Suppression of thermotolerance in mumps virus-infected cells is caused by lack of HSP27 induction contributed by STAT-1.

Viral infection modulates the regulation of apoptosis in host cells. Here, we report a novel mechanism by which human cells infected with mumps virus become susceptible to apoptosis caused by extracellular stresses. Mumps virus stimulates proteasome-dependent degradation of STAT-1 by action of viral accessory protein V, resulting in a severe decrease in STAT-1 protein in infected cells. We exposed mumps virus-infected and uninfected cells to heat and chemical stress. The infected cells failed to acquire resistance to apoptotic stimuli (thermotolerance) after exposure to these mild stresses. The induction of HSP27 by stress exposure was dramatically suppressed in the infected cells, but HSP70 induction was not affected. STAT-1 was required for transcriptional activation of the HSP27 gene, but not for the HSP70 gene, and cDNA transfection of STAT-1 in mumps virus-infected cells restored thermotolerance. Phosphorylated heat shock factor-1 (HSF-1) and STAT-1 phosphorylated on neither tyrosine nor serine residues were co-transported to the nucleus in response to stress. Furthermore, overexpression of unphosphorylatable mutants of STAT-1 also restored thermotolerance in mumps virus-infected cells. These lines of evidence indicate that the induction of HSP27 by stress requires STAT-1 in addition to the activated HSF-1. Furthermore, STAT-1 required for the induction of HSP27 worked independent to its phosphorylation. Thus, HSP27-dependent thermotolerance is suppressed by mumps virus infection through the destruction of STAT-1. The lack of thermotolerance should allow the infected cells to be eliminated by apoptosis and might be a host defense against viral infection.

To protect against viral infection, the host makes every effort to prevent viral replication and dissemination. The induction of cell death in virus-infected cells represents an effective host defense mechanism to eliminate the infected cells. Interferon (IFN) 1 is known to play various roles in host defense. One is the establishment of an antiviral state, which is achieved through translational suppression of protein synthesis by the action of 2Ј,5Ј-oligoadenylate synthetase and double-stranded RNA-dependent protein kinase (PKR) (1)(2)(3). Another role is the facilitation of cell death (4). Many IFN-induced gene products, such as PKR, TRAIL (tumor-necrosis factor-related apoptosis-inducing ligand), Fas (CD95/APO1), and IRFs (interferon regulatory factors), are also known to be mediators of apoptosis (4,5). Apoptotic cells infected with virus are rapidly eliminated by phagocytes such as macrophage and dendritic cells, and the phagocytosed virus is digested and carried into antigen presentation (6,7).
On the other hand, viruses suppress host defense mechanisms including the action of IFN by various strategies (8,9). The IFN signaling pathways have been well characterized as have those of various cytokines (10,11). Janus kinases (JAKs), which associate with certain IFN receptors, are activated when IFN binds to the specific receptors. The activated JAKs Tyrphosphorylate and thereby activate a family of transcription factors named signal transducer and activator of transcription (STAT). After stimulation with type I IFN, namely IFN-␣ or IFN-␤, the transcription factor forms a complex called ISGF3, which consists of Tyr-phosphorylated STAT-1, Tyr-phosphorylated STAT-2, and IRF9/p48/ISGF3␥. This complex then binds to the IFN-stimulated response element in the IFN-inducible gene promoter. After stimulation with IFN-␥, ␥-activated factor, which is a dimer of phosphorylated STAT-1, binds to the ␥ activation sequence. Many viruses, for example those belonging to family Paramyxoviridae (12,13), are reported to shut off or suppress the IFN signaling pathway (8,9). Viruses belonging to the genus Respirovirus within the family Paramyxoviridae, such as Sendai virus and human parainfluenza virus type 3, reduce the Tyr phosphorylation of STAT-1 (14 -17). More recently, we showed that the measles virus, a Morbillivirus, inhibits the Tyr phosphorylation of Jak1 induced by IFN-␣, but not that induced by IFN-␥ (18). Viruses belonging to the genus Rubulavirus have been shown to dramatically reduce the basal level of STAT proteins. Mumps virus and simian virus 5 accelerate the degradation of STAT-1 protein in a proteasome-dependent manner (19 -21). Human parainfluenza virus type 2 reduces STAT-2 protein levels (22,23). Accelerated degradation of STAT-1 by mumps virus is necessary and sufficient for the expression of viral accessory protein V (20,21,24). These STAT proteins act as transcription factors for IFN-inducible genes, so that the decrease in the proteins leads not only to a reduction of the antiviral state but also to other IFN-induced cellular responses. In fact, we reported that augmentation of apoptosis by IFN is diminished in mumps virus-infected cells (25,26). This suppression of apoptosis is also considered to be a mechanism by which the virus can escape from the host defense, because induction of apoptosis is a strategy that host cells use to eliminate infected virus.
