The major vault protein (MVP) is the main component of vaults, large ribonucleoprotein particles implicated in the regulation of cellular signaling cascades and multidrug resistance. Here, we identify MVP as an interferon γ (IFN-γ)-inducible protein. Treatment with IFN-γ resulted in a significant upregulation of MVP promoter activity as well as mRNA and protein levels. Activation of MVP expression by IFN-γ involved transcriptional upregulation through the JAK/STAT pathway based on an interaction of STAT1 with an interferon-γ-activated site (GAS) within the proximal MVP promoter. Mutation of this site distinctly reduced basal as well as IFN-γ-stimulated MVP transcription. IFN-γ also significantly enhanced the translation rate of MVP. Ectopic MVP overexpression in the MVP-negative lung cancer cell model H65 led to a downregulation of three known IFN-γ-regulated genes, namely ICAM-1, CD13 and CD36. Additionally, presence of MVP in H65 cells blocked both basal and IFN-γ-induced ICAM-1 expression whereas downmodulation of endogenous MVP levels by shRNA enhanced IFN-γ-induced ICAM-1 expression in U373 glioblastoma cells. MVP-mediated IFN-γ insensitivity was accompanied by significantly reduced STAT1 phosphorylation at Y701 and diminished translocation of STAT1 into the nucleus. Summarizing, we identify MVP as an IFN-γ-responsive gene interfering with IFN-γ-activated JAK/STAT signals. These data further substantiate that the vault particle functions as a general interaction platform for cellular signaling cascades.
The 110 kDa major vault protein (MVP) represents the main component of evolutionary highly conserved ribonucleoprotein particles called vaults and accounts for 70% of the total particle mass (Mossink et al., 2003). Vaults are barrel-shaped structures with a hollow interior. They are composed of multiple copies of three proteins and small, untranslated vault RNAs. Vaults were found in species as diverse as the slime mold Dictyostelium discoideum, electric ray, sea urchin and mammals, pointing towards essential cellular functions of vaults especially in multicellular organisms. The localization of vaults in mammalian cells is mainly cytoplasmic, however, a small portion was also found to be associated with the nuclear membrane (Chugani et al., 1993). Vaults are present in a wide range of cell types at numbers of over 104 particles per cell (Kickhoefer et al., 1998).
Although the precise function of this highly conserved structure has yet to be determined, vault morphology suggests a role in transport and sequestration processes of xenobiotics and/or cellular molecules (Suprenant, 2002). Vaults are generally present at notably high levels in tissues chronically exposed to xenobiotics, and MVP was also found to be overexpressed in certain tumor types when compared with their normal counterpart (Berger et al., 2001; Izquierdo et al., 1996; Zurita et al., 2003). For example, low levels of MVP were detected in normal human astrocytes, whereas all astrocytic brain tumors, including low-grade astrocytomas, expressed MVP/vaults at very high levels (Berger et al., 2001). Furthermore, it has been demonstrated that the tumor suppressor phosphatase and tensin homolog deleted on chromosome 10 (PTEN) binds to MVP, and that a significant quantity of endogenous PTEN associates with vault particles in HeLa cells, thus suggesting a possible regulatory role for MVP in the phosphoinositide 3-kinase/Akt (PI 3-kinase/Akt) signaling pathway (Yu et al., 2002). Recently, MVP has been shown to function as a scaffold protein for both SHP-2 and ERK, thus regulating the activation of the Ras-MAP kinase pathway (Kolli et al., 2004). Altogether, vaults have been linked to several processes involved in the malignant phenotype of cancer cells like multidrug resistance, cell motility and malignant transformation (reviewed in Suprenant, 2002; van Zon et al., 2003).
To date, only limited information is available regarding the molecular mechanisms regulating MVP expression in normal and tumor cells. MVP is upregulated under several cellular conditions including malignant transformation (Berger et al., 2001; Meijer et al., 1999), acquisition of drug resistance (Kickhoefer et al., 1998), and differentiation of dendritic cells (Schroeijers et al., 2002). MVP expression has been shown to be induced by differentiating agents including phorbol 12-myristate 13-acetate (Komarov et al., 1998), histone deacetylase (HDAC) inhibitors (Steiner et al., 2004), as well as cytotoxic drugs (Berger et al., 2000; Komarov et al., 1998; Stein et al., 2005). Recently, regulatory roles of chromatin remodeling (Emre et al., 2004) as well as alternative splicing (Holzmann et al., 2001) have been suggested. An initial analysis of the human MVP gene revealed a TATA-less promoter, which also lacks other core-promoter elements but harbors several putative transcription factor binding sites, including an inverted CCAAT box, a p53-binding site, and a GC box element (Lange et al., 2000).
In a recent report, we have identified an essential role of SP transcription factors in basal- and HDAC-inhibitor-induced MVP expression (Steiner et al., 2004). In this study, we show that IFN-γ enhances MVP expression through transcriptional and post-transcriptional mechanisms. Additionally, our data indicate for the first time a role of MVP/vaults in the regulation of IFN-γ-induced gene expression.
