|Home | About | Journals | Submit | Contact Us | Français|
Bax is a pro-apoptotic protein that mediates intrinsic cell-death signaling. Using a yeast-based functional screening approach, we identified interferon gamma receptor beta chain (IFNγR2) as a new Bax suppressor. IFNγR2 is a component of the IFNγ receptor complex along with the IFNγR alpha chain (IFNγR1). Upon IFNγ binding, a conformational change in the receptor complex occurs that activates the Jak2/STAT1 signaling cascade. We found that the C-terminal region (amino acids 296–337) of IFNγR2 (IFNγR2296-337) contains a novel Bax inhibitory domain. This portion does not contain the Jak2-binding domain; therefore, the antiapoptotic function of IFNγR2 is independent of JAK/STAT signaling. IFNγR2296-337 rescued human cells from apoptosis induced by overexpression of Bax but not Bak. Overexpression of IFNγR2 (wild type and IFNγR2296-337) rescued cells from etopo-side and staurosporine, which are known to induce Bax-mediated cell death. Interestingly, IFNγR2 inhibited apoptosis induced by the BH3-only protein Bim-EL, suggesting that IFNγR2 inhibits Bax activation through a BH3-only protein. Bax and IFNγR2 were co-immunoprecipitated from cell lysates prepared from HEK293 and DAMI cells. Furthermore, direct binding of purified recombinant proteins of Bax and IFNγR2 was also confirmed. Addition of recombinant Bcl-2 protein to cell lysates significantly reduced the interaction of IFNγR2 and Bax, suggesting that Bcl-2 and IFNγR2 bind a similar domain of Bax. We found that the C-terminal fragment (cytoplasmic domain) of IFNγR2 is expressed in human cancer cell lines of megakaryocytic cancer (DAMI), breast cancer (MDA-MD-468), and prostate cancer (PC3 cells). The presence of the C-terminal fragment of IFNγR2 may confer on cancer cells resistance to apoptotic stresses. Our discovery of the anti-Bax activity of the cytoplasmic domain of IFNγR2 may shed new light on the mechanism of how cell death is controlled by IFNγ and Bax.
Apoptosis is a cellular self-elimination mechanism essential for maintaining homeostasis (reviewed in ref. 1). Abnormal regulation of apoptosis is a cause of several diseases, including cancer and neurodegenerative disorders among others.1-4 Bax is a 21-kDa member of the conserved Bcl-2 family of proteins involved in regulating programmed cell death.5,6 Bax plays a key role in the intrinsic pathway of apoptosis.7 Bcl-2 family proteins are characterized by the presence of four Bcl-2 homology (BH) domains (reviewed in ref. 8). Antiapoptotic members (e.g., Bcl-2, Bcl-XL and Mcl-1) have all four BH domains (BH1-4).8 The proapoptotic members are further divided into multi-domain proteins (e.g., Bax, Bak and Bok) containing three BH domains (BH 1-3) or BH3-only proteins (e.g., Bim, Bid and PUMA, etc.,) containing just the BH-3 domain.8 The molecular mechanisms by which these proteins function and interact is not fully understood, but their role in apoptosis is indisputable (reviewed in refs. 9 and 10). Although it has been extensively studied how Bcl-2 family proteins influence each other, it is not well known how these proteins are regulated by non-Bcl-2 family proteins.
Recent studies have reported the ability of various non-Bcl-2 family members to control Bax-mediated cell death (e.g., Ku70, Humanin, ARC and Clusterin).11-15 Here we show that interferon gamma receptor beta chain (IFNγR2) is also a Bax inhibitor not belonging to the Bcl-2 family of proteins.
IFNγR2 is part of the interferon γ (IFNγ) receptor complex composed of IFNγR alpha chain (IFNγR1) and IFNγR2.16-19 IFNγR2 interacts with Jak2 prior to IFNγ binding. Upon IFNγ binding, a conformational change in the receptor complex occurs, followed by auto-phosphorylation of Jak kinase, phosphorylation of IFNγR1, and recruitment of STAT1, leading to STAT1 activation.16-19 IFNγR2 is expressed in the plasma membrane, endoplasmic reticulum (ER) and mitochondria.20,21 At present, the biological significance of the mitochondrial localization of IFNγR2 is not known. IFNγR2 knock-out mice show no sensitivity to IFNγ and are unable to prevent infection by Listeria monocytogenes.22 Previous studies showed that IFNγR2 plays a role in apoptosis regulation as a signal-transduction molecule of IFNγ (reviewed in ref. 19), but, to our knowledge, there is no report describing the apoptosis-regulating activity of IFNγR2 itself.
Here we report that the C-terminus of IFNγR2 has a Bax-inhibiting activity that is independent of the Jak/ STAT signal transduction pathway. We also found that certain cancer cell lines (DAMI cells, MDA-MD468 cells and PC3) express a truncated form of IFNγR2 containing the C-terminal Bax-inhibitory domain. The presence of this C-terminal fragment of IFNγR2 in the cytosol may help such cancer cells increase their resistance to cytotoxic stresses, including those elicited by chemo- and radiotherapy.
