Romo1 is a negative-feedback regulator of Myc

Myc protein levels are increased in response to mitogenic stimuli to stimulate G1–S phase progression of the cells, and the expression level of Myc is tightly controlled through transcriptional, translational and post-translational mechanisms (Arnold and Sears, 2008). Although Myc transcription is induced during G0–G1 transition, fine modulation of the Myc protein level occurs at the post-translational level through regulation of its stability (Bhatia et al., 1993). The main mechanism for Myc degradation involves ubiquitin-mediated proteolysis (Gross-Mesilaty et al., 1998). Myc is polyubiquitylated by E3 ubiquitin ligases, including the F-box proteins Fbw7 and Skp2, and a series of sequential phosphorylation events is required for Fbw7-mediated proteasomal degradation. Phosphorylation sites of Myc include Thr58 and Ser62. Phosphorylation at Ser62 occurs via the Ras–Raf–MEK–ERK pathway (Alvarez et al., 1991; Seth et al., 1991). Ras activation also inhibits the phosphatidylinositol-3-OH-kinase (PI3K)/Akt pathway to inhibit GSK-3β, resulting in stabilization of the Myc protein (Cross et al., 1995; Sears et al., 2000). During the later stages of G1, Ras activity decreases, and GSK-3β is reactivated to phosphorylate Myc at Thr58 (Henriksson et al., 1993; Saksela et al., 1992). Myc phosphorylated at Thr58 is ubiquitylated by the SCF^Fbw7 ubiquitin machinery for degradation by the 26S proteasome (Welcker et al., 2004; Yada et al., 2004).

Another important mechanism for ubiquitin-mediated proteasomal degradation of Myc is the Skp2 pathway (Kim et al., 2003; von der Lehr et al., 2003). Skp2 is also reported to ubiquitylate Cdk inhibitors and tumor suppressor proteins such as p27^Kip1 (Carrano et al., 1999), p57^Kip2 (Kamura et al., 2003), p130 (Tedesco et al., 2002) and Tob1 (Hiramatsu et al., 2006). Skp2 is overexpressed in cancer (Bashir and Pagano, 2003). The expressions of Skp2 and Myc are induced by mitogenic stimulation; however, Skp2 expression continues into S phase (Lisztwan et al., 1998). Skp2 interacts with two domains of Myc (residues 129–147: the N-terminal Myc box II domain and residues 379–418: the C-terminal bHLHZip domain) at the G1 to S phase transition to induce Myc degradation and turnover (Kim et al., 2003; von der Lehr et al., 2003). Skp2-mediated ubiquitylation does not correlate with Myc phosphorylation because Myc mutated at Thr58 is well ubiquitylated (Kim et al., 2003).

Romo1 (reactive oxygen species modulator 1) was first identified in 2006, and forced expression of Romo1 increases the level of cellular ROS that originate from mitochondria (Chung et al., 2006). Romo1 is localized to the mitochondria and releases mitochondrial ROS through complex III of the mitochondrial electron transport chain (Chung et al., 2008). Although the ROS produced by cytosolic enzymes such as NADPH oxidase have a role in cell proliferation, mitochondrial ROS are not known to be involved in cell proliferation. Recently, we reported that ROS originating from the endogenous Romo1 protein are necessary for both normal and cancer cell proliferation (Chung et al., 2009; Na et al., 2008). Suppression of Romo1 expression inhibits cell growth through inhibition of ERK activation and p27^Kip1 expression, demonstrating that ROS derived from Romo1 are required for cell proliferation. Romo1 expression is enhanced in senescent cells and in most cancer cells (Chung et al., 2006; Chung et al., 2008). Romo1 is also upregulated by serum deprivation and contributes to the serum-deprivation-mediated increase in ROS (Lee et al., 2010). Furthermore, a recent paper demonstrated that Romo1 modulates ROS production in the mitochondria (Kim et al., 2010). Romo1 recruits the anti-apoptosis regulator Bcl-X_L to decrease the mitochondrial membrane potential in response to tumor necrosis factor-α (TNF-α), resulting in ROS production. Although many studies have been conducted on Romo1, its physiological function is not well elucidated. In the present study, we investigated the role of Myc-induced Romo1 in Myc turnover after serum stimulation.