In this report, we show a novel mechanism that mumps virus-infected cells use to fight infection. Infected cells do not acquire thermotolerance and easily undergo apoptosis by stress. This phenomenon probably developed as a measure to counteract the inhibition of IFN-induced apoptosis by the action of the virus. Thermotolerance is a form of resistance to cell death acquired after exposure to mild, nonlethal, extracellular stresses (27,28). On the molecular level, the induction of a series of heat shock proteins (HSPs) is involved in the development of thermotorlerance (29 -32). Most HSPs, also called stress proteins, act as molecular chaperones, which assist in folding newly synthesized proteins and denatured proteins formed by stress. HSPs consume ATP as an energy source to fold proteins (33). Some HSPs, such as HSP70 and HSP27, are also known to be suppressors of cell death including apoptosis (34 -36). The suppressive activity of the stress-induced HSPs give rise to thermotolerance. Here, we investigated whether the lack of thermotolerance in mumps-virus infected cells is related to HSP27 induction. We found that expression of HSP27 was STAT-1-dependent, so that the induction of HSP27 disappeared in mumps virus-infected cells lacking STAT-1.

EXPERIMENTAL PROCEDURES
Cells and Viruses-The human oral squamous cell carcinoma lines, OSC19 and OSC70, and the stable transformant clones of OSC70 overexpressing human HSP27 were described previously (37,38). The 2fTGH cell line and its STAT1-deficient mutant, U3A, were kindly donated by Dr G. R. Stark (The Cleveland Clinic Foundation). These cells were maintained as described elsewhere (39,40).
Cells persistently infected with mumps virus were established as described previously (41,42). Briefly, OSC19 and OSC70 cells, infected with mumps virus vaccine strain Torii at a multiplicity of infection of 0.1, were cultured and passaged every 5 or 7 days. Persistent infection in cells was confirmed by virus titer in the culture supernatant, and viral V protein expression in cells was determined by Western blotting.

FIG. 1. Thermotolerance and stressinduced HSP27 suppress in oral squamous carcinoma cells persistently infected with mumps virus.
A and B, effects of AZC treatment on camptothecin-induced apoptosis and inducibility of HSP27 and HSP70 in OSC19 and mumps virus-infected OSC19 (OSC19-MP) cells. A, OSC19 and OSC19-MP cells were treated with 10 mM AZC for 2 h, cultured in normal medium for 5 h, and then treated with 5 g/ml camptothecin for 16 h. Cell viability was determined by gentian violet staining. B, cells were treated with 10 mM AZC for 2 h and then cultured in normal medium for various times. The protein levels of HSP27 and HSP70 were determined by Western blotting. Quantified results (mean Ϯ S.D. from triplicate experiments) are presented in graphs. Open circles, uninfected cells; closed circles, infected cells. C and D, effects of pre-heat shock on camptothecin-induced cells death and HSP induction in OSC70 and OSC70-MP cells. C, OSC70 and mumps virus-infected OSC70 (OSC70-MP) cells were cultured at 42°C for 1 h and then at 37°C for 5 h. The resulting cells were treated with 5 g/ml camptothecin for 16 h, and cell viability was determined. D, the protein levels of HSP27 and HSP70 during recovery culture after heat shock were determined by Western blotting at various time points. Quantified results are presented in graphs as above. E and F, exogenous HSP27 protects OSC70 cells from apoptosis. E, recombinant OSC70 clones stably overexpressing HSP27 were established. The HSP27 expression was confirmed by Western blotting. HSP70 levels were determined as a control. F, the clones were treated with 5 g/ml camptothecin for 16 h, and cell viability was determined.
otherwise mentioned, camptothecin was used at a concentration of 5 g/ml.
