Identification of Critical Amino Acids within the Nucleoprotein of Tacaribe Virus Important for Anti-interferon Activity*

Background: With the exception of Tacaribe virus, all arenavirus nucleoproteins are thought to inhibit type I interferon production. Results: Variation in nucleoprotein residues 389–392 of Tacaribe virus was characterized as a critical region regulating interferon inhibition. Conclusion: Some Tacaribe virus variants contain the important nucleoprotein residues necessary for interferon antagonism. Significance: Anti-interferon activity of nucleoproteins appears to be a conserved feature of all arenaviruses. The arenavirus nucleoprotein (NP) can suppress induction of type I interferon (IFN). This anti-IFN activity is thought to be shared by all arenaviruses with the exception of Tacaribe virus (TCRV). To identify the TCRV NP amino acid residues that prevent its IFN-countering ability, we created a series of NP chimeras between residues of TCRV NP and Pichinde virus (PICV) NP, an arenavirus NP with potent anti-IFN function. Chimera NP analysis revealed that a minimal four amino acid stretch derived from PICV NP could impart efficient anti-IFN activity to TCRV NP. Strikingly, the TCRV NP gene cloned and sequenced from viral stocks obtained through National Institutes of Health Biodefense and Emerging Infections (BEI) resources deviated from the reference sequence at this particular four-amino acid region, GPPT (GenBank KC329849) versus DLQL (GenBank NC004293), respectively at residues 389–392. When efficiently expressed in cells through codon-optimization, TCRV NP containing the GPPT residues rescued the antagonistic IFN function. Consistent with cell expression results, TCRV infection did not stimulate an IFNβ response early in infection in multiple cells types (e.g. A549, P388D1), and IRF-3 was not translocated to the nucleus in TCRV-infected A549 cells. Collectively, these data suggest that certain TCRV strain variants contain the important NP amino acids necessary for anti-IFN activity.


The arenavirus nucleoprotein (NP) can suppress induction of type I interferon (IFN). This anti-IFN activity is thought to be shared by all arenaviruses with the exception of Tacaribe virus (TCRV). To identify the TCRV NP amino acid residues that prevent its IFN-countering ability, we created a series of NP chimeras between residues of TCRV NP and Pichinde virus (PICV) NP, an arenavirus NP with potent anti-IFN function. Chimera
NP analysis revealed that a minimal four amino acid stretch derived from PICV NP could impart efficient anti-IFN activity to TCRV NP. Strikingly, the TCRV NP gene cloned and sequenced from viral stocks obtained through National Institutes of Health Biodefense and Emerging Infections (BEI) resources deviated from the reference sequence at this particular four-amino acid region, GPPT (GenBank KC329849) versus DLQL (GenBank NC004293), respectively at residues 389 -392. When efficiently expressed in cells through codon-optimization, TCRV NP containing the GPPT residues rescued the antagonistic IFN function. Consistent with cell expression results, TCRV infection did not stimulate an IFN␤ response early in infection in multiple cells types (e.g. A549, P388D1), and IRF-3 was not translocated to the nucleus in TCRV-infected A549 cells. Collectively, these data suggest that certain TCRV strain variants contain the important NP amino acids necessary for anti-IFN activity.
The Arenaviridae family contains more than 29 virus species, which are currently subdivided into two major groups: Old World (OW) 3 and New World (NW) arenaviruses (1,2). The OW arenaviruses constitute a single lineage, whereas the NW arenaviruses are divided into three phylogenetically distinct clades designated A, B, and C. OW and NW clade B arenaviruses include several significant human pathogens capable of causing hemorrhagic fever. Nonpathogenic arenaviruses such as NW clade B Tacaribe virus (TCRV) also merit significant attention because close phylogenetic relationships with family members causing hemorrhagic fever provide an excellent opportunity to identify determinants of pathogenesis. In addition, Pichinde virus (PICV) is a nonpathogenic arenavirus that upon long term passaging acquires virulence in guinea pigs and causes a disease similar to pathogenic human Lassa fever virus (LASV) infection in that model system (3).
Arenaviruses cause chronic infections of rodents, and each arenavirus species is associated with a particular rodent host species, except for the TCRV, which has been isolated only from fruit-eating bats in Trinidad. Asymptomatically infected rodents may invade human dwellings and transmit virus. Humans are most frequently infected through contact with infected rodent excreta, commonly via inhalation of dust or aerosolized virus-containing materials or by ingestion of contaminated foods. Infections with OW arenavirus LASV can occur at frequencies of 300,000 -500,000 per year in western Africa, and although most infections are typically subclinical or associated with mild febrile illness, a small percentage of cases progress to Lassa fever which is associated with serious morbidity and mortality (4). In addition, some South American outbreaks with NW arenaviruses such as Junin virus have been reported to cause severe hemorrhagic fever in humans with case fatality rates of 15-35% (5,6).