To perform a yeast-based functional screen for Bax inhibitors,23,24 yeast expression cDNA libraries were generated from purified mRNAs of human HeLa cells and mouse brain tissue using pJG4-5 and pYES2 vectors, respectively. As previously reported, two clones encoding the C-terminus of Ku70 were found as Bax suppressors in this screening15 (Fig. 1A). In the same experiment we also obtained a clone from the HeLa cell library encoding the cytoplasmic domain of IFNγR2 (IFNγR2263-337; amino acids 263–337 of IFNγR2) (Fig. 1A and B) as a Bax suppressor. IFNγR2263-337 contains a Jak2-binding domain (amino acids 284–295). To determine the role of the Jak2-binding domain in Bax inhibition, two IFNγR2 mutants were generated (Fig. 1B) and tested for their anti-Bax activity in human cells as described below. One mutant, IFNγR2296-337, encodes amino acids 296–337 of IFNγR2 and does not contain the Jak2-binding domain; the other mutant, IFNγR21-295, encodes amino acids 1–295, retaining the Jak2-binding domain but not the C-terminal 41 amino acids of the receptor subunit.
IFNγR2296-337 as well as IFNγR2wild type were able to inhibit apoptosis induced by Bax overexpression in HEK293 cells (Fig. 2A). On the other hand, IFNγR21-295 could not protect cells from Bax (Fig. 2A). These results suggest that the Bax-inhibiting domain localizes to the 41 amino acid sequence of the C-terminus of IFNγR2, and that Jak2-STAT1 signaling activated by IFNγ is not necessary for Bax inhibition. To confirm that IFNγR2 does not require Jak2-mediated signaling for Bax inhibition, human cell lines lacking IFNγR2 and Jak2 were examined. These cell lines were derived from the HT1080 human fibrosarcoma cell line.25 In these experiments, etoposide, a DNA topoisomerase II inhibitor, was used to induce apoptosis because etoposide is known to activate the Bax-mediated intrinsic cell death pathway.7 IFNγR2296-337 and IFNγR2wild type were both able to inhibit etoposide-induced apoptosis in these cells, but IFNγR21-295 could not (Fig. 2B–E). These results support the hypothesis that IFNγR2 can rescue cells from apoptosis independent of Jak2-mediated signal transduction.
To determine whether endogenously expressed IFNγR2 has a physiological role in suppressing apoptosis, IFNγR2 was knocked down by shRNA. HeLa cells were transfected with lentivirus that expresses shRNA targeting IFNγR2 mRNA. HeLa cells transfected with empty vector (pLKO1) or an shRNA targeting GFP mRNA were used as controls (Fig. 3A). IFNγR2 knock-down increased the sensitivity of HeLa cells to etoposide-induced apoptosis (Fig. 3B). Importantly, the basal level of apoptosis was also increased by IFNγR2 knock-down (Fig. 3B). These results suggest that IFNγR2 has a significant role in determining the cell-death sensitivity in HeLa cells.
Bim is a BH3-only protein which triggers Bax-mediated apoptosis.9,10 IFNγR2wild type as well as IFNγR2296-337 were able to inhibit apoptosis induced by Bim overexpression (Fig. 4A). This result suggests that IFNγR2 is able to suppress Bim-dependent Bax activation. On the other hand, IFNγR2 could not rescue cells from apoptosis induced by Bak overexpression (Fig. 4B), suggesting that IFNγR2 specifically inhibits Bax-mediated apoptosis.
Figure 5 shows HeLa (Fig. 5A–H) and HEK293 (Fig. 5I and J) cells expressing IFNγR2-GFP fusion proteins. IFNγR2wild type-GFP was detected in the plasma membrane, cytosol and a mitochondrion-like structure (Fig. 5C, D, I and J) as previously reported.20,21 In HeLa cells expressing IFNγR2wild type-GFP, GFP signal was detected mostly in the cytosol and plasma membrane (Fig. 5C and D), though a weak punctate pattern of GFP signal suggestive of mitochondrial localization was also detected (the image of this pattern was very difficult to capture due to the strong GFP fluorescence in the cytosol and plasma membrane). In the case of HEK293 cells, IFNγR2wild type-GFP localized to a more definite mitochondrion-like structure that was captured in the image (Fig. 5I and J). IFNγR2296-337-GFP was detected in the cytosol of both HeLa (Fig. 5E and F) and HEK293 cells (not shown). IFNγR21-295-GFP was detected in the cytosol, plasma membrane and the mitochondria-like structures in HeLa (Fig. 5G and H) and HEK293 cells (not shown). In HeLa cells, GFP signal from the mitochondria-like structure was more evident in cells expressing IFNγR21-295-GFP than cells expressing IFNγR2wild type-GFP (Fig. 5C, D and K).
