Role of the Host Cell's Unfolded Protein Response in Arenavirus Infection

Abstract

Arenaviruses are enveloped RNA viruses with a nonlytic life cycle that cause acute and persistent infections. Here, we investigated the role of the host cell's unfolded protein response (UPR) in infection of the prototypic arenavirus lymphocytic choriomeningitis virus (LCMV). In mammalian cells, the endoplasmic reticulum (ER) chaperone protein GRP78/BiP functions as the principal sensor for the induction of the UPR and interacts with three mediators: kinase/endonuclease inositol-requiring protein 1 (IRE1), PKR-like ER kinase (PERK), and activating transcription factor 6 (ATF6). Acute infection with LCMV resulted in a selective induction of the ATF6-regulated branch of the UPR, whereas pathways controlled by PERK and IRE1 were neither activated nor blocked. Expression of individual LCMV proteins revealed that the viral glycoprotein precursor (GPC), but not that of other viral proteins, was responsible for the induction of ATF6. Rapid downregulation of the viral GPC during transition from acute to persistent LCMV infection restored basal levels of UPR signaling. To address a possible role of ATF6 signaling in LCMV infection, we used cells deficient in site 2 protease (S2P), a metalloprotease required for the activation of ATF6. Cells deficient in S2P showed significantly lower levels of production of infectious virus during acute but not persistent infection, indicating a requirement for ATF6-mediated signaling for optimal virus multiplication. In summary, acute LCMV infection seems to selectively induce the ATF6-regulated branch of the UPR that is likely beneficial for virus replication and cell viability, but it avoids induction of PERK and IRE1, whose activation may be detrimental for virus and the host cell.

The arenaviruses are a large and diverse family of viruses that merit significant attention as powerful experimental models and important emerging human pathogens. Arenaviruses are enveloped viruses with a bisegmented negative-strand RNA genome and a nonlytic life cycle restricted to the cytoplasm (9, 13). Each genomic RNA L and S segment uses an ambisense coding strategy to direct the synthesis of two polypeptides in an opposite orientation, separated by a noncoding intergenic region (IGR). The S RNA encodes the viral glycoprotein precursor, GPC, and the nucleoprotein, NP, whereas the L RNA encodes the viral RNA-dependent RNA polymerase (RdRp, or L polymerase), and a small RING finger protein, Z. Arenavirus GPC is synthesized as a single polypeptide chain (ca. 75 kDa) and posttranslationally cleaved by the cellular proprotein convertase (PC) subtilisin kexin isozyme 1 (SKI-1)/site 1 protease (S1P) to yield the mature virion glycoproteins GP1 and GP2 (5, 21, 24). The GP1 part mediates virus interaction with host cell surface receptors and is located at the top of the mature glycoprotein (GP) spike present in the viral envelope. GP1 is associated via ionic interactions with the transmembrane GP2 that forms the stalk of the spike and resembles the fusion-active membrane-proximal parts of other enveloped viruses (17). The cellular protease SKI-1/S1P implicated in arenavirus GPC processing also plays a key role in the regulation of cholesterol homeostasis (6, 35) and the host cell's unfolded protein response (UPR) (36, 43).

The cellular UPR, also called the endoplasmic reticulum (ER) stress response, is an adaptive cellular response elicited when the influx of nascent, unfolded polypeptides into the ER exceeds its folding capacity under a variety of conditions, including viral infection (36). In mammalian cells, the ER chaperone protein GRP78/BiP functions as the principal sensor for the induction of the UPR and interacts with three mediators: kinase/endonuclease inositol-requiring protein 1 (IRE1), PKR-like ER kinase (PERK), and activating transcription factor 6 (ATF6). The UPR is triggered by the recognition of partially unfolded proteins by BiP and leads to the release and activation of PERK, IRE1, and ATF6. Activation of PERK results in phosphorylation of eukaryotic translation initiation factor 2α (eIF2α), leading to the attenuation of cap-dependent translation. PERK also induces the transcription factor ATF4 that upregulates the proapoptotic bZIP transcription factor GADD153/CHOP and GADD34. Upon activation, the endonuclease activity of IRE1 removes a 26-nucleotide intron from the mRNA of X-box binding protein 1 (XBP1). This spliced form of XBP1 mRNA codes for a potent transcription factor inducing many UPR genes (36). ATF6 is a type II ER transmembrane protein that contains a bZIP domain in the cytosol and an ER stress-sensing domain in the ER lumen. After ER stress triggers its release from BiP, ATF6 translocates to the Golgi apparatus where it is processed first by SKI-1/S1P at the luminal site and then by the metalloprotease site 2 protease (S2P) at the cytoplasmic site. Proteolytic processing of ATF6 results in release of the N-terminal fragment of ATF6 (nATF6) that translocates to the nucleus and contributes to transcriptional activation of genes involved in the UPR, including BiP itself, the chaperone GRP94, and CHOP (44). The main role of the UPR is to bring the folding capacity of the ER in line with the folding demand. However, if these cellular countermeasures fail and ER stress occurs over prolonged periods of time, the UPR switches from a prosurvival to a proapoptotic signal and induces programmed cell death.