Arenaviruses are enveloped viruses with a bisegmented RNA genome and a unique ambisense coding strategy. Despite this coding strategy, the Arenaviridae are classified as segmented single-strand, negative sense RNA viruses. The large (L) genome segment encodes two proteins, the viral RNA-dependent RNA polymerase (L protein) and the Z protein, a matrixlike protein important for virus assembly and budding (7). On the small (S) genome segment, the glycoprotein precursor and the nucleoprotein (NP) are encoded. The NP is the most abundant viral protein in both infected cells and the virus particles (1). The NP binds to both the genomic and antigenomic RNA and together with the L polymerase, forms the ribonucleoprotein core necessary for virus transcription and replication. Several groups have now demonstrated that the NP also has the ability to inhibit host innate immune defenses, particularly the type I interferon response (8 -11).
Among the most important cytokines produced by the host to defend against viral infection are the type I interferons (IFN-I) (12). IFN-I is produced very rapidly in virus-infected cells, inhibits viral replication, and regulates the activation of cells involved in innate and adaptive immunity (13). IFN-I produces an antiviral state in cells by inducing interferon-stimulated genes, which influence protein synthesis, cell growth, and cell death. IFN-I activation is mediated by host pathogen recognition receptors that sense pathogen-associated molecular patterns, which are components of foreign micro-organisms such as viral nucleic acids. Some pathogen recognition receptors are expressed on endosomal compartments (e.g. Toll-like receptors) whereas others such as retinoic acid-inducible gene I (RIG-I) and melanoma differentiation-associated protein 5 (MDA5) are expressed in the cytoplasm. Arenaviruses, like other negative strand RNA viruses, possess 5Ј-triphosphate (5Ј-PPP) motifs on their genomes which can serve as RIG-I triggers (9,14). RIG-I activation leads to nuclear translocation of the transcription factor IFN regulatory factor 3 (IRF-3), which in turn induces expression of the IFN␤ gene. The NP is thought to inhibit IFN production by blocking activation and translocation of IRF-3, potentially through direct protein-protein interactions (9,15). LASV NP was recently crystallized, revealing that the C-terminal domain contains 3Ј-5Ј exonuclease activity important for immune suppression (10,11). These results suggest that arenavirus NP may also prevent IFN-I induction by degrading viral pathogen-associated molecular patterns produced during viral replication and infection.
All arenavirus NPs are believed to contain anti-IFN activity with the exception of TCRV NP (8,15). Understanding why the TCRV NP lacks efficient anti-IFN activity may lead to insights into the potential pathogenic function of other arenavirus NPs. Based on the NP crystal structure and sequence alignments, it is not clear why TCRV lacks IFN inhibition activity because it shares all exonuclease catalytic residues (10,11). To address this problem, we generated NP chimeras to map the regions in TCRV that prevent its IFN-countering ability. This method of analysis used a gain-of-function assay thought to preserve the structural integrity of TCRV NP, to narrow down regions important for IFN-countering function. Through NP chimera analysis, we identified key TCRV NP residues between amino acids 383-407 involved in regulating anti-IFN function.

EXPERIMENTAL PROCEDURES
Cells and Culture Conditions-All cell lines were maintained in culture medium supplemented with 10% fetal bovine serum (FBS), 100 units/ml penicillin, and 100 g/ml streptomycin (Invitrogen) at 37°C in 5% CO 2 . HEK293T (human embryonic kidney) and A549 (human lung epithelial) cells were cultured in Dulbecco's modified Eagle's medium (DMEM); Vero (African green monkey kidney) cells were maintained in Minimum Essential Medium ␣. The IFN␤ reporter cells were generated from the P388D1 (mouse macrophage-like, a gift from Dr. Norbert Herzog, University of Texas Medical Branch) cell line by stable integration of an mCherry reporter under the control of the mouse IFN␤ promoter. The reporter P388D1 cells were grown in RPMI 1640 medium supplemented with 400 g/ml hygromycin B.
Plasmids and Chimeric Nucleoprotein Constructs-Arenavirus NP nucleotide sequences of TCRV and PICV were codonoptimized and synthesized by Epoch Life Sciences (Sugarland, TX) based on GenBank data base accession numbers NC_004293.1 and EF529746.1, respectively. TCRV and PICV nucleoproteins were tagged at the C terminus with hemagglutinin (HA), and chimeric TCRV NP and PICV NP constructs were made using an overlap PCR strategy. Each nucleoprotein HA-tagged construct was subcloned into the pcDNA3.1 vector (Invitrogen) under the control of a cytomegalovirus (CMV) promoter. All NP constructs were confirmed by sequencing. IFN␤ minimal promoters (human and mouse) were kind gifts from Dr. Adolfo Garcia-Sastre (Mount Sinai School of Medicine, New York, NY) (16). Each of these was excised from parental vectors and inserted into pcDNA3.1 upstream of the mCherry gene. The CMV promoter was subsequently removed so that only the cytokine promoter remained.
Viruses-TCRV strain 11573 was obtained through the National Institutes of Health BEI Research Resources Repository, NIAID, National Institutes of Health (NR-10175) and through ATCC (VR-1272). The BEI TCRV stock was derived from tissue culture (Vero) infection, whereas the ATCC TCRV stock was prepared from a newborn mouse brain inoculation. PICV strain P18 was provided by Dr. Norbert Herzog (University of Texas Medical Branch). Sendai virus (SeV) (Cantell strain) stocks were purchased from Charles River Laboratories (North Franklin, CT). The recombinant Rift Valley fever virus vaccine strain MP12 generated to carry a GFP gene (RVFV-GFP) in place of NSs gene was described previously (17).
