such as interleukin-15 and IFN-; and upregulating the expression of major histocompatibility complex class I molecules (3). The in vivo protection mediated by IFNs during viral infection is due in part to the direct inhibition of viral replication and in part to these immune-modulatory effects. While studies of protein kinase RNA activated (PKR), RNAse L, and Mx have shown these ISGs to play an important role during certain viral infections, these ISGs do not explain all of the effects mediated by IFN-/ (1, 46, 50, 51), suggesting the existence of additional IFN-induced antiviral pathways.
ISG15 is one of the earliest ISGs to be induced (up to 100-fold) following IFN stimulation (4, 7, 19, 20). It is expressed in a variety of cell types at low levels and induced within 6 h of stimulation with IFN-/, and to a lesser degree by IFN- (26). ISG15 has two ubiquitin-like domains, including a C-terminal LRLRGG motif through which it forms covalent conjugates with intracellular proteins (10, 26, 31). The formation of ISG15-protein conjugates parallels other ubiquitin-like proteins (UBLs), SUMO-1 and RUB-1, in the utilization of conjugating and deconjugating enzymes, many of which are distinct from the ubiquitin pathway (17, 28, 31, 48). The first member of the pathway to be identified was UBP43, a deconjugating enzyme for ISG15 (25, 28). This was followed by identification of UBE1L as an E1 (48) and the identification of UbcM8 as an E2 for ISG15 (17, 49). An E3 has yet to be reported. All members of this cascade, including ISG15, are tightly regulated by IFN-/. To date, only a few of the intracellular targets of ISG15 have been identified, including phospholipase C1, Jak1, Stat1, ERK1, and serpin 2a (8a, 11, 27, 49a). However, the fates of these and other ISGylated proteins remain to be elucidated. An extracellular form of ISG15 can also be found in the media of IFN-stimulated cells and in the sera of patients treated with IFN (6, 18, 35), and human recombinant ISG15 has been reported to exert cytokine activity (5, 18, 38). Despite these findings, the functional significance of ISG15 expression remains unclear.
There are, however, three types of data suggesting that ISG15 may play an important role during viral infection. First, ISG15 is rapidly upregulated by IFN, lipopolysaccharide (LPS), and viral infection. This has been shown both in vitro with microarray analysis (4, 7, 19, 20) and in vivo following viral infection (13, 20, 39). Infection of human fibroblasts with human cytomegalovirus (HCMV) resulted in a 150-fold increase in the expression of ISG15 mRNA (52). ISG15 RNA was also induced by herpes simplex virus infection (33). Infection of neonatal mice with Sindbis virus resulted in a >100-fold increase in ISG15 RNA expression in infected brain tissue (13, 20). ISG15 protein expression has also been localized to the meningeal, ependymal, and choroid plexus cells following lymphocytic choriomeningitis virus (LCMV) infection (39). Therefore the expression of ISG15 RNA and protein is markedly upregulated during viral infection. Second, mice lacking the deconjugating enzyme UBP43 demonstrate an increased resistance to infection with both LCMV and vesicular stomatitis virus (39). UBP43–/– cells have elevated basal levels of ISG15-protein conjugates, which are dramatically increased following IFN stimulation (28, 29). There is also an alteration in interferon signaling, with the UBP43–/– cells being hyperresponsive to interferon stimulation with prolonged STAT-1 phosphorylation, DNA binding, and IFN-mediated gene activation (29). While these data are consistent with a possible antiviral role of ISG15, they are not conclusive, since it is unclear if the antiviral state seen in the UBP43–/– mouse is due to increased levels of ISG15, to an enhanced IFN response that results in the upregulation of multiple ISGs, or to an as-yet-uncharacterized effect of UBP43. Finally, the influenza B protein NS1B binds to ISG15 and inhibits the coupling of UBE1L to ISG15 (48). This inhibits the formation of ISG15 conjugates in virus-infected cells. Therefore, at least one virus has evolved specific mechanisms to disrupt ISG15 function. These findings, in conjunction with the tight regulation of this conjugation cascade by IFN, suggest ISG15 may play an important role during viral infection. However, to date, ISG15 has no known function, and there has been no direct evidence that ISG15 is antiviral in vivo.
In this study, we utilize a recombinant, chimeric Sindbis virus system to overexpress ISG15 in IFN-/R–/– mice, which lack the ability to efficiently upregulate ISGs, including ISG15, in response to IFN-/. We show that expression of ISG15 in IFN-/R–/– mice attenuates Sindbis virus infection, providing in vivo evidence that ISG15 can function as an antiviral molecule.
MATERIALS AND METHODS
Mice. Mice were maintained at Washington University School of Medicine in accordance with all federal and university guidelines. 129, IFN-/R–/–, and IFN-//R–/– mice on the 129/SvPas background were initially obtained from M. Auget and then bred and maintained in our mouse colony (8, 9, 30). Unless otherwise noted, all animals were between 8 and 12 weeks of age and were age and sex matched within experiments.
Generation of ISG15 antibodies. Recombinant murine ISG15 was generated by cloning the open reading frame of murine ISG15 into pET-30a(+) (Novagen, Madison, WI) to express ISG15 with an amino-terminal His6 tag. His-tagged ISG15 protein (recombinant ISG15) was column purified on Ni-NTA His Bind Resin (Novagen) according to the manufacturer's instructions. Polyclonal rabbit antiserum (Cocalico, Reamstown, PA) was generated as previously described (14). Briefly, an initial inoculation of 100 μg recombinant ISG15 in complete Freund adjuvant was followed by 10 inoculations with 50 μg of recombinant ISG15 in incomplete Freund adjuvant 1 month apart. Sera were collected 7 days after boosts. Monoclonal antibodies were generated in Armenian hamsters in the Washington University School of Medicine Hybridoma Center using standard procedures (43).
Recombinant virus strains. Recombinant double-subgenomic Sindbis virus, dsTE12Q, was produced from a viral cDNA clone by in vitro transcription and RNA transfection of baby hamster kidney-21 (BHK) cells (12). ISG15-expressing Sindbis viruses were generated as follows.