Mutation of a putative STAT-binding site in the promoter of the human MVP gene diminishes basal MVP promoter activity
During our analysis of the promoter sequence of the human MVP gene for possible transcription-factor-binding sites (MatInspector software), we identified a putative STAT-binding site at position –33 to –25 bp relative to the TSP (Steiner et al., 2004). This site displayed 100% homology to the core STAT consensus sequence and resembles an IFN-γ-activated site (GAS) (Fig. 1). To investigate the involvement of this STAT-binding site in human MVP gene promoter activity, two luciferase reporter gene constructs, one comprising the full-length MVP promoter (pMVP1670) and one with a 5′-end deletion of the promoter (pMVP124), were mutated in their STAT-binding sites (pMVP1670-STATmut and pMVP124-STATmut, respectively). Mutations of this site resulted in significantly decreased promoter activities of pMVP1670-STATmut and pMVP124-STATmut by 50% and 75%, respectively (P<0.01 and P<0.001, respectively) compared with the wild-type counterparts (Fig. 1). These findings suggest an involvement of the investigated GAS element in the constitutive transcriptional activation of the human MVP gene.
IFN-γ enhances promoter activity of the human MVP gene through a GAS element binding STAT1
IFN-γ is a cytokine that transmits its signals predominantly through the binding of STAT1 homodimers to GAS elements within the promoter region of target genes (Levy and Darnell, 2002). Thus, the presence of a GAS element in the basal promoter region of the human MVP gene prompted us to investigate the involvement of IFN-γ in the regulation of MVP expression. The human cancer cell line Hep3B was transiently transfected with wild-type MVP (pMVP1670 and pMVP124) and mutated MVP (p1670-STATmut and pMVP124-STATmut) promoter constructs as well as the empty control vector pGL3-Basic. As shown in Fig. 2A, IFN-γ treatment (6 hours) enhanced wild-type MVP promoter activity up to 1.4-fold (P<0.01), whereas this stimulating effect was significantly diminished (P<0.05) when using MVP promoter constructs that contained the mutated STAT-binding site. By contrast, the vector control pGL3-Basic showed no IFN-γ response. Since the full-length (pMVP1670) and the 5′-end-deletion (pMVP124) promoter construct responded similarly to IFN-γ, the GAS element at position –33 to –25 bp relative to the TSP is suggested as the key promoter element transmitting the IFN-γ response of the human MVP gene. To further investigate whether STAT1 indeed interacts with the proximal MVP promoter in vitro and in vivo, we performed EMSA and ChIP assays, respectively. In Fig. 2B a representative EMSA experiment is shown comparing a high-affinity STAT1-binding sis-inducible-element (SIE) site with the respective GAS site of the MVP promoter (STAT-wt). STAT1 was specifically binding to the STAT-wt oligonucleotide, although at distinctly reduced affinity compared with the high-affinity SIE consensus site. Specificity of binding was demonstrated by the IFN-γ-dependent appearance of the respective DNA-protein complex and the efficient competition of the unlabeled STAT-wt but also SIE oligonucleotide. Furthermore, addition of a STAT1 antibody clearly reduced the respective electrophoretic mobility shift assay (EMSA) band for both oligonucleotides. When using a mutated STAT1 sequence (STAT-mut), no band shift was observed. In addition to STAT1, two further DNA-protein complexes were detected for the STAT-wt oligonucleotide. Both bands were specific, as demonstrated by competition of the unlabeled oligonucleotide, but their appearance did not depend on pretreatment with IFN-γ and was not influenced by the presence of the STAT1 antibody.
To demonstrate that STAT1 interacts also in vivo with the MVP gene promoter following IFN-γ treatment, we performed chromatin immunoprecipitation (ChIP) analysis. Fig. 2C illustrates that immunoprecipitations (IP) of crosslinked DNA-protein complexes derived from IFN-γ-treated Hep3B cells with a STAT1 antibody contained the respective core-promoter region of the human MVP gene.
IFN-γ treatment enhances endogenous MVP mRNA and protein levels by modulation of STAT1 signals
Northern blots were performed to determine whether IFN-γ-stimulated promoter activity also leads to enhanced levels of MVP mRNA. As shown representatively in Fig. 3A, treatment of Hep3B cells with IFN-γ (250 IU/ml) significantly increased the amount of MVP mRNA. At 6 hours of IFN-γ treatment the MVP mRNA level started to rise, peaking at 24 hours of IFN-γ exposure with a threefold increase in mRNA.
The impact of IFN-γ on MVP expression on the protein level was analyzed by western blot and compared with ICAM-1, a known IFN-γ-inducible protein, as a positive control for the IFN-γ response. A range of human cancer cell lines with differing basal MVP levels were tested for MVP expression following 24 hours and 48 hours of IFN-γ treatment. Hep3B, HeLa, KB-3-1 and THP-1 cells, all cell lines with relatively low basal MVP expression, showed up to 11-fold increased MVP levels after IFN-γ treatment (Fig. 3B,C). Since the amount of MVP also increases when cells reach confluence (data not shown), controls untreated for 24 hours and 48 hours were included in these western blot experiments. Cell density also explains the rather strong difference in basal MVP expression between the 24-hour- and 48-hour-control samples of Hep3B cells. In the three cell lines A549, F2000 and T98G, however, the already high basal MVP expression level did not further increase in response to IFN-γ treatment. The efficacy of IFN-γ signaling, however, was confirmed in parallel by upregulation of ICAM-1 (data not shown).