Virus Propagation and Titer Determination-Fresh TCRV and PICV stocks were prepared by diluting original stocks with a 1:500 dilution in 5 ml of medium and infecting 70% confluent Vero cells in T75 flask for 1 h at 37°C in 5% CO 2 . Following virus absorption, the inocula were replaced with fresh medium containing 2% FBS and penicillin/streptomycin. Supernatant was collected at day 8 for TCRV and day 5 for PICV. RVFV-GFP was also propagated in Vero cells as previously described (18). New TCRV, PICV, and RVFV-GFP stocks were clarified, purified over a 20% sucrose cushion through ultracentrifugation, and titered by plaque assay using standard methods (7,18). Viral Growth Kinetics-Growth kinetics for TCRV and PICV were measured using 50% tissue culture infective dose (TCID 50 ) measurements. P388D1 IFN reporter and A549 cells were infected in 12-well plates with TCRV or PICV at a multiplicity of infection (m.o.i.) of 0.2 or 2 in 1 ml of medium for 1 h at 37°C in 5% CO 2 . Cells were then washed two times with serum-free medium, and fresh medium containing 2% FBS and antibiotics were added. Samples were collected at several time points after infection and clarified before serial dilutions were made to infect Vero cells in triplicate. 6 -8 days after infection, the cells were examined visually for cytopathic effects and TCID 50 /ml was calculated using the Reed and Muench method (30).
IFN␤ Reporter Assay-293T cells were seeded into a 12-well tissue culture plate and co-transfected with an IFN␤ promoter plasmid expressing mCherry upon activation, a GFP expression plasmid for transfection efficiency normalization, and the NP constructs using TransIT-LT1 transfection reagent (Mirus Bio). For both positive and negative control experiments, the NP expression plasmid was replaced with an empty plasmid vector. All conditions, except for the negative control, were then infected with SeV 24 h after transfection. A 1:1000 dilution of the SeV supplied by Charles River Laboratories, containing an HA titer of 16,000 HA units/ml, was made for all infections involving SeV. At 16 h after infection, cell lysates were prepared and analyzed using a fluorescent plate reader (Tecan M1000). The IFN␤ activity measured by mCherry fluorescence was normalized to the GFP transfection efficiency signal. The positive control experiment was set to 100% activation, and all results were normalized to this control. Alternatively, 10 g/ml of low molecular mass poly(I:C) (Invivogen) or purified TCRV RNA was transfected into cells using TransIT-LT1 instead of SeV infection to stimulate the IFN␤ reporter. Purified TCRV RNA was obtained using QIAamp UltraSens Virus kit (Qiagen).
Statistical Analysis for Fig. 3E-Data for IFN␤ activation induced by SeV (measured by mCherry fluorescence normalized to transfection efficiency measured by GFP fluorescence) were compared using a two-tailed t test for each individual experiment. Values obtained with cells transfected with indicated NP plasmids and infected with SeV were compared with cells transfected with empty plasmid vector and infected with SeV, designated (ϩ) in Fig. 3E. p values were considered significant when they were Ͻ0.03 (single asterisk) and very significant when they were Ͻ0.01 (double asterisk). The p values shown in the figures and the text were based on the highest p values obtained from three independent experiments.
Western Blot Analysis-Clarified cell lysates expressing NPs were prepared 24 h after transfection from HEK293T cells using 1% Triton X-100 in phosphate-buffered saline (PBS) as the lysis buffer. NuPAGE LDS sample buffer (Invitrogen) was added to the lysate samples, heated at 99°C for 10 min, followed by separation on 4 -12% NuPAGE Bis-Tris precast polyacrylamide gels (Invitrogen). Samples were transferred onto polyvinylidene difluoride (PVDF) membranes, and blots were probed with anti-HA-HRP antibodies (Thermo Scientific) and HRPconjugated rabbit anti-actin polyclonal antibody (Novus) to detect the HA-tagged NPs and actin simultaneously.
NP Sequencing from Viral RNA-Viral RNA from sucrose cushion-purified TCRV passaged once in Vero cells from original BEI resources or ATCC stocks was extracted using QIAamp UltraSens Virus kit (Qiagen). cDNA synthesis using random primers was performed using a MMLV reverse transcriptase first-strand cDNA synthesis kit (Epicenter). The NP gene from the BEI-or ATCC-derived strains was amplified by PCR using forward primer 5-atggctcaatccaaggaagtg-3 and reverse primer 5-tcacagtgcaaaagctgttttg-3. PCR products were cloned into the pCR2.1-topo sequencing vector (Invitrogen). Ten clones from each TCRV source were sent for sequencing to Quintarabio (Albany, CA). The BEI TCRV NP clone with 98.2% sequence homology to the reference TCRV sequence (NC_004293.1) at the protein level was further subcloned into the mammalian expression vector pcDNA3.1 for expression analysis and referred to as construct vTNP.