IFNγR2-GFP expression in IFNγR2-null (mutant HT1080) cells was determined by western blot analysis using GFP antibody (Fig. 6A). Estimated molecular weights of IFNγR2wild type-GFP, IFNγR21-295-GFP and IFNγR2296-337-GFP, are approximately 67 kDa, 64 kDa and 34 kDa, respectively. Proteins that have similar molecular weights were detected by GFP antibodies (Fig. 6A), suggesting that IFNγR2-GFP were expressed in these cells. Interestingly, we observed that IFNγR21-295-GFP migrated slower than IFNγR2wild type-GFP in every western blotting experiment performed in this study (Fig. 5A and C). Since the estimated molecular weight of IFNγR21-295-GFP is smaller than IFNγR2wild type-GFP, this observation was unexpected. At present, we do not know the exact reason for this phenomenon, but a posttranslational modification such as glycosylation may be the cause of the slower migration of this mutant protein in SDS-PAGE.
Western blot analysis of IFNγR2-GFP expression was also performed using HEK293 cells (Fig. 6B–D). Although IFNγR2296-337-GFP was detected at its estimated molecular weight (Fig. 6B), IFNγR2wild type-GFP and IFNγR21-295-GFP could not be detected in a simple western blot using GFP antibodies in HEK293 cell lysates. To verify the expression of IFNγR2-GFP fusion proteins (both wt and mutants), cell lysates were subjected to GFP immunoprecipitation and samples were further analyzed by IFNγR2 antibodies (Fig. 6C and D). After enrichment of the GFP-tagged proteins, expression of IFNγR2wild type-GFP and IFNγR21-295-GFP was confirmed by monoclonal antibody recognizing the N-terminus of IFNγR2 (Fig. 6C), and IFNγR2296-337-GFP expression was confirmed by antibodies detecting the C-terminus of IFNγR2-GFP (Fig. 6D).
There were GFP-antibody-positive bands with slightly higher and lower molecular weight than GFP (29 kDa) in cells transfected with pEGFP C2-IFNγR2wild type and pEGFP C2-IFNγR21-295 (bands marked with * in Fig. 6A and B). We speculate that protease-dependent cleavage of IFNγR2-GFP fusion proteins produced these fragments. Protease inhibitors were present in the cell lysis buffer; therefore, it is likely that this cleavage occurred in the cells prior to preparation of the cell lysate, though further careful study will be needed to reveal the reasons for the appearance of these bands.
Bax activation involves exposure of the protein's N-terminus by a conformational change followed by Bax translocation from the cytosol to mitochondria.14,26,27 Exposure of the N-terminus of Bax can be monitored by immunohistochemistry using 6A7 Bax monoclonal antibody (6A7 Ab) recognizing an epitope in the N-terminus of Bax.28,29 Staurosporine (STS), a pan-kinase inhibitor,30 was used to induce the Bax conformational change. STS treatment (100 nM, 3 h) induced Bax activation that was detected by 6A7 Ab as shown in Figure 7A. GFP expression itself did not inhibit Bax activation (Fig. 7A). IFNγR2296-337-GFP as well as IFNγR2wild type-GFP (Fig. 7C and D) inhibited STS-induced Bax activation. On the other hand, IFNγR21-295-GFP did not inhibit Bax activation (Fig. 7B). The percentages of 6A7 Ab-positive cells in GFP-positive cells were calculated and the results are shown in Figure 7E. The inhibition of Bax activation by IFNγR2296-337-GFP as well as IFNγR2 wild type-GFP was statistically significant (Fig. 7E).
To examine whether IFNγR2 can bind Bax, we performed co-immunoprecipitation of endogenously expressed Bax and IFNγR2 in HEK293T cells (Fig. 8). It is known that certain detergents such as NP40 artificially activate Bax whereas CHAPS does not.28,29 Interestingly, Bax and IFNγR2 were co-immunoprecipitated by anti-Bax antibody in buffers containing either NP40 or CHAPS (Fig. 8A). This interaction was also observed when anti-IFNγR2 was used for immunoprecipitation and anti-Bax was used for Bax detection in western blot (Fig. 8B). Furthermore, the direct interaction of purified recombinant proteins of BaxΔTM (in which the c-terminal transmembrane (TM) domain is deleted) and IFNγR2263-337 (tagged with thioredoxin (rTrx)) was confirmed (Fig. 8C). These results suggest that the C-terminus of IFNγR2 directly binds Bax.
Since it is known that Bcl-2 binds and inhibits Bax, we examined whether Bcl-2 has any influence on the IFNγR2-Bax interaction in vitro. Interestingly, addition of recombinant Bcl-2 protein (a truncated form without the C-terminal transmembrane domain to increase solubility in the buffer) to the HEK293T cell lysate interferes with the interaction of Bax and IFNγR2 (Fig. 8D). In this experiment, NP40-based buffer was used because the Bcl-2-Bax interaction is known to be observed in this buffer.31 This result suggests that Bcl-2 and IFNγR2 recognize the same domain of Bax. We also examined the effects of Bax inhibiting peptide (BIP) designed from the Bax-binding domain of Ku70.13,15 Because Ku70 and BIP are known to bind the inactive form of Bax in CHAPS-based buffer, we used CHAPS-based buffer to examine the effects of BIP. As shown in Figure 8E, BIP did not cause a significant inhibition of the Bax-IFNγR2 interaction. Three independent experiments were performed, and we observed results similar to that in Figure 8E in two of the experiments. In one experiment, BIP caused a detectable reduction in the amount of IFNγR2 protein pulled down by Bax antibody (data not shown); however, this effect was not reproducible.
