Ubiquitin-like protein MNSFβ covalently binds to Bcl–G and enhances lipopolysaccharide/interferon γ-induced apoptosis in macrophages

Authors


Correspondence

M. Nakamura, Department of Cooperative Medical Research, Collaboration Center, Shimane University, Izumo 693–8501, Japan

Fax: +81 853 20 2913

Tel: +81 853 20 2916

E–mail: nkmr0515@med.shimane–u.ac.jp

Abstract

Monoclonal non-specific suppressor factor β (MNSFβ) is a ubiquitously expressed member of the ubiquitin-like family that is involved in various biological functions. Previous studies have demonstrated that MNSFβ covalently binds to intracellular pro-apoptotic protein Bcl–G and regulates the extracellular signal-regulated kinase (ERK)/mitogen-activated protein kinase (MAPK) cascade in the mouse macrophage cell line Raw264.7. In this study, we demonstrate that MNSFβ promotes lipopolysaccharide (LPS)/interferon γ (IFNγ)-induced apoptosis of Raw264.7 macrophages. In Raw264.7 cells treated with MNSFβ small interfering RNA (siRNA), LPS/IFNγ- or NO donor S–nitrosoglutathione-induced apoptosis was inhibited. siRNA-mediated knockdown of MNSFβ did not affect inducible nitric-oxide synthase (iNOS) expression in LPS/IFNγ-stimulated Raw264.7 cells. Conversely, co-transfection with MNSFβ and Bcl–G greatly enhanced LPS/IFNγ- induced apoptosis in Raw264.7 cells, accompanied by increased expression of p53 and decreased Cox–2 activity. Unlike co-transfection with wild-type MNSFβ, co-transfection of a mutant MNSFβ (G74A) and Bcl–G did not result in enhancement of LPS/IFNγ-induced apoptosis. Co-over-expression of MNSFβ and Bcl–G reduced S–nitrosoglutathione-induced ERK1/2 phosphorylation. Furthermore, electrophoretic mobility shift assay experiments revealed that MNSFβ down-regulates the ERK/activator protein 1 (AP–1) signaling cascade which leads to Cox–2 activation. We also observed that MNSFβ–Bcl–G promotes LPS/IFNγ-induced apoptosis of mouse peritoneal macrophages, together with a decrease in Cox–2 expression. Taken together, our data indicate an apoptosis-enhancing effect of MNSFβ–Bcl–G is due in part to down-regulation of Cox–2 activation in macrophages.

Structured digital abstract

Abbreviations
AP-1

activator protein 1

BCL-G

B-cell CLL/lymphoma-G

EMSA

electrophoretic mobility shift assay

ERK

extracellular signal-regulated kinase

Fau

Finkel-Biskis-Reilly murine sarcoma virus (FBR-MuSV) ubiquitously expressed

GSNO

S–nitrosoglutathione

IFNγ

interferon γ

IKK

IκB kinase

iNOS

inducible nitric-oxide synthase

LPS

lipopolysaccharide

MEK

mitogen-activated protein kinase/extracellular signal-regulated kinase kinase

MAPK

mitogen-activated protein kinase

MNSF

monoclonal non-specific suppressor factor

NFκB

nuclear factor κB

TLR2

Toll-like receptor 2

Introduction

In eukaryotes, a major mechanism for regulating protein activity involves covalent attachment of ubiquitin or ubiquitin-like proteins to the ε–amino group of lysine via formation of an isopeptide bond. Post-translational modification by ubiquitin-like proteins regulates a variety of important eukaryotic processes, such as cell division, nuclear transport, the stress response and immune responses [1-8]. Monoclonal non-specific suppressor factor β (MNSFβ) was originally identified as a cytokine [9]. MNSFβ (gene symbol FAU) is ubiquitously expressed as a fusion protein with the ribosomal protein S30 at its terminus. MNSFβ shares 57% homology with ubiquitin when conservative substitutions are considered [9, 10]. The C–terminal amino acids of ubiquitin (Gly-Gly), which are involved in isopeptide bond formation during conjugation, are conserved in the MNSFβ sequence [9]. Unlike ubiquitin, MNSFβ may not be involved in protein degradation. MNSFβ is covalently attached to certain lysines of specific target proteins, including Bcl–G, a pro-apoptotic member of the Bcl2 family, and endophilin II. We have demonstrated that MNSFβ binds to endophilin II, with a linkage between the C–terminal Gly74 and Lys294 [11]. MNSFβ covalently conjugates to endophilin II in liver and cells of the macrophage cell line Raw264.7, and the MNSFβ–endophilin II complex formation may be implicated in phagocytosis in macrophages [11, 12]. More recently, we found that MNSFβ–endophilin II inhibits the signal pathway upstream of IKB kinase (IKK) activation, but not downstream of toll-like receptor (TLR)2 signaling [13]. The B-cell CLL/lymphoma-G (BCL–G) gene is a pro-apoptotic p53 target gene, and Bcl–G induces apoptosis in human kidney HEK293T cells [14]. MNSFβ conjugates to Bcl–G, with a linkage between the C–terminal Gly74 and Lys110 [15], and regulates the mitogen-activated protein kinase (MAPK) pathway by inhibiting activation of extracellular signal-regulated kinase (ERK) [16]. Recently, it has been reported that MNSFβ, also known as Finkel-Biskis-Reilly murine sarcoma virus (FBR-MuSV) ubiquitously expressed (Fau), and Bcl–G are implicated in UV-induced apoptosis in human T–lymphoblastic leukemia cells [17].