Quantitative Real-time RT-PCR-Relative IFN␤ expression in A549 cells during virus infection was measured by collecting total RNA from the TCRV, PICV, SeV, RVFV-GFP, or no virus mock infections using TRIzol (Invitrogen). The total RNA was further purified using a Norgen RNA clean up and concentration kit. The real-time RT-PCR primer and probe set for human IFN␤ and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) housekeeping gene were purchased from Applied Biosystems, and gene expression measurements were analyzed on a Bio-Rad CFX384. Each experimental condition was run in triplicate with 25 ng of RNA per well using QuantiFast Probe RT-PCR reagents (Qiagen). Average IFN␤ C t values were normalized to the average C t values of GAPDH, and ⌬⌬C t -based -fold change calculations were set relative to no virus mock infection at the 12-h time point using Bio-Rad CFX manager software. To normalize data from three separate experiments, the SeV time point at 12 h was set to 100% activation, and all the values were normalized as a percentage of this reference value.
IRF-3 Translocation Imaging-A549 cells were plated overnight in complete medium on collagen-coated 24-well glass bottom plates (MatTek) and then infected with SeV, RVFV-GFP (1 m.o.i.), TCRV (2 m.o.i.) (BEI resources-derived), and PICV (2 m.o.i.) for 6, 9, 12, or 24 h. At the end of each infection time point, cells were washed with PBS and fixed with 4% paraformaldehyde for 20 min at room temperature followed by a wash with TBS. Cells were permeabilized with 90% methanol for 10 min, washed with PBS, put in block (BD BSA stain buffer ϩ 3% normal goat serum ϩ 0.5% Triton X-100) for 1 h. Next, cells were incubated for 1 h with anti-SeV chicken polyclonal antibody (Abcam), mouse anti-TCRV antibodies (ascitic fluid from ATCC), and mouse anti-PICV antibodies (ascitic fluid kindly provided by Drs. Stuart Nichol and Eric Bergeron, Centers for Disease Control) for each respective virus. RVFV-GFP virus infection and replication were measured through GFP reported expression. Each condition was also stained with anti-human IRF-3 from Santa Cruz Biotechnology (FL-425) at a 1:50 dilution. After washing two times with PBS, Alexa Fluorconjugated secondary antibodies were added in block for 30 min at room temperature. These samples were washed twice with PBS, incubated with 300 nM DAPI in PBS for 2 min at room temperature, washed two times with PBS, and imaged at a magnification of ϫ40 using an Olympus IX70 inverted microscope and indexed using ImagePro 6.3 software. For each well, 10 series of images were taken with each series containing three separate fluorescent channel images.

RESULTS
Chimeric Nucleoprotein Analysis Identifies Minimal Residues Important for Gaining Anti-interferon Activity in Tacaribe NP-All arenavirus NPs, with the exception of TCRV NP, are thought to inhibit type I interferon production and IRF-3 activation (8,10,11). Because the mechanisms by which arenaviruses subvert host innate immune defenses may be determinants of pathogenicity, we decided to map the specific amino acid residues within TCRV NP involved in its lack of anti-IFN activity. To accomplish this, the GenBank NP gene sequences from PICV (EF529746.1) and TCRV (NC_004293.1) were codon-optimized for expression in mammalian cells. Of all of the arenavirus NPs, PICV NP is one of the most potent inhibitors of type I interferon (8). Chimeric NPs were designed to insert PICV NP residues into homologous positions in the TCRV NP with the intent of acquiring anti-IFN function within the TCRV NP. The exact residues exchanged between PICV and TCRV NPs are illustrated in Fig. 1A. To test the ability of chimeric proteins to inhibit IFN␤ activation, 293T cells were transfected with NP constructs and an IFN␤ promoter reporter plasmid expressing the mCherry fluorophore upon activation. Twenty-four hours after transfection, cells were infected with SeV, a strong inducer of the IFN response, to trigger the IFN␤ promoter. As seen in Fig. 1B, transfection with PICV NP inhibited IFN␤ activation whereas TCRV NP permitted the IFN␤ response. The chimeric TCRV NP containing C-terminal residues of PICV NP (TP377-561) efficiently inhibited IFN␤ activation by SeV as predicted by the C-terminal location of the NP exonuclease domain (10,11). Further down-selection of NP regions by chimeric analysis revealed that TCRV NP expressing amino acids 377-388 from PICV NP (construct labeled TP377-388 in Fig. 1) gained anti-IFN function.
We next determined the level of NP expression by Western blot analysis following transfection. As shown in Fig. 1C, the NP expression levels between the different constructs varied; however, TP377-388 still had anti-IFN␤ activity equivalent to PICV NP, despite a 3.5-fold decrease in protein expression. Therefore, PICV NP is most likely expressed at much higher levels than necessary for anti-IFN activity.