(i) ISG15 FULL. Nucleotides 1 to 486 of murine ISG15 were PCR amplified using a 5' primer that contained a BstEII restriction site and FLAG epitope sequence and a 3' primer containing a BstEII site.
(ii) ISG15 MUTANT. The initial start codon of the ISG15 gene was deleted. The 5' primer contained a BstEII restriction site and a FLAG epitope, and the 3' primer contained a BstEII site.
(iii) ISG15 LRLRGG. Nucleotides 1 to 465 of murine ISG15 were PCR amplified using a 5' primer that introduced a BstEII restriction site but no FLAG epitope and a 3' primer containing GGT GGG TAA sequence and a BstEII site.
(iv) ISG15 LRLRAA. Nucleotides 1 to 465 of murine ISG15 were PCR amplified using a 5' primer that introduced a BstEII restriction site but no FLAG epitope and a 3' primer containing GCG GCG TAA sequence and a BstEII site.
The correct sequence for each virus was confirmed by sequencing. Viral stocks were generated by in vitro transcription of linearized cDNA templates, followed by transfection of the transcripts with Lipofectamine (Gibco-BRL, Gaithersburg, MD) into BHK cells. Supernatants were harvested after 24 to 48 h, clarified by centrifugation, and stored at –80°C. Titers of the stocks were determined by standard plaque assay on BHK cells. To examine the growth characteristics of these viruses, BHK cells were infected at a multiplicity of infection of 5.0 in a volume of 0.150 ml at 37°C for 1 h. The cells were then washed two times with 1x phosphate-buffered saline, and then 1 ml of Dulbecco's modified Eagle's medium (DMEM) supplemented with 1% fetal calf serum was added to each well. Samples were freeze-thawed three times and then titered by plaque assay on BHK cells. To determine the expression of ISG15 from the recombinant viruses, BHK cells (2 x 105) were infected at a multiplicity of infection of 5.0 with mock supernatant, dsTE12Q, ISG15 LRLRGG, or ISG15 LRLRAA. Total cell lysates were harvested 18 h postinfection and analyzed by Western blotting.
In vivo induction of ISG15. 129 or IFN-/R–/– mice were infected at 7 days of age with 1,000 PFU dsTE12Q intracranially (i.c.) into the right cerebral hemisphere. The right cerebral hemisphere was harvested into 1 ml of DMEM and then homogenized with 100 μl of 1.0-mm-diameter zirconia-silica beads (Biospec, Inc., Bartlesville, OK) at 3,200 rpm for 2 min with a Mini-Bead-Beater-8 (Biospec, Inc.). The brain homogenate was mixed with sodium dodecyl sulfate loading buffer and boiled for 30 min prior to Western blot analysis.
293T transfection system. Plasmids encoding UBE1L (pCAGGS-HA-UBE1L) and UbcM8 (pFlagCMV2-UbcM8) were kindly provided by Dong-Er Zhang (The Scripps Research Institute, La Jolla, California) (17). To generate ISG15 LRLRGG, nucleotides 1 to 465 of ISG15 were PCR amplified with a 5' primer containing a Lumiotag (6-amino-acid sequence) (New England Biolabs, Beverly, MA) and a HindIII site and a 3' primer containing a TAA sequence and a KpnI site. To generate ISG15 LRLRAA, nucleotides 1 to 465 of ISG15 were PCR amplified with the same 5' primer containing a Lumio tag and a HindIII site and a 3' primer containing GCA GCA TAA sequence and a KpnI site. Both products were cloned into pcDNA3.1/Hygro (Invitrogen, Carlsbad, California). The correct sequences were confirmed by sequencing, and expression of ISG15 was confirmed by Western blot analysis with anti-ISG15 antibodies.
For transfections, 293T cells were plated on 12-well plates and transfected using Lipofectamine 2000 (Invitrogen) following the manufacturer's instructions. Forty-eight hours following transfection, samples were harvested into 100 μl Laemmli sample buffer, boiled, separated on a 4 to 15% gradient gel (Bio-Rad, Hercules, CA), and subjected to Western blotting.
Western blot analysis. Samples were subjected to protein electrophoresis on 4 to 15% Tris gradient gels (Bio-Rad). The gels were then transferred to a polyvinylidene difluoride membrane. For ISG15 expression, the blots were either incubated with a rabbit anti-ISG15 polyclonal serum diluted 1:3,000 and then developed with a horseradish peroxidase (HRP)-conjugated goat anti-rabbit antiserum (Jackson Immunoresearch, West Grove, Pennsylvania) diluted 1:5,000 or probed with 17 μg of anti-ISG15 monoclonal antibody (MAb) (3C2) and then developed with goat anti-Armenian hamster HRP antibody diluted 1:5,000 (Jackson Immunoresearch). For loading controls, parallel blots were probed with an anti--actin MAb (clone AC-74) and then developed with a goat anti-mouse HRP-conjugated secondary antibody (Jackson Immunoresearch). All blots were developed with ECL PLUS chemiluminescent reagent (Amersham, Piscataway, NJ).
Viruses and viral studies. For some experiments (see Fig. 2), 8- to 10-week-old mice were infected with either 0.1 or 1.0 PFU of recombinant chimeric virus diluted in 10 μl of Hank's balanced salt solution and injected i.c. into the right cerebral hemisphere. For subsequent experiments, 8- to 12-week-old mice were infected subcutaneously (s.c.) with either 1 x 106 or 5 x 106 PFU of recombinant chimeric virus diluted in 50 μl of Hank's balanced salt solution in the left hind footpad. Viral titers were determined in organs harvested in 1 ml of DMEM without fetal bovine serum and homogenized with 100 μl of 1.0-mm-diameter zirconia-silica beads at 3,200 rpm for 2 min with a Mini-Bead-Beater-8 prior to plaque assay on BHK cells. The limit of detection of the assay is 50 PFU for all organs except the liver, for which it is 500 PFU.