Cell Viability-Cell viability was determined by a gentian violet dye binding assay as previously described (20). The binding dye was solubilized by 2-methoxyethanol, and absorbance at 595 nm was measured.
Viability was also measured with a modified 3-(4,5-dimethylthiazol-2yl)-2,5-diphenyltetrazolium bromide assay using a Cell Counting Kit 8 (Dojin Chemical, Kumamoto, Japan) Stress Experiments-Cells were exposed to stress by two methods, exposure to heat shock and treatment with a proline analogue, AZC. For heat shock experiments, the culture flasks were exposed to 42°C for 1 h, and then the cells were cultured under normal conditions for recovery from stress. For chemical stress experiments, cells were treated with 10 mM AZC for 2 h (43). After the removal of AZC, the cells were cultured under normal conditions for recovery.
The plasmids (0.3 or 1 g/10 5 cells) were transfected using Superfect reagent (Qiagen, Hilden, Germany) according to the manufacturer's instructions. Twenty-four hours after transfection, the cells were exposed to 10 mM AZC for 2 h and then cultured for 5 h under normal conditions. The resulting cells were treated with 5 g/ml camptothecin for 16 h, and then cell viability was measured as described above.
Western Blotting-SDS-polyacrylamide gel electrophoresis and Western blotting were carried out as described previously (44,45).
Reporter Assay with Luciferase-The human hsp27 promoter was kindly donated by Dr. L. A. Weber (University of Nevada). The human hsp27 promoter (Ϫ762 to Ϫ12), which was obtained from pKSm vector (38), and the human hsp70b promoter, which was obtained from the pD3SX vector (StressGen), were inserted into pGL3-basic luciferase reporter vectors harboring firefly luciferase (Promega). The luciferase plasmid and reference plasmid pRL-SV40 (Promega), which was used as a control for transfection efficacy, were co-transfected into cells using the Superfect reagent. Twenty-four hours after transfection, the cells were further incubated with 1 h under heat shock (42°C) and then 4 h under normal conditions, 1 h normal conditions and then 4 h of IFN-␥ treatment, or 1 h heat shock and then 4 h of IFN-␥ treatment. The

FIG. 2. Transient expression of STAT-1 restores the inducibility of thermotolerance and HSP27 expression in OSC19 cells infected with mumps virus.
A human STAT-1 expression plasmid or control plasmid (mock) was transfected into OSC19 and OSC19-MP by lipofection. A, protein levels of STAT-1 in OSC19 and OSC19-MP cells transiently overexpressing STAT-1 determined by Western blotting. Relative amounts of STAT-1␣ proteins quantified from three experiments are shown below each band. Actin levels were determined as a control for protein loading. B, induction of thermotolerance. Cells were treated with 10 mM AZC for 2 h, cultured in normal medium for 5 h, and then treated with 5 g/ml camptothecin for 16 h. Cell viability was determined. C, induction of HSP27 and HSP70. Cells were treated with 10 mM AZC for 2 h and then cultured in normal medium for various lengths of time. The protein levels of HSP27 and HSP70 were determined by Western blotting. Quantified results (mean Ϯ S.D. from triplicate experiments) are presented as graphs. Open circles, OSC19-MP(mock); closed circles, OSC19-MP(STAT-1).
FIG. 3. The hsp27 promoter, but not the hsp70 promoter, is activated by heat shock in a STAT-1-dependent manner determined by luciferase reporter gene assay. An hsp27 promoter-luciferase or an hsp70 promoter-luciferase reporter plasmid and a control reporter plasmid, pRL-SV40, were transfected into 2fTGH or its STAT-1-deficient mutant U3A. Cells were untreated (bar 1), treated with 1000 IU/ml IFN-␥ for 4 h (bar 2), or cultured at 42°C for 1 h and then at 37°C for 4 h without (bar 3) or with (bar 4) 1000 IU/ml IFN-␥. The treated cells were lysed, and luciferase activities were measured. The experiment was performed in triplicate. experiments were carried out in triplicate. The cells were lysed, and luciferase activity in the lysate was assessed using the Dual-Luciferase Reporter Assay System (Promega). Reporter activity was calculated as the ratio of firefly luciferase activity to Renilla luciferase activity. The results are expressed as -fold induction relative to the value of the untreated control.