Identification of Minimal TCRV NP Residues Required to Eliminate IFN␤ Antagonism Function in PICV NP-Having identified that residues 377-388 from PICV NP were enough to confer anti-IFN function to TCRV NP, we hypothesized that replacing those same residues in PICV NP with homologous TCRV NP residues would eliminate anti-IFN function. To test this theory, the PICV NP chimera with residues 383-396 from TCRV NP designated PT383-396 was constructed and analyzed for IFN␤ inhibition. Strikingly, using the IFN␤ reporter assay described for Fig. 1, this chimera retained anti-IFN activity (Fig. 2, A and B, lane PT383-396). Thus, to map the minimal region in TCRV NP required to abolish anti-IFN activity in PICV NP, we made a second series of chimeras similar to those made in Fig. 1A. Using PICV NP as the template, homologous FIGURE 1. The TCRV NP chimera with residues 377-388 from PICV NP rescues anti-interferon activity. A, schematic representation of the TCRV NP-based chimeric constructs containing C-terminal residues from PICV NP. The NP chimeras were labeled according to the amino acid residues from PICV NP that replaced homologous regions in the TCRV NP protein. B, inhibition of the IFN␤ reporter by NP chimeras. 293T cells were transfected with an IFN␤ promoter plasmid expressing mCherry upon activation, a GFP expression plasmid, and 1 g of the indicated HA tagged arenavirus NP or chimeric construct plasmid. For control experiments, the NP expression plasmid was replaced with empty vector. All wells except for the negative control were then infected with SeV 24 h after transfection. Cell lysates were analyzed 16 h after infection using a fluorescent plate reader. The positive control experiment was set to 100% activation, and all experiments were normalized to this control. Error bars represent S.E. of three experiments conducted in triplicate. C, protein expression levels of the NPs and chimeras. 293T clarified cell lysates were prepared 24 h after transfection using co-transfection conditions described above. The lysates were analyzed by Western blotting using anti-HA (HA) and anti-actin (Actin) antibodies. The relative reduction index (RI) was calculated as the quotient of the densitometry signal for the NP band and that for actin and then normalized by the ratio obtained with PICV NP (considered 1). The PICV NP* condition was transfected with 0.5 g instead of 1 g. Data shown represent one of three experiments with similar results.
TCRV NP residues were swapped into the C-terminal domain creating the PICV NP chimera constructs detailed in Fig. 2A. Using this approach, we found that substitution of residues 383-407 from TCRV NP into PICV NP eliminated anti-IFN function (Fig. 2B, lane PT383-407). Although the expression levels of the chimeric proteins varied, as determined by Western blot analysis (Fig. 2C), only the amounts equivalent to TCRV NP are necessary for function as seen in Fig. 1C. The amino acid region containing TCRV NP residues 383-407 was aligned with other representative arenavirus NPs (Fig. 2D). Interestingly, most of the residues in this domain are well conserved between OW and NW arenaviruses (Fig. 2D, red residues). However, two regions containing three-or four-amino acid stretches seem to deviate from the rest of the arenavirus NP residues within this region (Fig. 2D, black residues). Both of these regions from TCRV are necessary to abolish anti-IFN function in PICV NP, whereas only the DLQL residues from TCRV NP need to be replaced by PICV NP GPTN to gain antagonistic IFN activity.
Upon close inspection of the DLQL alignment region, the aspartic acid (D) residue 389 from TCRV NP appears to be a conserved glycine (G) across other OW and NW arenaviruses (Fig. 2D). Therefore, to test whether a substitution from an aspartic acid to a glycine can gain anti-IFN function in the TCRV NP, we made a construct containing a D389G point mutation (TNP-D389G). Using the IFN␤ reporter assay described above, we found that the TNP-D389G construct could not inhibit SeV-mediated IFN activation (Fig. 2E). Western blot analysis indicated that the construct was expressed at levels similar to TCRV NP (Fig. 2F). Therefore, the glycine substitution alone at TCRV NP residue 389 was not sufficient to gain IFN inhibitory activity. 293T clarified cell lysates were prepared 24 h after transfection from the conditions described above (Fig. 2B), run under reducing SDS-PAGE conditions, and analyzed by Western blotting using anti-HA (HA) and anti-actin (Actin) antibodies. One representative experiment of three is shown. D, alignment of the protein sequences corresponding to TCRV NP residues 383-407 and PICV NP residues 377-401 sequence with OW arenavirus Lassa, NW clade A (NW-A) arenavirus Parana, NW-B viruses Junin and Machupo, and NW-C viruses Latino and Oliveros. Nonconserved residues among the arenavirus family are shown in black. Asterisk (*) represents conserved residues in all aligned arenaviruses NP with the exception of the aspartic acid (D) residue 389 of TCRV NP. E, measurements of IFN␤ reporter activity during expression of TCRV NP containing a point mutation D389G (TNP-D389G). Induction of the IFN␤ promoter was performed by infection with SeV. F, levels of NP expression analyzed 24 h after transfection by Western blotting using anti-HA (NP) and anti-actin (Actin) antibodies. D and F, relative reduction index (RI) calculated as the quotient of the densitometry signal for the target protein band and that for actin and then normalized by the ratio obtained with PICVNP from transfection (considered 1). Data represent one of three experiments with similar results.