To further prove that the upregulation of MVP expression by IFN-γ involves Janus kinase (JAK)/STAT-mediated signals, IFN-γ treatment was performed in the presence of the JAK/STAT inhibitor curcumin and the JAK2 inhibitor AG490. Both inhibitors suppressed dose-dependently the stimulating effect of IFN-γ on MVP expression in Hep3B cells (Fig. 3D). The inhibitory activity on MVP expression was comparable to the one on ICAM-1, which is known to be activated by IFN-γ through STAT1 (Look et al., 1994). Additionally, Hep3B cells were transiently transfected with a DN-STAT1 expression construct by lipofection which led to a transfection rate of about 60%. As shown in Fig. 3E, presence of DN-STAT1 distinctly inhibited basal and to a lesser extent IFN-γ-induced MVP expression. The weaker effect in case of IFN-γ treatment might be explained by a pronounced stimulation of endogenous STAT1 expression by IFN-γ (Fig. 3E).
IFN-γ increases the translation rate of MVP but does not enhance protein stability
The six- to 11-fold increase of MVP cannot solely be explained by a 1.4-fold increased promoter activity and threefold increase in mRNA expression following IFN-γ treatment (compare Figs 1, 2, 3). Thus, pulse and pulse-chase experiments were performed to investigate the impact of IFN-γ on the translation rate and protein stability of MVP, respectively. Immunoprecipitation of MVP after a short [35S]methionine pulse resulted in a marked increase of newly synthesized, labeled MVP following IFN-γ treatment, with a peak at around 15 hours of IFN-γ pre-treatment (Fig. 4A, left panel). For a control, equal amounts of the immunoprecipitation supernatant were subjected to SDS-PAGE and stained with Coomassie Blue (Fig. 4A, middle panel), followed by exposure to X-ray film (Fig. 4A, upper panel). This indicated that the general [35S]methionine incorporation diminished in response to IFN-γ. When normalizing labeled MVP to the overall 35S-incorporation, 15 hours pre-treatment with IFN-γ led to an eightfold increase in the de-novo synthesis of MVP compared with untreated cells (Fig. 4A, right panel). To test whether the IFN-γ-mediated increase in the MVP translation rate was reflected on the total MVP level, in parallel, lysates of Hep3B cells treated for increasing time periods with IFN-γ were blotted and probed for MVP, ICAM-1 and β-actin (Fig. 4B). The level of IFN-γ-induced MVP increased at treatment for up to 24 hours and is comparable to the ICAM-1 induction. This strong upregulation of MVP at 24 hours IFN-γ treatment corresponds well to the peak in MVP translation at treatment for 15 hours (Fig. 4A).
Since the stimulation of MVP translation peaked at 15 hours IFN-γ treatment, this time-point was used for 2.5 hours pulse-labeling followed by 24-hour chase experiments (Fig. 5). Corresponding to published results (Zheng et al., 2005), our data suggested a rather long half-life of MVP with virtually no reduction of labeled protein after a 24-hour chase in untreated control cells. At time point zero of the chase, a higher amount of newly synthesized MVP could be detected in IFN-γ-treated cells confirming an enhanced MVP translation rate. Interestingly, IFN-γ treatment reduced the stability of the newly synthesized MVP, which was obvious from a diminished MVP stimulation ratio at 24-hour-compared with 12-hour-chase (Fig. 5, right panel).
IFN-γ-induced MVP assembles into intact vault particles
Vault particles typically pellet at 100,000 g and in sucrose-equilibrium gradient centrifugation intact vault particles can be found in the 45% and/or 50% sucrose fractions (Kickhoefer et al., 1998). To investigate whether IFN-γ-induced MVP assembles into vaults, we fractionated cell lysates from IFN-γ treated and untreated cells into a 100,000 g particle fraction (P) and a high-speed supernatant soluble fraction (S). The absence of a visible MVP band in the soluble fraction of 24-hour IFN-γ-treated Hep3B and KB-3-1 cells (Fig. 6A) as well as localization of the major part of MVP to the 45% sucrose fraction (data not shown), together with the typical granular staining pattern of vaults in IFN-γ-treated KB-3-1 cells (Fig. 6B), proved that IFN-γ-induced MVP completely incorporates into intact vault particles.
Effects of MVP on the expression of IFN-γ-regulated proteins
Next we established stable MVP-overexpressing cell clones of the small-cell lung cancer cell line H65, which completely lacks endogenous MVP. FACS analysis of several MVP-positive and -negative control clones revealed that three of the multiple investigated cell-surface markers were constantly and distinctly changed (downregulated) due to MVP expression, i.e. the cell adhesion molecule ICAM-1, the ectopeptidase CD13 (aminopeptidase N) and the thrombospondin receptor CD36 (Fig. 7A,B). All three proteins are known to be induced by IFN-γ (Look et al., 1994; Swerlick et al., 1992; van der Velden et al., 1998). Subsequent immunoblot analysis of ICAM-1 in several MVP-overexpressing cell clones confirmed the initial FACS experiments. Incubation of vector control clones with IFN-γ led to a distinct activation of ICAM-1 expression, whereas all MVP-positive clones were still devoid of detectable ICAM-1 expression following 24 hours of IFN-γ exposure (Fig. 7C). To investigate whether the MVP-mediated loss of IFN-γ-induced ICAM-1 expression is caused by downregulated expression of ICAM-1 or reduced protein stability, ICAM-1 mRNA levels were determined by reverse transcriptase (RT)-PCR. All MVP-overexpressing clones had lost detectable ICAM-1 mRNA expression (Fig. 7D).