Immunoprecipitation Analysis-Immunoprecipitation was carried out as described previously (18,45). Briefly, cells were lysed in radioimmune precipitation buffer (1% Nonidet P-40, 100 mM NaCl, 4 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, 5 g/ml aprotinin, 1 mM Na 3 VO 4 , 1 mM NaF, and 50 mM HEPES-NaOH, pH 7.5). The cell lysate and appropriate antibody (1 g) were incubated together, and then the complex containing the antibody was precipitated by protein G-Sepharose 4B (Amersham Biosciences). The resin was washed with radioimmune precipitation buffer three times, and the bound materials were eluted by boiling with SDS-PAGE sample buffer.
Extraction of Nuclear Fraction-Nuclear fractions were prepared using NE-PER Nuclear and Cytoplasmic Extraction Reagents (Pierce).

Stress Does Not Induce Thermotolerance or HSP27 Expression in Mumps
Virus-infected Cells-We examined the ability of stress to induce an anti-apoptotic state in cells infected with mumps virus (Fig. 1). The heat shock response was induced by incubation at 42°C for 1 h followed by recovery cultivation at 37°C for 5 h. The chemical stress response was induced by treatment with 10 mM AZC for 2 h followed by recovery cultivation at 37°C for 5 h. OSC19 and OSC70 cells were significantly (p Ͻ 0.01) resistant to camptothecin-induced apoptosis after recovery from stress. However, mumps virus-infected OSC19 (OSC19-MP) and OSC70 (OSC70-MP) cells did not show altered sensitivity to camptothecin after stress (Fig. 1, A  and C). Protein levels of HSPs during the stress response in the cells with or without mumps virus infection were determined by Western blotting. HSP70 was faintly induced after chemical stress in both infected and uninfected OSC19 cells. On the other hand, HSP27 was induced by chemical stress with AZC in the uninfected OSC19 cells but not in the mumps virus-infected cells (Fig. 1B). Both HSP27 and HSP70 were induced in OSC70 cells during the recovery period after heat shock. However, the level of HSP27 did not change significantly in mumps virusinfected cells (OSC70-MP) (Fig. 1D), whereas the level of HSP70 was increased by heat shock in the infected cells. These results suggest that the response of HSP27 to stress is suppressed in mumps-virus infected cells, but that of HSP70 is not. The lack of thermotolerance should be due to suppression of HSP27 inducibility. Recombinant OSC70 clones stably overexpressing HSP27 were resistant to camptothecin-induced apoptosis (Fig. 1, E and F). Not only the HSP27 levels but also cell viability after camptothecin treatment in the clones were similar to those in OSC70 cells after heat shock (Fig. 1, C to F). The results support the induction of HSP27 as a main contributor to thermotolerance in OSC19 and OSC70 cells.
Exogenous STAT-1 Expression Restores Thermotolerance and HSP27 Induction-Previously, we reported that STAT-1 protein expression diminishes or dramatically decreases in mumps virus-infected cells (21,46). Mumps virus accessory FIG. 4. Activation and nuclear translocation of HSF-1 and STAT-1 during heat shock and IFN treatment in OSC19 and OSC19-MP. A, OSC19, OSC19-MP transfected with a STAT-1 expression plasmid, and OSC19-MP transfected with a control plasmid (mock) were treated with heat shock (42°C, 1 h) or IFN-␥ (1000 IU/ml, 20 min). Lanes: C, untreated control; HS, heat shock-treated; IFN, IFN-␥-treated. The cells were lysed, and the lysates were analyzed by Western blot. Phosphorylated HSF-1 was detected by change of mobility. Phosphorylated STAT-1 was detected by specific antibodies for the Tyr-phosphorylated (pY) and Ser-phosphorylated (pS) forms. B, immunoprecipitation (IP) of an OSC19 cell lysate with an anti-HSF-1 antibody or rabbit IgG obtained from non-immunized sera (unrelated Ab) as a control. The precipitates were analyzed by Western blot. Western blotting results for total cell lysate used this experiment were shown in panel A. C and D, nuclear translocation of HSF-1 and STAT-1 after heat shock or IFN-␥ treatment in OSC19 (C) and OSC19-MP (D). The nuclear and cytosolic fractions were extracted from OSC19 and OSC19-MP after treatment as described in A. Each fraction was analyzed by Western blot.