Histopathology. Mice were infected s.c. with 5 x 106 PFU, and organs were harvested 5 days postinfection and fixed with 4% paraformaldehyde. The left cerebral hemisphere was embedded in paraffin, and a series of 4-μm sagittal sections were cut from medial to lateral. Sequential sections were stained with hematoxylin and eosin to detect histopathology, and immunostaining was done to detect Sindbis antigen. All slides were evaluated and scored blindly by one of the authors (R.E.S.) for the presence of pathological changes consistent with meningoencephalitis. Sindbis virus antigen was detected using a polyclonal anti-Sindbis antibody kindly provided by Dianne Griffin (22). Briefly, anti-Sindbis antibody was diluted 1:400 in phosphate-buffered saline with 2% goat sera, and tissue sections were incubated overnight at 4°C. HRP-conjugated goat anti-rabbit antibody (Jackson Immunoresearch) was diluted 1:250 in blocking buffer, and sections were incubated overnight at 4°C. Antigen was visualized by a 2-min staining with a solution of 3,3'-diaminobenzidine (Vector Laboratories, Burlingame, CA). No staining was observed in infected tissue incubated with control rabbit sera or mock-infected tissues incubated with immune sera. To determine the percentage of sections positive for Sindbis virus antigen, four independent sections from each animal were stained for anti-Sindbis virus and scored for the presence or absence of staining in the cerebellum, cerebral cortex, and hippocampus. The number of sections scored as positive was then divided by the number of total sections counted for each group and was reported as the percentage of positive sections.
Statistical methods. All data were analyzed with Prizm software (GraphPad Software, San Diego, CA). Survival data were analyzed by the Mantel-Haenzsel test, with death as the primary variable. Acute titer data were analyzed using the Mann-Whitney test. Error bars in figures represent the standard error of the mean.
RESULTS
Induction of ISG15 following in vivo Sindbis virus infection is markedly decreased in IFN-/R–/– mice compared to wt mice. Previous work has shown that ISG15 is induced in vitro by IFN-/, LPS, and viral infections (20, 26, 39). It has also been shown that ISG15 protein is induced in vivo during LCMV infection in the meningeal, ependymal, and choroid plexus cells following i.c. infection (39). Following infection of neonatal mice with Sindbis virus (dsTE12Q), ISG15 RNA is upregulated by >100-fold over mock-infected mice (13, 20). We wanted to determine if ISG15 protein was upregulated in vivo during Sindbis virus infection and whether this induction was dependent upon IFN-/. We therefore examined the expression of ISG15 in wild type (wt) and IFN-/R–/– mice following intracerebral Sindbis virus infection. In uninfected wt and IFN-/R–/– mice, there was no detectable ISG15 in brain lysates (Fig. 1). Following Sindbis virus infection of wt mice, free ISG15 and a small amount of ISG15 conjugates were detected 1 day postinfection. In contrast, neither free ISG15 nor ISG15 conjugates were detected in IFN-/R–/– mice 1 day after Sindbis infection. By 3 days postinfection, ISG15 and its conjugates could readily be detected in the brains of infected wt mice, and this response was further increased at 6 days postinfection (Fig. 1). Expression could not be analyzed after 1 day postinfection in IFN-/R–/– mice, as they succumbed to infection within 26 h. In adult IFN-/R–/– mice infected with Sindbis virus by the subcutaneous route, barely detectable amounts of ISG15 and ISG15 conjugates were found within the brain at 5 days postinfection (data not shown). These results demonstrate that ISG15 is induced following Sindbis virus infection, and this induction is markedly attenuated in animals lacking the IFN-/ receptor.
Expression of ISG15 by recombinant chimeric Sindbis virus in IFN-/R–/– mice attenuates infection. To evaluate the ability of ISG15 to function as an antiviral molecule in vivo, we utilized a recombinant chimeric Sindbis virus to express ISG15 in IFN-/R–/– mice and then evaluated these animals for attenuation of infection. The recombinant chimeric Sindbis virus system has been well characterized and was previously used to overexpress a variety of genes both in vitro and, more importantly, in vivo (12, 13, 22-24), providing a system to identify the potential antiviral effects of specific genes.
To examine the ability of ISG15 to function as an antiviral molecule, we first generated two recombinant Sindbis viruses. The first virus expressed a FLAG-tagged full-length ISG15 molecule (ISG15 FULL), and the second virus contained the ISG15 open reading frame with the initial ATG deleted (ISG15 MUTANT) (Fig. 2A). These viruses, as well as the parental virus dsTE12Q (no insert), were used to infect IFN-/R–/– mice, and attenuation of infection was determined by following the animals for lethality. While animals infected with either dsTE12Q or ISG15 MUTANT rapidly succumbed to the infection, IFN-/R–/– mice infected with ISG15 FULL were protected from lethality at both 0.1 PFU i.c. (P = 0.0020) (Fig. 2B) and 1.0 PFU i.c. (P = 0.0145) (Fig. 2C). One- to 2-day-old CD1 mice infected with ISG15 FULL were also partially protected from lethality, with 20% of the mice infected with ISG15 FULL surviving at 21 days postinfection compared to 2% survival in the mice infected with ISG15 MUTANT (data not shown). These results provide the first direct evidence that ISG15 can function as an antiviral molecule in vivo when expressed in infected cells.
The C-terminal LRLRGG domain of ISG15 is required for ISG15 conjugate formation. Structure-function studies of ubiquitin have previously demonstrated that the LRLRGG motif, and the terminal diglycine in particular, are required for ubiquitin to form intracellular conjugates (44). As ISG15 shares the LRLRGG motif with ubiquitin and other UBLs, such as RUB-1 and SUMO (10), we wanted to ascertain if the LRLRGG domain of ISG15 was also required for intracellular-conjugate formation.