To investigate whether endogenously expressed MVP also had an influence on ICAM-1 expression, we developed adenoviruses that expressed short hairpin RNA (shRNA) to knock-down MVP expression. In several cell models, MVP-shRNA presence for >72 hours led to a distinct downmodulation of MVP expression as representatively shown for the glioblastoma cell line U373 (Fig. 7E). Inhibition of MVP expression led to a distinctly enhanced ICAM-1 induction by IFN-γ, suggesting that also endogenously expressed MVP is involved in the regulation of IFN-γ-mediated signals.
Impact of MVP overexpression on STAT1 signaling
Since induction of ICAM-1 expression by IFN-γ is mediated through the JAK/STAT pathway leading to translocation of STAT1 into the nucleus (Tessitore et al., 1998), we determined the impact of MVP on the cellular localization of STAT1. As detected by western blot analysis and immunofluorescence staining (Fig. 8A and B, respectively), the translocation of STAT1 to the nucleus in response to IFN-γ was substantially reduced in MVP-expressing H65 cells. In search for factors that limit the nuclear translocation of STAT1, phosphorylation of STAT1 at Y701 – which is essential for the IFN-γ-induced translocation into the nucleus (Schroder et al., 2004) – was determined. A representative immunoblot is shown in Fig. 8C, illustrating that the presence of MVP in H65 cells drastically reduced phosphorylation of STAT1 at Y701 in response to IFN-γ. A similarly attenuating effect was also observed in the MVP-negative human melanoma cell line VM-48 when stably transfected with an EGFP-MVP expression construct (Fig. 8D).
Vaults and their major component MVP have been implicated in a variety of cellular processes including drug resistance, subcellular transport and cell differentiation (Mossink et al., 2003; Suprenant, 2002) as well as regulation of intracellular signaling cascades including the PI 3-kinase (Yu et al., 2002) and the MAP kinase pathway (Kolli et al., 2004). Here, we identify MVP as an IFN-γ-upregulated gene and present profound evidence that MVP interferes with IFN-γ-mediated STAT signal transduction. Herewith, the IFN-γ-induced JAK/STAT pathway is the third intracellular phosphorylation cascade demonstrated to be regulated by MVP/vaults. These observations suggest that the complex vault ribonucleoprotein particle represents a versatile platform for the intracellular regulation of multiple signal pathways transmitted from the cell surface to the nucleus.
In a previous study, we have identified Sp-family members as important transcription factors activating the MVP promoter by binding to a conserved GC-box element in the proximal promoter region (Steiner et al., 2004). In this study, we set out to further analyze the proximal MVP promoter region, which harbors the main promoting activity. In silico analysis identified a putative STAT-binding site that strongly resembles a GAS (IFN-γ-activated site) element, which is known to bind preferentially STAT1 homodimers (Schroder et al., 2004). Using luciferase-reporter constructs, we demonstrated that disruption of this STAT-binding site strongly reduced basal promoter activity, suggesting a role of JAK/STAT signals in activating MVP expression. The fact that constitutive activation of STAT signals is observed during several forms of malignant transformation by oncogenes, cytokines and tumor viruses (Takeda and Akira, 2000) indicates that STAT-dependent signals might contribute to the frequent upregulation of MVP in diverse human malignancies (Berger et al., 2001; Meijer et al., 1999).
One type of cytokine that transmits its signals mainly through GAS elements within the promoters of target genes is the type-II interferon IFN-γ (Sato et al., 2001). Consequently, we demonstrate that MVP is encoded by an IFN-γ-stimulated gene activated through the GAS site in the very proximal MVP promoter sequence. Interferons are a family of multifunctional cytokines that activate target genes and mediate antiviral, antiproliferative and immunomodulatory activities (for reviews, see Levy and Darnell, 2002; Sato et al., 2001). Whereas promoters that contain GAS elements (like that present in the MVP promoter) are preferentially activated by IFN-γ, type-I interferons induce transcription of target genes through interferon-stimulated response elements (ISRE) (Lau and Horvath, 2002). Correspondingly, IFN-α showed only a weak activating effect on the MVP promoter (data not shown). This corroborates oligonucleotide array data (Der et al., 1998), which demonstrated stimulation of MVP mRNA expression in a human fibrosarcoma cell line by IFN-γ but not IFN-α. Interaction of STAT1 with the respective site in the MVP promoter was demonstrated both in vitro and in vivo by EMSA and ChIP assays, respectively. It has to be mentioned, however, that the interaction was relatively weak when compared with the control SIE consensus site. This corresponds to the relatively weak (up to 1.4-fold) but significant stimulation of the MVP promoter by IFN-γ.
Further proof for an involvement of the JAK/STAT pathway in the activation of MVP expression came from experiments with pharmacological inhibitors. Curcumin, a known inhibitor of STAT-mediated transcription, blocked IFN-γ-induced MVP expression. Curcumoids are natural products of the Indian spice turmeric, and potent antioxidant and anti-inflammatory agents by suppressing pro-inflammatory cytokines, including TNF-α, IL-1β and NOs. One molecular mechanism underlying the anti-inflammatory activity of curcumin is the inhibition of JAK1/2 phosphorylation by activating the JAK-inhibiting phosphatase SHP-2 (Kim et al., 2003). Additionally to curcumin, also the more specific JAK2 inhibitor tyrphostin AG490 (De Vos et al., 2000) dose-dependently suppressed IFN-γ-induced MVP expression similarly to that of the well-known JAK/STAT target-gene ICAM-1. The data derived from pharmacological inhibitors were further corroborated by transient transfection experiments with a DN-STAT1 expression construct (Bromberg et al., 1998). Presence of DN-STAT1 substantially reduced basal and, in part, also IFN-γ-stimulated MVP expression. The less pronounced effect in case of IFN-γ stimulation is probably owing to a distinct stimulation of endogenous wild-type STAT1 expression by IFN-γ, partly overruling the dominant-negative effects of the transgene.