IFNγR2 C20 antibody (C20 Ab) recognizes the C-terminal 20 amino acids of IFNγR2 as an epitope. This antibody detected a small fragment (approximately 10 kDa) in western blot analysis of cell lysates prepared from megakaryocytic cancer cells (DAMI) and SV40-transformed kidney cells (HEK293T), but not from normal primary cultured cells (HUVECs) (Fig. 9A). This small fragment was enriched by immunoprecipitation (Fig. 9B), digested by trypsin and its identity was determined by targeted LC-MS/ MS analysis. As a result, it was confirmed that a tryptic peptide derived from the C-terminus of IFNγR2, DPTQPILEALDK, was present in the sample (Fig. 9C and D). Expression of this C-terminal fragment was also detected in two cancer cell lines, MDA-MD468 (breast cancer cell line) and PC3 (prostate cancer cell line) (Fig. 9E). Interestingly, the C-terminal fragment of IFNγR2 was not detected in normal mammary epithelial cells (4A100) or in non-tumorigenic immortalized breast (HME-1) and prostate (RWPE-1) cells. These results suggest that a certain protease expressed in malignant tumorigenic cells may produce the antiapoptotic cytoplasmic fragment derived from IFNγR2.
The approximate intracellular protein concentrations of IFNγR2 and Bax were determined by densitometric analysis of western blots using purified recombinant proteins as standards (Fig. 10A and B). For IFNγR2, densitometric analysis was performed on a band corresponding to non-glycosilated full-length IFNγR2 (approximately 37 kDa). For Bax, the density of a band corresponding to the full length of Bax (approximately 21 kDa) was measured. As there are other forms (glycosylated, truncated, etc.,) of IFNγR2 and Bax, the estimated protein concentration in this experiment may underestimate the actual total expression levels of these proteins in cells. However, our attempt to obtain an estimate of the stoichiometry of Bax and IFNγR2 will help determine our working hypothesis of how IFNγR2 regulates Bax-mediated apoptosis in the cell. The results of the measurements are shown in Table 1.
First, the concentrations of Bax and IFNγR2 in total cell lysate (i.e., no fractionation) were measured. The ratio of Bax to IFNγR2 was approximately 1:1, 3:1 and 2:1 in HUVECs, DAMI cells and HEK293T cells, respectively. Next, the concentrations of Bax and IFNγR2 in the cytosol, nucleus and heavy membrane (mitochondria-rich fraction) were measured using DAMI cells and HEK293T cells. The Bax:IFNγR2 ratio in the cytosolic fraction was 2:1 and 3.5:1 in DAMI cells and HEK293T cells, respectively. The ratio in the heavy membrane fraction of DAMI cells and HEK293T cells was approximately 1.5:1 and 6:1, respectively. Interestingly, Bax and IFNγR2 were also detected in the nuclear fraction (Fig. 10). Because the nuclear fraction contains ER membranes attached to the cytosolic surface of the nucleus, the estimated concentration of Bax and IFNγR2 in the nuclear fraction is expected to be higher than the actual concentration in the nucleus. Therefore, the data for the nuclear fraction may not be reliable for drawing a conclusion about the interaction of Bax and IFNγR2 inside of the nucleus.
The present study showed that the C-terminal portion of IFNγR2 has Bax-inhibiting activity, and that this activity does not require the Jak2-STAT1 pathway. In addition, we found that two cancer cell lines (MDA-MD-468 and PC3) express the C-terminal fragment of IFNγR2 (Fig. 9). Interestingly, this fragment was not detected in normal (HUVEC) or non-tumorigenic immortalized breast (HME-1) and prostate (RWPE-1 and RWPE-2) cells. At present, the mechanism of how this fragment is generated is not clear. According to the results of the mass spectrometry analysis, the sequence of this fragment begins in the middle of the transmembrane domain. As there is no potential start codon (ATG) in the corresponding cDNA sequence, the fragment is not likely be produced as a result of alternative start codon of IFNγR2 mRNA. More likely, protease-dependent cleavage of IFNγR2 probably generates the cytoplasmic fragment. Interestingly, presenilin, a membrane-associated protease, is reported to be able to cleave IFNα receptor and the C-terminal fragment plays a role in signal transduction.32 IFNγR2 may also be a substrate of metaloproteinases that are expressed at high levels in certain types of cancer cells.33 In general, IFNγ is known as a proapoptotic cytokine that induces cell death of abnormal cells, including cancer cells.34-36 The cleavage of IFNγR2 may have two implications for cancer cell survival: shutting down of IFNγ-mediated apoptosis signaling and production of the antiapoptotic cytoplasmic fragment. Further studies investigating the mechanism and biological significance of IFNγR2 cleavage are now underway.