Myc has been reported to stimulate ROS generation, which in turn induces DNA damage (Vafa et al., 2002). Romo1 has been also reported to recruit Bcl-X_L to reduce the mitochondrial membrane potential in response to TNF-α, resulting in mitochondrial ROS generation (Chung et al., 2006; Kim et al., 2010). Therefore, we investigated the correlation between Myc and Romo1 after serum stimulation. The expressions of Myc, Romo1 and p27^Kip1 (CDKN1B) were examined after addition of serum to cultures of normal human lung fibroblasts (IMR-90 and WI-38 cells) and human embryo kidney cells (HEK293). Low basal levels of Myc and Romo1 and high levels of p27^Kip1 were detected in unstimulated cells by western blot analysis (Fig. 1A). When the cells were stimulated with serum, Myc expression was induced at 1 hour and its level peaked at 3–6 hours after serum treatment. Interestingly, Romo1 expression was enhanced after Myc induction, peaking at 9–24 hours (Fig. 1A). By contrast, p27^Kip1 was downregulated at 6 hours after serum stimulation.

To observe whether Myc induced Romo1 expression, HEK293 and HeLa cells were transfected with Myc, and Romo1 expression was observed by western blot analysis. As shown in Fig. 1B, Myc increased the Romo1 protein level, demonstrating that Romo1 is downstream of Myc. Next, we examined whether serum-stimulated Myc expression also increased Romo1 expression. To observe whether knockdown of Myc blocked serum-induced Romo1 expression, WI-38 cells and IMR-90 cells were transfected with MYC siRNA. MYC siRNA transfection efficiently inhibited Myc induction by serum stimulation (Fig. 1C), and Myc knockdown blocked the serum-induced Romo1 expression (Fig. 1D). Furthermore, semi-quantitative RT-PCR and real-time PCR analyses demonstrated that serum stimulation upregulated ROMO1 expression transcriptionally (Fig. 1E). These results demonstrate that serum stimulation enhances Myc-mediated Romo1 expression.

To investigate whether knockdown of Romo1 suppressed serum-induced ROS generation, IMR-90 cells were transfected with ROMO1 siRNA and ROS levels were measured by staining the cells with MitoSOX, a probe for superoxide in the mitochondria. First, Romo1 knockdown in cells transfected with ROMO1 siRNA was examined by western blot analysis (supplementary material Fig. S1A). To exclude off-target effects of the ROMO1 siRNA construct, a rescue experiment was performed. The N-terminal deletion mutant of Romo1 (Romo1-ΔN), which does not include the ROMO1 siRNA-1 sequence, is known to induce mitochondrial ROS generation (Kim et al., 2010). ROMO1 siRNA-1 transfection efficiently caused Romo1 knockdown and decreased ROS levels in various cell lines (Hwang et al., 2007; Na et al., 2008). However, Romo1-ΔN was resistant to ROMO1 siRNA-1 (supplementary material Fig. S1B). In the present study, we showed that both wild-type Romo1 and Romo1-ΔN decreased Myc expression, but Romo1-ΔC did not (Fig. 4D). Therefore, we examined whether Romo1-ΔN can downregulate Myc expression in cells transfected with ROMO1 siRNA-1. As shown in supplementary material Fig. S1C, FLAG–Romo1 (wt) decreased Myc expression and ROMO1 siRNA-1 transfection blocked Romo1-induced Myc downregulation. However, ROMO1 siRNA-1 transfection failed to suppress Romo1-ΔN-induced Myc downregulation. This result showed the specificity of the ROMO1 siRNA construct. Next, we examined whether Romo1 knockdown suppressed serum-induced ROS production. As shown in Fig. 2A, high ROS levels were observed in serum-deprived cells and Romo1 knockdown inhibited serum deprivation-induced ROS production. The ROS increases were suppressed by serum addition. These results are consistent with a previous report (Lee et al., 2010). In Fig. 1A, we showed that serum stimulation increased Romo1 expression. This Romo1 induction was also observed by fluorescence microscopy (Fig. 2A). The serum-stimulated ROS increases were completely blocked by Romo1 knockdown (Fig. 2A). The decreases in Romo1 expression and ROS formation in primary human fibroblasts IMR-90 (Fig. 2B,C) or WI-38 (supplementary material Fig. S2A) were quantified using MetaMorph software. In Fig. 1D, we showed that knockdown of Myc blocked the serum-induced Romo1 expression. Therefore, we also measured the ROS formation after serum stimulation in cells transfected with MYC siRNA. As shown in supplementary material Fig. S2B, the serum-stimulated ROS increases were also blocked by Myc knockdown. Next, we examined Myc expression in IMR-90, WI-38 and HEK293 cells transfected with ROMO1 siRNA. Interestingly, Romo1 knockdown blocked the elimination of Myc after 6 hours of serum stimulation (Fig. 2D,E). Instead, Myc expression was gradually increased until 24 hours. From this result, we suggest that the Romo1 expression induced by Myc during G1 phase is necessary for elimination of Myc and we assume that increased Romo1 expression might be involved in Myc degradation in a negative-feedback mechanism.

