Human Cytomegalovirus Infection Activates and Regu
http://www.100md.com
病菌学杂志 2005年第11期
Department of Cancer Biology, Abramson Family Cancer Research Institute
Department of Pathology and Laboratory Medicine, School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania
ABSTRACT
Viral infection causes stress to the endoplasmic reticulum. The response to endoplasmic reticulum stress, known as the unfolded protein response (UPR), is designed to eliminate misfolded proteins and allow the cell to recover by attenuating translation and upregulating the expression of chaperones, degradation factors, and factors that regulate the cell's metabolic and redox environment. Some consequences of the UPR (e.g., expression of chaperones and regulation of the metabolism and redox environment) may be advantageous to the viral infection; however, translational attenuation would not. Thus, viruses may induce mechanisms which modulate the UPR, maintaining beneficial aspects and suppressing deleterious aspects. We demonstrate that human cytomegalovirus (HCMV) infection induces the UPR but specifically regulates the three branches of UPR signaling, PKR-like ER kinase (PERK), activating transcription factor 6 (ATF6), and inositol-requiring enzyme 1 (IRE-1), to favor viral replication. HCMV infection activated the eIF2 kinase PERK; however, the amount of phosphorylated eIF2 was limited and translation attenuation did not occur. Interestingly, translation of select mRNAs, which is dependent on eIF2 phosphorylation, did occur, including the transcription factor ATF4, which activates genes which may benefit the infection. The endoplasmic reticulum stress-induced activation of the transcription factor ATF6 was suppressed in HCMV-infected cells; however, specific chaperone genes, normally activated by ATF6, were activated by a virus-induced, ATF6-independent mechanism. Lastly, HCMV infection activated the IRE-1 pathway, as indicated by splicing of Xbp-1 mRNA. However, transcriptional activation of the XBP-1 target gene EDEM (ER degradation-enhancing -mannosidase-like protein, a protein degradation factor) was inhibited. These results suggest that, although HCMV infection induces the unfolded protein response, it modifies the outcome to benefit viral replication.
INTRODUCTION
Human cytomegalovirus (HCMV) is a betaherpesvirus which can cause significant medical problems in individuals with immature or compromised immune systems. The genome of HCMV is 230 kb of double-stranded DNA with the potential to encode over 200 proteins. Like that of other herpesviruses, HCMV viral gene expression occurs in an ordered temporal pattern having immediate-early, early, delayed-early, and late kinetics, with increasing viral protein synthesis over time. The burden of HCMV infection on the host cell has been shown to initiate a number of cellular stress responses. Here, we investigated the stress of HCMV infection on the endoplasmic reticulum (ER).
The ER is an extensive membranous network that provides a unique environment for the synthesis, folding, and modification of secretory and cell surface proteins. To ensure the success of these processes in the ER, a quality control mechanism exists to select proteins which have been improperly folded or modified. Accumulation of misfolded proteins causes ER stress and leads to activation of a complex signal transduction cascade known as the unfolded protein response (UPR) (reviewed in references 10 and 16). ER stress and the UPR are induced by physiological conditions known to cause protein misfolding, such as altered metabolic conditions (e.g., glucose deprivation) (19), expression of mutant proteins (e.g., influenza virus hemagglutinin) (17), and infection by viruses (e.g., hepatitis C virus) (36). The UPR can also be induced using drugs such as tunicamycin, which inhibits N-linked glycosylation in the ER, and thapsigargin, which disrupts calcium homeostasis in the ER.
Activation of the UPR is designed to eliminate misfolded proteins in the ER in two ways: by upregulating the expression of chaperone proteins and degradation factors to refold or eliminate misfolded proteins and by attenuating translation to reduce incoming protein traffic in the ER (reviewed in reference 32). Genes whose products are involved in metabolism and resistance to oxidative stress are also upregulated to aid the cell in recovering from ER stress. However, under severe conditions, when the cell is unable to recover from ER stress, apoptosis occurs.
Three ER-resident transmembrane proteins have been identified as sensors of ER stress: PKR-like ER kinase (PERK), activating transcription factor 6 (ATF6), and inositol-requiring enzyme 1 (IRE-1) (Fig. 1). Under normal conditions, the ER chaperone immunoglobulin heavy-chain-binding protein (BiP), also referred to as glucose regulated protein-78 (GRP78), is bound to the ER luminal domain of each sensor. However, when misfolded proteins accumulate in the ER, BiP is sequestered away from these sensors to bind to misfolded proteins (1). BiP release from PERK and IRE-1 leads to homodimerization of their luminal domains, causing autophosphorylation and activation, while BiP release from ATF6 unmasks a Golgi localization signal allowing relocation of ATF6 to the Golgi, where it is cleaved and activated (34).
Activation of PERK leads to attenuation of translation. Once activated by ER stress, PERK phosphorylates the eukaryotic initiation factor eIF2 (11, 12). In its phosphorylated form, eIF2 inhibits protein translation by tightly binding to another initiation factor, eIF2B, preventing it from catalyzing GDP-GTP exchange (18). Interestingly, phosphorylation of eIF2 results in increased translation of activating transcription factor 4 (ATF4) (11). ATF4 transcriptionally activates a number of genes encoding metabolism and redox regulatory factors thought to be involved in recovery from ER stress (13). These include the gene encoding GADD34 (26), which interacts with protein phosphatase 1 to dephosphorylate eIF2, thereby acting as part of a negative feedback loop to regulate eIF2 phosphorylation (4, 26).
Activation of ATF6 by ER stress causes transcriptional upregulation of ER chaperone proteins (40). ER stress causes the inactive 90-kDa ATF6 precursor to relocalize to the Golgi, where it is cleaved by the site-1 and site-2 proteases into the active 50-kDa protein (41) which translocates to the nucleus and activates the promoters of ER chaperone genes (e.g., BiP) (43). These newly synthesized chaperones refold misfolded proteins in the ER in an effort to relieve ER stress. ATF6 has also been shown to increase the amount of Xbp-1 mRNA (discussed below), providing a link between the ATF6 and IRE-1 pathways (21, 42).
Activation of the IRE-1 pathway upregulates the transcription of genes involved in protein degradation via X box binding protein 1 (XBP-1). In response to ER stress, IRE-1, which has intrinsic endonuclease activity, removes a 26-nucleotide intron from the Xbp-1 transcript (21). This spliced form of Xbp-1 mRNA encodes an active transcription factor which induces expression of a subset of genes encoding chaperones as well as genes encoding protein degradation enzymes (e.g., EDEM) (20).
Several viruses have been shown to induce ER stress and activate UPR signaling; these include three members of the flavivirus family, hepatitis C virus, Japanese encephalitis virus, and bovine viral diarrhea virus (15, 35, 36). Flaviviruses utilize the ER as the primary site of envelope glycoprotein biogenesis, genomic replication, and particle assembly (reviewed in reference 24); thus it is perhaps not surprising that infected cells experience ER stress. In addition, induction of ER stress has been associated with the pathology of other viruses. For example, respiratory syncytial virus induces apoptosis in infected cells through ER stress-mediated activation of caspase 12 (2), and the ability of the mouse retrovirus FrCasE to cause spongiform neurodegeneration is thought to depend upon activation of the UPR (9).
Viruses which induce ER stress are faced with the consequences of UPR activation, some of which can be detrimental to viral replication. As described above, activation of the PERK pathway leads to phosphorylation of eIF2, which results in translation attenuation. Thus, a virus that activates the PERK pathway must circumvent the effects of phospho-eIF2 in order to express viral and cellular gene products necessary for viral replication. However, other consequences of activating PERK (e.g., the expression of ATF4) may be beneficial for the viral infection, since ATF4 transcriptionally activates genes involved in reestablishing cellular metabolism (as part of the UPR recovery effort) and translation. Likewise, the expression of chaperones as a result of ATF6 activation may benefit virus replication by ensuring the proper folding of viral proteins. Activation of XBP-1, however, leads to expression of proteins involved in ER stress-associated degradation and may hamper viral replication by degrading necessary viral proteins. Hence, it may be advantageous for the virus to modulate the UPR, inhibiting the effects that would be detrimental to the infection, while maintaining those that may be beneficial. Here, we demonstrate that HCMV induces the UPR but regulates the three branches of the UPR signaling cascade to benefit the viral infection.
MATERIALS AND METHODS
Cells and viruses. Life-extended human foreskin fibroblasts (HFFs) (3) were propagated and maintained at 37°C in 5% CO2 in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum, Glutamax, and antibiotics (all reagents from Gibco, Carlsbad, CA). Cells were infected with the Towne strain of HCMV at a multiplicity of infection of 3. In experiments where thapsigargin was used, cells were treated with 2 μM thapsigargin for the indicated amount of time.
Metabolic labeling. Cells were washed with Dulbecco's modified Eagle's medium which lacked methionine and cysteine (methionine and cysteine-free medium; Gibco, Carlsbad, CA). After washing, cells were incubated in this medium for 15 min. Cells were then pulse-labeled for 30 min by the addition of methionine- and cysteine-free medium containing [35S]methionine and [35S]cysteine at 125 μCi/ml. At the time of harvest, radioactive medium was removed and cells were washed twice with cold phosphate-buffered saline (PBS), and proteins were extracted using radioimmunoprecipitation assay (RIPA) buffer. Equal volumes of lysate were analyzed by sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (11% polyacrylamide gels). Gels were dried and exposed to autoradiography film. In some experiments, cells were treated with 2 μM thapsigargin for 1 h prior to labeling; thapsigargin was included in the methionine- and cysteine-free medium used for washing and labeling. Incorporation of [35S]methionine and [35S]-cysteine into cells was quantitated using standard trichloroacetic acid precipitation methods.
Western analysis. Cell lysates were prepared in RIPA buffer. Proteins were quantitated using the Bio-Rad protein assay (Bio-Rad, Hercules, CA). Equal amounts of protein were separated by SDS polyacrylamide gel electrophoresis (11% polyacrylamide gels). Proteins were transferred to nitrocellulose membranes which were blocked by incubation for 30 min at room temperature in 5% nonfat milk diluted in PBS containing 0.5% Tween 20 (PBS-T). The following phospho-specific antibodies were diluted in PBS-T containing 5% bovine serum albumin (BSA): phospho-PERK (Cell Signaling Technologies) and phospho-eIF2 (Biosource, Carmillo, CA). The following antibodies were diluted in PBS-T containing 5% nonfat milk: total eIF2, ATF4/CREB2, BiP/GRP78, and GRP94 (all from Santa Cruz Biotechnologies, Santa Cruz, CA), ATF6 (generously provided by K. Mori, Kyoto University), and actin (Chemicon, Temecula, CA).