TCRV Strain Variant Contains GPPT Residues in the NP with
Functional Anti-IFN Activity-Interestingly, sequence variation in the DLQL region of TCRV NP was reported previously. Martínez-Sorbido et al. found that when RNA from TCRVinfected cells was used to clone the TCRV NP gene, multiple independent clones contained amino acids GPPT in place of the DLQL at residues 389 -392 indicated in the published sequence for TCRV NP (GenBank NC_004293) (19). However, their TRCV NP expression plasmid containing GPPT did not have anti-IFN activity (8,19). To further investigate sequence variations that were reported to occur at regions near the important residues DLQL of TCRV NP, we obtained a TCRV stock (strain TRVL-11573) from BEI resources. According to the BEI resources reference material, both the S and L RNA segments of TCRV strain TRVL-11573 have been sequenced (GenBank M20304 and J04340, respectively). The GenBank reference TCRV sequences were derived from M20304 and J04340. When we cloned and sequenced the TCRV NP gene from the BEI resources stock, we found that all of our 10 inde-pendent clones contained a TCRV NP protein sequence with a key substitution of residues GPPT in the DLQL region (Fig. 3A). The TCRV NP gene that we sequenced with 98.2% homology to the original sequence has been deposited in GenBank (accession no. KC329849).
The BEI TCRV stock was derived from cell culture passage in Vero cells, whereas another TCRV stock from ATCC was prepared from a newborn mouse brain inoculation. We additionally sequenced the NP gene from the ATCC stock but again found the GPPT substitution in all of our clones.
To test whether these GPPT residues in TCRV NP have functional anti-IFN activity when expressed using a codon-optimized sequence, the original TCRV NP plasmid used in Figs. 1 and 2 was mutated to substitute GPPT in place of the DLQL amino acids (TNP-GPPT). Similar to experiments performed in Figs. 1 and 2, 293T cells were transfected with an NP construct, with the IFN␤-mCherry reporter plasmid and a GFP plasmid to serve as a transfection control. Twenty-four hours after transfection, the IFN␤ promoter was activated either by SeV infection (Fig. 3B), a synthetic dsRNA ligand poly(I:C) (Fig.  3C), or purified TCRV genomic RNA (Fig. 3D). In each IFN activation method, the TNP-GPPT construct was able to efficiently reduce IFN␤ induction to levels comparable with PICV NP.
Next, to understand whether the new virus-derived (unoptimized) TCRV NP sequence (vTNP) could inhibit IFN␤ promoter induction, the vTNP, TNP-GPPT, and the PICV NP constructs were transfected at increasing concentrations into 293Ts and tested using the standard SeV activation protocol used previously (Fig. 3E). The protein expression levels of these constructs were also measured by Western blotting (Fig. 3F). We found that the TNP-GPPT was able to effectively inhibit the IFN␤ response induced by SeV infection to levels similar to PICV NP (Fig. 3E) even though the PICV NP was expressed at greater levels than TNP-GPPT (Fig. 3F). In fact, 0.25 g of TNP-GPPT was enough to decrease the IFN␤ response to SeV infection by 91 Ϯ 2.0% relative to SeV infection alone (Fig. 3E). When protein expression with 1 g of TNP-GPPT was set as 100%, expression levels around 50% (obtained with 0.25 g of TNP-GPPT) were sufficient to confer maximum anti-IFN activity (Fig. 3, E and F). Interestingly, the vTNP that contains a natural GPPT domain was only able to decrease the IFN␤ response by a maximum of 60 Ϯ 7.0% with 1 g of vTNP; however, transfection of 1 g of vTNP resulted in 25% NP expression relative to NP expression with 1 g of codon-optimized TNP-GPPT (Fig. 3, E and F). Therefore, the alterations in IFN␤ inhibition from virus-derived sequences compared with codonoptimized sequences appeared to be directly due to the differences in protein expression levels.
Lack of IFN␤ Induction during TCRV or PICV Infection in a Mouse Macrophage Reporter Cell Line-Because the TCRV strain variant described above contains the important GPPT residues in NP critical for IFN-I inhibition, we wanted to determine whether TCRV infection could prevent IFN induction. IFN␤ activation by TCRV infection was measured in a P388D1 mouse macrophage cell line with an IFN␤ reporter stably integrated (Fig. 4). Reporter P388D1 cells were also infected with PICV, SeV, and recombinant RVFV virus vaccine strain expressing GFP in place of the NSs gene (RVFV-GFP), as controls. As mentioned earlier, PICV NP has potent anti-IFN activity; SeV induces a strong IFN response; and the NSs gene in RVFV inhibits IFN activation, therefore its deletion leads to IFN␤ production during RVFV-GFP infection (17). A range of inoculum for PICV and TCRV was used to infect the P388D1 cells. IFN␤ activation was monitored through expression of a fluorescent reporter via the IFN␤ promoter. For a period of 24 -72 h, the cells did not activate the IFN␤ promoter above the uninfected condition. Infection with SeV or RVFV-GFP induced IFN␤ promoter activation with an elevated signal that was maintained for 72 h. In contrast, the IFN␤ response was not activated in P388D1 reporter cells infected with PICV or TRCV. Virus titers from these infections were measured, and PICV infection was productive at all starting virus concentrations (Fig. 4B), whereas TCRV infection produced moderate amounts of virus in P388D1 at 48 and 72 h after infection (Fig.  4C). In all, neither TCRV nor PICV infection of P388D1 reporter cells activated the IFN␤ response, consistent with NPinduced IFN-I antagonism.

