The CDK inhibitor p21waf1/cip1 is degraded by a ubiquitin‐independent proteolytic pathway. Here, we show that MDM2 mediates this degradation process. Overexpression of wild‐type or ring finger‐deleted, but not nuclear localization signal (NLS)‐deleted, MDM2 decreased p21waf1/cip1 levels without ubiquitylating this protein and affecting its mRNA level in p53−/− cells. This decrease was reversed by the proteasome inhibitors MG132 and lactacystin, by p19arf, and by small interfering RNA (siRNA) against MDM2. p21waf1/cip1 bound to MDM2 in vitro and in cells. The p21waf1/cip1‐binding‐defective mutant of MDM2 was unable to degrade p21waf1/cip1. MDM2 shortened the half‐life of both exogenous and endogenous p21waf1/cip1 by 50% and led to the degradation of its lysine‐free mutant. Consequently, MDM2 suppressed p21waf1/cip1‐induced cell growth arrest of human p53−/− and p53−/−/Rb−/− cells. These results demonstrate that MDM2 directly inhibits p21waf1/cip1 function by reducing p21waf1/cip1 stability in a ubiquitin‐independent fashion.
Proteolysis plays a crucial role in regulating the pathway of the tumor suppressor p53. At least three proteins in this pathway are regulated through proteasome‐mediated mechanisms. One of them is p53 itself. The nuclear p53 transcriptional activator is degraded through a ubiquitin‐dependent proteasomal system mediated by an E3 ubiquitin ligase called MDM2 (Haupt et al., 1997; Honda et al., 1997; Kubbutat et al., 1997; Fuchs et al., 1998). Thus the level of p53 in a normal cell is extremely low. In response to various cellular stresses, p53 is chemically modified through processes such as phosphorylation or acetylation at its N‐ and/or C‐terminal domains (Appella and Anderson, 2001; and references therein). Some of these modifications shield p53 from attack by MDM2 and also enhance its activity (Gu and Roeder, 1997; Shieh et al., 1997; Canman et al., 1998; Giaccia and Kastan, 1998; Keller et al., 2001). p53 becomes stabilized and activated to induce the transcription of many target genes whose protein products either initiate apoptosis or cause cell growth arrest (Crook et al., 1994; Vogelstein et al., 2000). One of these p53 targets is MDM2 that in turn serves as a negative feedback regulator (Barak et al., 1993; Wu et al., 1993) and finely tunes p53 activity (Momand et al., 1992). Besides ubiquitylating p53 and mediating its degradation, MDM2 itself is also regulated by the ubiquitin‐dependent proteasome (Michael and Oren, 2003).
The third protein in the p53 pathway, which is also regulated through a proteolytic mechanism (Blagosklonny et al., 1996; Maki and Howley, 1997; Fukuchi et al., 1999; Nakanishi et al. 2000; Sheaff et al., 2000), is the cyclin‐dependent kinase (CDK) inhibitor p21waf1/cip1 (el‐Deiry et al., 1993; Harper et al., 1993; Bunz et al., 1998). p21waf1/cip1 is one of the prime transcriptional targets of p53 and mediates p53‐dependent cell growth arrest and senescence. Unlike p53 and MDM2, which are degraded in either the nucleus or cytoplasm (Yu et al., 2000), p21waf1/cip1 appears to be degraded solely in the nucleus (Sheaff et al., 2000). Also different from p53 and MDM2, the proteasomal turnover of p21waf1/cip1 is independent of its ubiquitylation (Sheaff et al., 2000), although this protein can be stabilized by proteasome inhibitors and is ubiquitylated in cells (Blagosklonny et al., 1996; Maki and Howley, 1997; Fukuchi et al., 1999; Nakanishi et al., 2000; Sheaff et al., 2000). These studies suggest that ubiquitylation may not necessarily be a prerequisite for proteolysis of some protein targets by the proteasome pathway. Although it was shown recently that the 20S proteasome binds to p21waf1/cip1 and leads to its degradation in vitro (Touitou et al., 2001), it is still unknown what may regulate p21waf1/cip1 degradation in cells. This study as described here reveals our new finding that MDM2 mediates the proteasomal turnover of p21waf1/cip1 without ubiquitylating this protein in cells.