protein V accelerates the proteasome-dependent degradation of STAT-1 (21,21). We therefore examined the contribution of STAT-1 to the lack of thermotolerance and HSP27 induction in infected cells. OSC19-MP cells transfected with a STAT-1 expression plasmid expressed about 50% of the amount of STAT-1 protein in uninfected OSC19 cells (Fig. 2A). The STAT-1 expression in OSC19-MP also restored the stress-induced resistance to apoptosis (Fig. 2B) and the inducibility of HSP27 (Fig.  2C). The induction of HSP70 was not affected. Uninfected OSC19 cells overexpressing STAT-1 showed similar inducibility of HSP27 and HSP70 by the stress compared with OSC19 cells transfected with plasmids. The mumps virus may therefore suppress HSP27-related thermotolerance through degradation of STAT-1.
STAT-1 Is Necessary for Transcriptional Activation of HSP27-To reveal the contribution of STAT-1 to HSP27 expression, a luciferase reporter assay was carried out using 2fTGH cells and the STAT-1-deficient mutant U3A cells derived from them. Although the hsp27 promoter was activated about 4-fold by heat shock in 2fTGH cells, its activation was dramatically suppressed in the STAT-1-deficient U3A cells (Fig. 3). IFN-␥ treatment did not affect basal or heat shockinduced hsp27 promoter activity in either cell type. In contrast, the hsp70 promoter was strongly activated by heat shock in both cell types. However, there was ϳ30% less hsp70 promoter activity in response to heat shock induction in U3A cells than in 2fTGH cells. The hsp70 promoter was activated by IFN-␥ treatment compared with basal activity in both cell types; however, the promoter activity induced by heat shock was significantly suppressed by IFN-␥ treatment.
Neither Tyr nor Ser Phosphorylation of STAT-1 Is Necessary for HSP27 Induction-STAT-1 and HSF-1 are activated by Tyr phosphorylation induced by IFN and stress, respectively. We determined the activation status of STAT-1 by Western blotting using specific antibodies against the phosphorylated forms of STAT-1. The phosphorylated forms of HSF-1 were detected as slower migrating bands compared with unphosphorylated HSF-1. HSF-1 was phosphorylated by heat shock but not by IFN-␥ in OSC19, OSC19-MP, and OSC19-MP expressing STAT-1 (Fig. 4A). STAT-1 was phosphorylated on the Tyr-701 residue by IFN-␥ but not by heat shock in OSC19 cells and OSC19-MP expressing STAT-1. STAT-1 that was Tyr-phosphorylated by IFN-␥ treatment translocated into the nucleus (Fig.  4C). STAT-1 phosphorylation on the Ser-727 residue occurred strongly following heat shock and weakly following IFN-␥ treatment (Fig. 4A); and the Ser-phosphorylated STAT-1 was retained in the cytosol but did not translocate into the nucleus (Fig. 4C). Immunoprecipitation analysis using an anti-HSF-1 antibody revealed that STAT-1 interacted with HSF-1 in OSC19 cells (Fig. 4B). However, neither Tyr-nor Ser-phosphorylated STAT-1 was observed in the precipitates obtained with the anti-HSF-1 antibody. The interaction with STAT-1 was not affected by heat shock (and therefore by phosphorylation of HSF-1) or IFN-␥ treatment (Fig. 4B). Furthermore, Western blotting of nuclear extracts showed that the phosphorylated HSF-1 clearly translocated into the nucleus in OSC19 and OSC19-MP cells (Fig. 4, C and D). Interestingly, unphosphorylated STAT-1 co-translocated with phosphorylated HSF-1 into the nucleus in OSC19 cells. These lines of evidence suggest that STAT-1, but neither the Tyr-nor the Ser-phosphorylated form, is necessary for the transcriptional induction of HSP27 by stress. We then examined whether phosphorylation of STAT-1 would affect the induction of thermotolerance by using STAT-1-unphosphorylatable mutants. Expression plasmids encoding wild-type STAT-1, the unphosphorylatable dominantnegative mutants Y701F and S727A, or the constitutively active mutant Y710E were transfected into OSC19 and OSC19-MP. Induction of thermotolerance in OSC19 cells was not affected by expression of exogenous wild-type STAT-1, the unphosphorylatable dominant negative mutants, or the constitutively active mutant (Fig. 5, upper panel). Protein levels of HSP27 and HSP70 were increased 2.2-3.5-fold and 1.1-1.3fold, respectively, after recovery from the AZC stress. OSC19-MP cells transfected with a control plasmid did not display thermotolerance or HSP27 induction during recovery culture from stress. However, overexpression of the unphosphorylatable STAT-1 mutants in OSC19-MP cells restored upregulation of HSP27 (1.9 -2.8-fold compared with untreated cells) and thermotolerance after recovery from AZC stress to a similar extent as OSC19 cells and