Whereas IFN-γ enhanced the MVP promoter activity significantly but relatively modestly, the increase at the protein level was distinct (up to 11-fold). This suggests that enhanced transcription can not solely account for the stimulation of MVP by IFN-γ. Indeed, pulse and pulse-chase experiments and comparison with respective MVP mRNA levels demonstrated that IFN-γ does stimulate the MVP translation rate in a time-dependent and transient manner. An impact on translation and/or stabilization of target proteins is not unexpected because interferons were shown to regulate multiple genes involved in regulation of translation and protein degradation (de Veer et al., 2001).
In contrast to enhanced MVP transcription and translation we found evidence for reduced MVP stability in IFN-γ-treated cells. Compared with the relatively long half-life of endogenous MVP – recently reported to be at least three days (Zheng et al., 2005) – in pulse-chase experiments the amount of [35S]methionine-labeled MVP was higher at a 12-hour than a 24-hour chase in the case of IFN-γ-treated Hep3B cells. Therefore, the half-life of the newly translated MVP protein might be limited by delayed association into vaults. However, cell fractionation and immunofluorescene experiments demonstrated a complete, IFN-γ-independent incorporation of MVP into vault particles within 24 hours. Thus, the shorter half-life of IFN-γ-induced MVP protein might rather reflect a general feedback-mechanism controlling upper levels of MVP expression in human cells. Correspondingly, we found no IFN-γ-mediated upregulation of MVP in cell types that already express high endogenous levels of MVP. Comparable observations have been reported in the case of TSA- and drug-selection-stimulated MVP overexpression (Emre et al., 2004; Kickhoefer et al., 1998). The underlying mechanisms regulating MVP stability are unknown. However, very recently an active mechanism of MVP degradation by the proteasome has been described in mammalian oocytes and zygotes (Sutovsky et al., 2005).
The observation that IFN-γ activates MVP on the levels of transcription and translation, together with the recent report (Kolli et al., 2004) that MVP interacts with the JAK/STAT inhibitory tyrosine phosphatase SHP-2 (You et al., 1999), led us to the hypothesis that MVP might play an autoregulatory role in IFN-γ response signals. This hypothesis was corroborated by data derived from the analysis of MVP-transfected cell clones of the MVP-negative small-cell lung cancer cell line H65. When screening for changes in the expression of cell-surface proteins by FACS analysis, we found three proteins whose levels were distinctly changed (all downregulated) in MVP-positive clones: ICAM-1, CD13 and CD36. All three proteins are encoded by genes whose expression is stimulated by IFN-γ (Look et al., 1994; Swerlick et al., 1992; van der Velden et al., 1998). In particular, ICAM-1 has been very well characterized as an IFN-γ-induced protein that is upregulated by binding of STAT1 homodimers to GAS elements within the ICAM-1 promoter (Tessitore et al., 1998). Thus, we focused our experiments on the impact of MVP expression on basal and IFN-γ-stimulated ICAM-1 expression. Notably, ICAM-1 expression was completely blocked by MVP overexpression even after IFN-γ stimulation. To clarify whether endogenously expressed MVP also interferes with IFN-γ-mediated signals we knocked-down MVP expression with adenoviruses that express shRNA. In agreement with the data from our overexpression experiments, cell cultures with distinctly reduced MVP levels displayed a significantly enhanced IFN-γ-mediated ICAM-1 expression. Taken together, these data strongly suggest that MVP is involved in the regulation of IFN-γ-mediated gene expression by blocking essential steps in the IFN-γ-stimulated JAK/STAT signals.
Searching for the underlying mechanisms, we found a reduced translocation of STAT1 to the nucleus and a distinct reduction of phosphorylation of STAT1 at Y701 in response to IFN-γ in all MVP-positive H65 cell clones. A similar inhibition of STAT1 phosphorylation by MVP overexpression was also detectable in the human melanoma cell line VM-48 but not the brain tumor cell line KMYH (data not shown). Phosphorylation at Y701 is essential for IFN-γ-mediated gene expression because it allows dissociation of the STAT1 homodimer from the receptor, followed by nuclear translocation and binding to the respective GAS in the promoters of target genes (Schroder et al., 2004). This suggests that MVP activation by IFN-γ might be part of an autoregulatory, inhibitory feedback-loop in H65 and VM-48 cells.
MVP knock-out mice have no major abnormalities, suggesting that MVP is dispensable for normal development (Mossink et al., 2002). However, MVP-negative MEFs are hypersensitive against growth-factor starvation (Kolli et al., 2004), which corresponds with a growth defect of Dictyostelium following MVP gene destruction (Vasu and Rome, 1995). These observations suggest that MVP is generally engaged in supporting cell survival and anti-apoptotic signals. Correspondingly, MVP is upregulated during malignant progression and was shown to correlate with an unfavourable prognosis in several tumor types, suggesting proto-oncogenic function of vaults (Berger et al., 2001; Meijer et al., 1999; Zurita et al., 2003). Our observations add a completely new aspect to the proposed function of MVP as an oncogene, namely the block of IFN-γ-mediated JAK/STAT signals, which are known to protect against tumor development, inhibit proliferation of a variety of tumor cells and activate several pro-apoptotic proteins (Calo et al., 2003).