A transient-transfection assay (17) was performed in which an ISG15 E1 (UBE1L) and E2 (UbcM8) and either wild-type ISG15 with an intact LRLRGG domain or a mutant form of ISG15 in which LRLRGG was mutated to LRLRAA were transfected into 293 T cells. Protein ISGylation was assessed by Western blot analysis. Cells transfected with just E1 and E2 did not display any background ISG15 expression or conjugation (Fig. 3A, lanes 1 and 2). ISG15 LRLRGG transfection alone produced free ISG15 expression (Fig. 3A, lane 3), while the addition of E1 and E2 resulted in protein ISGylation (Fig. 3A, lane 4). In contrast, if ISG15 LRLRAA was transfected into the cells, only free ISG15 was detected, even in the presence of E1 and E2 (Fig. 3A, lanes 5 and 6). These data demonstrate that, similar to ubiquitin, the LRLRGG domain of ISG15 is necessary for intracellular conjugate formation.
The C-terminal LRLRGG motif is important for ISG15's antiviral activity. Since the C-terminal LRLRGG motif is critical for ISG15 to form intracellular conjugates, we wanted to determine whether the LRLRGG motif was also important for ISG15's antiviral activity. We generated additional recombinant chimeric Sindbis viruses that expressed ISG15 either with an intact LRLRGG motif (ISG15 LRLRGG) or with a diglycine-to-dialanine mutation (ISG15 LRLRAA) (Fig. 2A). To avoid potential complications, we also removed the FLAG epitope tag from these viruses and detected the expression of ISG15 protein with a monoclonal antibody generated against murine ISG15. Following the generation of these recombinant viruses, we infected BHK cells with either the parental dsTE12Q virus or ISG15 containing Sindbis viruses. There was no evidence of ISG15 expression in either mock- or dsTE12Q-infected BHK cells (Fig. 3B, lanes 1 and 2). In contrast, BHK cells infected with either ISG15 LRLRGG or ISG15 LRLRAA expressed similar levels of ISG15 (Fig. 3B, lanes 3 and 4). All of the viruses (dsTE12Q, ISG15 LRLRGG, and ISG15 LRLRAA) grew with similar kinetics and to similar final titers under single-step growth conditions in both BHK cells and IFN-/R–/– murine embryonic fibroblasts (MEFs) (Fig. 3C).
We next wanted to determine what effect the LRLRAA mutation would have on ISG15's ability to protect IFN-/R–/– mice from Sindbis virus infection. Since previous studies evaluating Sindbis virus pathogenesis in IFN-/R–/– mice had utilized a subcutaneous route of infection, we infected mice by the s.c. route for the remainder of the experiments in this study (9, 40). Similar to the ISG15 FULL virus, ISG15 LRLRGG protected IFN-/R–/– mice from lethality, with 100% of the mice surviving at 1 x 106 PFU s.c. (P = 0.0291) (Fig. 3D) and >80% surviving at 5 x 106 PFU s.c. (P = 0.0106) (Fig. 3E) compared to 60% and 20% survival with dsTE12Q. Mutating the LRLRGG domain to LRLRAA abrogated this protection, as mice infected with ISG15 LRLRAA had only 60% survival at 1 x 106 PFU s.c. (P = 0.9612) (Fig. 3D) and only 30% survival at 5 x 106 PFU s.c. (P = 0.7332) (Fig. 3E). A similar loss of protection was seen when a virus expressing ISG15 lacking the entire C-terminal LRLRGG motif was used (data not shown). We also saw no attenuation of infection when a Sindbis virus expressing a different ISG (ISG12) was used to infect IFN-/R–/– mice (data not shown).
It has been shown that, following subcutaneous infection with Sindbis virus strain TR339, mice with an intact interferon system limited viral replication and dissemination, with clearance of nearly all organs by 72 to 96 h postinfection (40). In contrast, IFN-/R–/– mice develop a rapid dissemination of TR339, with high titers (>104 PFU) found in multiple organs, and eventually succumb to the infection (40). We therefore wanted to determine if the IFN-/R–/– mice that survived infection with ISG15 LRLRGG had limited viral replication and dissemination or still developed a high-titer systemic Sindbis infection. IFN-/R–/– mice were infected with either dsTE12Q, ISG15 LRLRGG, or ISG15 LRLRAA and were sacrificed at either 1, 3, or 5 days postinfection. Viral titers were then determined for various organs. As expected, IFN-/R–/– mice infected with either dsTE12Q or ISG15 LRLRAA developed a systemic infection with titers of >104 PFU detected in the serum, spleen, liver, and lung as early as 1 day postinfection and in the brain and spinal cord by 3 days postinfection (Fig. 4). When viral titers were analyzed in mice infected with ISG15 LRLRGG, we found that they also developed a systemic infection, with equivalent titers being found in all organs at 1 day postinfection. This indicates that ISG15 expression did not alter the initial Sindbis replication and dissemination.
However, by 3 days postinfection, mice infected with ISG15 LRLRGG had at least 10-fold-lower viral titers (compared to mice infected with ISG15 LRLRAA) in the serum (P < 0.05), spleen (P < 0.005), lung (P < 0.05), and liver (P < 0.0005). By 5 days postinfection, these effects were even more dramatic. Viral titers could not be detected in the serum or liver of mice infected with ISG15 LRLRGG. In the brain (P < 0.005), spinal cord (P < 0.0005), spleen (P < 0.0005), and lung (P < 0.005), viral titers were decreased by 10- to 100-fold compared to the titers found in mice infected with dsTE12Q or ISG15 LRLRAA (Fig. 4). Interestingly, in both the brain and spinal cord, there was also a statistically significant (P < 0.005) decrease in viral titers in mice infected with ISG15 LRLRAA compared to those infected with dsTE12Q. Thus, while the overexpression of ISG15 did not limit the initial high-titer, rapid systemic dissemination of Sindbis virus infection observed in IFN-/R–/– mice, it did result in a significant decrease in viral titers in multiple organs at both 3 and 5 days postinfection. Both lethality data and viral titers demonstrate that ISG15's LRLRGG motif plays an important role in its antiviral activity in this in vivo system.