wild-type STAT-1-overex- FIG. 5. Induction of thermotolerance in OSC19 and OSC19-MP cells transfected with expression vectors encoding wild-type STAT-1 or STAT-1 Tyr-701 and Ser-727 unphosphorylatable mutants. Plasmids expressing wild-type STAT-1 (wt), STAT-1-Y701F, STAT-1-Y701E, or STAT-1-S727A (0.3 or 1 g/10 5 cells), or control plasmids (mock) were transfected into OSC19 and OSC19-MP cells. The STAT-1 plasmid and its mutants carry FLAG tags. After transfection for 24 h, cells were treated with 10 mM AZC for 2 h and then cultured in normal medium for 5 h. The treated cells were treated with 5 g/ml camptothecin for 16 h. Cell viability was determined by gentian violet dye staining. The HSP protein levels were determined by Western blotting. The expression of STAT-1 and its mutants was determined by Western blotting using an anti-FLAG tag antibody. Actin levels were determined as a control for protein loading.
pressing OSC19-MP cells (Fig. 5, lower panel). On the other hand, HSP70 upregulated by stress in OSC19-MP cells transfected with plasmid containing STAT-1 and its unphosphorylatable mutants (1.3-1.7-fold) did not show a significant difference compared with cells transfected with the control plasmid (1.4-fold). These findings suggest that phosphorylation on neither the Tyr nor Ser residues of STAT-1 is necessary for the induction of HSP27 as a stress response. DISCUSSION In the present study, we found that oral squamous carcinoma cells infected with mumps virus lacked thermotolerance. Mild, namely nonlethal, stress induces HSPs, which act as cell death suppressors. The cells exposed to mild stress become resistant to more severe, potentially lethal stress (29 -32). HSP70 and HSP27 are well characterized as cell death suppressors (34,(47)(48)(49)(50). The mumps virus-infected cells were able to induce HSP70 during recovery from stress but were unable to induce HSP27. This indicates that transcriptional regulation of HSP27 is different from that of HSP70. Expression of these HSPs is commonly transcriptionally regulated by HSFs, especially HSF-1 (51)(52)(53)(54). HSF-1 is activated, through phosphorylation, by extracellular stress. Phosphorylated HSF-1 forms trimers, translocates to the nucleus, and binds to conserved regulatory sequences known as heat shock elements (HSEs) on the promoter region of HSP genes (53,54). In addition, an imperfect estrogen response element has also been identified in the hsp27 promoter (51,52). We found that transcriptional activation of HSP27 was dependent on STAT-1. HSP27 was not induced by stress in mumps virus-infected cells, which have a defect in the STAT-1 protein, and exogenous expression of STAT-1 restored the inducibility of HSP27. STAT-1 is an essential component of a transcriptional activator complex that is induced by both IFN-␣ and IFN-␥ (10,11). STAT-1 is activated by phosphorylation on its Tyr-701 and/or Ser-727 residues. Tyr-701 is phosphorylated by Jak1, which is associated with the cytoplasmic domain of IFN receptors, in response to IFN stimulation (10,11). Ser is phosphorylated mainly by p38 mitogen-activated protein kinase (55,56). We observed that the Ser residue, but not the Tyr residue, was phosphorylated after heat shock and chemical stress. However, data from the transfection of STAT-1 unphosphorylatable mutants, including dominant-negative and constitutively active mutants, indicated that phosphorylation of STAT-1 is not required for the transcriptional activation of HSP27. How then does STAT-1 work to increase HSP27 expression? One possibility is that STAT-1 is a direct transcriptional activator that is associated with an un-known element of the hsp27 promoter. However, the STAT-1 phosphorylation status of both Tyr and Ser seems to be unrelated to the up-regulation of HSP27. It is possible that another activation mechanism of STAT-1, such as methylation on Arg (57), contributes to the stress response. The most probable mechanism is that the activated HSF-1 and STAT-1 complex binds to the HSE(s) of the hsp27 promoter (Fig. 6). In contrast, this complex is likely not to be required for activation of hsp70 promoter. Supporting this idea, STAT-1 associates with HSF-1 in cells, and Tyr-or Ser-unphosphorylated STAT-1 cotranslocates with phosphorylated HSF-1. However, we have tried and failed in observing the interaction of STAT-1 with HSF-1 bound to HSE, for example, by electrophoretic mobility shift assay (data not shown). Alternatively, some negative regulators that are usually kept in check by STAT-1 could be predominantly activated in mumps virus-infected cells in the absence of STAT-1.