Several ubiquitin-like proteins are implicated in interferon signaling. MNSFβ, ISG15, FAT10 and NUB1L are induced by interferon γ (IFNγ) [18-22]. Among them, MNSFβ and FAT10 are closely involved in apoptosis [15, 23]. In this study, we demonstrate that MNSFβ covalently conjugates to Bcl–G and promotes LPS/IFNγ-induced apoptosis in Raw264.7 cells.

Results

MNSFβ enhances LPS/IFNγ-induced apoptosis in Raw264.7 macrophages

In a previous study, we showed that MNSFβ covalently conjugates to Bcl–G and regulates the ERK/MAPK cascade [16]. As Bcl–G is a ubiquitously expressed pro-apoptotic protein, we examined the mechanism of action of the MNSFβ–Bcl–G complex in LPS/IFNγ-induced apoptosis in Raw264.7 cells. We first addressed whether MNSFβ siRNA affects the apoptosis of LPS/IFNγ-stimulated cells. In Raw264.7 cells transfected with MNSFβ–1 siRNA, LPS/IFNγ-induced apoptosis was significantly inhibited (Fig. 1A). RT–PCR analysis demonstrated that MNSFβ-1 siRNA, but not control scrambled siRNA, specifically reduced the expression of MNSFβ. Conversely, transfection with expression construct pcDNA3.1–MNSFβ resulted in the enhancement of apoptosis induced by LPS/IFNγ (Fig. 1B). These results are in accordance with the findings of the siRNA experiments (Fig. 1A), indicating that this ubiquitin-like protein may mediate pro-apoptotic signal transduction. We next examined the involvement of Bcl–G in the regulation of apoptosis by MNSFβ. We have shown that the covalent complex of MNSFβ and Bcl–G is observed in unstimulated Raw264.7 cells [16]. Transfection with expression construct pcDNA3.1–Bcl–G resulted in a small enhancement of LPS/IFNγ-induced apoptosis in Raw264.7 cells (Fig. 1B), and, conversely, the apoptosis was slightly inhibited by Bcl–G siRNA knockdown (data not shown). Co-transfection of the expression vectors encoding Bcl–G and MNSFβ caused a marked enhancement of apoptosis (83 ± 5%). Together with previous findings regarding covalent conjugation of MNSFβ to Bcl–G [15], these results suggest that the MNSFβ–Bcl–G complex may play an essential role in the positive regulation of apoptosis in macrophages. Supporting this notion, mutant proteins in which the C–terminal Gly-Gly of MNSFβ was mutated to Gly-Ala showed no pro-apoptotic activity (Fig. 1B). We have previously shown that MNSFβ fails to bind to mutant Bcl–G (K110R) [15]. As shown in Fig. 1B, transfection of the mutant construct had no effect on pro-apoptotic function. Thus, MNSFβ conjugation to Bcl–G through a linkage between the C–terminal Gly74 and Lys110 is responsible for the regulation of apoptosis. The over-expression of MNSFβ and Bcl–G was confirmed by western blot. An anti-Bcl–G IgG recognized Bcl–G and the MNSFβ–Bcl–G complex in cell extracts (Fig. 1B), as previously described [15]. An anti-MNSFβ Ig recognized multiple bands, including those with similar electrophoretic mobility to the MNSFβ–Bcl–G complex, as previously described [16]. Thus, the MNSFβ–Bcl–G complex was detected by immunoprecipitation and western blot analyses. The data in Fig. 1C show that co-transfection with Bcl–G and MNSFβ cDNAs remarkably increased formation of MNSFβ–Bcl–G complexes (33.5 kDa MNSFβ adduct). Expression of free (unconjugated) 8.5 kDa MNSFβ was also observed (Fig. 1B). The levels of endogenous MNSFβ and Bcl–G did not change upon LPS and/or IFNγ treatment (Fig. 1B).

Figure 1.