ISG15 expression reduces Sindbis virus-induced pathology in the brain. To further study the effects of ISG15 expression on Sindbis virus infection in vivo, we examined brain sections 5 days after infection for histopathology and for viral-antigen expression as detected by immunohistochemistry (IHC). IFN-/R–/– mice infected with dsTE12Q had clear evidence of meningoencephalitis, with inflammation of the meninges, vascular cuffing, and focal parenchymal involvement characterized by neuronal loss, pyknotic nuclei, and gliosis. The localization of these focal lesions correlated with the presence of viral antigen detected by IHC. Mice infected with ISG15 LRLRAA also demonstrated evidence of meningoencephalitis, although the severity was decreased compared to animals infected with dsTE12Q, consistent with the decreased viral titers detected in the brains of mice infected with ISG15 LRLRAA on day 5 (Fig. 4). When brain sections were analyzed by IHC, foci of viral antigen were detected in the cerebral cortex and cerebellum, similar to dsTE12Q (Fig. 5A, B, and D). In contrast, animals infected with ISG15 LRLRGG not only had 100-fold-lower viral titers in their brains (Fig. 4), but also demonstrated decreased histopathological evidence of meningoencephalitis and decreased viral antigen immunoreactivity in the cerebral cortex and, most strikingly, in the cerebellum (Fig. 5C and D). No viral staining was detected in any of the sections of the cerebellum in 11 out of 13 animals infected with ISG15 LRLRGG. These data, in combination with the lethality and viral-titer data, provide strong evidence that ISG15 functions as an antiviral molecule in this system and that the LRLRGG motif of ISG15 plays an important role in this activity.
DISCUSSION
IFN-/ play a critical role during viral infection through their direct inhibition of viral replication and by mediating a variety of immunomodulatory effects. While PKR, RNAse L, Mx, and inducible nitric oxide synthase have been shown to inhibit the replication of certain viruses, there are several lines of evidence that indicate there are additional IFN-induced antiviral genes yet to be identified. First, the action of the known IFN-induced antiviral genes is virus specific. For example, PKR–/– mice are more susceptible to vesicular stomatitis virus and encephalomyocarditis virus but clear influenza and vaccinia virus infections (1, 46). Pretreatment of triply deficient mice (Mx–/–, PKR–/–, and RNAse L–/–) with IFN protects them from lethal encephalomyocarditis virus infection (51). Furthermore, while IFN-/R–/– mice rapidly succumb to fatal Sindbis virus infection, triply deficient mice develop only subclinical infection (41). In this report, we explore the ability of ISG15 to function as an antiviral molecule. Following Sindbis virus infection, we demonstrate that mice rapidly upregulate the expression of ISG15 and ISG15 conjugates and that this upregulation is markedly decreased in animals lacking the IFN-/ receptor. Reconstituting the expression of ISG15 in virus-infected cells with a recombinant Sindbis virus system attenuated viral infection in IFN-/R–/– mice, as manifested by decreased lethality, decreased viral titers, and decreased viral antigen staining within the brains of infected mice. The antiviral activity was largely dependent upon the LRLRGG motif, since mutating this motif abrogated the protection. These data provide in vivo evidence that ISG15 functions as an antiviral molecule.
Possible mechanism(s) of ISG15's antiviral activity. Previous studies have suggested two mechanisms by which ISG15 may exert its antiviral activity—protein conjugation and cytokine activity (5, 18, 26, 31, 35, 38, 39, 48). Similar to ubiquitin, the LRLRGG motif is required for ISG15 conjugate formation (Fig. 3A). However, only a limited number of ISG15 targets have been defined, and the fates of these ISGylated proteins are still unknown (8a, 11, 27, 49a). Furthermore, the relevance of these known targets to in vivo antiviral activity is currently unclear. Studies are ongoing by several groups to identify additional ISG15 conjugation targets and to understand the fates of these ISGylated proteins. In vivo analysis of Sindbis viruses engineered to express either ISG15 LRLRGG or LRLRAA demonstrated that the LRLRGG motif is important for ISG15's antiviral activity (Fig. 3D and E). These results support the possibility that ISG15's antiviral activity is due to its conjugation to as-yet-unidentified targets. These targets could be host proteins that either directly effect viral replication or in some way modulate the immune response. Alternatively, the targets could be viral proteins that are conjugated to ISG15 and alter various aspects of the viral life cycle or the immune response to the virus. Modification of viral proteins by either ubiquitin or UBLs, such as SUMO, has been seen with both RNA and DNA viruses. However, the functional consequences of these modifications in many cases are unknown. The Gag protein of human immunodeficiency virus is ubiquitinated, and this modification increases its ability to recruit the endosomal complexes required for transport (ESCRT-I) to the plasma membrane for viral budding (37). The IE1-72kDa and IE-2-86kDa proteins of HCMV are sumoylated (32). The generation of a mutant virus in which the sumoylation site of IE1-72kDa was disrupted resulted in a virus with decreased in vitro growth and reduced accumulation of the IE2 transcripts and protein (32). It is therefore possible that ISG15, similar to other UBLs, can modify viral proteins and alter the viral life cycle. One hypothesis is that ISG15 functions by antagonizing ubiquitin's activity by competing for ubiquitin conjugation sites on specific host or viral proteins, for example, those required for human immunodeficiency virus budding or for HCMV IE1 function.
All members of the ISG15 conjugation pathway are regulated by IFN-/. This is evident in IFN-/R–/– cells and mice, in which there is markedly decreased expression of ISG15 and ISG15 conjugates (Fig. 1). Therefore, despite driving expression of ISG15 in infected cells, the efficient induction of the conjugation cascade may not occur in this system. While IFN-/ are potent inducers of the ISG15 conjugation cascade, previous work has also shown that IFN- can induce low-level expression of ISG15 and ISG15 conjugates (10, 19). Analysis of animals 5 days postinfection with any of the recombinant Sindbis viruses demonstrates the presence of IFN- in their sera (data not shown). IFN- induced by the viral infection may allow low-level induction of the ISG15 conjugation cascade. ISG15 and ISG15 conjugate expression in IFN-/R–/– MEFs can also be induced by pI:pC and LPS (D. J. Lenschow and N. V. Giannakopoulos, unpublished results), albeit at much lower levels than in wild-type cells. Therefore, the overexpression of ISG15 in infected cells, which will also be receiving signals via Toll-like receptor ligation or additional cytokine stimuli, such as IFN-, may reconstitute sufficient ISG15 conjugation to mediate an antiviral effect that is dependent upon conjugation of ISG15 to target proteins.