RT-PCR analysis. Total RNA was extracted using RNA STAT-60 (Tel Test, Friendswood, TX) according to the manufacturer's protocol. cDNA was reverse transcribed from 1.0 mg of total RNA with random hexamer primers using MultiScribe reverse transcriptase (Applied Biosystems, Foster City, CA) as recommended by the supplier. Quantitative real-time PCR was performed on equal volumes of cDNA reverse transcription product using TaqMan chemistry as recommended by the manufacturer (Applied Biosystems) on the ABI 7000 system. Fold change was calculated using the CT method of relative quantitation with 18S rRNA as the endogenous control for normalization, as described in the Applied Biosystems User Bulletin. Primer and probe sets used for 18S, GADD34, ATF4, Xbp-1, and EDEM are available from Applied Biosystems as Assays on Demand. For experiments in which splicing of Xbp-1 was analyzed, RNA was extracted as described above and spliced and unspliced Xbp-1 were amplified using reverse transcription (RT)-PCR as described previously (33).
RESULTS
HCMV induces phosphorylation of PERK and eIF2. To determine whether the PERK pathway is activated by HCMV infection, HFFs were infected with HCMV and harvested at the indicated days postinfection (Fig. 2). Cell lysates were analyzed by Western analysis using antibodies which recognize the phosphorylated forms of PERK and eIF2. Phosphorylated PERK (P-PERK, Fig. 2) was first detected at 3 days postinfection and accumulated significantly throughout the remainder of the 5-day time course. (Note that the phosphorylated form of PERK runs slightly faster than a nonspecific protein present prior to 3 days postinfection.) Correspondingly, phosphorylated eIF2 (P-eIF2, upper band of doublet in Fig. 2) was also observed by 3 days postinfection. However, in striking contrast to PERK phosphorylation, eIF2 phosphorylation peaked at 3.5 days postinfection and decreased thereafter. These results indicate that PERK is phosphorylated and active in HCMV-infected cells and that it phosphorylates eIF2. However, the discrepancy between the increasingly higher levels of phosphorylated PERK and the relatively low and decreasing levels of phosphorylated eIF2 suggests that the viral infection limits the amount of phosphorylated eIF2.
In addition, using an antibody that recognizes both phosphorylated and unphosphorylated eIF2, we determined that, compared to mock-infected cells, total eIF2 increased significantly by 2 days postinfection and remained elevated throughout the infection time course (Fig. 2). Control samples (on the right) demonstrate that 1 h treatment with thapsigargin results in phosphorylation of PERK and eIF2 but does not affect the total level of eIF2. These data suggest that the ratio of total eIF2 to phosphorylated eIF2 is high in infected cells due to the induction of eIF2. The induction of eIF2, together with its limited phosphorylation, suggests high levels of unphosphorylated eIF2 are maintained during HCMV infection, potentially resulting in an environment favoring continued translation rather than translational attenuation.
HCMV does not cause translation attenuation despite phosphorylation of eIF2. Using pulse-labeling experiments with 35S-labeled methionine and cysteine (see Materials and Methods) we determined whether the presence of phosphorylated eIF2 in HCMV-infected cells affected global translation. In Fig. 3 the total pulse-labeled proteins are shown separated by SDS-polyacrylamide gel electrophoresis. In addition, the labeling was further quantified by trichloroacetic acid precipitation; the incorporated counts in each sample are compared to the incorporation in the mock-infected sample, which is set at 100%. As a positive control for UPR induction we analyzed 35S-labeled methionine and cysteine incorporation in cells treated with thapsigargin (Thap) for 1 h. Consistent with previous reports, induction of the UPR by thapsigargin treatment resulted in translation attenuation as indicated by a 50% reduction in 35S-labeled methionine and cysteine incorporation into total protein (Fig. 3).
We then compared 35S-labeled methionine and cysteine incorporation in mock-infected cells with that in HCMV-infected cells at a time when PERK activation and eIF2 phosphorylation was not detected (6 h postinfection) and at a time when they were detected (4 days postinfection). At both infection time points total 35S-labeled methionine and cysteine incorporation was similar to the mock-treated samples (95% and 85%, respectively). It is clear from the autoradiogram that by 4 days postinfection the synthesis of specific cellular proteins is up- and down-regulated by the infection and viral proteins are accumulating; however, global protein synthesis does not appear to be inhibited.
HCMV increases expression of ATF4, whose translation is dependent on phosphorylated eIF2. We next asked whether the limited amount of phospho-eIF2 in HCMV-infected cells was sufficient to induce translation of ATF4. Western analysis (Fig. 4A) demonstrated that ATF4 protein levels were induced by 3 days postinfection, increased by 3.5 and 4 days postinfection, and remained high throughout the course of infection. The time of induction of ATF4 at 3 days postinfection coincides with the time at which eIF2 phosphorylation was first detected (Fig. 2).
In response to ER stress, translation of ATF4 mRNA increases through a phospho-eIF2-dependent mechanism by which inefficient ribosome scanning of the ATF4 mRNA allows translation to initiate at the bona fide AUG for ATF4 rather than at upstream AUGs present in the 5' end of the ATF4 mRNA (25). Consistent with this mechanism, 12 h treatment with thapsigargin resulted in a significant increase in ATF4 protein levels relative to untreated cells (Fig. 4A), but only a small (approximately 2-fold) increase in the level of ATF4 mRNA (Fig. 4B) in treated cells relative to untreated cells at the same time point.
To confirm that the increase in ATF4 protein levels observed during HCMV infection results from increased ATF4 translation, we measured the level of ATF4 mRNA over a time course of HCMV infection using quantitative RT-PCR. It is important to note that the fold change in ATF4 mRNA levels shown in Fig. 4 is calculated by comparing RNA levels in infected cells to the levels in mock-infected controls taken at each time point (infected/mock, where the mock is set at 1 for each time point). At no time during HCMV infection was there a greater than 1.5-fold increase in the level of ATF4 mRNA (Fig. 4B), indicating that induction of ATF4 expression by HCMV resulted from increased translation as previously proposed (25). Thus, phospho-eIF2-dependent translation of ATF4, as well as global translation, is maintained in HCMV infected cells.
HCMV induces the expression and ER stress-associated alteration in glycosylation of ATF6 but prevents ATF6 cleavage and activation. To determine whether HCMV activates the ATF6 pathway of the UPR, we assessed the status of ATF6 protein in HCMV-infected cells. The 90-kDa precursor of ATF6 exists in the ER as a glycosylated protein. In response to ER stress, there is an alteration in the glycosylation which is detectable by increased electrophoretic mobility (14). After translocation to the Golgi, the ATF6 precursor is cleaved into a 50-kDa product which functions as a transcription factor for many genes encoding ER chaperones (e.g., BiP and GRP94).
Using an antibody which recognizes the N terminus of ATF6 (common to both the 90-kDa and 50-kDa proteins; Fig. 5A), we determined the status of ATF6 over a 5-day time course of HCMV infection. As previously reported, thapsigargin treatment for up to 2 h resulted in a migration shift due to altered glycosylation (ATF6 to ATF6, Fig. 5A) and cleavage of the 90-kDa precursor into the 50-kDa "N"-terminal product ATF6(N). In HCMV-infected cells, we observed a significant induction in the level of ATF6 precursor by 1 day postinfection as well as a shift in the mobility of the precursor (from ATF6 to ATF6) between 3.5 and 4 days postinfection, indicative of an ER stress-induced change in the ATF6 glycosylation state. However, the 50-kDa cleavage product was not detected at any time during HCMV infection, suggesting that HCMV infection inhibits the cleavage of ATF6.
Since cleavage of ATF6 is necessary to produce the transcriptionally active form of ATF6, we predicted that ATF6 target genes would not be expressed in HCMV-infected cells. To test this, we analyzed the level of BiP and GRP94 by Western analysis (Fig. 5B). HCMV infection caused a transient increase in BiP at 1 to 2 days postinfection, times prior to the induction of the UPR; this may be an ATF6-independent induction mediated by immediate early or early viral proteins. However, intracellular BiP returned to basal levels by 3 days postinfection and remained low thereafter, in agreement with the lack of active ATF6.
The level of GRP94 gradually increased over the 5-day infection time course (Fig. 5B) and eventually attained a level equivalent to that seen in thapsigargin-treated cells (compare CMV 5d to 48 h Thap, Fig. 5B). Since the induction of GRP94 occurred by 1 day postinfection (prior to induction of the UPR by HCMV at 3 days postinfection), we conclude that an early viral function, not involving ATF6, accounts for the gradual increase in GRP94, a chaperone that may be useful during infection.
We also determined whether we could detect ATF6-dependent transcriptional activation of the gene encoding Xbp-1 by measuring total Xbp-1 mRNA in HCMV-infected and thapsigargin-treated cells using quantitative RT-PCR (Fig. 5C). As expected, a 2-h treatment with thapsigargin increased the amount of Xbp-1 mRNA to approximately 5-fold that detected in untreated cells; the mRNA remained elevated at 12 h posttreatment. The level of Xbp-1 mRNA in HCMV-infected cells relative to mock-infected cells at each time point showed an average induction of only 2-fold (Fig. 5C) over the course of the infection, significantly below that attainable with thapsigargin treatment. This low-level induction again agrees with the lack of active ATF6.
HCMV induces splicing of Xbp-1 mRNA. In response to ER stress, IRE-1 removes a 26-bp intron from the full-length Xbp-1 mRNA [Xbp-1(u)] which results in the spliced form of Xbp-1 [Xbp-1(s)]. Xbp-1(s) encodes XBP-1, a transcription factor which activates the expression of genes encoding chaperones as well as genes encoding proteins involved in UPR-mediated protein degradation, such as EDEM (ER degradation-enhancing -mannosidase-like protein). To determine whether the IRE-1 pathway is activated in response to HCMV-infection, we evaluated the splicing of Xbp-1 mRNA by RT-PCR using primers which amplify both Xbp-1(u) and Xbp-1(s) (Fig. 6A; upper and lower bands, respectively). In mock-infected cells, only the unspliced form of Xbp-1 was detected over the 5-day time course. However, in HCMV-infected cells, in addition to the unspliced form which was detected at all times, the spliced form of Xbp-1 was detectable as a faint band by 2 days postinfection and a more intense band by 3.5-days postinfection and thereafter. As expected, splicing of Xbp-1 was also observed in cells treated with thapsigargin for 30 min or 2 h. These results indicate that HCMV infection activates the IRE-1 pathway leading to the splicing of Xbp-1 RNA.
Since the spliced form of Xbp-1 encodes a transcription factor, we asked whether XBP-1 target genes were transcriptionally activated in HCMV-infected cells. Unlike transcription of BiP and GRP94, which can be activated by either ATF6 or XBP-1, transcription of EDEM is entirely dependent on XBP-1 (20). Thus, the level of EDEM mRNA in HCMV-infected cells was measured using real-time RT-PCR (Fig. 6B). In contrast to treatment with thapsigargin, which resulted in a 4- to 5-fold increase in the level of EDEM mRNA by 12 h posttreatment, no significant increase in the level of EDEM mRNA was detected at any time during the 5-day HCMV time course. Thus, the transcriptional activity of XBP-1, as indicated by the level of EDEM mRNA, is inhibited in HCMV infected cells.