Although Myc expression was increased in cells transfected with ROMO1 siRNA after 6 hours of serum stimulation compared with control cells, its expression was very low at early times after mitogenic stimulation (0–3 hours, Fig. 2D). Recently, we reported that ROS derived from Romo1 expression also regulate cell proliferation through activation of ERK in various normal and cancer cell lines (Chung et al., 2009; Na et al., 2008). Therefore, we examined whether ROS derived from Romo1 expression were required for induction of Myc in early G1 phase. As shown in Fig. 3A, Myc induction for 2 hours after serum stimulation was suppressed by Romo1 knockdown. Treatment with antioxidants also inhibited Myc induction. However, hydrogen peroxide (H₂O₂) treatment of cells transfected with ROMO1 siRNA recovered Myc expression (Fig. 3B). Myc expression was also examined using various kinase inhibitors. MEK1/2-specific inhibitors, PD98059 and U0126, blocked Myc induction and ERK activation (Fig. 3B). These results demonstrate that ROS derived from Romo1 in response to serum stimulation are necessary for Myc induction and cell cycle progression.

To further investigate the correlation between Romo1 expression and cell cycle transition triggered by serum stimulation, flow cytometric analysis was carried out in IMR-90 cells transfected with ROMO1 siRNA. In this experiment, Romo1 knockdown delayed cell cycle progression into S phase (Fig. 3C). This finding was also confirmed in WI-38 cells (supplementary material Fig. S2C). These results indicate that Romo1 expression has an important role in cell cycle entry triggered by mitogenic stimulation via ERK activation and Myc induction. We also suggest that the basal level of ROS derived from the steady-state level of Romo1 is required for ERK activation and Myc stabilization. By contrast, enhanced ROS levels generated from Romo1 expression, which is induced by Myc, trigger the elimination of Myc.

As shown in Fig. 2D, knockdown of Romo1 blocked the elimination of Myc. Therefore, we investigated whether increased Romo1 expression caused Myc downregulation. Romo1 was transfected into HeLa cells, and Myc expression was measured by western blot and immunofluorescence analysis. Romo1 overexpression triggered downregulation of Myc (Fig. 4A and supplementary material Fig. S3A). To confirm this finding, cells were co-transfected with Myc and Romo1, and Myc expression was again measured by western blot analysis. As shown in Fig. 4B, Romo1 also induced the downregulation of Myc that was expressed exogenously. Expression of Romo1 also decreased the expression of Myc in Huh-7, HeLa, A549 and H1299 cells (supplementary material Fig. S3B).

Next, we analyzed which domain of Myc was responsible for its expression. Romo1 and Myc deletion constructs (Herbst et al., 2004; Tworkowski et al., 2002) were co-transfected into the cells, and Myc expression was measured by western blot analysis. We found that Romo1 promoted the downregulation of wild-type (WT) Myc, a ΔA construct including Myc box (Mb) I, a ΔB construct, a ΔE construct and a ΔG construct. However, Romo1 failed to downregulate a ΔC construct including the Mb II domain, a ΔD construct including the PEST domain, and a ΔF construct (Fig. 4C). This result demonstrates that the Mb I domain, which is needed for Fbw7-mediated proteasomal degradation of Myc, is not required for Myc degradation and that another mechanism exists for ubiquitin-mediated proteasomal degradation of Myc. Recently, we reported that the C-terminal region of Romo1 is important for TNF-α-induced ROS production (Kim et al., 2010). To examine the effects of Romo1 deletion constructs on Myc expression, two deletion constructs of Romo1, designated FLAG–Romo1-ΔC (deletion of the C-terminal 48–79 residues) and FLAG–Romo1-ΔN (deletion of the N-terminal 1–16 residues), were transfected into HeLa cells and Myc expression was assessed. As shown in Fig. 4D, both wild-type Romo1 and Romo1-ΔN decreased Myc expression, but Romo1-ΔC did not. From this result, we suggest that ROS derived from the C-terminal domain of Romo1 induced Myc degradation through the Mb II domain, Mb III domain and a ΔF construct (316–378 residues).

Romo1 is localized in the mitochondria and induces mitochondrial ROS production through complex III of the mitochondrial electron transport chain (Chung et al., 2008). To determine whether mitochondrial ROS production through complex III is required for downregulation of Myc, HeLa cells were cultured in the presence of an antioxidant (trolox), mitochondrial respiratory chain complex III inhibitors (myxothiazol and stigmatellin), complex I inhibitor (rotenone), complex II inhibitor (malonate) or complex IV inhibitor (sodium azide), and Myc downregulation was assessed by western blot analysis (Fig. 4E). Romo1-triggered downregulation of Myc was blocked by myxothiazol and stigmatellin and by trolox. By contrast, the other inhibitors failed to inhibit Romo1-mediated Myc downregulation. Next, we treated the cells with increasing amounts of H₂O₂, and Myc downregulation was analyzed by western blot analysis. As shown in Fig. 4F, treatment with a low concentration of H₂O₂ increased the amount of Myc protein. By contrast, treatment with higher concentrations of H₂O₂ decreased the amount of Myc protein. These findings indicate that Romo1-derived ROS have an important role in Myc regulation.