Lack of IFN␤ Transcription and IRF-3 Nuclear Translocation during TCRV or PICV Infection of A549
Cells-Additionally, we tested the IFN-I response in A549 cells during TCRV infection because this human lung epithelial cell line has a functional IFN␤ response and has been used extensively in studies of arenavirus infection (8,19,20). We first determined whether these cells were permissive to infection by TCRV and PICV. A549 cells were infected with TCRV or PICV at varying concentrations, and samples were collected at 12, 24, 48, and 72 h after infection. The results shown in Fig. 5, A and B, indicate that TCRV and PICV productively infect A549 cells. Next, we measured IFN␤ promoter activation by quantitative real-time RT-PCR. Total RNA from the infected cells was isolated at 12, 24, 48, and 72 h after infection. As a positive control for IFN␤ activation, the A549 cells were infected with SeV or RVFV-GFP. The total RNA samples were subjected to analysis by quantitative RT-PCR for IFN␤ mRNA expression, and the levels of IFN␤ induction were compared with the uninfected control (Fig. 5C). SeV produced a strong IFN response early after infection (12 h), whereas RVFV-GFP produced a peak response at 24 h after infection. Interestingly, both TCRV and PICV failed to stimulate an IFN␤ response at 12 or 24 h after infection at all m.o.i..
To determine whether IFN␤ inhibition correlated with lack of IRF-3 nuclear translocation, we infected A549 cells with TCRV, PICV, and control viruses SeV and RVFV-GFP and then used immunofluorescence microscopy to monitor IRF-3 translocation. As expected, SeV and RVFV-GFP induced IRF-3 nuclear translocation at all measured time points (Fig. 6B). Interestingly, both PICV and TCRV infection at 6,9,12, and 24 h failed to induce IRF-3 translocation. In summary, neither TCRV nor PICV infection activates IFN␤ transcription or IRF3 nuclear translocation in A549 cells, consistent with functional NP anti-IFN activity.

DISCUSSION
Arenavirus NPs function primarily in regulating viral transcription and genome replication but also play a role in type I IFN inhibition. With the exception of TCRV, all NPs of NW and OW arenaviruses examined to date have the ability to inhibit IFN␤ production and IRF-3 translocation. Why TCRV NP lacks this important innate immune antagonistic function had not been previously reported. With the goal of identifying the TCRV NP protein domains and specific residues responsible for its lack of anti-IFN function, we used chimeric protein analysis. By exchanging homologous C-terminal domain residues between PICV and TCRV NPs and testing for IFN inhibition, we identified key TCRV NP residues between amino acids 383-407 involved in regulating anti-IFN function.
Sequence alignment of TCRV NP residues 383-407 to NPs from representative NW and OW arenaviruses show that this region is highly conserved and contains only two nonconserved short amino acid stretches, one containing three residues and the other containing four residues (Fig. 2D). In the LASV NP crystal structure, these two aligned short amino acid stretches lie in separate loop regions between three ␤-strands as indicated in Fig. 7A. The 3Ј-5Ј exonuclease activity discovered in the C-terminal domain of LASV NP, essential for in vitro anti-IFN activity, involves catalytic residues located in one of these ␤-sheets highlighted in Fig. 7, A and B. The other two highlighted ␤-sheets contain residues that coordinate zinc ion binding and are thought to be important for C-terminal domain stabilization and/or substrate specificity (10). We speculate that TCRV NP residues DLQL located in predicted loop regions surrounding catalytic residues may alter the ability to coordinate the divalent ion essential for exonuclease activity.
In the context of chimeric PICV NP, both short amino acid stretches in predicted loop regions from TCRV NP (DLQL and AKKQ) were necessary to abolish anti-IFN function. Remark- . Total RNA was extracted, and IFN␤ levels were determined by quantitative RT-PCR and normalized to GAPDH mRNA levels. ⌬⌬C t -based -fold change calculations were set relative to mock infection as described under "Experimental Procedures." -Fold changes were converted to percentage activation, and the SeV-infected level at 12 h was set to 100% activation. Error bars represent S.E. for three experiments conducted in triplicate. , and PICV (m.o.i. ϭ 2) for 12 h and then fixed with 4% paraformaldehyde. IRF-3 protein was labeled in infected cells by immunofluorescence using conditions indicated under "Experimental Procedures." Images were cropped, but relative cell sizes were maintained. The images shown are representative of results from three independent experiments with similar results. B, using the infection conditions and immunofluorescence assay described above, images were captured from cells fixed at 6, 9, 12, and 24 h after infection. The percentage IRF-3 nuclear translocation was quantified by manually assessing IRF-3 translocation in 100 infected cells from several images. n ϭ 3, Ϯ S.D. (error bars).