DISCUSSION
Viral infection triggers cellular stress responses, potentially resulting in circumstances which could attenuate viral replication. In the case of ER stress and induction of the UPR, protein synthesis is attenuated and under severe ER stress conditions the cell undergoes apoptosis; neither condition would be beneficial to the viral infection. However, other aspects of the UPR may be beneficial to the viral infection. Thus, the question is whether the virus can limit harmful aspects of the UPR while maintaining beneficial ones. In the present work we provide evidence that HCMV infection induces the UPR but regulates the three signal transduction pathways of the UPR (PERK, ATF6, and IRE-1) in order to eliminate deleterious effects while maintaining potentially beneficial ones (Fig. 7).
HCMV infection activates the UPR between 2 and 3 days postinfection. HCMV encodes over 200 proteins, many of which are abundantly expressed and undergo glycosylation and other modifications in the ER. Under the experimental conditions used here (HFF cells, HCMV Towne, and multiplicity of infection of 3), phosphorylation of PERK and eIF2 was first detected at 3 days postinfection and Xbp-1 splicing was detected at 2 days postinfection. Thus, UPR activation by HCMV likely occurs at some time between 2 and 3 days postinfection. This is a time when there is abundant synthesis of late viral glycoproteins (29). It is possible that improper folding or aggregation of these proteins triggers the UPR. Indeed, a similar phenomenon has been demonstrated with the hepatitis C virus envelope protein E2, whose expression has been shown to activate transcription of ER stress-associated chaperone genes (22).
It is also possible that immunomodulatory proteins encoded by HCMV (US2, US3, US6, and US11) could induce the UPR through their disruption of major histocompatibility complex class I antigen presentation. In particular, the US2 and US11 glycoproteins, which are expressed with late kinetics, are located in the ER membrane of infected cells and redirect nascent major histocompatibility complex class I proteins from the ER into the cytosol (reviewed in reference 28). This activity may disrupt other transport mechanisms in the ER leading to ER stress and induction of the UPR. Furthermore, a recent report (23) indicates that HCMV US11 directly interacts with ER proteins which function in the export of proteins from the ER for degradation (e.g., Derlin-1) and protein refolding (i.e., BiP), suggesting US11 itself may modify ER function.
Yet another possible ER stress stimulus is HCMV virion envelopment, a process which utilizes the host secretory network and may perturb the ER (reviewed in references 27). Analysis of HCMV-infected cells by electron microscopy demonstrates large numbers of ER membrane-derived vesicles (our unpublished data); thus, virion envelopment may induce ER stress and the UPR. Consistent with this idea, flaviviruses are thought to induce ER stress as a result of particle assembly and virus maturation at ER membranes (31).
Precise control of BiP levels may allow the viral infection to control when the UPR is activated. Although HCMV clearly activates the UPR, the temporal nature of the induction may be specifically controlled. Our data suggest that HCMV infection causes a transient increase in BiP levels at 1 to 2 days postinfection, but by 3 days postinfection, BiP returns to basal levels. It is interesting to consider that the increased BiP present at 1 and 2 days could significantly inhibit the activation of the UPR by maintaining PERK, ATF6m and IRE-1 in their inactive forms.
The rapid disappearance of BiP between 2 and 3 days postinfection coincides with the appearance of UPR activation in infected cells. Thus, a precise, virally mediated control of BiP levels could determine when the UPR is activated. Interestingly, it has recently been shown that BiP is included in the HCMV virion (39); this may account in part for the elevated BiP levels at 1 and 2 days; however, it is equally likely that a viral immediate-early or early protein activates the BiP expression seen at 1 and 2 days postinfection. Regardless, the rapid decrease in BiP between 2 and 3 days postinfection may be virally mediated in order to allow UPR activation during the late phase of the infection. In addition, a recent report demonstrates a physical interaction between BiP and the HCMV glycoprotein US11 (23), suggesting a mechanism by which HCMV may target BiP for inactivation or degradation.
HCMV maintains translation despite UPR activation. Our data show that HCMV infection significantly activates PERK during the viral infection; however, the expected phosphorylation of eIF2 is significantly limited and global translation is not inhibited in infected cells. Consistent with our results, it has previously been reported that HCMV prevents the phosphorylation of eIF2, and it has been suggested that the immediate early proteins encoded by TRS1 and the closely related IRS1 genes are involved (5). The disproportionate levels of phosphorylated PERK compared to phosphorylated eIF2 has also been observed during hepatitis C virus infection where the hepatitis C virus E2 protein binds to and inhibits phosphorylated PERK, thus inhibiting the phosphorylation of eIF2 and relieving the global translation inhibition normally induced by PERK activation (30).
HCMV also causes a significant increase in the amount of total eIF2, a condition which would favor an increased level of unphosphorylated eIF2, especially under the conditions of limited phosphorylation as described above. Together, all of the data support a model in which HCMV limits the amount of phosphorylated eIF2 in order to maintain global translation.
Translation dependent on phosphorylated eIF2 is maintained during HCMV infection. Despite the limited phosphorylation of eIF2 in HCMV-infected cells, we detected a significant increase in the protein level of ATF4, whose mRNAs depend on phosphorylated eIF2 for translation (11). Expression of ATF4 leads to transcriptional activation of target genes involved in metabolism and redox maintenance, and thus may benefit the course of HCMV infection by maintaining a permissive cellular environment. In addition, ATF4 has been shown to transcriptionally activate GADD34 (26), which interacts with protein phosphatase 1 to dephosphorylate eIF2 (4). Induction of GADD34 may benefit the infection through dephosphorylation of eIF2 and may be one means by which the amount of phosphorylated eIF2 is limited. In this regard, HSV-1 encodes a GADD34 homolog, ICP34.5, suggesting that GADD34 function may play an important role in herpesvirus biology (6).
Our data suggest that during an HCMV infection, both global translation, dependent on unphosphorylated eIF2, and phospho-eIF2-dependent translation can be maintained simultaneously. However, the ability to support both global and phospho-eIF2-dependent translation is not supported by the present model for phospho-eIF2 inhibition of translation. Phosphorylated eIF2 has a much greater binding affinity for eIF2B than unphosphorylated eIF2, and the stable complex of phospho-eIF2 and eIF2B effectively eliminates eIF2B's guanine nucleotide exchange function. It has been proposed that the cell's limited amount of eIF2B can all be bound by only a small amount of phospho-eIF2 (18), inhibiting global translation and allowing phospho-eIF2-dependent translation.
By this mechanism, the appearance of phospho-eIF2 in infected cells suggests that there should be inhibition of global translation, yet this is not what we observed. However, it has been suggested that translation of ATF4 occurs at levels of eIF2 that have only minimal inhibitory effects on global translation (7, 8, 38). Thus, HCMV's ability to increase total levels of eIF2 while limiting its phosphorylation may ensure that both global translation and phospho-eIF2a-dependent translation are maintained. What is clear is that HCMV has evolved a means to manipulate translation to its advantage, eliminating the negative effects of PERK activation (translation attenuation) while maintaining the positive effects that result from induction of ATF4.
HCMV inhibits the activation of ATF6. The transcriptional activation function of ATF6 is critical to the UPR, since ATF6 increases the expression of many ER chaperone genes whose products are necessary for protein refolding. In addition, ATF6 increases the level of Xbp-1 mRNA, which, once spliced by IRE-1, encodes a transcriptional activator of genes whose products are involved in protein degradation (discussed further below). In HCMV-infected cells, the levels of the ATF6 precursor protein increased significantly by 1 days postinfection, and between 3.5 and 4 days postinfection the ATF6 precursor displayed a mobility shift consistent with altered glycosylation that occurs when there is ER stress (14). However, in HCMV infected cells the ATF6 precursor is not cleaved to the active form.
Inhibition of ATF6 activation by HCMV was somewhat surprising, as one might expect that its transcriptional activation of chaperone encoding genes might benefit the virus by aiding in protein folding and preventing protein aggregation. However, our data suggest that the virus can increase the expression of ATF6 target genes in a gene specific manner independent of ATF6 activation. The transient increases in BiP levels detected at 1 to 2 days postinfection are discussed above and are likely independent of ATF6. Although GRP94 protein levels increased during the infection, the induction occurred early in the infection (1 day postinfection) and GRP94 gradually increased over the course of the 5-day infection. This early induction (prior to the detection of UPR activation) and the gradual increase of the protein over time, is not consistent with UPR activation but is consistent with an immediate early or early viral transcriptional activation mechanism which does not involve ATF6. A third ATF6 target gene, Xbp-1, was activated by no more than 2-fold during the infection, in agreement with there being no active ATF6. As described below, the inhibition of transcription of the Xbp-1 gene may be beneficial to the infection.
HCMV inhibits XBP-1-mediated transactivation. The transcriptional activation function of XBP-1 is also critical to the UPR since it stimulates expression of proteins involved in ER stress-induced protein degradation, such as EDEM, as well as chaperones also controlled by ATF6 (20). Our data show that splicing of Xbp-1 mRNA is stimulated by HCMV infection, consistent with the activation of the IRE-1 pathway. However, transcriptional activation of the XBP-1 target gene EDEM is not detected in HCMV-infected cells. This suggests that either the levels of spliced Xbp-1 mRNA in HCMV-infected cells are too low to produce sufficient XBP-1 protein or the transcriptional activity of XBP-1 is inhibited by HCMV. At this point either or both possibilities could contribute to HCMV's inhibition of XBP-1 transcriptional activation. In cells containing hepatitis C virus replicons, the transcriptional activity of XBP-1 is inhibited despite the presence of significant levels of spliced XBP-1 (37). In any event, the resulting inhibition of expression of genes such as EDEM may be beneficial to viral infection, since this would inhibit ER stress-induced protein degradation, which may otherwise hinder the accumulation of necessary viral proteins.
In sum, our data show that during an HCMV infection the UPR is induced but is modified in ways which benefit the viral infection, inhibiting those aspects which may be deleterious to the progress of the viral infection and maintaining those which may be beneficial.
ACKNOWLEDGMENTS
We thank Kazutoshi Mori for providing antisera to ATF6; the members of the Alwine laboratory for helpful discussions and critical evaluation of the data; and Sherri Adams for critical reading of the manuscript.
J.A.I. is supported by an NIH postdoctoral training grant (F32 AI062055-01). A.H.S. is supported by a Howard Hughes Medical Institute predoctoral fellowship. This work was supported by NIH grant CA28379-24 awarded to J.C.A. by the National Cancer Institute and by a Tobacco Settlement Grant awarded to J.C.A. by the Abramson Cancer Center at the University of Pennsylvania.