The level of p21waf1/cip1 is inversely proportional to that of MDM2 independent of p53
In our attempt to identify the regulator of p21waf1/cip1 stability, we tested whether MDM2 is involved in modulating the p21waf1/cip1 level in cells, because both MDM2 and p21waf1/cip1 are always induced by p53 whenever p53 is activated (Vogelstein et al., 2000). Thus, MDM2 would have the chance to target p21waf1/cip1 for degradation. Also, when examining the level of p21waf1/cip1 in MDM2‐deficient or proficient cell lines, we found that its level is inversely proportional to that of MDM2 (Figure 1). p21waf1/cip1 protein was low in p53−/− mouse embryonic fibroblast (MEFs), but was markedly elevated in p53−/−/mdm2−/− MEFs (Figure 1A). This alteration was not due to the elevation of its mRNA level (Figure 1A). Interestingly, MG132 (Figure 1A) or lactacystin (Figure 1B) treatment only restored the level of p21waf1/cip1 in p53−/− MEFs, but not that in p53−/−/mdm2−/− MEFs, suggesting that the protein level of p21waf1/cip1 is regulated by MDM2. Thus, blocking the proteasome pathway may not be necessary for stabilizing p21waf1/cip1 in the absence of MDM2. Also, the increase of p21waf1/cip1 in these p53−/−/mdm2−/− cells was independent of p53. Consistently, ectopic expression of MDM2 in p53−/−/mdm2−/− MEFs markedly decreased endogenous p21waf1/cip1 (Figure 1C). These results suggest that MDM2 may be involved in regulating the proteasomal turnover of p21waf1/cip1 in cells.
MDM2 does not require its ring finger to mediate p21waf1/cip1 degradation in the nucleus
To test this possibility, we first checked whether overexpression of MDM2 leads to a decrease of p21waf1/cip1 level by comparing wild‐type MDM2 with its deletion mutants. Human p53‐null lung adenocarcinoma H1299 cells were transfected with plasmids encoding p21waf1/cip1 alone or together with wild‐type and deletion mutant MDM2s. At 48 h after transfection, cells were harvested for western blot (WB) and northern blot (NB) analyses. In line with the results in Figure 1, overexpression of MDM2 reduced the protein, but not the mRNA, level of p21waf1/cip1 (Figure 2B). Surprisingly, this reduction did not appear to absolutely require the C‐terminal ring finger domain of MDM2, as the 1–441 MDM2 mutant that was unable to mediate p53 degradation (Figure 2D) also reduced p21waf1/cip1 (Figure 2B). However, the nuclear localization signal (NLS) domain of MDM2 seems to be critical as p21waf1/cip1 levels did not change in the presence of the Δ‐NLS (Δ150–230) MDM2 mutant (Figure 2B). In contrast, this mutant was still able to lead to p53 degradation (Figure 2D). Remarkably, MG132 treatment reversed the reduction of p21waf1/cip1 by both the wild‐type and 1–441 mutant MDM2s (Figure 2C). The N‐terminal p53‐binding domain of MDM2 was also dispensable for decreasing p21waf1/cip1 (data not shown). Furthermore, wild‐type and 1–441 MDM2s localized in the nucleus, whereas the Δ150–230 MDM2 stayed in the cytoplasm (Figure 2E). Additionally, MDM2 reduced the level of p21waf1/cip1 in the presence of the nuclear export inhibitor leptomycin B (LMB) in H1299 cells (Figure 2F). LMB blocks the nuclear export of MDM2 (Freeman and Levine, 1998; and our unpublished data). These results suggest that MDM2 may mediate the degradation of both exogenous and endogenous p21waf1/cip1 independently of its E3 ubiquitin ligase domain in the nucleus.
MDM2 does not mediate p21waf1/cip1 ubiquitylation in cells
The observation that the ring finger domain of MDM2 was dispensable for the MDM2‐mediated p21waf1/cip1 decrease (Figure 2) suggests that MDM2 might not ubiquitylate p21waf1/cip1. To test this possibility, p53‐null H1299 cells and p53−/−/mdm2−/− MEFs were transfected with vectors encoding p21waf1/cip1, His6‐ubiquitin and hemagglutinin (HA)‐MDM2 in different combinations. Ubiquitylated MDM2 or p21waf1/cip1 molecules were pulled‐down by Ni‐NTA–agarose and probed with anti‐MDM2 or anti‐p21waf1/cip1 antibodies. Consistent with the results in Figure 2, MDM2 did not ubiquitylate p21waf1/cip1, though p21waf1/cip1 was ubiquitylated in these cells (Figure 3A and B) as expected (Maki and Howley, 1997; Fukuchi et al., 1999; Sheaff et al., 2000). p21waf1/cip1 ubiquitylation must be catalyzed by an unknown E3 ubiquitin ligase, as p21waf1/cip1 ubiquitylation was detected in MEFs that do not have the mdm2 gene (Figure 3A). In contrast, exogenous MDM2 was ubiquitylated efficiently in both cell lines (Figure 3A and B). The reduction of p21waf1/cip1 or MDM2 ubiquitylation when the two proteins were co‐expressed might be due to the competition of these proteins for the limited pool of His6‐tagged ubiquitins (Figure 3A and B). Alternatively, p21waf1/cip1 may also inversely regulate the MDM2 level, though this idea needs to be clarified. Nevertheless, these results clearly show that MDM2 does not ubiquitylate p21waf1/cip1 in cells. Consistent with this notion, overexpression of MDM2 as well as its ring finger‐truncated mutant also led to degradation of a p21waf1/cip1 mutant with six lysine–arginine (6KR) substitutions (Figures 3C and 5A). The p21waf1/cip1 6KR mutant was shown to be degraded as efficiently as wild‐type p21waf1/cip1 in cells, though the mutant was not ubiquitylated in cells (Figure 3B; Sheaff et al., 2000). These results demonstrate that MDM2 does not ubiquitylate p21waf1/cip1 in cells.