MNSFβ down-regulates LPS/IFNγ-induced apoptosis. (A) Raw264.7 cells were transfected with siRNA directed against MNSFβ or scrambled siRNA. After 48 h of siRNA transfection, Raw264.7 cells were stimulated with LPS (1 μg·mL−1) plus IFNγ (100 U·mL−1) for 24 h, and stained using APOPercentage™ (right panel). Stained cells were analyzed as described in Experimental procedures (left panel). Values are means ± SD of triplicate samples for one of three independent experiments with similar results. The asterisk indicates a statistically significant difference (< 0.05) versus treatment with control siRNA. MNSFβ mRNA expression was analyzed by RT–PCR after treatment with siRNAs for 48 h (middle panel). (B) Raw264.7 cells were transfected with MNSFβ and mutant MNSFβ (G74A) with or without Bcl-G or mutant Bcl-G (K110R). Raw264.7 cells were stimulated 48 h after transfection and analyzed as described in (A) (upper panel). Values are means ± SD of triplicate samples for one of three independent experiments with similar results. Asterisks indicate statistically significant differences (< 0.05) versus treatment with control cDNA. The presence of MNSFβ and Bcl–G was detected by western blotting (lower panel). (C) The MNSFβ–Bcl–G complex was detected by immunoprecipitation and western blot analyses, as described in Experimental procedures. The same membrane was reprobed with anti-Bcl–G Ig.

MNSFβ enhances NO-induced apoptosis

It has been reported that LPS/IFNγ-induced apoptosis in Raw264.7 cells is mediated by NO [24]. Thus, we focused on NO-induced apoptosis in Raw264.7 cells. We first examined the effect of MNSFβ siRNA on iNOS expression in LPS/IFNγ-activated Raw264.7 cells. Western blot analysis showed that iNOS expression was not affected by treatment with MNSFβ siRNA (Fig. 2A). To confirm the results observed by western blot analysis, we determined the effect of siRNAs directed against MNSFβ and Bcl–G on NO production in LPS/IFNγ-activated cells. No difference was observed between control and knockdown cells (Fig. 2B). To further study the mechanism of action of MNSFβ, we examined the effect of MNSFβ siRNA on apoptosis in cells treated with the NO donor S–nitrosoglutathione (GSNO)-treated cells. As shown in Fig. 2C, NO-induced apoptosis was significantly inhibited in cells transfected with MNSFβ siRNA. In contrast, the effect of Bcl–G siRNA was small. Double knockdown of MNSFβ and Bcl–G strongly reduced NO-induced apoptosis. Co-transfection with MNSFβ and Bcl–G strongly enhanced LPS/IFNγ- induced apoptosis, as described above (Fig. 1B). Thus, our data suggest that the MNSFβ–Bcl–G complex may be involved in the signal pathway downstream of NO production in LPS/IFNγ-stimulated macrophages.

Figure 2.

MNSFβ down-regulates NO-induced apoptosis. (A) Raw264.7 cells transfected with siRNAs were stimulated with LPS (1 μg·mL−1) plus IFNγ (100 U·mL−1) for 24 h. Cell extracts were subjected to immunoblot analysis with anti-iNOS Ig, or anti-α-tubulin. (B) Raw264.7 cells transfected with siRNAs were stimulated with LPS (1 μg·mL−1) plus IFNγ (100 U·mL−1) for 24 h. NO production in the culture medium was determined as described in Experimental procedures (left panel). Values are means ± SD of triplicate samples. Bcl–G mRNA expression was analyzed by RT–PCR after treatment with siRNAs for 48 h (right panel). (C) Raw264.7 cells were transfected with siRNA directed against MNSFβ or Bcl–G. After 48 h of siRNA transfection, Raw264.7 cells were stimulated with 1 mm GSNO for 24 h and stained using APOPercentage™. Stained cells were analyzed as described in the legend to Fig. 1A. Values are means ± SD of triplicate samples. Asterisks indicate statistically significant differences (< 0.05).

Special emphasis was placed on p53, because the BCL–G gene is a pro-apoptotic p53 target gene [14]. In addition, NO induces p53 accumulation in Raw264.7 cells [25, 26]. Double knockdown of Bcl–G and MNSFβ resulted in a marked decrease in the amount of p53 in LPS/IFNγ-stimulated cells (Fig. 3A). Conversely, expression of p53 was strongly increased in the cells co-transfected with MNSFβ and Bcl–G (Fig. 3B). We also observed inhibition of both poly (ADP-ribose) polymerase (PARP) cleavage and caspase–3 activation by the double knockdown of Bcl–G and MNSFβ (Fig. 3C,D). These results indicate that the MNSFβ–Bcl–G complex may act upstream of p53 and positively regulate the level of p53.

Figure 3.