As shown in Figure 8, IFNγR2 interacted with Bax in NP40-based buffer as well as in CHAPS-based buffer. Previous studies showed that NP40, but not CHAPS, induces artificial activation of Bax associated with a conformational change (i.e., exposure of the N-terminus that can be detected by antibodies recognizing this portion as an epitope).28,29 Our observation suggests the possibility that IFNγR2 can bind both the active and inactive forms of Bax. There are at least three possible mechanisms to explain such an interaction. First, the IFNγR2-binding domain of Bax may not show major structural changes before and after exposure of the N-terminus, thus permitting binding in either conformation. Second, IFNγR2 may bind to different domains of Bax in different buffers, i.e., IFNγR2 may be able to bind more than two sites of Bax. Third, IFNγR2 may bind only primed (or partially active) Bax in which the N-terminus is already exposed. Because a small fraction of Bax molecules are reported to be partially activated in living cells,37 IFNγR2 may be able to co-immunoprecipitate with these active Bax molecules present in cell lysates prepared in CHAPS-based buffer. The present study showed that IFNγR2 is able to inhibit the conformational change and mitochondrial translocation of Bax that are triggered by apoptotic stresses such as STS treatment (Fig. 7). As discussed earlier, IFNγR2 may be able to bind both inactive and N-terminus-exposed Bax molecules. We speculate that the binding of IFNγR2 to the inactive form of Bax stabilizes the inactive status of Bax, just as Ku70 or BIP binding does.13,15 However, IFNγR2 is also able to bind to N-terminus-exposed Bax molecules, as Bcl-2 does. This interaction between IFNγR2 and N-terminus-exposed Bax may suppress further activation of Bax such as its mitochondrial translocation and multimerization, analogous to the Bcl-2-Bax interaction (reviewed in refs. 9 and 10). Further biochemical studies are clearly needed to explain the mechanism of how IFNγR2 inhibits the Bax activation process.
Bcl-2 is known to heterodimerize with Bax, and this interaction inhibits the multimerization of Bax that is required for the permeabilization of the mitochondria outer membrane.38 It has been shown that the Bcl-2-Bax interaction can be observed only in buffer (e.g., NP40-based buffer) that induces N-terminus exposure of Bax.31 Thus Bcl-2 is thought to bind primed (partially activated) Bax.39 On the other hand, Ku70 protein and BIP are known to bind Bax only in buffer (CHAPS-based or detergent-free buffer) that maintains the inactive conformation of Bax.15 Therefore, Ku70 is thought to bind the inactive form of Bax. In this study, we observed that Bcl-2 competed with IFNγR2 to interact with Bax in NP40-based buffer; however, BIP did not show significant influence on the IFNγR2-Bax interaction in CHAPS-based buffer (Fig. 8D). These results suggest that IFNγR2 and Bcl-2 may recognize the same domain of Bax, and that the mechanisms of interaction of Bax-BIP (and probably Ku70) and Bax-IFNγR2 are different. At present, we cannot identify any homologous amino acid sequence between Bcl-2 and IFNγR2 by using publicly available homology search software. IFNγR2 may have an uncharacterized domain that is able to show Bcl-2-like Bax-binding activity. As discussed earlier, the results of the present study also imply the possibility that IFNγR2 binds two different domains of Bax, since IFNγR2 can bind Bax both in NP40- and CHAPS-based buffer. Although it is highly speculative, one of the domains may be the Bcl-2-binding domain of Bax, and the other one may be the domain that is already exposed to the molecular surface in CHAPS-based buffer.
To determine whether endogenous levels of IFNγR2 have roles in controlling Bax activation in cancer cell lines, IFNγR2 expression was knocked down by shRNA in HeLa cells. As shown in Figure 3, IFNγR2 shRNA-expressing cells showed an increase in basal cell death (i.e., cell death in normal cell culture) and these cells showed higher sensitivity to apoptotic stresses such as etoposide treatment. These results support our hypothesis that IFNγR2 has a physiologically significant role in the maintenance of cellular life as an inhibitor of a potentially cytotoxic protein, Bax. However, the anti-Bax activity of IFNγR2 may become physiologically significant only in particular cell types such as cancer cells. At present, there is no report describing a significant increase of apoptosis in IFNγR2 knock-out mice, suggesting that IFNγR2-deficiency does not cause severe defects in apoptosis regulation at least in normal mouse cells. Further careful studies will be needed to determine the physiological significance of the anti-Bax activity of IFNγR2 in both normal and cancer cells.
As shown in Figure 10 and Table 1, the number of IFNγR2 molecules in the cell is similar to or less than the number of Bax molecules. This observation implies that IFNγR2 will not be able to bind and control all Bax molecules in a cell if these proteins interact in a 1:1 ratio. Probably, only a certain portion of Bax molecules are under the control of IFNγR2, and other Bax molecules are controlled by different Bax regulators11-15 such as Ku70, ARC, Humanin, Clusterin and Bcl-2 family proteins (reviewed in refs. 9 and 10). There are multiple mechanisms tightly controlling cellular fate to survive or die by regulating Bax activity. IFNγR2 may participate in Bax regulatory mechanisms as one of the negative regulators of Bax, even if its role is a minor one.