REFERENCES
Bertolotti, A., Y. Zhang, L. M. Hendershot, H. P. Harding, and D. Ron. 2000. Dynamic interaction of BiP and ER stress transducers in the unfolded-protein response. Nat. Cell Biol. 2:326-332.
Bitko, V., and S. Barik. 2001. An endoplasmic reticulum-specific stress-activated caspase (caspase 12) is implicated in the apoptosis of A549 epithelial cells by Respiratory Syncytial Virus. J. Cell Biochem. 80:441-454.
Bresnahan, W. A., G. E. Hultman, and T. Shenk. 2000. Replication of wild type and mutant human cytomegalovirus in life-extended human diploid fibroblasts. J. Virol. 74:10816-10818.
Brush, M. H., D. C. Weiser, and S. Shenolikar. 2003. Growth arrest and DNA damage-inducible protein GADD34 targets protein phosphatase 1 to the endoplasmic reticulum and promotes dephosphorylation of the subunit of eukaryotic translation initiation factor 2. Mol. Cell. Biol. 23:1292-1303.
Child, S. J., M. Hakki, K. L. De Niro, and A. P. Geballe. 2004. Evasion of cellular antiviral responses by human cytomegalovirus TRS1 and IRS1. J. Virol. 78:197-205.
Chou, J., and B. Roizman. 1994. Herpes simplex virus 1 134.5 gene function, which blocks the host response to infection, maps in the homologous domain of the genes expressed during growth arrest and DNA damage. Proc. Natl. Acad. Sci. USA 91:5247-5251.
Deng, J., P. D. Lu, Y. Zhang, D. Sheuner, R. J. Kaufman, N. Sonenberg, H. P. Harding, and D. Ron. 2004. Translational repression mediates activation of nuclear factor kappa B by phosphorylated translation initiation factor 2. J. Virol. 24:10161-10168.
Dever, T. E., L. Feng, R. C. Wek, A. M. Cigan, T. F. Donahue, and A. G. Hinnebusch. 1992. Phosphorylation of initiation factor 2 alpha by protein kinase GCN2 mediates gene-specific translational control of GCN4 in yeast. Cell 68:585-596.
Dimcheff, D. E., S. Askovic, A. H. Baker, C. Johnson-Fowler, and J. L. Portis. 2003. Endoplasmic reticulum stress is a determinant of retrovirus-induced spongiform neurodegeneration. J. Virol. 77:12617-12629.
Harding, H. P., M. Calfon, F. Urano, I. Novoa, and D. Ron. 2002. Transcriptional and translational control in the mammalian unfolded protein response. Annu. Rev. Cell Dev. Biol. 18:575-599.
Harding, H. P., Y. Zhang, A. Bertolotti, H. Zeng, and D. Ron. 2000. PERK is essential for translation regulation and cell survival during unfolded protein response. Mol. Cell. 5:897-904.
Harding, H. P., Y. Zhang, and D. Ron. 1999. Protein Translation and folding are coupled by an endoplasmic-reticulum-resident kinase. Nature 397:271-274.
Harding, H. P., Y. Zhang, H. Zeng, I. Novoa, P. D. Lu, M. Calfon, N. Sadri, C. Yun, B. Popko, R. Paules, D. F. Stojdl, J. C. Bell, T. Hettmann, J. M. Leiden, and D. Ron. 2003. An integrated stress response regulates amino acid metabolism and resistance to oxidative stress. Mol. Cell. 11:619-633.
Hong, M., S. Luo, P. Baumeister, J. Huang, R. K. Gogia, M. Li, and A. S. Lee. 2004. Underglycosylation of ATF6 as a novel sensing mechanism for activation of the unfolded protein response. J. Biol. Chem. 279:11354-11363.
Jordan, R., L. Wang, T. M. Graczyk, T. M. Block, and P. R. Romano. 2002. Replication of a cytopathic strain of Bovine Viral Diarrhea Virus Activates PERK and induces endoplasmic reticulum stress-mediated apoptosis of MDBK cells. J. Virol. 76:9588-9599.
Kaufman, R. J. 1999. Stress signaling from the lumen of the endoplasmic reticulum: coordination of gene transcription and translational controls. Gen. Dev. 13:1211-1233.
Kozutsumi, Y., M. Segal, K. Normington, M. J. Gething, and J. Sambrook. 1988. The presence of malfolded proteins in the endoplasmic reticulum signals the induction of glucose-regulated proteins. Nature 332:462-464.
Krishnamoorthy, T., G. D. Pavitt, F. Zhang, T. E. Dever, and A. G. Hinnebusch. 2001. Tight binding of the phosphorylated subunit of initiation factor 2 (eIF2) to the regulatory subunits of guanine nucleotide exchange factor eIF2B is required for inhibition of translation initiation. Mol. Cell. Biol. 21:5018-5030.
Lee, A. S. 1992. Mammalian stress response: induction of the glucose-regulated protein family. Cur. Opin. Cell Biol. 4:267-273.
Lee, A. S., N. N. Iwakoshi, and L. H. Glimcher. 2003. XBP-1 regulates a subset of endoplasmic reticulum-resident chaperone genes in the unfolded protein response. Mol. Cell. Biol. 23:7448-7459.
Lee, K., W. Tirasophon, X. Shen, M. Michalak, R. Prywes, T. Okada, H. Yoshida, K. Mori, and R. J. Kaufman. 2002. IRE-1-mediated unconventional mRNA splicing and S2P-mediated ATF6 cleavage merge to regulate XBP1 in signaling the unfolded protein response. Genes Dev. 16:452-466.
Liberman, E., Y. Fong, M. J. Selby, W. Choo, L. Cousense, M. Houghton, and T. S. B. Yen. 1999. Activation of the grp78 and grp94 promoters by Hepatitis C Virus E2 envelope protein. J. Virol. 73:3718-3722.
Lilley, B. N., and H. L. Ploegh. 2004. A membrane protein required for dislocation of misfolded proteins from the ER. Nature 249:834-840.
Lindenbach, B. D., and C. M. Rice. 2001. Flaviviridae: the viruses and their replication, p. 991-1041. In P. M. Howe, D. M. Knipe, D. E. Griffin, R. A. Lamb, M. A. Martin, B. Roizman, and S. E. Straus (ed.), Fields virology, 4th ed. Lippincott Williams & Wilkins, Philadelphia, Pa.
Lu, P. D., H. P. Harding, and D. Ron. 2004. Translation reinitiation at alternative open reading frames regulates gene expression in an integrated stress response. J. Cell Biol. 167:27-33.
Ma, Y., and L. M. Hendershot. 2003. Delineation of a negative feedback regulatory loop that controls protein translation during endoplasmic reticulum stress. J. Biol. Chem. 278:34864-34873.
Mettenleiter, T. C. 2002. Herpesvirus assembly and egress. J. Virol. 76:1537-1547.
Mocarski, E. S. 2002. Immunomodulation by cytomegaloviruses: manipulative strategies beyond invasion. Trends Microbiol. 10:332-339.
Mocarski, E. S., and C. T. Courcelle. 2001. Cytomegaloviruses and their replication, p. 2629-2673. In P. M. Howe and D. M. Knipe (ed.), Fields virology. Lippincott Williams & Wilkins, Philadelphia, Pa.
Pavio, N., P. R. Romano, T. M. Graczyk, S. M. Feinstone, and D. R. Taylor. 2003. Protein synthesis and endoplasmic reticulum stress can be modulated by the Hepatitis C Virus envelope protein E2 through the eukaryotic initiation factor 2 kinase PERK. J. Virol. 77:3578-3585.
Rice, C. M. 1996. Flaviviridae: the viruses and their replication, p. 931-959. In P. M. Howe and D. M. Knipe (ed.), Fields virology. Lippincott-Raven Publishers, Philadelphia, Pa.
Rutkowski, D. R., and R. J. Kaufman. 2004. A trip to the ER: coping with stress. Trends Cell Biol. 14:20-28.
Shang, J., and M. A. Lehrman. 2004. Discordance of UPR signaling by ATF6 and IRE1p-XBP1 with levels of target transcripts. Biochem. Biophys. Res. Commun. 317:390-396.
Shen, J., X. Chen, L. M. Hendershot, and R. Prywes. 2002. ER stress regulation of ATF6 localization by dissociation of BiP/GRP78 binding and unmasking of golgi localization signals. Dev. Cell. 3:99-111.
Su, H. L., C. L. Liao, and Y. L. Lin. 2002. Japanese Encephalitis Virus infection initiates endoplasmic reticulum stress and an unfolded protein response. J. Virol. 76:4162-4171.
Tardif, K. D., K. Mori, and A. Siddiqui. 2002. Hepatitis C virus subgenomic replicons induce endoplasmic reticulum stress activating an intracellular signaling pathway. J. Virol. 76:7453-7459.
Tardif, K. D., K. I. Mori, R. J. Kaufman, and A. Siddiqui. 2004. Hepatitis C Virus suppresses the IRE-1-Xbp1 pathway of the unfolded protein response. J. Biol. Chem. 279:17158-17164.
Tzamarias, D., I. Roussou, and G. Thireos. 1989. Coupling of GCN4 mRNA translational activation with decreased rates of polypeptide chain initiation. Cell 57:947-954.
Varnum, S. M., D. N. Streblow, M. E. Monroe, P. Smith, K. J. Auberry, L. Pasa-tolic, D. Wang, D. G. Camp, K. Rodland, S. Wiley, W. Britt, T. Shenk, R. D. Smith, and J. A. Nelson. 2004. Identification of proteins in human cytomegalovirus (HCMV) particles: the HCMV proteome. J. Virol. 78:10960-10966.
Wang, Y., J. Shen, N. Arenzanna, W. Tirasaphon, R. Kaufman, and R. Prywes. 2000. Activation of ATF6 and an ATF6 DNA binding site by the endoplasmic reticulum stress response. J. Biol. Chem. 275:27013-27020.
Ye, J., R. B. Rawson, R. Komuro, X. Chen, U. P. Dave, R. Prywes, M. S. Brown, and J. L. Goldstein. 2000. ER stress induces cleavage of membrane-bound ATF6 by the same proteases that process SREBPs. Mol. Cell. 6:1355-1364.
Yoshida, H., T. Matsui, A. Yamamoto, T. Okada, and, M. K. 2001. Xbp1 mRNA is induced by ATF6 and spliced by IRE-1 in response to ER stress to produce a highly active transcription factor. Cell 107:881-891.