MDM2 reduces the half‐life of both exogenous and endogenous p21waf1/cip1
Next we wanted to determine whether MDM2 truly affects the turnover of p21waf1/cip1 by performing pulse–chase experiments using [35S]methionine. As shown in Figure 4A, MDM2 markedly reduced the half‐life of p21waf1/cip1 from ∼2 h to ∼45 min in H1299 cells (Figure 4A). Similarly, 1–441 MDM2 also reduced the half‐life of p21waf1/cip1, though to a lesser extent (Figure 4A). Again, the total level of p21waf1/cip1 drastically decreased in the presence of wild‐type or 1–441 mutant MDM2 (lower panels of Figure 4A). In agreement with these results, the half‐life of endogenous p21waf1/cip1 was also shortened to ∼30 min from ∼100 min when MDM2 was expressed in p53−/−/mdm2−/− MEFs (Figure 4B). These results demonstrate that MDM2 accelerates the turnover of p21waf1/cip1 in cells.
MDM2 binds to the C‐terminal half of p21waf1/cip1 in vitro
Then, we determined whether MDM2 interacts with p21waf1/cip1 by performing a set of GST–protein interaction assays. Purified GST–p21 (Figure 5A) or GST–MDM2 (Figure 5C) were incubated with either His6‐MDM2 or His6‐p21 purified from bacteria. As shown in Figure 5, MDM2 physically bound to p21waf1/cip1 as well as its C‐terminal domain half (Figure 5B). Interestingly, p21waf1/cip1 appeared to bind preferentially to the MDM2 central region (amino acids 222–425) that contains a zinc finger domain (amino acids 290–390), because neither the N‐terminal domain (amino acids 1–150) nor the C‐terminal domain (amino acids 425–491) was able to bind to the p21waf1/cip1 protein as well as did wild‐type MDM2 or the 285–384 region (Figure 5C). The central p21waf1/cip1‐binding domain (amino acids 222–437) of MDM2 is crucial for MDM2‐mediated p21waf1/cip1 degradation, as deletion of this domain rendered the MDM2 mutant unable to degrade p21waf1/cip1 in cells (Figure 5D), even though both this mutant and p21waf1/cip1 were in the nucleus (data not shown). These results demonstrate that MDM2 physically associates with p21waf1/cip1 in vitro and this association is crucial for MDM2 to degrade p21waf1/cip1 in cells.
MDM2 binds to p21waf1/cip1 in cells
In order to determine whether endogenous MDM2 and p21waf1/cip1 proteins indeed associate with each other in cells, we performed two experiments. First, we tested whether this association is detectable upon DNA damage, which induces p53 and thus MDM2 and p21waf1/cip1. To do so, human astrocytoma SJSA cells that contain wild‐type p53 were irradiated with 7 Gy of γ‐irradiation and harvested at different time points for WB and immunoprecipitation (IP)–WB analyses. As shown in Figure 6A, both MDM2 and p21waf1/cip1 were induced by p53 upon irradiation (left panels). Interestingly, the endogenous MDM2 protein was co‐immunoprecipitated with anti‐p21waf1/cip1 (lane 3) or anti‐MDM2 (lane 8) antibodies 4 h post‐irradiation (lanes 1–3). This result was reproducible and not non‐specific, as MDM2 was not detected in the reaction with an anti‐actin antibody (lanes 4–6). Secondly, we tested whether MDM2 and p21waf1/cip1 associate in human cervical carcinoma HeLa or SJSA cells after treatment with MG132. Indeed, this was true for both of the cell lines. As shown in Figure 6B for HeLa cells, MDM2 and p21waf1/cip1 were co‐immunoprecipitated with anti‐p21waf1/cip1 (lane 4) or anti‐MDM2 (lane 8) antibodies but not with anti‐actin antibodies (lane 6), only when the p21waf1/cip1 level was induced by MG132. Hence, these results using two different human cell lines demonstrate that MDM2 binds to p21waf1/cip1 when both of the proteins are induced.