Many enveloped DNA and RNA viruses can induce the UPR in mammalian cells, including the flaviviruses Japanese encephalitis virus (39), hepatitis C virus (30), and West Nile virus (28), retroviruses (14, 15), Borna disease virus (42), coronaviruses (3, 4, 12, 40), human cytomegalovirus (CMV) (7, 8, 19, 20), reoviruses (38), and Semliki Forest virus (2). Virus-induced UPR frequently involves proapoptotic signaling contributing to pathogenesis (2, 14, 15, 28, 39, 42). Some viruses are also capable of manipulating the host cell's UPR to benefit their replication (7, 8, 12, 19, 20, 30, 38).

A hallmark of arenaviruses is their nonlytic life cycle and their ability to establish persistent infections in cultured cells and in vivo (9). Since arenaviruses do not integrate into the host genome, their long-term persistence requires some level of continued replication and expression of the full repertoire of viral proteins, including the viral GPC. In this work, we have examined the hypothesis that the ability of arenaviruses to establish persistent productive infection in the host cell may involve the capability of these viruses to differentially modulate the distinct branches of the UPR: maintaining the protective responses favorable for virus multiplication and evading and/or blocking the proapoptotic pathways. We found that during acute LCMV infection, high expression levels of the viral glycoprotein precursor (GPC) selectively induced the ATF6-regulated branch of the UPR, whereas pathways controlled by PERK and IRE1 remained silent. Rapid downregulation of GPC during the transition to persistent infection restored basal levels of UPR signaling. ATF6-mediated signaling was required for optimal virus multiplication in acute but not in persistent infection.

MATERIALS AND METHODS

Antibodies.

Mouse monoclonal antibodies (MAbs) 113 (anti-lymphocytic choriomeningitis virus [anti-LCMV] NP) and 83.6 (anti-LCMV GP) were described previously (10, 41), as were rabbit polyclonal Abs recognizing LCMV Z (23). Guinea pig immune serum to LCMV was a kind gift of Michael Oldstone, The Scripps Research Institute. Mouse MAb anti-eIF2α and rabbit MAb anti-phospho-eIF2α (Ser51) were from Cell Signaling (Danvers, MA), while mouse MAb anti-α-tubulin and mouse MAb anti-FLAG M2 were from Sigma-Aldrich (St. Louis, MO). Horseradish peroxidase (HRP)-conjugated secondary antibodies were from Pierce (Rockford, IL), while Rhodamine Red-X (RRX)-labeled goat anti-mouse IgG was from Jackson ImmunoResearch Laboratories (Suffolk, United Kingdom).

Cells and viruses.

Human lung adenocarcinoma epithelial cells (A549), human hepatoma cells (Huh7), and Syrian golden hamster kidney cells (BHK-21), were maintained in Dulbecco's modified Eagle medium (DMEM) (Gibco BRL, NY) containing 10% fetal bovine serum (FBS). Chinese hamster ovary (CHO)-K1 cells were grown in DMEM nutrient mixture Ham F-12 (Gibco) supplemented with 10% FBS. The S2P-deficient CHOK1-derived cell line M19 (18) was maintained in complete medium supplemented with 5 μg/ml cholesterol (Sigma-Aldrich), 1 mM sodium mevalonate (Sigma-Aldrich), and 20 μM sodium oleate (Sigma-Aldrich). LCMV Armstrong 53b (ARM53b) has been already described (22). Seed stocks of LCMV were prepared by growth in BHK-21 cells, and titers were determined as reported previously (16). Briefly, BHK-21 cells were infected with 10-fold serial dilutions of the virus stock and fixed 16 h postinfection (p.i.), and infected cells were stained using MAb 113 combined with a fluorescence-labeled secondary antibody.

Virus infection.

For infection of cells with LCMV, seed stocks were diluted to the multiplicity of infection (MOI) indicated in the figure legends and added to cells for 1 h at 37°C. After 1 h of incubation, the inoculum was removed, cells were washed twice with medium, and normal medium was added. Cells were cultivated at 37°C and 5% (vol/vol) CO₂ for the time points indicated in the figure legends and then further processed.

Western blot analysis.

Cells were lysed in radioimmunoprecipitation assay (RIPA) buffer (50 mM Tris, 150 mM NaCl, 0.5% deoxycholate [DOC], 1% NP-40, 0.1% SDS, pH 8.0) supplemented with protease inhibitor cocktail (Roche, Rotkreuz, Switzerland) and 1 mM sodium orthovanadate. Total protein concentrations were quantified using a Micro BCA protein assay kit (Pierce) according to the manufacturer's instructions. Lysates were then analyzed by SDS-PAGE and immunoblotting as described previously (32). Densitometric analysis of blots was performed as reported previously (21).

Real-time PCR.