The C-terminal fragment of IFNγR2 is expected to control Bax in the cytosol since this fragment is a cytosolic domain of IFNγR2. In fact, IFNγR2-GFP296-337 localized to the cytosol (Fig. 5). IFNγR2-GFPwild type was found in the plasma membrane, ER and mitochondria (Fig. 5). These observations are consistent with those of previous reports describing the subcellular localization of IFNγR2.20,21 We speculate that IFNγR2 is able to inhibit Bax activation in the mitochondrion as Bcl-2 controls Bax in this organelle, though further careful study is needed to test our working hypothesis.
The yeast strain EGY48 was used for the yeast functional screening for Bax inhibitors as described previously.15,23,24 Mouse Bax was expressed under a galactose-inducible promoter using pGilda vector as reported.40 Yeast expression libraries of cDNAs of mouse brain and a human cell line (HeLa) were prepared in pYES2 and pJG4-5 vectors, respectively.
HT1080 (wild type, IFNγR2 null mutant and JAK2 null mutant) cells were kindly provided by Dr. George Stark. All HT1080 cell lines were cultured in Dulbecco's Modified Eagle's Medium (DMEM) supplemented with 10% fetal bovine serum (FBS). HeLa, Human Embryonic Kidney (HEK) 293 and HEK293T cells were purchased from ATCC, and cultured in DMEM supplemented with 10% FBS. LNCap and PC3 cells were purchased from ATCC and cultured in DMEM F12 medium supplemented with 10% FBS. RWPE1 and RWPE2 were purchased from ATCC and cultured in keratinocytes-SFM plus supplements medium (Gibco). Primary mammary epithelial (4A100) cultures were derived from organoids isolated from discarded mammary tissue acquired from patients undergoing reduction mammoplasty surgery. Anonymized specimens were acquired from patients who given written consent, through the Tissue Procurement and Histology Core Facility of the Case Comprehensive Cancer Center (Case CCC), under a Case CCC IRB approved protocol. Primary epithelial cultures were established using a protocol described by Dr. Martha Stampfer (as described by P. Novac et al.41) and were grown in M87A + X medium. Human HME1 cells (Clontech) were grown in medium 171 with mammary epithelial growth supplement (Cascade Biologics) and penicillin-streptomycin.42 Human breast cancer cell line MDA-MD468 was cultured in RPMI, 5% FBS, supplemented with L-glutamine, penicillin-streptomycin and fungizone (Gibco). Human megakaryocytic cell line DAMI cells was cultured in Iscove's Modified Dulbecco's Medium (IMDM) supplemented with 10% horse serum. Human umbilical vein endothelial cells (HUVEC) were cultured in endothelial cell growth medium with supplements (EGM®-2-Endothelial Cell Medium-2-Lonza).
Cells were cultured overnight in DMEM supplemented with 10% FBS. The transfections were performed using Superfect® (Qiagen, Valencia, CA) in accordance with the manufacturer's instructions. Transfection efficiency was analyzed by the expression of the EGFP-tagged proteins.
Five E. coli clones expressing pLKO1-shRNA IFNγR2 plasmids were purchased from Open Biosystems (cat # RHS4533-NM_005534). Lentiviruses were produced in HEK293T cells by transfection using each of pLKO1-shRNA IFNγR2, pCMV DR 8.76 and pMD2G. Viruses were produced and used to infect HeLa cells with a 1:3 dilution of stock lentivirus for 16 h. Cells were then cultured for 24 h in complete medium, and then stable clones expressing the shRNA against IFNγR2 and shRNA against GFP (control shRNA) were selected using puromycin. To select the best shRNA targeting IFNγR2 mRNA, cell lysates were analyzed by western blotting, and the best clone showing the lowest IFNγR2 protein expression was used to determine the effects of IFNγR2 knock-down in HeLa cells.
Apoptosis was induced by transfecting the cells with pcDNA3-human Bax, or pcDNA3-human Bak, or pcDNA3-human Bim EL, or by treatment with etoposide (10 uM) or staurosporine (100 nM). To determine the induction of apoptosis by different apoptotic stresses, cells were stained with Hoechst 33258 dye, and the numbers of cells with apoptotic nuclei were counted using fluorescence microscopy. Three hundred cells were analyzed in triplicate samples. The data presented in the figures showed the percentage of apoptosis ±SE of three independent experiments. Caspase activity was measured by using a fluorogenic caspase 3 substrate II (Calbiochem), as previously described.43 IFNγR2 constructs were cloned in pEGFP-C2 (IFNγR2wild type, IFNγR21-295 and IFNγR2296-337) vector.
For Bax, Bak or Bim EL overexpression, cells were transfected with either 1 μg pcDNA3-human Bax, 1 μg pcDNA3-human Bak, or 1 μg pcDNA3-human Bim EL, and 4 μg of pEGFP plasmid encoding IFNγR2, and the apoptosis or caspase activity was determined 24 h after the transfection.