Yoshida, H., T. Okada, K. Haze, H. Yanagi, T. Yura, M. Negishi, and K. Mori. 2001. Endoplasmic reticulum stress-induced formation of transcription factor complex ERSF including NF-Y (CBF) and activating transcription factors 6 and 6? that activates the mammalian unfolded protein response. Mol. Cell. Biol. 21:1239-1248.(Jennifer A. Isler, Alison)
Department of Pathology and Laboratory Medicine, School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania
ABSTRACT
Viral infection causes stress to the endoplasmic reticulum. The response to endoplasmic reticulum stress, known as the unfolded protein response (UPR), is designed to eliminate misfolded proteins and allow the cell to recover by attenuating translation and upregulating the expression of chaperones, degradation factors, and factors that regulate the cell's metabolic and redox environment. Some consequences of the UPR (e.g., expression of chaperones and regulation of the metabolism and redox environment) may be advantageous to the viral infection; however, translational attenuation would not. Thus, viruses may induce mechanisms which modulate the UPR, maintaining beneficial aspects and suppressing deleterious aspects. We demonstrate that human cytomegalovirus (HCMV) infection induces the UPR but specifically regulates the three branches of UPR signaling, PKR-like ER kinase (PERK), activating transcription factor 6 (ATF6), and inositol-requiring enzyme 1 (IRE-1), to favor viral replication. HCMV infection activated the eIF2 kinase PERK; however, the amount of phosphorylated eIF2 was limited and translation attenuation did not occur. Interestingly, translation of select mRNAs, which is dependent on eIF2 phosphorylation, did occur, including the transcription factor ATF4, which activates genes which may benefit the infection. The endoplasmic reticulum stress-induced activation of the transcription factor ATF6 was suppressed in HCMV-infected cells; however, specific chaperone genes, normally activated by ATF6, were activated by a virus-induced, ATF6-independent mechanism. Lastly, HCMV infection activated the IRE-1 pathway, as indicated by splicing of Xbp-1 mRNA. However, transcriptional activation of the XBP-1 target gene EDEM (ER degradation-enhancing -mannosidase-like protein, a protein degradation factor) was inhibited. These results suggest that, although HCMV infection induces the unfolded protein response, it modifies the outcome to benefit viral replication.
INTRODUCTION
Human cytomegalovirus (HCMV) is a betaherpesvirus which can cause significant medical problems in individuals with immature or compromised immune systems. The genome of HCMV is 230 kb of double-stranded DNA with the potential to encode over 200 proteins. Like that of other herpesviruses, HCMV viral gene expression occurs in an ordered temporal pattern having immediate-early, early, delayed-early, and late kinetics, with increasing viral protein synthesis over time. The burden of HCMV infection on the host cell has been shown to initiate a number of cellular stress responses. Here, we investigated the stress of HCMV infection on the endoplasmic reticulum (ER).
The ER is an extensive membranous network that provides a unique environment for the synthesis, folding, and modification of secretory and cell surface proteins. To ensure the success of these processes in the ER, a quality control mechanism exists to select proteins which have been improperly folded or modified. Accumulation of misfolded proteins causes ER stress and leads to activation of a complex signal transduction cascade known as the unfolded protein response (UPR) (reviewed in references 10 and 16). ER stress and the UPR are induced by physiological conditions known to cause protein misfolding, such as altered metabolic conditions (e.g., glucose deprivation) (19), expression of mutant proteins (e.g., influenza virus hemagglutinin) (17), and infection by viruses (e.g., hepatitis C virus) (36). The UPR can also be induced using drugs such as tunicamycin, which inhibits N-linked glycosylation in the ER, and thapsigargin, which disrupts calcium homeostasis in the ER.
Activation of the UPR is designed to eliminate misfolded proteins in the ER in two ways: by upregulating the expression of chaperone proteins and degradation factors to refold or eliminate misfolded proteins and by attenuating translation to reduce incoming protein traffic in the ER (reviewed in reference 32). Genes whose products are involved in metabolism and resistance to oxidative stress are also upregulated to aid the cell in recovering from ER stress. However, under severe conditions, when the cell is unable to recover from ER stress, apoptosis occurs.
Three ER-resident transmembrane proteins have been identified as sensors of ER stress: PKR-like ER kinase (PERK), activating transcription factor 6 (ATF6), and inositol-requiring enzyme 1 (IRE-1) (Fig. 1). Under normal conditions, the ER chaperone immunoglobulin heavy-chain-binding protein (BiP), also referred to as glucose regulated protein-78 (GRP78), is bound to the ER luminal domain of each sensor. However, when misfolded proteins accumulate in the ER, BiP is sequestered away from these sensors to bind to misfolded proteins (1). BiP release from PERK and IRE-1 leads to homodimerization of their luminal domains, causing autophosphorylation and activation, while BiP release from ATF6 unmasks a Golgi localization signal allowing relocation of ATF6 to the Golgi, where it is cleaved and activated (34).
Activation of PERK leads to attenuation of translation. Once activated by ER stress, PERK phosphorylates the eukaryotic initiation factor eIF2 (11, 12). In its phosphorylated form, eIF2 inhibits protein translation by tightly binding to another initiation factor, eIF2B, preventing it from catalyzing GDP-GTP exchange (18). Interestingly, phosphorylation of eIF2 results in increased translation of activating transcription factor 4 (ATF4) (11). ATF4 transcriptionally activates a number of genes encoding metabolism and redox regulatory factors thought to be involved in recovery from ER stress (13). These include the gene encoding GADD34 (26), which interacts with protein phosphatase 1 to dephosphorylate eIF2, thereby acting as part of a negative feedback loop to regulate eIF2 phosphorylation (4, 26).
Activation of ATF6 by ER stress causes transcriptional upregulation of ER chaperone proteins (40). ER stress causes the inactive 90-kDa ATF6 precursor to relocalize to the Golgi, where it is cleaved by the site-1 and site-2 proteases into the active 50-kDa protein (41) which translocates to the nucleus and activates the promoters of ER chaperone genes (e.g., BiP) (43). These newly synthesized chaperones refold misfolded proteins in the ER in an effort to relieve ER stress. ATF6 has also been shown to increase the amount of Xbp-1 mRNA (discussed below), providing a link between the ATF6 and IRE-1 pathways (21, 42).
Activation of the IRE-1 pathway upregulates the transcription of genes involved in protein degradation via X box binding protein 1 (XBP-1). In response to ER stress, IRE-1, which has intrinsic endonuclease activity, removes a 26-nucleotide intron from the Xbp-1 transcript (21). This spliced form of Xbp-1 mRNA encodes an active transcription factor which induces expression of a subset of genes encoding chaperones as well as genes encoding protein degradation enzymes (e.g., EDEM) (20).
Several viruses have been shown to induce ER stress and activate UPR signaling; these include three members of the flavivirus family, hepatitis C virus, Japanese encephalitis virus, and bovine viral diarrhea virus (15, 35, 36). Flaviviruses utilize the ER as the primary site of envelope glycoprotein biogenesis, genomic replication, and particle assembly (reviewed in reference 24); thus it is perhaps not surprising that infected cells experience ER stress. In addition, induction of ER stress has been associated with the pathology of other viruses. For example, respiratory syncytial virus induces apoptosis in infected cells through ER stress-mediated activation of caspase 12 (2), and the ability of the mouse retrovirus FrCasE to cause spongiform neurodegeneration is thought to depend upon activation of the UPR (9).
Viruses which induce ER stress are faced with the consequences of UPR activation, some of which can be detrimental to viral replication. As described above, activation of the PERK pathway leads to phosphorylation of eIF2, which results in translation attenuation. Thus, a virus that activates the PERK pathway must circumvent the effects of phospho-eIF2 in order to express viral and cellular gene products necessary for viral replication. However, other consequences of activating PERK (e.g., the expression of ATF4) may be beneficial for the viral infection, since ATF4 transcriptionally activates genes involved in reestablishing cellular metabolism (as part of the UPR recovery effort) and translation. Likewise, the expression of chaperones as a result of ATF6 activation may benefit virus replication by ensuring the proper folding of viral proteins. Activation of XBP-1, however, leads to expression of proteins involved in ER stress-associated degradation and may hamper viral replication by degrading necessary viral proteins. Hence, it may be advantageous for the virus to modulate the UPR, inhibiting the effects that would be detrimental to the infection, while maintaining those that may be beneficial. Here, we demonstrate that HCMV induces the UPR but regulates the three branches of the UPR signaling cascade to benefit the viral infection.
MATERIALS AND METHODS
Cells and viruses. Life-extended human foreskin fibroblasts (HFFs) (3) were propagated and maintained at 37°C in 5% CO2 in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum, Glutamax, and antibiotics (all reagents from Gibco, Carlsbad, CA). Cells were infected with the Towne strain of HCMV at a multiplicity of infection of 3. In experiments where thapsigargin was used, cells were treated with 2 μM thapsigargin for the indicated amount of time.
Metabolic labeling. Cells were washed with Dulbecco's modified Eagle's medium which lacked methionine and cysteine (methionine and cysteine-free medium; Gibco, Carlsbad, CA). After washing, cells were incubated in this medium for 15 min. Cells were then pulse-labeled for 30 min by the addition of methionine- and cysteine-free medium containing [35S]methionine and [35S]cysteine at 125 μCi/ml. At the time of harvest, radioactive medium was removed and cells were washed twice with cold phosphate-buffered saline (PBS), and proteins were extracted using radioimmunoprecipitation assay (RIPA) buffer. Equal volumes of lysate were analyzed by sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (11% polyacrylamide gels). Gels were dried and exposed to autoradiography film. In some experiments, cells were treated with 2 μM thapsigargin for 1 h prior to labeling; thapsigargin was included in the methionine- and cysteine-free medium used for washing and labeling. Incorporation of [35S]methionine and [35S]-cysteine into cells was quantitated using standard trichloroacetic acid precipitation methods.
Western analysis. Cell lysates were prepared in RIPA buffer. Proteins were quantitated using the Bio-Rad protein assay (Bio-Rad, Hercules, CA). Equal amounts of protein were separated by SDS polyacrylamide gel electrophoresis (11% polyacrylamide gels). Proteins were transferred to nitrocellulose membranes which were blocked by incubation for 30 min at room temperature in 5% nonfat milk diluted in PBS containing 0.5% Tween 20 (PBS-T). The following phospho-specific antibodies were diluted in PBS-T containing 5% bovine serum albumin (BSA): phospho-PERK (Cell Signaling Technologies) and phospho-eIF2 (Biosource, Carmillo, CA). The following antibodies were diluted in PBS-T containing 5% nonfat milk: total eIF2, ATF4/CREB2, BiP/GRP78, and GRP94 (all from Santa Cruz Biotechnologies, Santa Cruz, CA), ATF6 (generously provided by K. Mori, Kyoto University), and actin (Chemicon, Temecula, CA).