Total RNA was purified with an RNeasy mini kit (Qiagen, Chatsworth, CA), and cDNA was synthesized using a QuantiTect reverse transcription kit (Qiagen). TaqMan probes specific for BiP (Hs99999174_m1), GRP94 (Hs00427665_g1), CHOP (Hs01090850_m1), GADD34 (Hs00169585_m1), S1P (Hs00186886_m1), and glyceraldehyde-3-phosphate dehydrogenase (GAPDH; Hs99999905_m1) were obtained from Applied Biosystems (Foster City, CA). Real-time PCR was performed using a StepOne real-time PCR system (Applied Biosystems), and gene expression levels relative to GAPDH were determined according to the 2^−ΔΔCT method (25).

Pharmacological treatment.

In order to artificially trigger the UPR in positive controls, cells were treated 1 h with 5 mM dithiothreitol (DTT), 4 h with 600 nM thapsigargin (Tg), or 4 h with 5 μg/ml tunicamycin (all from Sigma-Aldrich) as described previously (43).

XBP-1 splicing assay.

Total RNA was purified with an RNeasy mini kit (Qiagen, Chatsworth, CA), and cDNA was synthesized using a QuantiTect reverse transcription kit (Qiagen). XBP1 cDNA was amplified using primers encompassing the IRE-I splicing site (5′-CATGGCCTTGTAGTTG-3′, 5′-CTGGGTCCACCAAGTTGT-3′). PCR products of 270 bp and 244 bp, representing unspliced (u) and spliced (s) XBP-1, respectively, were separated on 2% agarose gels.

Plasmids and transfection.

Expression plasmids for LCMV NP, GPC, Z, and L were described previously (11, 23). In transfection experiments, 4 × 10⁶ Huh7 cells were electroporated at 150 V for 70 ms in 4-mm gap cuvettes using an ECM830 generator (BTX Harvard Apparatus, Holliston, MA) and the amount of plasmid DNA indicated in the legend to Fig. 5. Cells were analyzed 48 h after transfection.

Caspase activation assay.

Cells were infected with LCMV ARM53b at an MOI of 3, and activation of caspases 3/7, 8, and 9 was tested at 24, 36, and 48 h postinfection with Caspase-Glo 3/7, 8, and 9 assays (Promega Dübendorf, Switzerland), respectively. As positive control, cells were treated with 5 μM Staurosporin (St) for 3 h to induce the activation of caspases 3/7 and 9 or with 500 nM soluble tumor necrosis factor (TNF)-related apoptosis-inducing ligand (TRAIL) for 5 h to trigger caspase 8.

Statistical analysis.

Real-time PCR data were analyzed using two-way analysis of variance (ANOVA) followed by a Bonferroni post hoc test with GraphPad Prism (San Diego, CA). Statistical significance compared with the uninfected cells (see Fig. 2 and 4) or mock-transfected cells (see Fig. 5E) is indicated.

RESULTS

Differential expression of the viral GPC and NP during acute and persistent LCMV infection.

Previous studies of LCMV infection in vitro and in vivo revealed that expression levels of the viral NP and GPC are differently regulated during acute and persistent infections (29). NP expression levels remained high and relatively constant during the acute and persistent phases of LCMV infection. In contrast, GPC was highly expressed during the first days of acute infection, followed by a marked downregulation during the transition to persistent infection. We hypothesized that differential expression of GPC during acute and persistent infection posed likely distinct challenges to the host cells' protein folding machinery and the cellular UPR. To study the role of the UPR during different phases of LCMV infection, we first established the temporal expression of GPC in suitable cell lines. The human lung epithelial cell line A549 and the hepatocyte cell line Huh7 were infected with the LCMV strain ARM53b, and expression of GPC and NP was examined over time by Western blotting. In line with published data, we observed maximal GPC expression levels at 36 to 48 h p.i., followed by a marked reduction after 72 h (Fig. 1), whereas NP reached a plateau at 24 h of infection and remained constant thereafter. At 96 h p.i., >98% of cells remained positive for NP, and high levels of infectious virus production were detected in cell culture supernatants (data not shown).

FIG. 1.

Differential regulation of LCMV NP and GP expression during acute infection. Huh7 and A549 cells were infected with LCMV ARM53b (MOI, 0.1). At the indicated time points postinfection (p.i.), cells were lysed in SDS-PAGE sample buffer, and proteins were separated and probed by Western blotting using monoclonal antibody (MAb) 83.6 to LCMV GP2 (GP2), polyclonal guinea pig serum to LCMV NP (NP), and a MAb to α-tubulin (α-Tu), combined with HRP-conjugated secondary antibodies and enhanced chemiluminescence (ECL) as described previously (32). Alpha-tubulin was used as the loading control.

LCMV infection transiently induces the ATF6-regulated branch of the UPR.