The exact molecular base, how MVP interferes with STAT1 phosphorylation inhibition, needs to be determined in ongoing studies. The fact that we observed this activity only in two out of three investigated cell models suggests the involvement of additional factors in the observed repression-mechanism, which are apparently not provided by all cell types. So far, three main negative regulators of IFN-γ signaling have been described: first, the suppressor of cytokine signaling (SOCS) family, which bind to catalytic sites of receptor kinases and JAK-family kinases, thus inhibiting STAT1-binding and -activation; second, the protein inhibitors of activated STATs (PIAS), which directly interact with STAT1-binding to DNA; third, the already mentioned tyrosine phosphatase SHP-2, which dephosphorylates members of the STAT family. Interestingly, it has been observed recently that MVP, as a scaffold protein for the ERK/MAPK pathway, is a SHP-2 substrate involved in mediating ERK- and especially ELK-1-activation by epidermal growth factor signals (Kolli et al., 2004). In contrast to the inhibitory role in the JAK/STAT signal pathway, SHP-2 activity is required for full activation of the ERK/MAPK pathway by several growth-factor-receptor tyrosine kinases (Neel et al., 2003). These data, together with our findings, suggest that MVP generally cooperates with SHP-2, leading in parallel to activation of MAPK and inhibition of JAK/STAT1 signals. The observation that also a second tyrosine phosphatase, PTEN–a major inhibitor of the PI 3-kinase pathway and a potent tumor suppressor – interacts with MVP (Yu et al., 2002) suggests that vault particles function as a general interaction platform for multiple cellular signaling cascades in the cytoplasm.
Materials and Methods
Cell culture and treatment
Tumor cell lines Hep3B (hepatocellular carcinoma), HeLa (cervix adenocarcinoma), KB-3-1 (epidermoid carcinoma), U373 (glioblastoma) and THP-1 (monocytic leukemia) were obtained from the American Type Culture Collection (Rockville, Maryland, USA) and cultured in RPMI-1640 Medium (Sigma, St Louis, MO). H65 and VM-48 cells were established from brain metastases derived from a small-cell lung cancer and a malignant melanoma, respectively, and were extensively characterized with regard to patho-histological and immunohistochemical markers. All media were supplemented with 10% fetal bovine serum (FBS) and cells were maintained at 37°C in a humidified atmosphere containing 5% CO2. Cell cultures were frequently checked for Mycoplasma contamination. Cells were treated with 250-1000 IU/ml of IFN-γ (human recombinant interferon-gamma-1b, IMUKIN, Boehringer Ingelheim, Germany) as described in the respective figure legends. Curcumin (C-1386, Sigma) and the tyrphostin AG490 [α-cyano-(3,4-dihydroxy)-N-benzylcinnamide, Alexis Biochemicals, Lausen, Switzerland] (De Vos et al., 2000) were dissolved in DMSO and added at varying concentrations 30 minutes before IFN-γ treatment.
Cloning of the MVP promoter and construction of luciferase reporter gene vectors
Cloning of the human MVP gene promoter and of luciferase reporter constructs has been described previously (Steiner et al., 2004). Briefly, the human MVP promoter was amplified with the Expand Long Template PCR System (Roche, Mannheim, Germany) using genomic DNA as template. A PCR product spanning the promoter region from –1670 bp to +47 bp relative to the transcription start point (TSP) was then cloned into the pCR2.1-TOPO vector with the TOPO TA Cloning Kit (Invitrogen, Groningen, The Netherlands) and further subcloned into the HindIII site of the pGL3-Basic vector (Promega, Madison, WI), generating pMVP1670. The 5′-end deletion construct pMVP124 was obtained by restriction digest of pMVP1670 with PstI and SacI.
To set point mutations within the predicted transcription factor binding site for STAT1 the wild-type promoter construct pMVP1670 was mutated applying the QuickChange site-directed mutagenesis kit (Stratagene, La Jolla, CA), generating pMVP1670-STATmut. The two point mutations were set within the putative STAT-binding site (–33 to –25 bp relative to the TSP) changing the sequence from TGCCGGGAA to TGCCGacAA. The mutation was confirmed by sequencing. The 5′-end deletion construct pMVP124-STATmut was generated as described above for pMVP124.
Transient transfection of human cells with 0.4 μg reporter plasmid was performed using the Lipofectamin Plus reagent (Invitrogen) according to the manufacturer's protocol. Forty-eight hours after transfection cells were lysed in 0.25 M Tris-HCl pH 7.5 with 0.5% Triton X-100 and luciferase assays were performed according to standard protocols. Protein concentration was determined with the Micro BCA protein assay reagent kit (Pierce, Rockford, IL). The protein content was used for normalisation of luciferase activity. IFN-γ (250 IU/ml) was added 6 hours before harvesting cells for the assay.