Myc degradation is regulated by Fbw7 and Skp2, and Myc degradation by Fbw7 is dependent on the phosphorylation of Thr58 and Ser62 in the MB1 domain (Welcker et al., 2004; Yada et al., 2004). To ascertain whether Myc degradation controlled by Romo1 is related to the phosphorylation of Thr58 and Ser62, wild-type (WT) Myc or Myc mutants (T58A, S62A or T58AS62A) were transfected into HeLa cells, and Myc expression was examined by western blot analysis. As shown in Fig. 5A, Myc expression was decreased in cells expressing Romo1, Skp2 or Fbw7. As expected, Fbw7 expression failed to degrade the Myc protein in cells transfected with Myc mutants (Fig. 5B). This result is consistent with a previous report (Welcker et al., 2004; Yada et al., 2004). However, Romo1 and Skp2 efficiently degraded the Myc protein in cells transfected with Myc mutants, demonstrating that Romo1-triggered Myc degradation is not mediated by Fbw7 (Fig. 5B,C). H₂O₂ treatment also efficiently triggered Myc degradation in cells transfected with Myc mutants (Fig. 5D). Next, we investigated whether Romo1 stimulated Myc degradation through Skp2. SKP2 siRNA was transfected into cells to knock down Skp2 (supplementary material Fig. S4) and Myc expression was examined. Interestingly, Skp2 knockdown suppressed Romo1-induced Myc degradation (Fig. 5E).

Recent reports have shown that the phosphorylation of Skp2 at Ser72 by Akt leads to cytoplasmic translocation of Skp2 (Gao et al., 2009; Lin et al., 2009). To investigate whether Romo1 expression regulates Skp2 cytoplasmic translocation, we observed HeLa cells by fluorescence microscopy after Romo1 transfection. As shown in Fig. 6A, Romo1 expression induced the cytoplasmic translocation of Skp2. Cytoplasmic Skp2 levels were quantified by fluorescence microscopy and analysis with MetaMorph software. This finding was also confirmed in HEK293 cells (supplementary material Fig. S5A). To determine whether H₂O₂ treatment also led to the cytoplasmic translocation of Skp2, the cells were treated with H₂O₂. H₂O₂ treatment promoted the cytoplasmic translocation of Skp2 (Fig. 6B), which was confirmed in cells exogenously transfected with Skp2 (Fig. 6C). We also confirmed the localization of Skp2 by cellular fractionation of HeLa cells. The cells were treated with H₂O₂ or transfected with FLAG–Romo1. Both Romo1 and H₂O₂ induced cytoplasmic translocation of Skp2 (Fig. 6D). These results demonstrate that ROS derived from Romo1 promote the cytoplasmic translocation of Skp2.

Next, we performed an immunofluorescence assay to detect the subcellular localization of Skp2 in response to serum stimulation in normal human fibroblasts, IMR-90 cells. The cells were serum-starved for 48 hours and then treated with serum for 15 hours. The cells were then harvested for immunofluorescence analysis. Skp2 was observed in the nucleus before serum stimulation. However, Skp2 was translocated into cytoplasm after serum addition for 15 hours (Fig. 6E). A similar finding was also observed in other normal human fibroblasts, WI-38 cells (supplementary material Fig. S5B).

We showed that downregulation of Myc by Romo1 is required for mitochondrial ROS production through complex III of the mitochondrial electron transport chain (Fig. 4E). Therefore, we explored whether cytoplasmic translocation of Skp2 by Romo1 was blocked by inhibitors of complex III of the mitochondrial electron transport chain. After HeLa cells were transfected with FLAG–Romo1, the cells were incubated with various inhibitors of the mitochondrial electron transport chain. Cytoplasmic translocation of Skp2 was detected in cells transfected with Romo1 (Fig. 6F). However, the Romo1-induced Skp2 cytoplasmic translocation was inhibited by myxothiazol and stigmatellin, but was not affected by other inhibitors. Trolox was used as a positive control. We also investigated whether Romo1 expression regulated cytoplasmic translocation of Myc in HeLa cells after Romo1 transfection. Romo1 expression enhanced the cytoplasmic translocation of Myc (Fig. 6G). We also confirmed the localization of Myc by cellular fractionation of HeLa cells transfected with FLAG–Romo1 (Fig. 6H).