To examine induction of the ATF6-regulated branch of the cellular UPR during LCMV infection, we monitored expression of a specific set of ATF6-regulated genes, including the chaperones BiP and GRP94, as well as the transcription factor CHOP (36). We further included GADD34, which is regulated by CHOP (27). Huh7 cells were infected with LCMV, and expression levels of GPC were assessed over time by Western blot analysis (Fig. 2A). To monitor expression of ATF6-regulated genes, we determined transcription levels of BiP, GRP94, CHOP, and GADD34 by real-time quantitative PCR using specific TaqMan probes; we also included a probe for the housekeeping gene GAPDH for normalization. At 36 to 48 h p.i., we detected a significant induction of the expression of BiP, GRP94, and CHOP (Fig. 2B to D), concomitant with the transient expression of high GPC levels (Fig. 2A). The marked reduction of GPC expression after 72 h (Fig. 2A) correlated with a drop in transcription levels of BiP, GRP94, and CHOP to levels similar to those for uninfected cells (Fig. 2B to D). The UPR gene GADD34 was also induced but showed a delay in upregulation, with maximal levels reached at 48 h p.i. and significantly elevated mRNA levels after 72 h of infection (Fig. 2E). These delayed kinetics of GADD34 induction are consistent with the fact that GADD34 is regulated by CHOP, which is itself a downstream target of ATF6 (36). Interestingly, despite similar GPC expression levels at 36 and 48 h p.i., we consistently observed lower induction levels of BiP, GRP94, and CHOP at 48 h (Fig. 2B to D). The reasons for this apparent downregulation of ATF6 activity are unknown, but it may be due to the enhanced folding capacity of the ER upon induction of the ATF6-regulated branch of the UPR at 36 h, which partially releases ER stress. Levels of SKI/S1P mRNA remained unchanged throughout LCMV infection (not shown); therefore, changes in levels of SKI/S1P expression were unlikely to have contributed to the regulation of ATF6 activation.

FIG. 2.

LCMV infection induces UPR signaling mediated by ATF6. (A) Kinetics of GPC expression. Huh7 cells were either left uninfected (u) or were infected with LCMV ARM53b (MOI, 0.1) and lysed at the indicated time points p.i.; expression of the viral GPC was detected by Western blotting as in Fig. 1. Alpha-tubulin (α-Tu) was used as the loading control. (B to E) Analysis of the induction of the UPR genes BiP, GRP94, CHOP, and GADD34 during LCMV infection. Cells infected as in panel A were lysed at the indicated time points p.i., total RNA was extracted, and cDNA was synthesized by reverse transcription. Real-time PCR was performed using a StepOne real-time PCR system (Applied Biosystems) and TaqMan probes specific for BiP, GRP94, CHOP, GADD34, and GAPDH. Gene expression levels relative to GAPDH were determined according to the 2^−ΔΔCT method (25). Data presented are the fold induction above levels for uninfected cells (means ± standard deviations [SD]; n = 3). Statistical significance was assessed by two-way ANOVA, and significant differences are indicated. (F) Activation of the UPR genes BiP, GRP94, and CHOP by pharmacological treatment with tunicamycin and thapsigargin. Huh7 cells were treated for 4 h with dimethyl sulfoxide (DMSO), 600 nM thapsigargin (Tg), or 5 μg/ml tunicamycin (Tm), total RNA was extracted, and the mRNA levels of BiP, GRP94, and CHOP were determined by real-time PCR as in panels B to D. Data are presented as the fold induction above levels for mock (DMSO)-treated cells (means ± SD; n = 3).

Next, we compared the profile of the virus-induced upregulation of the ATF6-regulated genes BiP, GRP94, and CHOP with the induction profiles obtained with pharmacological agents classically used to trigger the cellular UPR: tunicamycin (Tm), an inhibitor of N-linked glycosylation, which causes accumulation of underglycosylated unfolded proteins in the ER lumen and thapsigargin (Tg), an inhibitor of the Ca²⁺-dependent ATPase in the ER (44). Huh7 cells were treated for 4 h with 600 nM Tg or 5 μg/ml Tm, and mRNA levels of BiP, GRP94, and CHOP were determined by real-time PCR (Fig. 2F). LCMV infection (Fig. 2B and C) and treatment with Tm or Tg (Fig. 2F) resulted in similar levels of induction of the chaperones BiP and GRP94, whereas exposure to Tg or Tm resulted in a much stronger upregulation of CHOP (Fig. 2F) than that seen in results for virus infection (Fig. 2D). These findings likely reflected the fact that BiP and GRP94 are regulated mainly by ATF6, whereas the regulation of CHOP involves ATF6 and PERK (26). This difference in the relative activation of CHOP by LCMV is of interest, as CHOP plays a prominent role in the induction of apoptosis as a result of sustained ER stress (36, 44). The weak induction of CHOP by LCMV infection compared to that by Tm and Tg further suggested that LCMV may be able to differentially induce the branches of the UPR regulated by ATF6 and PERK.

LCMV infection neither activates nor blocks PERK and IRE1.

Activation of the UPR sensor PERK in response to ER stress results in a shut-off host cell translation by phosphorylation of eIF2α, and prolonged activation of PERK triggers proapoptotic signaling involving CHOP (34). To examine LCMV-induced activation of PERK, we monitored eIF2α phosphorylation in LCMV-infected Huh7 cells by Western blotting using an antibody specifically recognizing the Ser51 phosphorylated form of eIF2α that is generated upon PERK activation. Phospho-eIF2α signals were normalized to total eIF2α. Under our experimental conditions, we were unable to detect a significant increase in eIF2α phosphorylation at any time throughout the first 72 h of infection (Fig. 3A). Treatment of uninfected cells with Tg for 4 h induced a robust increase in eIF2α phosphorylation, proving that the PERK pathway is functional in these cells (Fig. 3A).