HEK293T cells were lysed in 300 μl NP40 buffer (150 mM NaCl, 10 mM HEPES at pH 7.4 and 1% NP40) or CHAPS buffer (150 mM NaCl, 10 mM HEPES at pH 7.4 and 1% CHAPS) supplemented with protease inhibitors (1:100 dilution of protease inhibitor Cocktail; Sigma) and PMSF, as previously reported.14 Samples were pre-cleared by incubating 300-μl (1,000 mg total protein) cell lysates with 20 μl protein-G-sepharose (Amersham Biosciences) at 4°C for 1 h. Then, the samples were incubated with 20 μl protein-G-sepharose pre-absorbed with 2 μg of Bax monoclonal antibody (B9, Santa Cruz) or IFNγR2 monoclonal antibody (Fitzgerald) at 4°C for 2 h. After the incubation, sepharose beads were washed with lysis buffer. Beads were then boiled in 30 μl Laemmli buffer, and 15 μl of the sample was analyzed by western blotting. Western blotting analysis of pre-immunoprecipitation (Input) (100 μg total protein) and immunoprecipitated samples (IP) were performed with a Bax monoclonal antibody (B9 antibody, Santa Cruz), Bax polyclonal antibody (N20 antibody, Santa Cruz), or IFNγR2 polyclonal antibody (Fitzgerald).
Recombinant human Bax ΔTM (Bax C-terminal transmembrane truncated human Bax) was produced by using pHMTc vector downstream of maltose binding protein (MBP), separated by the TEV protease site. Overexpressed MBP-Bax was purified through a maltose-binding column (NEB) and subsequently cleaved by TEV protease (Invitrogen), followed by Ni-affinity purification to remove the protease and the His-tagged MBP. IFNγR2 cytoplamic domain (amino acids 263–337) was fused with thioredoxin (rTrx) to increase the recovery rate from bacterial lysates. The production and the purification of this fusion protein were performed by Protein X Laboratory (San Diego, CA). Recombinant Bax (25 ng) was loaded onto Sepharose G beads pre-equilibrated with anti-Bax antibody (Bax B9, Santa Cruz) or pre-immune IgG (control IgG) at 4°C for 2 h. The excess Bax molecules were washed 3 times with buffer (50 mM phosphate buffer, pH 7.4). Recombinant IFNγR2 (263-337)-rTrx (25 ng) was added to the beads preloaded with Bax and anti-Bax or control IgG. Beads and IFNγR2263-337-rTrx were incubated at 4°C for 2 h. After the incubation, beads were extensively washed with the loading buffer (50 mM phosphate buffer, pH 7.4). Beads were boiled in Laemmli buffer, and the supernatant was collected as a sample. Samples were analyzed by western blot using anti IFNγR2 antibody (C20, Santa Cruz) and Bax (N20, Santa Cruz).
HEK293T cells were lysed by using either NP40 buffer (150 mM NaCl, 10 mM HEPES at pH 7.4 and 1% NP40) or CHAPS buffer (150 mM NaCl, 10 mM HEPES at pH 7.4 and 1% CHAPS) supplemented with protease inhibitors (1:100 dilution of protease inhibitor Cocktail; Sigma) and PMSF, as previously reported.14 To determine if recombinant Bcl-2 (Prospect cat # PRO-630) protein competes with endogenous IFNγR2 for binding to endogenous Bax, HEK293T cell lysate prepared in NP40 buffer was used. Three-hundred microliters (1,000 μg total protein) of the sample was pre-cleared by incubating in 20 μl protein G-sepharose (Amersham Biosciences) at 4°C for 1 hour. Cleared samples (300 μl) were incubated (4°C for 2 h) with or without recombinant Bcl-2 (75 or 150 nM final concentration) in the presence of protein G sepharose (20 μl) preabsorbed with 2 μg of Bax polyclonal antibody (BD Biosciences). Beads were washed and then boiled in 30 μl Laemmli buffer, and 15 μl of the eluted protein solution was analyzed by western blotting. Western blotting of pre-immunoprecipitation (pre-treated) (100 μg total protein) and immunoprecipitated samples (IP) were performed with IFNγR2 monoclonal antibody (Fitzgerald-WB), Bcl-2 monoclonal antibody (Santa Cruz), and Bax polyclonal antibody (HRP-conjugated N20 antibody, Santa Cruz). To determine the effects of BIP15 on the interaction of endogenous IFNγR2 and Bax, HEK293T cell lysate prepared in CHAPS buffer was used. Three-hundred microliters of the sample was pre-cleared by incubating with 20 μl protein G-sepharose (Amersham Biosciences) at 4°C for 1 h. Cleared samples (300 μl, 1,000 μg total protein) were incubated (37°C for 2 h) with or without BIP (40 or 200 μM final concentration) in the presence of protein G sepharose (20 μl) preabsorbed with 2 μg of IFNγR2 monoclonal antibody (Fitzgerald). Beads were washed and then boiled in 30 μl Laemmli buffer, and 15 μl of the eluted protein solution was analyzed by western blotting. Western blot analysis of Bax was performed by using Bax polyclonal antibody (N20, Santa Cruz).