The relationship between HSF-1 and STAT-1 is very interesting because it indicates cross-talk between the stress response and cytokine signaling. However, only a few studies describing such a relationship have been reported. Stephanou et al. (58,59) reported that the HSEs of HSP70 and HSP90␤ share a STAT-binding motif, so that STAT-1 synergistically activates the HSF-1-dependent induction of HSP70 and HSP90␤ and STAT-3 antagonizes it (60,58). They used IFN-␥ for STAT-1 activation (58) and interleukin-6 for STAT-3 activation (60), and clearly showed the contribution of STAT-1 to this system by using 2fTGH cells and STAT-1-deficient U3A mutant cell (58). They also indicated that STAT-1 and HSF-1 directly interact in Hep-G2 cells (58). We also have observed an interaction between STAT-1 and HSF-1 in OSC19 cells (Fig. 4). In addition, the hsp70 promoter was less efficiently activated by heat shock in STAT-1-deficient U3A cells compared with 2fTGH cells (Fig. 3). The STAT-1 could have some roles in not only in HSP27 induction but also HSP70 induction. However, we have not observed a synergistic effect of IFN-␥ treatment and stress exposure on HSP70 expression, despite our use of several cell types and several techniques such as Western blotting (data not shown) and reporter gene assay (Fig. 3). Conversely, the stress induction of HSP70 tended to be suppressed by treatment with IFN-␥. The reason for this contradiction is not clear. From our results, the interaction between HSF-1 and STAT-1 is important for the stress-related induction of HSP27 but not HSP70. Furthermore, IFN-␥ does not affect induction of HSPs by stress exposure. This is supported by the observation that the contribution of STAT-1 to HSP27 FIG. 6. Schematic diagram of the effect of mumps virus infection on IFN signaling and transcriptional activation of HSP27. Activated HSF-1 and unphosphorylated STAT-1 are required for HSP27 transcriptional activation. Mumps virus V protein enhances degradation of STAT-1 in a proteasome-dependent manner (21). Because of the degradation of STAT-1, mumps virus-infected cells lack both IFN signal transduction and HSP27dependent thermotolerance (this study). expression does not require its phosphorylation on either the Tyr or Ser residues.
Host cells produce various immune responses to prevent viral replication and dissemination. Activation of the IFN system is one such major response. Conversely, the virus tries to suppress the host IFN system (8,9). The most effective route to global suppression would be to shut off the IFN signal transduction pathway through adjacent IFN receptors, such as the JAK/STAT pathway, rather than to inhibit effector molecules such as 2Ј,5Ј-oligoadenylate synthetase and double-stranded RNA-dependent protein kinase, which are induced by IFN. For example, the measles virus suppresses IFN-␣-induced Jak1 phosphorylation but not IFN-␥-induced Tyr phosphorylation (18), which seems to have a mild effect on host cells. In contrast, the accelerated degradation of STAT1 proteins by the mumps virus seems to have a very severe effect to host cells, causing almost complete suppression of both IFN-␣ and IFN-␥ signal transduction. Through this strategy, the mumps virus escapes from an IFN-induced antiviral state and apoptosis; in response, the host cell fails to induce thermotolerance, namely resistant to stress (Fig. 6). The lack of thermotolerance could be a by-product of mumps virus infection, or it could be a viral strategy for establishing infection by suppressing the stress response. However, we propose that host cells severely affected by mumps virus should be eliminated easily by apoptosis induced by fever and other stresses in infectious diseases. Therefore, a lack of thermotolerance can be regarded as a host defense mechanism for viral infection.