Enhancement of NO-induced p53 accumulation by MNSFβ. Raw264.7 cells transfected with the indicated siRNA (A, C, D) or cDNA (B) were stimulated with LPS (1 μg·mL−1) plus IFNγ (100 U·mL−1) for 24 h. Following total protein extraction, equal amounts of protein were subjected to 12% SDS/PAGE, blotted onto polyvinylidene fluoride membranes, and probed using antibodies against p53, caspase–3, PARP, α–tubulin or β–actin. The presence of free MNSFβ was detected by western blotting (B).

Cox–2 expression is involved in MNSFβ-mediated apoptosis regulation

Because it has been reported that induction of Cox–2 represents an essential regulator of NO-mediated apoptosis in Raw264.7 cells [27], we investigated whether Cox–2 activation is involved in an anti-apoptosis mechanism in which MNSFβ and Bcl–G form a complex in macrophages. Raw264.7 cells were transfected with MNSFβ cDNA with or without Bcl–G cDNA. After 48 h, Raw264.7 cells were treated with 10 μg·mL−1 LPS for 24 h. Then western blotting was performed as described in Experimental procedures. To simplify interpretation of the involvement of Cox–2 expression in the pro-apoptotic mechanism of MNSFβ, the transfected cells were stimulated with LPS alone. As can be seen in Fig. 4A, ectopic over-expression of MNSFβ cDNA significantly reduced the Cox–2 level in LPS-stimulated cells. Co-over-expression of MNSFβ and Bcl–G led to strong inhibition of the Cox–2 expression. Bcl–G over-expression did not affect the expression of Cox–2. We did not observe significant changes in the level of Cox–2 expression in cells transfected with mutant MNSFβ (G74A) cDNA. To confirm the pro-apoptotic function of MNSFβ, Raw264.7 cells transfected with MNSFβ cDNA were stimulated with 1 μg·mL−1 LPS plus 100 U·mL−1 IFNγ for 24 h, and Cox–2 expression was analyzed by western blot. As shown in Fig. 4B, Cox–2 expression was significantly decreased in MNSFβ cDNA-transfected cells. Co-over-expression of MNSFβ and Bcl–G potently reduced the Cox–2 expression. We next examined whether MNSFβ–Bcl–G inhibits GSNO-induced Cox–2 expression. Cox–2 protein was induced by GSNO at concentrations of 200 μm (data not shown) and 1 mm (Fig. 4C). Cox–2 expression was remarkably inhibited by co-over-expression of MNSFβ and Bcl–G (Fig. 4C). These results clearly demonstrate that MNSFβ–Bcl–G promotes apoptosis partly through inhibiting Cox–2 expression.

Figure 4.

Inhibition of LPS/IFNγ-induced Cox–2 expression by MNSFβ. (A) Raw264.7 cells were transfected with MNSFβ with or without Bcl–G. After over-expression, Raw264.7 cells were stimulated with LPS (10 μg·mL−1) for 24 h, and the presence of Cox–2 protein in the cell lysates was detected by Western blotting. (B) Raw264.7 cells were transfected with MNSFβ or mutant MNSFβ (G74A). After over-expression, Raw264.7 cells were stimulated with LPS (1 μg·mL−1) plus IFNγ (100 U·mL−1) for 24 h, and the presence of Cox–2 protein in the cell lysates was detected by western blotting. The blots also show the over-expression level of free MNSFβ. (C) Raw264.7 cells were transfected with MNSFβ or mutant MNSFβ (G74A). After over-expression, Raw264.7 cells were stimulated with GSNO (1 mm) for 24 h, and the presence of Cox-2 protein in the cell lysates was detected by western blotting. The blots also show the over-expression level of free MNSFβ.

To further investigate the mechanism of Cox–2 regulation by MNSFβ–Bcl–G, we examined the signal transduction pathway leading to the induction of Cox–2 expression. ERK activity has been implicated in the induction of Cox–2 expression in response to LPS or the NO donor GSNO in Raw264.7 cells [28, 29]. We have previously shown that MNSFβ covalently binds to Bcl–G and regulates the ERK–MAPK cascade [16]. Thus, we focused on the involvement of MNSFβ–Bcl–G-mediated ERK regulation in Cox–2 production by LPS-stimulated Raw264.7 cells. The level of ERK activation was evaluated by western blotting using phosphospecific antibody. In Raw264.7 cells, LPS-stimulated ERK1/2 phosphorylation peaked at 20 min after LPS stimulation as described previously [16]. In Raw264.7 cells co-transfected with MNSFβ and Bcl–G cDNA, LPS-induced ERK1/2 phosphorylation was strongly inhibited (Fig. 5A), and LPS-induced p38 and Jun N-terminal kinase (JNK) phosphorylation were unaffected (data not shown). It should be noted that MNSFβ siRNA enhanced LPS-induced ERK1/2 phosphorylation without affecting p38 and JNK phosphorylation [16]. In Raw264.7 cells co-transfected with mutant MNSFβ (G74A) and Bcl–G (K110R) cDNA, the phosphorylation of ERK1/2 was not affected. We next investigated whether MNSFβ–Bcl–G affects GSNO-induced ERK activation. In preliminary experiments, we observed that ERK1/2 phosphorylation peaked at 8 h, and then decreased to the baseline level after treatment. As shown in Fig. 5B, ERK1/2 phosphorylation was significantly inhibited in MNSFβ cDNA-transfected cells. Co-over-expresion of MNSFβ and Bcl–G strongly inhibited GSNO-induced ERK1/2 phosphorylation. This inhibition was not observed when Gly74 of MNSFβ was mutated to alanine. Thus, the MNSFβ–Bcl–G complex functions as a negative regulator of Cox–2 expression, probably by down-regulating ERK activity following LPS or GSNO treatment.