ably, the chimeric PICV NP containing (essentially) only the DLQL residues from TCRV NP retained IFN suppression (Fig.  2B, lane TP377-388). Because the DLQL residues located near the 3Ј-5Ј exonuclease residues (Fig. 7B) most likely destabilize the 3Ј-5Јexonuclease domain and activity, as described above, the PICV NP containing the DLQL residues potentially inhibits IFN␤ activation through an alternative method. The arenavirus NP was recently reported to bind IRF-activating kinase IKK⑀ through direct protein-protein interactions (21). Functional NP-IKK⑀ interactions were shown to prevent downstream IRF3 phosphorylation and IFN␤ transcription. Therefore, one possibility is that a PICV NP with DLQL loses the 3Ј-5Ј exonuclease activity but retains IKK⑀ binding, whereas chimeric PICV NP with both short amino acid stretches (DLQL and AKKQ) does not have 3Ј-5Ј activity nor IKK⑀ binding. Future studies measuring both NP-IKK⑀ interactions and 3Ј-5Јexonuclease activity may resolve this hypothesis.
Interestingly, we confirm that sequence variation exists in the region involved in TCRV NP IFN antagonism. We showed that the reference sequence of TCRV NP does not have anti-IFN activity, but when DLQL residues (amino acids 389 -392) were mutated to the GPPT residues observed in several independent clones of TCRV, TCRV NP was able to inhibit IFN␤ production (Fig. 3). Martínez-Sobrido et al. also reported this sequence variation in the TCRV NP; however, their results with TCRV NP containing the GPPT sequence showed that the TCRV NP still lacked IFN inhibition function (8,19). The fact that we used codon-optimized plasmids expressing the TCRV NP for cell expression studies may explain the discrepancy in our results. The TCRV NP appears to be weakly expressed com-pared with PICV NP. A comparison of expression levels between codon-optimized plasmids of both NPs indicated that TCRV NP was expressed at 30% of PICV NP (Fig. 1C). Additionally, by comparing unoptimized with codon-optimized TCRV NP expression levels in 293Ts cells, we found that the unoptimized viral sequences were expressed at ϳ23% of codonoptimized sequence (Fig. 3F). Therefore, studies with poorly expressed GPPT containing TRCV NP plasmids may account for apparent lack of anti-IFN function in previous reports.
The frequency at which alternative TCRV sequences exist in nature is currently unknown. Several TCRV isolates were originally obtained from bats in Trinidad in the 1950s; however, only TCRV strain 11573 remains available (22,23). Although the reference sequence, BEI resources, and ATCC TCRV stocks are all derived from strain 11573, variation in sequences may have occurred due to different methods or duration of passaging in cell culture. The TCRV NP clone sequence for BEI resources stock used in this paper has been deposited in Gen-Bank (accession no. KC329849).
Our results also show that TCRV infection does not activate IFN␤ production in P388D1 or A549 cells. Additionally, IRF-3 translocation was not apparent in A549 cells at 6, 9, 12 or 24 h post-infection and similar results were seen with PICV. Although these results are consistent with a functional TCRV NP IFN inhibition mechanism, we cannot rule out the contribution of other proteins or other mechanisms of inhibition. The arenavirus Z protein for instance has also been reported to inhibit IFN␤ activation through IRF-3 interactions (24). In addition to the active suppression mechanism thought to involve both NP and Z during infection, potential stealth strategies may also exist. The complementary untranslated regions of the arenavirus genomes are thought to form a doublestranded RNA structure that acts as a RIG-I innate immune sensor decoy/antagonist (25). Finally, arenaviruses have recently been shown to replicate in discrete cytosolic bodies that can potentially shield the virus from innate immune detection (26). Our results are also consistent with studies that show TCRV infection in primary human monocytes and macrophages did not activate IFN␤ production (27).
In conclusion, we demonstrated that efficiently expressed TCRV NP with the GPPT residues at positions 389 -392 is capable of IFN␤ inhibition. Therefore, this in vitro phenotype of immunosuppression induced by the arenavirus NPs appears to be a conserved feature of all arenaviruses, both pathogenic and nonpathogenic, and may be required for arenavirus replication in IFN-competent cells. In support of this conclusion, the new arenavirus (Golden gate virus) discovered in snakes shares homology to TCRV residues 383-407 (Fig. 7) and contains all catalytic (DEDDh motif) residues responsible for NP exonuclease activity (28).  The partial C-terminal domain of the LASV NP is shown as a ribbon structure. A, highlighted in the center of the NP structure are three ␤-strands (red) and two connecting loop regions (blue) near the TCRV NP DLQL/GPPT residues with an alignment for LASV to TCRV, PICV and the newly discovered Golden gate virus (Snake) arenavirus nucleoproteins. The residues (ϩ) that coordinate zinc ion (blue sphere) binding are thought to be important for C-terminal domain stabilization and/or substrate specificity. The divalent manganese cation (gray sphere) important for in vitro exonuclease activity is positioned using residues (*) in the ␤1-strand (10,11). B, the LASV NP Gly-392, corresponding to TCRV NP residue Asp-389 in the reference (REF) sequence, is located in the active site of the 3Ј-5Ј exonuclease domain. Residue changes in loop domains highlighted in blue are hypothesized to affect divalent cation coordination and influence IFN␤ inhibition efficiency. The image was created using VMD software and PDB file 3MWT (29).