Alternatively, ISG15's antiviral activity may be completely independent of its ability to modify target proteins. Human ISG15 has been reported to function as a cytokine. Recombinant human ISG15 induces NK cell proliferation, augments lymphokine-activated-killer activity, and stimulates the production of IFN- (5, 38). Additional studies have suggested that ISG15 may induce dendritic cell maturation and neutrophil recruitment (34, 35). Therefore, in addition to its capacity to complex with intracellular proteins, ISG15 may also function as a cytokine whose activity could impact upon either the innate or adaptive immune response. If unconjugated ISG15 either within the cell or functioning as a cytokine is responsible for the antiviral activity, it appears that the LRLRGG motif is also important for this activity. Studies of recombinant human ISG15 demonstrated that the LRLRGG motif had to be exposed for full cytokine activity, since full-length recombinant ISG15 had no activity (5). This motif may be required for processing and/or release of ISG15 from the cell or may be important for receptor binding.
Utilization of recombinant Sindbis virus system as a screen for IFN-induced antiviral molecules. IFN-/ induce hundreds of genes, a subset of which are likely to be antiviral. The recombinant chimeric Sindbis virus has been used to express antiapoptotic genes and antibodies in vivo to evaluate the roles of these proteins during Sindbis virus pathogenesis (12, 23, 24). In this study, we utilized this system to express ISG15 in IFN-/R–/– mice to determine if the expression of this ISG could restore any of the antiviral activity mediated by IFN-/. Expression of individual ISGs in the recombinant Sindbis virus system could be used to screen a variety of ISGs to determine if they exhibited antiviral activity. By utilizing an in vivo approach, we can identify not only those genes that directly inhibit viral replication, but also those genes that can impact upon the immune system. Testing these viruses in the IFN-/R–/– mice allows evaluation of other ISGs in isolation. Finally, since this system is easy to manipulate, structure-function analysis can be readily performed.
There are also limitations to this in vivo genetic screen for IFN-induced antiviral molecules. First, many antiviral genes are virus specific. Since the readout for this screen is the attenuation of Sindbis infection, it may miss antiviral genes that have activity against other viruses. This can be addressed by using other viral systems to express host genes in IFN-/R–/– mice. Second, the initial readout utilized in this screen is lethality, which may be too stringent to identify subtle effects. Using less stringent screens, such as the analysis of viral titers and immunohistochemistry, may be beneficial in this regard. In addition, this screen depends on the overexpression of the IFN-induced molecule, perhaps resulting in nonphysiological effects. In this study, the comparison between LRLRGG and LRLRAA addresses this concern by allowing a comparison between two overexpressed proteins in vivo. Furthermore, since this screen is performed in mice lacking IFN-/ receptors, expressed genes are evaluated in isolation from other maximally induced ISGs. Given the potential redundancy between ISGs, an effect seen in an IFN-/R–/– mouse may be less significant in an intact animal. The ultimate validation of this as an in vivo genetic approach to the identification of novel IFN-induced antiviral genes will be the analysis of mice deficient in selected ISGs and determining if these mice display altered susceptibility to viral infections.
ACKNOWLEDGMENTS
H.W.V. was supported by NIH grant U54 AI057160 and the Pfizer Biomedical Agreement. D.J.L. was supported by a Pfizer Postdoctoral Fellowship in Immunology and NIH Career Development Award K08 AI59390-01. B.L. was supported by NIH grant R01 A144157. The Washington University Hybridoma Core Facility was supported in part by NIH grant P30 AR048335.
We thank Darren Kreamalmayer for his outstanding expertise in animal care.
REFERENCES
Abraham, N., D. F. Stojdl, P. I. Duncan, N. Methot, T. Ishii, M. Dube, B. C. Vanderhyden, H. L. Atkins, D. A. Gray, M. W. McBurney, A. E. Koromilas, E. G. Brown, N. Sonenberg, and J. C. Bell. 1999. Characterization of transgenic mice with targeted disruption of the catalytic domain of the double-stranded RNA-dependent protein kinase, PKR. J. Biol. Chem. 274:5953-5962.
Arnheiter, H., S. Skuntz, M. Noteborn, S. Chang, and E. Meier. 1990. Transgenic mice with intracellular immunity to influenza virus. Cell 62:51-61.
Biron, C. A. 2001. Interferons alpha and beta as immune regulators—a new look. Immunity 14:661-664.
Blomstrom, D. C., D. Fahey, R. Kutny, B. D. Korant, and E. Knight, Jr. 1986. Molecular characterization of the interferon-induced 15-kDa protein. Molecular cloning and nucleotide and amino acid sequence. J. Biol. Chem. 261:8811-8816.
D'Cunha, J., E. Knight, Jr., A. L. Haas, R. L. Truitt, and E. C. Borden. 1996. Immunoregulatory properties of ISG15, an interferon-induced cytokine. Proc. Natl. Acad. Sci. USA 93:211-215.
D'Cunha, J., S. Ramanujam, R. J. Wagner, P. L. Witt, E. Knight, Jr., and E. C. Borden. 1996. In vitro and in vivo secretion of human ISG15, an IFN-induced immunomodulatory cytokine. J. Immunol. 157:4100-4108.
Der, S. D., A. Zhou, B. R. Williams, and R. H. Silverman. 1998. Identification of genes differentially regulated by interferon alpha, beta, or gamma using oligonucleotide arrays. Proc. Natl. Acad. Sci. USA 95:15623-15628.
Dunn, G. P., A. T. Bruce, K. C. F. Sheehan, V. Shankaran, R. Uppaluri, J. D. Bui, M. S. Diamond, C. M. Koebel, C. Arthur, J. M. White, and R. D. Schreiber. 2005. A critical function for type I interferons in cancer immunoediting. Nat. Immunol. 6:722-729.