Figure 5.

MNSFβ conjugates to Bcl–G and regulates NO-induced ERK activity. Raw264.7 cells were transfected with MNSFβ with or without Bcl–G. After over-expression, Raw264.7 cells were treated with 1 μg·mL−1 LPS for 20 min (A) or 1 mm GSNO for 8 h (B). The presence of phospho-ERK1/2 and total ERK1/2 in the cell lysates was detected by western blotting. The intensity of the signals as determined by densitometric scanning is expressed as the fold change relative to that of untreated cells. Values are means ± SD (= 4). A representative autoradiograph is shown. Asterisks indicate statistically significant differences (< 0.05) versus the controls. The blots also show the over-expression level of free MNSFβ.

NFκB and AP–1 are involved in the LPS signaling cascade leading to Cox–2 production [30]. We have previously reported that MNSFβ slightly affects LPS-induced NFκB signaling [16]. In this study, an electrophoretic mobility shift assay (EMSA) was performed to determine whether MNSFβ is involved in LPS-induced AP–1 activation. MNSFβ siRNA-transfected Raw264.7 cells were treated with 1 μg·mL−1 LPS for 1 h. Then an EMSA was performed as described in Experimental procedures. AP–1 activation in the transfected cells was significantly enhanced (Fig. 6). Thus, it is conceivable that MNSFβ affects ERK/AP–1 signaling rather than NFκB signaling. We also examined whether Bcl–G siRNA affects LPS-induced AP–1 activation in Raw264.7 cells. Bcl–G siRNA transfection did not significantly affect AP–1 activation induced by LPS. Double knockdown of MNSFβ and Bcl–G strongly enhanced AP–1 activation.

Figure 6.

MNSFβ siRNA enhances LPS-induced AP–1 activation. Raw264.7 cells were transfected with MNSFβ siRNA with or without Bcl–G siRNA. After 48 h of siRNA transfection, Raw264.7 cells were treated with 1 μg·mL−1 LPS for 60 min. An EMSA was performed using a consensus AP–1 probe in the presence or absence of a 100-fold excess of unlabeled competitor. A supershifted band in lane 5 is indicated by an arrowhead. In lane 6, unlabeled oligonucleotide was added. A representative blot is shown in the upper panel. Means ± SD of three experiments are shown in the lower panel. Asterisks indicate statistically significant differences (< 0.05) versus the control (LPS plus scrambled).

MNSFβ promotes LPS/IFNγ-induced apoptosis in peritoneal macrophages

To further investigate the pro-apoptotic mechanism of MNSFβ–Bcl–G, we prepared peritoneal macrophages from mice and examined LPS/IFNγ-induced apoptosis in these cells. Co-over-expresion of MNSFβ and Bcl–G caused a marked enhancement of apoptosis (Fig. 7A). In contrast, co-over-expression of mutant MNSFβ (G74A) and mutant Bcl–G (K110R) did not affect LPS/IFNγ-induced apoptosis. Furthermore, western blot analysis showed that Cox–2 expression was strongly inhibited by co-over-expression of MNSFβ and Bcl–G (Fig. 7B). Thus, our observation that MNSFβ–Bcl–G enhances LPS/IFNγ-induced apoptosis in Raw264.7 cells is confirmed in primary cultured peritoneal macrophages.

Figure 7.