RT-PCR analysis. Total RNA was extracted using RNA STAT-60 (Tel Test, Friendswood, TX) according to the manufacturer's protocol. cDNA was reverse transcribed from 1.0 mg of total RNA with random hexamer primers using MultiScribe reverse transcriptase (Applied Biosystems, Foster City, CA) as recommended by the supplier. Quantitative real-time PCR was performed on equal volumes of cDNA reverse transcription product using TaqMan chemistry as recommended by the manufacturer (Applied Biosystems) on the ABI 7000 system. Fold change was calculated using the CT method of relative quantitation with 18S rRNA as the endogenous control for normalization, as described in the Applied Biosystems User Bulletin. Primer and probe sets used for 18S, GADD34, ATF4, Xbp-1, and EDEM are available from Applied Biosystems as Assays on Demand. For experiments in which splicing of Xbp-1 was analyzed, RNA was extracted as described above and spliced and unspliced Xbp-1 were amplified using reverse transcription (RT)-PCR as described previously (33).
RESULTS
HCMV induces phosphorylation of PERK and eIF2. To determine whether the PERK pathway is activated by HCMV infection, HFFs were infected with HCMV and harvested at the indicated days postinfection (Fig. 2). Cell lysates were analyzed by Western analysis using antibodies which recognize the phosphorylated forms of PERK and eIF2. Phosphorylated PERK (P-PERK, Fig. 2) was first detected at 3 days postinfection and accumulated significantly throughout the remainder of the 5-day time course. (Note that the phosphorylated form of PERK runs slightly faster than a nonspecific protein present prior to 3 days postinfection.) Correspondingly, phosphorylated eIF2 (P-eIF2, upper band of doublet in Fig. 2) was also observed by 3 days postinfection. However, in striking contrast to PERK phosphorylation, eIF2 phosphorylation peaked at 3.5 days postinfection and decreased thereafter. These results indicate that PERK is phosphorylated and active in HCMV-infected cells and that it phosphorylates eIF2. However, the discrepancy between the increasingly higher levels of phosphorylated PERK and the relatively low and decreasing levels of phosphorylated eIF2 suggests that the viral infection limits the amount of phosphorylated eIF2.
In addition, using an antibody that recognizes both phosphorylated and unphosphorylated eIF2, we determined that, compared to mock-infected cells, total eIF2 increased significantly by 2 days postinfection and remained elevated throughout the infection time course (Fig. 2). Control samples (on the right) demonstrate that 1 h treatment with thapsigargin results in phosphorylation of PERK and eIF2 but does not affect the total level of eIF2. These data suggest that the ratio of total eIF2 to phosphorylated eIF2 is high in infected cells due to the induction of eIF2. The induction of eIF2, together with its limited phosphorylation, suggests high levels of unphosphorylated eIF2 are maintained during HCMV infection, potentially resulting in an environment favoring continued translation rather than translational attenuation.
HCMV does not cause translation attenuation despite phosphorylation of eIF2. Using pulse-labeling experiments with 35S-labeled methionine and cysteine (see Materials and Methods) we determined whether the presence of phosphorylated eIF2 in HCMV-infected cells affected global translation. In Fig. 3 the total pulse-labeled proteins are shown separated by SDS-polyacrylamide gel electrophoresis. In addition, the labeling was further quantified by trichloroacetic acid precipitation; the incorporated counts in each sample are compared to the incorporation in the mock-infected sample, which is set at 100%. As a positive control for UPR induction we analyzed 35S-labeled methionine and cysteine incorporation in cells treated with thapsigargin (Thap) for 1 h. Consistent with previous reports, induction of the UPR by thapsigargin treatment resulted in translation attenuation as indicated by a 50% reduction in 35S-labeled methionine and cysteine incorporation into total protein (Fig. 3).
We then compared 35S-labeled methionine and cysteine incorporation in mock-infected cells with that in HCMV-infected cells at a time when PERK activation and eIF2 phosphorylation was not detected (6 h postinfection) and at a time when they were detected (4 days postinfection). At both infection time points total 35S-labeled methionine and cysteine incorporation was similar to the mock-treated samples (95% and 85%, respectively). It is clear from the autoradiogram that by 4 days postinfection the synthesis of specific cellular proteins is up- and down-regulated by the infection and viral proteins are accumulating; however, global protein synthesis does not appear to be inhibited.
HCMV increases expression of ATF4, whose translation is dependent on phosphorylated eIF2. We next asked whether the limited amount of phospho-eIF2 in HCMV-infected cells was sufficient to induce translation of ATF4. Western analysis (Fig. 4A) demonstrated that ATF4 protein levels were induced by 3 days postinfection, increased by 3.5 and 4 days postinfection, and remained high throughout the course of infection. The time of induction of ATF4 at 3 days postinfection coincides with the time at which eIF2 phosphorylation was first detected (Fig. 2).
In response to ER stress, translation of ATF4 mRNA increases through a phospho-eIF2-dependent mechanism by which inefficient ribosome scanning of the ATF4 mRNA allows translation to initiate at the bona fide AUG for ATF4 rather than at upstream AUGs present in the 5' end of the ATF4 mRNA (25). Consistent with this mechanism, 12 h treatment with thapsigargin resulted in a significant increase in ATF4 protein levels relative to untreated cells (Fig. 4A), but only a small (approximately 2-fold) increase in the level of ATF4 mRNA (Fig. 4B) in treated cells relative to untreated cells at the same time point.
To confirm that the increase in ATF4 protein levels observed during HCMV infection results from increased ATF4 translation, we measured the level of ATF4 mRNA over a time course of HCMV infection using quantitative RT-PCR. It is important to note that the fold change in ATF4 mRNA levels shown in Fig. 4 is calculated by comparing RNA levels in infected cells to the levels in mock-infected controls taken at each time point (infected/mock, where the mock is set at 1 for each time point). At no time during HCMV infection was there a greater than 1.5-fold increase in the level of ATF4 mRNA (Fig. 4B), indicating that induction of ATF4 expression by HCMV resulted from increased translation as previously proposed (25). Thus, phospho-eIF2-dependent translation of ATF4, as well as global translation, is maintained in HCMV infected cells.
HCMV induces the expression and ER stress-associated alteration in glycosylation of ATF6 but prevents ATF6 cleavage and activation. To determine whether HCMV activates the ATF6 pathway of the UPR, we assessed the status of ATF6 protein in HCMV-infected cells. The 90-kDa precursor of ATF6 exists in the ER as a glycosylated protein. In response to ER stress, there is an alteration in the glycosylation which is detectable by increased electrophoretic mobility (14). After translocation to the Golgi, the ATF6 precursor is cleaved into a 50-kDa product which functions as a transcription factor for many genes encoding ER chaperones (e.g., BiP and GRP94).
Using an antibody which recognizes the N terminus of ATF6 (common to both the 90-kDa and 50-kDa proteins; Fig. 5A), we determined the status of ATF6 over a 5-day time course of HCMV infection. As previously reported, thapsigargin treatment for up to 2 h resulted in a migration shift due to altered glycosylation (ATF6 to ATF6, Fig. 5A) and cleavage of the 90-kDa precursor into the 50-kDa "N"-terminal product ATF6(N). In HCMV-infected cells, we observed a significant induction in the level of ATF6 precursor by 1 day postinfection as well as a shift in the mobility of the precursor (from ATF6 to ATF6) between 3.5 and 4 days postinfection, indicative of an ER stress-induced change in the ATF6 glycosylation state. However, the 50-kDa cleavage product was not detected at any time during HCMV infection, suggesting that HCMV infection inhibits the cleavage of ATF6.
Since cleavage of ATF6 is necessary to produce the transcriptionally active form of ATF6, we predicted that ATF6 target genes would not be expressed in HCMV-infected cells. To test this, we analyzed the level of BiP and GRP94 by Western analysis (Fig. 5B). HCMV infection caused a transient increase in BiP at 1 to 2 days postinfection, times prior to the induction of the UPR; this may be an ATF6-independent induction mediated by immediate early or early viral proteins. However, intracellular BiP returned to basal levels by 3 days postinfection and remained low thereafter, in agreement with the lack of active ATF6.
The level of GRP94 gradually increased over the 5-day infection time course (Fig. 5B) and eventually attained a level equivalent to that seen in thapsigargin-treated cells (compare CMV 5d to 48 h Thap, Fig. 5B). Since the induction of GRP94 occurred by 1 day postinfection (prior to induction of the UPR by HCMV at 3 days postinfection), we conclude that an early viral function, not involving ATF6, accounts for the gradual increase in GRP94, a chaperone that may be useful during infection.
We also determined whether we could detect ATF6-dependent transcriptional activation of the gene encoding Xbp-1 by measuring total Xbp-1 mRNA in HCMV-infected and thapsigargin-treated cells using quantitative RT-PCR (Fig. 5C). As expected, a 2-h treatment with thapsigargin increased the amount of Xbp-1 mRNA to approximately 5-fold that detected in untreated cells; the mRNA remained elevated at 12 h posttreatment. The level of Xbp-1 mRNA in HCMV-infected cells relative to mock-infected cells at each time point showed an average induction of only 2-fold (Fig. 5C) over the course of the infection, significantly below that attainable with thapsigargin treatment. This low-level induction again agrees with the lack of active ATF6.
HCMV induces splicing of Xbp-1 mRNA. In response to ER stress, IRE-1 removes a 26-bp intron from the full-length Xbp-1 mRNA [Xbp-1(u)] which results in the spliced form of Xbp-1 [Xbp-1(s)]. Xbp-1(s) encodes XBP-1, a transcription factor which activates the expression of genes encoding chaperones as well as genes encoding proteins involved in UPR-mediated protein degradation, such as EDEM (ER degradation-enhancing -mannosidase-like protein). To determine whether the IRE-1 pathway is activated in response to HCMV-infection, we evaluated the splicing of Xbp-1 mRNA by RT-PCR using primers which amplify both Xbp-1(u) and Xbp-1(s) (Fig. 6A; upper and lower bands, respectively). In mock-infected cells, only the unspliced form of Xbp-1 was detected over the 5-day time course. However, in HCMV-infected cells, in addition to the unspliced form which was detected at all times, the spliced form of Xbp-1 was detectable as a faint band by 2 days postinfection and a more intense band by 3.5-days postinfection and thereafter. As expected, splicing of Xbp-1 was also observed in cells treated with thapsigargin for 30 min or 2 h. These results indicate that HCMV infection activates the IRE-1 pathway leading to the splicing of Xbp-1 RNA.
Since the spliced form of Xbp-1 encodes a transcription factor, we asked whether XBP-1 target genes were transcriptionally activated in HCMV-infected cells. Unlike transcription of BiP and GRP94, which can be activated by either ATF6 or XBP-1, transcription of EDEM is entirely dependent on XBP-1 (20). Thus, the level of EDEM mRNA in HCMV-infected cells was measured using real-time RT-PCR (Fig. 6B). In contrast to treatment with thapsigargin, which resulted in a 4- to 5-fold increase in the level of EDEM mRNA by 12 h posttreatment, no significant increase in the level of EDEM mRNA was detected at any time during the 5-day HCMV time course. Thus, the transcriptional activity of XBP-1, as indicated by the level of EDEM mRNA, is inhibited in HCMV infected cells.