[35S]methionine incorporation and immunoprecipitation of MVP
For pulse-chase experiments 106 Hep3B cells were treated with IFN-γ (250 IU/ml) for 15 hours prior to the pulse. Before [35S]methionine incorporation cells were washed with warm PBS and starved for 15 minutes at 37°C in DMEM without L-methionine and L-cysteine (Gibco) supplemented with glutamine (2 mM) and 10% dialysed FBS (Gibco). Labeling was performed with 12.5 μl of TRAN35S-LABEL (∼10 mCi/ml; MP Biomedicals, Irvine, CA) in 2 ml of methionine-free DMEM with 10% dialyzed FBS, glutamine and 20 mM HEPES (pH 7.6) per flask and an incubation time of 2.5 hours at 37°C. After the pulse the medium was removed and cells were washed with warm RPMI-1640 medium. The chase was performed in RPMI-1640 medium supplemented with 300 mg/l L-methionine (Sigma) and 10% FBS. Cells were harvested, washed with PBS, frozen in liquid N2 and stored at –80°C until performing IP. For pulse experiments the labeling step was reduced to a duration of 0.5 hours. IFN-γ treatment was maintained throughout pulse and chase in the respective flasks.
For the IP of MVP, the cell pellet was lysed in 250 μl ice-cold modified RIPA buffer [50 mM Tris-HCl, pH 7.4, 1% NP-40, 0.25% Na-deoxycholate, 150 mM NaCl, 1 mM PMSF, 1 mM Na3VO4, 1 mM NaF, Complete protease-inhibitor-cocktail (Roche)] and incubated on ice for 15 minutes. After centrifuging the lysate for 15 minutes at 15,800 g at 4°C the supernatant fraction was diluted 1:1 with PBS and pre-cleared by adding 25 μl of a 50% slurry of Protein A Sepharose 4 Fast Flow (Amersham, Uppsala, Sweden) in PBS and incubation at 4°C for 10 minutes. The beads were removed by centrifugation at 15,800 g at 4°C for 10 minutes. Pre-cleared cell lysates were incubated with 0.5 μg of the immunoprecipitating antibody for 2 hours at 4°C. MVP was immunoprecipitated with the LRP-56 antibody (Alexis Biochemicals, Lausen Switzerland). A purified Mouse IgG1 (Ancell, Bayport, MN) was used as control antibody. The immunocomplex was captured by adding 25 μl of protein A sepharose (50% slurry) and 2 μg of rabbit anti-mouse IgG (Upstate, Lake Placid, NY) for 1 hour at 4°C. Beads were collected by pulse centrifugation, washed three times with 800 μl ice-cold modified RIPA buffer and resuspended in 80 μl sample buffer (40% glycerine, 5% 2-mercaptoethanol, 9.2% SDS, 250 mM Tris-HCl pH 6.8, Bromphenol Blue). Samples were heated to 95°C for 5 minutes and separation was performed by 8% SDS-PAGE overnight at 40 V. After 1 hour in fixing solution (10% acetic acid, 30% methanol) and 30 minutes in Enlightning enhancer solution (NEN), the gel was vacuum-dried at 78°C for 1.5 hours and exposed to X-ray film at –80°C.
Northern blot and RT-PCR
Total RNA was extracted from control and IFN-γ-treated cells by the Trizol extraction method (Invitrogen). Northern blots were performed as described previously (Steiner et al., 2004). MVP mRNA was detected using a 32P-labeled probe corresponding to a 1225 bp XbaI fragment from DKFZp434L1720 (Holzmann et al., 2001); glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and β-actin mRNA were analysed for control purposes. Northern blots were analysed with a PhosphoImager SI and quantified with the ImageQuant software (Molecular Dynamics, Sunnyvale, CA). 28s RNA was stained with ethidium bromide and visualized with a FluorImager 595 (Molecular Dynamics). The expression of ICAM-1 mRNA was detected using the primers 5′-AGAACCTTACCCTACGCTGC-3′ and 5′-CAGTATTACTGCACACGTCAGC-3′ and GAPDH as house-keeping control as described (Berger et al., 2001).
Cells grown on chamber slides in the presence or absence of IFN-γ were fixed for 12 minutes in methanol/acetone (1:1) at –20°C. Staining for MVP and STAT1 was performed with the LRP-56 (1:20) and the Stat1 p84/p91 (1:400; sc-346 X, Santa Cruz, CA) antibody, respectively, diluted in 1% bovine serum albumin (BSA) in PBS and a fluorescein isothiocyanate (FITC)-labeled anti-mouse (F-1010, Sigma) or anti-rabbit (F-9887, Sigma) antibody diluted 1:200 in 1% BSA in PBS. Slides were mounted in Vectashield mounting medium with DAPI (Vector Laboratories, Burlingame, CA). Staining was analyzed with a DMRXA Leica with epifluorescence equipment (Leica, Cambridge, UK). All digital micrographs were taken with a COHU charged coupled device camera at room temperature using a Leica HCX PL APO 40x/1.25-0.75 oil objective and Leica CW4000 FISH software version 1.2.
Cell fractionation, immunoblot analysis and flow cytometry
Cell fractionation and MVP detection by western blot were performed as described (Berger et al., 2000) using the following antibodies: LRP (BD, San Jose, CA), β-actin (A-5441, Sigma), ICAM-1 (sc-8439, Santa Cruz), p-Stat1 (Tyr 701, sc-7988) and Stat1 p84/p91 (E-23, sc-346 X). Expressions of multiple cell surface marker proteins were determined by flow cytometry as described (Aldrian et al., 2003; Steinberger et al., 2004) using the mouse monoclonal antibodies 5-92 (CD13), 5-390 (CD13), 5-271 (CD36), VIAP (isotype control) (all supplied by O. Majdic, Institute of Immunology, Vienna), and ICAM-1 (Biodesign, Saco, ME).