Suppression of MDM2 induces p21waf1/cip1 in p53‐null cells
Next, we wanted to determine whether suppressing MDM2 would induce p21waf1/cip1 in a p53‐independent fashion by conducting two different experiments. First, because mouse p19arf inhibits MDM2‐mediated p53 degradation (Honda and Yasuda, 1999; Xirodimas et al., 2001), we tested whether this protein also affects MDM2‐mediated p21waf1/cip1 degradation. H1299 cells were transfected with the p21waf1/cip1 plasmid, and/or the MDM2 plasmid or together with the p19arf plasmid. Protein levels were detected by WB analysis 48 h after transfection. As expected, MDM2 again led to a reduction of the p21waf1/cip1 level, but p19arf reversed this reduction (Figure 7A). It has been suggested that p19arf inhibits p53 degradation by MDM2 either by moving MDM2 to the nucleolus and thus sequestering it away from p53 (Zhang et al, 1998; Tao and Levine, 1999; Weber et al., 1999) or through direct interaction with MDM2 (Kamijo et al., 1998; Llanos et al., 2001; Clark et al., 2002). Thus, p19arf may do the same to reverse p21waf1/cip1 degradation by MDM2. Indeed, in the presence of p19arf, both p21waf1/cip1 and MDM2 were detected in the nucleolus (the bottom row of Figure 7B). MDM2 was moved to the nucleolus by p19arf (second row from top), because MDM2 and/or p21waf1/cip1 were scattered in the nucleus in the absence of p19arf (the second bottom row of Figure 7B, and Figure 2E). Consistently, these three proteins formed a ternary complex in cells (Figure 7C). This complex formation requires MDM2, as p21waf1/cip1 did not bind p19arf in the absence of MDM2 (Figure 7D). Furthermore, being in the nucleolus appears to be critical for p19arf inhibition of MDM2‐mediated p21waf1/cip1 degradation, as p19arf, which was unable to bring the 1–441 MDM2 into the nucleolus (Figure 7F; Weber et al., 2000), failed to protect p21waf1/cip1 from degradation by this mutant MDM2 (Figure 7E). Finally, two oncoproteins, c‐Myc and E1A, which have been shown to induce p19arf (de Stanchina et al., 1998; Zindy et al., 1998), also induced p21waf1/cip1 in p53−/− MEFs (Figure 7G). These results confirm that the degradation of p21waf1/cip1 is mediated by MDM2, which is blocked by p19arf through a subcellular separation mechanism.
Secondly, we tested whether direct suppression of MDM2 by small interfering RNA (siRNA) could elevate p21waf1/cip1 levels. H1299 cells were transfected with two different concentrations of the anti‐MDM2 siRNA oligomers or scrambled RNA oligomers as a control using Oligofectamine (Invitrogen). Cells were harvested for WB analysis with antibodies against MDM2 or p21waf1/cip1. As shown in Figure 8A, the MDM2 level was markedly reduced by the anti‐MDM2 siRNA in a dose‐dependent manner, but not by the scrambled RNA. Consistent with the results in Figures 1,2,3,4, the p21waf1/cip1 level was inversely elevated when the MDM2 level decreased (Figure 8A). As clearly depicted in the graphic presentation, 300 nM of anti‐MDM2 siRNA oligomers resulted in a >2‐fold induction of p21waf1/cip1 while reducing the level of MDM2 by >2‐fold. We can rule out the possibility that p21waf1/cip1 was induced through activation of endogenous p73 by anti‐MDM2 siRNA, as there was no detectable p73 in H1299 cells (Zeng et al., 1999). Also, p73 was not induced non‐specifically by the stress of siRNA transfection (data not shown). Furthermore, the p21waf1/cip1 induction was not due to activation of p63 either, as MDM2 was shown previously to enhance p63 activity (Calabro et al., 2002; and data not shown). Thus, suppressing MDM2 would reduce p63 activity. Finally, because this cell line lacks p53, these results demonstrate that suppressing MDM2 function or reducing its expression can rescue p21waf1/cip1 from degradation independently of p53.