Giannakopoulos, N. V., J. K. Luo, V. Papoz, W. Zou, D. J. Lenschow, B. S. Jacobs, E. C. Borden, J. Li, H. W. Virgin, and D. E. Zhang. 2005. Proteomic identification of proteins conjugated to ISG15 in mouse and human cells. Biochem. Biophys. Res. Commun. 336:496-506.
Gil, M. P., E. Bohn, A. K. O'Guin, C. V. Ramana, B. Levine, G. R. Stark, H. W. Virgin, and R. D. Schreiber. 2001. Biologic consequences of Stat1-independent IFN signaling. Proc. Natl. Acad. Sci. USA 98:6680-6685.
Haas, A. L., P. Ahrens, P. M. Bright, and H. Ankel. 1987. Interferon induces a 15-kilodalton protein exhibiting marked homology to ubiquitin. J. Biol. Chem. 262:11315-11323.
Hamerman, J. A., F. Hayashi, L. A. Schroeder, S. P. Gygi, A. L. Haas, L. Hampson, P. Coughlin, R. Aebersold, and A. Aderem. 2002. Serpin 2a is induced in activated macrophages and conjugates to a ubiquitin homolog. J. Immunol. 168:2415-2423.
Hardwick, J. M., and B. Levine. 2000. Sindbis virus vector system for functional analysis of apoptosis regulators. Methods Enzymol. 322:492-508.
Johnston, C., W. Jiang, T. Chu, and B. Levine. 2001. Identification of genes involved in the host response to neurovirulent alphavirus infection. J. Virol. 75:10431-10445.
Kapadia, S. B., B. Levine, S. H. Speck, and H. W. Virgin. 2002. Critical role of complement and viral evasion of complement in acute, persistent, and latent gamma-herpesvirus infection. Immunity 17:143-155.
Karupiah, G., J. H. Chen, C. F. Nathan, S. Mahalingam, and J. D. MacMicking. 1998. Identification of nitric oxide synthase 2 as an innate resistance locus against ectromelia virus infection. J. Virol. 72:7703-7706.
Karupiah, G., Q. Xie, R. M. L. Buller, C. Nathan, C. Duarte, and J. D. MacMicking. 1993. Inhibition of viral replication by interferon-gamma induced nitric oxide synthase. Science 261:1445-1448.
Kim, K. I., N. V. Giannakopoulos, H. W. Virgin, and D. E. Zhang. 2004. Interferon-inducible ubiquitin E2, Ubc8, is a conjugating enzyme for protein ISGylation. Mol. Cell. Biol. 24:9592-9600.
Knight, E., Jr., and B. Cordova. 1991. IFN-induced 15-kDa protein is released from human lymphocytes and monocytes. J. Immunol. 146:2280-2284.
Korant, B. D., D. C. Blomstrom, G. J. Jonak, and E. Knight, Jr. 1984. Interferon-induced proteins. Purification and characterization of a 15,000-dalton protein from human and bovine cells induced by interferon. J. Biol. Chem. 259:14835-14839.
Labrada, L., X. H. Liang, W. Zheng, C. Johnston, and B. Levine. 2002. Age-dependent resistance to lethal alphavirus encephalitis in mice: analysis of gene expression in the central nervous system and identification of a novel interferon-inducible protective gene, mouse ISG12. J. Virol. 76:11688-11703.
Leib, D. A., M. A. Machalek, B. R. Williams, R. H. Silverman, and H. W. Virgin. 2000. Specific phenotypic restoration of an attenuated virus by knockout of a host resistance gene. Proc. Natl. Acad. Sci. USA 97:6097-6101.
Levine, B., J. E. Goldman, H. H. Jiang, D. E. Griffin, and J. M. Hardwick. 1996. Bcl-2 protects mice against fatal alphavirus encephalitis. Proc. Natl. Acad. Sci. USA 93:4810-4815.
Liang, X. H., Jiang, H. H., and B. Levine. 1997. Expression of a biologically active antiviral antibody using a sindbis virus vector system. Molecular Immunology 34:907-917.
Liang, X. H., L. K. Kleeman, H. H. Jiang, G. Gordon, J. E. Goldman, G. Berry, B. Herman, and B. Levine. 1998. Protection against fatal Sindbis virus encephalitis by beclin, a novel Bcl-2-interacting protein. J. Virol. 72:8586-8596.
Liu, L. Q., R. Ilaria, Jr., P. D. Kingsley, A. Iwama, R. A. van Etten, J. Palis, and D. E. Zhang. 1999. A novel ubiquitin-specific protease, UBP43, cloned from leukemia fusion protein AML1-ETO-expressing mice, functions in hematopoietic cell differentiation. Mol. Cell. Biol. 19:3029-3038.
Loeb, K. R., and A. L. Haas. 1992. The interferon-inducible 15-kDa ubiquitin homolog conjugates to intracellular proteins. J. Biol. Chem. 267:7806-7813.
Malakhov, M. P., K. I. Kim, O. A. Malakhova, B. S. Jacobs, E. C. Borden, and D. E. Zhang. 2003. High-throughput immunoblotting: ubiquitin-like protein ISG15 modifies key regulators of signal transduction. J. Biol. Chem. 278:16608-16613.
Malakhov, M. P., O. A. Malakhova, K. I. Kim, K. J. Ritchie, and D. E. Zhang. 2002. UBP43 (USP18) specifically removes ISG15 from conjugated proteins. J. Biol. Chem. 277:9976-9981.
Malakhova, O. A., M. Yan, M. P. Malakhov, Y. Yuan, K. J. Ritchie, K. I. Kim, L. F. Peterson, K. Shuai, and D. E. Zhang. 2003. Protein ISGylation modulates the JAK-STAT signaling pathway. Genes Dev. 17:455-460.