Electrophoretic mobility shift and chromatin immunoprecipitation assays
Electrophoretic mobility shift assay (EMSA) was performed as described previously (Steiner et al., 2004) but in a changed binding buffer (HEPES 5 mM, MgCl2 0.75 mM, EGTA 0.05 mM, glycerol 2.5%, NaCl 75 mM, Ficoll 5%, BSA 2.5 mg/ml, 0.5 mg/ml tRNA). The sequences of the oligonucleotides were: high affinity STAT1 binding site (sis-inducible element, SIE, Santa Cruz, sc-2535) 5′-GTGCATTTCCCGTAAATCTTGTCTACA-3′, STAT binding site of the MVP promoter (STAT-wt) 5′-GCCTTGCCCTGCCGGGAAGTGATCCCC-3′ and mutated STAT site (STAT-mut) 5′-GCCTTGCCCTGCCGacAAGTGATCCCC-3′. For competition experiments unlabeled oligonucleotides were added at a 100-fold molar excess to the binding reaction. For supershift experiments 2 μg of polyclonal STAT1 (E-23) gel-shift antibody (Santa Cruz) was used.
Chromatin immunoprecipitation (ChIP) assays were performed using the ChIP assay kit from Upstate according to the manufacturer's instructions. Chromatin of 106 Hep3B cells, untreated or treated with 250 IU/ml IFN-γ, was cross-linked by addition of formaldehyde to a final concentration of 1% (10 minutes, 37°C) and resuspended in 200 μl SDS lysis buffer. Each sample was sonicated at 30% intensity for 5 × 30 seconds bursts using a Bandelin sonoplus sonicator (Berlin, Germany) with a 2 mm tip. For immunoprecipitation, 0.5 μg of rabbit polyclonal Stat1 p84/p91 antibody (sc-346 X) was employed. PCR was performed using the forward primer 5′-TTAACTCCCAAGCCCCACCCCTGGGCTT-3′ and the reverse primer 5′-GAGAAGCTTCTAGAAGTGCAGGTAGC-3′ obtaining a product of 126bp. PCR conditions included one cycle (5 minutes at 95°C) and 30 cycles (30 seconds at 94°C, 30 seconds at 56°C, and 45 seconds at 72°C) with a final extension step (8 minutes at 72°C).
MVP expression, shRNA constructs and cell transfection
Stable MVP-overexpressing cell clones were established using the construct S-MVP-3 and a puromycin resistance vector as described previously (Holzmann et al., 2001). Transient transfection with a dominant-negative (DN)-STAT1 (Bromberg et al., 1998), lacking the terminal tyrosine 701, was done with Lipofectamine plus reagent (Invitrogen) following the manufacturer's instructions. The DN-STAT1 construct was gratefully obtained from J. E. Darnell Jr (Rockefeller University, New York, NY). To generate the pEGFP-MVP expression plasmid, the MVP cDNA was obtained from S-MVP-3, subcloned into the EcoRI and HindIII sites of pBluescript SK(+/–) (Stratagene) to gain the appropriate sites for inserting MVP in frame into pEGFP-C1 (Clontech) by using BglII and HindIII. Thus pEGFP-MVP encodes a chimeric protein of MVP and EGFP at the N-terminus. Establishment of EGFP-MVP expressing VM-48 cell clones was done by electroporation followed by selection in G418 at 500 μg/ml (Sigma). shRNA expression constructs directed against human MVP and Renilla luciferase as a non-expressed control gene were generated and cloned into recombinant adenoviruses through the Gateway cloning system (Invitrogen). The respective oligonucleotides for construction of MVP shRNA were sense 5′-GGCCGTGGAGGTCGTGGAGATCATTCAAGAGATGATCTCCACGACCTCCACTTTTTT-3′ and antisense 5′-AATTAAAAAAGTGGAGGTCGTGGAGATCATCTCTTGAATGATCTCCACGACCTCCAC-3′ targeting nucleotides 551-573 of the MVP transcript (NM_005115). Briefly, single-stranded oligonucleotides were synthesized and annealed to result dsDNA with sense-loop-antisense structure and respective overhangs for direct cloning into pENTR plasmids containing hH1 and hU6 promoter sequences for shRNA expression. Expression cassettes were verified by direct sequencing. Adenoviral shRNA expression plasmids were obtained through LR-reaction using pENTR and pAD/PL-DEST (Invitrogen). Resulting constructs were transfected and recombinant adenoviruses amplified in HEK293 cells by standard methods. pENTR-hH1 and pAD/PL-DEST were kindly provided by H. Takemori, Osaka.
We thank Otto Majdic and Ingela Kindas-Mügge for supplying antibodies and/or performing flow cytometry analyses as well as Birgit Strobl for technical hints regarding EMSA. The technical assistance of Vera Bachinger, Christian Balcarek, Doris Mejri and Ilse Fröhlich is gratefully acknowledged. This work was supported by the Jubiläumsfonds der Österreichischen Nationalbank (Grant number: 8817) and by the Herzfelder'sche Familienstiftung.
- Accepted October 18, 2005.
- © The Company of Biologists Limited 2006