Degradation of p21waf1/cip1 mediated by MDM2 is independent of MDMx
A recent discovery that deletion of the p53 gene can rescue the lethal phenotype of MDMx‐deficient mice (Parant et al., 2001; Migliorini et al., 2002;) suggests that MDMx may be in the same pathway as MDM2. To test whether the p21waf1/cip1 degradation mediated by MDM2 is dependent upon MDMx, we introduced p21waf1/cip1 alone or together with MDM2 into p53−/−/mdmx−/− MEFs. As shown in Figure 8B and C, ectopic expression of MDM2 reduced the level of either exogenous or endogenous p21waf1/cip1 in the absence of MDMx. These results suggest that MDM2 can mediate p21waf1/cip1 degradation independently of MDMx.
MDM2 functionally reverses the cell growth suppression induced by p21waf1/cip1
To determine the functional consequence of the degradation of p21waf1/cip1 by MDM2, fluorescence‐activated cell sorting (FACS) analysis was carried out after transfection of H1299 cells with either the p21waf1/cip1 plasmid alone or together with the MDM2 plasmid. At 48 h after transfection and 16 h after nocodazole treatment, cells were harvested for cell sorting and cytometry. As expected, ectopic expression of p21waf1/cip1 in these cells induced cell growth arrest at G1 phase (∼25%) in comparison with the vector control (∼7%) (Figure 9A). Overexpression of MDM2 conversely brought G1 cells down to ∼7 or ∼5% even in the presence of p21waf1/cip1, indicating that MDM2 directly inhibits p21waf1/cip1 function. Consistent with this result, MDM2 also reversed p21waf1/cip1‐induced suppression of cell growth in a colony formation assay using p53‐ and Rb‐null human osteosarcoma Saos‐2 cells (Figure 9B). Saos‐2 cells were used to exclude the possibility that MDM2 might inhibit p21waf1/cip1 function by inactivating p53 and Rb (Momand et al., 1992; Xiao et al., 1995). Taken together, these results demonstrate that MDM2 negatively affects p21waf1/cip1‐mediated cell growth arrest in a p53‐ or Rb‐independent fashion.
Our study identifies MDM2 as a cellular p21waf1/cip1 regulator that mediates its proteasomal turnover. The significance of this study is 2‐fold. First, this finding adds another level of regulation to the p53 pathway. Because p21waf1/cip1 mediates p53‐induced G1 arrest in response to DNA damage (el‐Deiry et al., 1993; Harper et al., 1993), MDM2 would inhibit G1 arrest more effectively by degrading both p53 and p21waf1/cip1 simultaneously. Several lines of evidence support this hypothesis. First, ectopic expression of MDM2 in p53‐null cells inhibited p21waf1/cip1‐mediated growth arrest (Figure 9). Also the p21waf1/cip1 protein level, but not the mRNA level, was elevated in p53−/−/mdm2−/− MEFs in comparison with that in p53−/− MEFs (Figure 1). Consistently, the level of p21 was inversely proportional to that of MDM2 when the MDM2 expression was suppressed by siRNA targeting in p53‐null cells (Figure 8). Furthermore, ectopic expression of MDM2 resulted in the degradation of both exogenous and endogenous p21waf1/cip1 (Figures 1,2,3) as well as shortening the half‐life of both exogenous and endogenous p21waf1/cip1 (Figure 4). Moreover, this degradation was rescued by overexpression of p19arf (Figure 7). Also, overexpression of c‐Myc or E1A oncoproteins, which induce p19arf (de Stanchina et al., 1998; Zindy et al., 1998) induced p21waf1/cip1 in p53‐null MEFs (Figure 7G). Finally, MDM2 specifically mediates p21waf1/cip1 degradation, as MDM2 did not affect the level of another CDK inhibitor p27 (data not shown). These results strongly demonstrate that MDM2 regulates p21waf1/cip1 proteasomal turnover in cells. Hence, targeting two components of the p53 pathway by MDM2 is biologically significant.