MNSFβ–Bcl–G down-regulates LPS/IFNγ-induced apoptosis in murine peritoneal macrophages. (A) Peritoneal macrophages were transfected with MNSFβ with or without Bcl–G. After over-expression, the macrophages were stimulated with LPS (1 μg·mL−1) plus IFNγ (100 U·mL−1) for 24 h. Apoptotic cells were analyzed as described in Experimental procedures. Values are means ± SD of triplicate samples. Asterisks indicate statistically significant differences (< 0.05) versus control treatment (LPS/IFNγ plus empty pcDNA3.1 vector). (B) Peritoneal macrophages were transfected with MNSFβ with or without Bcl–G. After 48 h transfection, cells were stimulated with LPS (10 μg·mL−1) for 24 h, and the presence of Cox–2 protein in the cell lysates was detected by western blotting. The blots also show the over-expression level of free MNSFβ.

Discussion

Although LPS/IFNγ has been demonstrated to induce apoptosis in murine macrophages, the mechanism is not completely understood. Cox–2 expression protects Raw264.7 cells from apoptosis induced by LPS/IFNγ [29]. In this study, we show that ubiquitin-like protein MNSFβ enhances LPS/IFNγ-induced apoptosis in Raw264.7 cells and murine peritoneal macrophages. MNSFβ covalently binds to the pro-apoptotic protein Bcl–G and promotes LPS/IFNγ-induced apoptosis by down-regulating Cox–2 activation. Transfection experiments using a point mutation showed that covalent interaction between MNSFβ and Bcl–G enhances this pro-apoptotic activity (Fig. 1B). We also showed that double knockdown of MNSFβ and Bcl–G enhanced Cox–2 expression in Raw264.7 cells activated with LPS alone (Fig. 4A). MNSFβ–Bcl–G may be also implicated in the Cox–2 pathway, which does not involve iNOS. We present data showing that MNSFβ–Bcl–G regulates the signal transduction pathway downstream of NO production in LPS/IFNγ-stimulated macrophages. Indeed, siRNA directed against MNSFβ and Bcl–G inhibited apoptosis in Raw264.7 cells stimulated using the NO donor GSNO (Fig. 2). It has been reported that Cox–2 expression and p53 accumulation are inversely related in Raw264.7 macrophages [29]. Consistent with this report, we observed that p53 accumulation was accompanied by a reduction in Cox–2 expression in LPS/IFNγ-stimulated Raw264.7 cells transfected with MNSFβ cDNA (Fig. 3B and Fig. 4B).

Several ubiquitin-like proteins are involved in IFN signaling: MNSFβ, FAT10, ISG15, and NUB1L are induced by IFNγ [18-22]. Among them, FAT10 is closely involved in apoptosis [23]. IFNγ induces the formation of a number of MNSFβ adducts in T cells [18] and macrophages (unpublished data). However, in this study, we have shown that the level of MNSFβ–Bcl–G did not change upon IFNγ treatment. Thus, it is conceivable that MNSFβ target molecule(s) other than Bcl–G may be responsible for the regulation of apoptosis in macrophages. We have previously demonstrated that post-translational modification of endophilin II by MNSFβ is involved in Dectin–1-mediated phagocytosis and inflammatory responses in macrophages [12]. Other endophilin family proteins such as endophilin A1 and endophilin B1 are involved in the ubiquitin or ubiquitin-like conjugation systems [31]. Endophilin B1 interacts with BAX, a pro-apoptotic Bcl–2 family member, and regulates Bcl–2-mediated programmed cell death [32]. Investigations are underway to determine whether MNSFβ–endophilin II is responsible for apoptosis regulation.

In this paper, we present data showing that MNSFβ covalently binds to the target protein Bcl–G and enhances apoptosis in macrophages. Likewise, the ubiquitin-like protein FAT10 forms covalent conjugates to target proteins and induces apoptosis. Uba6 (ubiquitin-like modifier activating 6) activates ubiquitin and FAT10 [33]. It appears unlikely that Uba6 recognizes and activates MNSFβ because formation of a complex between MNSFβ and Uba6 has not been observed in lysates from various tissues and cell lines, including Raw264.7. However, we cannot rule out the possibility that other ubiquitin-like activating enzymes may be involved in MNSFβ conjugation. Despite repeated attempts, we were unable to isolate and purify MNSFβ conjugation enzyme(s) because of the highly aggregative character of recombinant MNSFβ [9].

LPS/IFNγ–induced macrophage activation leads to a variety of signal transduction pathways, including all three MAPK cascades, i.e. the ERK, p38 and JNK pathways. Transfection experiments showed that MNSFβ–Bcl–G down-regulates LPS- or NO-induced ERK1/2 phosphorylation (Fig. 5). This observation suggests that MNSFβ–Bcl–G is involved in both early- and late-phase ERK1/2 phosphorylation leading to Cox–2 activation (Fig. 8). MNSFβ–Bcl–G directly binds to ERK1/2 and inhibits ERK1/2 phosphorylation by mitogen-activated protein kinase/extracellular signal-regulated kinase kinase (MEK), as described previously [16]. The interactions between p53 and the Raf/MEK/ERK are very complicated and probably feedback upon one another [34]. MNSFβ-mediated ERK regulation may explain an inverse expression of p53 and Cox–2 in Raw264.7 cells. Although the JNK pathway has been shown to play a critical role in stress-induced apoptosis in various cell types, neither JNK nor p38 MAPK phosphorylation were affected by MNSFβ, as previously described [16]. Thus, MNSFβ–Bcl–G may regulate LPS/IFNγ-induced apoptosis in macrophages partly through inhibition of ERK phosphorylation.