FIG. 3.

LCMV activation neither activates nor blocks IRE1 and PERK. (A) LCMV infection does not induce phosphorylation of eIF2α. Huh7 cells were mock infected (u) or infected with LCMV as in Fig. 2A. At the indicated time points, cells were lysed in SDS sample buffer containing 1 mM sodium orthovanadate. Proteins were separated by SDS-PAGE, blotted onto nitrocellulose membranes, and probed with rabbit MAb anti-phospho-eIF2α (Ser51) (P-eIF2α). Total eIF2α was detected with a MAb to eIF2α that is insensitive to phosphorylation. As a positive control, cells were left untreated (u) or treated for 4 h with 600 nM Tg (Tg) and lysed, and phosphorylation of eIF2α was detected as described above. Note the induction of phospho-eIF2α (P-eIF2α) in Tg-treated cells. (B) IRE1 is not activated during LCMV infection. Cells were mock infected (u) or infected with LCMV as in panel A. As a positive control, cells were treated for 4 h with 600 nM Tg. At the indicated time points p.i. or upon Tg treatment (Tg), cells were lysed and total RNA was extracted. After cDNA synthesis was performed as in panels B to E, XBP-1 cDNA was amplified using primers encompassing the IRE-I splicing site. The resulting PCR products of 270 bp and 244 bp, representing unspliced (u) and spliced (s) XBP-1, respectively, were separated on agarose gels. In addition to the bands corresponding to the unspliced form (uXBP1) and the spliced form (sXBP1), we consistently detected a third band in Tg-treated samples. This phenomenon has been reported before and can be explained by the formation of hybrids (hXBP1) between the spliced and unspliced forms during annealing. For a detailed discussion, see the work by Shang and Lehrman (37). (C) Induction of eIF2α phosphorylation by Tg is not blocked in LCMV-infected cells. Huh7 cells were either mock infected (u) or infected with LCMV as in panel A. At the indicated time points p.i., cells were treated with 600 nM Tg for 4 h and the phosphorylation of eIF2α was detected as in panel A. (D) Induction of XBP1 splicing by IRE1 is not inhibited in LCMV-infected cells. Mock-infected (u) and LCMV-infected cells were treated with Tg as in panel C, and splicing of XBP1 was detected as in panel B.

To examine LCMV-induced activation of IRE1, we monitored splicing of the mRNA of XBP1 in LCMV-infected Huh7 cells at different times p.i., by reverse transcription (RT)-PCR using primers encompassing the IRE-I splicing site. The resulting PCR products of 270 bp and 244 bp, representing unspliced (u) and spliced (s) XBP-1, respectively, were resolved by agarose gel electrophoresis and visualized by ethidium bromide staining (Fig. 3B). We did not observe detectable splicing of XBP1 at any time point p.i., whereas treatment of uninfected parallel cultures with Tg for 4 h resulted in robust detection of spliced XBP-1 mRNA, demonstrating normal function of IRE1 in Huh7 cells (Fig. 3B).

The induction of ATF6, but that of not PERK or IRE1, during acute LCMV infection suggested that LCMV either selectively avoided activation of PERK and IRE1 or was able to actively block these pathways. To address the latter possibility, we studied the function of the PERK and IRE1 pathways in LCMV-infected cells. Huh7 cells were infected with LCMV or mock infected and, at different time points p.i., treated with Tg for 4 h, followed by the detection of eIF2α phosphorylation and XBP1 splicing as described above (Fig. 3C and D). LCMV infection perturbed neither Tg-induced eIF2α phosphorylation nor XBP1 splicing at any time point, respectively, making a virus-mediated block of PERK or IRE1 unlikely. In summary, our data indicate that LCMV infection selectively induces ATF6-regulated UPR genes, but neither activates nor blocks PERK and IRE1.

LCMV infection is nonlytic and normally not associated with overt cytopathic effects. However, the transient induction of CHOP observed in our study opened the possibility of induction of apoptosis. The extrinsic pathway of apoptosis involves early activation of caspase 8, whereas triggering of the intrinsic apoptotic pathway results in activation of caspase 9. Both pathways activate downstream caspases, including caspase 3. To assess activation of cellular apoptosis, we assessed activation of caspases 3, 8, and 9 during LCMV infection using Caspase-Glo 3/7, Caspase-Glo 8, and Caspase-Glo 9 assays (for details, see Materials and Methods). At no time during LCMV infection did we detect significant activation of the tested caspases (data not shown), indicating that the transient induction of CHOP is insufficient to induce significant apoptosis.

LCMV infection affects the sensitivity of the ATF6-regulated branch of the UPR.