To investigate whether Romo1 promotes the cytoplasmic translocation of Skp2 through the PI3K–Akt pathway, HeLa cells were transfected with FLAG–Romo1. As shown in Fig. 7A, Romo1 expression triggered Akt activation and Myc degradation. The Akt activation and Myc degradation were inhibited by the PI3K inhibitor LY294002. Immunofluorescence experiments revealed that Romo1 expression induced the cytoplasmic translocation of Myc and Skp2, and their translocations were suppressed by LY294002 (Fig. 7B). We further investigated whether the H₂O₂-induced cytoplasmic translocation of Skp2 was blocked by LY294002. Immunofluorescence analysis showed that cytoplasmic translocation of exogenous Skp2 by H₂O₂ treatment was inhibited by LY294002 and trolox (Fig. 7C). These results suggest that the cytoplasmic translocation of Skp2 and Myc induced by Romo1 is mediated by Akt.

To determine whether Romo1 expression enhances the interaction between Skp2 and Myc, FLAG–Romo1 was transfected into HeLa cells and a co-immunoprecipitation experiment was performed. As shown in Fig. 8A, Romo1 expression increased Skp2 binding to Myc and trolox treatment inhibited this interaction. To confirm this interaction, Skp2 was immunoprecipitated with its antibody and western blot analysis was performed with anti-Myc antibody (Fig. 8B). Next, we examined whether Romo1 enhanced Myc ubiquitylation. Myc was transfected into HeLa cells, together with FLAG–Romo1, FLAG–Fbw7α or FLAG–Skp2. To identify the extent of Myc ubiquitylation, ubiquitylated-Myc was immunoprecipitated with anti-HA antibody and subjected to western blot analysis using anti-Myc antibody. As shown in Fig. 8C,D, Romo1 significantly increased the amount of Myc-ubiquitin conjugates, and trolox treatment suppressed Myc ubiquitylation. Fbw7α and Skp2 were used as positive controls. Fig. 8E showed that Romo1-ΔN significantly increased the amount of Myc-ubiquitin conjugates, but Romo1-ΔC had no effect on Myc ubiquitylation. This finding was also examined in HEK 293 cells (supplementary material Fig. S6). We also explored whether H₂O₂ enhances ubiquitylation of Myc (Fig. 8F). These results indicate that ROS increase modulated by Romo1 expression induces an interaction between Skp2 and Myc and then enhances Myc ubiquitylation. To identify whether Romo1 regulates the stability of Myc protein, Romo1 was expressed by transient transfection in HeLa cells and the cells were treated with cycloheximide (CHX). The half-life of Myc was decreased in cells transfected with Romo1 (Fig. 8G, upper panel). Romo1 also reduced the stability of exogenously transfected Myc in the cells (Fig. 8G, lower panel). These results suggest that Romo1 negatively controls Myc stability and that it is an important post-translational regulator of Myc expression.

Myc is an unstable protein with a half-life of 20–30 minutes (Hann and Eisenman, 1984) and Myc degradation during G₁-S phase progression has been well identified. Serum stimulation enhances Ras activation in the PI3K–Akt pathway during early G1, resulting in GSK-3β inhibition (Gregory and Hann, 2000; Sears et al., 2000). During late G1 phase, the Ras activity decreases and GSK-3β phosphorylates Myc on Thr 58, resulting in its ubiquitylation and degradation (Sears et al., 2000). Phosphorylation of Myc on Thr 58 plays a key role in Myc degradation, and a mutation on Thr 58 contributes to tumorigenesis (Bhatia et al., 1993). However, Myc degradation by GSK-3β-mediated Myc phosphorylation on Thr 58 is not sufficient for Myc degradation during late G1 phase. Indeed, Myc mutated at Thr 58 was reported to have a half-life of 30–40 minutes, compared to the 20–30 minutes half-life of wild-type Myc (Bader et al., 1986; Sears et al., 1999; Salghetti et al., 1999). Moreover, Myc mutated at Thr 58 is still targeted for ubiquitylation and degradation (Hann, 2006). Therefore, an additional pathway should exist for Myc degradation. In the present study, we showed that Myc expression reached a peak at 3–6 hours and declined at 9 hours. However, Myc expression levels continuously increased until 24 hours when Romo1 expression was suppressed (Fig. 2D). We also showed that Romo1 expression promoted the ubiquitylation and degradation of Myc through cytoplasmic translocation of Skp2 and Myc (Figs 6 and 8). Therefore, we suggest that the Romo1/ROS/Skp2 pathway is another pathway for Myc turnover. The Romo1-mediated pathway appears to be one of the main pathways for Myc degradation, because the Myc level was significantly decreased when the cells were transfected with Romo1 to enhance Romo1 expression (Fig. 4).