Figure 8.

Intracellular signaling pathways involving inhibition of Cox–2 by the MNSFβ–Bcl–G complex. MNSFβ–Bcl–G inhibits both early- and late-phase ERK phosphorylation leading to Cox–2 activation.

It has been demonstrated that the expression of Cox–2 induced by LPS requires NFκB activation in Raw264.7 cells [28]. NFκB is activated by ubiquitin-dependent degradation of its inhibitory partner IκBα. We have shown that MNSFβ slightly inhibits LPS-induced NFκB activation [16]. More recently, we have also shown that MNSFβ siRNA enhanced the degradation of inhibitor of kappa B α (IκBα) induced by zymosan [12]. Unlike ubiquitin, attachment of MNSFβ does not appear to target proteins for degradation, and instead has been proposed to change the ability of the modified protein to react with other cellular proteins, including Bcl–G. Thus, it seems unlikely that MNSFβ directly regulates the degradation of IκBα protein. Instead, we demonstrated that MNSFβ negatively regulates the LPS-induced ERK/AP–1 signaling cascade leading to Cox–2 activation (Fig. 6). Careful experiments are required to clarify the role of MNSFβ in LPS-induced AP–1 activation.

Taken as a whole, the present study demonstrates that MNSFβ–Bcl–G formation may be important for potent regulation of apoptosis in macrophages.

Experimental procedures

Materials

Rabbit polyclonal antibodies to Cox–2, iNOS, p53, caspase–3, ERK1/2 and phospho-ERK1/2 were purchased from Cell Signaling Technology (Danvers, MA, USA). Rabbit polyclonal antibodies to MNSFβ and Bcl–G were prepared as described previously [9, 15]. Anti-Fos Ig was obtained from Santa Cruz Biotechnology (Dallas, TX, USA).

Cell culture, siRNAs and transfection of cells

The Raw264.7 macrophage-like cell line (TIB–71, American Type Culture Collection, Manassas, VA, USA) was cultured routinely in Dulbecco's modified Eagle's medium with 10% fetal bovine serum and penicillin 100 U/mL, streptomycin 100 μg/mL at 37 °C and 5% CO2. siRNA duplexes (siRNAs) were synthesized and purified by Qiagen (Chatsworth, CA, USA). Two different siRNAs were employed for MNSFβ knockdown (Qiagen ID SI00999525 and ID SI00999544; termed MNSFβ–1 and MNSFβ–2, respectively) and for Bcl–G knockdown (Qiagen ID SI00928179 and ID SI00928193; termed Bcl–G–1 and Bcl–G–2, respectively). Scrambled siRNA directed against 5′-GGACTCGACGCAATGGCGTCA-3′ was used as a negative control [16]. Cells were treated with siRNA according to the instructions provided with the HiPerFect transfection reagent (Qiagen). Raw264.7 cells (1.2 × 105) were treated with 3 μg of siRNA in RPMI–1640 medium (Sigma Aldrich, St. Louis, MO, USA) supplemented with 10% fetal bovine serum in the presence of HiPerFect transfection reagent. After a 48 h incubation at 37 °C, medium containing the mixture of HiPerFect and siRNA was replaced by Dulbecco's modified Eagle's medium containing 10% fetal bovine serum, and cells were incubated for the indicated periods of time.

Apoptosis assay

Apoptotic cells were stained using an APOPercentage™ apoptosis assay (Biocolor, Carrickfergus, UK) according to the manufacturer's instructions. Stained cells were analyzed using imagej software (http://rsb.info.nih.gov/ij/) and Adobe Photoshop (San Jose, CA, USA).

Nitric oxide detection

NO production in culture medium was determined using 50 μL Griess reagent (Sigma, St Louis, MO, USA) added to 50 μL of culture medium. After 15 min of incubation at room temperature under dark conditions, the nitrite concentration was measured at 540 nm on a plate reader. Nitrite concentrations were calculated by comparison with a nitrite standard reference curve.