DISCUSSION
Viral infection triggers cellular stress responses, potentially resulting in circumstances which could attenuate viral replication. In the case of ER stress and induction of the UPR, protein synthesis is attenuated and under severe ER stress conditions the cell undergoes apoptosis; neither condition would be beneficial to the viral infection. However, other aspects of the UPR may be beneficial to the viral infection. Thus, the question is whether the virus can limit harmful aspects of the UPR while maintaining beneficial ones. In the present work we provide evidence that HCMV infection induces the UPR but regulates the three signal transduction pathways of the UPR (PERK, ATF6, and IRE-1) in order to eliminate deleterious effects while maintaining potentially beneficial ones (Fig. 7).
HCMV infection activates the UPR between 2 and 3 days postinfection. HCMV encodes over 200 proteins, many of which are abundantly expressed and undergo glycosylation and other modifications in the ER. Under the experimental conditions used here (HFF cells, HCMV Towne, and multiplicity of infection of 3), phosphorylation of PERK and eIF2 was first detected at 3 days postinfection and Xbp-1 splicing was detected at 2 days postinfection. Thus, UPR activation by HCMV likely occurs at some time between 2 and 3 days postinfection. This is a time when there is abundant synthesis of late viral glycoproteins (29). It is possible that improper folding or aggregation of these proteins triggers the UPR. Indeed, a similar phenomenon has been demonstrated with the hepatitis C virus envelope protein E2, whose expression has been shown to activate transcription of ER stress-associated chaperone genes (22).
It is also possible that immunomodulatory proteins encoded by HCMV (US2, US3, US6, and US11) could induce the UPR through their disruption of major histocompatibility complex class I antigen presentation. In particular, the US2 and US11 glycoproteins, which are expressed with late kinetics, are located in the ER membrane of infected cells and redirect nascent major histocompatibility complex class I proteins from the ER into the cytosol (reviewed in reference 28). This activity may disrupt other transport mechanisms in the ER leading to ER stress and induction of the UPR. Furthermore, a recent report (23) indicates that HCMV US11 directly interacts with ER proteins which function in the export of proteins from the ER for degradation (e.g., Derlin-1) and protein refolding (i.e., BiP), suggesting US11 itself may modify ER function.
Yet another possible ER stress stimulus is HCMV virion envelopment, a process which utilizes the host secretory network and may perturb the ER (reviewed in references 27). Analysis of HCMV-infected cells by electron microscopy demonstrates large numbers of ER membrane-derived vesicles (our unpublished data); thus, virion envelopment may induce ER stress and the UPR. Consistent with this idea, flaviviruses are thought to induce ER stress as a result of particle assembly and virus maturation at ER membranes (31).
Precise control of BiP levels may allow the viral infection to control when the UPR is activated. Although HCMV clearly activates the UPR, the temporal nature of the induction may be specifically controlled. Our data suggest that HCMV infection causes a transient increase in BiP levels at 1 to 2 days postinfection, but by 3 days postinfection, BiP returns to basal levels. It is interesting to consider that the increased BiP present at 1 and 2 days could significantly inhibit the activation of the UPR by maintaining PERK, ATF6m and IRE-1 in their inactive forms.
The rapid disappearance of BiP between 2 and 3 days postinfection coincides with the appearance of UPR activation in infected cells. Thus, a precise, virally mediated control of BiP levels could determine when the UPR is activated. Interestingly, it has recently been shown that BiP is included in the HCMV virion (39); this may account in part for the elevated BiP levels at 1 and 2 days; however, it is equally likely that a viral immediate-early or early protein activates the BiP expression seen at 1 and 2 days postinfection. Regardless, the rapid decrease in BiP between 2 and 3 days postinfection may be virally mediated in order to allow UPR activation during the late phase of the infection. In addition, a recent report demonstrates a physical interaction between BiP and the HCMV glycoprotein US11 (23), suggesting a mechanism by which HCMV may target BiP for inactivation or degradation.
HCMV maintains translation despite UPR activation. Our data show that HCMV infection significantly activates PERK during the viral infection; however, the expected phosphorylation of eIF2 is significantly limited and global translation is not inhibited in infected cells. Consistent with our results, it has previously been reported that HCMV prevents the phosphorylation of eIF2, and it has been suggested that the immediate early proteins encoded by TRS1 and the closely related IRS1 genes are involved (5). The disproportionate levels of phosphorylated PERK compared to phosphorylated eIF2 has also been observed during hepatitis C virus infection where the hepatitis C virus E2 protein binds to and inhibits phosphorylated PERK, thus inhibiting the phosphorylation of eIF2 and relieving the global translation inhibition normally induced by PERK activation (30).
HCMV also causes a significant increase in the amount of total eIF2, a condition which would favor an increased level of unphosphorylated eIF2, especially under the conditions of limited phosphorylation as described above. Together, all of the data support a model in which HCMV limits the amount of phosphorylated eIF2 in order to maintain global translation.
Translation dependent on phosphorylated eIF2 is maintained during HCMV infection. Despite the limited phosphorylation of eIF2 in HCMV-infected cells, we detected a significant increase in the protein level of ATF4, whose mRNAs depend on phosphorylated eIF2 for translation (11). Expression of ATF4 leads to transcriptional activation of target genes involved in metabolism and redox maintenance, and thus may benefit the course of HCMV infection by maintaining a permissive cellular environment. In addition, ATF4 has been shown to transcriptionally activate GADD34 (26), which interacts with protein phosphatase 1 to dephosphorylate eIF2 (4). Induction of GADD34 may benefit the infection through dephosphorylation of eIF2 and may be one means by which the amount of phosphorylated eIF2 is limited. In this regard, HSV-1 encodes a GADD34 homolog, ICP34.5, suggesting that GADD34 function may play an important role in herpesvirus biology (6).
Our data suggest that during an HCMV infection, both global translation, dependent on unphosphorylated eIF2, and phospho-eIF2-dependent translation can be maintained simultaneously. However, the ability to support both global and phospho-eIF2-dependent translation is not supported by the present model for phospho-eIF2 inhibition of translation. Phosphorylated eIF2 has a much greater binding affinity for eIF2B than unphosphorylated eIF2, and the stable complex of phospho-eIF2 and eIF2B effectively eliminates eIF2B's guanine nucleotide exchange function. It has been proposed that the cell's limited amount of eIF2B can all be bound by only a small amount of phospho-eIF2 (18), inhibiting global translation and allowing phospho-eIF2-dependent translation.
By this mechanism, the appearance of phospho-eIF2 in infected cells suggests that there should be inhibition of global translation, yet this is not what we observed. However, it has been suggested that translation of ATF4 occurs at levels of eIF2 that have only minimal inhibitory effects on global translation (7, 8, 38). Thus, HCMV's ability to increase total levels of eIF2 while limiting its phosphorylation may ensure that both global translation and phospho-eIF2a-dependent translation are maintained. What is clear is that HCMV has evolved a means to manipulate translation to its advantage, eliminating the negative effects of PERK activation (translation attenuation) while maintaining the positive effects that result from induction of ATF4.
HCMV inhibits the activation of ATF6. The transcriptional activation function of ATF6 is critical to the UPR, since ATF6 increases the expression of many ER chaperone genes whose products are necessary for protein refolding. In addition, ATF6 increases the level of Xbp-1 mRNA, which, once spliced by IRE-1, encodes a transcriptional activator of genes whose products are involved in protein degradation (discussed further below). In HCMV-infected cells, the levels of the ATF6 precursor protein increased significantly by 1 days postinfection, and between 3.5 and 4 days postinfection the ATF6 precursor displayed a mobility shift consistent with altered glycosylation that occurs when there is ER stress (14). However, in HCMV infected cells the ATF6 precursor is not cleaved to the active form.
Inhibition of ATF6 activation by HCMV was somewhat surprising, as one might expect that its transcriptional activation of chaperone encoding genes might benefit the virus by aiding in protein folding and preventing protein aggregation. However, our data suggest that the virus can increase the expression of ATF6 target genes in a gene specific manner independent of ATF6 activation. The transient increases in BiP levels detected at 1 to 2 days postinfection are discussed above and are likely independent of ATF6. Although GRP94 protein levels increased during the infection, the induction occurred early in the infection (1 day postinfection) and GRP94 gradually increased over the course of the 5-day infection. This early induction (prior to the detection of UPR activation) and the gradual increase of the protein over time, is not consistent with UPR activation but is consistent with an immediate early or early viral transcriptional activation mechanism which does not involve ATF6. A third ATF6 target gene, Xbp-1, was activated by no more than 2-fold during the infection, in agreement with there being no active ATF6. As described below, the inhibition of transcription of the Xbp-1 gene may be beneficial to the infection.
HCMV inhibits XBP-1-mediated transactivation. The transcriptional activation function of XBP-1 is also critical to the UPR since it stimulates expression of proteins involved in ER stress-induced protein degradation, such as EDEM, as well as chaperones also controlled by ATF6 (20). Our data show that splicing of Xbp-1 mRNA is stimulated by HCMV infection, consistent with the activation of the IRE-1 pathway. However, transcriptional activation of the XBP-1 target gene EDEM is not detected in HCMV-infected cells. This suggests that either the levels of spliced Xbp-1 mRNA in HCMV-infected cells are too low to produce sufficient XBP-1 protein or the transcriptional activity of XBP-1 is inhibited by HCMV. At this point either or both possibilities could contribute to HCMV's inhibition of XBP-1 transcriptional activation. In cells containing hepatitis C virus replicons, the transcriptional activity of XBP-1 is inhibited despite the presence of significant levels of spliced XBP-1 (37). In any event, the resulting inhibition of expression of genes such as EDEM may be beneficial to viral infection, since this would inhibit ER stress-induced protein degradation, which may otherwise hinder the accumulation of necessary viral proteins.
In sum, our data show that during an HCMV infection the UPR is induced but is modified in ways which benefit the viral infection, inhibiting those aspects which may be deleterious to the progress of the viral infection and maintaining those which may be beneficial.
ACKNOWLEDGMENTS
We thank Kazutoshi Mori for providing antisera to ATF6; the members of the Alwine laboratory for helpful discussions and critical evaluation of the data; and Sherri Adams for critical reading of the manuscript.
J.A.I. is supported by an NIH postdoctoral training grant (F32 AI062055-01). A.H.S. is supported by a Howard Hughes Medical Institute predoctoral fellowship. This work was supported by NIH grant CA28379-24 awarded to J.C.A. by the National Cancer Institute and by a Tobacco Settlement Grant awarded to J.C.A. by the Abramson Cancer Center at the University of Pennsylvania.