Two main pathways of ubiquitin-mediated degradation of Myc exist for Myc turnover. One is mediated by Fbw7. The other pathway is mediated by Skp2. Myc ubiquitylation through Fbw7 has been well elucidated. A series of sequential phosphorylation events occur after mitogenic stimulation and are followed by Fbw7-mediated degradation of Myc (Sears et al., 1999; Yeh et al., 2004). However, Skp2-mediated Myc degradation is not well understood. In the present study, we demonstrate that Romo1 induces Myc degradation through a novel mechanism not previously reported. Romo1 induced by the enhanced Myc level increased the cellular ROS level to trigger the cytoplasmic translocation of Skp2 (Fig. 6). Skp2 was induced by mitogenic stimulation and reached a peak in S phase. Skp2 binds to two domains of Myc (Kim et al., 2003; von der Lehr et al., 2003). It is unlikely that Romo1 induced Skp2 expression because increased Romo1 expression did not increase the Skp2 level (Fig. 8A). Instead, Romo1 contributed to the cytoplasmic translocation of Skp2. Skp2 is reportedly located in the nucleus (Miura et al., 1999). However, a recent report showed that Skp2 translocates into the cytoplasm after Akt-mediated phosphorylation of Ser 72 (Gao et al., 2009; Lin et al., 2009). Romo1 also regulated the cytoplasmic translocation of Myc in the presence of the proteosomal inhibitor, MG-132 (Fig. 6G,H). In addition to enhancing the cytoplasmic translocation of Skp2 and Myc, Romo1 promoted the interaction between Skp2 and Myc, resulting in Myc ubiquitylation (Fig. 8). It seems that Romo1 does not directly interact with Skp2 or Myc in Myc degradation because H₂O₂ treatment increased the cytoplasmic translocation of Skp2 and Myc (Fig. 6B). Antioxidant treatment also suppressed the cytoplasmic translocation of Skp2 (Fig. 6F). Previously, we reported that ROS originate from complex III of the mitochondrial respiratory chain (Chung et al., 2008). Therefore, we examined whether cytoplasmic translocation of Skp2 was blocked by mitochondrial complex III inhibitors, and we showed that treatment of complex III inhibitors efficiently suppressed the cytoplasmic translocation of Skp2 (Fig. 6F). From these results, we suggest that ROS derived from Romo1 play an important role in Myc turnover. Although we showed that Myc degradation occurs via Romo1-mediated cytoplasmic translocation of Skp2, the exact mechanism by which Romo1 regulates the cytoplasmic translocation of Skp2 remains to be studied in the future.

Appropriate ROS levels maintained inside the cells play an important role in cell growth and survival, and the physiological range of H₂O₂ concentrations is 0.001 to 0.7 μM (Burdon and Rice-Evans, 1989). Although excessive ROS production reportedly contributes to many pathological disorders, including cancer, aging, and neurological diseases, ROS are required for redox signaling and their main source is NADPH oxidase (Finkel, 2003; Turrens, 2003). This enzyme responds to growth or cell survival signals to induce ROS production and is subsequently eliminated by antioxidant enzymes. Although the increase in ROS triggered by Romo1, which is induced by Myc, contributes to Myc degradation, appropriate levels of ROS are critical for Myc stabilization. The results presented in this study are consistent with a previous report implicating ROS produced by hematopoietic cytokines in G1 to S progression in Myc stabilization (Iiyama et al., 2006). This previous study also showed that NAC treatment reduced the stability of Myc protein, while H₂O₂ treatment of the cells enhanced its stability. H₂O₂ treatment induced ERK-dependent Myc phosphorylation on Ser 62 (Benassi et al., 2006). Recently, we reported that Romo1 is necessary for cell growth and that Romo1 knockdown induces cell cycle arrest at G1 through inhibition of ERK activation and induction of p27^Kip1 expression, demonstrating that ROS originating from the mitochondria play a key role as signaling mediators (Chung et al., 2009; Na et al., 2008). In the present study, we also showed that Romo1 knockdown inhibits ERK activation and Myc expression (Fig. 3A), resulting in cell cycle arrest at G1 (Fig. 3C). Therefore, we suggest that the basal level of ROS derived from the steady-state level of Romo1 is required for ERK activation and Myc stabilization. In contrast, excessive ROS produced from increased Romo1 expression induced by Myc trigger Myc degradation. This suggestion is supported by Fig. 3B and Fig. 4F. When the cells were treated with a low amount of H₂O₂, Myc expression was increased; however, treatment with additional H₂O₂ down-regulated Myc expression.