To our surprise, the degradation of p21waf1/cip1 by MDM2, unlike that of p53 (Haupt et al., 1997; Honda et al., 1997; Kubbutat et al., 1997), does not need its intrinsic E3 ubiquitin ligase activity (Figures 2,3,4). Thus, this finding suggests a novel ring finger‐independent mechanism for MDM2‐mediated proteolysis of some target proteins such as p21waf1/cip1. Although how MDM2 exactly leads to p21waf1/cip1 degradation still remains to be investigated, it is plausible that MDM2 may recruit the 26S proteasomal machinery to degrade p21waf1/cip1, as MDM2 directly interacted with this protein (Figure 5), and the proteasome inhibitor MG132 was able to rescue this degradation, subsequently increasing the level of p21waf1/cip1 (Figure 2). In line with this idea, some yeast E3 ligases have been shown to interact directly with the 19S complex of the 26S proteasome (Xie and Varshavsky, 2000), and mammalian p21waf1/cip1 could bind to the 20S complex directly (Touitou et al., 2001). Thus, MDM2 might bind to the 19S complex and p21waf1/cip1, recruiting p21waf1/cip1 to the 20S complex. Alternatively, MDMx may serve as a partner of MDM2 in regulating p21waf1/cip1 stability, though MDMx is not essential for MDM2‐mediated p21waf1/cip1 degradation (Figure 8B). It is also possible that other unknown proteins might bind to the MDM2–p21waf1/cip1 complex and thus serve as an adaptor to bring the 26S proteasome complex to p21waf1/cip1. Though all of these hypotheses remain to be studied, it is conceivable that this event exclusively occurs in the nucleus, because (i) p21waf1/cip1 co‐localized with MDM2 in the nucleus, and the mutant MDM2 expressed in cytoplasm, though able to degrade p53, was unable to degrade p21waf1/cip1 (Figure 2); and (ii) MDM2 was able to degrade p21waf1/cip1 even in the presence of the nuclear export inhibitor LMB (Figure 2F). This notion is in accordance with the earlier report showing that p21waf1/cip1 was degraded in the nucleus (Sheaff et al., 2000). In addition, the MDM2–p21waf1/cip1 interaction was detected in p53‐proficient cells after irradiation or treatment with the proteasomal inhibitor MG132 (Figure 6). Therefore, it is certain that MDM2 binds to and degrades p21waf1/cip1 in the nucleus. Our finding does not exclude the present concept that MDM2 degrades p53 through a ubiquitin‐mediated proteasome system (Honda et al., 1997; Fang et al., 2000). Ubiquitin‐independent proteolysis has begun to emerge as a common pathway, parallel to the ubiquitin‐dependent one, for regulating the turnover of some cellular proteins (Sheaff et al., 2000; Verma and Deshaies, 2000; David, 2002; Hoyt et al. 2003; Orlowski and Wilk, 2003). Therefore, elucidating the detailed mechanism underlying MDM2‐mediated p21waf1/cip1 degradation would be instrumental for better understanding of this pathway.
Identification of MDM2 as a regulator of p21waf1/cip1 turnover may provide insight into understanding how MDM2 regulates p53‐independent pathways. It has been shown that transforming growth factor‐β (TGF‐β) induces the expression of p21waf1/cip1 through a p53‐independent mechanism (Datto et al. 1995) and that MDM2 can rescue TGF‐β‐induced growth arrest in a p53‐indedepent manner (Sun et al. 1998). It is possible that MDM2 may inhibit TGF‐β‐dependent growth arrest by mediating the proteasomal degradation of p21waf1/cip1. Thus further investigating the likely involvement of the MDM2‐mediated p21waf1/cip1 turnover in the TGF‐β signaling pathway would be an interesting topic for future research.
Materials and methods
Human lung adenocarcinoma H1299 cells, human embryonic kidney epithelial 293 cells, mouse embryonic testicular carcinoma F9 cells and mouse p53−/−, p53−/−/mdm2−/− or p53−/−/mdmx−/− MEFs were cultured as previously described (Zeng et al., 2000). The 293‐HA‐MDM2 stable cell line was established by introducing pCDNA3‐HA‐MDM2 into 293 cells and selecting with 0.5 mg/ml G418 for 3–4 weeks. Cell clones that expressed HA‐MDM2 were confirmed by immunofluorescent staining and WB analyses with anti‐HA antibodies.
Lysis buffer consisted of 50 mM Tris–HCl pH 8.0, 0.5% NP‐40, 1 mM EDTA, 150 mM NaCl and 1 mM phenylmethylsulfonyl fluoride (PMSF). Buffer C 100 (BC100) included 20 mM Tris–HCl pH 7.9, 0.1 mM EDTA, 10% glycerol, 100 mM KCl, 4 mM MgCl2, 0.2 mM PMSF, 1 mM dithiothreitol (DTT) and 0.25 mg/ml of pepstatin A. The 1× SSC consisted of 0.15 M NaCl, 15 mM sodium citrate pH 7.0.