The dynamic and transient induction of ATF6 observed during acute LCMV infection opened the possibility that LCMV may modulate the sensitivity of this pathway to ER stress. To investigate this possibility, we compared the sensitivity levels of the ATF6-regulated branch of the UPR toward Tg in LCMV-infected and uninfected control Huh7 cells. At the time points indicated in Fig. 4, infected and uninfected cells were exposed to Tg for 4 h, and transcription levels of BiP, GRP94, CHOP, and GADD34 were determined by quantitative RT-PCR. At 36 h p.i., Tg treatment of infected cells resulted in significantly higher levels of induction of BiP, GRP94, and CHOP than those for uninfected cells (Fig. 4). This apparent hypersensitivity of the ATF6-regulated pathway disappeared after 48 h of infection, probably due to the enhanced folding capacity in the ER as a result of the prior induction of ATF6-regulated UPR genes. Likewise, ATF6-independent induction of GADD34 in response to Tg was also enhanced in infected cells at 36 h p.i., albeit to a lesser extent. Our data rather indicate that LCMV infection causes a transient hypersensitivity of the ATF6-regulated branch of the cellular UPR and that downregulation of viral GPC expression restores normal levels of ATF6 responsiveness in persistently infected cells. At no time point p.i. did we observe a reduced sensitivity of the ATF6 pathway, excluding perturbation of ATF6-induction by the virus.

FIG. 4.

LCMV infection transiently changes the sensitivity of the ATF6-regulated branch of the UPR. At the indicated time points, Huh7 cells that had been mock infected (uninfected) or infected with LCMV ARM53b (MOI, 1) were treated with 600 nM thapsigargin for 4 h. Cells were lysed, total RNA was extracted, and the mRNA levels of BiP, GRP94, CHOP, and GADD34 were assessed by real-time PCR as in Fig. 2B to E. Data shown are expression levels relative to GAPDH as determined using the 2^−ΔΔCT method (means ± SD; n = 3). Statistical significance was assessed by two-way ANOVA, and significant differences are indicated.

Expression of recombinant LCMV GPC, but not that of NP, Z, and L, activates the ATF6 branch of the cellular UPR.

To identify the viral protein responsible for the observed induction of ATF6, we expressed each LCMV protein individually in recombinant form and assessed activation of ATF6 by monitoring expression of its downstream target BiP. Briefly, Huh7 cells were transfected with expression constructs for LCMV GPC, NP, Z, and L using electroporation that resulted in circa 60% transfection efficiency. After 48 h, cells were lysed and recombinant proteins detected by Western blotting, using α-tubulin for normalization. Compared to results for cells infected with LCMV for 36 h in parallel, we observed similar expression levels of GPC and Z (Fig. 4B and 5A), whereas expression levels of NP were higher in transfected cells (Fig. 5C). In case of polymerase L, we were able to detect the recombinant protein, but not L, in infected cells (Fig. 5D). To assess induction of BiP, parallel specimens were subjected to RNA extraction, followed by quantitative real-time PCR analysis as described above. As shown in Fig. 5E, transfection of cells with GPC, but not that with NP, Z, and L, significantly induced BiP mRNA. These experiments suggest that expression of GPC is responsible for the induction of the ATF6-regulated branch of the cellular UPR during acute LCMV infection.

FIG. 5.

Expression of LCMV GPC, but not that of NP, Z, and L, induces the cellular UPR. (A to D) Expression of recombinant LCMV proteins in Huh7 cells. Huh7 cells were transfected with the expression plasmids pC-LCMV-GPC, pC-LCMV-NP, pC-LCMV-Z-HA, and pC-LCMV-L-FLAG using electroporation. For LCMV GPC and NP, increasing amounts of plasmid were used (40, 60, and 80 μg). Mock-transfected cells were electroporated with an empty vector. Transfection efficiency was assessed using the plasmid pC-EGFP containing an enhanced green fluorescent protein (EGFP) reporter. The electroporation procedure resulted in significant cell death immediately after transfection, and α-tubulin (α-Tu) was detected for normalization. After 48 h, cells were lysed and recombinant proteins were detected by Western blotting using specific antibodies. For comparison, LCMV-infected cells at 36 h postinfection (LCMV 36hpi) were used. Note that for LCMV Z, the hemagglutinin (HA) tag results in a significant shift in the apparent molecular mass of the recombinant protein (Z-HA) compared to the wild-type (Z) present in infected cells. The unspecific band (*) was observed for all samples and is due to a cross-reactivity of the antibody used. (E) Expression of recombinant LCMV GPC, but not that of NP, Z, and L, induces ATF6. Huh7 cells were transfected with recombinant proteins as in panels A to D. Total RNA was extracted, and the induction of the ATF6 downstream target BiP was examined as in Fig. 2B. Data presented are the fold induction above levels for uninfected cells (means ± SD; n = 3). Statistical significance was assessed by two-way ANOVA, and significant differences are indicated. (F) Expression of recombinant GPC results in induction of IRE1. Huh7 cells were either transfected with recombinant LCMV GPC as in panel A, infected with LCMV for 36 h as in Fig. 2B, or treated with Tg as in Fig. 3B. Total RNA was extracted, and splicing of XBP1 in response to activation of IRE1 was monitored as in Fig. 3B. (G) Expression of recombinant GPC results in only mild induction of PERK. Total lysates of Huh7 cells treated as in panel F were probed for phospho-eIF2α as in Fig. 3A, and signals were normalized to total eIF2α levels. Determination of the relative signal intensities for phospho-eIF2α and total eIF2α (P-eIF2α/eIF2α) by densitometry revealed only mild induction of PERK in transfected cells.