Myc is known to play an important role in cell proliferation and its level is controlled transcriptionally, post-transcriptionally or post-translationally. Abnormal regulation of Myc contributes to tumor formation. In the present study, we identified a novel pathway of negative feedback regulation for Myc during G1 phase. Upon mitogenic stimulation, Myc expression is increased for cell cycle progression. Myc induces Romo1 expression to enhance cellular ROS levels. The ROS promote the cytoplasmic translocation of Skp2 to cause Myc ubiquitylation, resulting in Myc degradation (Fig. 9). Myc has also been reported to stimulate ROS generation, which induces DNA damage (Vafa et al., 2002). Enhanced ROS levels were also observed in Myc transgenic animals. We showed that Myc up-regulation, which was induced by serum stimulation, increased the ROS level via Romo1 expression. Our results presented in this study are the first to elucidate the mechanism of Myc-induced ROS production. An ROS imbalance can cause cellular DNA damage and genomic instability, which can contribute to the initiation, promotion and malignancy of tumors (Finkel and Holbrook, 2000). Therefore, the results presented in this study provide important information regarding the mechanism of Myc-stimulated oncogenesis associated with ROS.

The human lung fibroblast IMR-90 and WI-38 cells were obtained from the American Type Culture Collection (ATCC, Manassas, VA) and cells ranging from 29 to 34 in population doubling level (PDL) were used. WI-38 VA-13 cells were cultured in Eagle's minimal essential media (EMEM, Gibco-Invitrogen, Grand Island, NY). Human embryonic kidney (HEK) 293 cells, HeLa cervix carcinoma cells and Huh-7 human hepatocarcinoma cells were cultured in Dulbecco's modified Eagle's media (DMEM, Gibco-Invitrogen). Human non-small cell lung cancer (NSCLC) cell lines A549 and H1299 were cultured in Ham's F12 and RPMI 1640 medium (Gibco-Invitrogen), respectively. All media contained 10% heat-inactivated FBS (Gibco-Invitrogen), sodium bicarbonate (2 mg/ml; Sigma-Aldrich, St Louis, MO), penicillin (100 units/ml), and streptomycin (100 μg/ml; Gibco-Invitrogen). PD98059, U0126, LY294002, Wortmannin, SP600215 and SB203580 were purchased from StressGen (Victoria, BC, Canada). 3-[[6-(3-amino-phenyl)-1H-pyrrolo[2,3-d] pyrimidin-4-yl]oxy]-phenol (TWS119) was obtained from Calbiochem (La Jolla, CA). N-acetyl-cysteine (NAC), hydrogen peroxide (H₂O₂), nocodazol, 6-hydroxy-2,5,7,8-tetramethylchroman-2-carboxylic acid (Trolox), stigmatellin, myxothiazol, malonate, rotenone, sodium azide and N-carbobenzoxy-l-leucinyl-l-leucinyl-l-norleucinal (MG132) were purchased from Sigma-Aldrich. 2′,7′-dichlorofluorescein diacetate (DCF-DA) and MitoSOX were obtained from Molecular Probes (Eugene, OR).

cDNAs encoding FLAG–Romo1 Wild-type (Wt) and deletion mutants, ΔN and ΔC, were prepared in our laboratory and have been validated previously (Kim et al., 2010). Complementary DNA encoding Myc (human) was cloned into pcDNA3 (Invitrogen). pCGN-HA–Myc (Wt) and deletion mutants (ΔA, ΔB, ΔC, ΔD, ΔE, ΔF and ΔG) were described previously (Herbst et al., 2004; Tworkowski et al., 2002). pCl-FLAG–Myc (Wt) and substitution mutants (T58A, S62A and T58AS62A) and the pCGN-HA–Fbw7 construct were kindly provided by Keiichi I. Nakayama and Masaki Matsumoto (Department of Molecular and Cellular Biology, Kyushu University, Japan) and have been described earlier (Yada et al., 2004). The p3-FLAG–Myc-Fbw7α was kindly provided by Professor Bruce E. Clurman (Fred Hutchinson Cancer Research Center, University of Washington School of Medicine, Seattle) and was described previously (Welcker et al., 2004). CMV-FLAG–Skp2 was kindly provided by Professor Tae Jun Park (Department of Biochemistry and Molecular Biology, Ajou University, Republic of Korea) and was described previously (Park et al., 2009).

The sequences of Romo1 siRNA were unique to their intended targets, based on BLAST searches. The Romo1 siRNA sequence was 5′-GGGCUUCGUGAUGGGUUG-3′ (sense strand). The other siRNA against Romo1 was described previously (Hwang et al., 2007). The Myc siRNA sequences (Grandori et al., 2005), Skp2 siRNA sequences (Carrano et al., 1999; Nishitani et al., 2006; Zhang et al., 2004), and control siRNA sequence (Chung et al., 2009) were described previously. siRNAs were purchased from Bioneer (Taejon, Republic of Korea).