Antibodies and plasmids
Monoclonal anti‐Flag and anti‐α‐tubulin antibodies were purchased from Sigma. The polyclonal anti‐p53 antibody was purchased from Santa Cruz Biotech. Monoclonal anti‐MDM2 antibodies 4B11 and 2A10 were described previously (Zeng et al., 1999). Monoclonal anti‐p21waf1/cip1 antibodies were purchased from Neomarker Biotech and Oncogene Science, respectively. We also obtained some anti‐p21waf1/cip1 antibodies from Wafik S.el‐Deiry (University of Pennsylvania). pCDNA3‐HA‐MDM2 and pCMV‐p53 were previously described (Zeng et al., 1999). pCEP4‐p21waf1/cip1 was described (el‐Deiry et al., 1993). The MDM2 mutant plasmids were described previously (Chen et al., 1993). pCDNA3‐p19arf was a gift from Yue Xiong (University of North Carolina) (Zhang et al., 1998). The p21waf1/cip1 6KR mutant was generously provided by Bruce E.Clurman (Fred Hutchinson Cancer Research Center) (Sheaff et al., 2000). GST–p21waf1/cip1N and GST–p21waf1/cip1C were generously provided by Anindya Dutta (Harvard Medical School) (Chen et al., 1995).
Transient transfection and WB analysis
H1299 cells, or p53−/−/mdm2−/− or p53−/−/mdmx−/− MEFs (60% confluence in a 60 mm plate) were transfected with pCEP4‐p21waf1/cip1 (1 μg) alone or together with pCDNA3‐HA‐MDM2 or deletion mutant MDM2 (see figure legends for the amount of plasmids used). At 48 h post‐transfection, cells were harvested for preparation of whole‐cell lysates. Whole‐cell lysates containing 50 or 100 μg of protein were loaded directly onto an SDS–gel and proteins were detected by enhanced chemiluminescence (ECL) reagents (Bio‐Rad) after WB using antibodies, as indicated in the figure kegends.
NB analysis was conducted as described (Zeng et al., 1999). Total RNA was isolated from transfected H1299 cells using the Trizol reagent (Life Technologies, Inc.). A 15 μg aliquot of RNA was loaded onto a 1.5% agarose gel and transferred to a nitrocellulose membrane. The membrane after UV cross‐linking was incubated with 32P‐labeled cDNA probes encoding human p21waf1/cip1 at 42°C overnight. After washing with 4× SSC once and 1× SSC twice, the blot was exposed to X‐ray film overnight.
Analysis of p21waf1/cip1's half‐life in cells
H1299 cells were transfected with the p21waf1/cip1 plasmid alone or together with the MDM2 or 1–441 MDM2 expression plasmid as described above. At 48 h after transfection, transiently transfected H1299 cells in 100 mm plates were labeled with 100 μCi/ml Easy Expression (PerkinElmer Life sciences) in 4 ml of Dulbecco's modified Eagle's medium (DMEM) with 2% dialyzed methionine‐free calf serum for 45 min at 37°C. Cells were then washed with phosphate‐buffered saline (PBS), incubated in DMEM containing10% fetal bovine serum (FBS) and harvested at different time points. Cell lysates were prepared for IP as described above. The level of p21waf1/cip1 from the transfected cells was quantified by scanning the blot and using the Adobe Photoshop program, and plotted using the Cricket Graph (Figure 4A). These experiments were repeated twice.
In vivo ubiquitylation assay
The in vivo ubiquitylation assay was conducted as described previously (Xirodimas et al., 2001). H1299 cells or p53−/−/mdm2−/− MEFs in 100 mm plates were transfected with His6‐ubiquitin (2 μg), wild‐type p21waf1/cip1 (2 μg) or HA‐MDM2 (2 μg) expression plasmids using the lipofectAMINE reagent (Invitrogen). At 48 h after transfection, cells from each plate were harvested and split into two aliquots, one for straight WB and the other for detection of ubiquitylated proteins using Ni‐NTA beads (Qiagen). The eluted proteins from the beads were analyzed by WB for polyubiquitylation of p21waf1/cip1 with monoclonal p21waf1/cip1 antibodies (NeoMarker, Ab11).