LCMV GPC expression from our expression plasmid likely has different kinetics than GPC expression during viral infection and may result in a higher expression level of GPC in some cells. We therefore assessed the effect of GPC expression upon electroporation of Huh7 cells with increasing amounts of plasmid DNA on activation of IRE1 and PERK. In contrast to results for infected cells at 36 h p.i., we detected significant splicing of XBP1 in cells transfected with increasing amounts of recombinant GPC (Fig. 5F), but only a very mild increase in phosphorylation of eIF2α (Fig. 5G). Considering the similar overall expression levels of GPC in transfected and infected cell cultures (Fig. 5A), the induction of IRE1 upon expression of recombinant GPC seems surprising. However, this discrepancy may be explained by the different expression kinetics of LCMV GPC upon transfection and/or higher GPC expression on a per-cell basis in some transfected cells. In summary, our data show that the viral GPC is the major component responsible for induction of the cellular UPR and that the specific induction of ATF6, but not that of IRE1, observed with infected cells critically depended on accurate regulation of GPC expression during infection. In contrast, the GPC levels reached during infection as well as transfection in our system were insufficient to induce significant activation of PERK.

ATF6 activation by S2P is required for optimal LCMV multiplication in acute but not persistent infection.

Activation of ATF6 in cells upon ER stress is mediated by the sequential action of the proteases SK1-I/S1P in the Golgi apparatus and the metalloprotease S2P in the cytoplasm. While SKI-1/S1P is also needed to process the arenavirus GPC, S2P has not been implicated in viral GPC processing. To specifically address the role of ATF6 signaling in LCMV replication without affecting GPC processing, we used the mutant CHOK1-derived cell line M19 that is deficient in S2P but has normal levels of SKI-1/S1P activity (18, 31). Upon treatment with a number of ER stress stimuli, including Tm and Tg, M19 cells are unable to activate ATF6 (43). Since S2P is also implicated in the regulation of cholesterol biosynthesis (31), we cultured M19 cells in medium supplemented with cholesterol and lipids as described previously (31). To address a possible role for S2P-dependent ATF6 activation in cell entry and early replication of LCMV, M19 and wild-type CHOK1 cells were infected with LCMV at different MOIs, and infection was detected after one round of replication (16 h p.i.) by intracellular immunofluorescence staining for NP. Similar levels of infection were observed for M19 and CHOK1 cells (Fig. 6A), indicating that S2P-dependent ATF6 activation is not required for LCMV cell entry and early replication. To look at the role of S2P-dependent ATF6 activation in a later phase of the LCMV life cycle, we determined infectious virus production from infected cells. As a control, we included a variant of LCMV bearing a furin recognition site in its GPC (LCMV-RRRR) that does not depend on SKI-1/S1P for GPC processing (33). M19 and CHOK1 cells were infected with wild-type LCMV and LCMV-RRRR at MOI of 3, which resulted in infection of >95% of cells after 16 h (data not shown), and infectious virus titers in the cell supernatant were monitored. Up to 48 h postinfection, S2P-deficient cells showed significantly lower virus production than wild-type cells, but they reached almost wild-type levels after 72 h (Fig. 6B). We also addressed the ability of LCMV to establish persistent infection in S2P-deficient cells. For this purpose, M19 and CHOK1 cells were infected with wild-type LCMV and LCMV-RRRR, and infected cells were passaged every 4 days for 3 weeks. Detection of infectious virus titers in cell culture supernatants revealed similar levels of virus production for the two cell types (Fig. 6C), indicating that LCMV can establish persistence in the absence of a functional ATF6 pathway. The data obtained with S2P-deficient cells suggest that S2P-mediated activation of ATF6 is required for optimal multiplication of LCMV during acute infection but not in persistently infected cells.

FIG. 6.