Antibodies were: anti-Myc mouse monoclonal (Santa Cruz Biotechnology, Santa Cruz, CA) and rabbit polyclonal (Santa Cruz Biotechnology), anti-p27^kip1 mouse monoclonal (BD Pharmingen, San Diego, CA) and rabbit polyclonal (Zymed Laboratories, San Francisco, CA), anti-phospho-ERK rabbit polyclonal (Cell Signaling Technology, Beverly, MA), anti-ERK rabbit polyclonal (Cell Signaling Technology), anti-phospho-Akt rabbit polyclonal (Cell Signaling Technology), anti-Akt rabbit polyclonal (Cell Signaling Technology), anti-Skp2 rabbit polyclonal (Santa Cruz Biotechnology) and anti-Lamin B1 rabbit polyclonal (Santa Cruz Biotechnology), β-actin mouse monoclonal (Sigma-Aldrich), anti-FLAG (M2) (Sigma-Aldrich) and anti-HA (Sigma-Aldrich). Mouse monoclonal antibody (mAb) against Romo1 was described previously (Kim et al., 2010).

For serum stimulation experiments, human lung primary fibroblast IMR-90 and WI-38 cells and human embryo kidney (HEK) 293 cells were washed twice with serum-free media and further incubated in EMEM with 0.05% FBS for 48 h (Lee et al., 2010). EMEM containing 30% FBS was then added and cells were collected at the indicated time points.

Semi-quantitative RT-PCR analysis was performed as described previously (Chung et al., 2008). SYBR Green PCR amplifications were performed using an iCycler iQ Real-Time Detection System (Bio-Rad Laboratories, USA) associated with the iCycler Optical System Interface software (version 2.3; Bio-Rad). All PCR experiments were carried out in triplicate with a reaction volume of 25 μl, using iCycler IQ 96-well optical grade PCR plates (Bio-Rad) covered with iCycler optical-quality sealing film (Bio-Rad). Data analyses (calculations), including determining the relative amounts of each target mRNA, were performed with the iCycler IQ real-time detection system (Bio-Rad).

Cells were transfected with the indicated constructs or siRNA using Lipofectamine™ (Gibco-Invitrogen). The immunoprecipitation and western blot analysis were described previously (Kim et al., 2010).

Intracellular ROS production was measured using a fluorescence microscope (Olympus LX71 microscope), and the images were analyzed using MetaMorph software (Universal Imaging, Westchester, PA) for quantification purposes as described earlier (Kim et al., 2010; Lee et al., 2010). For immunofluorescence assays, cells were fixed in 3.7% paraformaldehyde (Sigma-Aldrich) for 10 minutes at room temperature and stained using standard protocols. For quantification of protein translocation, 100–200 cells were monitored in each experiment by fluorescence microscopy and were validated as described previously (Gao et al., 2009; Lin et al., 2009).

For analysis of cell cycle profile by FACS, cells were harvested in a time-dependent manner after induction, fixed with ethanol, stained with propidium iodide (PI, 50 μg/ml, Sigma-Aldrich) containing RNase A (100 mg/ml, Sigma-Aldrich) for 30 minutes at room temperature. The DNA content was analyzed using a FACScan flow cytometer (Becton Dickinson, San Jose, CA).

The Nuclear extract kit (California, USA) was used to perform cellular fractionation in accordance with the manufacturer's instructions. The purity of the extract was confirmed by western blot analysis against anti-cytosol-specific-β-actin (Sigma-Aldrich) or anti-nuclear-specific-lamin B1.

For protein stabilization analysis, HeLa cells were transfected with the indicated constructs. After transfection for 48 h, cells were treated with cycloheximide (CHX, 20 μg/ml). The cell lysates were prepared and analyzed by western blot analysis. After CHX treatment, endogenous or exogenous Myc levels were quantified by densitometric scanning in the image J program. For Myc ubiquitylation, cells were transfected with Ubiquitin (Ub)-HA plasmid together with various constructs for 2 days and then treated with MG132 (10 μM) for 6 h. The immunoprecipitates were subjected to western blot analysis as described previously (Kim et al., 2010).

Each assay was performed in triplicate and independently repeated at least three times. Statistical significance was defined as P<0.05. Means, SEs and Ps were calculated using GraphPad PRISM version 4.02 for Windows (GraphPad Software, San Diego, CA).

This research was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education, Science and Technology (2010-0021371), by a grant from the National R&D Program for Cancer Control, Ministry for Health, Welfare and Family Affairs, Republic of Korea (1020180), by National Nuclear R&D Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education, Science and Technology (20100018574) and by a grant of the Korea Healthcare Technology R&D Project, Ministry for Health, Welfare & Family Affairs, Republic of Korea (A084537-0902-0000100).