Immunofluorescent staining and fluorescent microscopic analysis
H1299 cells were co‐transfected with the plasmid encoding p21waf1/cip1 alone, or together with the MDM2‐, 1–441 MDM2 or Δ150–230 MDM2 expression plasmids. At 48 h after transfection, cells were fixed for immunofluorescent staining with monoclonal anti‐p21waf1/cip1 antibodies and polyclonal anti‐MDM2 antibodies, as well as for DNA staining with 4′,6‐diamidino‐2‐penylindole (DAPI). The Alexa Fluor 488 (green) goat anti‐mouse antibody and the Alexa Fluor 546 (red) goat anti‐rabbit antibody (Molecular Probes, OR) were used for p21waf1/cip1 and MDM2, respectively. Stained cells were analyzed under an Zeiss Axiovert 25 fluorescent microscope.
GST fusion protein association assay
The fusion proteins were expressed in Escherichia coli and purified on a glutathione–Sepharose 12B column. Protein–protein association assays were conducted as reported using fusion protein‐containing beads (Zeng et al., 1999). Purified p21waf1/cip1 was incubated with the glutathione–Sepharose 4B beads (50% slurry) containing ∼500 ng of GST–MDM2, its deletion mutants as shown in Figure 4B, or GST, respectively. For a reverse GST pull‐down assay, GST–p21waf1/cip1, GST–p21waf1/cip1N, GST–p21waf1/cip1C and purified MDM2 were used. At 1 h after incubation at room temperature, the mixtures were washed intesively. Bound proteins were analyzed on a 12% SDS–gel and detected by WB.
H1299 cells were transfected with an empty plasmid or the p21waf1/cip1 plasmid or with p21waf1/cip1 and MDM2. At 40 h post‐transfection and 16 h after nocodazole treatment (0.4 μM), cells were harvested and re‐suspended in 100 μl of PBS, and transferred to a polystyrene tube. Into the tube, 200 μl of pH 7.2 propidium iodide (PI) stain (50 μl/ml PI, 30 μg/ml polyethylene glycol 8000, 2 μg/ml RNase A, 0.1% Triton X‐100, 0.38 M NaCl) was added. Samples were incubated for 10 min and analyzed for DNA content using a Becton Dickinson FACScan flow cytometer. Data were analyzed by the multicycle software program using the polynomial S‐phase algorithm.
Total RNA was isolated using TRIzol reagent (Invitrogen Corp, Carlsbad, CA) from p53−/− and p53−/−/mdm2−/− MEFs after being treated or not with MG132. RT–PCRs were conducted as previously desribed (Zeng et al., 2002). PCR products were analyzed on a 4.5% polyacrylamide gel followed by autoradiography, or on an agarose gel followed by ethidium bromide staining. The following primers were used: human p21waf1/cip1, 5′‐ATGTCAGAACCGGCTGGGGATG‐3′, 5′‐TTAGGGCTTCCTCTT GGAGAAG‐3′; human GAPDH, 5′‐TCTAGACGGCAGGTCAGGTCC ACC‐3′, 5′‐CCACCCATGGCAAATTCCATGGCA‐3′; mouse p21waf1/cip1, 5′‐ATGTCCAATCCTGGTGATGTCC‐3′, 5′‐TCAGGGTTTTCTCTTG CAGAAG‐3′; mouse GAPDH, 5′‐ATGGAGAAGG‐CCGGG‐GCCCA CTT‐3′, 5′‐TTACTCCTTGGAGGCCATGTA‐3′.
SiRNA design and transfection
A twenty‐one nucleotide RNA targeting human MDM2 mRNA 5′‐AAG CCA UUG CUU UUG AAG UUA‐3′ and a scrambled small RNA were chemically synthesized and supplied in the 2′‐deprotected, annealed and desalted form by Dharmacon (Lafayette, CO). The sequence of the siRNA duplex was designed according to the manufacturer's recommendations and subjected to a BLAST search against the human genome sequence to ensure that endogenous non‐mdm2 genes of the genome were targeted.
H1299 cells were split 16 h before transfection at 70% confluency. Oligofectamine (Invitrogen)‐mediated transient transfection of siRNA was performed in 60 mm plates. siRNA (100 or 300 nM) and 9 μl of oligofectamine were used for each plate in serum‐free DMEM. Cells were incubated in transfection mixture for 5 h and further cultured in DMEM containing 10% FBS. The second siRNA transfection was carried out on the next day. Scrambled siRNAs were used as a control.
We thank David M.Keller and Jayme Gallegos for proofreading this manuscript. We also thank Guillinao Lozano, Wafik S.el‐Deiry, Jiandong Chen, Charles J.Sherr, Yue Xiong, Bruce E.Clurman and Anindya Dutta for generously providing us with cell lines, antibodies and plasmids. This work is supported by grants to H.L. from the NIH (CA095441, CA93614 and CA079721).
- Copyright © 2003 European Molecular Biology Organization