HeLa cells were transiently transfected with pEGFP-C2-IFNγR2 (wild type, 1–295, or 296–337) using Superfect® (Qiagen, Valencia, CA) in accordance with the manufacturer's instructions. Four mico-grams of the plasmid were used to transfect cells cultured in 6-cm diameter dishes. After 24 h of transfection, the cells were treated with staurosporine (100 nM) for 3 h. Then, the cells were washed with phosphate buffer pH 7.4 (PBS), fixed using paraformaldehyde (1%), permeabilized with Triton X-100 (0.02%), blocked with goat serum, and the activation of Bax was analyzed by immunocytochemistry using monoclonal Bax 6A7 antibody (BD Pharmingen) and Alexa Fluor® 568-labeled anti-mouse IgG secondary antibody (Invitrogen).15
Recombinant BaxΔTM and IFNγR2263-337-rTrx were used as standards. Cells were harvested and lysed in NP40 buffer (10 mM HEPES, 150 mM NaCl and 1% NP40 pH 7.4), or with hypotonic buffer for subcellular fractionation (20 mM HEPES, 10 mM KCl, 1.5 mM MgCl2, 1 mM EDTA and 250 mM sucrose); both buffers were supplemented with protease inhibitors cocktail (Sigma) and PMSF (Sigma). LDH, F1α and YY1 proteins were used as makers of the cytosolic, mitochondrial and nuclear fractions, respectively. Cell lysates from equivalent cell numbers, and sequential dilutions of protein standards were subjected to SDS-PAGE (BioRad). Bax antibody conjugated with horseradish peroxidase (HRP) (anti-Bax N20-HRP, Santa Cruz) was used to detect Bax, and IFNγR2 antibody (C-20, Santa Cruz) was used to detect IFNγR2. HRP-conjugated anti-rabbit goat IgG was used as a secondary antibody. Signal intensities were analyzed by using BioRad Gel Doc and Quantity One 4.5.1 software from BioRad.
To enrich for the immunoreactive (ir)-IFNγR2 fragment expressed in transformed cells, DAMI and HEK293T cell lysates were incubated with anti-IFNγR2 antibody (C-20, Santa Cruz) overnight at 4°C, and the antibody-protein complexes were recovered by incubation of the mixture with protein G sepharose. The sepharose gels were then boiled in 30 μl Laemmli buffer, and 15 μl of the eluted protein solution was used for 1D-SDS-PAGE and western analysis. From a Coomassie blue-stained Tris-HCl gel, bands running between the protein markers for 10 and 15 kDa were collected. Proteins were reduced by DTT, alkylated by iodoacetamide and digested by trypsin overnight. The tryptic peptides were extracted from the gel by using 60% acetonitrile in 0.1% formic acid.44 Recombinant Trx-tagged IFNγR2263-337 was used as positive control for the LC-MS/MS analysis. The tryptic peptides were analyzed by LC-MS/MS using a LTQ Orbitrap XL linear ion trap mass spectrometer (Thermo Fisher Scientific, Waltham, MA) coupled to an Ultimate 3000 HPLC system (Dionex) in the Case Center for Proteomics. The LC-MS/MS analysis was performed as follows: peptide solutions were injected into a reverse phase Aclaim PepMap 100 C18 column (3 mm, 100 Å, 150 mm × 75 mm, Dionex Corporation, Sunnyvale, CA). Mobil phases used were: 2% acetonitrile, 0.1% formic acid in water (solvent A), and 80% acetonitrile, 0.1% formic acid (solvent B). A linear gradient of solvent B from 0% to 60% over a period of 60 min was used at a flow rate of 300 nL/min. Three specific peptide ions that are expected to be produced from IFNγR2 were selected with a mass window of 3 amu and subjected to MS/MS analysis with a normalized collision energy of 35%. These ions were: m/z 670.4 (z = 2, DPTQPILEALDK), 868.9 (z = 2, DDVWDSVSIISFPEK), and 754.4 (z = 3, YWFHTPPSIPLQIEEYLK). LTQ injection time was set to 2 s and automatic gate control target was 10,000 ions. The results from the LC-MS/MS analysis were subjected to an NCBI nr (version 20070216, containing 4626804 sequences) database search using Mascot Daemon Version 2.2.0 with a mass tolerance set to 2 Da for the precursor and 1 Da for the product ions. In addition to the targeted analysis the remaining digest was also analyzed by data-dependent LC-MS/MS.
We are grateful to Dr. George Stark of The Lerner Research Institute-Cleveland Clinic, for providing the HT1080 IFNγR2 null cells and the HT1080 Jak2 null cells used in this study, and for his comments and suggestions about the project. This work was supported in part by National Institutes of Health (NIH) grants PC0CA1037 and R01AG031903 (to S.M.) and an American Heart Association fellowship 0615139B (to J.A.G.). The study was also supported in part by the Flow Cytometry and Confocal Microscopy core facilities of the Comprehensive Cancer Center of Case Western Reserve University and University Hospital P30CA43703, and by the Translational Research Core and the Tissue Procurement and Histology Core Facility of the Case Comprehensive Cancer Center (P30 CA43703).