Muller, U., S. Steinhoff, L. F. L. Reis, S. Hemmi, J. Pavlovic, R. M. Zinkernagel, and M. Aguet. 1994. Functional role of type I and type II interferons in antiviral defense. Science 264:1918-1921.
Narasimhan, J., J. L. Potter, and A. L. Haas. 1996. Conjugation of the 15-kDa interferon-induced ubiquitin homolog is distinct from that of ubiquitin. J. Biol. Chem. 271:324-330.
Nevels, M., W. Brune, and T. Shenk. 2004. SUMOylation of the human cytomegalovirus 72-kilodalton IE1 protein facilitates expression of the 86-kilodalton IE2 protein and promotes viral replication. J. Virol. 78:7803-7812.
Nicholl, M. J., L. H. Robinson, and C. M. Preston. 2000. Activation of cellular interferon-responsive genes after infection of human cells with herpes simplex virus type 1. J. Gen. Virol. 81:2215-2218.
Owhashi, M., Y. Taoka, K. Ishii, S. Nakazawa, H. Uemura, and H. Hambara. 2003. Identification of a ubiquitin family protein as a novel neutrophil chemotactic factor. Biochem. Biophys. Res. Commun. 309:533-539.
Padovan, E., L. Terracciano, U. Certa, B. Jacobs, A. Reschner, M. Bolli, G. C. Spagnoli, E. C. Borden, and M. Heberer. 2002. Interferon stimulated gene 15 constitutively produced by melanoma cells induces e-cadherin expression on human dendritic cells. Cancer Res. 62:3453-3458.
Player, M. R., and P. F. Torrence. 1998. The 2-5A system: modulation of viral and cellular processes through acceleration of RNA degradation. Pharmacol. Ther. 78:55-113.
Pornillos, O., J. E. Garrus, and W. I. Sundquist. 2002. Mechanisms of enveloped RNA virus budding. Trends Cell Biol. 12:569-579.
Recht, M., E. C. Borden, and E. Knight, Jr. 1991. A human 15-kDa IFN-induced protein induces the secretion of IFN-gamma. J. Immunol. 147:2617-2623.
Ritchie, K. J., C. S. Hahn, K. I. Kim, M. Yan, D. Rosario, L. Li, J. C. de la Torre, and D. E. Zhang. 2004. Role of ISG15 protease UBP43 (USP18) in innate immunity to viral infection. Nat. Immunol. 10:1374-1378.
Ryman, K. D., W. B. Klimstra, K. B. Nguyen, C. Biron, and R. E. Johnston. 2000. Alpha/beta interferon protects adult mice from fatal Sindbis virus infection and is an important determinant of cell and tissue tropism. J. Virol. 74:3366-3378.
Ryman, K. D., L. J. White, R. E. Johnston, and W. B. Klimstra. 2002. Effects of PKR/RNase L-dependent and alternative antiviral pathways on alphavirus replication and pathogenesis. Viral Immunol. 15:53-76.
Sen, G. C. 2001. Viruses and interferons. Annu. Rev. Microbiol. 55:255-281.
Sheehan, K. C., N. H. Ruddle, and R. D. Schreiber. 1989. Generation and characterization of hamster monoclonal antibodies that neutralize murine tumor necrosis factors. J. Immunol. 142:3884-3893.
Sloper-Mould, K. E., J. C. Jemc, C. M. Pickart, and L. Hicke. 2001. Distinct functional surface regions on ubiquitin. J. Biol. Chem. 276:30483-30489.
Stark, G. R., I. M. Kerr, B. R. Williams, R. H. Silverman, and R. D. Schreiber. 1998. How cells respond to interferons. Annu. Rev. Biochem. 67:227-264.
Stojdl, D. F., N. Abraham, S. Knowles, R. Marius, A. Brasey, B. D. Lichty, E. G. Brown, N. Sonenberg, and J. C. Bell. 2000. The murine double-stranded RNA-dependent protein kinase PKR is required for resistance to vesicular stomatitis virus. J. Virol. 74:9580-9585.
Yang, Y. L., L. F. Reis, J. Pavlovic, A. Aguzzi, R. Schafer, A. Kumar, B. R. Williams, M. Aguet, and C. Weissmann. 1995. Deficient signaling in mice devoid of double-stranded RNA-dependent protein kinase. EMBO J. 14:6095-6106.
Yuan, W., and R. M. Krug. 2001. Influenza B virus NS1 protein inhibits conjugation of the interferon (IFN)-induced ubiquitin-like ISG15 protein. EMBO J. 20:362-371.
Zhao, C., S. L. Beaudenon, M. L. Kelley, M. B. Waddell, W. Yuan, B. A. Schulman, J. M. Huibregtse, and R. M. Krug. 2004. The UbcH8 ubiquitin E2 enzyme is also the E2 enzyme for ISG15, an IFN-/-induced ubiquitin-like protein. Proc. Natl. Acad. Sci. USA 101:7578-7582.
Zhao, C., C. Denison, J. M. Huibregtse, S. Gygi, and R. M. Krug. 2005. Human ISG15 conjugation targets both IFN-induced and constitutively expressed proteins functioning in diverse cellular pathways. Proc. Natl. Acad. Sci. USA 102:10200-10205.
Zhou, A., J. Paranjape, T. L. Brown, H. Nie, S. Naik, B. Dong, A. Chang, B. Trapp, R. Fairchild, C. Colmenares, and R. H. Silverman. 1997. Interferon action and apoptosis are defective in mice devoid of 2',5'-oligoadenylate-dependent RNase L. EMBO J. 16:6355-6363.
Zhou, A., J. M. Paranjape, S. D. Der, B. R. Williams, and R. H. Silverman. 1999. Interferon action in triply deficient mice reveals the existence of alternative antiviral pathways. Virology 258:435-440.
Zhu, H., J. P. Cong, G. Mamtora, T. Gingeras, and T. Shenk. 1998. Cellular gene expression altered by human cytomegalovirus: global monitoring with oligonucleotide arrays. Proc. Natl. Acad. Sci. USA 95:14470-14475.(Deborah J. Lenschow, Nadi)