S2P-mediated ATF6 activation is required for optimal LCMV multiplication in acute infection but in not persistently infected cells. (A) Cell entry and early infection of LCMV are not affected in S2P-deficient cells. M19 and wild-type CHOK1 cells were plated in eight-well LabTek tissue culture chambers (2 × 10⁴ cells/well) and infected with LCMV at the indicated MOI. After 16 h, cells were fixed and stained for intracellular LCMV NP using MAb 113 anti-LCMV NP and a Rhodamine Red-X-labeled secondary antibody as described previously (22). Cell nuclei were counterstained with 4′,6-diamidino-2-phenylindole (DAPI), and specimens were examined by immunofluorescence. For each specimen, 100 DAPI-labeled cells were examined and NP-positive cells were scored (means ± SD; n = 4). (B) LCMV production in M19 and CHOK1 cells during the initial 72 h of infection (acute infection). M19 and CHOK1 cells plated in M6 tissue culture plates (4 × 10⁵ cells/well) were infected with wild-type LCMV (LCMV-WT) and an LCMV variant whose GPC is processed by furin (LCMV-RRRR) at an MOI of 3. Cells were extensively washed, and fresh medium was added. At the indicated time points, cell culture supernatants were taken and virus titers were determined by an immunofocus assay on fresh monolayers of Vero E6 cells as described previously (33). Results are means from duplicate experiments, and one of three representative experiments is shown. (C) LCMV production in persistently infected M19 and CHOK1 cells. M19 and CHOK1 cells plated in M6 tissue culture plates (4 × 10⁵ cells/well) were infected with wild-type LCMV (LCMV-WT) and LCMV-RRRR at an MOI of 3. Cells were extensively washed, fresh medium was added, and cells were cultured for a total of 20 days. Every 4 days, cells were passaged, and virus titers in cell culture supernatants were determined as in panel B. Results are means from duplicate experiments, and one of two representative experiments is shown.

DISCUSSION

The present study addresses for the first time the role of the cellular UPR in arenavirus infection using the prototypic LCMV. Infection of cells with LCMV resulted in a transient induction of the ATF6-regulated branch of the UPR concomitant with the peak of GPC expression, whereas the pathways regulated by PERK and IRE1 were neither activated nor blocked by the virus. Expression of individual viral proteins revealed that the induction of the cellular UPR was caused mainly by the viral GPC but not by the other viral proteins, NP, Z, and L. Rapid downregulation of GPC during the transition to persistent infection restored basal levels of ATF6 signaling. We further demonstrate a requirement for ATF6-mediated signaling for optimal virus multiplication in acute, but not persistent, infection.

A major trigger of the UPR upon infection with many viruses is the expression of structural proteins of the viral envelope in the ER, likely overwhelming the folding capacity and resulting in the accumulation of unfolded proteins. In case of arenaviruses, the translocation of the GPC from the ER to the Golgi apparatus is subject to tight regulation (1). It is thus conceivable that unfolded GPC accumulates in the ER of infected cells during the phase of maximal GPC expression and triggers the UPR. A crucial role for GPC in the induction of the cellular UPR in arenavirus infection is supported by the observation that only GPC, but not the other viral proteins NP, Z, and L, could induce the cellular UPR when expressed as individual protein in recombinant form. The rapid downregulation of GPC during the transition from acute to persistent infection likely releases ER stress and resets UPR signaling and gene expression back to basal levels.

Acute infection with LCMV resulted in a selective induction of the ATF6-regulated branch of the UPR, whereas pathways controlled by PERK and IRE1 remained silent. This selective and reversible activation of ATF6, but not that of PERK and IRE1, during acute LCMV infection suggests that the virus evolved to fine-tune the dynamics of GPC expression, allowing transient induction of ATF6-regulated UPR genes, such as BiP and GRP94, that augment the folding capacity in the ER. This adaptive response of the host cell may facilitate the biosynthesis of large amounts of viral GPC required for optimal virus production during acute infection and protect the cell against virus-inflicted damage. At the same time, GPC expression seems at all time points sufficiently low to prevent the induction of PERK and IRE1, whose activation contributes to translational arrest and proapoptotic signaling (36). The importance of tight regulation of GPC expression for the selective induction of the ATF6 branch of the UPR is illustrated by the detectable activation of IRE1 upon transfection of cells with increasing amounts of recombinant GPC. The UPR branch regulated by PERK was neither induced in infected cells nor induced upon transfection of recombinant LCMV GPC in our system. The latter finding may be due to limited transfection efficiency in our system, such that GPC expression levels sufficient to induce PERK are not reached, and we cannot at present exclude the possibility that excessive GPC expression may induce the PERK pathway.

A beneficial role of the ATF6 branch of the UPR for optimal virus multiplication in acutely infected cells is indeed suggested by the reduced virus production in LCMV-infected S2P-deficient cells that are unable to activate ATF6. Since S2P-dependent ATF6 activation was dispensable for cell entry and early replication assessed by NP expression, ATF6-regulated UPR genes may play a role in optimal GPC biosynthesis and other steps late in the viral life cycle. LCMV was able to establish chronic productive infection in S2P-null cells, indicating that the ATF6-regulated UPR is not essential for viral persistence that is characterized by markedly reduced GPC expression levels. In summary, our data suggest that acute LCMV infection seems to selectively induce the ATF6-regulated branch of the host cell's UPR that is likely beneficial for virus replication and cell viability. Expression of the viral GPC seems to be regulated in a way to avoid induction of PERK and IRE1, whose activation may be detrimental for the virus and the host cell. During the transition to persistent infection, downregulation of the viral GPC resets the cellular UPR to normal levels. This may allow the persistent virus to “merge” into the normal background level of ER stress of the host cell and may be crucial for the establishment of a long-term asymptomatic chronic infection.