Western blot analysis

Whole-cell lysates were prepared by extracting proteins using a buffer containing 50 mm Tris/HCl, pH 7.4, 150 mm NaCl, 1% NP–40 (Sigma-Aldrich) and 0.1% SDS supplemented with protease inhibitors. The protein concentrations of the cell lysates were determined by Bradford assay (Bio–Rad, Hercules, CA, USA). Equal amounts of protein were subjected to 12% SDS/PAGE and transferred onto polyvinylidene fluoride membranes. The membrane was incubated overnight at 4 °C in Tris-buffered saline solution with 5% milk to block non-specific binding sites. Membranes were incubated with the primary antibodies for a minimum of 2 h at room temperature in Tris-buffered saline with 0.1% Tween–20 (Tris/Tween). Horseradish peroxidase secondary antibodies were incubated for 1 h at room temperature in Tris/Tween with 5% milk. Labeled proteins were visualized by chemiluminescence according to the manufacturer's instructions (GE Healthcare, Waukesha, WI, USA).

Immunoprecipitation

Immunoprecipitation was performed using a horseradish peroxidase-conjugated antibody that recognizes native rabbit IgG (Rabbit TrueBlot®, eBioscience, Inc., San Diego, CA, USA), according to the manufacturer's instructions. Extracts of Raw264.7 cells in RIPA buffer (50 mm Tris, 1% Nonidet P–40, 0.25% deoxycholate, 150 mm NaCl, 1 mm EDTA, 1 mm phenylmethylsulfonyl fluoride, pH 7.4, containing 1 μg·mL−1 each of the protease inhibitors aprotinin, leupeptin and pepstatin) were pre-cleared using 50 μL anti-rabbit IgG beads for 1 h on ice. Subsequently, 5 μg of primary antibody against Bcl–G was added to pre-cleared lysates, which were incubated on ice for an additional 1 h. Samples were then incubated overnight at 4 °C with 50 μL anti-rabbit IgG beads. The beads were washed five times with RIPA buffer (2 500 g for 30 seconds at 4 °C), and immunoprecipitates were released from the beads by boiling for 10 min in NuPAGE LDS sample buffer (Invitrogen, Carlsbad, CA, USA). Immunoblotting was performed using antibody against MNSFβ. A Rabbit IgG TrueBlot® was used as a second antibody.

RT–PCR

RT–PCR was performed for 30 cycles as described previously [15]. The PCR primers used to detect mRNA are as follows: MNSFβ, 5'-CGCCCAGGAACTACACACC-3' (sense) and 5'-GCCTGCTACTTCCAGAGTGG-3' (antisense) (222 bp); Bcl–G, 5'-CCCAAGCTCTCCAGAACAAG-3' (sense) and 5'-CTGAGCTCGGATCTCCTTTG-3' (antisense) (213 bp). All amplified PCR products were isolated and sequenced to verify their identity. PCR products were separated by 2% agarose gel electrophoresis and stained with ethidium bromide.

Mutagenesis and transfection

Mutant Bcl–G (K110R) was generated by replacing the codon for lysine 110 with the codon for arginine, as previously described [15]. cDNAs encoding MNSFβ and Bcl–G were sub-cloned into the vector pcDNA3.1(+) (Invitrogen). Transient DNA transfections were performed using Lipofectamine Plus reagent (Invitrogen) according to the manufacturer's instructions and using 5 μg plasmid DNA per six-well plate.

Electrophoretic mobility shift assay (EMSA)

Nuclear extracts were prepared as described by Dignam et al. [22]. Protein–DNA complexes were detected using biotin end-labeled double-stranded DNA probes. The sequence for the AP–1 site was CGCTTGATGACTCAGCCGGAA. The binding reaction was performed using an EMSA ‘Gel Shift’ kit according to the manufacturer's instructions (Affymetrix, Santa Clara, CA, USA). The reaction products were separated on a 6% polyacrylamide gel in 0.5% Tris borate/EDTA, transferred onto a nylon membrane, and fixed on the membrane by UV cross-linking. The biotin-labeled probe was detected using chemiluminescence. In a supershift assay, 2 μg anti-Fos Ig was added to EMSA reaction mixtures after the binding reaction, followed by incubation at 25 °C for 30 min prior to electrophoresis.

Collection of peritoneal macrophages

Peritoneal macrophages were obtained using BALB/c mice injected 4 days previously with 2 mL of a sterile 4% Brewer's thioglycollate broth (Difco, Detroit, MI, USA). The cells were collected by centrifugation at 400 g for 5 min, washed, and resuspended in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum. The collected macrophages were used for apoptosis studies. All experiments were approved and performed in accordance with the guidelines of the Animal Care Committee of Shimane University.

Statistical analysis

Statistical significance was analyzed by Student's t test and expressed as a P value.

Acknowledgements

This work was supported by a grant-in-aid for scientific research (C) to M.N. from the Ministry of Education, Culture, Sports, Science and Technology of Japan.

Ancillary