REFERENCES
Bertolotti, A., Y. Zhang, L. M. Hendershot, H. P. Harding, and D. Ron. 2000. Dynamic interaction of BiP and ER stress transducers in the unfolded-protein response. Nat. Cell Biol. 2:326-332.
Bitko, V., and S. Barik. 2001. An endoplasmic reticulum-specific stress-activated caspase (caspase 12) is implicated in the apoptosis of A549 epithelial cells by Respiratory Syncytial Virus. J. Cell Biochem. 80:441-454.
Bresnahan, W. A., G. E. Hultman, and T. Shenk. 2000. Replication of wild type and mutant human cytomegalovirus in life-extended human diploid fibroblasts. J. Virol. 74:10816-10818.
Brush, M. H., D. C. Weiser, and S. Shenolikar. 2003. Growth arrest and DNA damage-inducible protein GADD34 targets protein phosphatase 1 to the endoplasmic reticulum and promotes dephosphorylation of the subunit of eukaryotic translation initiation factor 2. Mol. Cell. Biol. 23:1292-1303.
Child, S. J., M. Hakki, K. L. De Niro, and A. P. Geballe. 2004. Evasion of cellular antiviral responses by human cytomegalovirus TRS1 and IRS1. J. Virol. 78:197-205.
Chou, J., and B. Roizman. 1994. Herpes simplex virus 1 134.5 gene function, which blocks the host response to infection, maps in the homologous domain of the genes expressed during growth arrest and DNA damage. Proc. Natl. Acad. Sci. USA 91:5247-5251.
Deng, J., P. D. Lu, Y. Zhang, D. Sheuner, R. J. Kaufman, N. Sonenberg, H. P. Harding, and D. Ron. 2004. Translational repression mediates activation of nuclear factor kappa B by phosphorylated translation initiation factor 2. J. Virol. 24:10161-10168.
Dever, T. E., L. Feng, R. C. Wek, A. M. Cigan, T. F. Donahue, and A. G. Hinnebusch. 1992. Phosphorylation of initiation factor 2 alpha by protein kinase GCN2 mediates gene-specific translational control of GCN4 in yeast. Cell 68:585-596.
Dimcheff, D. E., S. Askovic, A. H. Baker, C. Johnson-Fowler, and J. L. Portis. 2003. Endoplasmic reticulum stress is a determinant of retrovirus-induced spongiform neurodegeneration. J. Virol. 77:12617-12629.
Harding, H. P., M. Calfon, F. Urano, I. Novoa, and D. Ron. 2002. Transcriptional and translational control in the mammalian unfolded protein response. Annu. Rev. Cell Dev. Biol. 18:575-599.
Harding, H. P., Y. Zhang, A. Bertolotti, H. Zeng, and D. Ron. 2000. PERK is essential for translation regulation and cell survival during unfolded protein response. Mol. Cell. 5:897-904.
Harding, H. P., Y. Zhang, and D. Ron. 1999. Protein Translation and folding are coupled by an endoplasmic-reticulum-resident kinase. Nature 397:271-274.
Harding, H. P., Y. Zhang, H. Zeng, I. Novoa, P. D. Lu, M. Calfon, N. Sadri, C. Yun, B. Popko, R. Paules, D. F. Stojdl, J. C. Bell, T. Hettmann, J. M. Leiden, and D. Ron. 2003. An integrated stress response regulates amino acid metabolism and resistance to oxidative stress. Mol. Cell. 11:619-633.
Hong, M., S. Luo, P. Baumeister, J. Huang, R. K. Gogia, M. Li, and A. S. Lee. 2004. Underglycosylation of ATF6 as a novel sensing mechanism for activation of the unfolded protein response. J. Biol. Chem. 279:11354-11363.
Jordan, R., L. Wang, T. M. Graczyk, T. M. Block, and P. R. Romano. 2002. Replication of a cytopathic strain of Bovine Viral Diarrhea Virus Activates PERK and induces endoplasmic reticulum stress-mediated apoptosis of MDBK cells. J. Virol. 76:9588-9599.
Kaufman, R. J. 1999. Stress signaling from the lumen of the endoplasmic reticulum: coordination of gene transcription and translational controls. Gen. Dev. 13:1211-1233.
Kozutsumi, Y., M. Segal, K. Normington, M. J. Gething, and J. Sambrook. 1988. The presence of malfolded proteins in the endoplasmic reticulum signals the induction of glucose-regulated proteins. Nature 332:462-464.
Krishnamoorthy, T., G. D. Pavitt, F. Zhang, T. E. Dever, and A. G. Hinnebusch. 2001. Tight binding of the phosphorylated subunit of initiation factor 2 (eIF2) to the regulatory subunits of guanine nucleotide exchange factor eIF2B is required for inhibition of translation initiation. Mol. Cell. Biol. 21:5018-5030.
Lee, A. S. 1992. Mammalian stress response: induction of the glucose-regulated protein family. Cur. Opin. Cell Biol. 4:267-273.
Lee, A. S., N. N. Iwakoshi, and L. H. Glimcher. 2003. XBP-1 regulates a subset of endoplasmic reticulum-resident chaperone genes in the unfolded protein response. Mol. Cell. Biol. 23:7448-7459.
Lee, K., W. Tirasophon, X. Shen, M. Michalak, R. Prywes, T. Okada, H. Yoshida, K. Mori, and R. J. Kaufman. 2002. IRE-1-mediated unconventional mRNA splicing and S2P-mediated ATF6 cleavage merge to regulate XBP1 in signaling the unfolded protein response. Genes Dev. 16:452-466.
Liberman, E., Y. Fong, M. J. Selby, W. Choo, L. Cousense, M. Houghton, and T. S. B. Yen. 1999. Activation of the grp78 and grp94 promoters by Hepatitis C Virus E2 envelope protein. J. Virol. 73:3718-3722.
Lilley, B. N., and H. L. Ploegh. 2004. A membrane protein required for dislocation of misfolded proteins from the ER. Nature 249:834-840.
Lindenbach, B. D., and C. M. Rice. 2001. Flaviviridae: the viruses and their replication, p. 991-1041. In P. M. Howe, D. M. Knipe, D. E. Griffin, R. A. Lamb, M. A. Martin, B. Roizman, and S. E. Straus (ed.), Fields virology, 4th ed. Lippincott Williams & Wilkins, Philadelphia, Pa.
Lu, P. D., H. P. Harding, and D. Ron. 2004. Translation reinitiation at alternative open reading frames regulates gene expression in an integrated stress response. J. Cell Biol. 167:27-33.
Ma, Y., and L. M. Hendershot. 2003. Delineation of a negative feedback regulatory loop that controls protein translation during endoplasmic reticulum stress. J. Biol. Chem. 278:34864-34873.
Mettenleiter, T. C. 2002. Herpesvirus assembly and egress. J. Virol. 76:1537-1547.
Mocarski, E. S. 2002. Immunomodulation by cytomegaloviruses: manipulative strategies beyond invasion. Trends Microbiol. 10:332-339.
Mocarski, E. S., and C. T. Courcelle. 2001. Cytomegaloviruses and their replication, p. 2629-2673. In P. M. Howe and D. M. Knipe (ed.), Fields virology. Lippincott Williams & Wilkins, Philadelphia, Pa.
Pavio, N., P. R. Romano, T. M. Graczyk, S. M. Feinstone, and D. R. Taylor. 2003. Protein synthesis and endoplasmic reticulum stress can be modulated by the Hepatitis C Virus envelope protein E2 through the eukaryotic initiation factor 2 kinase PERK. J. Virol. 77:3578-3585.
Rice, C. M. 1996. Flaviviridae: the viruses and their replication, p. 931-959. In P. M. Howe and D. M. Knipe (ed.), Fields virology. Lippincott-Raven Publishers, Philadelphia, Pa.
Rutkowski, D. R., and R. J. Kaufman. 2004. A trip to the ER: coping with stress. Trends Cell Biol. 14:20-28.
Shang, J., and M. A. Lehrman. 2004. Discordance of UPR signaling by ATF6 and IRE1p-XBP1 with levels of target transcripts. Biochem. Biophys. Res. Commun. 317:390-396.
Shen, J., X. Chen, L. M. Hendershot, and R. Prywes. 2002. ER stress regulation of ATF6 localization by dissociation of BiP/GRP78 binding and unmasking of golgi localization signals. Dev. Cell. 3:99-111.
Su, H. L., C. L. Liao, and Y. L. Lin. 2002. Japanese Encephalitis Virus infection initiates endoplasmic reticulum stress and an unfolded protein response. J. Virol. 76:4162-4171.
Tardif, K. D., K. Mori, and A. Siddiqui. 2002. Hepatitis C virus subgenomic replicons induce endoplasmic reticulum stress activating an intracellular signaling pathway. J. Virol. 76:7453-7459.
Tardif, K. D., K. I. Mori, R. J. Kaufman, and A. Siddiqui. 2004. Hepatitis C Virus suppresses the IRE-1-Xbp1 pathway of the unfolded protein response. J. Biol. Chem. 279:17158-17164.
Tzamarias, D., I. Roussou, and G. Thireos. 1989. Coupling of GCN4 mRNA translational activation with decreased rates of polypeptide chain initiation. Cell 57:947-954.
Varnum, S. M., D. N. Streblow, M. E. Monroe, P. Smith, K. J. Auberry, L. Pasa-tolic, D. Wang, D. G. Camp, K. Rodland, S. Wiley, W. Britt, T. Shenk, R. D. Smith, and J. A. Nelson. 2004. Identification of proteins in human cytomegalovirus (HCMV) particles: the HCMV proteome. J. Virol. 78:10960-10966.
Wang, Y., J. Shen, N. Arenzanna, W. Tirasaphon, R. Kaufman, and R. Prywes. 2000. Activation of ATF6 and an ATF6 DNA binding site by the endoplasmic reticulum stress response. J. Biol. Chem. 275:27013-27020.
Ye, J., R. B. Rawson, R. Komuro, X. Chen, U. P. Dave, R. Prywes, M. S. Brown, and J. L. Goldstein. 2000. ER stress induces cleavage of membrane-bound ATF6 by the same proteases that process SREBPs. Mol. Cell. 6:1355-1364.
Yoshida, H., T. Matsui, A. Yamamoto, T. Okada, and, M. K. 2001. Xbp1 mRNA is induced by ATF6 and spliced by IRE-1 in response to ER stress to produce a highly active transcription factor. Cell 107:881-891.
Yoshida, H., T. Okada, K. Haze, H. Yanagi, T. Yura, M. Negishi, and K. Mori. 2001. Endoplasmic reticulum stress-induced formation of transcription factor complex ERSF including NF-Y (CBF) and activating transcription factors 6 and 6? that activates the mammalian unfolded protein response. Mol. Cell. Biol. 21:1239-1248.(Jennifer A. Isler, Alison)